POTENTIAL IMPACTS OF CLIMATE CHANGE IN
DRY COASTAL ECOSYSTEMS OF BRITISH COLUMBIA

by

Heather Klassen
B.Sc., Royal Roads University, 2004

THESIS SUBMITTED IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
IN
NATURAL RESOURCES AND ENVIRONMENTAL STUDIES

UNIVERSITY OF NORTHERN BRITISH COLUMBIA
April 2012

© Heather Klassen, 2012

1+1

Library and Archives
Canada

Bibliotheque et
Archives Canada

Published Heritage
Branch

Direction du
Patrimoine de I'edition

395 Wellington Street
Ottawa ON K1A0N4
Canada

395, rue Wellington
Ottawa ON K1A 0N4
Canada
Your file Votre reference
ISBN:

978-0-494-94160-7

Our file Notre reference
ISBN:

978-0-494-94160-7

NOTICE:

AVIS:

The author has granted a non­
exclusive license allowing Library and
Archives Canada to reproduce,
publish, archive, preserve, conserve,
communicate to the public by
telecommunication or on the Internet,
loan, distrbute and sell theses
worldwide, for commercial or non­
commercial purposes, in microform,
paper, electronic and/or any other
formats.

L'auteur a accorde une licence non exclusive
permettant a la Bibliotheque et Archives
Canada de reproduire, publier, archiver,
sauvegarder, conserver, transmettre au public
par telecommunication ou par I'lnternet, preter,
distribuer et vendre des theses partout dans le
monde, a des fins commerciales ou autres, sur
support microforme, papier, electronique et/ou
autres formats.

The author retains copyright
ownership and moral rights in this
thesis. Neither the thesis nor
substantial extracts from it may be
printed or otherwise reproduced
without the author's permission.

L'auteur conserve la propriete du droit d'auteur
et des droits moraux qui protege cette these. Ni
la these ni des extraits substantiels de celle-ci
ne doivent etre imprimes ou autrement
reproduits sans son autorisation.

In compliance with the Canadian
Privacy Act some supporting forms
may have been removed from this
thesis.

Conform em ent a la loi canadienne sur la
protection de la vie privee, quelques
formulaires secondaires ont ete enleves de
cette these.

W hile these forms may be included
in the document page count, their
removal does not represent any loss
of content from the thesis.

Bien que ces formulaires aient inclus dans
la pagination, il n'y aura aucun contenu
manquant.

Canada

Abstract
Climate change is impacting forest ecosystems. Climatic envelopes were developed for
dry coastal ecosystems and 18 diagnostic plant species in southwestern British Columbia to
project current and future suitable climate space. Future projections suggest a northward
shift for ecosystem and species, with a reduction in ecosystem climate space and variable
results for species climate space. Results suggest that ecosystem climatic envelopes
represent cumulative biological complexity and that the ecosystem-level processes and
functions cannot be allocated among the species within the plant community. A monitoring
network was established to improve understanding and to detect changes in climate, soil,
and vegetation relationships, and hence the distribution of ecosystems and species, over
time. Baseline summaries detect climatic differences between monitored ecosystems. This
climatic envelope research provides a foundation for theoretical development and the field
study provides site-specific datasets to improve our understanding of forest ecosystems and
our ability to manage land and resources.

Table of Contents
Abstract............................................................................................................................................ ii
Table of Contents.......................................................................................................................... ill
List of Tables.................................................................................................................................. vii
List of Figures.................................................................................................................................. x
Acknowledgements......................................................................................................................xv
1.

Background and General Introduction..............................................................................1

1.1
1.2
1.3
2.

Research Purpose and Study Area.................................................................... 1
The Climatic Basis for Ecosystem Classification............................................... 2
Thesis Overview................................................................................................7

Ecosystem-Based Climatic Envelope Models: Current and Future Distribution of Dry
Coastal Ecosystems.............................................................................................................8

2.1
ABSTRACT.........................................................................................................8
2.2
INTRODUCTION................................................................................................. 8
2.2.1 Climate Change and Impacts to Ecosystems....................................................8
2.2.2 Global Climate Models and Interpolation Tools............................................. 10
2.2.3 Ecosystem Classification and Ecosystem-Based Climatic Envelopes.............. 13
2.2.4 Research Objectives........................................................................................ 17
2.3
METHODS........................................................................................................ 18
2.3.1 Study Area.......................................................................................................18
2.3.2 Data and Analysis............................................................................................21
2.3.2.1 Global climate model ensemble analysis and selection......................... 23
2.3.2.2 Ecosystem classification.......................................................................... 26
2.3.2.3 Ecosystem-based climatic envelopes: baseline andfuture distribution 28
2.3.2.4 Areas of ecosystem overlap, persistence, and shift in southwestern
British Columbia.................................................................................... 31
2.3.2.5 Constraints on extended CDF baseline suitable climate space on
Vancouver Island................................................................................... 32
2.4
RESULTS.......................................................................................................... 34
2.4.1 Global Climate Model Ensemble Analysis and Selection................................ 34
2.4.2 Ecosystem Classification.................................................................................38
2.4.3 Ecosystem-Based Climatic Envelopes: baseline andfuture distribution
41
2.4.4 Areas of Ecosystem Overlap, Persistence, and Shiftin Southwestern British
Columbia....................................................................................................... 47
2.4.5 Constraints on the Extended CDF Baseline Suitable ClimateSpace on
Vancouver Island.......................................................................................... 54
2.5
DISCUSSION.................................................................................................... 58
2.5.1 Global Climate Model Selection and Interpretation....................................... 58
2.5.2 The Ecosystem-Based Climatic Envelopes......................................................60
2.5.3 Areas of Overlap, Persistence, and Shift of Suitable Climate Space............. 65
2.5.4 Constraints on the Extended CDF Climatic Envelope.....................................67
iii

2.5.5
2.6
3.

Limitations of the Climatic Envelope Technique........................................... 68
CONCLUSIONS AND RECOMMENDATIONS................................................... 69

Species Climatic Envelope Models: Current and Future Distribution of 18 Species
Common to the Coastal Douglas-fir zone...................................................................... 71

3.1
ABSTRACT..................................................................................................... 71
3.2
INTRODUCTION.............................................................................................. 72
3.2.1 Climate and Vegetation Relationships........................................................... 72
3.2.1.1 Individual species and plant assemblages...............................................72
3.2.1.2 Species distribution, climate change, and model projections.................74
3.2.2 Research Objectives....................................................................................... 77
3.3
METHODS......................................................................................................78
3.3.1
Study Area..................................................................................................... 78
3.3.2 Data and Analysis........................................................................................... 80
3.3.2.1 Species selection and location data........................................................ 80
3.3.2.2 Species baseline and future climatic envelope models.......................... 83
3.3.2.3 Sensitivity analysis for the baseline species climatic envelopes............. 84
3.3.2.4 Comparison of the CDFmm ecosystem mapping to modelled species
overlap and CDF climatic envelopes in the British Columbia study area
............................................................................................................... 85
3.4
RESULTS..........................................................................................................85
3.4.1 Species Climatic Envelopes: Current and Future Distributions..................... 85
3.4.2 Sensitivity Analysis for the Baseline Species Climatic Envelopes.................. 99
3.4.3 Comparison of the CDFmm Ecosystem Mapping to Modelled Species
Overlap and CDF Climatic Envelopes in the British Columbia Study Area ..103
3.5
DISCUSSION..................................................................................................105
3.5.1 Species Climatic Envelopes.......................................................................... 105
3.5.2 Sensitivity Analysis....................................................................................... 110
3.5.3 CDFmm Ecosystem Mapping, Individual Species Overlap,and CDF Climatic
Envelopes.................................................................................................... 112
3.5.4 Limitations of the Climatic Envelope Technique..........................................113
3.
6
CONCLUSIONS AND RECOMMENDATIONS.........................................115
4.

Site Descriptions and Preliminary Microclimate Differences in a Long-Term
Monitoring Network Established in the Transition between Coastal Douglas-fir and
Coastal Western Hemlock Forests............................................................................... 117

4.1
4.2
4.2.1
4.2.2
4.2.3
4.2.4
4.3
4.3.1

ABSTRACT.....................................................................................................117
INTRODUCTION.............................................................................................118
Vegetation and Microclimate...................................................................... 118
Boundary Layer Climate................................................................................120
Soil Water Balance....................................................................................... 122
Research Objectives......................................................................................130
METHODS......................................................................................................132
The Vancouver Island Study Area................................................................ 132
iv

4.3.2
Study Site Selection......................................................................................133
4.3.3 Field Sampling Design..................................................................................134
4.3.4 Field Sampling Equipment and Data Collection........................................... 137
4.3.4.1 Dataloggers........................................................................................... 137
4.3.4.2 Soil moisture.......................................................................................... 139
4.3.4.3 Soil and air temperature........................................................................ 141
4.3.4.4 Additional microclimate measures on CDFmm and CWHxm mesic sites
............................................................................................................. 144
4.3.5 Data and Analysis......................................................................................... 148
4.3.5.1 Study site description............................................................................. 148
4.3.5.2 Microclimate data analysis.................................................................... 150
4.4
RESULTS....................................................................................................... 153
4.4.1
Coastal Douglas-fir moist maritime (CDFmm) and Coastal Western Hemlock
very dry maritime (CWHxm) Mesic Site Descriptions.................................153
4.4.2
Baseline Microclimate Summaries for September 15, 2010 - September 14,
2011............................................................................................................ 158
4.4.3 Growing Season Microclimate Summaries: Comparisons of the Coastal
Douglas-fir moist maritime (CDFmm) and Coastal Western Hemlock very dry
maritime (CWHxm) subzones......................................................................167
4.5
DISCUSSION AND APPLICATIONS..................................................................171
4.5.1 Microclimate Data........................................................................................ 171
4.5.2 Research and Management Applications.................................................... 174
4.5.3 Long-term Data Collection and Monitoring Plan......................................... 175
4.6
CONCLUSIONS.............................................................................................. 177
5.

Current and Future Biogeography of the Coastal Douglas-fir Zone on Vancouver
Island: Synthesizing Climatic Envelope Models and Microclimate Field Study
Findings........................................................................................................................... 179

5.1
5.2
5.2.1
5.2.2
5.2.3
5.2.4
5.2.5
5.3
5.3.1
5.3.2
5.4
6.

INTRODUCTION............................................................................................ 179
SYNTHESIS DISCUSSION............................................................................... 181
Extending the Definitions of the Coastal Douglas-fir Biogeoclimatic Zone. 181
Climatic Envelope Projections for the Study Area....................................... 182
Some Insights into Plant Community Organization and Distribution.......... 185
Climatic Constraints on Ecosystem and Species Distributions.................... 187
Establishment of a Forest Microclimate Monitoring Network.................... 189
APPLICATIONS.............................................................................................. 191
Management Applications........................................................................... 191
Research Applications.................................................................................. 195
CONCLUSIONS.............................................................................................. 197

References....................................................................................................................... 200

v

Appendix A: Global climate models (GCMs) and carbon emission scenarios available for
future projections using the ClimateWNA interpolation tool (ClimateWNA v4.52 2010).
.......................................................................................................................................... 215
Appendix B: Annual, seasonal, and monthly climate variables available from ClimateWNA
for current and future projections (ClimateWNA v4.52 2010). Annual climate variables
used for climatic envelope definitions in bold.................................................................. 216
Appendix C: Total area (km2) and climate summary statistics (mean, standard deviation (SD),
minimum, and maximum) for baseline and future extended Coastal Douglas-fir (CDF)
suitable climate space (SCS) in southwestern British Columbia. The climatic envelope
definition values (minimum and maximum) appear in parentheses below the climate
variable..............................................................................................................................217
Appendix D: Paired t-test results summary table for the 5-km buffer within and outside the
extended Coastal Douglas-fir (CDF) baseline suitable climate space (SCS) on Vancouver
Island. Climate variables ranked from most significant (highest t-value) to least significant
difference between the means. Sample number includes 30 paired random samples;
latitude (°N) is the mid-point of the sample window........................................................219
Appendix E: Summary statistics (mean and standard deviation, SD, n= 298 077) for each of
the five global climate models (GCM) used for the five-model GCM ensemble projecting
climate for three future timeslices (2010-2039, 2040-2069,2070-2099) for the Pacific
Northwest (southwest British Columbia, western Washington, and Oregon)................. 243
Appendix F: Scientific and common names for 18 species common to the Coastal Douglas-fir
moist maritime (CDFmm) subzone listed by tree, shrub, herb, and moss strata............. 245
Appendix G: Current and future species suitable climate space (SCS) area (km2) for the full
extent (minimum-maximum), the confidence percentiles ( l st-99th and 5th-95th), and core
percentiles (25th-75th) for the southwestern British Columbia study area....................... 246
Appendix H: Seasonal (Fall, September 15 - December 14; Winter, December 15 - March
14; Spring, March 15 - June 14; and Summer, June 15 - September 14) microclimate
summary for the Coastal Douglas-fir moist maritime (CDFmm) and Coastal Western
Hemlock very dry maritime (CWHxm) mesic study sites on the south, central, and north
transects (n=3) for the baseline year (September 15, 2010 - September 14, 2011) with
the mean, standard error (S.E.), minimum (Min.) and maximum (Max.) values and the
results of the paired t-test (df=2, p<0.05).........................................................................248
Appendix I: Monthly microclimate summary for the Coastal Douglas-fir moist maritime
(CDFmm) and Coastal Western Hemlock very dry maritime (CWHxm) mesic study sites on
the south, central, and north transects (n=3) for the baseline year (September 15,2010 September 14, 2011) with the mean, standard error (S.E.), minimum (Min.) and maximum
(Max.) values and the results of the paired t-test (df=2, p<0.05)..................................... 249

List of Tables
Table 2-1: Baseline climate normals (1961-1990) for Coastal Douglas-fir moist maritime
(CDFmm) and Coastal Western Hemlock dry maritime (CWHxm) biogeoclimatic
subzones in British Columbia. The mean and standard deviation were calculated by
overlaying a 1-km grid over BEC v.7 subzone mapping (BC MFR 2008) and using
ClimateWNA (Wang et al. 2010) to output climate variables*................................... 20
Table 2-2: Total area (km2) with private and public* land percentages (%) for Coastal
Douglas-fir moist maritime (CDFmm) and Coastal Western Hemlock very dry
maritime (CWHxm) subzones (BC LRDW 2011).......................................................... 21
Table 2-3: Percent cover rules and rationale for filtering species (common and scientific
names) to identify plots from the Pacific Northwest United States ecological plot
database that can be considered equivalent to the Coastal Douglas-fir forest zone. 28
Table 2-4: Mean annual air temperature (°C) and mean daily precipitation (mm/day) for the
study area (as defined in Figure 2-4) as derived from baseline global climate model
(GCM) climate normals from Pacific Climate Impacts Consortium (PCIC) Regional
Analysis Tool (RAT) and interpolated local climate normals from ClimateWNA
34
Table 2-5: Summary table of the relationship between the observed (climate station) and
interpolated (ClimateWNA) mean monthly temperatures (°C) and total monthly
precipitation (mm) at three meteorological stations distributed latitudinally within
the study area (Victoria, Capital Regional District, Cedar).......................................... 38
Table 2-6: Summary of minimum (Min.) and maximum (Max.) climate values defining the
climatic envelopes for the extended Coastal Douglas-fir (CDF) and Coastal Western
Hemlock very dry maritime (CWHxm) biogeoclimatic units....................................... 41
Table 2-7: The total area (km2) for minimum-maximum and percentile projections of
suitable climate space (SCS) for the extended Coastal Douglas-fir (CDF) and Coastal
Western Hemlock very dry maritime (CWHxm) ecosystems baseline climate normals
(1961-1990) using ClimateWNA interpolated climate data........................................ 44
Table 2-8: Total area (km2) in the minimum-maximum climatic envelopes for the extended
Coastal Douglas-fir (CDF) and Coastal Western Hemlock very dry maritime (CWHxm)
using 16 annual climate variables from ClimateWNA over a 1-km grid of
southwestern British Columbia, western Washington, and western Oregon............ 47
Table 2-9: Area of overlap (km2), total suitable climate space (SCS) area (km2), and
percentage SCS within the overlap area (%) for the extended Coastal Douglas-fir
(CDF) and Coastal Western Hemlock very dry maritime (CWHxm) suitable climate
space (SCS) in southwestern British Columbia for baseline to future timeslices
(baseline to 2010 - 2039, baseline to 2040 - 2069, baseline to 2070 - 2099) using
the five-model global climate model (GCM) ensemble and minimum-maximum
climatic envelope........................................................................................................ 48
Table 2-10: Areas of persistence (km2) in southwestern British Columbia as projected for
baseline to future timeslices (baseline to 2010 - 2039, baseline to 2040 - 2069,
baseline to 2070 - 2099) in extended Coastal Douglas-fir (CDF) and Coastal Western
Hemlock very dry maritime (CWHxm) suitable climate space (SCS) using the five-

model global climate model (GCM) ensemble and minimum-maximum climatic
envelope...................................................................................................................... 50
Table 2-11: Area (km2) projected to shift from Coastal Western Hemlock very dry maritime
(CWHxm) suitable climate space (SCS) into extended Coastal Douglas-fir (CDF) and
out of the extended CDF into undefined SCS in southwestern BC over time (a)
Baseline to the 2010 - 2039 timeslice; (b) 2010-2039 to 2040-2069 timeslice; (c)
2040-2069 to 2070-2099 timeslice), using the five-model global climate model (GCM)
ensemble and minimum-maximum climatic envelopes............................................. 52
Table 2-12: Comparing the extended Coastal Douglas-fir (CDF) to the undefined zone in
future timeslices using the minimum (Min.) and maximum (Max.) values for climate
variables of points projected to shift from the extended CDF to undefined suitable
climate space (SCS) in southwestern British Columbia over time (Baseline to 20102039; 2010-2039 to 2040-2069; and 2040-2069 to 2070-2099 timeslices), climatic
envelopes using the five-model global climate model (GCM) ensemble and
minimum-maximum climatic envelope definitions. Variables out of CDF baseline
climatic envelope range are highlighted in bold; BLUE = low, RED= high.................. 54
Table 3-1: Summary of vegetation (scientific and common name listed by tree, shrub, herb,
and moss strata) that characterizes the zonal sites of the Coastal Douglas-fir moist
maritime (CDFmm) and the adjacent Coastal Western Hemlock very dry maritime
(CWHxm) subzones*. Prominence bars measure the combination of cover and
frequency of occurrence............................................................................................. 81
Table 3-2: The climatic envelope definitions using the minimum (min) and maximum (max)
values for the 16 annual climate variables* for the 18 diagnostic species of zonal
sites in the Coastal Douglas-fir moist maritime subzone............................................ 87
Table 3-3: Shift in species suitable climate space (SCS) over four timeslices within the
current mapped Coastal Douglas-fir moist maritime (CDFmm) (BEC v.7) in the British
Columbia study area using the minimum - maximum climatic envelope definitions
and the five-model global climate model (GCM) ensemble (CDFmm total area in BC
study area = 5 631 km2).............................................................................................. 98
Table 3-4: Change in area (km2) of baseline minimum-maximum suitable climate space (SCS)
for each species and the species overlap area with ± adjustment to each climate
variable (one at a time) by 10% of the minimum - maximum climate variable range
within the British Columbia study area..................................................................... 100
Table 4-1: Potential impacts to components of the soil water balance with conditions of
related components, if applicable (Varallyay 1990).................................................. 128
Table 4-2: Summary of the HOBO® microclimate sensors and their technical
specifications............................................................................................................. 148
Table 4-3: Vegetation strata*, species, frequency (percentage occurrence in n=3 plots), and
mean percent cover (± standard error (S.E.)) with paired t-test results (df = 2, p <
0.05) for plant species found in the Coastal Douglas-fir moist maritime (CDFmm) and
Coastal Western Hemlock very dry maritime (CWHxm) mesic study sites on the
south, central, and north transects. Significant results in bold.................................154
Table 4-4: Mean and standard error (S.E.) canopy gap fraction, openness, and leaf area
index (LAI) values calculated using hemispherical photos taken directly above the
viii

three soil moisture subplots for each of the three study sites in the Coastal Douglasfir moist maritime (CDFmm) and Coastal Western Hemlock very dry maritime
(CWHxm)....................................................................................................................156
Table 4-5: Site and soil characteristics for each of the Coastal Douglas-fir moist maritime
(CDFmm) and Coastal Western Hemlock very dry maritime (CWHxm) study sites on
the south, central, and north transects.....................................................................157
Table 4-6: Annual microclimate summary and paired t-test (df = 2, p < 0.05) results for the
Coastal Douglas-fir moist maritime (CDFmm, n=3) and Coastal Western Hemlock very
dry maritime (CWHxm, n=3) mesic study sites on the south, central, and north
transects for the baseline year (September 15, 2010 - September 14, 2011) with the
mean, standard error (S.E.), minimum (Min.) and maximum (Max.) values and the
results of the paired t-test (df=2, p<0.05). Significant differences in bold............... 159
Table 4-7: Growing season (May 15 - September 14, 2011) mean and standard error (S.E.)
and paired t-test results testing for a difference in the mean for each microclimate
variable (df =2, p < 0.05) in the Coastal Douglas-fir moist maritime (CDFmm,n=3) and
Coastal Western Hemlock very dry maritime (CWHxm, n=3) mesic study sites.
Significant results in bold...........................................................................................167
Table 4-8: Comparisons of the extremes (minimum and maximum) for the mean daily
growing season microclimate attributes in the Coastal Douglas-fir moist maritime
(CDFmm, n=3) and Coastal Western Hemlock very dry maritime (CWHxm, n=3) mesic
study sites, between May 15, 2011 and September 14,2011. N/A values indicate
multiple occurrences of minimum or maximum values............................................168
Table 4-9: Short-term and long-term objectives for the climate change monitoring network
established in the Coastal Douglas-fir moist maritime (CDFmm) and Coastal Western
Hemlock very dry maritime (CWHxm) study area on southeastern Vancouver Island,
British Columbia........................................................................................................ 176
Table 5-1: Global climate model (GCM) five-model ensemble mean annual climate variables
and the overall change in climate variables for the baseline (1961-1990) and
projected for future (2010-2039, 2040-2069, and 2070-2099) timeslices for the
southwestern British Columbia study area............................................................... 183

ix

List of Figures
Figure 2-1: The geographic area covered by the ClimateWNA interpolation tool (from
http://www.genetics.forestry.ubc.ca/cfcg/ClimateWNA/ClimateWNA.html)
13
Figure 2-2: Relationship between static climatic envelope models and other factors
contributing to dynamic models. (Adapted from Guisan and Thuiller 2005, Franklin
2009)............................................................................................................................16
Figure 2-3. Current distribution of the Coastal Douglas-fir moist maritime (CDFmm) and the
Coastal Western Hemlock very dry maritime (CWHxm) biogeoclimatic subzones (BEC
v.7) on southeastern Vancouver Island, the Gulf Islands, and adjacent mainland
coast, British Columbia................................................................................................ 19
Figure 2-4: The Vancouver Island and adjacent mainland British Columbia "custom area"
used in Pacific Climate Impacts Consortium (PCIC) Regional Analysis Tool (RAT) for
global climate model (GCM) climate data baseline outputs
(http://tools.pacificclimate.org/select); inset diagram represents a portion of the
ClimateWNA 1-km point grid locations used to output baseline interpolated climate
data over the entire custom area for GCM selection. Note that grid cell size and
location vary for each GCM output from the PCIC RAT, but all overlay the custom
area..............................................................................................................................24
Figure 2-5: The 5-km buffer within and outside of the extended Coastal Douglas-fir (CDF)
suitable climate space (SCS) for the baseline timeslice (1961-1990) with inset figure
of the East-West transects systematically bisecting the full extent of the extended
CDF boundary on Vancouver Island, British Columbia................................................ 33
Figure 2-6: Scatter plot of the standardized mean annual temperature (MAT, °C) and mean
daily precipitation (MDP, mm) values for the global climate model (GCM) carbon
emission scenarios considered for five-model GCM ensemble based on the
availability from ClimateWNA to project future timeslices. Note 1. The Ensemble
GCM scenarios circled in red and the ClimateWNA value as red "x". [GCM data
source = Pacific Climate Impacts Consortium; interpolated climate station data
source = ClimateWNA]. Note 2. The Hadley CM3 A2 and B2 and the UKMO Had CM3
A2 GCM values are very close, so only three red ensemble circles appear on the
figure........................................................................................................................... 35
Figure 2-7: Comparison of mean monthly temperatures (MMT, °C) for observed values
recorded by 3 meteorological stations (Victoria, Capital Regional District (CRD) 14G,
and Cedar) and interpolated values calculated by ClimateWNA for the geo­
referenced locations. Solid line is the linear regression for each data series............. 36
Figure 2-8: Comparison of means for total monthly precipitation (TMP, mm) for observed
values recorded by 3 meteorological stations (Victoria, Capital Regional District (CRD)
14G, and Cedar) and interpolated values calculated by ClimateWNA for the geo­
referenced locations. Solid line is the linear regression for each data series............. 37
Figure 2-9: Modelled Potential Natural Vegetation Zones of Washington and Oregon
(Ecoshare 2011), showing the distribution of 96 United States ecological plots
classified as Coastal Douglas-fir (CDF) zone ecological units used in combination with

x

BEC v.7 mapping to create the climatic envelope definitions for the extended CDF
suitable climate space................................................................................................. 39
Figure 2-10: Projection of the suitable climate space (SCS) for the (a) extended Coastal
Douglas-fir (CDF) and (b) Coastal Western Hemlock very dry maritime (CWHxm)
zonal ecosystems for the baseline annual climate normals (1961-1990) using
ClimateWNA interpolated climate station data. Each map is created by overlaying
the minimum - maximum SCS, 1st - 99th, 5th - 95th, and 25th - 75th percentile SCS......43
Figure 2-11: Future distributions for the extended Coastal Douglas-fir (CDF) zonal ecosystem
climates, as projected for (a) 2010 - 2039, (b) 2040 - 2069, and (c) 2070 - 2099
timeslices, using the five-model GCM ensemble. Distributions for the minimum maximum, 1st - 99th, 5th - 95th, and 25th - 75th percentile climatic envelopes included
(when present) for each timeslice...............................................................................45
Figure 2-12: Future distributions for the Coastal Western Hemlock very dry maritime
(CWHxm) zonal ecosystem climates, as projected for (a) 2010 - 2039, (b) 20402069, and (c) 2070 - 2099 timeslices, using the five-model GCM ensemble.
Distributions for the minimum - maximum, 1st - 99th, 5th - 95th, and 25th - 75th
percentile climatic envelopes included (when present) for each timeslice................46
Figure 2-13: The extended Coastal Douglas-fir (CDF) and Coastal Western Hemlock very dry
maritime (CWHxm) suitable climate space (SCS) overlap area as projected over time
(a) Baseline; (b) 2010 - 2039; (c) 2040 - 2069; and (d) 2070 - 2099 timeslices, using
the five-model global climate model (GCM) ensemble and minimum-maximum
climatic envelopes.......................................................................................................49
Figure 2-14: Extended Coastal Douglas-fir (CDF) ecosystem persistent area over time (a)
Baseline to the 2010 - 2039 timeslice; (b) Baseline to the 2040-2069 timeslice; (c)
Baseline to the 2070-2099 timeslice, using the five-model global climate model
(GCM) ensemble and minimum - maximum climatic envelopes................................ 50
Figure 2-15: Coastal Western Hemlock very dry maritime (CWHxm) ecosystem persistent
area projected over time (a) Baseline to the 2010-2039 timeslice; (b) Baseline to
the 2040-2069 timeslice; (c) Baseline to the 2070-2099 timeslice), using the fivemodel global climate model (GCM) ensemble and minimum - maximum climatic
envelopes.................................................................................................................... 51
Figure 2-16: Area projected to shift from Coastal Western Hemlock very dry maritime
(CWHxm) suitable climate space (SCS) to the extended Coastal Douglas-fir (CDF) SCS
over time (a) Baseline to 2010-2039; (b) 2010-2039 to 2040-2069; and (c) 2040-2069
to 2070-2099 timeslices, using the five-model global climate model (GCM) ensemble
and minimum-maximumclimatic envelopes............................................................... 52
Figure 2-17: Areas projected to shift from the extended Coastal Douglas-fir (CDF) suitable
climate space (SCS) to undefined SCS overtime (a) Baseline to 2010-2039, (b) 20102039 to 2040-2069, and (c) 2040-2069 to 2070-2099 timeslices, using the five-model
global climate model (GCM) ensemble and minimum-maximum climatic
envelopes.................................................................................................................... 53
Figure 2-18: Smoothed t-value (a) increasing, (b)decreasing, and (c) neutral trends for the 5km buffer moving window paired t-test (n=30, 2-tailed, p<0.05) for each climate
variable within and outside of the extended Coastal Douglas-fir (CDF) baseline
xi

suitable climate space (SCS) on Vancouver Island, t-values for each climate variable
are smoothed using a 10-sample weighted average starting at the south end of the
boundary..................................................................................................................... 56
Figure 2-19: The three most significant climate variables* from the south to the north end
of the baseline extended Coastal Douglas-fir (CDF) minimum-maximum suitable
climate space (SCS) boundary on Vancouver Island. Climate variables are ranked by
highest (1-3; most significant) using a 10-sample average of t-value results from the
paired t-test (n=30, 2-tailed, p<0.05).......................................................................... 57
Figure 3-1: The southwest British Columbia (BC) study area (Vancouver Island, small islands
of the Salish Sea, and adjacent mainland BC coast) with the full Pacific Northwest
coast (west coast of Washington and Oregon and the BC study area) domain area
inset............................................................................................................................. 79
Figure 3-2: Map of western North America with location points for all 18 diagnostic species
of the Coastal Douglas-fir moist maritime subzone used as inputs to the climatic
envelope definitions (total number of plots > 33 000)............................................... 82
Figure 3-3: Percent of the British Columbia study area (total area = 97 857 km2) occupied by
the suitable climate space (SCS) for 18 plant species characteristic of the Coastal
Douglas-fir moist maritime (CDFmm) zonal site series as based on the minimum maximum climatic envelope, the 1st - 99th, 5th - 95th, and 25th - 75th percentiles,
projected using a five-model global climate model (GCM) ensemble. Species listed by
stratum: trees (5), shrubs (7), herbs (3), and bryophytes (3)......................................92
Figure 3-4: Mapped distributions of suitable climate space (SCS) using the minimum and
maximum values for 16 annual climate variables on the south coast of British
Columbia (with Pacific Northwest inset in upper right corner representing the
domain area for the analysis) for 18 diagnostic species of the Coastal Douglas-fir
moist maritime (CDFmm) zonal ecosystem for the baseline (1961-1990) and future
(2070-2099) time slices. Model projections for baseline conditions use ClimateWNA
baseline normals and future use the five-model global climate model (GCM)
ensemble..................................................................................................................... 97
Figure 3-5: The number of species whose range changes by more than 100 km2 with an (a)
increase and (b) decreased climate variable adjustments by 10% of the current range
of climate in the British Columbia study area for the baseline (1961-1990) timeslice.
A positive value (green bar) for the number of species indicates an increase in
species SCS range, while a negative value (red bar) for the number of species
indicates a decrease in species SCS range...................................................................99
Figure 3-6: Current mapped distribution of the Coastal Douglas-fir moist maritime (CDFmm)
in the British Columbia study area (BEC v.7, BC MFR 2008)......................................103
Figure 3-7: Comparison of the suitable climate space (SCS) for the (a) overlap area of 18
individual diagnostic species, and (b) extended Coastal Douglas-fir (CDF) SCS
(Chapter 2) for the baseline and three future timeslices in the British Columbia study
area using a five-model global climate model (GCM) ensemble...............................104
Figure 4-1: Soil moisture is an integration of climate, soil, and vegetation relations (adapted
from Rodriguez-lturbe and Porporato 2004)............................................................ 119

xii

Figure 4-2: Hydrological processes in the water balance model: water enters as
precipitation, moving through the system by interception, throughfall, infiltration
processes, and is lost from the system by evapotranspiration, surface runoff, and
change in water storage............................................................................................ 123
Figure 4-3: Biogeoclimatic Ecosystem Classification (BEC) subzone map with the south,
central, and north sampling transect locations on the southeastern end of Vancouver
Island, British Columbia............................................................................................. 135
Figure 4-4: Site layout with the three soil moisture and soil temperature subplots, the
Campbell Scientific CR10 data logger box, and air temperature sensors positioned
around the plot center...............................................................................................137
Figure 4-5: (a) Data logger box and air temperature sensors adjacent to the soil moisture
and soil temperature sensor at subplot 3 of the CDFmm hygric site on the central
transect and (b) the Campbell Scientific CR10 data logger and Panasonic 12 V battery
in the data logger box................................................................................................138
Figure 4-6: (a) Soil pit (35 cm depth) dug to insert the Decagon EC-5 soil moisture sensor and
(b) the soil moisture sensor installed into the rooting zone soil profile (25-30 cm
depth) at the CDFmm-CWHxm transition site on the north transect....................... 141
Figure 4-7: Thermocouple air temperature sensors with radiation shields at 0.3 m and 1.5 m
above the ground surface, mounted on a metal stake adjacent to the Campbell
Scientific CR10 data logger box and subplot 3 at the CDFmm mesic site on the south
transect...................................................................................................................... 142
Figure 4-8: Thermochron iButton® sensor (small metal disc) being inserted into rooting zone
soil horizon (ca. 27 cm depth) to measure soil temperature adjacent to the Decagon
EC-5 soil moisture sensor.......................................................................................... 143
Figure 4-9: HOBO® microclimate station at the south transect CDFmm mesic site, showing
the tripod and mast structure that house the wind speed, precipitation, air
temperature and relative humidity, and solar radiation sensors, and with a HOBO®
U30 data logger powered by a solar panel................................................................145
Figure 4-10: Mean seasonal air temperature (1.5 m and 0.3 m), soil moisture, and soil
temperature microclimate (with ±1 standard error bars) for the Coastal Douglas-fir
moist maritime (CDFmm) and Coastal Western Hemlock very dry maritime (CWHxm)
mesic sites (n=3) for the baseline year (September 15, 2010 - September 14, 2011).
* below the x-axis denotes significant difference between seasonal means (paired ttest, df=2, p < 0.05)....................................................................................................161
Figure 4-11: Mean seasonal total precipitation, relative humidity, solar radiation, and wind
speed microclimate (with ±1 standard error bars) for the Coastal Douglas-fir moist
maritime (CDFmm) and Coastal Western Hemlock very dry maritime (CWHxm) mesic
sites (n=3) for the baseline year (September 15, 2010 - September 14, 2011). *
below the x-axis denotes significant difference between seasonal means (paired ttest, df=2, p < 0.05)....................................................................................................162
Figure 4-12: Walter and Lieth climatic diagram for the CDFmm (n=3) and CWHxm (n=3)
mesic sites for the baseline year (September 15, 2010 - September 14, 2011). Blue =
total monthly precipitation (mm) and Red = mean monthly air temperature (°C). The
dry season (red dotted area) occurs when the precipitation curve undercuts the
xiii

temperature curve; the moist season (blue vertical lines) occurs when the
precipitation curve supersedes the temperature curve, while the solid blue area
denotes a wet period. The blue bars below the x-axis indicate potential frost
period.........................................................................................................................163
Figure 4-13: Time series of the mean daily soil moisture, soil temperature, precipitation, and
air temperature for the Coastal Douglas-fir moist maritime (CDFmm) mesic study
sites (n=3) for the baseline year (September 15, 2010 - September 14, 2011)....... 165
Figure 4-14: Time series of the mean daily soil moisture, soil temperature, precipitation, and
air temperature for the Coastal Western Hemlock very dry maritime (CWHxm) mesic
study sites (n=3) for the baseline year (September 15,2010 - September 14,
2011)
166
Figure 4-15: Soil moisture persistence in the growing season on the Coastal Douglas-fir
moist maritime (CDFmm) mesic sites (n=3). Blue dotted line is the 95% confidence
interval for a 0 (random) ACF score. The x-axis (time lag) denotes the number of days
for which the soil moisture persists.......................................................................... 170
Figure 4-16: Soil moisture persistence in the growing season on the Coastal Western
Hemlock very dry maritime (CWHxm) mesic sites (n=3). Blue dotted line is the 95%
confidence interval for a 0 (random) ACF score. The x-axis (time lag) is the number of
days for which the soil moisture persists.................................................................. 171

xiv

Acknowledgements
I would like to thank the BC Ministry of Environment, the Capital Regional District, and
the BC Ministry of Forests, Lands, and Natural Resource Operations (MFLNR, previously the
Ministry of Forests and Range) for climate data; the US Forest Service (particularly Cindy
McCain), Jim Thorne at the University of California Davis, and the US EcoShare Project for
US ecological plot data and mapping resources; the BC MFLNR Research Ecology program
for the Biogeoclimatic Ecosystem Classification database and mapping, field assistance, and
expertise on ecological classification techniques. Many thanks to Peter Ott, BC MFLNR for
statistical advice and Todd Davis, MFLNR for GIS support. Nancy-Anne Rose patiently
explained the climatic modelling approach and Tongli Wang, Trevor Murdock, and Dave
Spittlehouse responded to questions on climate change projections, climate data, and field
equipment.
Field site access and logistics were provided by BC Parks, Capital Regional District, BC
Timber Sales, and North Cowichan Regional District. Field equipment and sampling support
came from Jiquan Chen, University of Toledo, Ross Benton, Canadian Forest Service (CFS),
Elizabeth Campbell from MFLNR (CFS), and Bill Floyd and Melissa Todd, MFLNR. Radio
communication support was provided by the MFLNR Communication Team.
Funding for my graduate research was provided by BC PSA Pacific Leaders Scholarship,
the Future Forests Ecosystem Initiative (FFEI), and MFLNR and CFS research funds.
The support of my M.Sc. supervisor, Phil Burton, and committee members, Stephen Dery
and Scott Green, is gratefully acknowledged. Many thanks to my colleagues at the MFLNR
Coast Area office, particularly Sari Saunders, Andy MacKinnon, Marty Kranabetter, and Rod
Negrave, who provided valuable words of encouragement, field expertise, and research
support. Ongoing support from family and friends was invaluable and much appreciated,
especially Wes Klassen and Sarah West who not only gave endless words of encouragement
but also provided field assistance in a moment's notice.

xv

1.

Background and General Introduction

1.1

Research Purpose and Study Area

There is strong evidence that global climatic change will impact the current distribution
of ecosystems. North American seasonal climate is projected to warm in the next century
with increasing precipitation in the winter and spring and decreasing precipitation in the
summer (Christensen et al. 2007). The province of British Columbia (BC) as a whole is on an
upward warming trend. This increase in temperature aligns with global projections reported
in the 1990s using coupled ocean-atmosphere general circulation models (OAGCMs)
developed for climate change studies over a decade ago (Johns et al. 1997). Impacts to
ecosystems are projected to be most severe at high latitudes, high elevations, and for dry
ecosystems of the province (Hamann and Wang 2006, Campbell and Wang in press). The dry
coastal ecosystems investigated in this study are unique relative to the remainder of wet
coastal ecosystems in BC where there is little climate change impact projected (Hamann and
Wang 2006). The study ecosystem is characterized by a Mediterranean-type climate and
includes a variety of rare and endangered species (CDC 2009). Dominant species of the dry
coastal ecosystem such as coastal Douglas-fir (Pseudotsuga menziesii var. menziesii), grand
fir (Abies grandis), and western redcedar (Thuja plicata) are predicted to expand their
ranges by 100 km per decade (Hamann and Wang 2006, Wang et al. in press). Trends of
increasing temperatures and shifts in precipitation regimes suggest the potential for a wider
distribution of dry ecosystem types on the BC coast.
The study ecosystem is the Coastal Douglas-fir moist maritime (CDFmm) subzone. The
CDFmm is geographically limited to the southeastern side of Vancouver Island, small islands

1

of the Salish Sea, and a narrow strip of the adjacent mainland BC (Nuszdorfer et al. 1991).
Similar climate and vegetation communities are also located in the Puget Trough and San
Juan Islands of Washington, and the Willamette Valley of Oregon in the United States (US,
Franklin and Dyrness 1988). The potential impacts of climate change in the CDFmm are
higher than for other low-elevation coastal forest ecosystems because the CDF has warm,
dry summers and mild, wet winters; it is drier and milder than other coastal ecosystems of
BC. To explore potential impacts of changes in climate in this dry coastal ecosystem of BC,
the published Biogeoclimatic Ecosystem Classification (BEC) system vegetation hierarchy
was used to select plant communities and individual species that commonly occur within a
homogeneous climatic unit on the coast (Green and Klinka 1994).

1.2

The Climatic Basis for Ecosystem Classification

The natural environment is complex and dynamic. The discipline of ecology studies the
distribution, abundance, and productivity of living organisms and their interactions with
each other and their physical environment (Kimmins 1987). Research involving both the
biotic and abiotic environment is specific to the realm of "ecosystem ecology". An
ecosystem is defined by Odum (1971) as a unit that includes all organisms in a given
geographic area interacting with the physical environment so that the exchange of materials
between living and non-living parts within the system can be defined. Ecosystems are
complex in that they can be defined at any spatial scale and they change over time,
resulting in a system where it is difficult to define boundaries.

2

Ecosystem classification (e.g., by physiognomy, function, and/or structure) provides
order to the matrix of physiographic and vegetation patterns across the landscape
(Shimwell 1971). Forest ecosystems have been defined and classified for a variety of
conservation and forest management objectives, including, but not limited to: timber sales,
forest age inventories, species inventories, wildlife management, recreation opportunities,
and non-timber forest product harvesting. The objective of the classification determines the
components of the ecosystem used to create the classification structure. The best approach
to classification is one that is user-friendly, responds to current needs, and meets the
objective of the user (Kimmins 1987). Three general approaches to ecosystem classification
focus on the physical environment (climate- Koppen 1923; landform- Burger 1972, Wertz
and Arnold 1975), biological environment (vegetation- Daubenmire 1976, Whittaker 1973,
Mueller-Dombois and Ellenberg 1974), or both the physical and biological components
(Meidinger and Pojar 1991, Willoughby et al. 2005, GO MNR 2007).
Ecosystem classification systems have evolved over the last century. Globally, the original
biome-level classification systems were developed at large spatial scales as representations
of vegetation distribution driven by regional climate (Woodward 1987, Pearson et al. 2002).
Forest zones were being classified in BC and the Pacific Northwest (PNW) US in the early
twentieth century (Meidinger and Pojar 1991). Canadian forests were first classified by B.E.
Fernow into units that represented climate regions (Fernow 1912). Since then, studies of
ecosystem structure and function have led to development of numerous classification and
mapping systems based on multi-scale hierarchical classifications. British Columbia forests
have been classified into forest types based on vegetation structure and composition,

3

climate, soil conditions, and physiography, ultimately leading to the development of the
current BEC system (Meidinger and Pojar 1991).
Alberta uses a similar approach to ecosystem classification, particularly in the
mountainous terrain of western Alberta (Willoughby et al. 2005). Other Canadian provinces,
where elevational diversity and climatic gradients are less extreme, have applied different
approaches to ecosystem classification and mapping. For example, the provincial ecological
land classification system in Ontario uses bedrock, climate, and vegetation as the basis for
ecological units (GO MNR 2007), with more emphasis on the bedrock and substrate at finer
scales (Crins et al. 2007) compared to the terrain and elevation emphasis at finer scales in
the BEC system.
The PNW US has an extensive ecological plot database that has been used with various
vegetation classification methods. Major vegetation types with the states of Oregon and
Washington are described in Franklin and Dyrness's (1988) compilation of over 50 years of
classification work in the PNW. The PNW is equivalent to BC in diversity of climate, terrain,
and vegetation and therefore the ecological classification of this area includes a multi-scale
hierarchical ecological classification system, though not one directly comparable to the BC
BEC system.
The BEC system is based on ecological climax theory. This theory assumes that a climax
ecosystem represents a dynamic equilibrium or steady state that is the culmination of
ecological succession and is a product of the climate and local topography (Clements 1936).
The BEC system builds on Koppen's climate zones (1923), by incorporating site, vegetation,
and soil characteristics to create the current ecosystem classification (Kimmins 1987). The

4

BEC database (BECdb) was built using undisturbed mature (80-250 yrs old) and old growth
(>250 yrs old) field plot data to represent the climax plant communities of climatic regions
(Green and Klinka 1994, BC MFR 2009). The field data are used in combination with aerial
photographs, thematic maps, and visual inspections to create elevation, slope, and aspect
rules to map the subzone units (Eng and Meidinger 1999).
Biogeoclimatic subzones are the foundation of the system and are delimited by regional
climate (i.e., without local topographic effects). The subzone units are grouped together to
form BEC zones (broader ecosystem units defined by dominant tree species) and are sub­
classified into variants based on vegetation and characteristic plant species (MacKinnon et
al. 1992). Within each biogeoclimatic unit (subzone or variant), differences in terrain and
slope position result in different combinations of soil moisture regime (which can range
from xeric to hygric) and soil nutrient regime (which can range from poor to rich). Those
combinations of soil moisture regime and soil nutrient regime in each biogeoclimatic unit
are reflected in characteristic plant associations and constitute distinct "site series." Zonal
sites are mesic (modal) on the soil moisture gradient and average in soil nutrient regime, so
their vegetation is considered to representative of the climate more than topographic
factors. Like all site series within a subzone, the zonal sites have a distinct plant association
(Meidinger and Pojar 1991).
Biogeoclimatic units are strongly linked to climate and are therefore suitable ecological
units within the classification system for climatic envelope modelling (DeLong et al. 2010).
Climatic envelopes describe the climate niche or suitable climate space, but do not consider
biological or physiological factors. The climatic envelope approach is best applied to

5

ecological units with plenty of location data, such as the ecosystems and species used in this
study. Previous simulations and bioclimatic envelope modelling suggest a northerly shift of
suitable climatic space for dry coastal ecosystems up the coast and perhaps a shift toward a
different climate regime in the area currently occupied by the CDF (Burton and Cumming
1995, Hamann and Wang 2006). Climatic envelopes are useful to project areas of
persistence and areas of change that can be considered for long-term monitoring programs
and sustainable management planning.
Climate regime shifts and subsequent forest redistribution may occur over the long-term
(e.g., centuries). Indicators of these regional changes may be detected at the site level.
Some of the earliest evidence of change may come as inter-annual variability in phenology,
seasonal timing of leaf emergence in spring, and senescence in autumn (Waring and
Running 2007). Long-term forest microclimate datasets are largely unavailable due to the
expense and labour intensity associated with field sampling, though it is generally
recognized that field data are required to increase our ability to predict future conditions.
Climate fluctuations, the interaction between land and atmospheric conditions, and their
effects on vegetation can be linked using soil moisture measures (Rodriguez-lturbe 2000,
Entin et al. 2000). The dynamics of soil moisture play a dominant role in vegetation stress
and suitability of vegetation to climate and soil conditions (Guswa 2002). Establishment of
long-term sampling transects collecting microclimate measurements to increase the
confidence in descriptions of mechanistic relationships and the predictive utility of
quantitative models is an essential component of climate change research (Montaldo et al.
2008). Such transects could facilitate future studies to calculate and compare potential

6

water availability to vegetation among site types (Campbell 2008), quantify climatevegetation relationships, detect changes over time, and validate climate change model
projections.

1.3

Thesis Overview

This study investigates plant communities (assemblages of different plant species) at the
ecosystem and species level and the microclimate conditions where the study ecosystems
currently occupy space on the southwestern coast of BC. The first two chapters of this study
explore the utility of defining geographic boundaries of ecosystems and species using
current climate conditions to project spatial and temporal shifts in suitable climate space
with a five-model ensemble of climate change scenarios. Specifically, Chapter 2 explores
current and future ecosystem distributions for the CDF and the adjacent ecosystem unit, the
Coastal Western Hemlock very dry maritime (CWHxm), using BEC subzone mapping and
PNW US ecological plot data. Chapter 3 explores current and future species distributions for
18 species commonly found in the CDFmm. Chapter 4 describes the establishment of the
long-term monitoring network in the CDFmm and CWHxm and presents preliminary results
for one baseline year of field data. Results from these three chapters are discussed in the
final synthesis chapter to conclude the study.

7

2.

Ecosystem-Based Climatic Envelope Models: Current and Future
Distribution of Dry Coastal Ecosystems

2.1

ABSTRACT

Ecosystems are impacted by changes in climate. Based on the relationship between
mature forests and climate regimes, the dry coastal ecosystems of British Columbia are
investigated using climatic envelope models to project current and future suitable climate
space. The projections suggest a northward shift in ecosystem distribution, with the
introduction of new, undefined suitable climate space in southern Vancouver Island. The
projected Coastal Douglas-fir ecosystem boundary is constrained by different climate
variables across its range on Vancouver Island. The south is constrained by warm
temperatures and dry soil conditions, while the north is constrained by cool seasonal
temperatures. Areas of persistence and change are evaluated to consider impacts to
conservation and biodiversity. Results of the climatic envelope modelling are important for
improved understanding of the ecology of this biogeoclimatic zone, the potential for shifts
in ecosystem distribution, and in providing guidance to land and resource management.

2.2

INTRODUCTION

2.2.1

Climate Change and Impacts to Ecosystems

Ecosystems are impacted by climatic change globally. Here in North America, projections
suggest a warming trend with increases in winter precipitation and decreases in summer
precipitation (Christensen et al. 2007). The province of British Columbia (BC) is currently on

8

a warming trend that aligns with global projections reported in the 1990s (Johns et al.
1997). Recent studies suggest that high latitude, high elevation, and dry ecosystems are
most vulnerable to impacts of climatic change (Hamann and Wang 2006, Campbell and
Wang in press). Coastal temperate rainforests are projected to be relatively less impacted,
though species common to the dry coastal ecosystems are projected to gain suitable
climate in the next century (Hamann and Wang 2006, Rehfeldt et al. 2006). Trends of
increasing temperatures and shifts in precipitation regimes suggest a redistribution of dry
ecosystem types on the BC coast. Results of the research presented here define the spatial
and temporal scope of projected impacts of climatic change at a regional scale, indicate
areas of greatest (and least) climatic change over the next century, and explore the climatic
constraints on ecosystem distribution in the Vancouver Island study area.
Research into the ecological consequences of climatic change utilizes a variety of
modelling and analytical approaches, as well as the application of ecological theory. This
study uses the climatic envelope approach to define baseline and future distributions of dry,
coastal ecosystems in southwestern BC, to explore the relationships of climate and
ecosystem distribution, and to investigate how climatic change will impact the distribution
of suitable climate space (SCS). Suitable climate space is defined as the geographic area
that is within the defined range of the climatic envelope for an ecosystem (Rose and Burton
2011). The climatic envelope is an empirical definition of the climatic conditions acceptable
for an ecosystem based on its current spatial extent and the climate attributes associated
with the ecosystem at its known locations. Climatic envelopes are constructed using climate

9

and ecosystem models and data inputs that are analysed with geographic information
system (GIS) and statistical software.

2.2.2

Global Climate Models and Interpolation Tools

The natural environment is a complex and dynamic system that reflects the interaction of
biotic and abiotic components. Models can be used to study relevant aspects of dynamic
systems to understand individual components and the relationships between components
of the system (Soetaert and Herman 2009). As such, models have the dual role of
generating, to varying degree, both understanding of system functioning and predictions of
future system state (Bunnell 1989). Climate predictions are based on models and although
the estimates from global climate models (GCMs) have greatly improved over the years,
confidence remains higher for some variables (e.g., temperature) compared to others (e.g.,
precipitation) (Randall et al. 2007).
There are a range of GCMs available from various climate research institutions (e.g.,
Hadley Center, UK; Bjerknes Center, Norway; Canadian Center for Climate Modelling and
Analysis, Canada; Commonwealth Scientific and Industrial Research Organisation, Australia;
and others, Appendix A) and each institution uses a variety of carbon emission scenarios
(CES) to simulate baseline and future climate (IPCC 2007). GCMs are created using physical,
chemical, and biological properties of the climate system, as well as the interactions of
these properties and the feedback processes. The output of a GCM is limited by the
assumptions, initial climate conditions, statistical inputs, calibration uncertainty, temporal
and spatial scaling, and the underlying framework inherent in developing and exercising a

10

global model (Knutti 2008). Because GCMs are each based on different assumptions and
different parameters selected to describe climate behaviours, a range of model projection
outputs is expected. The ensemble approach for exploring multiple GCM and CES
combinations captures the range of possible responses while reducing the overall statistical
error in projected climatic change (Berry et al. 2002, Gleckler et al. 2008, Pierce et al. 2009,
Radic and Clarke 2011). Though error is presumably reduced by averaging across the
ensemble, the confidence in the model projection is not increased. This study employs a
five-model GCM ensemble based on GCM and CES combinations that best describe the
current climate of the study area, based on the assumption that the models that best
represent current climate will also project the most relevant future climate scenarios
(Gleckler etal. 2008).
GCMs are large-scale global models. There is a range of climate download and
interpolation tools available for regional climate change research. Two tools used in this
study include the Pacific Climate Impacts Consortium (PCIC) Regional Analysis Tool (RAT)
and Climate Western North America (ClimateWNA v.4.52; Hamann and Wang 2005, Wang
et al. 2010).
The PCIC online tool provides spatial and statistical GCM data for baseline and future
timeslices (PCIC 2011). The PCIC RAT data source for the baseline data is the Lawrence
Livermore National Laboratory (T. Murdock, pers. comm., January 29, 2012). The user
defines custom regions for data outputs, however, the data outputs are limited to mean
annual air temperature and precipitation variables.

11

The ClimateWNA tool provides scale-free, interpolated climate data for baseline
timeslices and scale-free, interpolated GCM projections for future timeslices (Wang et al.
2012). ClimateWNA is the updated version of ClimateBC, which was created to improve the
relationship between the scale of the available climate data and the spatial scale of
resource databases (Hamann and Wang 2005, Wang et al. 2006, 2012). ClimateWNA
processes climate data using bilinear interpolation and elevation adjustment based on the
2.5 arcmin x 2.5 arcmin cell grids of the Parameter-elevation Regressions on Independent
Slopes Model (PRISM, Daly et al. 2002). The result is a continuous layer of scale-free climate
data projected for one baseline climate normal timeslice (1961-1990) and three future
climate normal timeslices (2010-2039; 2040-2069; 2070-2099) using GCMs from the third
and fourth Intergovernmental Panel on Climate Change (IPCC) Assessment Reports (Wang
et al. 2006, in press).
ClimateBC demonstrated an improved ability to simulate independent climate station
data when compared to the original PRISM data when the tool was tested for regional
application (Wang et al. 2006). ClimateWNA includes statistical and data input upgrades to
ClimateBC and it covers a greater geographic area of Canada and the United States (Latitude
24.5° N - 80.0° N; Longitude 100° W -179° W), facilitating climate research beyond the BC
provincial boundary (Figure 2-1).

12

Figure 2-1: The geographic area covered by the ClimateWNA interpolation tool (from
http://www.genetics.forestry.ubc.ca/cfcg/ClimateWNA/ClimateWNA.html).
Climate normals are calculated using the arithmetic mean of climate station data over a
30-year time period (WMO 1989). This study uses the 1961-1990 normals as the baseline
timeslice. The ecological dataset used for this study was collected between 1978 - 2009,
with the majority of the plot data collected between 1980 -1999 (Pojar et al. 1987). Three
future timeslices of projected climate normals are also used (2010 - 2039; 2040 - 2069; and
2070 - 2099) to evaluate potential change of SCS over time.

2.2.3

Ecosystem Classification and Ecosystem-Based Climatic Envelopes

Ecosystem classification (e.g., by physiognomy, function, and structure) provides order to
the matrix of physiographic and vegetation patterns across the landscape (Shimwell 1971).
The characteristic structure of an ecosystem is a result of the interactions of organisms and
the physical environment within a given area (Odum 1969). Original biome-level

13

classification systems were developed at large spatial scales because vegetation distribution
is driven by regional climate (Woodward 1987, Pearson et al. 2002).
Forest zones were being classified in BC and the Pacific Northwest (PNW) of the United
States (US) in the early twentieth century (Meidinger and Pojar 1991). Since then, studies of
ecosystem structure and function have led to the development of numerous classification
and mapping systems based on multi-scaled hierarchical classifications to be used for
ecosystem-based management, planning, conservation, and research.
BC forests have been classified into forest types based on composition, soil conditions,
physiography, and climate, ultimately leading to the development of the currently used BC
Biogeoclimatic Ecosystem Classification (BEC) system (Meidinger and Pojar 1991). The PNW
US has an extensive ecological plot database that has been used to describe ecosystems in
the PNW using various vegetation classification methods (Franklin and Dyrness 1988). The
PNW is equivalent to BC in diversity of climate, terrain, and vegetation and there is a multi­
scale hierarchical ecological classification system used there, though not one directly
comparable to the BC BEC system. To facilitate climatic envelope modelling of dry, coastal
ecosystems as classified in BC, the BEC ecosystem unit concept is extended into the PNW US
using ecological plot data.
The BEC system is based on ecological climax theory. This theory assumes that a climax
ecosystem represents a dynamic equilibrium or steady state that is the culmination of
ecological succession and is a product of the climate and local topography (Clements 1936).
The BEC database (BECdb) was built using undisturbed mature (80-250 yrs old) and old
growth (>250 yrs old) field plot data to represent the climax plant communities for climatic

14

regions (Green and Klinka 1994, BC MFR 2009). The field data are used in combination with
aerial photographs, thematic maps, and visual inspections to create elevation, slope, and
aspect rules to map the subzone units (Eng and Meidinger 1999).
Biogeoclimatic subzones are the foundation of the system and are delimited by regional
climate (i.e., without local topographic effects). The subzone units are grouped together to
form BEC zones (broader ecosystem units defined by dominant tree species) and are sub­
classified into variants based on vegetation and characteristic plant species (MacKinnon et
al. 1992). Within each biogeoclimatic unit (subzone or variant), differences in terrain and
slope position result in different combinations of soil moisture regime (which can range
from xeric to hygric) and soil nutrient regime (which can range from poor to rich). Those
combinations of soil moisture regime and soil nutrient regime in each biogeoclimatic unit
are reflected in characteristic plant associations and constitute distinct "site series." Zonal
sites are mesic (modal) on the soil moisture gradient and average in soil nutrient regime, so
their vegetation is considered to representative of the climate more than topographic
factors. Like all site series within a subzone, the zonal sites have a distinct plant association
(Meidinger and Pojar 1991).
Biogeoclimatic units are strongly linked to climate and are therefore a suitable scale
within the classification system for climatic envelope modelling (DeLong et al. 2010).
Previous bioclimatic envelope modelling suggests a slight expansion (Rose and Burton 2009)
and a northerly shift of SCS for dry coastal ecosystems up the coast and perhaps a shift
towards a different climate regime in our study area (Hamann and Wang 2006).

15

To accommodate the northerly shift in SCS, the BC ecosystem classification system is
extended south for the study ecosystem to better define the southern components of
ecosystem climatic envelopes. The climatic envelopes will not specifically predict change to
ecosystems, nor will these empirical models explain the cause and effect of model climate
parameters and ecosystem response, but they will quantify potential SCS for ecosystems
under climate change scenarios (Pearson and Dawson 2003, Hijmans and Graham 2006).
The climatic envelopes are directive in that they suggest how one of the factors in the
realized distribution of ecosystems will shift relative to baseline conditions so that
ecosystem processes and functions can further be evaluated based on the changes in SCS
(Figure 2-2).
,------ --! Climate Factors
i
___

j

•—— C^CHmatic Envelope Models^)
Suitable Climate space

i
1 Dispersal Factors

i
k

POTE NTIAL DISTRIBUTION
V/

i Disturbance Factors r " ^ C ^ D y n a m ic Models
•
i Resource Factors
•

t/
'{

REALIZED DISTRIBUTION

i

Suitable climate space modulated
by dispersal, disturbance, and
biotic interactions
Figure 2-2: Relationship between static climatic envelope models and other factors contributing to
dynamic models. (Adapted from Guisan and Thuiller 2005, Franklin 2009).
It is customary to exclude potential outliers and resultant overestimations of SCS by
defining climatic envelopes with l st-99th or 5th-95th percentiles, but it is also recognized that
this approach may also remove valid observations and thus underestimate the SCS (Skov

16

and Svenning 2004). Climatic envelopes are defined here using minimum and maximum
values to represent the full extent of the climatic boundaries of ecosystem distribution, with
the understanding that the SCS may be overestimated slightly with this approach. These
empirical definitions could be further applied to investigate the relative importance of
climatic thresholds and other resource factors such as terrain, soils, and the history of
disturbance and dispersal using field experimentation and process-based modelling
techniques (Landsberg 2003). The climatic envelope definitions include both temperature
and water balance climate variables to reflect the energy requirements, heat and cold
tolerance, and the amount and timing of precipitation in relation to evapotranspiratory
demands (Skov and Svenning 2004).

2.2.4

Research Objectives

To determine the baseline relationships of climate and ecosystem distribution and how
projected climatic change will impact the distribution of the SCS, this chapter addresses five
research objectives: (1) determine five GCMs that closely depict baseline climate normals
for the study area; (2) extend the Coastal Douglas-fir biogeoclimatic unit into the PNW US
using ecological plot data; (3) develop baseline ecosystem climatic envelopes and project
SCS for three future timeslices; (4) determine locations in southwestern BC where SCS for
study ecosystem units is projected to overlap, persist, and shift overtime; and (5)
determine the climate variables that constrain the baseline extended CDF SCS on Vancouver
Island.

17

2.3

METHODS

2.3.1

Study Area

The focus of the climatic envelope analysis is the dry coastal ecosystems in southwestern
BC, specifically areas currently within and adjacent to the moist maritime subzone of the
Coastal Douglas-fir biogeoclimatic zone (CDFmm) on Vancouver Island, BC. Data are
compiled from a larger geographic area to capture the full extent of the CDF climate range,
including the smaller islands of the Salish Sea and the adjacent mainland coast of BC as well
as Washington and Oregon in the PNW US. The Vancouver Island study area consists of two
ecosystem types, recognized as distinct subzones in the BC BEC system, with equivalent
ecosystems located in the US (Nuszdorfer et al. 1991).
Within BC, the CDFmm subzone is the smallest subzone on the coast, (257 000 ha, <0.3%
of the province) and the only subzone within the Coastal Douglas-fir (CDF) zone. Located on
the southeastern side of Vancouver Island, the Gulf Islands, the southwest Lower Mainland
and part of the Sunshine Coast on the BC mainland (Figure 2-3), this subzone consists of a
Mediterranean-type climate with vegetation that thrives in the rain shadow of the Olympic
and Vancouver Island Mountains and is characterized by warm, dry summers and mild, wet
winters (Green and Klinka 1994).
The Coastal Western Hemlock (CWH) zone is distributed up the latitudinal gradient of the
provincial coast, but the Coastal Western Hemlock very dry maritime (CWHxm) subzone is
the smallest (896 300 ha, 1.0% of the province), warmest, and driest of all CWH subzones in
BC. The CWHxm is currently distributed adjacent to the CDFmm, westward (typically at
higher elevations) and northward of it (Figure 2-3).

18

CWH xm

0

25

90

~4i t, I.,!.

lOOKDomataf*

Figure 2-3. Current distribution of the Coastal Douglas-fir moist maritime (CDFmm) and the Coastal
Western Hemlock very dry maritime (CWHxm) biogeoclimatic subzones (BEC v.7) on southeastern
Vancouver Island, the Gulf Islands, and adjacent mainland coast, British Columbia.
Annual and seasonal trends for these subzones are similar, though overall, the CWHxm is
slightly cooler and wetter (Table 2-1). These climatic differences result in a shorter growing
season in the CWHxm and a summer soil moisture deficit in the CDFmm. These differences
in air temperature and moisture regime are also sufficiently distinct to differentiate
dominant tree species composition in the two BEC zones: the CDF is dominated by coastal
Douglas-fir (Pseudotsuga menziesii var. menziesii) and the CWH is dominated by Douglas-fir
and coastal western hemlock (Tsuga heterophylla) on zonal sites (Green and Klinka 1994).
This study area, with low topographical relief of the east side of Vancouver Island, is a good
candidate for climatic envelope modelling as ecosystems at higher elevations are more
frequently misclassified (Hamann and Wang 2006) when employing bioclimatic envelope
modelling.

Table 2-1: Baseline climate normals (1961-1990) for Coastal Douglas-fir moist maritime (CDFmm)
and Coastal Western Hemlock dry maritime (CWHxm) biogeoclimatic subzones in British Columbia
The mean and standard deviation were calculated by overlaying a 1-km grid over BEC v.7 subzone
mapping (BC MFR 2008) and using ClimateWNA (Wang et al. 2010) to output climate variables*.
CDFmm

CWHxm

Mean

Standard

Mean

Standard

(n=3 597)

Deviation

(n=10 515)

Deviation

Mean Annual Temperature (°C)

9.6

0.33

8.7

0.77

Mean Warm Month Temperature (°C)

16.9

0.41

16.5

0.66

Mean Cold Month Temperature (°C)

3.2

0.53

1.8

0.86

13.7

0.74

14.7

0.78

Mean Annual Precipitation (mm)

1065.0

169.03

1901.2

647.12

Mean Summer Precipitation (mm)

195.9

44.64

335.3

104.41

Annual Heat Moisture Index

18.9

3.14

10.8

3.19

Summer Heat Moisture Index

90.2

17.48

54.0

16.34

Degree Days below 0°C (days)

50.6

13.09

96.6

35.55

Degree Days above 5°C (days)

1923.6

78.97

1733.7

165.57

Number of Frost-Free Days (days)

300.1

12.99

276.2

20.23

Frost-Free Period (days)

218.2

19.28

191.1

23.22

Precipitation as Snow (mm)

43.0

14.55

143.6

98.81

Extreme Minimum Temperature (°C)

-16.6

1.48

-19.1

1.96

Hargreaves reference evaporation

654.7

20.15

639.0

37.47

Hargreaves climatic moisture deficit

276.4

48.34

162.5

70.74

Climate Variable

Temperature Difference (TD) between
MWMT and MCMT (°C)

*Climate variable definitions in Appendix B.
The CDFmm-CWHxm transition is an area of interest both ecologically and economically.
With general trends of increasing temperatures and potentially greater moisture stress,
there is the potential for CDFmm species and communities to occupy lands currently in the
CWHxm subzone. The CDFmm is inhabited by 218 rare and endangered plant and wildlife
species and 35 recognized plant communities (CDC 2009). While much of the CDFmm is
developed or in small, private land holdings, a larger proportion of the CWHxm is on public

20

land (Table 2-2) and is an important, productive part of the timber harvesting land base for
commercial forestry.
Table 2-2: Total area (km2) with private and public* land percentages (%) for Coastal Douglas-fir
moist maritime (CDFmm) and Coastal Western Hemlock very dry maritime (CWHxm) subzones (BC
LRDW 2011)
Land Area

CDFmm

CWHxm

Total area (km2)

2 520

9 583

Percentage (%)

100

100

Private land (km2)

2 016

5 270

Percentage (%)

80

55

Public land* (km2)

504

4 313

Percentage (%)

20

45

* Includes municipal, provincial, and federal land.
Research in the CDFmm and CWHxm subzones is important because the timber,
recreation, and urban development demands on these ecosystems are high. Research is
required to inform decision making processes for land and natural resource management,
particularly by providing guidance in identifying areas of persistence where climate will
potentially be relatively stable and areas of transition where rate of change will potentially
be the greatest.

2.3.2

Data and Analysis

The climatic envelope models used secondary data sets for both climate and ecological
variables. The climate data were assembled using tools available through the world wide
web as part of the PCIC Regional Analysis Tool (RAT) and the ClimateWNA interpolation
tool. The PCIC RAT provided climate variable meta-data downloads for baseline (1961-1990)
GCM projections. These data were used in the selection of GCMs for the study. ClimateWNA
21

provided interpolated baseline data as well as projections to evaluate scenarios in three
future timeslices. These data were used in the baseline and future projections of SCS for the
ecosystems under study. Local climate station data were also used to evaluate ClimateWNA
interpolations within the study area. Climate station data sources included the BC Ministry
of Environment (MOE), the Capital Regional District (CRD), and the BC Ministry of Forests
and Range (MFR) Wildfire Branch climate stations.
Sources of ecological data included the BEC database (BECdb) plot data (BC MFR 2009)
and BEC version 7 (BEC v. 7) subzone mapping (BC MFR 2008) from BC, and Ecoshare
datasets from the US (Ecoshare 2011). The BECdb included hundreds of plots in the CDFmm
and CWHxm (BCMFR 2009) and the US datasets included thousands of plots in the PNW
that have been sampled by a variety of field personnel over the past 30 years to improve
ecosystem classification using stratified sampling. The BC ecological plot data included: plot
location, site description (e.g., slope, elevation, aspect), and vegetation (e.g., species code,
percent cover) variables sampled as outlined in the Describing Ecosystems in the Field
manual (BC MOELP and BC MOF 1998). The US Ecoshare database was a compilation of
ecological plot data from a variety of sources including the US Forest Service (USFS)
ecological and inventory plots and academic research projects. Vegetation data collection
followed similar methods to the BC plots, and data were recorded using standardized
species, structure, and composition codes within the Ecoshare database. These datasets
were optimal for climatic envelope modelling because of the large sample size (Hirzel and
Guisan 2002).

22

2.3.2.1 Global climate model ensemble analysis and selection
The first objective of this chapter was to select five GCMs that were best calibrated to
baseline (1961-1990) climate normals for the study area. Both the PCIC RAT and
ClimateWNA were required for GCM selection because GCM projection data were only
available for future timeslices in ClimateWNA. The PCIC RAT was used for baseline GCM
data and ClimateWNA was used for interpolated scale-free baseline climate data.
In the PCIC RAT, pre-defined regions were provided, or the user may choose a custom
area. A custom area for the study area, including Vancouver Island and the adjacent
mainland BC, was selected to download baseline data (Figure 2-4). Only the GCM scenarios
available in the ClimateWNA tool were downloaded and evaluated for the five-model GCM
ensemble (Appendix A). Mean annual air temperature, precipitation, maximum air
temperature, and minimum air temperature (with standard deviations) were provided by
the PCIC RAT for each GCM.
A 1-km point grid overlaying the PCIC RAT custom area (Figure 2-4) was prepared and run
through ClimateWNA to output the annual temperature and precipitation data for the
baseline climate normals. The mean and standard deviation for annual temperature and
precipitation were calculated.
The PCIC RAT and ClimateWNA climate data outputs were used to select the five GCM
and CES combinations that best match the baseline climate for the study area.

23

Figure 2-4: The Vancouver Island and adjacent mainland British Columbia "custom area" used in
Pacific Climate Impacts Consortium (PCIC) Regional Analysis Tool (RAT) for global climate model
(GCM) climate data baseline outputs (http://tools.pacificclimate.org/select); inset diagram
represents a portion of the ClimateWNA 1-km point grid locations used to output baseline
interpolated climate data over the entire custom area for GCM selection. Note that grid cell size and
location vary for each GCM output from the PCIC RAT, but all overlay the custom area.
To compare measurements made in different units (e.g., temperature in °C and
precipitation in mm), the data were standardized using Equation 2-1. This equation adjusted
the mean to zero and the variance to 1.0 using a combination of centering the values on the
ClimateWNA mean and normalizing the values (Wildi 2010). The standardized values were
used to calculate the Euclidean distance (Equation 2-2; Wildi 2010) between two points (the
GCM point and the ClimateWNA point).

Equation 2-1

x[ =

24

where Xj'= standardized value; Xj= the GCM or ClimateWNA mean annual value; x = the

ClimateWNA mean annual variable; and J~ECxi ~ x )2=: the ClimateWNA standard
deviation for the climate variable.

Equation 2-2

Deu =

~ x2 j)2

where Dei,2= the Euclidean distance between two points (ClimateWNA value and GCM
value); x lj= the standardized GCM value and; x2j= the standardized ClimateWNA value.
Based on the equations above, the origin (0,0) represents the standardized mean annual
temperature (MAT) and the mean annual precipitation (MAP) ClimateWNA values. The five
shortest Euclidean distances from the origin were the GCMs determined to be the most
similar to the ClimateWNA values and were included in the five-model GCM ensemble.
To evaluate the utility of the interpolated ClimateWNA data outputs for the study area,
these data were compared to local climate station data. Local climate station data (sourced
from climate stations not included in the current normals calculation processing for
ClimateWNA outputs) were gathered from three sites that were distributed across the
latitudinal gradient of the study area: The MOE Victoria (Topaz) climate station at the south
end (49° 20' 18.6" N, 124° 27' 5.76" W, 30 m elev.), the CRD 14G climate station located
between Victoria and Duncan (48° 29' 39.4" N, 123° 36' 52.6" W, 490 m elev.), and the MFR
Cedar climate station located just south of Nanaimo (49° 2' 51.4" N, 123° 52' 29.0" W, 30 m
elev.). ClimateWNA data were downloaded for the geo-referenced location of the climate
station over the time period in which the climate station data were available (ranges from

25

1989-2006; start date varied by climate station; end date limited by ClimateWNA as historic
monthly data availability ends in 2006). The 17-year time period is relatively short to
evaluate climate normals (30 years), but it was the most complete and accurate observation
data set available in the study area. The climate station data and the ClimateWNA output
data were compared using linear techniques.

23.2.2 Ecosystem classification
To capture the full extent of the focal ecosystem and to define its climatic envelope, the
concept of the CDF BEC zone was extended into the PNW US. This was accomplished by
filtering geo-referenced ecological plot data from the PNW using BEC rules for zone and
subzone delineation and species composition criteria specific to CDF zone forest unit.
The first step was to select plots representing zonal sites, as the objective was to
delineate ecosystems representative of the regional climate and zonal sites are where soil
moisture conditions are primarily controlled by climate (Kimmins 1987). Crest, upper slope,
lower slope, and depression positions reflect the influence of local site conditions as well as
regional climate, and hence were excluded from further consideration.
US ecological plots were further filtered to include only those meeting the following
criteria: homogeneity, mature forest (climatic climax), and size (> 400 m2). US plots located
west of the height of land of the Cascade Mountain Range and north of the OregonCalifornia border were selected based on biogeography and distribution theory. The study
area remained west of the Cascade Range to maintain the maritime climate and north of
the Siskiyou Mountains in the Klamath Range which extends west-east forming a mountain

26

barrier near the Oregon - California border. This region is a unique geographic area in age,
history, and geological character as well as a complex of climate and vegetation diversity
where the PNW and California flora meet and mix together (Whittaker 1960).
The second filtering process was based on CDF zone vegetation, focussing primarily on
tree species as the tree canopy is indicative of long term climate regimes. The CDF
vegetation filter rules and rationale are outlined in Table 2-3. The plots that met all rules
were included in the extended CDF zone projections.
After the filtering process, zonal plots from BC and the US databases were compiled and
loaded into PC-ORD (McCune and Mefford 2006) to complete a cluster analysis using the
Euclidean (Pythagorean) distance measure and Ward's Method for group linkage. This
analysis was based on the fact that vegetation is indicative of regional climate and if results
indicated the vegetation from all zonal plots to be similar (e.g., one cluster), they could be
considered one subzone. If different (e.g., BC clustered into separate group from the US
plots) then the plots represented a larger forest zone unit, with potentially multiple subzone
units not found in BC. The CWHxm climatic envelope was defined using BEC mapping and BC
climate data only, similar to model inputs for other bioclimatic envelope studies (Wang et
al. 2006). Ecosystem classification of US plot data was limited to the CDF forest ecosystems
and was referred to as the "extended CDF". The extended CDF zone and the CWHxm
subzone are explored in this thesis.

27

Table 2-3: Percent cover rules and rationale for filtering species (common and scientific names) to
identify plots from the Pacific Northwest United States ecological plot database that can be
considered equivalent to the Coastal Douglas-fir forest zone.
Rule

Species (layer, if applicable)

Rationale

(percent
cover)
Mature Coastal Douglas-fir (Pseudotsuga

2 1%

Dominant tree species in CDF zone

tl%

Regeneration indicates suitable climate

menziesii)
Regeneration Coastal Douglas-fir;

currently persists

(Pseudotsuga menziesii)
Western hemlock (Tsuga heterophylla)

<2%

To distinguish from adjacent CWH zone

White fir (Abies concolor)

0%

To distinguish from higher elevation
forests

Amabilis fir (Abies amabilis)
Subalpine fir (Abies lasiocarpa)
Noble fir (Abies procera)
Yellow cedar (Chamaecyparis
nootkatensis)
Engleman spruce (Picea engelmannii)
Whitebark pine (Pinus albicaulis)
Mountain hemlock (Tsuga mertensiana)
California incense-cedar (Calocedrus

0%

decurrens)

To distinguish from California-type
forests

Tanoak (Lithocarpos densiflorus)
Sugar pine (Pinus lambertiana)
Western larch (Larix occidentialis)

0%

To distinguish from Interior Douglas-fir
forests

0%

Sitka spruce (Picea sitchensis)

To distinguish from coastal fog belt

2.3.2.3 Ecosystem-based climatic envelopes: baseline and future distribution

Climatic envelope definitions
Climatic envelopes were defined using the baseline range of climate conditions and the
definitions were used to project baseline and future SCS. Baseline climatic envelope
28

definitions were parameterized using a 1-km point grid overlay on the BEC v.7 subzone
mapping for BC and the filtered US plot locations. These points were run through
ClimateWNA to derive baseline statistics for the set of corresponding climate variables.
The climatic envelopes were constructed using annual climate variable outputs from
ClimateWNA that were selected based on biological relevance (Appendix B). For example,
the annual variables that were defined using a specific Julian date on which the frost-free
period begins and ends (bFFP and eFFP, respectively) were eliminated from the annual
climate variable list as they restricted potential shifts in the timing of the frost-free period
(FFP). The FFP variable, measured in days, was retained as it was not married to a specific
date of the year.
Seasonal and monthly climate variables were also eliminated from the climatic envelope
definition of SCS for the same reasons as above (in that they were tied to specific dates of
the year). Seasonal variation in temperature and precipitation were captured using annual
climate variables such as mean warm monthly temperature (MWMT) and mean summer
precipitation (MSP) as changes in seasonal temperature and precipitation were projected
for the study area, though annual means may remain the same (Chmura et al. 2011). The
minimum and maximum values of 16 annual climate variables were used to define the
climatic envelopes.

Baseline and future distribution
To project the SCS for an ecosystem unit (extended CDF or CWHxm) a 1-km point grid
was distributed over the PNW (western Oregon and Washington) and southwestern BC
(Vancouver Island and adjacent mainland). The geo-referenced grid points were input to
29

ClimateWNA resulting in an output file with the baseline (1961 -1990) annual climate
variables for each point on the grid. Each of these points was evaluated for climatic
suitability using a series of conditional statements in R statistical software (R Core
Development Team 2010) based on the climatic envelope definitions. If all 16 climate
variables for one point on the grid were within the suitable range for the target ecosystem,
then the point was classified as "suitable" for the ecosystem.
The suitable points, with location data, were tallied and saved to a file to plot the spatial
distribution of the SCS. Percentiles ( l st-99th, 5th-95th, and 25th-75th) were calculated for each
climate variable using all points included in the baseline SCS. The grid points were evaluated
using percentiles to represent specified confidence levels ( l st-99th and 5th-95th percentiles)
and core area (25th-75th percentiles) of SCS. Some studies used the 1st - 99th or 5th - 95th
percentiles for local climate variables to define the climatic envelope to remove the effects
of rare microsites and potential errors in the input location data (Rose and Burton 2011).
Here the minimum and maximum values were used to explore the full potential extent of
the ecosystem distribution, including the rare occurrences. The rigorous 16 climate variable
filter was a trade-off to the potential errors in the input data set (e.g., conservative model
outputs due to the stringent set of conditional statements).
To project future distributions, the 1-km point grid was run through ClimateWNA to
generate the output files with the climate variables for each of the five GCMs selected for
the five-model GCM ensemble for the three future timeslices (2010-2039, 2040-2069, 20702099). The mean for each climate variable for each point on the grid was calculated using R
statistical software (R Core Development Team 2010) to create a single five-model GCM

30

ensemble file for each future timeslice. Using the climatic envelope definitions, each point
in the five-model GCM ensemble file was evaluated for suitability for the target ecological
unit (extended CDF or CWHxm). SCS for each ecosystem was determined for the three
future timeslices, using the minimum-maximum definitions of SCS created from the baseline
climate normals. The SCS was projected over space using Arc GIS software (ESRI2009). The
l st-99th, 5th-95th, and 25th-75th percentiles were also projected to evaluate both the
confidence in the model output and the core area over time.

23.2.4 Areas of ecosystem overlap, persistence, and shift in southwestern British Columbia
The extended CDF and CWHxm climatic envelope projections for four timeslices
(baseline; 2010-2039; 2040-2069; 2070-2099) using the five-model GCM ensemble were
evaluated for overlap, persistence (e.g., no change), and areas of ecosystem shift over time.
Each point in the SCS was representative of 1 km2 and the tally of these points per SCS
represented change in area over time. Areas of overlap are points that were within the
climatic envelope definition of both the extended CDF and CWHxm for the same timeslice
and were explored to identify climatic ecotone areas over time.
Areas of climatic envelope persistence were evaluated point by point, in a similar manner
to the temporal corridor approach (Rose and Burton 2009) with an emphasis from baseline
to 2010-2039, baseline to 2040-2069, and baseline to 2070-2099 timeslices.
The areas of ecosystem shift were determined by examining changes in the patterns of
suitability with an emphasis from baseline to 2010-2039,2010-2039 to 2040-2069, and
2040-2069 to 2070-2099. The purpose of this exercise was to identify locations in

31

southwestern BC most suitable for long-term climate change monitoring installations
(Chapter 4). In particular, it was hypothesized that some locations currently in the CWHxm,
and close to the CDF boundary, may undergo a transition to a climate condition more like
the CDF SCS.

2.3.2.S Constraints on extended CDF baseline suitable climate space on Vancouver Island
The climatic constraints were evaluated across the boundary of the distribution of the
baseline extended CDF SCS on Vancouver Island. To determine the climate variable(s) that
most constrains the distribution of the extended CDF climatic envelope, a 5-km buffer area
within and outside the baseline extended CDF SCS boundary was analysed using a moving
window from the south to the north end of the boundary (48.42 °N - 50.10 °N latitude
range). Each window consisted of 30 paired samples. A sample pair was composed of one
point within the extended CDF SCS and one point outside of the extended CDF SCS. The
sample pairs were randomly selected from transects that bisect the boundary (east to west)
and were systematically distributed > 1 km apart from the south to north end of the
extended CDF SCS boundary (Figure 2-5).
A paired t-test analysis was used to test for a difference between the mean values within
and outside the boundary (2-tailed, n=30, p<0.05) for each iteration of the moving window.
The null hypothesis was that there is not a difference between the climate variable values
within and outside of the boundary; the alternative hypothesis was that there is a
difference between the climate variables within ar;d outside of the boundary. If the
absolute value of the test result was less than the critical t-value (df = 29), the null

32

hypothesis was not rejected, but if the test value was greater than the critical t-value (df =
29), the null hypothesis was rejected at the 5 % significance level.
The sampling window was stepped from the south to the north end of the boundary by
0.2 ° latitude, resulting in a total of 56 paired t-tests. Results were summarized by climate
variable t-test results across the latitudinal gradient using a 10 sample weighted average to
smooth the results line. The climate variables with most significant difference (e.g., highest
t-value) in each sampling window imply that they have the greatest constraint on locating
that boundary segment.
East-west transects

50Kfem«t*r*

Figure 2-5: The 5-km buffer within and outside of the extended Coastal Douglas-fir (CDF) suitable
climate space (SCS) for the baseline timeslice (1961-1990) with inset figure of the East-West
transects systematically bisecting the full extent of the extended CDF boundary on Vancouver Island,
British Columbia.

33

2.4

RESULTS

2.4.1

Global Climate Model Ensemble Analysis and Selection

The GCM-modelled values for the historical mean annual temperature (MAT) across the
custom area (as defined in Figure 2-4) range between -0.03 “C and 8.81 °C, while the
ClimateWNA-interpolated MAT average is 5.77 °C for the baseline normals (Table 2-4). The
GCM-modelled values for the mean daily precipitation (MDP) range between 4.07 mm/day
and 5.95 mm/day, while the ClimateWNA-interpolated MDP is greater than 1 mm above
this range at 7.2 mm/day (Table 2-4).
Table 2-4: Mean annual air temperature (°C) and mean daily precipitation (mm/day) for the
study area (as defined in Figure 2-4) as derived from baseline global climate model (GCM)
climate normals from Pacific Climate Impacts Consortium (PCIC) Regional Analysis Tool
(RAT) and interpolated local climate normals from ClimateWNA.
Mean Annual

Mean Daily

Temperature (°C)

Precipitation (mm/day)

AR4 BCCR BCM20, A2, run 1

-0.03

5.54

AR4 CCCMA CGCM3, A1B, run 1

4.52

4.13

AR4 CCCMA CGCM3, A2, run 1

4.52

4.13

AR4 CCCMA CGCM3, Bl, run 1

4.52

4.13

TAR CSIR02, A2

4.35

4.69

TAR CSIR02, B2

4.35

4.69

TAR ECHAM4, A2

4.3

4.64

TAR ECHAM4, B2

4.31

4.62

TAR HADCM3, A2

5.15

4.49

TAR HADCM3, B2

5.15

4.48

AR4 MIROC32 HIRES, Bl, run 1

8.81

5.81

AR4 MIROC32 MEDRES, A2, run 2

5.19

5.32

AR4 PCM, A2, run 1

1.64

4.95

AR4 PCM, Bl, run 1

1.7

4.92

AR4 UKMO HadCM3, A2, run 1

5.15

4.48

AR4 UKMO HadGEMl, A1B, run 1

4.36

5.95

ClimateWNA

5.77

7.21

Global Climate Model, Scenario, Run

34

The standardized (Equation 2-1, Wildi 2010) MAT and MDP values for the five GCMs used
in this study are compared to the standardized ClimateWNA MAT and MDP in Figure 2-6.
The five GCM scenarios with the shortest Euclidean distances to ClimateWNA values
(located at the origin) are: AR4 MIROC32 Medres A2, AR4 UKMO HadGEMl A1B, TAR
HADCM3 A2, TAR HADCM3 B2, and TAR UKMO HadCM3 A2 (Appendix A for GCM details).
Note that the TAR HADCM3 A2, TAR HADCM3 B2, and TAR UKMO HadCM3 A2 are so close
in value that they overlap in Figure 2-6, therefore it appears that only three points were
circled in red indicating selection for the five-model GCM ensemble for future climate
projections. All models and scenarios appear to under-estimate MDP 0.4 to 1.1 mm/day,
and even the best models and scenarios under-estimate MAT by 0.2 to 0.5 °C.

C

fl

X C Iim W N A

■ M IROC32 MEDRES_A2

* UKMO H a d G E M l_ A lB

-H A D C M 3 _ B 2

a

UKMO H adC M 3_A 2

+ C S IR 02_A 2

O E C H A M 4_A 2

• ECHAM4_B2

•C C C M A C G C M 3 _ A 1 B

O CCCMA CGCM 3_A2

- C CCMACGCM3_B1

- M IR 0 C 3 2 HIRES_B1

+ PCM_B1

X PC M _A 2

*C BCCR B C M 20

O Ensemble

a.
'U (...........
2! -1.2
-1
Q.

H A D C M 3_A 2
♦ CSIR02_B2

A2

,...........

r..

-0.8

-0.4

-0.6

0#
-

> ...................................................
re
_

“
©
re
a> ................................................
s

0.4

0.6

0.8

1.2

0.2

-0,4

<§>

#
+ • ....

0.2

- 0.2

-

0.6

-

0.8

+ ....

M ean Annual Temperature

Figure 2-6: Scatter plot of the standardized mean annual temperature (MAT, °C) and mean daily
precipitation (MDP, mm) values for the global climate model (GCM) carbon emission scenarios
considered for five-model GCM ensemble based on the availability from ClimateWNA to project
future timeslices. Note 1. The Ensemble GCM scenarios circled in red and the ClimateWNA value as
35

red "x". [GCM data source = Pacific Climate Impacts Consortium; interpolated climate station data
source = ClimateWNA], Note 2. The Hadley CM3 A2 and B2 and the UKMO Had CM3 A2 GCM values
are very close, so only three red ensemble circles appear on the figure.
Observed climate station data and interpolated ClimateWNA data for the location of the
three independent climate stations are compared below (Figures 2-7 and 2-8). Interpolated
projections of mean monthly temperature are very similar to the observed climate data (R2
£ 0.98 for all three stations; Figure 2-7, Table 2-5).

o
9
H
< 5
Z f
0J ~o
ro a»
ra

I

u

I
4at
->
c

Climate Station (observed) M M T (°C)
Victoria (2 0 0 1 -2 0 0 6 )

- CRD (19 98 -2 0 06 )

a Cedar (1 9 89 -2 0 06 )

Figure 2-7: Comparison of mean monthly temperatures (MMT, °C) for observed values recorded by 3
meteorological stations (Victoria, Capital Regional District (CRD) 14G, and Cedar) and interpolated
values calculated by ClimateWNA for the geo-referenced locations. Solid line is the linear regression
for each data series.
The relationships between interpolated projections and the observed climate station
total monthly precipitation are more variable: Victoria (R2 = 0.44); CRD (R2 = 0.72); and
Cedar (R2 = 0.90) (Figure 2-8).

36

2 0 0 jk

600

Climate Station (observed) TMP (mm)
Victoria (2 0 0 1 -2 0 0 6 )

■ CRD (1 9 98 -2 0 06 )

* Cedar (1 9 89 -2 0 06 )

Figure 2-8: Comparison of means for total monthly precipitation (TMP, mm) for observed values
recorded by 3 meteorological stations (Victoria, Capital Regional District (CRD) 14G, and Cedar) and
interpolated values calculated by ClimateWNA for the geo-referenced locations. Solid line is the
linear regression for each data series.
The mean monthly temperature and total monthly precipitation for observed and
interpolated data are compared in Table 2-5. For temperature, the southern station
(Victoria) is most similar to the interpolated values; and the values become less similar as
comparisons move northward (CRD to Cedar). Precipitation is more variable in its similarity
between interpolated and observed values, with high standard deviations for both observed
and interpolated values.

37

Table 2-5: Summary table of the relationship between the observed (climate station) and
interpolated (ClimateWNA) mean monthly temperatures (°C) and total monthly precipitation (mm)
at three meteorological stations distributed latitudinally within the study area (Victoria, Capital
Regional District, Cedar).
Location

Victoria (2001-2006)

CRD (1998-2006)

Cedar (1989-2006)

[N]

[61]

[105]

[211]

Temperature (°C)
Observed

Interpolated

Observed

Interpolated

Observed

Interpolated

Mean

10.7

10.6

8.3

9.2

14.2

10.1

St. Dev.

4.2

4.4

5.2

4.9

6.2

5.4

R2

0.99

0.98

0.98

Slope

1.03

0.94

0.87

Precipitation (mm)
Observed

Interpolated

Observed

Interpolated

Observed

Interpolated

Mean

69.2

64.4

119.2

126.9

79.5

95.3

St. Dev.

80.7

61.2

123.9

116.9

81.1

73.3

R2

0.34

0.70

0.90

Slope

0.44

0.72

0.90

2.4.2

Ecosystem Classification

The US ecological plot data were filtered using the BEC rules for zone and subzone
delineation and species composition rules as outlined in Table 2-3 of the Methods section.
The filtering process resulted in 96 zonal, CDF-equivalent plots distributed through the
Puget Trough, WA, and Willamette Valley, OR, as well as on the lower slopes of the Coast
Mountain and Cascade Mountain ranges on each side of the valley lowlands (Figure 2-9).
The US CDF plots are mostly found within areas of WA and OR mapped as Western Hemlock
and Douglas-fir zones based on Ecoshare's (2011) Modelled Potential Natural Vegetation
(MPNV) Zones system (Figure 2-9). The 24 plots located in the Western Hemlock MPNV

38

Zone are classified as CDF due to the low percent cover of Western Hemlock and the high
percentage of Douglas-fir recorded in both the mature and regeneration layers, indicating
that plots in these stands are more analogous to BC's CDF zone. Other plots in the Western
Hemlock MPNV Zone were rejected due to >2 % cover of Western Hemlock.
7?“
^

*

L .V .J ,

M

N

'■>J/.:’

A

-

<

'-Vw

Vancouver V .
island, BC

£_•>

V -*-

L«g«nd
•

US CDF plots

Hi
HI
Hi
Hi

Salt Desert
Grassland

EZj

Hi
Hi
m

Hi
Hi
Hi
■1
Hi
Hi

CD
H il
■■
1

Hi
EO

Steppe
Western Juniper Zone
Lodgepole Pine
Sitka Spruce
Ponderosa Pine
Oregon W hite Oak
Jeffrey Pine
Douglas-fir
Tanoak
Grand Fir
W estern Hemlock
W hite Fir
Shasta Red Fir
Pacific Fir
M ountain Hemlock
Subalpine Fir
Parkland
Alpine

0 37.5 75

150 Kilometers

1i i i I i i i I

Figure 2-9: Modelled Potential Natural Vegetation Zones of Washington and Oregon (Ecoshare
2011), showing the distribution of 96 United States ecological plots classified as Coastal Douglas-fir
(CDF) zone ecological units used in combination with BEC v.7 mapping to create the climatic
envelope definitions for the extended CDF suitable climate space.

39

The elevation for the US CDF zonal plots ranges between 115 -1300 m above sea level.
This range is broader than the BC CDF where the elevation range is limited to 0-315 m
above sea level for plots in the BECdb (BC MFR 2009). The US CDF plot elevation mean is
420 m (standard deviation = 225 m) with the highest elevation plots located in Oregon on
south aspect slopes.
The vegetation data for BC and US CDF zonal plots were analysed using PC-ORD cluster
analysis, resulting in two clusters. The two clusters were then summarized using Vegetation
and Environment NexUS Professional (VPro, BC MFR 2007) vegetation summary tools to
compare the two groupings. Results of this analysis confirm that the US CDF plots fall into
two clusters: (1) CDFmm subzone and (2) an additional group of plots that extend the CDF
concept to another drier CDF subzone, based on vegetation composition and structure.
Although the two clusters were mapped together as the "extended CDF", the US CDF
extension expands the BC CDFmm to a broader ecological unit.
The extended CDF plot summaries indicate an additional vegetation assemblage with less
western red cedar (Thuja plicata) and grand fir (Abies grandis) in the tree layer and less salal
(Gaultheria shallon) and Oregon grape (Mahonia nervosa) in the understory. These plots
include more grass species and herbaceous cover, likely due to drier conditions and the
decrease in shrub species. The extended CDF ecological unit remains dominated by coastal
Douglas-fir canopy and therefore it is appropriate to classify as CDF at the BEC zone level of
the hierarchy. Further ecosystem classification and correlation work is required to classify
the plots to the subzone level (a step beyond the scope of this project), but the additional

40

zonal ecosystem is provisionally called Coastal Douglas-fir dry maritime (CDFdm) subzone
using BEC terminology.

2.4.3

Ecosystem-Based Climatic Envelopes: baseline and future distribution

The climatic envelope definitions for the extended CDF and CWHxm ecological units are
similar and show strong overlap for many climate variables (Table 2-6). In general, the
extended CDF is warmer and drier, resulting in fewer chilling degree-days (DD<0) and a
greater number of growing degree-days (DD>5).
Table 2-6: Summary of minimum (Min.) and maximum (Max.) climate values defining the climatic
envelopes for the extended Coastal Douglas-fir (CDF) and Coastal Western Hemlock very dry
maritime (CWHxm) biogeoclimatic units.

Extended CDF

Annual Climate Variable
MAT (°C)

Mean Annual Temperature

MWMT

Mean Warmest Monthly

CC)

Temperature

MCMT

Mean Coldest Monthly Temperature

CC)
TD (°C)

Continentality, temperature
difference between MWMT and

CWHxm

Min.

Max.

Min.

Max.

8.2

13.0

5.7

11.7

15.6

21.3

14.1

19.3

1.6

7.7

-1.7

6.3

11.1

18.4

11.2

17.9

438

3473

750

5379

81

429

134

883

6.2

45.9

3.2

25.2

43.3

248.1

17.8

120.1

4

102

8

314

MCMT
MAP

Mean Annual Precipitation

(mm)
MSP

Mean Annual Summer (May -

(mm)

September) Precipitation

AH:M

Annual Heat Moisture Index
(MAT+10)/(MAP/1000)

SH:M

Summer Heat Moisture Index
(MWMT)/( MSP/1000)

DD<0

Degree-Days below 0°C, chilling
degree-days

41

T a b le 2 -6 (c o n tin u e d ).

DD>5

CWHxm

Extended CDF

Annual Climate Variable
Degree-Days above 5°C, growing
degree-days

Min.

Max.

Min.

Max.

1574

3005

1157

2550

NFFD

Number of Frost Free Days

256

346

197

328

FFP

Frost Free Period (consecutive days)

162

303

122

266

PAS

Precipitations as Snow between

(mm)

August in previous year and July in

11

132

28

975

-21.7

-7.4

-26.9

-11.4

current year
EMT (°C)

Extreme Minimum Daily
Temperature over 30 years

Eref

Hargreaves Reference Evaporation

571

1001

508

961

CMD

Hargreaves Climatic Moisture Deficit

152

641

0

509

The baseline distribution of the extended CDF and CWHxm ecosystem climatic envelopes
in BC and the PNW US are shown in Figure 2-10. The extended CDF ecosystem SCS is
distributed further to the south and further inland on mainland BC (Figure 2-10a) while the
CWHxm ecosystem SCS extends further to the north (Figure 2-10b) than those ecosystems
are currently mapped. The extended CDF SCS is limited to the east side of Vancouver Island,
small islands of the Salish Sea and the adjacent mainland in BC, similar to the BEC v.7
mapped distribution. In the PNW US, the extended CDF SCS is limited to lower elevations in
the Puget Sound area and down through the Willamette Valley in Oregon.
The CWHxm SCS extends over much of Vancouver Island, smaller islands and onto the
adjacent mainland BC, a larger geographical area than currently described as CWHxm
(Green and Klinka 1994). The CWHxm SCS in the PNW US covers most of the Olympic
Peninsula and Puget Sound area, but not at the lowest or the highest elevations of either of
these geographic areas of Washington. Into Oregon, the CWHxm SCS is above the lowest

42

elevations of the Willamette Valley, but similar to the CDF SCS, the CWHxm SCS does not
extend into the fog belt of the west coast.

Legend
Minimum-MaximumSCS
l*'-99* percents*SCS
S
»_9 5 * percentile S
CS
YA 25*-7S*' percemBeSCS

Legend
MMmum-MaxlmutnSCS
percentsSCS
S*-95* percents SCS

■

I

0 50 100

0 50 100

200Klom et*f»

WOKIwneter*

Figure 2-10: Projection of the suitable climate space (SCS) for the (a) extended Coastal Douglas-fir
(CDF) and (b) Coastal Western Hemlock very dry maritime (CWHxm) zonal ecosystems for the
baseline annual climate normals (1961-1990) using ClimateWNA interpolated climate station data.
Each map is created by overlaying the minimum - maximum SCS, 1st - 99th, 5th - 95th, and 25th - 75th
percentile SCS.

The full extent (minimum-maximum climatic envelope) of the extended CDF covers
76 725 km2with the majority of this area located in the PNW US (Figure 2-10a, Table 2-7).
The l st-99th and 5th-95th percentile climatic envelope distributions are presented to consider
impacts of outliers and errors, as well as microsites and rarities in the input dataset. The
distribution of these percentiles is similar to the full extent, but the edges of the distribution
(typically at higher elevations) drop out of the percentile climatic envelopes. The extended

43

CDF SCS is reduced by 10 927 km2 in the l st-99th percentile range and by 35 074 km2 in the
5th 95th percentj|e range. The 25th-75th percentile climatic envelope distribution is presented

to consider the core area of this climatic envelope. Interestingly, the core (4 808 km2) is
located only in the PNW US, primarily at lower elevations in the Puget Trough and
Willamette Valley.
The full extent of the CWHxm is 119 050 km2, covering the full latitudinal range of the
domain area evaluated in this project (from the Smith Inlet and northern tip of Vancouver
Island, BC, to the south end of Oregon; Figure 2-10b, Table 2-7). Regions excluded from the
SCS are coastal areas and high elevations. Portions of the small islands of the Salish Sea and
Puget Sound are also excluded. The southern Oregon portion of the SCS drops out of the 1st99th and 5th-95th percentile envelopes. The CWHxm SCS is reduced by 19 441 km2 in the 1st99th percentile range and 53 179 km2 in the 5th-95th percentile range. The core (25th-75th
percentile climatic envelope) is further reduced to 4 820 km2 and omits much of BC
currently mapped as CWHxm (Figure 2-3) on Vancouver Island, BC.
Table 2-7: The total area (km2) for minimum-maximum and percentile projections of suitable climate
space (SCS) for the extended Coastal Douglas-fir (CDF) and Coastal Western Hemlock very dry
maritime (CWHxm) ecosystems baseline climate normals (1961-1990) using ClimateWNA
interpolated climate data.

Minimum-

l st-99,h

5th-95th

maximum

Percentile

Percentile

Percentile

Extended CDF SCS (km2)

76 725

65 798

41 651

4 808

CWHxm SCS (km2)

119 050

99 609

65 871

4 820

Biogeoclimatic Unit

25 -75th

Future projections indicate that the distribution of the extended CDF SCS is expected to
decrease over time (Figure 2-11), almost completely disappearing from the PNW US by the

44

end of the century (Figure 2 -llc ). Areas currently mapped as CDF in BC (BEC v. 7, Figure 2-3)
are projected to drop out of the SCS, particularly on the southeast end of Vancouver Island
and the south coast of mainland BC, while the distribution of the extended SCS can be
expected to move northwards. The total area for the full extent (minimum - maximum) of
the climatic envelopes for each timeslice is listed in Table 2-8. The extended CDF percentile
climatic envelopes for the 2010-2039 timeslice decrease in area from 73 664 km2
(minimum-maximum SCS) to 39 559 km2 ( l st-99th percentile range) to 8 092 km2 (5th-95th
percentile range) and to 23 km2 (25th-75th percentile range). The decreasing pattern is
similar in the 2040-2069 and 2070-2099 timeslices where climate conditions do not fit the
core (25th - 75th percentile) definition for the 2040-2069 timeslice and climate conditions do
not fit all percentile climatic envelope definitions for the 2070-2099 timeslice.

Ucend

I

H V '4 9 *p *rc *itt»sa

M wriw um Mm muro SCS

. S*-W*|»rc*itt»sa

Q

3

Mvrs*p«r<*it**$cs

0 » 109

Figure 2-11: Future distributions for the extended Coastal Douglas-fir (CDF) zonal ecosystem
climates, as projected for (a) 2010 - 2039, (b) 2040 - 2069, and (c) 2070 - 2099 timeslices, using the
five-model GCM ensemble. Distributions for the minimum - maximum, 1st - 99th, 5th - 95th, and 25th 75th percentile climatic envelopes included (when present) for each timeslice.
45

Future projections for the CWHxm climatic envelope are similar to the extended CDF SCS,
with a reduction in area over time (Figure 2-12). The climatic envelopes for the 2010-2039
timeslice is 84 229 km2 (minimum-maximum SCS) to 62 007 km2 ( l st-99th percentile range),
28 803 km2 (5th-95th percentile range) and 150 km2 (25th-75th percentile range). The
respective climatic envelopes decrease in the 2040-2069 timeslice with the 44 064 km2 in
the full extent, 27 944 km2 ( l st-99th percentile range), 9 850 km2 (5th-95th percentile range),
and 0 km2 (25th-75th percentile range). The last timeslice of the century is less than one-third
the area of the 2040-2069 timeslice with 15 094 km2 in the minimum - maximum SCS, 7 049
km2 ( l st-99th percentile range), 2 256 km2 (5th-95th percentile range) and 0 km2 (25th-75th
percentile range). Most of the current CWHxm (as mapped in BEC v.7) drops out of the
climatic envelope as the SCS shifts northward towards the end of the century (Figure 2-12c).

Ugtnd

I

M Inlw om M w iw w m SCS
1*169* p c rc m itta SCS

2 5 *-7 5 * p « ro M ttlt SCS

Figure 2-12: Future distributions for the Coastal Western Hemlock very dry maritime (CWHxm) zonal
ecosystem climates, as projected for (a) 2010 - 2039, (b) 2040 - 2069, and (c) 2070 - 2099
timeslices, using the five-model GCM ensemble. Distributions for the minimum - maximum, 1st - 99th,
5th - 95th, and 25th - 75th percentile climatic envelopes included (when present) for each timeslice.

46

Note that both the extended CDF and CWHxm SCS total area decreases over time, but by

different proportions (Table 2-8). The extended CDF SCS decrease in the first future
timeslice is relatively small (4 %) compared to the CWHxm decrease (29.2 %). By 2040 2069, both the extended CDF SCS and the CWHxm SCS are reduced to less than half of the
baseline SCS (52.6 % and 63 %, respectively). By the end of the century, the extended CDF
and the CWHxm SCS are expected to be reduced by 91.5 % and 87.3 %, respectively.
Table 2-8: Total area (km2) in the minimum-maximum climatic envelopes for the extended Coastal
Douglas-fir (CDF) and Coastal Western Hemlock very dry maritime (CWHxm) using 16 annual climate
variables from ClimateWNA over a 1-km grid of southwestern British Columbia, western
Washington, and western Oregon.
CWHxm SCS

Extended CDF SCS

Percent of

Percent of
baseline SCS (%)

Area (km2)

76 725

100.0

119 050

100.0

2010-2039

73 664

96.0

84 229

70.8

2040-2069

36 334

47.4

44 064

37.0

2070-2099

6 542

8.5

15 094

12.7

Timeslice
Baseline
(1961-1990)

2.4.4

Area (km2)

baseline SCS (%)

Areas of Ecosystem Overlap, Persistence, and Shift in Southwestern British Columbia

As noted in Table 2-6, the definitions of climatic envelopes for the extended CDF and
CWHxm overlap for many climate variables. The SCS for both study ecosystems were
evaluated to determine trends in the area of overlap for each timeslice. The overlap area for
the SCS of the two units increases in the 2010-2039 timeslice, then decreases over the next
two future timeslices into the end of the century (Table 2-9).

47

Table 2-9: Area of overlap (km2), total suitable climate space (SCS) area (km2), and percentage SCS
within the overlap area (%) for the extended Coastal Douglas-fir (CDF) and Coastal Western Hemlock
very dry maritime (CWHxm) suitable climate space (SCS) in southwestern British Columbia for
baseline (1961-1990) to future timeslices (baseline to 2010 - 2039, baseline to 2040 - 2069, baseline
to 2070 - 2099) using the five-model global climate model (GCM) ensemble and minimummaximum climatic envelope.
Baseline

2010 - 2039

2040 • 2069

2070 • 2099

5 239 km2

6 092 km2

2 522 km2

15 km2

Extended CDF SCS total area

5 453 km2

7 597 km2

10 100 km2

5 558 km2

(% of SCS in overlap)

(96.1 %)

(80.1%)

(25.0%)

(0.3%)

CWHxm SCS total area

40 089 km2

46 148 km2

38 557 km2

14 243 km2

(% of SCS in overlap)

(13.1%)

(13.2%)

(6.5%)

(0.1%)

Suitable Climate Space

Extended CDF and CWHxm
Overlap Area

The baseline overlap area (Figure 2-13a) covers much of the current distribution of the
BC BEC CDFmm as per BEC v.7 mapping (Figure 2-3). The overlap area shifts north slightly
and expands predominantly on the mainland into the 2010-2039 timeslice (Figure 2-13b).
The overlap area continues to shift northward but remains mostly on Vancouver Island into
the 2040-2069 timeslice (Figure 2-13c), then almost totally disappears with only 15 km2
remaining in the 2070-2099 timeslice (Figure 2-13d), approximately half-way up the east
coast of Vancouver Island, west of Campbell River.

48

Lagand
U 8 * nd

|£j] Overlaparea ^ 3 //

0 Overlaparea

0 50 100
« .i

0 50 100

200Kilometer*!""-'. _ , 1L

....U - I _ L _ 1 _ 1 ....1

\

200KBometenT -

........................................J

Legend

Legend

vJX .

\

V

. V

f l

Q Overlaparea

Overlaparea

0 50 100

^

0

200Kilometer^^ v __ A

50 100

200KHometerep ^ ....

Figure 2-13: The extended Coastal Douglas-fir (CDF) and Coastal Western Hemlock very dry maritime
(CWHxm) suitable climate space (SCS) overlap area as projected over time (a) Baseline; (b) 2010 2039; (c) 2040 - 2069; and (d) 2070 - 2099 timeslices, using the five-model global climate model
(GCM) ensemble and minimum-maximum climatic envelopes.
The areas of persistence were evaluated from baseline to each future timeslice. The
results continue to demonstrate a similar trend to the overall distribution of the total and
overlap area, as both the extended CDF and CWHxm persistent SCS decrease in size towards
the end of the century (Table 2-10).

49

Table 2-10: Areas of persistence (km2) in southwestern British Columbia as projected for baseline to
future timeslices (baseline to 2010 - 2039, baseline to 2040 - 2069, baseline to 2070 - 2099) in
extended Coastal Douglas-fir (CDF) and Coastal Western Hemlock very dry maritime (CWHxm)
suitable climate space (SCS) using the five-model global climate model (GCM) ensemble and
minimum-maximum climatic envelope.
Baseline to

Baseline to

Baseline to

2010-2039

2040-2069

2070-2099

Extended CDF SCS (km2)

5 451

5 319

310

CWHxm SCS (km2)

36 942

26 368

6115

Biogeoclimatic Unit

Future SCS projections of both the extended CDF and CWHxm suggest very little to no
area of persistence in areas currently mapped as CDFmm or CWHxm in BEC v.7 mapping
(Figure 2-3) by the end of the century (Figure 2-14 and 2-15). The extended CDF area of
persistence remains within the currently mapped CDFmm (BEC v.7) for two future
timeslices, but then ca. 5 000 km2 are projected to drop out of the persistent area in the
2070-2099 timeslice, resulting in a few small areas of extended CDF persistent SCS. The
persistent areas include the Jordan River area on the south end of Vancouver Island,
Qualicum Beach and Campbell River areas of the eastern side of Vancouver Island, a small
part of Texada Island, and the Powell River area on the Sunshine Coast on the BC mainland.

Figure 2-14: Extended Coastal Douglas-fir (CDF) ecosystem persistent area over time (a) Baseline to
the 2010 - 2039 timeslice; (b) Baseline to the 2040-2069 timeslice; (c) Baseline to the 2070-2099
timeslice, using the five-model global climate model (GCM) ensemble and minimum - maximum
climatic envelopes.
50

The CWHxm SCS is expected to persist in the currently mapped CDFmm and CWHxm
(BEC v.7) locations in the 2010-2039 timeslice (Figure 2-14a), but then is projected to shift
out of the CDFmm mapped area (BEC v.7) in the 2040-2069 timeslice (Figure 2-14b). The
persistent area remains in lower elevations of Vancouver Island and the mainland coast. By
the end of the century (Figure 2-14c), the CWHxm SCS can be expected to persist in areas
currently north of the mapped CWHxm (BEC v.7).

Figure 2-15: Coastal Western Hemlock very dry maritime (CWHxm) ecosystem persistent area
projected over time (a) Baseline to the 2010 - 2039 timeslice; (b) Baseline to the 2040-2069
timeslice; (c) Baseline to the 2070-2099 timeslice), using the five-model global climate model (GCM)
ensemble and minimum - maximum climatic envelopes.
Towards the end of the century, most of the southeastern end of Vancouver Island is
projected to be outside of areas of persistence for either the extended CDF or CWHxm. The
areas of persistence for the extended CDF and CWHxm do not overlap and the extended
CDF areas of persistence generally are to the south and lower elevations when compared to
the CWHxm areas of persistence.
Areas of ecosystem SCS shift from the CWHxm SCS to the extended CDF SCS on
southwestern BC are projected to increase over for the first two future timeslices, and then
decrease at the end of the century. The area shifting out of the extended CDF SCS to an
undefined SCS increases over time (Table 2-11).

51

Table 2-11: Area (km2) projected to shift from Coastal Western Hemlock very dry maritime (CWHxm)
suitable climate space (SCS) into extended Coastal Douglas-fir (CDF) and out of the extended CDF
into undefined SCS in southwestern BC over time (a) Baseline to the 2010 - 2039 timeslice; (b) 20102039 to 2040-2069 timeslice; (c) 2040-2069 to 2070-2099 timeslice), using the five-model global
climate model (GCM) ensemble and minimum-maximum climatic envelopes.
Biogeodimatic Unit SCS Shift

CWHxm to extended CDF (km2)
Extended CDF to undefined zone (km2)

Baseline to

2010-2039

2040-2069

2010-2039

to 2040-2069

to 2070-2099

1293

6 613

4 872

2

132

5 009

The area shifting from CWHxm SCS to extended CDF SCS is projected to increase over
time (Figure 2-16). The shift from CWHxm SCS to extended CDF in the first two timeslices
(Figure 2-16 a,b) includes area currently mapped as CDF (BEC v.7) and areas of persistence
for the extended CDF SCS in the respective timeslices. A substantial portion of the area
becoming CDF SCS in the 2070-2099 timeslice is from the area currently mapped as CWHxm
on Vancouver Island in BEC v.7 mapping (Figure 2-16c).

4

»

1M

U »««M

Figure 2-16: Area projected to shift from Coastal Western Hemlock very dry maritime (CWHxm)
suitable climate space (SCS) to the extended Coastal Douglas-fir (CDF) SCS over time (a) Baseline to
2010-2039; (b) 2010-2039 to 2040-2069; and (c) 2040-2069 to 2070-2099 timeslices, using the fivemodel global climate model (GCM) ensemble and minimum-maximum climatic envelopes.
The area projected to shift out of the CDF SCS is minimal in the first future timeslice
(2 km2, Figure 2-17a) and the second future timeslice (132 km2, Figure 2-17b). The area is

52

greatest in the 2070-2099 timeslice when a large portion of the area currently mapped as
CDFmm in the BEC v.7 mapping shifts to an undefined SCS (Figure 2-17c).

Figure 2-17: Areas projected to shift from the extended Coastal Douglas-fir (CDF) suitable climate
space (SCS) to undefined SCSover time (a) Baseline to 2010-2039, (b) 2010-2039 to 2040-2069, and
(c) 2040-2069 to 2070-2099 timeslices, using the five-model global climate model (GCM) ensemble
and minimum-maximum climatic envelopes.
The area shifting into the undefined SCS is compared to the extended CDF SCS by
evaluating the climate variable values over time (Table 2-12). In the first future timeslice,
the frost-free period (FFP) minimum and maximum values are higher than the current range
of FFP values for the climatic envelope, indicating a longer frost free period. During the next
timeslice, the mean warm monthly temperature (MWMT), continentality (TD), number of
frost free days (NFFD), and FFP are all higher than the SCS maximum value, while the
precipitation falling as snow (PAS) is lower than the SCS minimum value. In the last timeslice
of the century, the same climate variables remain out of range, with the addition of the
mean annual temperature (MAT) and growing degree-days (DD>5) that are also higher than
the SCS maximum values. The future projections for the southwestern BC portion of the
extended CDF SCS imply warmer, drier conditions, similar to the climate currently found in
the grasslands of the Willamette Valley in Oregon (Appendix C).

53

Table 2-12: Comparing the extended Coastal Douglas-fir (CDF) to the undefined zone in future
timeslices using the minimum (Min.) and maximum (Max.) values for climate variables of points
projected to shift from the extended CDF to undefined suitable climate space (SCS) in southwestern
British Columbia over time (Baseline to 2010-2039; 2010-2039 to 2040-2069; and 2040-2069 to
2070-2099 timeslices), climatic envelopes using the five-model global climate model (GCM)
ensemble and minimum-maximum climatic envelope definitions. Variables out of CDF baseline
climatic envelope range are highlighted in bold; BLUE = low, RED= high.

Baseline

Climate

Baseline to

Baseline to

2010-2039

2040-2069

2070-2099

Max.

Min.

Max.

Min.

Max.

Min.

Max.

MAT (°C)

8.2

13.0

11.4

11.4

12.2

12.5

12.0

14.1

MWMT (°C)

15.6

21.3

18.4

18.1

19.3

21.7

21.0

23.7

MCMT ("C)

1.6

7.7

5.5

5.5

3.2

6.1

3.8

6.9

TD (°C)

11.1

18.4

12.7

12.7

13.5

18.4

14.8

18.7

MAP (mm)

438.0

3473.0

746.6

750.2

728.4

1659.4

788.8

2343.2

MSP (mm)

81.0

429.0

150.6

151.0

109.2

345.4

113.6

341.4

AHM

6.2

45.9

28.5

26.7

13.5

30.6

9.6

30.2

SHM

43.3

248.1

120.8

121.1

63.7

181.5

69.3

195.5

DD<0

4

102

13

13

9

51

7

39

DD>5

1574

3005

2445

2450

2748

2871

2688

3399

NFFD

256

346

341

342

320

348

314

353

FFP

162

303

305

307

252

322

236

326

PAS (mm)

11.0

132.0

11.6

11.6

8.6

39.2

6.2

46.4

EMT (°C)

-21.7

-7.4

-10.2

-10.0

-16.0

-15.5

-8.3

Eref

571.0

1001.0

623.8

626.8

662.8

820.2

672.2

880.0

CMD

152.0

641.0

306.8

310.2

235.4

425.0

236.4

466.4

2.4.5

t

Min.

00

Variable

Baseline to

Constraints on the Extended CDF Baseline Suitable Climate Space on Vancouver
Island

Each climate variable was evaluated using a paired t-test analysis to determine the
climatic constraints of the boundary of the baseline extended CDF SCS on Vancouver Island.
Results are evaluated across a latitudinal gradient using a moving window of 30 paired
54

samples. Figure 2-18 represents the (a) increasing, (b) decreasing, and (c) neutral trends of
the significance of the climate variables from the southern to the northern end of the
baseline extended CDF minimum-maximum SCS boundary. The trend lines are smoothed by
averaging the t-values for every 10 paired tests along the boundary. The significance
increases with the increases in t-values. The location of each sample window is defined
using the midpoint latitude (decimal degrees).
Continentality (TD) is not significantly different within and outside of the boundary at the
southern end, but then increases in significance moving northward. Mean cold monthly
temperature (MCMT), cooling degree days (DD<0), number of frost free days (NFFD), frost
free period (FFP), and extreme minimum temperature (EMT) all become increasingly
significant in constraining the extended CDF boundary towards the northern end (Figure 218 a). FFP, EMT, NFFD, MCMT, and DD<0 are the most significant limitations on the
northern end of the boundary (Figure 2-19). Summer heat moisture (SHM), Hargreave's
climatic moisture deficit (CMD), annual heat moisture (AHM), mean annual precipitation
(MAP), precipitation as snow (PAS), and Hargreave's reference evaporation (Eref) all
decrease in significance from the south end to the north end of the boundary (Figure 2-18
b). Eref is not significantly different within and outside of the boundary at the north end.
SHM, CMD, AHM, MSP, and MAP are the most significant in constraining the southern end
of the boundary (Table 2-12). The mean annual temperature (MAT), mean warm monthly
temperature (MWMT), mean summer precipitation (MSP), growing degree days (DD>5) are
neutral in the significance of the climate variables constraint on the boundary (e.g., do not

55

increase or decrease strongly), but the MAT, MWMT, DD>5 do gain significance mid-latitude
on the boundary (Figure 2-18 c).

(a)
14

— MCMT
TD
--DD0
- NFFD
— FFP
- -EMT
— T-CRIT

12
at

10

_3 8

5I 6
4
2

0
48.8

49.0

49.4

49.2

49.6

MAP
AHM
SHM
PAS
Eref
CMD
T-CRIT
48.8
16

49.0

49.2

49.4

49.6

(C)

14

12

at
J3

sI

10

MAT

8

MWMT

6

MSP

4

DD5

2

T-CRIT

0
48.8

49.0

49.2

49.4

49.6

Latitude (°N)
Figure 2-18: Smoothed t-value (a) increasing, (b) decreasing, and (c) neutral trends for the 5-km
buffer moving window paired t-test (n=30, 2-tailed, p<0.05) for each climate variable within and
outside of the extended Coastal Douglas-fir (CDF) baseline suitable climate space (SCS) on
Vancouver Island, t-values for each climate variable are smoothed using a 10-sample weighted
average starting at the south end of the boundary.
56

MCM T
FFP /

DD<0

EMT
NFFD

Toflno

IFFD
AHM

MT

North"

SHM

Legend
IN extam tend CDF SCS

OUT extended CDF SCS
0

12.5 25

SHM
CMD
AHM
Figure 2-19: The three most significant climate variables* from the south to the north end of the
baseline extended Coastal Douglas-fir (CDF) minimum-maximum suitable climate space (SCS)
boundary on Vancouver Island. Climate variables are ranked by highest (1-3; most significant) using
a 10-sample average of t-value results from the paired t-test (n=30, 2-tailed, p<0.05).
*EMT = extreme monthly temperature, FFP = frost free period, MCMT = mean cold monthly
temperature, NFFD = number of frost-free days, DD<0 = degree-days less than 0°C, AHM = annual
heat:moisture index, SHM = summer heat:moisture index, MAT = mean annual temperature, and
CMD = climatic moisture deficit.
There is a shift in significance level for the climate variables when evaluating the
differences in climate on the inside and the outside of the extended CDF boundary on
Vancouver Island (Figure 2-19). Generally, the south end is constrained by moisture
variables and the north end is constrained by cool temperature variables. Specifically, the
57

south end of the boundary (Victoria-Duncan area), the summer heat moisture index (SHM),
climatic moisture deficit (CMD), and annual heat moisture index (AHM) are most significant,
while at the north end of the boundary (Courtenay-Campbell River area), the frost free
period (FFP), extreme minimum temperature (EMT), and number of frost free days (NFFD)
are most significant in defining the boundary. Summary statistics and test results for all
paired t-tests (e.g., before smoothing averages) with climate variables ordered by
significance are presented in Appendix D.

2.5

DISCUSSION

2.5.1

Global Climate Model Selection and Interpretation

This study uses a five-model ensemble approach for future GCM projections to explore
the implications of future climate scenarios. The GCM selection was based on a close match
to the current climate within the study area. This approach focuses on the application of
GCMs created using climate inputs, calibration, and scaling best suited to the coastal study
area. This approach differs from the "bookend approach" where a range of scenarios are
averaged to account for all extremes of modelled future conditions. Both of these
approaches have application to risk management and planning tools, to prepare for
uncertainty of future conditions.
The GCM ensemble approach is used by other studies but selection using mean annual
temperature (MAT) and mean daily precipitation (MDP) criteria to narrow the ensemble to
those models which do the best job of depicting current conditions has not yet been
reported elsewhere in the literature. More complex analyses have been conducted to

58

evaluate GCM ability to project Western North American climate (Gleckler et al. 2008,
Pierce et al. 2009, Radic and Clarke 2011). All of these studies agree that a multi-model
ensemble is superior to individual models in simulating mean annual climate cycles. They
found that the statistical errors in the model tend to be centered around zero and by using
the mean of multiple models, the overall error is reduced. Radic and Clarke (2011) also
ranked the two AR4 models used in this study (HadCM3 and MIROC3.2 medres) in their top
five (out of 22) performing models for northwest North America. Their study was based on
statistical metrics of six climate variables, including precipitation and temperature.
The GCM selection is based on the theory that this approach is also best applied to small
study areas where variability in baseline conditions is minimized before exploring potential
future climate conditions. Note that four of the five selected GCMs are from the Hadley
Center for Climate Prediction and Research/Met Office, UK, though with different carbon
emission scenarios (2 x A 2 ,1 x B2, and 1 x A1B) for future projections (Figure 2-6) suggesting
that the UK models are well-calibrated to Vancouver Island climate. The other GCM selected
for the ensemble is from the Center for Climate System Research (University of Tokyo),
National Institute for Environmental Studies, and Frontier Research Center for Global
Change (JAMSTEC), Japan. Three of the GCMs are from the third IPCC assessment report
generation and two of the GCMs are from the fourth IPCC assessment report generation.
Coincidently, this ensemble results in the application of models from institutes that are
located in a maritime environment, although not here in North America. Generally speaking,
the Hadley GEM1A1B model projects greater mean annual temperatures and lower mean
annual precipitation when compared to the other GCM scenarios included in the GCM

59

ensemble, but the trends in climate variables per GCM vary over time for the study area
(Appendix E).
When ClimateWNA interpolated data are compared to the GCM modelled data,
ClimateWNA mean annual temperature is within the upper range and mean daily
precipitation average is 1 mm greater of GCM values. This suggests that the GCMs with
higher precipitation and temperatures are selected for the five-model GCM ensemble to
project future climate and ecosystem SCS. When interpolated ClimateWNA and local
climate station data are compared, ClimateWNA interpolated data indicate that the mean
monthly temperature is well correlated with the local climate station data, but that the
mean monthly precipitation correlations are much poorer. This reinforces comments that
precipitation is difficult to model and is also variable across the landscape (Bates et al.
2008). Overall, the evaluation results presented in section 2.4.1 suggest that the five GCM
scenarios chosen for the ensemble are a reasonable match to the baseline data from
ClimateWNA, which is also a close match to local climate data, but that more confidence
should be placed on temperature-based constraints than any which depend on
precipitation.

2.5.2

The Ecosystem-Based Climatic Envelopes

The focal ecological units explored in this study are based on BC BEC forest ecosystem
units. The CDF zone was extended into the US to capture the full distribution of this
ecosystem within the PNW. This approach follows up on previous work of Hamman and
Wang (2006) which found that the south end of Vancouver Island quickly is excluded from

60

CDF SCS in future projections, and the projected climatic conditions in the geographic area
do not match the SCS for current BC-defined dry coastal ecological units. There is an
international correlation project for ecosystem classification in progress that will facilitate
future climatic envelope modelling, but there remains much work to be completed in this
area of study (C. Cadrin pers.comm., March 9,2011; D. Meindingerpers.comm., March 9,
2011). When completed, the climatic envelope definitions will include the confirmed extent
of ecosystems across the border.
The ecosystem classification extension used here is primarily based on slope position and
tree species composition. The climatic envelope for the extended CDF is broader than the
BC CDF climate range, with an increase in the range of all climate variables, particularly the
growing degree days (DD>5), resulting in the inclusion of an additional vegetation
assemblage that occurs on warmer sites, but remaining adequately similar in tree species
composition to meet the overall definition of a CDF zone ecological unit.
The projected distribution of the extended CDF is similar to established climate zones
(Koppen 1923), continent-wide Ecoregion mapping (Bailey 1997), and more recently
developed CDF bioclimatic envelope models built based on the BEC zones and Random
Forests machine-learning decision tree modelling methods (T. Wang pers. comm. March 28,
2010). The dry coastal ecosystems fall within the Subtropical Dry Summer and/or Humid
Subtropical Climate zones (Koppen 1923). These climate zones are further divided into
ecoregions, including the Marine Mixed Forests (WA and OR lowland areas) and the Marine
Mountains Mixed and Coniferous Forests and Meadows (WA and OR coast, Vancouver
Island, and adjacent mainland BC; Bailey 1997).

61

Other researchers have projected CDF climatic envelopes into the PNW US based on the
distribution of the CDF within BC (T. Wang, pers. comm., March 28, 2010). The baseline
extended CDF SCS is similar to Wang's projections with BC in that they are distributed over
the same area on Vancouver Island, small islands of the Salish Sea, and the adjacent
mainland. In the PNW US, the two projections of suitable climate space overlap, however
the extended CDF SCS, created using US ecological plot data, includes a larger area and
extends beyond Wang's projections which are limited to the lowest elevation portions of
the Puget Trough and the Willamette Valley.
Although the objective to extend the CDF to the south is met and supported by other
studies, future projections continue to result in an undefined SCS zone on southern
Vancouver Island using this classification and model methodology. This is similar to the
findings of Hamman and Wang (2006) as well as recent regional climatic envelope work by
Campbell and Wang (in press), all of which resulted in some anomalous (undefined) climatic
units with future projections. As these "anomalous" ecological units may nonetheless have
analogues in more southerly locales, these findings support the efforts for international
ecosystem classification for research and management for future climate scenarios, but also
for further research on climate and vegetation relationships to determine climatic
thresholds for ecosystem distributions.
Baseline distribution of the extended CDF SCS is very close to the BEC v.7 mapping in BC.
The baseline CWHxm SCS covers a larger area compared to the BEC mapping and it includes
the most of the mapped CDFmm, only excluding the southern tip of Vancouver Island. This
suggests that the distribution of the CDFmm could be limited by climate-driven conditions

62

(e.g., fire; BC MOF 1995, Wong et al. 2003), whereas the distribution of the CWHxm is
limited by other factors (e.g., dispersal, competition, disturbance, resources) not captured
using the climatic envelope approach.
The core area (25th - 75th percentile) occupies, on average, 5% of the area occupied by
the full extent (minimum-maximum) of the climatic envelope for both ecosystem units. The
extended CDF core is completely in the US while the CWHxm is distributed throughout BC
and the PNW US. The core percentile areas for the extended CDF and CWHxm are mostly
out of the currently mapped CDF and CWHxm, confirming that BC is the northern climatic
extent of those ecosystems. This suggests that within the Vancouver Island study area, the
dry coastal ecosystems may be more resilient to climatic change initially if climate regimes
are shifting northward and established vegetation assemblages are able to persist in
southern coastal areas with climate similar to the projected climate for Vancouver Island.
Future distributions of the extended CDF SCS are within close proximity of the current
CDFmm distribution up until the third future timeslice (2070-2099), but the climate
conditions may be more like the PNW US as this area becomes the southern extent of the
baseline projection of CDF SCS. The CWHxm SCS moves northward, mostly out of the
current mapped CWHxm by the end of the century. These results are comparable to those
of Hamann and Wang (2006) that projected decreases in SCS for both the CDF and CWH BEC
zones within BC, with greater overall changes in the CDF zone when compared to the CWH
zone in BC. Though the trends are similar, these results are not directly comparable as the
scale is not the same (Hamman and Wang 2006 modelled BEC zones while this study models
an extended CDF zone and the CWHxm subzone).

63

More recent projections of the CDF climatic envelopes in the 2040-2069 timeslice were
developed for the southeastern side of Vancouver Island based on current BEC mapping and
two individual GCM CES. Their results suggest either a decrease of 15.5% in CDF suitable
climate space using the CGCM3 A2 or an increase of 13.7 % CDF suitable climate space using
the HadGEMl A1B (Campbell and Wang in press). The mapped distribution of the results are
similar to the area of extended CDF persistence (Figure 2-14) where the SCS on
southeastern Vancouver Island persists into the 2040-2069 timeslice.
The question of whether or not ecosystems as we know them today will persist has been
asked for decades. Climate change projections in the study area in the 1990s suggested that
southwestern BC (i.e., the CDFmm) may not support forest stands in the future, and that
other ecosystems would thrive and expand (e.g., the CWH) (Burton and Cumming 1995).
Brown and Hebda (2002) discuss the origin and potential future of biodiversity on southern
Vancouver Island, and note that southern Vancouver Island was once a forest composed of
western hemlock and pine woodland (during the early late-glacial interval). Then the pine,
spruce, fir, western hemlock, and mountain hemlock replaced the pine woodland (late lateglacial interval). As fires became more common, Douglas-fir forests became dominant on
mesic and xeric sites, while the western hemlock dominated stands remained on the hygric
sites (Brown and Hebda 2002). Though it seems that the composition of forests has not
changed much over 12 000 years, increases in technological resources is increasing our
ability to detect small-scale change. With results suggesting less than 15% of the suitable
climate space for the CDF remaining at the end of the century, increased research and
monitoring are required to inform land management decision makers.

64

2.5.3

Areas of Overlap, Persistence, and Shift of Suitable Climate Space

As previously noted, the baseline SCS for the extended CDF is overlapped by the CWHxm,
suggesting that there is a large climatic ecotone or transition area between the two
ecosystem units. The SCS occupied by both ecosystem units (i.e., the "overlap area")
increases in 2010-2039, though both the extended CDF and CWHxm SCS areas decrease,
suggesting climate conditions are changing within the shared climate tolerance of the
extended CDF and CWHxm SCS. Then, as climate conditions are projected to change further
over time, the overlap area is expected to decrease while the extended CDF and the
CWHxm SCS both decrease in area and shift northward on Vancouver Island. It is projected
that there will be less than 1 % of the overlap area remaining at the end of the century,
suggesting that the two ecosystem SCS will be diverging in space, potentially resulting in
vegetation assemblages less similar in composition and structure and more distinct from
one another.
The areas of ecosystem persistence for extended CDF and CWHxm SCS are projected to
be greatly reduced by the end of the century. The projections based on the five-model GCM
ensemble suggest that, over time, conditions could remain suitable for both ecosystems
until 2040-2069. The 2070-2099 timeslice conditions are projected to result in a loss of
> 5 000 km2 from the extended CDF SCS and > 20 000 km2 from the CWHxm SCS. As stated
by Rose and Burton (2011), the areas of persistent SCS can be considered climate refugia
and can be important areas to consider for conservation planning.

65

The SCS shifts from CWHxm to the extended CDF, suggesting that the extended CDF
could occupy areas of higher elevation and latitude, specifically areas adjacent to the
CDFmm and currently mapped as CWHxm (BEC v.7) on Vancouver Island. As this shift in
ecosystem suitability occurs, the total SCS for both ecosystems is reduced, suggesting that
the climate conditions are diverging from climatic envelope definitions. As areas now
characterized by the extended CDF shift to the undefined SCS climate, the area currently
mapped as CDF will almost totally shift into a different, and undefined, climatic envelope.
To explore how climate variables are diverging from the climatic envelope definitions and
to better understand the undefined SCS, the component changes in projected CDF
suitability were tracked over time (Table 2-12). Initially, the FFP lengthens beyond the
baseline normals (2010-2039), then the MWMT, TD, NFFD all increase and the PAS
decreases beyond the baseline climatic envelope. By the end of the century, the MAT and
DD>5 values are higher than the baseline range, resulting in the need to assign some SCS to
a new or currently undefined climatic envelope.
This evaluation of the space that was once suitable for the extended CDF suggests that
the currently mapped CDFmm region may experience an increase in temperatures, with the
precipitation amount remaining the same in quantity, but some of it shifting from snow to
rain. These findings are supported by the results of other climatic envelope mapping on
Vancouver Island that suggest the current CDFmm could shift to a warmer ecosystem type
within this century (Jones and Brown 2008). This suggests that water storage will be even
more important for plants to make it through the longer and warmer growing season.

66

The extended CDF includes the southern extent of the CDF forest zone, and an additional
subzone provisionally determined to be drier than the CDFmm (therefore called the
CDFdm). Projections suggest that most of the area currently supporting the CDF could shift
beyond even the climatic conditions of the CDFdm, potentially resulting in climatic
conditions suitable to an ecosystem unit more like the mixed Pseudotsuga menziesii and
Pinus ponderosa forests or the Quercus woodlands of the Willamette Valley, OR (Franklin
and Dyrness 1988).

2.5.4

Constraints on the Extended CDF Climatic Envelope

The constraint analysis on baseline SCS for the extended CDF on Vancouver Island
suggests that the SCS is limited by different climate variables along the latitudinal gradient.
The southern end of the ecosystem is most limited by moisture availability (AHM, SHM,
CMD) while the northern end is most limited by (cooler) temperature indicators (NFFD, FFP,
EMT). This suggests that ecosystem polygons are bound by a variety of limiting factors that
shift in significance throughout the spatial distribution of the ecosystem. The climate
variables on either side of the CDF boundary were investigated and findings suggest that if
temperature increases and precipitation decreases (particularly in the growing season), the
zonal CDF SCS could shift to the north and west on Vancouver Island. Increased
representation by plant species of CDF zonal ecosystems in these areas is possible given
that many of these species currently occur on dry sites of the adjacent CWHxm ecosystem.
Further investigation of species distributions may provide insight to the species-specific
climate constraints to ecosystem distribution.

67

2.5.5

Limitations of the Climatic Envelope Technique

The climatic envelope projections were developed using the ClimateWNA interpolation
tool. This tool enables scale-free climate projections, but biases in the outputs, such as the
distribution of climate station input data and the uncertainty associated with the GCM input
data were considered for in the utility of the tool output data. The selection of the the five
GCMs for the five-model GCM ensemble was based on the GCMs most representative of the
baseline climate normals for the study area. Though the five-model GCM ensemble most
closely represents the current climate normals for the study area, the uncertainty around
the projections increases over time slices (e.g., the further into the future, the more
uncertainty). This is one of many potential combinations of GCMs that could be applied to
explore the potential future climate regime for the study area.
Climatic envelope models are static representations of the baseline climatic extremes
associated with the geographic range of ecosystems, but with no consideration of dynamic
interactions. Envelope models assume a steady state or equilibrium relationship while
adaptation, dispersal, and physiological responses to changes in atmospheric conditions,
and inter-species interactions are not considered (Berry et al. 2002, Pearson et al. 2002).
Envelope models are used to determine fundamental niche space without the influence of
localized land-use and microclimatic factors. The shifts in ecosystem SCS may be beyond the
ability of component species to migrate and therefore, predicted SCS is often broader than
the actual SCS distribution (Berry et al. 2002). Climatic envelopes are useful to inform areas

68

of persistence and ecosystem shift as well as the climatic constraints on ecosystem
boundaries, with the understanding of the limitations of the approach.

2.6

CONCLUSIONS AND RECOMMENDATIONS

Overall, results of climatic envelope models using a five-model GCM ensemble and an
extended classification and distribution of the BC CDF forest ecosystem indicate that
distributions of both the extended CDF and CWHxm dry coastal ecosystems could shift
northward in the next century. The SCS for each BEC zone seems to shift further apart
spatially, suggesting that the current ecosystem transition zone could become wider and
may result in climatic conditions not currently found on Vancouver Island (or within the
PNW US). The results also suggest that there are potential areas of ecosystem persistence
and ecosystem shifts on Vancouver Island that could be included in a network of climate
change monitoring.
Outcomes from this project suggest that international correlation of ecosystem
classification not only has applications to management and current understanding of the
forest ecology in dry coastal ecosystems, but could provide opportunities to explore climate
and ecosystem relationships along the southern border of BC. Trends in ecosystem SCS may
provide a framework for further climate change modelling work that combines climatic
envelope modelling and other ecosystem drivers, such as natural and anthropogenic
disturbance regimes. To support further climate change research, a network of monitoring
transects is recommended. By monitoring changes in climate, soil conditions, and
vegetation composition, the data required for adaptive management in dry coastal

69

ecosystems will be available. Results presented here suggest that the south end of the CDF
on Vancouver Island is a good candidate for monitoring because it is the area of most rapid
change in SCS. The central and northern portions of the CDFmm include important areas of
persistence for conservation objectives. This is important research in the dry coastal
ecosystems where timber, recreation, and urban development demands on these
ecosystems are challenging current decision-making processes for land and natural resource
management.

70

3.

Species Climatic Envelope Models: Current and Future Distribution of
18 Species Common to the Coastal Douglas-fir zone

3.1

ABSTRACT

Climate change is impacting the distribution of species. The ranges of 18 diagnostic
species of the rare Coastal Douglas-fir moist maritime (CDFmm) subzone in British Columbia
(BC) are explored using climatic envelopes for current and future time periods. Though
many of these species have the potential to shift northward up the Pacific Northwest coast
over time, the distribution of the 18 species within the southwestern BC study area is
projected to remain relatively unchanged. Seven species (Abies grandis, Gaultheria shallon,
Lonicera ciliosa, Achlys triphylla, Hylocomium splendens, Kindbergia oregana, and
Rhytidiadelphus triquetrus) are expected to lose suitable climate space from the currently
mapped CDFmm. Results of a sensitivity analysis suggest that geographic limits to the
distribution of the study species are most constrained by precipitation inputs and the
characteristic summer moisture deficit for the Coastal Douglas-fir forest. The climatic
envelope for Cornus nuttaliii constrains the CDFmm species overlap area the most,
suggesting it to be the limiting species for the CDFmm zonal plant community as currently
defined, making it a potentially sensitive indicator of shifts in climate regime. The modelled
range of combined species distributions is greater than that modelled for the modal CDF
ecosystem as a whole. Results suggest individual species patterns vary with changes in
climate and that the composition of the CDFmm zonal plant assemblage may change over
time. This research provides guidance for conservation and sustainable resource planning
71

and a theoretical foundation to direct hypotheses for further research into climate, soil, and
vegetation relationships.

3.2

INTRODUCTION

3.2.1

Climate and Vegetation Relationships

3.2.1.1 Individual species and plant assemblages
The primary driver of the geographic distribution of plants is generally considered to be
climate (Forman 1964, Box 1995). Significant shifts in climate have drastically altered the
distribution of plant species in the past (Brown and Hebda 2002, Delcourt and Delcourt
1988, Williams et al. 2005). Species distribution models have been developed to explore the
potential impacts of climate change on flora and fauna around the world (Franklin 1995,
Guisan and Zimmerman 2000, Skov and Svenning 2004, Morin and Thuiller 2009) using a
variety of approaches (Iverson and Prasad 2001, Pearson et al. 2002, Hamann and Wang
2006, McKenney et al. 2007), yet all approaches suggest reorganization of plant
communities accordingly. These studies assume that macroclimate is a major determinant
of plant distribution, though it is recognized that other factors, such as terrain and soils
(abiotic) and competition and genetic adaptation (biotic), also influence species ranges at a
local spatial scale (Pearson and Dawson 2003, Skov and Svenning 2004).
Plant communities (also termed plant associations) are recognized, recurrent
combinations of vegetation across the landscape (Whittaker 1960, Mueller-Dombois and
Ellenberg 1974). The fundamental level of organization and environmental response (and

72

hence the distribution, succession, and evolution) of ecosystems and their components is a
discussion topic within the discipline of ecology that was initiated almost a century ago,
often presented as a dichotomous choice between predicted patterns of individualistic and
community units (Gleason 1926, Whittaker 1951). The argument that individual species
patterns are independent was posed by Gleason (1926), while the plant community concept
was championed by Clements (1916). Often, studies supporting either the individualistic or
the community perspective are proven by falsifying the opposing perspective (Gleason
1926, Curtis 1955, McIntosh 1967). But more recent literature suggests that there are more
than the two options to explain vegetation patterns occurring across environmental
gradients (Whittaker 1967, Shipley and Keddy 1987) and that communities are determined
by complex interactions and are therefore functioning both as individuals and as mutually
influenced assemblages of species (Lortie et al. 2004). Over the course of time, species have
redistributed independently, forming new plant assemblages (Webb 1987, Williams et al.
2005) and the impacts of climate change will likely result in additional independent changes
in distribution (Iverson et al. 2009). In this chapter, the climatic gradients for species and
ecosystems are explored using climatic envelopes to examine and compare the projected
patterns of individuals and plant assemblages over time.
The Biogeoclimatic Ecosystem Classification (BEC) system is a hierarchical classification
used to define and interpret ecosystem structure and distribution in British Columbia (BC).
The climax successional theory (Clements 1936) is the basis of the BEC system. This theory
assumes that at mature and old-growth successional stages, forest ecosystems have
developed to a self-maintaining "climax" state of vegetation at equilibrium with the site,

73

soil, and climate features of that location. The zone and subzone levels of the hierarchy are
primarily defined by climate conditions. Variants and site series are finer classification units
that consider floristic, site, and soil features within a climate zone/subzone. Within the BEC
system, subzones are defined by delineating the boundaries of ecosystems based on zonal
sites with average soil moisture and soil nutrient conditions for that subzone. Based on
these average conditions, climate is the primary driver of ecosystem development, rather
than site, soil, or terrain features (Pojar et al. 1987, McKinnon et al. 1992). The 18 species
analyzed in this chapter (see common and scientific names in Appendix F) occur together in
mature and old-growth forests and are considered diagnostic of zonal sites in the Coastal
Douglas-fir moist maritime (CDFmm) subzone currently located on the south coast of BC.

3.2.1.2 Species distribution, climate change, and model projections
There are many complex interactions between climate, site, and vegetation on the
landscape. The number of modelling publications projecting future vegetation patterns
increased rapidly in the mid-1990s (Drew et al. 2011). The availability of climate data and
improved climate models is facilitating development of species distribution models and
research in climate and vegetation relationships and the climatic niche (Franklin 2009). This
chapter develops species-specific climatic envelopes for all 18 species diagnostic of zonal
sites in the Coastal Douglas-fir moist maritime (CDFmm) subzone forest on the south coast
of BC. The climatic envelope is an empirical definition of the climatic conditions acceptable
for a species, based on the current spatial extent and associated climate variables for the
species' current locations. The minimum and maximum values define the most generous

74

climatic envelope definitions, while various percentiles (e.g., 5th to 95th or 25th to 75th) of
climate attributes describing the full range of current distributions are calculated to express
the confidence and core area of the suitable climate space projections. Climatic envelopes
are constructed and evaluated using geographic information system (GIS) and statistical
software.
The climatic envelopes are projected for baseline (1961-1990) and future (2010-2039,
2040-2069, and 2070-2099) timeslices using 30-year climate normals. The future projections
use a five-model global climate model (GCM) ensemble created by calculating the mean of
five GCMs best suited to characterize the baseline climate of the study area (Chapter 2). The
general approach is similar to that of Chapter 2, but individual species are modelled rather
than ecosystem units. As in Chapter 2, the suitable climate space (SCS) for a species is a
modelled geographic area that is within the defined range of the climatic envelope for that
species.
Climate data collections are becoming more frequent across the landscape with
improved climate sensor and data logging technology. However, the climate data collection
is skewed to populated areas and is not evenly distributed across the diverse landscape of
the southwestern coast of BC. The ClimateWNA computer program (ClimateWNA v.4.52,
Wang et al. 2010) is a freely available tool that provides interpolated climate data for
baseline normals and interpolated GCM data for future time periods across the diverse
landscape of western North America. When modelling future climate, a range of projections
is expected because GCMs are each based on different assumptions and selected
parameters to describe climate behaviours. The ensemble approach for exploring multiple

75

GCM and carbon emission scenario (CES) combinations captures the range of likely climatic
changes (Berry et al. 2002) and is expected to reduce the overall statistical error in
simulated climate (Gleckler et al. 2008). This study employs a five-model GCM ensemble
based on five GCM and CES combinations that best describe the baseline climate of the
study area. Climatic envelope models, ClimateWNA, GCMs, and the five-model GCM
ensemble are all detailed in Chapter 2.
An advantage to the climatic envelope modelling approach is that species presence and
absence of data are sufficient to build the model (Pearson et al. 2002). Because of the
abundance of documented presence location data available for the study species, this
component of the thesis focuses on climatic envelope model development and projects the
expected distribution of SCS (Rose and Burton 2011) for each diagnostic species. Because
species may respond to climatic change in different ways, climatic envelopes are
appropriate to predict the potential SCS for a species.
Climatic envelope models are static representations of the baseline climatic extremes
associated with the geographic range of a species or other ecological feature, and lack any
dynamic interactions. The climatic envelope approach is based on ecological niche theory.
The fundamental ecological niche represents the environmental conditions within which a
species can survive (Hutchinson 1957). The realized niche represents the space within which
a species actually lives and survives, within the confines of biotic interactions (Austin et al.
1990, Guisan and Zimmermann 2000). Climatic envelope models are used to determine the
climatic niche, the climatic component of the fundamental niche space without the

76

influence of localized land-use and microclimatic factors (Pearson and Dawson 2003, Lo et
al. 2010).
Climatic envelope models assume a steady state or equilibrium relationship, while
adaptation, dispersal, physiological response to changes in atmospheric conditions, and
inter-species interactions are not considered (Berry et al. 2002, Pearson et al. 2002). The
shifts in species SCS may be beyond the species' ability to migrate and therefore, a
predicted SCS is often broader than the actual distribution of a species (Berry et al. 2002).
One approach to understanding the limitations of the climatic envelope approach is to do a
sensitivity analysis in which one climate variable is increased or decreased while the other
predictor variables are held constant. The impact of climatic variability and subsequent
species response can be quantified using this technique (Hamann and Wang 2006,
Heikkinen et al. 2006).

3.2.2

Research Objectives

A series of climatic envelope models were developed to explore the potential impacts of
climatic change on diagnostic plant species currently distributed in the study area, and to
compare the results of combined projections for individual species and those derived for
entire plant assemblages in Chapter 2. A sensitivity analysis was also conducted to explore
the impact of variance of GCM outputs and to determine the climate variables that most
constrain the distribution of each species within the study area. There are three research
objectives for this chapter: (1) determine the baseline (1961-1990) and future (2010-2039,
2040-2069, 2070-2099) SCS for 18 diagnostic species of the CDFmm zonal site series using a

77

five-model GCM ensemble; (2) evaluate individual species SCS distributions by completing a
sensitivity analysis of each climate variable for the baseline climatic envelopes, and (3)
project the overlap area for diagnostic species over time and compare this to the current
distribution of the CDF BEC v.7 mapping (BC MFR 2008) on Vancouver Island and projected
CDF ecosystem climatic envelopes (Chapter 2).

3.3

METHODS

3.3.1

Study Area

Climatic envelope projections were generated for the Pacific Northwest (PNW) of the
USA (western Washington and Oregon) and southwestern BC (latitude 42.2° - 51.9°; Figure
3-1 inset) to review the anticipated northerly shifts with climate change over time (Skov and
Svenning 2004, Hamann and Wang 2006). The area of most interest for projecting the
potential redistribution of species was the southwest coast of BC (latitude 48.3°N - 51.9°N;
Figure 3-1), specifically areas currently in or adjacent to the CDFmm on Vancouver Island.

78

Figure 3-1: The southwest British Columbia (BC) study area (Vancouver Island, small islands of the
Salish Sea, and adjacent mainland BC coast) with the full Pacific Northwest coast (west coast of
Washington and Oregon and the BC study area) domain area inset.
The Coastal Douglas-fir moist maritime (CDFmm) subzone and its distinctive ecosystems
are currently distributed on the southeast side of Vancouver Island, on adjacent smaller
islands within the Salish Sea, and a small area of mainland BC (Figure 2-3). This ecosystem is
unique in BC and Canada as it is the driest ecosystem on the Pacific coast. This ecosystem is
in the rain shadow of the Olympic Mountains and the Vancouver Island Mountains resulting
in a soil moisture deficit during the growing season (Nuszdorfer et al. 1991). The adjacent
biogeodimatic unit is the driest of the largest BEC zone on the coast, the Coastal Western
Hemlock very dry maritime (CWHxm) subzone. For more information on the CDFmm and
the adjacent CWHxm subzone, see Chapter 2. Interest in exploring the CDFmm is generated
by potential increases in temperature and shifts in precipitation regimes that could alter
growing conditions (e.g., the soil moisture deficit) in the currently mapped CDF forest

79

ecosystem. The adjacent CWHxm mapped unit could potentially become more similar to the
CDFmm (Chapter 2).

3.3.2

Data and Analysis

3.3.2.1 Species selection and location data
The BEC system uses a combination of species abundance and occurrence data to
evaluate diagnostic vegetation for ecosystems. Each site series within a BEC subzone has a
vegetation summary table to assist in the identification and characterization of the site
series unit (MacKinnon et al. 1992). The 18 species evaluated in this study are those that are
listed as characteristic species on the vegetation summary table for the zonal site series
within the CDFmm subzone (Green and Klinka 1994). These species are also found in other
ecosystem types (e.g., the adjacent CWHxm), but when found together on a site, they are
used to identify the zonal site of the CDFmm. Table 3-1 is an excerpt of the vegetation
summary table of selected species that best characterize the zonal site series of the CDFmm
and CWHxm based on the plot data used in the publication of the BEC system on the coast
(Green and Klinka 1994). Note that some, but not all, species are also found in the adjacent
CWHxm subzone. The species relative prominence value (represented by the size of the
black bar) for the CDFmm and CWHxm zonal site series (Table 3-1) is a combined measure
of mean cover and mean frequency of a species' occurrence in the ecological plot data and
it is used to describe the relative values of species in a site series (Green and Klinka 1994).

80

Table 3-1: Summary of vegetation (scientific and common name listed by tree, shrub, herb, and
moss strata) that characterizes the zonal sites of the Coastal Douglas-fir moist maritime (CDFmm)
and the adjacent Coastal Western Hemlock very dry maritime (CWHxm) subzones*. Prominence
bars measure the combination of cover and frequency of occurrence.
Scientific Name

Common Name

CDFmm

CWHxm

Abies grandis

Grand fir

■ 1

Acer macrophyllum

Bigleaf maple

■

Arbutus menziesii

Arbutus

1

Pseudotsuga menziesii

Douglas-fir

Thuja plicata

Western redcedar

Tsuga heterophylla

Western hemlock

Comusnuttallii

Western f loweri ng dogwood

Gaultheriashallon

Salal

Holodiscus discolor

Ocean spray

Loniceraciliosa

Western trumpet honeysuckle J ______

Mahonia nervosa

Dull Oregon-grape

Rosa gymnocarpo

Baldhip rose

■

Symphoricarpos spp.

Snowberry

■

Vaccinium parvifolium

Red huckleberry

Achlys triphylla

Vanilla leaf

Linnoea borealis

Twinflower

Polystichum munitum

Sword fern

1

■

Pteridium aquilinum

Bracken fern

■

■

Hylocomium splendens

Step moss

■ _____ w m

Kindbergia oregano

Oregon beaked moss

■

Rhtidiadelphus loreus

Lanky moss

■

Rhytidiadelphus triquetrus

Electrified cat's tail moss

■ ■
m

m

m

1
■ 1

■

1

■
■

■

*Based on the CDFmm and CWHxm Vegetation Table (page 76 and 106) from Green and
Klinka (1994).
Species location data were compiled from multiple data sources in BC and the PNW US.
BC sources included the BEC database plot data (BC MFR 2009) and Terrestrial Ecosystem
Mapping plot data (TEM, BC EcoCat 2010). US sources included the US Forest Service (USFS)

81

Ecological Plot data available on the Ecoshare website (http://ecoshare.info/category/datasets/) for Washington and Oregon and a compilation of USFS and research ecological plot
data for California (J. Thorne pers. comm. December 27, 2010). The Jepson Herbaria
database (http://ucjeps.berkeley.edu/) at the University of California and the Consortium of
Pacific Northwest Herbaria database (http://www.pnwherbaria.org/) were additional
sources for data in Alaska, BC, Washington, and Oregon. Most of the ecological plot data
were collected during the same time period as the baseline climate normals (1961-1990). In
total, data from >33 000 plots or collection locations situated between Alaska and California
were utilized for the development of climatic envelope models (Figure 3-2).

1

750

1,500 Kilometers

Figure 3-2: Map of western North America with location points for all 18 diagnostic species of the
Coastal Douglas-fir moist maritime subzone used as inputs to the climatic envelope definitions (total
number of plots > 33 000).

82

Species location data were limited to plots in mature and old-growth stands, when
possible, in keeping with the climax-based development and interpretation of BEC
ecological units. One exception to this was the use of herbarium data. These data are
species occurrence data without reference to stand age. The species climatic envelope
definitions were developed using the extent of each species' range from Alaska south to
California (data permitting) in an attempt to fully capture the entire Pacific Northwest
climatic envelope for each species. Without a complete data set for current species range
(e.g., if only using BC data), extrapolation to predict future species ranges could result in
overprediction and an inability to simulate southern range limit (Hamann and Wang 2006).

3.3.2.2 Species baseline and future climatic envelope models
The climatic envelopes for individual species were defined using the compiled geo­
referenced location data and associated climate variables output from ClimateWNA for the
baseline (1961-1990) timeslice. For each species, the location data were input to
ClimateWNA and annual climate variables for those locations were generated as output. Of
the 20 annual climate variables generated from ClimateWNA (Appendix B), 16 were used to
define the climatic envelope for each species (as per Chapter 2). The climate variables were
selected based on biological relevance and temporal flexibility. The climate variables fixed
to a specific calendar date were rejected from the definitions to accommodate potential
seasonal shifts within the calendar year over time. The minimum, maximum, and selected
percentile ranges ( l st-99th, 5th-95th, 25th-75th) of each climate variable were determined for
each species under observed "baseline" (1961-1990) conditions using R statistical software

83

(R Core Development Team 2010). The minimum and maximum values are used to create
the climatic envelope for the full extent of climate conditions for each species. The l st-99th
and 5th-95th percentile values are used to evaluate confidence and the 25th-75th percentile
values are used to define the core climatic envelope for each species.
To project the climatic envelope, a 1-km grid of locations was created across the study
area, for which climate attributes generated by ClimateWNA at each node were evaluated
for concordance with climatic envelope definitions using a series of conditional statements
in R statistical software (R Core Development Team 2010). The 1-km grid covers the Pacific
Northwest (as in Chapter 2). The conditional statements were constructed using the
minimum and maximum values for each of the 16 climate variables for each species. As per
Rose and Burton (2011), points were identified as SCS when conditions are met for each
timeslice. Suitable climate space was defined for each species under the baseline (19611990) climate timeslice using ClimateWNA. The three future (2010-2039, 2040-2069, 20702099) timeslices were projected using the five-model GCM ensemble created using the
mean of the best five GCM and scenario combinations for the study area (as determined in
Chapter 2).

3.3.2.3 Sensitivity analysis fo r the baseline species climatic envelopes
The sensitivity test was conducted to better understand the limitations of the climatic
envelopes for species. The sensitivity test involved an adjustment of each climate variable
(one at a time) to evaluate the influence of that particular variable on a species' climatic
envelope distribution. The climate variable adjustment was calculated based on the

84

baseline climate for the southwest BC study area. The adjustment consisted of increasing or
decreasing the range in each climate variable by 10%. The baseline climatic envelope was
run for each species 32 times: 16 times for the increased adjustment and 16 times for the
decreased adjustment for each climate variable, one at a time. The change in area (km2)
was calculated for each adjusted SCS and is reported in the summary table (Table 3-4).

3.3.2.4 Comparison o f the CDFmm ecosystem mapping to modelled species overlap and CDF
climatic envelopes in the British Columbia study area
The SCS for each of the 18 species diagnostic of the CDFmm were evaluated for each
timeslice to determine the geographic area occupied by all 18 species climatic envelopes.
The points on the 1-km grid were each evaluated for climatic suitability for all 18 species
within an individual timeslice. The suitable locations were referred to as the species overlap
area, representing the potential SCS for all 18 individual species. The overlap area was
compared to the current CDFmm BEC v.7 mapping (BC MFR 2008) and to the SCS
distribution for zonal plant communities of the CDF developed in Chapter 2. Hence, the
distribution of zonal plant communities for the CDF was compared on the basis of three
alternative methodologies: (1) current BEC v.7 mapping (BC MFR 2008); (2) species overlap
area; and (3) the SCS created for the CDF as a whole, as presented in Chapter 2.

3.4

RESULTS

3.4.1

Species Climatic Envelopes: Current and Future Distributions

85

Climatic envelope definitions for each species and the overlap area are based on the
minimum and maximum values for each of the 16 annual climate variables (Table 3-2).
Generally, the two fern species (Polystichum munitum and Pteridium aquilinum) have the
largest range in climatic envelope definitions, while two shrub species, Cornus nuttallii and
Lonicera ciliosa, and one moss species, Kindbergia oregana, have the narrowest range in
climatic envelope definitions.
The Cornus nuttallii climatic envelope definition does not include points where the mean
annual temperature (MAT) drops below freezing (the minimum value is 2.1 °C). Acer
macrophyllum has the warmest MAT and the mean warmest month temperature (MWMT)
values at the maximum end of the climatic envelope definition. Abies grandis, Achlys
triphylla, and Kindbergia oregana have the highest minimum mean annual precipitation
(MAP) values (> 400 mm).
The overlap area has the narrowest climatic envelope definition because it is defined
using the greatest value in the minimum and the lowest value in the maximum range of all
18 climatic envelope definitions. The overlap area describes the climate conditions suitable
to all 18 species.

86

Table 3-2: The climatic envelope definitions using the minimum (min) and maximum (max) values for the 16 annual climate variables* for the 18
diagnostic species of zonal sites in the Coastal Douglas-fir moist maritime subzone.
MWMT(°C)

MCMT(°C)

Plant species

min

max

min

max

min

max

min

max

min

Abies grandis
Acer macrophyllum

-13.5
-1.2

13.5
20.1

-3.9

-21.4

9.3

441

-11.6

14.1

5.3
5.3

25.4

10.8

21.3
30.1

24.9

194

Arbutus menziesii

-14.7

17.7

5.1

25.9

-33.1

13.8

5.6

38.2

166

Pseudotsuga menziesii

-18.0

16.0

-11.3

25.3

-33.1

11.2

5.5

38.2

Thuja plicata

-7.5

16.0

3.0

25.3

-13.7

9.2

6.8

26.9

166
288

Cornus nuttallii

2.1
-14.7

17.4

12.4

26.3

-8.1

11.7

8.2

24.8

13.9
18.1

5.1

-33.1
-33.1

8.7

6.5

38.2

0.8

22.3
27.2

14.3

5.3

38.2

15.5

4.5

-15.1

7.2

10.6

31.3

14.9
16.7

5.1

25.9
24.2
25.4

-33.1

9.7

5.5

26.8

-33.1
-21.2

11.1
11.7

Gaultheria shallon
Holodiscus discolor
Lonicera ciliosa

-14.7

Mahonia nervosa

-7.6
-14.7

Rosa gymnocarpa
Symphoricarpos albus

-14.7
-14.7

17.2

Achlys triphylla
Polystichum munitum

-13.9

13.3

-17.2

Pteridium aquilinum

0.1
-4.0

TD(°C)

MAP (mm)

MAT (°C)

AHM

SHM

max

MSP (mm)
min
max

min

max

min

max

3880
5474

48
16

970

-3.1

-11.7

860

3.5

45.6
149.8

19.5

308.3
1393.8

4090
5474

16

636

-28.3

88.8

22.6

1317.6

1170

-28.3

74.2

-41.9

1200.0

6054

15
27

1450

2.4

54.9

9.2

659.3

307

4855

21

4.2

77.5

22.6

1152.4

166
166

5866
5044

30
14

720
1299

44.3

719

-28.3
-28.3

91.4

11.0
2.7

613.3
1375.0

3598

71

574

2.8

79.2

24.2

304.2

38.2

240
166

5646

166

90.2

0.5

804.0
1158.8

5.4

33.8

178

5323
5587

-28.3
-28.3

15.1

38.2

897
964

53.6

5.5

19
17
13

1382

-8.2

96.5

-26.5

1800.0

22.0

-21.8

9.0

5.8

24.8

424

6054

30

979

-5.8

43.6

-8.3

17.8

-0.9
-5.7

26.2

-33.1

14.0

4.2

38.2

166

6054

15

1431

-28.3

91.7

-11.9

690.0
1311.8

-22.5

18.6

-15.6

25.7

-22.9

14.4

5.3

29.3

5474

15

1258

-7.5

134.5

-31.5

1406.7

Hylocomium splendens

-14.8

11.9

2.3

21.4

-34.3

5.3

9.1

46.3

165
184

6880

79

1737

-12.5

58.2

6.8

234.5

Kindbergia oregana
Rhytidiadelphus
triquetrus

-0.4
-10.4

14.8
12.3

10.6
2.3

22.4
22.2

-12.4
-28.1

8.5
6.7

9.1
9.3

25.1
44.2

420
276

5866
5299

91
68

1448
1227

2.9
-1.0

44.3
58.2

10.3
8.5

240.7
279.4

Species overlap area
2.1
11.9
12.4
21.3
-8.1
*AII climate variables and definitions listed in Appendix B.

5.3

10.6

24.8

441

3598

91

574

4.2

43.6

24.2

234.5

87

T a b le 3 - 2 (c o n tin u e d ).

DD<0

DD>5

NFFD

FFP

PAS (mm)

EMT(°C)

CMD

Eref

Plant species

min

max

min

max

min

max

min

max

min

max

min

max

min

max

min

max

Abies grandis

1

4937

136

3180

6

351

34

314

4

2075

-59.1

-6.4

0

1174

0

740

Acer macrophyllum

0

545

5557

71

365

3.8

415

1637

0

1436

138

4695

35

365

365
362

1050 -44.9

0

32
27

0

Arbutus menziesii

1660
5623

0

1009

-62.0

2.8

155

0

1126

Pseudotsuga menziesii
Thuja plicata

0
1

6518
2954

136
136

4074

362
357

27
48

364

1
2

3061
2313

-62.0
-45.7

-3.5

0

0

4080

3
28

1511
1300

-6.0

118

1300

0

939
939

Cornus nuttallii

0

897

782

4589

118

354

327

-37.6

-6.1

494

1034

5623

138

3338

35

355

332

3061

-62.0

155

0

739

Holodiscus discolor

5623

4848
3921

9
34

0

1593

9

-7.0

169

1233

8
7

1205

289

-62.0
-49.4

50

5623

138

3706

35

48
27

1995
1262

0

365
339
357

365

3120

136
136

27

Lonicera ciliosa
Mahonia nervosa

0
5

-5.0
-4.7

1439
1241

0

2

1
4

1115

Gaultheria shallon

72
27

2

1942

-62.0

-4.8

155

1329

0

970

Rosa gymnocarpa

0

5623

136

4326

12

360

27

336
354

1

2377

-62.0

-2.9

30

1443

0

1178

Symphoricarpos albus

0

5330

136

4530

4

360

18

359

1492

-59.5

-4.4

0

1537

0

1192

Achlys triphylla

1

136

3118

8

352

2472

-59.3

1222

0

831

136

4729

4

365

365

0

1608

0

1308

0

8090

136

5008

4

365

28

365

0

-64.0
-59.7

0

Pteridium aquilinum

2616
3659

-6.1
4.1

0

0

31
27

322

Polystichum munitum

5069
6230

0
4

5.3

0

1563

0

Hylocomium splendens

14

6078

136

2652

45

337

2

298

17

2797

0

2
6

1500
4722

606
136

3662
2766

113
74

337
333

60
36

298
283

9
16

2335
1953

0
297
0

931

Kindbergia oregana
Rhytidiadelphus
triquetrus

-67.8 -9.6
-41.6 -9.5
-60.4 -10.5

1203
591

1268
998

0
0

734
695

Species overlap area
14
897
782
2652
118
*AII climate variables and definitions listed in Appendix B.

333

72

283

17

1009

-37.6 -10.5

494

931

8

591

88

344

823

The geographic extent of the locations meeting species climatic envelope definitions
under the four timeslices (baseline and three future projections) as well as the l st-99th, 5th95th, and 25th-75th percentiles were determined for the BC study area (total area = 97 857
km2). The tabular results of the SCS in the BC study area are presented in Appendix G.
Projected trends in the species SCS areas are presented in Figure 3-3. The Acer
macrophyllum, Arbutus menziesii, Pseudotsuga menziesii, Cornus nuttallii, Holodiscus
discolor, Mahonia nervosa, Rosa gymnocarpa, Symphoricarpos albus, and Polystichum
munitum SCS percent occupancy of the BC study area is maintained or increases over the
four timeslices using the five-model GCM ensemble. Interestingly, the l st-99th percentile SCS
for Pseudotsuga menziesii, Cornus nuttallii, Holodiscus discolor, Mahonia nervosa, Rosa
gymnocarpa, Symphoricarpos albus, and Polystichum munitum increase for the first three
timeslices and then decrease at the end of the century (2070-2099 timeslice). Most of the
species SCS areas increase initially, with the exception of Abies grandis and Hylocomium
splendens, for which SCS steadily decreases over time. On average, the area occupied by the
l st-99th percentile envelopes is approximately < 60 % of the full BC study area, and the 5th95th percentile SCS areas are < 30 % of the full BC study area. The core (25th-75th percentile)
area is < 10 % of the BC study area for each species.

89

Acer macrophylium

Abies grandis
• 1st- 99th

"M in - Max"

— — 5 th -9 5 th

—

ls t - 9 9 th

— — S th -9 5 th

100

100
90
n

80

80

< 70

70

>

60

fta

60

50
40

S

40

\
8

30
20

30
20
10

...

0
Baseline

2010-2039

Baseline

2040-2069

2040-2069

20 10 -2 0 3 9

Tim eslice

T im eslice

Pseudotsugamenziesii

Arbutus menziesii
» 1st* 99th

— — 5 th -9 5 th

.......... 2 5 th -7 5 th

100

100

90

E »0

80

2

80

70

<

70

60

•f

60

40

H

40

30

|

30

“ Min - Max1* —

ls t - 9 9 th

— — 5 th -9 5 th

Baseline

2 0 1 0 -2 0 3 9

2 0 4 0 -2 0 6 9

-1 s t - 99th

— — 5 th -9 5 th

« 50

50

20
10

0
20 1 0 -2 0 3 9

2 0 4 0 -2 0 6 9

20 70 -2 0 9 9

Tim eslice

Cornusnuttaiiii

Thujapiicata
•1 s t - 99th — — 5th -9 5 th

100

100
#
5

90
80

70

<

70

ift 60
50
x 40
t 30
8 20
i 100

"I

60

3
6

50
40

*

30

8

20

#

90

*

S 80

£ 10
Baseline

2 0 1 0 -2 0 3 9

20 40 -20 69

Baseline

20 70 -2 0 9 9

20 40 -2 0 6 9

20 70 -2 0 9 9

T im eslice

Gaultheriashailon

Hoiodiscus discolor
• 1 s t - 99th

— — 5 th -9 5 th

"M in - Max*

- 2 5 th -7 5 th

-

1st- 99th

— — 5 th -9 5 th

100

100
90

*

90

80

Z

<

80
70

■f

60

* 70
Ift 60
50
X 40
€ 30
20

a so
tt

40

^

30

»

10

8 20

8

10

0
2010-2039

Baseline

2040-2069

90

2 0 1 0 -2 0 3 9

20 70 -2 0 9 9

Mahonia nervosa

Lonicera ciliosa

• 1 s t - 99th

— — 5 th -9 5 th

• 25th -7 5 th

100
7

90

S

80

X

70

■f

60

S

40

* SO

■s 30
£

S
«

a.

2010-2039

Baseline

2040-2069

2070-2099

2010-2039

T im eslice

Symphoricarposalbus

Rosagymnocarpa

“ 1 s t- 99th

“ 1 s t- 99th — — 5th -9 5 th

— — 5 th -9 5 th

100
*

90

«
«

80 «
70 -i
60

ft 50
a
r

40

*

30

8

20

£

10
0
2010-2039

Baseline

2040-2069

2010-2039

2040-2069

T im e t! ice

T im e t! ice

Achlys triphylla

Polystichummunitum
“ 1st- 99th — — 5 th -9 5 th

“ 1 s t- 99th — — 5th -9 5 th

100

100

90

90

80

80

70

70

60

60

40

40

30

30

20

20

10

10

50

50

0
Baseline

0

2040-2069

2010-2039

Baseline

2010-2039

2040-2069

T im eslice

Hylocomiumsplendens

Pteridiumaquilinum
- 1 s t - 99th

- 1 s t - 99th — — 5 th -9 5 th

— — 5 th -9 5 th

• 25th - 75th

100
*

90

2
5

80
70
60

2010-2039

2040-2069

*

50

6
■6

40
30

Baseiine

2070-2099

T im eslice

91

2010-2039

2070-2099

Rhytidiadelphus triquetrus

Kinbergia oregana

■Max"

1st- 99th

Baseline

2010-20

2 5 th -7 5 th

100

100
£

90

*

90

2

80

2

80

2010-2039

2040-2069

2070-2099

Figure 3-3: Percent of the British Columbia study area (total area = 97 857 km2) occupied by the
suitable climate space (SCS) for 18 plant species characteristic of the Coastal Douglas-fir moist
maritime (CDFmm) zonal site series as based on the minimum - maximum climatic envelope, the 1st
- 99th, 5th- 95th, and 25th- 75th percentiles, projected using a five-model global climate model
(GCM) ensemble. Species listed by stratum: trees (5), shrubs (7), herbs (3), and bryophytes (3).

The projected SCS distributions for individual species for the baseline (1961-1990) and
end of century (2070-2099) timeslices are presented spatially in Figure 3-4 with the tree,
shrub, herb, and moss layers, grouped and presented in that order. The BC study area is the
main window for each species with an inset of the PNW. The inset is useful to observe the
potential northerly shift in some SCS distributions. Note that when SCS is reduced over time,
it generally is lost from the lower elevation terrain along the coastline. Also note that many
of the SCS distributions do not indicate substantial change in the BC study area, but the
northerly shift in overall SCS is apparent in the inset figure of the PNW projection. Abies
grandis, Achlys triphylla, Gaultheria shallon, Lonicera ciliosa, Hylocomium splendens,
Kindbergia oregana, and Rhytidiadelphus triquetrus demonstrate the most dramatic
potential for northerly shifts in SCS, resulting in a loss of SCS in the study area.

92

M J*-

0

SB 100

200Wlomet«rt

U i— i. g . i. . A . - i- J

y

'

;■
Abies grandis (2070-2^99)

Aides grandis (baseline)

fi-

9 SO 100

N

0 SO 100

200

\

«s

JOOKHonMMn

l,,l-> ,.X „l

Acer macrophyllum (b&selin

Acer macrophyllum (2070-2

m *

0 »

100

i,

> j

200 Kilometers

*•

Arbutus menziesii {2070-209^

Arbutus menziesii (baseline

0 SO 100

0 SO 100

200K>k*n<t«ii

\

' 3

Pseudotsuga menziesii{basetj

aOOWkWHtw*

r

Pseudotsuga menziesii (2O70*p99)
93

e so too
1 * *. v

i

200HbrMtM
*

J A i

0 W 100

v

3DOMiom«wr*

••

Thuja plicata (baseline)

Thuja plicata (2070-2d99)

OCT""

0

s> 100

1 . . i i ,i,

mmmm

0 90 100

?

\

aoOMkxnMm

I, idl Ml ,1 I I ,.*■,■»,.■■»

XT

Cornus nuttallii (basel)n e)

Cornus nuttallii (2070-^099) ^

«g

0 50 100

30QK»om*#s

*

o 90 ioo

,

aoowtamim

.J * *

Gaultheria shallon (2070-20

Gauitheria shatlon (ba^eline^
Lf- a
*£ & r*rr

0 »

100

1 I I , I n i, . ! , In t

xammttm

0 90 100

200Kflom *tf*

j

" tiHolodiscus discolor (bd|S6 1in^

s
Holodiscus discolor (2070-2
94

*

^

m 100
U.

o

" H iLonicera ciliosa (baseline) ^
■

0 » 100

JOONtomatm

1, l.,l„„l„„l,.,.l,..,t..A.t

.,

Lonicera ciliosa (2070^ 099)^ 1

'

Mahonia nervosa (2070-20

0 50 100

200«tora**»

0 50 100

T

Rosa gymnocarpa (baseline^^

o so ioo

mmmm

jQOKHom#**

Rose,gymnocarpa (2 0 7 0 -2 0 9 #

<44

o so ioo

a & x fe tn ***

fcjuot.ii

L.jL^X .XutnnlnJm J

Symphoricarpos o to ^ b a s e t^ lj_________ Symphoricarpos albus {2 0 7 0 ^ 9 9 )
95

0 50 100

aoOKMNMttf*

Achlys triphylla (baselW)

0 50 100

Achlys triphylla (2070^099)

< 0

30OKHMMtWk ,

0 58 100

Polystichum munitum {2070^59)

Polystichum munitum (basell|

0 90 100

Pteridium aquilinum

Pteridium aquilinum (baselirfep

0 90 100
t „ *.1*. I

0 m kxj

a00Wtom«t«$
, 1, „ l •

.A

1 .»

J.inUn.i.nXi.if

Hylocomium splendens (b a s ^ e )________ Hylocomium $plendens{ {207&§&39)
96

0 50 wo

300Wfem*«m y-.

o

tn .......

aOKfemmra
> ..r

a»w ion«N n

Kindbergia oregana (i070-2(

Kindbergia oregana (^seli

0 50 100

s> wo

r

a 50 100

^

I , I .A . I . , 4 ..I..*,.!

Rhytidiadelphus triquetrus (b y lin e )

200

Rhytidiadelphus triquetrus (Z&M-2099)

Figure 3-4: Mapped distributions of suitable climate space (SCS) using the minimum and maximum
values for 16 annual climate variables on the south coast of British Columbia (with Pacific Northwest
inset in upper right corner representing the domain area for the analysis) for 18 diagnostic species of
the Coastal Douglas-fir moist maritime (CDFmm) zonal ecosystem for the baseline (1961-1990) and
future (2070-2099) time slices. Model projections for baseline conditions use ClimateWNA baseline
normals and future use the five-model global climate model (GCM) ensemble.
Focusing in on the current BEC v.7 ecosystem mapping (BC MFR 2008), the shift in
species SCS was evaluated relative to the current CDFmm subzone. Note that Table 3-3
indicates that all species' SCS distributions fully occupy the 5 631 km2 of the currently
mapped CDFmm (±4 km2 error). The five-model GCM ensemble suggests that very little SCS
area will be lost from the CDFmm initially, with the exception of one shrub species (Lonicera
ciliosa) and the three moss species (Hylocomium splendens, Kindbergia oregana, and
Rhytidiadelphus triquetrus) in the first future timeslice (2010-2039). The four previously

97

listed species, one tree (Abies grandis) and one herb (Achlys triphylla) SCS decrease in area
in the second future timeslice (2040-2069). The results for the 2070-2099 timeslice indicate
that the most SCS loss from the baseline CDFmm would be for Abies grandis, Gaultheria
shallon, Lonicera ciliosa, Achlys triphylla and the three moss species. Thuja plicata, Cornus
nuttallii, and Mahonia nervosa also decrease in SCS in the final timeslice of the century. The
Hylocomium splendens SCS decreases the most with a loss of 4 886 km2 from the current
CDFmm mapped area.
Table 3-3: Shift in species suitable climate space (SCS) over four timeslices within the current
mapped Coastal Douglas-fir moist maritime (CDFmm) (BEC v.7) in the British Columbia study area
using the minimum - maximum climatic envelope definitions and the five-model global climate
model (GCM) ensemble (CDFmm total area in BC study area = 5 631 km2).
Total SCS (km2)

Change in area (km2) from Baseline
CDFmm

Baseline
(1961-1990)

2010-2039

Cornus nuttallii
Gaultheria shallon
Holodiscus discolor
Lonicera ciliosa
Mahonia nervosa
Rosa gymnocarpa
Symphoricarpos albus

5 631
5 631
5 631
5 631
5 631
5 629
5 631
5 629
5 627
5 631
5 631
5 631

0
0
0
0
0
0
0
0
-14
0
0
0

Achlys triphylla
Polystichum munitum
Pteridium aquilinum
Hylocomium splendens
Kindbergia oregana
Rhytidiadelphus triquetrus

5 631
5 631
5 631
5 627
5 631
5 630

0
0
0
-116
-14
-87

Species
Abies grandis
Acer macrophyllum
Arbutus menziesii
Pseudotsuga menziesii
Thuja plicata

98

2040-2069
-12
0
0
0
0

2070-2099

0
0
0
-195
0
0
0

-36
-1886
0
-2 392
-14
0
0

-2
0
0
-2 014
-196
-737

-3 312
0
0

-4 699
0
0
0
-5

-4 886
-3 047
-4 490

3.4.2

Sensitivity Analysis for the Baseline Species Climatic Envelopes

The climate variables for each point on the 1-km grid were adjusted by 10 % of the
baseline range in climate variables for the BC study area. After the grid points were adjusted
(one adjustment at a time, all other climate variables remain unadjusted), the climatic
envelope analysis was run to determine the change in SCS area for the baseline timeslice for
each species. To evaluate the sensitivity of the species' response to the uncertainty in
climate variable predictions, the number of species whose ranges increased or decreased by
more than 100 km2 was calculated (Figure 3-5). The uncertainty associated with the fivemodel GCM ensemble projections impacts the climatic envelope models, generally resulting
in a decrease in species SCS.

Figure 3-5: The number of species whose range changes by more than 100 km2with an (a) increase
and (b) decreased climate variable adjustments by 10% of the current range of climate in the British
Columbia study area for the baseline (1961-1990) timeslice. A positive value (green bar) for the
number of species indicates an increase in species SCS range, while a negative value (red bar) for the
number of species indicates a decrease in species SCS range.
The sensitivity analysis also determines the climate variables that limit the speciesspecific climatic envelopes. The greater the change in species SCS, the more limiting that
climate variable is on the climatic envelope (Table 3-4).

99

Table 3-4: Change in area (km2) of baseline minimum-maximum suitable climate space (SCS) for each
species and the species overlap area with ± adjustment to each climate variable (one at a time) by
10% of the minimum - maximum climate variable range within the British Columbia study area.

Species

Adjust.

Abies grandis
Acer
macrophyllum
Arbutus
menziesii
Pseudotsuga
menziesii

+
+
+
+
+

Thuja plicata

-

Cornus nuttallii
Gaultheria
shallon
Holodiscus
discolor
Lonicera ciliosa
Mahonia
nervosa
Rosa
gymnocarpa
Symphoricarpos
albus
Achlys triphylla
Polystichum
munitum
Pteridium
aquilinum
Hylocomium
splendens
Kindbergia
oregana
Rhytidiadelphus
triquetrus
Species overlap
area

+
+
+
+
+
+
+
+
+
+
-

+
+
+
+
-

MAT
(“C)
0
0
0
0
0
0
0
0
0
0
0
-1965
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-2
0
0
0
-1368

MWMT
(*C)
0
0
515
-2252
0
-22
0
0
0
0
150
-10139
0
-846
0
0
0
0
0
-615
0
0
0
0
0
0
0
0
0
0
0
0
85
-10817
0
0
93
-7695

MCMT

TO

MAP

MSP

CO

CO

(mm)

(mm)

0
0
0
0
0
0
0
0
0
0
0
-1164
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-1543
0
0
0
0
0

-277
0
-881
0
0
0
0
0
0
-320
-990
-1711
0
-126
0
0
103
-5867
0
0
0
0
0
0
-1040
0
0
0
0
0
0
-10092
-579
-10065
0
-11241
-979
-5767

-18974
8068
-492
-129
-2559
116
-3058
400
-783
-43
-361
-134
-1171
44
-760
-91
-4629
342
-979
-58
-3213
267
-2010
239
-440
-75
-846
-16
-3517
426
-52
-79
-983
-116
-3825
646
-4346
281

-6627
-4183
-6584
-2341
-16881
-727
-3518
-1369
-59
-3026
-10342
-2487
-156
-3231
-14151
671
-21930
-3896
-16022
-1622
-13546
994
-177
-2180
-12907
174
-308
-2288
-2043
-1939
0
-7381
0
-7756
-1351
-5836
-18684
-4702

-444

-938

100

AHM
0
0
125
-21951
0
0
0
0
0
-39304
436
-18788
0
0
0
0
0
-4629
0
0
0
0
0
0
0
0
0
0
0
0
0
0
503
-43064
0
0
0
-10718

SHM
0
0
8286
-28984
673
-22047
0
0
210
-38428
4373
-23630
2779
-45766
0
-528
249
-22018
1750
-45214
0
-4929
0
0
0
0
0
0
0
0
4
-35023
331
-37672
0
-34147
37
-17216

Table 3-4 (continued).

Species
Abies grandis
Acer
macrophyllum
Arbutus
menziesii
Pseudotsuga
menziesii
Thuja plicata
Cornus nuttallii
Gaultheria
shallon
Holodiscus
discolor
Lonicera ciliosa
Mahonia
nervosa
Rosa
gymnocarpa
Symphoricarpos
albus
Achlys triphylla
Polystichum
munitum
Pteridium
aquilinum
Hylocomium
splendens
Kindbergia
oregana
Rhytidiadelphus
triquetrus
Species overlap
area

Adjust.
+
+
-

+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
-

DD<0
0
-36536
0
-35989
0
-28157
0
-44762
0
-45277
-3033
-29222
0
-45351
0
-29761
0
-24710
0
-42760
0
-43930
0
-44882
0
-44586
0
-45116
0
-44780
0
-47308
0
-45336
0
-45499
-2302
-25060

DD>5
0
-1053
0
-977
0
-129
0
-2447
0
0
1
-2282
0
-2149
0
-1516
0
0
0
-1479
0
-2183
0
-932
0
-1715
0
-2327
0
-2444
0
-2401
13
-7331
0
-294
1
-1651

NFFD
-617
-8
-9
-71
-9
-54
-17
-35
-182
0
-193
-1793
-280
-1293
-9
-36
-2306
0
-182
-904
-57
-98
-57
0
-516
-34
-9
-37
-9
-37
-5638
-2800
-5509
-9206
-7479
-4132
-4323
-1314

101

FFP
875
-7527
0
-542
0
-1363
58
-8383
4656
-8564
52
-3225
6
-7574
1
-5593
1322
-2821
1
-5640
11
-7477
0
-2506
547
-8364
33
-7999
187
-8745
-10
-919
538
-8864
135
-6570
-37
-2428

PAS
(mm)
-4519
-36407
-14718
-30258
-13298
-22999
-598
-42471
-3094
-41788
-8533
-29717
-62
-42730
-2985
-30957
-8345
-25444
-5085
-40894
-2933
-41473
-13308
-35068
-2263
-42081
-2393
-41718
-101
-42512
-1702
-43132
-1259
-43007
-6686
-40271
-8843
-24744

EMT
(*C)
0
0
-262
0
0
0
0
0
-215
0
-4735
504
0
0
0
0
0
-1
0
0
0
0
0
0
0
0
0
0
0
0
0
-338
-3443
-266
0
-1054
-3799
315

Eref
0
0
82
-4825
0
0
0
0
0
0
634
-21397
0
-6
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-69
0
0
178
-15793

CMD
0
-37585
0
-23162
0
-12298
0
-54913
0
-51657
0
-13749
0
-53278
7662
-22853
26
-9355
0
-42577
0
-49703
0
-43804
0
-50522
0
-55648
0
-55649
0
-56181
0
-48074
0
-49351
47
-8670

It is interesting that the increase of 10 % in mean annual temperature (MAT) did not
change the SCS for any of the 18 species and that the decrease of 10 % in MAT only
impacted the SCS area for two species (Cornus nuttallii and Kindbergia oregana). However,
adjustments to the mean annual precipitation (MAP) did impact the SCS results for all 18
species. The 10% increase resulted in a loss of SCS for all 18 species, while the decrease in
MAP resulted in a decrease of nine of the species and increase of the remaining nine species
SCS. The greatest losses in SCS resulted from the mean summer precipitation (MSP) and
precipitation as snow (PAS) adjustments. The decrease in climatic moisture deficit (CMD)
also resulted in large loss of SCS. With either an increased or a decreased adjustment of
climate variables across the BC study area, SCS is generally lost, not gained.
The species overlap area is altered consistently by a decrease of 10 % of any of the
climate variables. The decrease in MAP and extreme minimum temperature (EMT) both
result in an increase in species overlap SCS, while a decrease in 10% of the remaining
climate variables result in a decrease in species overlap SCS. The increase by 10 % of each
climate variable resulted in various changes. The increase in MAT and annual heat:moisture
index (AHM) do not alter the species overlap SCS, and the increase in growing degree-days
(DD >5) only increases SCS by 1 km2. Summer heat:moisture index (SHM), Hargreaves
annual reference evaporation (Eref), mean warm monthly temperature (MWMT), and CMD
increases by 10 % all result in increases in the overlap SCS by up to 178 km2. Increasing any
of the remaining nine climate variables result in decreases in the overlap SCS, by up to
18 684 km2.

102

3.4.3

Comparison of the CDFmm Ecosystem Mapping to Modelled Species Overlap and
CDF Climatic Envelopes in the British Columbia Study Area

The current BEC system maps the CDFmm as being restricted to the southeast coast of
Vancouver Island, small islands of the Salish Sea, and the southwestern mainland BC (Figure
3-6).

0

SO 100

200 KUomatere

Figure 3-6: Current mapped distribution of the Coastal Douglas-fir moist maritime (CDFmm) in the
British Columbia study area (BEC v.7, BC MFR 2008).
The species overlap (Figure 3-7a) for baseline SCS distributions includes an increase in
area, when compared to the mapped CDFmm (Figure 3-6), expanding north and east on the
mainland and north and west on Vancouver Island. In 2010-2039 timeslice, the species
overlap SCS shifts northward very slightly, but is similar to the baseline SCS. In 2040-2069,
the species overlap SCS is similar again to the baseline, but there is less of the currently
mapped CDFmm area within the polygon. In 2070-2099, the species overlap SCS shifts
northward and out of the currently mapped CDFmm.

103

Ik #

>

Baseline

•

.

J

b

a

(1 9 6 1 -1 9 9 0 )

*
»»m
' \

/ i

2 0 1 0 -2 0 3 9

rn m m m

}

x
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
■
■
■
-

/*%
iy^B*

■tsv^
‘
■
■ »

I

’'r~ *v

K
t _wm. _mm«m

^ ,|r
2 0 4 0 -2 0 6 9

L

i

t » W MWMn
A

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

ik ?

>

2 0 7 0 -2 0 9 9
I s fZ

......
e «o«* mmmmn

**0*J9

t,.4,

SOOWtorNW* f-..T
1,^.4
< --

■>

?

x
/*a

■ &

Figure 3-7: Comparison of the suitable climate space (SCS) for the (a) overlap area of 18 individual
diagnostic species, and (b) extended Coastal Douglas-fir (CDF) SCS (Chapter 2) for the baseline and
three future timeslices in the British Columbia study area using a five-model global climate model
(GCM) ensemble.
For the ecosystem-based approach, using the extended CDF zone developed in Chapter
2, the SCS baseline is similar to the current CDFmm mapping (Figure 3-6) though it extends
further north on the southeast side of Vancouver Island. In the 2010-2039 timeslice, the
104

projected SCS shifts northward and westward (Figure 3-7b). In the 2040-2069 timeslice, the
SCS shifts more northward and up the inlets on the mainland. In the 2070-2099 timeslice,
the SCS again shifts further northward and out of the currently mapped CDFmm. The
ecosystem-based approach results in a smaller SCS area compared to the individual species
overlap SCS area.
Overall, the species overlap and ecosystem-based approaches result in similar shifts out
of the currently mapped CDFmm by the end of the century, but the species overlap
approach begins shifting out of the currently mapped CDFmm in the first future timeslice
(2010-2039) and the ecosystem approach shifts in the last future timeslice (2070-2099).

3.5

DISCUSSION

3.5.1

Species Climatic Envelopes

This study uses the same approach to climatic envelope modelling as Chapter 2. The
trends over time for changes in SCS areas are variable for each species (Figure 3-3), as found
in other studies (Berry et al. 2002, Chen et al. 2011). Most species SCS areas increase over
the first future timeslice indicating that BC climate conditions appear to be in the process of
changing to better suit all 18 species diagnostic of the CDFmm, according to the five-model
GCM ensemble. Climate conditions continue to change through the second and third future
timeslice, but with variable effects on species SCS distributions. The confidence in each
species projection is indicated by differences obtained for climatic envelopes defined by the
different percentile ranges. Where the trends (Figure 3-3) are parallel over time, then the

105

confidence in the model remains the same among future projections. If the lines in Figure 33 diverge, then this indicates that the confidence in the projections diminishes over time.
The confidence in the models generally drops off by the end of the century, suggesting that
the SCS projections are less reliable for the more distant future, as expected.
The tree species distributions in the BC study area consistently increase or remain the
same (Figure 3-3) and cover the currently mapped CDFmm area (Figure 3-4) until the end of
the century, with one exception, Abies grandis (Figure 3-3). By the end of the century, the
southeastern tip of Vancouver Island is projected to not be within the Abies grandis SCS.
The projected change in SCS distribution is more apparent in the PNW inset figure, revealing
very little SCS for this species south of the border. The Abies grandis climatic envelope
definition has the lowest maximum value for mean annual temperature (MAT), mean warm
monthly temperature (MWMT), mean annual precipitation (MAP), annual heat:moisture
index (AHM), summer heat:moisture index (SHM), growing degree-days (DD>5), number of
frost-free days (NFFD), frost-free period (FFP), extreme minimum temperature (EMT),
Hargreaves reference evaporation (Eref), and Hargreaves climatic moisture deficit (CMD)
and the highest minimum value for MAP, chilling degree-days (DD<0), and precipitation as
snow (PAS), when compared to the other CDF tree species (Table 3-2). This suggests that
Abies grandis is the most sensitive of the five tree species to increases in annual and
seasonal air temperature climate variables, but is also sensitive to decreases in precipitation
(rain or snow) and the number of cooling days.
Current descriptions state that Abies grandis occurrence decreases with increasing
latitude, precipitation, and elevation, though it tolerates fluctuating groundwater tables

106

(Pojar and MacKinnon 1994, Klinka et al. 1989). The climatic envelope for Abies grandis
suggests that this species' tolerance to mean annual precipitation (441 mm minimum and
3880 mm maximum) is more limited than other tree species in the study area. However, it
is not the MAP that most limits the climatic envelope for this species, it is the climatic
moisture deficit (CMD) that does. The range of Abies grandis extends down into California,
but only where summer drought rarely occurs (USDA1965). It may be that the summer
drought on southern Vancouver Island becomes longer in duration with climate change,
beyond the acceptable range for this species.
Overall, the tree species climatic envelopes presented here are similar to the results of
Hamann and Wang (2006), where species also gain more habitat than is lost. The exception
to this is Abies grandis, as it shifts out of the currently mapped CDFmm in this study, but
remains in Hamann and Wang's projections (2006). The persistence of Pseudotsuga
menziesii is supported by climatic envelope projections created using the Random Forests
algorithm (Rehfeldt et al. 2006, Wang et al. in press). Another study by Burton and Cumming
(1995) suggested that the CDF would be replaced by chaparral community, dominated by
evergreen shrubs that are currently found in southern Oregon and California. A subsequent
study predicted that Pseudotsuga menziesii and Abies grandis would totally disappear
within 25-100 years (depending on the GCM), constituting a catastrophic collapse of the
CDF ecosystem (Cumming and Burton 1996). This differs from the projected species ranges
that suggest that most tree species currently found here will persist.
Suitable climate space for shrub species generally increases or remains the same, but
again there is a northward shift in distribution. Gaultheria shallon and Lonicera ciliosa SCS

107

shift northward out of the southeastern end of Vancouver Island. The Gaultheria shallon
climatic envelope definition has the lowest maximum values for MAT, MWMT, AHM, DD>5,
and CMD when compared to the other shrubs, suggesting that increases in temperatures,
growing degree-days, or climatic moisture deficits would decrease the SCS for this species
(Table 3-2). The current Gaultheria shallon distribution in BC decreases with increasing
elevation and continentality (Klinka et al. 1989) and is primarily associated with moist,
coastal conifer forests (Pojar and MacKinnon 1994). Gaultheria shallon grows under a
variety of soil moisture conditions, but is most vigourous in moist conditions and loses
vigour under drier soil conditions (Haeussler et al. 1990). The climatic envelope for
Gaultheria shallon was most limited by the CMD value, though some SCS was lost due to an
overall increase in TD using the five-model GCM ensemble.
The Lonicera ciliosa climatic envelope definition has the lowest maximum value for MAP,
MSP, SHM, NFFD, FFP, EMT, and Eref, and the highest minimum value forTD, MSP, SHM,
DD<0, and PAS when compared to the other shrubs (Table 3-2). This suggests that the SCS
of this species is more susceptible to changes in precipitation regimes and seasonal
extremes in temperature when compared to the other CDF shrubs. Lonicera ciliosa currently
occurs on very dry to moderately dry, and nitrogen-medium soils, and is characteristic of
moisture-deficient sites (Klinka et al. 1989). This suggests that the range of Lonicera ciliosa is
facilitated by moisture deficit, but the climatic envelopes suggest that the SCS is most
limited by chilling degree days (DD<0) and precipitation as snow (PAS) which both decrease
substantially using the five-model GCM ensemble.

108

The SCSs for individual herbaceous species cover most of the study area in the baseline
and future timeslices, with the exception of Achlys triphylla, where the distribution shifts
northward and out of the southeastern end of Vancouver Island and the Fraser Valley on
mainland BC. The Achlys triphylla climatic envelope definition consists of the narrowest
range (e.g., highest minimum and lowest maximum values) of climate variable values when
compared to the other herbaceous species, with the exception of its maximum MAP value
(Table 3-2). This suggests that Achlys triphylla is susceptible to changes in almost all climate
variables and therefore could be an important indicator for climatic change. The current
distribution of Achlys triphylla decreases with increasing latitude, elevation, and
continentality (Klinka et al. 1989). The Achlys triphylla climatic envelope is defined by one of
the lowest maximum values for continentality (TD), which increases over time using the
five-model GCM ensemble.
The three mosses (Hylocomium splendens, Kindbergia oregana, and Rhytidiadelphus
triquetrus) are more consistent in their decrease in SCS and northerly shift over time. All
moss SCS areas are projected to become reduced in the PNW US and the south end of
Vancouver Island and coastal areas by the end of the century, suggesting that mosses are
also important indicators of climatic change. Moss species are unique in that they are
capable of tolerating periods of desiccation, and begin photosynthesizing with increases in
moisture (Schofield 1992). The three moss species all decrease in SCS area with increases in
CMD values over time, suggesting that the seasonal dry period may shift in duration and
that the climate suitable for moss species will not persist.

109

Seven species SCSs (Abies grandis, Gaultheria shallon, Lonicera ciliosa, Achlys triphylla,
Hylocomium splendens, Kindbergia oregana, and Rhytidiadelphus triquetrus) are projected
to shift out of the current CDFmm distribution on Vancouver Island. Most of these species
SCSs were most sensitive to the increases in climatic moisture deficit (CMD), except for
Lonicera ciliosa which was most sensitive to potential decreases in precipitation as snow.
Based on the model projections using the five-model GCM ensemble, the seven species
highlighted above could be sensitive indicators of climate change in the study area, with the
moss species impacts occurring within the first future timeslice (Table 3-3). Overall, the
individual species SCS shifts differ for each species in response to climatic gradients, as were
the results in other forest systems (Austin 1987).

3.5.2

Sensitivity Analysis

The sensitivity analysis met two objectives: first, to evaluate the five-model GCM
projections and the impact of GCM uncertainty on SCS; and secondly, to evaluate the
climatic constraints on each species SCS. The GCM prediction uncertainty was evaluated by
the number of species SCS that changed by more than 100 km2 (Figure 3-5). The potential
error or variation in temperature related climate variables seems to have relatively little
impact on the SCS for all species. The precipitation and moisture related variables have
more impact on the distribution of species SCS. This is interesting as temperature is
relatively more predictable (e.g., has less uncertainty associated with it) in GCM modelling
than the precipitation variables (Randall et al. 2007). The adjustments in moisture-related
climate variables changed the area for some species by up to 55 000 km2, demonstrating

110

the importance of considering model uncertainty and the need for refined precipitation
projections in predicting climate space suitable for future species distributions.
The results of the 10% adjustments to climate variables on the individual species SCS
areas reveal the climate variables that most constrain the distribution of those species. If
the climate variable adjustment that impacts the SCS the most is considered to be the
dominant constraint on the distribution of a species, then three climate variables are the
most important: DD<0, PAS, and CMD. Fewer chilling degree-days (DD<0) constrains three
species: Acer macrohpyllum, Arbutus menziesii, and Symphoricarpos albus. The species
overlap area is also most constrained by DD<0. Reduced precipitation as snow (PAS)
constrains three species: Cornus nuttallii, Holodiscus discolor, and Lonicera ciliosa (all shrub
species). Reduced climate moisture deficit (CMD) constrains the remaining 12 species: Abies
grandis, Pseudotsuga menziesii, Thuja plicata, Galtheria shallon, Mahonia nervosa, Rosa
gymnocarpa, Achlys triphylla, Polystichum munitum, Pteridium aquilinum, Hylocomium
splendens, Kindbergia oregana, and Rhytidiadelphus triquetrus. Though the species listed
above have the same climate variable that most impacts the SCS, each species is secondarily
impacted by different climate variables.
Two species in particular, Cornus nuttallii and Kindbergia oregana, are impacted by
adjustments in all climate variables. Seven of the species (Arbutus menziesii, Pseudotsuga
menziesii, Holodiscus discolor, Rosa gymnocarpa, Symphoricarpos albus, Polystichum
munitum, and Pteridium aquilinum) are constrained by adjustments in seven climate
variables: MAP, SHM, DD<0, DD>5, NFFD, FFP, PAS, and CMD (i.e., other climate variables
do not change the net SCS area). Overall, it appears that Cornus nuttallii is most constricting

111

the overlap area and therefore could be the most susceptible to impacts of climate change
within the CDFmm subzone.

3.5.3

CDFmm Ecosystem Mapping, Individual Species Overlap, and CDF Climatic Envelopes

Generally, the climatic envelope definitions for individual species commonly found in the
CDFmm are broader than the definitions of the ecosystem as defined in Chapter 2. The
ecosystem and individual species overlap approaches produce similar projections with
reduced SCS in the currently mapped CDFmm over time (Figure 3-6 and 3-7). The species
overlap SCS is distributed over a larger area than the extended CDF zone SCS with broader
climatic envelope definitions. The projected species overlap area based on individual SCS
areas is also greater than the currently mapped CDFmm (BC MFR 2008). Perhaps the
ecosystem approach is bound to define a more restricted area in that it intrinsically reflects
other site or niche factors, such as soils or competition, while the individual species
approach only considers plant species relationships with climate (the basic climatic niche).
The mapped CDFmm comparison confirms that a number of other factors constrain the
diagnostic CDFmm plant association distribution from occupying its full potential, including
soil and terrain effects, competition, dispersal limitations, and the history of biogeographic
migrations and disturbance. Hence, the climatic niche only represents one component of
the ecological niche (Pearson and Dawson 2003). Consequently, climate is not currently the
greatest limitation to CDFmm ecosystem distribution, and likely will not be in the future
either.

112

Few studies have attempted to reconstruct community assemblages from individual
species predictions (Guisan and Thuiller 2005). One comparison of projected effects of
climate change on species and their respected biotic communities suggested that impacts to
individual species may not be consistent with impacts to the plant community distributions.
Contrarily, a study by Rehfeldt et al. (2006) projected similar increases in distributions of
both Pseudotsuga menziesii and coastal forests, though this study was limited to comparing
tree species (excluded understory vegetation) and forest zones.
However, another study including multiple species components of the plant community
found different results. Comparisons of species and community distribution models in
alpine environments found that the species have a broader niche and require more
individual statistical fitting than community models (Zimmerman and Kienast 1999). Results
of the Zimmerman and Kienast (1999) study suggest that community models better predict
vegetation patterns than the dominant species, as community patterns are more uniform in
response to environmental gradients while species patterns are predicted with lower
accuracy at local scales. The results found here and in the literature suggest that plant
assemblages are complex and include both individual and inter-species tolerances when
explored across potential future climatic gradients (Shipley and Keddy 1987, Lortie et al.
2004).

3.5.4

Limitations of the Climatic Envelope Technique

The climatic envelope current and future projections were developed using the
ClimateWNA interpolation tool. This tool enables scale-free climate projections, but with

113

imperfect outputs, reflecting the distribution of climate station input data and the
uncertainty associated with the GCM input data influence the output data. The selection of
the five GCMs for the five-model GCM ensemble was based on the GCMs most
representative of the baseline climate normals for the study area. Though the five-model
GCM ensemble most closely represents the current climate normals for the study area, the
uncertainty around the projections increases over time (e.g., the further into the future, the
more uncertainty). This is one of many potential combinations of GCMs that could be
applied to explore the potential future climate regime for the study area.
The climatic envelope model approach is an exploratory approach designed to evaluate
the relationships of climate and vegetation. The climatic envelope definitions were created
using the minimum and maximum values of geo-referenced locations for the species. The
potential errors in data collection, such as species identification and location records, could
result in climatic envelope definitions that are beyond the climatic limit for the species. To
reduce the chance of locations being falsely labelled as suitable for the species SCS, 16
climate variables were included in the climatic envelope definitions, and 98-percentile, 90percentile, and 50-percentile core ranges were also evaluated. The 16 variables are a
conservative filter on the evaluation of suitable points.
Other potential sources of error for the model outputs are associated with the
availability of species location data. The moss data were limited in the US, so the climatic
envelope definitions are skewed to conditions in BC and a small amount of US herbaria
data. This could be influencing the perceived shift in species distributions in the future
projections as these species shifted earlier and further than other species. Also, modelled

114

projections are limited by potential bias in the collection of observation data to more
accessible locations, compounded by the approach used to parameterize species
distributions (Pearson et al. 2006, Franklin 2009), resulting in variation among approaches
for predicting the impacts of climate change (Pearson and Dawson 2003, Hijmans and
Graham 2006). Despite the limitations of the climatic envelope approach (Guisan and
Zimmerman 2000), the application of these simple climate driven models provides some
useful preliminary guidelines for monitoring (of indicator species and locations) and nature
conservation (as also concluded by Zimmermann and Kienast 1999, Berry et al. 2002, and
Rose and Burton 2011).

3. 6

CONCLUSIONS AND RECOMMENDATIONS

Overall, this chapter has demonstrated that climate change will impact the suitable
habitat and range for individual species over time. The impacts are not uniform across the
landscape, suggesting that impacts may not be consistent through the current ecosystem
distribution. Monitoring individual species growth, abundance, and presence across site
types throughout the current distribution of the CDFmm to detect changes is important for
tracking the impacts of climate change to BC's distinctive Coastal Douglas-fir species and
ecosystems. Overall, it appears that Cornus nuttallii is most sensitive to both increased and
decreased adjustments of the current climate attributes and therefore could be the most
susceptible to impacts of climate change within the CDFmm subzone.
Exploring the projected species- and ecosystem-based distributions suggests that
ecosystem-based projections capture the community at the local spatial scale. Results also

115

support the paradigm shift beyond the historical dichotomy between ecosystem and
individualistic approaches, recognizing the complexity of plant communities functioning
both individualistically (with some species independent of the assemblage) and other
species interdependent (Lortie et al. 2004). The degree to which the local and regional
distributions of all species in the CDF are correlated and do or do not show significant
relationships to any of the several climatic gradients remains to be determined, but is
beyond scope of this study.
Based on the results of species climatic envelope projections, monitoring the seven
species (Abies grandis, Gaultheria shallon, Lonicera ciliosa, Achlys triphylla, Hylocomium
splendens, Kindbergia oregana, and Rhytidiadelphus triquetrus) with SCS areas that are
projected to shift from the current CDFmm is recommended. In particular, it is suggested
that the growth and vigour of Cornus nuttallii and Kindbergia oregana, the two species that
are identified as most limited to the CDFmm plant community, would be sensitive indicators
of a changing climate. Species may vary in their rates of change, depending on multiple
internal species traits and external drivers (Chen et al. 2011). Locations where individual and
overlap species SCS areas are projected to be most impacted in the near future are also
important to monitor, as such areas can be expected to exhibit vegetation change first.
Conversely, locations where SCS areas are projected to remain stable should be recognized
as valuable areas of persistence and therefore may serve as refugia for species and plant
associations (Rose and Burton 2009, 2011). Continued research on the assemblage of plant
communities and the ecological niche is required to better describe and predict the
complexity of natural systems over space and time.

116

4.

Site Descriptions and Preliminary Microclimate Differences in a LongTerm Monitoring Network Established in the Transition between
Coastal Douglas-fir and Coastal Western Hemlock Forests

4.1

ABSTRACT
Dry coastal forests are unique on the coast of British Columbia. The characteristic

summer soil moisture deficit may be impacted by changes in future climate. A long-term
monitoring network was established between the Coastal Douglas-fir moist maritime
(CDFmm) and Coastal Western Hemlock very dry maritime (CWHxm) subzones to quantify
microclimate across a soil moisture gradient and to track changes over time. The mesic sites
are most representative of regional climate, while some upper- and lower-slope sites are
also included to indicate how terrain is modifying the climate at the site-level. The baseline
summaries for the CDFmm and CWHxm mesic sites over the course of one year show subtle
differences in annual and seasonal microclimate. The findings support of the existing
biogeoclimatic classification, though continued data collection to compile a multi-year data
set is required to describe the microclimate regime. The study sites are available for
additional research and are currently included in other nutrient, plant growth and vigour,
vegetation change, and decomposition studies. The long-term data sets will be valuable for
climate change monitoring and for guiding research and management in these ecosystems.

117

4.2

INTRODUCTION

4.2.1

Vegetation and Microclimate

There is strong evidence that global climatic change will impact the distribution of
forested ecosystems. In North America, climate predictions for the next century suggest
increasing precipitation in winter and spring, with decreasing precipitation in the summer
(Christensen et al. 2007) and temperature increases of 5"C that could displace isotherms by
500 km (Manabe and Wetherald 1986, U.S. Committee of Global Change 1988). Such shifts
in precipitation regimes and increasing temperatures on the coast of British Columbia (BC)
could result in a greater abundance of dry ecosystem types as regional vegetation
distribution is largely determined by climate (Woodward 1987, Pearson and Dawson 2003).
Most climate change projections conclude that climate conditions suitable for the Coastal
Douglas-fir moist maritime subzone (CDFmm) will shift, over the next several decades, into
areas now classified as the Coastal Western Hemlock very dry maritime subzone (CWHxm,
Hamann and Wang 2006, Campbell and Wang in press).
Ecohydrology is the study of physical hydrology, soil moisture, and their interaction with
vegetation. These interactions are important when considering the impacts of climatic
change on soils and vegetation in ecosystems strongly influenced by soil moisture regimes
(Rodriguez-lturbe 2000), such as the CDFmm subzone on the southern coast of Vancouver
Island, BC. The CDFmm has a Mediterranean-type climate with vegetation that thrives in the
rainshadow of the Olympic and Vancouver Island Mountains, with warm dry summers and
mild wet winters (Green and Klinka 1994). The ecology of the CDFmm is characterized by a
summer moisture deficit.

118

Ecosystems are mapped at the subzone level on the basis of potential vegetation that is
suited to the climate, terrain, and soil parameters according to BC's Biogeoclimatic
Ecosystem Classification (BEC) system (Eng and Meidinger 1999). Climate change
projections suggest the current CDFmm distribution could shift northward and upward in
elevation and that a reduction in moisture availability in the summer could substantially
reduce growth of Douglas-fir, the dominant tree species, and other vegetation across site
types (Burton and Cumming 1995, Cumming and Burton 1996, Spittlehouse 2003).
Soil moisture balance is often the key functional expression of hydrological dynamics
articulated in ecological patterns and processes (Rodriguez-lturbe 2000). Soil moisture is an
integrator of climate and site conditions and is an important link in climate, soil, and
vegetation interactions (Figure 4-1). The BEC system divides subzone forest units (e.g.,
CDFmm) into site series which are based on relative soil moisture (xeric, mesic, hygric) and
nutrient (poor, medium, rich) regimes within the subzone (MacKinnon et al. 1992). Soil
moisture balance is a strong determinant of the relative vigour, reproduction, and
competitive abilities of plant species, resulting in the distinctive plant associations
associated with each site series. The potential impacts of climatic change to vegetation and
soil moisture availability can be quantified using soil water balance.

Climate

Soil Moisture

Soil

Vegetation
Figure 4-1: Soil moisture is an integration of climate, soil, and vegetation relations (adapted from
Rodriguez-lturbe and Porporato 2004).

119

4.2.2

Boundary Layer Climate

Linking climate to soil and vegetation processes at the sub-regional scale requires an
understanding of boundary layer climates. The boundary layer is adjacent to the Earth's
surface and is often delineated into layers to quantify the interaction between the soil
surface and the atmosphere. All atmospheric processes impact the soil surface, but the
boundary layer is the layer that influences the soil surface on a daily time scale. Boundary
layer climate is driven by turbulence resulting in regular mixing of air parcels and flux in heat
and moisture (Lemon et al. 1957, Oke 1987). The depth of the boundary layer is determined
by friction across rough surfaces and diurnal temperature gradients. The depth ranges from
1 km during the day when the surface is heated by the sun and drops to 200 m at night
when the surface cools (Oke 1987). The vertical layering of the atmosphere varies across
space and time with temperature and moisture gradients but can be generalized for
studying processes at different scales.
The boundary layer climate system is composed of energy and mass fluxes and balances.
Based on the First Law of Thermodynamics, there is no change in total energy in the system,
but energy changes form continuously (e.g., radiant, thermal, kinetic, and potential energy).
The soil-atmosphere system is a process-response system as there is linkage between
energy flow (process) and the change in temperature (response; Oke 1987). The physical
properties of the system determine the ability to (1) absorb, transmit, reflect, and emit
radiation, (2) conduct and convect heat, and (3) capacity to store energy (Oke 1987). For
example, site aspect and slope position impact solar radiation inputs, while soil texture,

120

coarse fragment content, humus form, and soil depth impact the infiltration rate of
precipitation to the rooting zone and moisture feedbacks to the system. In forested
environments, the canopy structure buffers solar radiation and convection (e.g.,
temperature and wind; Oke 1987).
Water content of the soil-atmosphere system is conserved in three states: water vapour,
liquid water, and solid (ice and snow). Water is transported through the system by
convection, precipitation, infiltration, and runoff. Convection, the movement or mixing of
air masses, transports water in liquid and gas form to and from the atmosphere in both
sensible and latent forms (Oke, 1987). Precipitation is the primary input of water to the soilatmosphere system. Infiltration and runoff are the movement of water into soil from the
surface and over the soil surface. In forests, interception and evapotranspiration are also
important components of water transport through the system (Rouse and Wilson 1972,
Rollenbeck and Anhuf 2007).
Energy fluxes (sensible and latent heat) influence evaporation and snowmelt, but also
the rate of water flow through the system (Dingman 2002). For example, an increase in
temperature, even with a decrease in precipitation, would increase evaporation from water
surfaces, leading to increased air humidity and, in time, could facilitate precipitation events
elsewhere (Varallyay 1990). This example scenario demonstrates how the macro-scale
(106 m horizontal scale) inputs to the system are impacted by meso-scale (104 m horizontal
scale) and micro-scale (10'2- 103 m horizontal scale) feedback processes resulting in a
system that is in a constant flux of energy and mass (Oke, 1987).

121

There is increased research in linking macro-scale to micro-scale processes driven by the
increase in climate change research (Dekker et al. 2007). Field studies in microclimate
research and monitoring provide the data required to quantify boundary layer dynamics
across sites, and to explore the potential impacts of forest management and climate change
at individual sites (Montaldo et al. 2008).

4.2.3

Soil Water Balance

The soil water balance is based on the conservation of soil water as a function of time
and is applied to quantify the physical movement of water (Rodriguez-lturbe and Porporato
2004). Water balance for a soil column is used at various spatial scales with a focus on soil
water storage for plant use (Equation 4-1). Precipitation is the primary input balanced with
evapotranspiration from the soil surface and vegetation, runoff across the soil surface, and
change in soil moisture storage. Ground water is not considered as a source of soil moisture
for plants as it is beyond the depth of the rooting zone sampling for this study.

Equation 4-1

p=ET + R + AS

where p = precipitation, ET = evapotranspiration, R = surface runoff, and AS = change in
water storage in soil column (Oke 1987). All quantities are volumetric measurements (L3).

122

Hydrological processes related to water balance
The hydrological processes involved in calculating water balance equations operate on
different time scales: precipitation inputs go directly to the ground surface or are
intercepted by vegetation (Figure 4-2); intercepted water either evaporates or accumulates
until the capacity of the leaf, branch, or stem to hold the water is surpassed and the water
falls to the ground (throughfall); once the water lands on the ground surface, it can infiltrate
from the surface into the soil or will run off over the surface to stream channels or to areas
where it may infiltrate to unsaturated soil zones (Oke 1987). The rate of each of these
processes is driven by physical forces (e.g., gravity, wind) and thermodynamics (e.g., solar
radiation, convection).

In te rc e p tio n

E v a p o tra n s p ira tio n (ET)
P re c ip ita tio n (P)

T h ro u g h fa ll
S u rfac e R u n o ff (R)

In filtr a tio n (I)
C h ange in w a t e r s to ra g e (AS)

Figure 4-2: Hydrological processes in the water balance model: water enters as precipitation, moving
through the system by interception, throughfall, infiltration processes, and is lost from the system
by evapotranspiration, surface runoff, and change in water storage.

123

Precipitation
The primary input to a water balance equation is precipitation, which is a random event
that varies over time scales. To consider this variability, stochastic differential equations are
applied in predictive water balance modelling (Rodriguez-lturbe 2000). Water vapour from
the Earth's surface is lifted in unstable air masses and through convection. When water
vapour cools to its dew point, it condenses to form cloud droplets or ice crystals (Oke 1987).
Precipitation events occur when droplets or crystals can no longer be held in suspension in
the atmosphere. On smaller spatial scales, dew, hoarfrost, rime, and fog-drip are
precipitation events that occur close to the Earth's surface (Oke 1987).

Interception and Throughfall
Interception is the process by which precipitation lands on vegetative surfaces before it
hits the ground. A significant fraction of total evapotranspiration in most (dry) regions is
intercepted water (Dingman 2002). This loss of water back to the atmosphere depends on
the structure and form of vegetation (specifically, leaf area index, LAI) and the intensity and
frequency of precipitation events, especially in dry ecosystems (Rodriguez-lturbe and
Porporato 2004). For example, interception rate is higher in temperate rainforests in
comparison to prairie grasslands. The water intercepted by vegetation accumulates on the
foliage, stems, and branches where it either evaporates or (when the vegetation is
saturated) falls off and drops to surface below (throughfall).

124

Evapotranspiration
Evapotranspiration is the composite loss of water to the atmosphere from all sources
including ground surface evaporation, intercepted water evaporation, and vegetation
transpiration. Evapotranspiration is determined largely by meteorologic variables such as
solar radiation, temperature, humidity, and wind speed (Dingman 2002). Evaporative loss is
strongest by day, particularly in early daylight hours. Evapotranspiration is greatest early in
the day as convective activity increases air parcel mixing and it declines into evening,
reaching its lowest levels at night (Oke 1987). When soil moisture content is high,
evapotranspiration rate also depends on plant type and climatic conditions (Rodriguezlturbe and Porporato 2004).
Actual evapotranspiration (AET) is difficult to measure and therefore, potential
evapotranspiration (PET) rates are often used in water balance calculations. Estimates of
actual evapotranspiration (AET) can be calculated using the ratio of average precipitation
(W) to potential evapotranspiration (PET, Equation 4-2, Pike 1964).

Equation 4-2

AET = W / [1 + (W/PET)2] *

Potential evapotranspiration is estimated using calculations such as the Bowen ratio or
the Penman-Monteith equation for microscale studies that focus on site-level processes
(Spittlehouse 1989). Evapotranspiration calculations depend on the ratio of surface-air
temperature difference to surface-air vapour-pressure difference. Vapour pressure is a
function of elevation (and varies slightly over time at a given location) and latent heat varies

125

slightly with temperature (Dingman 2002). The Penman-Montheith calculation also
considers soil and canopy heat fluxes and wind speed in and above the forest canopy
(Spittlehouse 2003).
Evapotranspiration rates are influenced by characteristics of the vegetative surface, such
as: (1) albedo of the surface, which determines the net radiation; (2) the maximum leaf
conductance; (3) the atmospheric conductance, which is largely determined by vegetation
height; (4) presence or absence of intercepted water; (5) soil type; and (6) terrain
characteristics, such as aspect, slope, and elevation.

Surface runoff
Surface runoff is all the precipitation and moisture that lands on the surface and does not
infiltrate into the soil profile. This occurs when rainfall volume exceeds the available soil
storage capacity (Rodriguez-lturbe and Porporato 2004). Surface runoff is most easily
quantified in watershed studies with a streamflow gage at the mouth of the stream
network. For landscape- or microsite-scale studies, the temporal distribution of runoff is
estimated using probabilistic analysis that considers precipitation inputs and infiltration
rates (e.g., when near surface soil moisture reaches 100% capacity, surface runoff occurs).

Infiltration
Infiltration is the movement of water from the soil surface into the soil where it is
available for plants or it enters ground water storage. The magnitude and rate of infiltration
depends on precipitation and soil moisture dynamics (Rodriguez-lturbe and Porporato
126

2004). Infiltrated water is either retained in storage, is evaporated from the surface, taken
in and transpired by vegetation, or it resurfaces by travelling through ground water
channels (Dingman 2002). Water flow occurs in response to spatial gradients of potential
energy from gravity and/or pressure into unsaturated soils (Dingman 2002). Soil surface
microtopography, soil texture, plant and animal activity, and thickness of the forest floor
impact infiltration rates, which vary over the landscape (Dingman 2002). Infiltration can be
quantified by monitoring soil moisture content at various depths in the soil profile over
time.

Groundwater storage
Groundwater is the lowest boundary of the soil profile. Loss to groundwater is
maximized when soil is saturated and rates decrease rapidly as soil dries out (Rodriguezlturbe and Porporato 2004). The change in storage is almost always assumed negligible in
water balance models, particularly over long measurement periods. To minimize the
variability in the groundwater term in strongly seasonal climates, it is recommended that
water balance calculations start and end at the same time of year, (e.g., in mid-September
when fall rains commence in coastal BC) or at least at times of the year when storage is
equal (Dingman 2002).

Soil moisture
Impacts to soil water balance as a result of changes in climate are complex, but can be
summarized through soil moisture content as an integrative and ecologically important
127

condition (Varallyay 1990, Rodriguez-lturbe 2000). Soil moisture dynamics depend on many
processes and interactions. Soil moisture is quantified by relative soil moisture content or
volumetric water content, depending on the objective of the analysis. The volumetric water
content (9) is the ratio of water volume to soil volume in the soil profile. Changes in
precipitation and temperature will impact components of the soil water balance (Table 4-1;
Varallyay 1990). Generally, a decrease in precipitation results in decreases in
evapotranspiration, runoff, infiltration, water storage, and groundwater recharge, while
increases in temperature increase evapotranspiration and decrease other related
hydrological processes (runoff, infiltration, and water storage).

Table 4-1: Potential impacts to components of the soil water balance with conditions of related
components, if applicable (Varallyay 1990).

Component o f Water Balance

Impact (with conditions, if applicable)

Decrease in average annual precipitation
Evaporation

Decrease, if infiltration is limited

Transpiration

Decrease, if plant canopy is not limited

Surface runoff

Decrease, if infiltration rate, permeability,
and water storage capacity of soil is limited

Infiltration and water storage

Decrease, if not limited

Groundwater recharge

Decrease, if soil profile has good vertical
drainage and permeability is not limited

Increase in annual temperature
Evaporation and transpiration

Increase, if plant canopy not suffering from
limited water supply

Surface runoff, infiltration, water storage,

Decrease, particularly if precipitation is low

groundwater recharge

128

Vegetation influences on water balance
Soil moisture dynamics are complex, with linkage to several physical, thermal, and
biological processes within the boundary layer and the soil root zone. In forested
ecosystems, vegetation influences water balance by altering air movement and
temperature regimes. In general, there is greater air mixing over a forest stand because of
the rough surface (e.g., compared to bare soil or herbaceous vegetation). During the day,
there is often a temperature inversion within the stand (i.e., with warmer air below the
canopy than above it) with very little temperature variation with depth and time (Oke
1987). The effects of mutual shading, multiple reflection, and long-wave radiation
interaction (feedback) are applicable at the community scale. Vegetation is strongly linked
to evapotranspiration as well as interception, plant water storage, and the structure of the
soil moisture zone as it relates to plant root morphology and distributions.
Vegetation cover impacts soil temperature and water content with slight feedbacks to
latent heat flux and sensible heat flux (El Maayar et al. 2002). Plant roots add to the
structure of the soil over the life cycle of the plant. Roots add organic matter and fibrous
materials to hold water in the soil but also create macropores that facilitate water
channelling through the soil profile.
Plants directly impact soil moisture by taking up water from the soil. The water available
to plants is the difference between field capacity of the soil and the permanent wilting point
for the plant. Although the permanent wilting point is frequently treated as a constant
(corresponding to moisture held at a tension of -1.5 MPa), in fact species differ in their
ability to withdraw water from soils across a wide range of soil water (matric) potentials

129

(Lambers et al. 2008). Plants indirectly impact soil moisture by affecting the partitioning of
precipitation to infiltration and surface runoff and the drainage of water to groundwater
and surface runoff, and the partitioning of available energy to latent and sensible heat.
Consequently, the growth of vegetation is often strongly related to soil moisture availability,
but as a nonlinear function (Teuling et al. 2006).
Soil water balance models can be used to explore potential impacts of climate change on
vegetation. Overtime, considerable changes in species distribution, dynamics, diversity, and
production capacity will lead to changes in ecosystem structure and function (Varallyay
1990). The changes in vegetation will trigger climatic feedback by modifying albedo, surface
roughness, micro-circulation of heat, and energy balances in the boundary climate that will
alter forest water cycles and soil formation processes over the long term (e.g., > 100 yrs;
Charney 1975, Varallyay 1990). Projecting such changes requires complex models and
extensive field data to calibrate site- or species-specific relationships and validate modelled
outcomes. Although soil water balance modelling is beyond the scope of this thesis, the
study design for a long-term monitoring network will provide data required for comparative
soil water balance modelling in the study area.

4.2.4

Research Objectives

In 2010, an integrated climate change project was established to investigate the
potential impacts of climate change by evaluating the relationships of macroclimate and
microclimate, nutrient regimes, vegetation productivity, and regeneration across the
transition from the CDFmm to the CWHxm subzone (Klassen et al. 2010; Klassen and

130

Saunders 2010,2011). This chapter describes the establishment of the integrated climate
change project and addresses four short-term objectives for the microclimate study to: (1)
complete a literature review on soil water balance modelling (above); (2) develop a field
sampling study design; (3) deploy and maintain equipment for monitoring forest
microclimate and soil moisture conditions; (4) devise a data management and
summarization protocol; and (5) complete a baseline description of the microclimate at
representative mesic sites.

The baseline microclimate summary of CDFmm and CWHxm mesic sites includes:
1. A vegetation, site, and soil description for the six mesic sites.
2. A comparison and description of the annual, seasonal, monthly, and daily
microclimate conditions in the CDFmm and the CWHxm for the baseline year
(September 15, 2010 - September 14,2011).
3. A comparison and description of the microclimate attributes for the CDFmm and
CWHxm subzones during the 2011 growing season (May 15 - September 14, 2011),
including the timing and amounts of extreme conditions and soil moisture
persistence.

This chapter focuses on the baseline (year 1) description of microclimate conditions on
mesic sites of the CDFmm and CWHxm, providing quantified field conditions associated with
the ecosystems and species explored in Chapters 2 and 3. Long-term data series are
required to fully characterize the moisture regime and microclimate trends and to detail the

131

relationships between macroclimate, microclimate, and soil moisture across sites, and to
thereby more fully comprehend growing conditions in a dry coastal forest ecosystem.
Although such a complete analysis is beyond the scope of this thesis, the work described
here provides a snapshot of some differences, and provides a legacy of baseline data and
installed instrumentation to be followed up on in the future.

4.3

METHODS

4.3.1

The Vancouver Island Study Area

The CDFmm subzone is the smallest subzone on the coast of BC (257 000 ha, <0.3% of
the province) and the only subzone within the Coastal Douglas-fir (CDF) zone. The CDFmm is
located on the southeastern side of Vancouver Island, small islands in the Salish Sea, and
the adjacent mainland BC in the rain shadow of the Olympic and Vancouver Island
Mountains (Chapter 2, Figure 2-1). This subzone consists of a Mediterranean-type climate
with vegetation that thrives with warm, dry summers and mild, wet winters (Green and
Klinka 1994). The ecology of the CDFmm is characterized by a summer moisture deficit.
The Coastal Western Hemlock (CWH) zone is distributed along the entire latitudinal
gradient of the provincial coast, but the CWHxm subzone is the smallest of all CWH
subzones (896 300 ha, 1.0 % of the province). The CWHxm is currently distributed adjacent
to the CDFmm (Chapter 2, Figure 2-1), typically at higher elevations where the climate is
slightly cooler and wetter (Chapter 2, Table 2-1). The slight shift in temperature and
moisture regime is more suited to climax vegetation dominated by coastal Tsuga
heterophylla (western hemlock) rather than Pseudotsuga menziesii var. menziesii (coastal

132

Douglas-fir), which takes on more of a successional role in the CWHxm (Green and Klinka
1994). Both of these subzones have some occurrence on mainland BC, but are primarily
distributed on Vancouver Island and the smaller islands in the Salish Sea. For more details
on the study area and BEC, see Chapter 2.

4.3.2

Study Site Selection

Study sites were selected to represent a range of soil moisture conditions (mesic, hygric,
and xeric) through the latitudinal gradient of the CDFmm subzone. Published BEC (BC MFR
2008) and Terrestrial Ecosystem Mapping (TEM; RIC 1998) digital map layers were
combined with Terrain Resource Information Management (TRIM) digital topographical
map layers to locate suitable site candidates. The site candidates were evaluated using
ecosystem delineation, forest age, contours, water bodies, roads, railways, aspect, and
elevation features available on the map layers. Climatic envelope projections (following
Rose and Burton 2011) were also reviewed to assess potential shifts in future climate
normals, with the aim of identifying areas projected to have the greatest changes in climatic
moisture deficit (CMD) over the next 30-60 years.
The eastern side of Vancouver Island is primarily private land and is the most populated
region of Vancouver Island. To establish field sites in this area, land ownership and site
security were also considered in the site selection process. Because of a long history of
timber harvesting and mixed land use, the forest stand age and structural stage varies
considerably across the fragmented landscape. Ideal study sites are representative of
mature CDFmm and CWHxm forests, specifically mesic sites within the two subzones and

133

the transition zone between, and xeric and hygric sites within the CDFmm. The study sites
also meet logistical criteria including truck accessibility and proximity to roads to reduce
travel time.

4.3.3

Field Sampling Design

The long-term sampling network is designed to investigate microclimate across the
CDFmm and CWHxm subzones, as well as site conditions within the CDFmm subzone. The
field measurement installation, which will continue operating for the foreseeable future,
consists of three transects that run from the CDFmm across the transition zone into the
CWHxm. There are five plots on each of the three transects to sample a range of soil
moisture conditions: (1) CDFmm mesic; (2) CDFmm xeric; (3) CDFmm hygric; (4) CDFmmCWHxm transition mesic; and (5) CWHxm mesic.
The CDFmm and CWHxm mesic sites are to be monitored for changes in ecosystem
distribution across the transition of CDFmm to CWHxm, while the edaphic sites (xeric and
hygric) are instrumented to monitor changes within the CDFmm. The transition sites are
positioned to detect changes along the current CDFmm boundary.
The three transects run from the eastern side of Vancouver Island in a westnorthwesterly direction into the CWHxm. The three transects are distributed latitudinally
through the CDFmm ecosystem: the southern transect is located in the Victoria area, the
central transect is located in the Duncan/Cowichan Valley area, and the northern transect is
located in the Qualicum/Coombs area (Figure 4-3).

134

SamplingTransect

Figure 4-3: Biogeoclimatic Ecosystem Classification (BEC) subzone map with the south, central, and
north sampling transect locations on the southeastern end of Vancouver Island, British Columbia.

Each study site is established in a stand of mature vegetation characteristic of the
subzone and site series. The site is equipped with a microclimate station including a
Campbell Scientific (CS) CR10 data logger with three Decagon EC-5 soil moisture sensors and
one soil temperature sensor (copper-constantan, CuCn, thermocouple wire) inserted in the
rooting zone (sampling at ca. 25-30 cm depth). Soil temperature is also measured with two
Thermachron iButtons® at subplot 1 and subplot 2 because the CS CR10 data logger is
limited in the number of sensors that can be wired to each data logger's channel board. Air
temperature is measured using CuCn thermocouple wire mounted at two heights (0.3 m
and 1.5 m) with radiation shields on a metal stake adjacent to the data logger box.

135

Additional microclimate attributes associated with soil water balance calculation are also
measured on the CDFmm and CWHxm mesic study sites to assess and track microclimate
dynamics and, over time, calculate soil water balances to assess impacts of climate change.
The mesic sites in the CDFmm and CWHxm also have a metal tripod and mast with an Onset
HOBO® U30 data logger with the following sensors: HOBO® wind speed sensor (at 3.5 m),
HOBO® rain gauge sensor (at 3.0 m), HOBO® temperature and relative humidity sensor (at
1.5 m), and HOBO® silicon pyranometer (solar radiation) sensor (at 1.0 m). One exception is
that the CDFmm mesic site on the north transect is equipped with CS-compatible sensors
and the data are recorded on the CS CR10 data logger. This site will be equipped with a
HOBO® climate station when funds are available. The equipment models and technical
specifications are detailed in Section 4.3.4.
The plot center of each study site is centered in a homogeneous forest stand that is a
minimum of 1 ha in area. The three soil moisture and soil temperature subplots were
established using the following protocol: (1) subplot 1 is 2 m @ 0° (north) from plot center,
(2) subplot 2 is 5 m from subplot 1 @ 150° (southeast), and (3) subplot 3 is 5 m from subplot
1 @ 210° (southwest). Subplot 3 is always closest to the CR10 data logger box (Figure 4-4).
Due to soil coarse fragment content, tree locations, or soil depth, the site layout was
modified as required. If modifications were required, the angle was offset first and then, if
necessary, the distance from subplot 1 was shortened to position the sensors at the
specified depth.

136

Subplot 1

A

1

N
5 m @ 210'

/

5 m @ 150'

/

/
J 2 m @ 0°

A
/

P lo t c e n te r

/

1 1 C R 10 d a ta lo g g e r bo x

1

I A ir te m p e r a t u r e sensor

S u b p lo t 2

S u b p lo t 3

Figure 4-4: Site layout with the three soil moisture and soil tem perature subplots, the Campbell
Scientific CR10 data logger box, and airtem perature sensors positioned around the plot center.

On the mesic sites in the CDFmm and CWHxm, microclimate stations also include a
tripod with solar radiation, wind, rain, and temperature and relative humidity sensors (for a
total of six sites with tripods). The tripod was located as close as possible to subplot 3 (but
outside the subplot layout), adjacent to the soil moisture and temperature sensors and the
CS CR10 data logger box.

4.3.4

Field Sampling Equipment and Data Collection

4.3.4.1 Data loggers
Campbell Scientific (CS) CR10 data loggers are used at all sampling sites to record the soil
moisture, soil temperature, and air temperature data. The data loggers are programmed
using CS LoggerNet v.3.3 software. On the north transect CDFmm mesic site, all
microclimate data are recorded on a CS CR10, while the remaining five mesic sites (north
137

CWHxm, central CDFmm and CWHxm, and south CDFmm and CWHxm) include a HOBO®
climate station to record the wind speed, precipitation, relative humidity, and solar
radiation.
The CS CR10 data loggers are housed in a water-resistant plastic box on the ground
surface next to subplot 3 on each study site (Figure 4-5). The HOBO® data loggers are
housed in a waterproof box that is mounted on the tripod at 1.8 m above the ground
surface. Each CS CR10 data logger is powered by a 12 V Panasonic rechargeable sealed leadacid battery. These older data loggers are limited by data storage capacity, not power
availability, and have to be downloaded every two months. Each HOBO® data logger is
powered by a 4 V rechargeable sealed lead-acid battery that is charged on site with a solar
panel mounted on the south side of the tripod at 2.3 m above the soil surface. Although
these data loggers do not have the same data capacity or power limitations, they were
downloaded and maintained every two months with the CS data loggers.

Figure 4-5: (a) Data logger box and air temperature sensors adjacent to the soil moisture and soil
tem perature sensor at subplot 3 of the CDFmm hygric site on the central transect and (b) the
Campbell Scientific CR10 data logger and Panasonic 12 V battery in the data logger box.

138

All data loggers were programmed to record data at 1-hour logging intervals, with
sampling ranging from 15 second (CS CR10 data loggers) to 4 minute (HOBO® 1130 data
loggers) intervals, with the mean values saved on the hour. The HOBO U30 data logger had
the 4-minute sampling to save power as the solar panel was not recharging the battery,
particularly in the late fall and winter months. The exceptions to the 1-hour logging interval
are the Thermochron iButton® temperature sensors, which record soil temperature at 3hour logging intervals due to limited memory capacity.

4.3.4.2 Soil moisture
There are two ways to measure soil water status in the field: by water volume content or
by energy or water potential. Field measurement techniques for soil water content include:
gravimetric, neutron scattering, time domain reflectometry, and capacitance sensors. Water
potential for vegetation can be inferred from soil moisture content measures if the soil
texture (particle size distribution) is known (Dingman 2002).
The gravimetric method removes soil samples from the sampling site. The samples are
taken to a lab where the moist soil sample is weighed, dried to constant weight, and then
weighed again to calculate the moisture content as the relative difference between wet and
dry weights. This method is often used to calibrate technical methods (e.g., data loggers and
soil moisture sensors; Brady and Weil 2008). The neutron scattering method is based on
two facts: (1) fast neutrons are slowed by collision with hydrogen atoms and (2) hydrogen is
most abundant in water in the soil (Oke 1987). A probe measures the slow neutrons
associated with water to measure soil moisture content within ca. 10 cm of the probe. Time

139

domain reflectometry (TDR) is an electromagnetic method that measures the rate at which
an electromagnetic pulse travels down and then back up a transmission rod and the degree
of dissipation of the impulse at the end of the lines. Capacitance sensors are thin metal rods
that determine the dielectric constant of the soil which is then converted to infer soil
moisture content.
The gravimetric method is time consuming and destructive in that it removes soil
samples and impacts the study area; but it remains the standard calibration method (Brady
and Weil 2008). The neutron scattering method requires a radiation permit, is expensive,
and the accuracy of this method is impacted by potential soil disturbance when placing the
probe in the soil column, and by high coarse fragment content in the soil (Oke 1987).
Overall, the TDR method is most widely used today, but the equipment is somewhat
expensive. The TDR equipment is portable, accurate over a wide range of soil water content
and soil types, and can be installed at a range of sampling depths. The capacitance methods
are less expensive but up until ca. 5 years ago, the sensors had to be calibrated with
changes in temperature and salinity.
Decagon EC-5 soil moisture sensors apply modern capacitance technology that measures
volumetric water content by determining the dielectric constant of the soil using
capacitance and frequency domain technology (Figure 4-6, Decagon Devices 2009, Ventura
et al. 2010). The improved technology (e.g., improvements to frequency settings and the
sensor architecture) has reduced the calibration requirements. The EC-5 sensors are
calibrated for mineral soils and measure a range of 0-100% volumetric water content within
a 0.18 L soil volume of influence. The sampling depth (25-30 cm depth) targets the rooting

140

zone for vegetation growing on the forested sites. The accuracy is rated at ± 3% (Table 4-5).
Note that under very dry conditions, occasional negative soil moisture content values can
be reported.

Figure 4-6: (a) Soil pit (35 cm depth) dug to insert the Decagon EC-5 soil moisture sensor and (b) the
soil moisture sensor installed into the rooting zone soil profile (25-30 cm depth) at the CDFmmCWHxm transition site on the north transect.

4.3A.3 Soil and air temperature
The copper-constantan thermocouples are deployed as both airtemperature (at 1.3 m
and 0.3 m above the surface, Figure 4-7) and soil temperature (25-30 cm below surface)
sensors. The thermocouples use standard metals that generate predictable output voltages
and can be used over a large range of temperatures. The two metals in the thermocouples
are soldered at one end. A voltage is generated at the junction and this voltage is a function
of temperature (Pico Technology 2011). The thermocouples deployed below the soil surface
are covered with electrical heat shrink tubing to protect them from moisture.

141

Figure 4-7: Thermocouple air temperature sensors with radiation shields at 0.3 m and 1.5 m above
the ground surface, mounted on a metal stake adjacent to the Campbell Scientific CR10 data logger
box and subplot 3 at the CDFmm mesic site on the south transect.

Thermochron iButton® (model DS1921G) units are also deployed to measure soil
temperature at subplots 1 and 2, because the CS CR10 data logger is limited in the number
of sensors that can be wired to each data logger's channel board. The iButtons® are selfcontained single-channel data loggers with a built-in digital thermometer and real-time
clock (Figure 4-8; Maxim Integrated Products 2011). The iButtons® record data within the
same unit as the sensor, so they are dug up every 2 months to download with the other
data loggers. A 30-cm hole is dug adjacent to the unit, data are downloaded, and then the
unit is placed back into the soil profile with the least disturbance as possible.

142

Figure 4-8: Thermochron iButton® sensor (small metal disc) being inserted into rooting zone soil
horizon (ca. 27 cm depth) to measure soil temperature adjacent to the Decagon EC-5 soil moisture
sensor.

The soil moisture and soil temperature sensors were placed in the soil at a single
standard depth in each of the three subplots at each study site. A 35 cm soil pit was dug and
the soil moisture sensor was inserted in a downward angle into the profile in the upslope
direction (if the site is sloped) from the side of the soil pit at a depth of 25-30 cm. If the
sensor hit rock or roots, the sensor was re-inserted into undisturbed soil. In rocky soils, a
soil knife was inserted to create a channel for inserting the sensor without damaging it, but
ensuring that the sensor maintained contact with the soil (e.g., no air pockets to capture soil
water).
The thermocouple soil temperature sensors are used in subplot 3 at each site, inserted
into the side of the soil pit at the mid-point depth of the soil moisture sensor (i.e., soil
moisture sensor at 25-30 cm, temperature sensor at 27.5 cm). The iButton® temperature
sensors were placed (in subplots 1 and 2) at the same depth as the soil moisture sensor, but
are offset horizontally by 25 cm to minimize disturbance to the soil moisture sensor when
the iButtons® are dug up to download bi-monthly.
143

Two air temperature thermocouple sensors were mounted on aluminum stakes (0.3 m
and 1.5 m height above the ground surface). The air temperature sensors are housed in
radiation shields so ambient air temperature is measured rather than the effects of
radiative heating associated with direct-beam solar radiation. Radiation shields are
camouflaged with sword fern fronds to reduce visibility and the risk of potential human and
animal interaction with equipment. All radiation shields were camouflaged to maintain
consistent protocol across all sites.

4.3.4.4 Additional microclimate measures on CDFmm and CWHxm mesic sites
As noted above, additional microclimate measurements are logged at six of the mesic
sites using Onset HOBO® data loggers and sensors (Figure 4-9), with the exception of one
station on the north transect CDFmm site that uses CS-compatible sensors so the data can
be logged on a CS CR10 data logger along with the soil moisture and soil temperature data.
The relative humidity, airtemperature, precipitation, solar radiation, and wind sensors are
all mounted on a 3-m tripod and mast adjacent to the soil moisture and soil temperature
sensors at subplot 3 and the CS CR10 data logger box. The accuracy and output range of the
CS-compatible equipment is greater than that of the HOBO® equipment, so the HOBO®
specifications (Table 4-2) were used to guide data quality control and report interpretations.

144

Figure 4-9: HOBO® microclimate station at the south transect CDFmm mesic site, showing the tripod
and mast structure that house the wind speed, precipitation, air tem perature and relative humidity,
and solar radiation sensors, and with a HOBO® U30 data logger powered by a solar panel.

Relative humidity and air temperature
Relative humidity (RH) and airtemperature measurements are combined in one sensor
that is mounted in a radiation shield 1.5 m above the ground on the north side of the
climate station tripod at each site. The HOBO® temperature and RH sensor is accurate to 0.2
°C at temperatures ranging from 0°C to 50 °C and ± 2.5% RH from 10% to 90% RH (Table 42). The Vaisalla HMP35 model 107F temperature and relative humidity probe that is

145

deployed on the north transect CDFmm mesic site is accurate to 0.2 °C within the -45 °C to
45 °C range.

Precipitation
Precipitation is measured with an Onset HOBO® and CS TE525 rain sensors, consisting of
tipping bucket rain gauges mounted on the tripod 3 m above the soil surface on the north
side of the tripod. The rain gauge has a funnel at the top that directs the precipitation onto
a mechanism that teeters back and forth when full. Each tip records a specified amount (0.2
mm for the HOBO®) of precipitation in millimetres (up to a maximum of 80 mm per 1-hr
logging interval).
The rain gauge sensors are not heated, therefore if temperatures drop below freezing,
the precipitation freezes within the bucket. When air temperatures rise above freezing, the
snow and ice melt and the amount is measured, although any sub-zero precipitation
introduces two sources of error. First, if the bucket fills with snow or ice, then any
subsequent precipitation runs off and the total amount is not measured. Secondly, the
hourly and daily time stamps for precipitation events are not accurate, as there is a time lag
in the melting period that skews and attenuates the data.
Additional potential sources of error with the tipping bucket are caused by debris
accumulation in the bucket and by wind and turbulence. Leaves and needles from the trees,
as well as pollen and dust, can accumulate and block the funnel in the bucket. The
equipment is routinely checked in conjunction with data downloads and, if possible, the
sites are visited after storm events to maintain equipment. Undercatch due to winds and

146

turbulence is also an issue with the tipping buckets, although the impact in the forested
sites is minimal.

Solar radiation
The amount of subcanopy solar radiation is measured with a HOBO® Silicon Pyranometer
sensor and a LiCor LI200S full spectrum radiation sensor (at the north transect CDFmm
mesic site) mounted 1.0 m above the soil surface on the south side of the tripod to monitor
light availability to the understory vegetation. The HOBO® Silicon Pyranometer sensor has a
300 -1 1 0 0 nm spectral range and the LiCor L1200S sensor has a 400 -1100 nm spectral
range. These sensors are typically used in open sites, but are suitable to measure light and
to predict soil warming and evaporation within the forest, but not to monitor or sample
plant growth (photosynthetically active radiation (PAR) sensors are used for such studies,
with a narrower spectral range of 400-700 nm). Silicon photodiodes are used to measure
radiation flux in terms of power per unit area (W/m2) with small measurement errors when
used in natural sunlight conditions, but have reduced accuracy when sensors are deployed
under a plant canopy (Onset 2009). Both the HOBO® and LiCor sensors are located under
the forest canopy, so the measures will be used to compare relative differences in
subcanopy radiation among sites.

Wind
Wind speed is measured at 3.3 m above the soil surface on the south side of the tripod at
each site. The HOBO® wind sensor and the RM Young wind sensor (on the north transect
147

CDFmm mesic site) are both anemometers that measure wind speed every three seconds
and the hourly average is recorded on the (1-hr) logging interval. The HOBO® wind sensor is
accurate within ± 0.5 m/s and measures wind speed up to 54 m/s.
Table 4-2: Summary o f the HOBO® microclimate sensors and their technical specifications.

Temperature

Microclimate
Parameter

Soil moisture

Manufacturer/Model

Decagon EC-5

Accuracy

Output Range

Range

0 to saturation

-40 °C to 60 °C

0.03 m3/m 3 (±
3%) in all soils; up
to 8 dS/m

Temperature

Copper-Constantan
thermocouple wire

Wind speed

HOBO® wind sensor

± 0.1 °C
± 0.5 m/s at
> lm /s

Precipitation

n/a
0 to 54 m/s

± 1.0% at up to

0 to 12.7 cm/hr;
maximum 4000

20 mm/hr

tips per logging

HOBO® rain

- 270 °C to 400
°C
-40°C to 75 °C

0° C to 50 °C

interval
Solar Radiation

HOBO® silicon

± 10 W /m 2 or

pyranometer

±5%, whichever
is greatest in

0 -1280 W /m 2

-40 °C to 75 °C

-40 °C to 75 °C

-40 °C to 75 °C

sunlight
Temperature

HOBO® temperature
and RH sensor

Relative
Humidity (RH)

4.3.5

0.2 °C over 0 -5 0
°C
± 2.5% from 10%

0 to 100% RH at

to 90% RH

-40 °C to 75 °C

-40 °C to 75 °C

Data and Analysis

4.3.5.1 Study site description
After the sites were selected, ecological plot descriptions were completed on each site,
including vegetation, site, and soil data as per BC MOELP and BC MOF (1998). The
vegetation data were collected in the 20 m x 20 m (400 m2) plot by vertical strata and

percent cover (estimated by eye, averaged between estimates of two field personnel during
June 2010 sampling). The strata were organized by vegetation height starting with the tree
canopy stratum (A, > 10 m height), the shrub stratum (B, woody plants 0 m - 1 0 m height),
the herb stratum (C), and the moss, lichen, and seedling stratum (D). Percent cover by
species and stratum was estimated as the percentage of the ground surface covered by the
crown.
The site location and description was recorded to characterize the forest stand. The geo­
referenced location, aspect, slope percentage, elevation, mesoslope position, and surface
topography were recorded to describe the physical characteristics of the study site (BC
MOEIP and BC MOF 1998). The successional status and structural stage of the forest stand
were recorded to characterize the forest. The soil moisture and soil nutrient regimes were
determined using the site and soil descriptions.
A soil pit was dug on each of the study sites to describe the forest floor and soil profile
for each site. The terrain and soil classification as well as the descriptions of the forest floor,
rooting depth, rooting zone particle size, and root restricting layer were recorded. The
water source and drainage class for the rooting zone were also included in the survey. The
mineral soil description included the depth, colour, texture, percent coarse fragments, and
soil structure. Evidence of biological and fungal activity in the soil was also noted, if present.
A hemispherical photo was taken above each of the three soil moisture subplots on each
study site to describe the canopy cover directly above the soil moisture sensors. The photos
were taken using a tripod with a Canon digital SLR with a Sigma 8-mm fish-eye lens levelled
1.5 m above the soil moisture subplot.

149

The canopy gap fraction, openness, and leaf area index (LAI) were calculated from the
hemispherical photos using WinSCANOPY® software (Regent Instruments Inc. 2006). The
gap fraction is the fraction of pixels classified as open sky in a sky grid region in the image
(computed as the proportion of pixels classified as sky). Openness is the relative amount of
open sky, emphasizing the region of the canopy above the lens. This is different from the
gap fraction, which is calculated on a two dimensional space, while the openness considers
canopy elevation in the area of sky projected in the image. Leaf Area Index (LAI) is the one­
sided (or projected) leaf area (m2) per unit of ground area (m2), estimated in WinSCANOPY®
from the gap fraction data at different zenith angles based on Bonhomme and Chartier
(1972).

4.3.5.2 Microclimate data analysis
Annual, seasonal, monthly, and daily microclimate data
To describe the CDFmm and CWHxm mesic site conditions for the baseline year
(September 15,2010-September 14, 2011), digital data downloaded from data loggers
were compiled and standardized to an hourly data format. Daily, monthly, seasonal, and
annual summaries for each research site were created using the plotrix package and the
stats package in R statistical software (Lemon 2006, R Core Development Team 2010). Mean
soil moisture, temperature, relative humidity, solar radiation, and wind speed, plus total
precipitation values, were calculated for each time period for the CDFmm mesic and
CWHxm mesic sites. The mean, standard error, minimum, and maximum value for each
subzone were calculated. A paired t-test to test for a significant difference (df=2, p < 0.05)

150

between CDFmm and CWHxm means was also calculated using the stats package in R
statistical software (R Core Development Team 2010). The samples were paired by transect
to remove some of the variance, resulting in a more powerful test, but also note that the
difference between means was difficult to detect with a small sample size (Gotelli and
Ellison 2004).
A climatic diagram for each subzone was created using the three-site average for
monthly total precipitation and mean, minimum, and maximum air temperature (1.5 m)
values for each subzone. The Walter and Lieth climatic diagrams were created using the
climatol package in R statistical software (Guijarro 2011). This figure provided a visualization
of the dry and moist seasons as well as the potential frost period and was used to define the
growing season for the 2010-2011 baseline year.
The daily soil moisture, precipitation, and temperature values were plotted on a time
series for the baseline year. The daily values for each subzone are the average of the three
mesic sites. The time series figure was also used to interpret and define the growing season
for the baseline year.

Growing season microclimate data
The growing season was defined to be the time period between peak rain events and
within the dry season (portrayed as a moisture deficit in the climatic diagram), with
continuous increase of daily mean temperatures that stayed above 5°C, and without risk of
sub-zero temperatures. The baseline year growing season (May 15 - September 14,2011)
data for the CDFmm and CWHxm mesic sites were summarized using the plotrix package

151

and the stats package in R statistical software (Lemon 2006, R Core Development Team
2010). A paired t-test to test for a significant difference (df=2, p < 0.05) between the
subzone means was also calculated using the stats package in R statistical software (R Core
Development Team 2010) as described above.
The 2011 growing season extremes and soil moisture persistence were further
characterized using the mean daily microclimate values for each subzone. The microclimate
extremes are defined by the maximum and minimum daily values for each climate variable
and the dates of occurrence during the growing season.
Soil moisture persistence is the tendency for high values of the variable to follow high
values, indicating that the soil moisture remains within the soil profile day after day
(Dingman 2002).The soil moisture persistence evaluation was initiated after the wet humid
period and the rain event that occurred in mid-May 2011 to determine the number of days
that the soil moisture content from the wet humid period (winter and spring) persists in the
rooting zone (25-30 cm depth).
The soil moisture persistence was calculated using the autocorrelation function in the
stats package in R statistical software (R Core Development Team 2010). Autocorrelation
describes how the initial daily soil moisture value (day 0) is related to subsequent daily soil
moisture values. The autocorrelation function was used with a time lag of one day and a
confidence level of 95% (default settings). The autocorrelation function (ACF) is defined by
equation 4-3, ranges between -1 and +1, and indicates the strength of the linear
relationship (Crawley 2007).

152

^

Equation 4-3

where p(/c) = the autocorrelation function at lag k, y(k) = the autocovariance function, and
y(0) = the variance function.
Positive autocorrelation function values indicate a positive correlation in the time series,
values near or equal to zero indicate no correlation, and values less than zero indicate that
the values become less related overtime. If values of autocorrelation are consistently close
to zero, then the data are interpreted as temporally random. Higher or lower values of
autocorrelation suggest trends in the data.

4.4

RESULTS

4.4.1

Coastal Douglas-fir moist maritime (CDFmm) and Coastal Western Hemlock very dry
maritime (CWHxm) Mesic Site Descriptions

The percent cover for the total vegetation in each stratum was estimated during the
survey and results were averaged among the three mesic sites to describe the overall
canopy structure in CDFmm and CWHxm forests. The mean total tree canopy (A) stratum
cover was equivalent for the CDFmm and CWHxm mesic sites at 66.7% cover. The mean
total shrub (B) stratum cover was greater in the CDFmm (94.3%) when compared to the
CWHxm (76.7%). The mean herbaceous stratum (C) was less in the CDFmm (4.0%) than the
CWHxm (10.3%). The mean total moss (D) stratum cover was similar, with 38.3% in the
CDFmm and 46.3% in the CWHxm.

153

There were species that occurred in both the CDFmm and CWHxm mesic sites (Table 4-

3). Species that were only found in the CDFmm include Acer macrophyllum, Arbutus
menziesii, Amelanchier alnifolia, Adenocaulon bicolour, Blechnum spicant, and Corallorhiza
maculata. Species found only in the CWHxm include Carex sp. Chimaphila umbellata,
Galium sp., grass sp., Linnaea borealis, Kindbergia praelongum, Plagiothecium undulatum,
and Rhytidiadelphus triquetrus. The standard error was greater for the species percent
cover means that were in the upper ranges (> 10 % cover). There was not a significant
difference in the percent cover for any of the species except for Trientalis borealis (t =5.196,
p = 0.035). Note that the sample size is very small (n=3) which limits detection of a
significant difference between the two subzones.
Table 4-3: Vegetation strata*, species, frequency (percentage occurrence in n=3 plots), and mean
percent cover (± standard error (S.E.)) with paired t-test results (df = 2, p < 0.05) for plant species
found in the Coastal Douglas-fir moist maritime (CDFmm) and Coastal Western Hemlock very dry
maritime (CWHxm) mesic study sites on the south, central, and north transects. Significant results in
bold.

Stratum*

A
A
A
A
A
6
B
B
B
B
B
B
B

Species

Acer macrophyllum
Arbutus menziesii
Pseudotsuga menziesii
var. menziesii
Thuja plicata
Tsuga heterophylla
Abies grandis
Acer macrophyllum
Amelanchier alnifolia
Arbutus menziesii
Gaultheria shallon
Holodiscus discolour
Lonicera hispidula
Mahonia nervosa

Paired t-test

CDFmm
%
mean % cover
plots
(± S.E.)
33
1.33 (1.33)
33
2.00 (2.00)

CWHxm
%
mean % cover
( i S.E.)
plots
0
0.00 (0.00)
0
0.00 (0.00)

t
1.000
1.000

P
0.423
0.423

100
33
33
67
33
33
33
100
67
33
100

100
67
100
67
67
0
0
100
33
33
100

2.747
1.386
1.375
1.000
0.000
1.000
1.000
0.101
0.961
> 4.303
1.288

0.111
0.300
0.303
0.423
1.000
0.423
0.423
0.929
0.438
1.000
0.327

63.33 (3.33)
1.67 (1.67)
1.00 (1.00)
1.33 (0.88)
1.00 (1.00)
0.33 (0.33)
0.67 (0.67)
59.00 (8.02)
2.33 (1.86)
0.33 (0.33)
36.67 (10.93)
154

32.00 (11.36)
21.67 (13.02)
19.33 (12.86)
1.00 (0.58)
1.00 (0.58)
0.00 (0.00)
0.00 (0.00)
61.67 (18.56)
0.33 (0.33)
0.33 (0.33)
18.33 (3.33)

Table 4-3 (continued).

Stratum*

Species

CDFmm
mean % cover
%
plots
(± S.E.)

CWHxm
%
mean % cover
plots
(± S.E.)

Paired t-test
t

Pseudotsuga menziesii
67
1.67 (1.67)
0.139
33
var. menziesii
1.33 (0.88)
0.17 (0.17)
33
0.17 (0.17)
>4.303
Rhamnus purshiana
33
67
2.767
100
1.50 (0.50)
0.37 (0.32)
Rosa gymnocarpa
1.000
Symphoricarpos albus
33
1.33 (1.33)
0
0.00 (0.00)
67
0.000
33
3.33 (3.33)
3.33 (2.40)
Thuja plicata
33
2.00 (2.00)
67
0.548
Tsuga heterophylla
4.33 (2.33)
0.000
Vaccinium parvifolium
100
3.33 (0.33)
100
3.33 (1.33)
67
67
0.638
0.37 (0.32)
0.83 (0.60)
Achlys triphylla
1.000
33
0.03 (0.03)
0
0.00 (0.00)
Adenocaulon bicolour
33
0.03 (0.03)
0
0.00 (0.00)
1.000
Blechnum spicant
Carex sp.**
33
1.000
0
0.00 (0.00)
0.03 (0.03)
0.67 (0.67)
1.000
0
0.00 (0.00)
33
Chimaphila umbellata
0
0.00 (0.00)
1.000
33
0.03 (0.03)
Corallorhiza maculata
Galium sp.**
0
33
0.07 (0.03)
2.000
0.00 (0.00)
33
0.10 (0.06)
1.732
Grass sp.**
0
0.00 (0.00)
2.67 (2.67)
1.000
Linnaea borealis
0
0.00 (0.00)
33
0.918
100
1.67 (0.67)
100
3.00 (1.53)
Polystichum munitum
0.233
Pteridium aquilinum
100
1.73 (1.63)
100
2.33(1.33)
67
0.597
100
0.87 (0.58)
0.37 (0.32)
Rubus ursinus
67
100
0.37 (0.07)
5.196
0.07 (0.03)
Trientalis borealis
67
27.33 (26.34)
100
Hylocomium splendens
27.67 (15.07)
0.011
100
100
17.00 (11.50)
1.288
Kindbergia oreganum
10.33 (6.36)
1.000
0
0.00 (0.00)
33
0.33 (0.33)
Kindbergia praelongum
Plagiothecium
D
0
1.000
undulatum
0.00 (0.00)
33
0.33 (0.33)
67
D
33
0.67 (0.67)
0.37 (0.32)
0.353
Rhytidiadelphus loreus
Rhytidiadelphus
D
0
0.00 (0.00)
33
2.67 (2.67)
1.000
triquetrus
D
Thuja plicata
0
0.00 (0.00)
33
1.000
0.03 (0.03)
* A = tree canopy stratum (> 10m height), B= the shrub stratum (0 m - 1 0 m height), C= the
B
B
B
B
B
B
B
C
C
C
C
C
C
C
C
C
C
C
C
C
D
D
D

herb stratum, and D= the moss stratum.
** Species for these genera were lumped.

The canopy above the soil moisture subplots was quantified using hemispherical photos.
The CDFmm canopy above the soil moisture sensors has fewer gaps and less open sky than

155

P
0.902
1.000
0.110
0.423
1.000
0.639
1.000
0.589
0.423
0.423
0.423
0.423
0.423
0.184
0.225
0.423
0.456
0.837
0.612
0.035
0.992
0.327
0.423
0.423
0.758
0.423
0.423

the CWHxm study sites, but the LAI values suggest that the CDFmm canopy is less dense

than the CWHxm study sites (Table 4-4). Leaf area index values ranged from 2.83-3.78, with
lower values representing open vegetation and higher values representing a dense conifer
forest.
Table 4-4: Mean and standard error (S.E.) canopy gap fraction, openness, and leaf area index (LAI)
values calculated using hemispherical photos taken directly above the three soil moisture subplots
for each of the three study sites in the Coastal Douglas-fir moist maritime (CDFmm) and Coastal
Western Hemlock very dry maritime (CWHxm).

CWHxm

CDFmm
Gap

Gap

Transect

Statistic

Fraction

Openness

LAI

Fraction

Openness

LAI

South

Mean

6.21

6.69

2.83

8.88

10.02

3.07

(n=3)

(S.E.)

(0.18)

(0.22)

(0.10)

(2.44)

(2.80)

(0.37)

Central

Mean

6.55

7.22

3.13

6.25

7.22

(n=3)

(S.E.)

(0.18)

(0.18)

(0.21)

(0.19)

(0.24)

3.78
(0.37)

North

Mean

9.10

10.10

3.13

6.91

7.73

3.40

(n=3)

(S.E.)

(0.92)

(1.01)

(0.04)

(0.60)

(0.70)

(0.18)

All sites

Mean

7.29

8.00

3.03

7.35

8.32

3.41

(n=9)

(S.E.)

(0.53)

(0.61)

(0.08)

(0.83)

(0.94)

(0.19)

The study sites varied in aspect, ranging from east to southwest aspect for CDFmm sites
and east to northwest aspect for CWHxm sites (Table 4-5). The sites on the more northfacing sites were on gentle slopes and it was determined in the field that the aspect for
these sites has minimal impacts on the vegetation and microclimate conditions for this
study. The sites range in slope from 5 % to 55 %. The steeper slopes were evaluated in the
field as ecologically equivalent to the mesic site conditions as these are in mid-slope
positions and it is inferred that they have equivalent moisture entering and exiting the site.

156

The CDFmm sites were generally at lower elevations than the CWHxm sites (as expected
from BEC mapping and characterizations), with the exception of the north transect CWHxm
site which is located northwest of the most northern extent of the currently mapped
CDFmm polygon where the CWHxm extends down to sea level.
Descriptions of the forest floors (in terms of humus form) and soil horizons in the rooting
zone (25-30 cm depth) are listed below. Forest floors were < 5 cm thick for these sites and
varied in humus form from mull to mor (Green et al. 1993). The mineral horizons were
either enriched with amorphous material (Bf) or altered by hydrolysis, oxidation, or solution
(Bm, BC MOELP and BC MOF 1998). The coarse fragment content was generally lower in the
CDFmm sites. The soil textures ranged from sand to loamy sand and the rooting depths
were from 40 cm in the CWHxm sites to over 55 cm deep in the CDFmm sites.
Table 4-5: Site and soil characteristics for each of the Coastal Douglas-fir moist maritime (CDFmm)
and Coastal Western Hemlock very dry maritime (CWHxm) study sites on the south, central, and
north transects.
CWHxm
CDFmm
Site and Soil
Characteristics
Central
South
Central
North
South

North

48”28'30.4"
123°28'58.8"

48°51'03.8"
123°37'13.3"

49°25'30.0"
124°40'00.3"

48°34'05.2"
123°39'46.3"

48'50'15.5"
123°49'23.5"

49“28'22.4"
124°48'39.2"

Middle

Middle

M iddle

Middle

Middle

Middle

Elevation (m)

226
27
75

96
55
130

75
7
42

315
13
250

335
19
238

98
5
40

Humus Form

Verm im ull

Verm im ull

M orm o der

Verm im ull

Verm im ull

H em im or

(VL)

(VL)

(RD)

(VL)

(VL)

(HR)

Bf

Bm

Bf

Bfj

Bm

Bm

40

20

25

40

60

60

Loamy Sand

Sandy Loam

Sandy Loam

(LS)

(SL)

(SL)

Latitude/
Longitude
Slope position
Aspect (")*
Slope (%)

Soil Horizon
D escription**

(25 -30 cm
depth)
Coarse
Fragment
C o n te n t(%)
Soil T e x tu re ***

(25-30 cm

Sandy Loam
(SL)

Sand (S)

157

Sand (S)

Table 4-5 (continued).

CDFmm
CWHxm
Site and Soil
Characteristics
Central
North
South
Central
South
depth)
Rooting Depth
55
45
40
40
55+
(cm)
* Aspect denoted by degrees; 0° = north, 90° = east, 180° = south, 270° = west.
**B f mineral horizon enriched with amorphous material, principally Al and Fe combined
with organic matter (Bfj, where the j denotes an expression of, but failure to meet, the
specified limits of Bf). Bm mineral horizon slightly altered by hydrolysis, oxidation, or
solution, or all three to give a change in colour or structure, or both.
* * * Sand (85 -100% sand); Loamy sand (70 -90 % sand; < 15 % clay); Sandy loam (45 - 80 %
sand; < 20 % clay)

4.4.2

Baseline Microclimate Summaries for September 15, 2010 - September 14, 2011

The microclimate data for the baseline year (September 15, 2010 - September 14, 2011)
were summarized in four time intervals: annual, seasonal, monthly, and daily. Soil moisture
values are presented as percent moisture by volume (%) and the precipitation is presented
as total sum (mm) for the specified timeframe.
The mean annual summary indicates that the CDFmm is generally warmer (in terms of
both air and soil temperatures), generally drier (in terms of precipitation and soil moisture),
significantly less humid, and significantly windier than the CWHxm (Table 4-6). The standard
error around the mean is greater for the soil moisture, air temperatures, and relative
humidity in the CDFmm, and greater for the soil temperature, precipitation, and solar
radiation in the CWHxm.

158

North
40

Table 4-6: Annual microclimate summary and paired t-test (df = 2, p < 0.05) results for the Coastal
Douglas-fir moist maritime (CDFmm, n=3) and Coastal Western Hemlock very dry maritime (CWHxm,
n=3) mesic study sites on the south, central, and north transects for the baseline year (September
15, 2010 - September 14, 2011) with the mean, standard error (S.E.), minimum (Min.) and maximum
(Max.) values and the results of the paired t-test (df=2, p<0.05). Significant differences in bold.
CWHxm

CDFmm

Microclimate
Parameter

Mean S.E.

Min.

Max. Mean

S.E.

Soil Moisture (%)

8.68

1.06

6.68

10.26 10.77

8.90

0.29

8.38

9.37

8.26

0.49

7.29

8.34

1.15

6.06

Soil Temperature
(°C)
Air Temperature
(1.5 m, °C)
Air Temperature
(0.3 m, °C)
Total Precipitation
(mm)

Paired t-test

Min.

Max.

0.38

P
10.19 11.49 2.263 0.152

7.87

0.32

7.40

8.48

1.708 0.230

8.84

7.80

0.06

7.73

7.87

0.508 0.662

9.77

8.04

0.16

7.77

8.33

0.229 0.841

t

653.7 85.7 523.9 815.6 1002.2 107.8 791.0 1145.2 1.867 0.203

Wind Speed (m/s)

0.26

0.06

0.14

0.36

0.06

0.04

0.00

0.12

6.969 0.020

Relative Humidity
(%)

87.6

1.5

84.5

89.6

92.2

1.0

90.4

93.6

7.854 0.016

11.5

0.5

10.9

12.6

12.2

2.6

8.4

17.2

0.321 0.779

Solar Radiation
(W/m2)

The CDFmm and CWHxm study sites were generally driest and warmest in the summer
months (June-September) and coldest and wettest in the winter months (DecemberMarch) for the baseline year (Figure 4-10, 4-11). Similar to the annual summary results, the
CDFmm was generally warmer and drier, with less humidity and greater wind speeds.
Exceptions to the general trends were the air temperature (1.5 m) in the fall and air
temperature (0.3 m) in the spring and fall, when the CWHxm sites were warmer than the
CDFmm sites. However, the CDFmm sites were significantly warmer than the CWHxm sites
in the winter (1.5 m, 3.18 °C and 1.95 °C respectively) and in the summer (0.3 m, 15.38 "C

159

and 14.70 °C). The soil temperature was generally warmer in the CDFmm, with a significant
difference in the winter season (4.67 °C and 3.33 °C).
The soil moisture content was generally less in the CDFmm sites, but there was not a
significant difference between the two subzone seasonal means. The variability of mean
seasonal soil moisture content is greater in the CDFmm than the CWHxm (Appendix H). The
precipitation amounts were less in the CDFmm, but without a significant difference for the
baseline year.
The relative humidity was always less in the CDFmm sites, with a significant difference in
the spring and fall seasons. The solar radiation measures were similar for the two subzones.
The CDFmm was generally windier than the CWHxm, with significant differences in the
winter (0.29 m/s vs. 0.07 m/s, respectively) and summer (0.25 m/s vs. 0.01 m/s,
respectively) season means.

160

CDFmm

CWHxm

CDFmm

CWHxm

I
CDFmm

|

Dec. 15-March 14

CWHxm

CDFmm

March 15-June 14

CWHxm

June 15-Sept. 14 ^ S e p t. 15-Dec. 14

Figure 4-10: Mean seasonal air temperature (1.5 m and 0.3 m), soil moisture, and soil temperature
microclimate (with ±1 standard error bars) for the Coastal Douglas-fir moist maritime (CDFmm) and
Coastal Western Hemlock very dry maritime (CWHxm) mesic sites (n=3) for the baseline year
(September 15, 2010 - September 14, 2011). * below the x-axis denotes significant difference
between seasonal means (paired t-test, df=2, p < 0.05).

161

CDFmm

CWHxm

CDFmm

CWHxm

0 .4 5
0 .4

UT

0.35

1

03
■ g O .2 5

<u

£ 0.2

■o

. £ 0 .1 5

§

0.1
0 .0 5

*

CDFmm

■ Dec. 15-March 14

*

CDFmm

CWHxm

March 15-June 14

*

*

CWHxm

June 15-Sept. 14 (|S e p t. 15-Dec. 14

Figure 4-11: Mean seasonal total precipitation, relative humidity, solar radiation, and wind speed
microclimate (with ±1 standard error bars) for the Coastal Douglas-fir moist maritime (CDFmm) and
Coastal Western Hemlock very dry maritime (CWHxm) mesic sites (n=3) for the baseline year
(September 15, 2010 - September 14, 2011). * below the x-axis denotes significant difference
between seasonal means (paired t-test, df=2, p < 0.05).

The baseline year was also summarized using climatic diagrams using monthly averages
for temperature and monthly total precipitation over the 2010-2011 year. The 2011 dry

162

season (implying the potential for a moisture defecit), indicated by the red dotted area,
commenced in May for the CDFmm and continued into September, while the CWHxm dry
season was shorter, beginning in early June and continuing through to late August (Figure 412). The precipitation amounts were higher in the CWHxm, resulting in a longer wet season
(implying a water surplus), denoted by the solid blue area in Figure 4-12. Both the CDFmm
and CWHxm subzones were susceptible to sub-zero air temperatures from November 2010
through to April 2011.
CDFmm

CWHxm

300

100
T>

^40

n
60

ai

Figure 4-12: W alter and Lieth climatic diagram for the CDFmm (n=3) and CWHxm (n=3) mesic sites
for the baseline year (September 15, 2010 - September 14, 2011). Blue = total monthly precipitation
(mm) and Red = mean monthly air temperature (°C). The dry season (red dotted area) occurs when
the precipitation curve undercuts the temperature curve; the moist season (blue vertical lines)
occurs when the precipitation curve supersedes the temperature curve, while the solid blue area
denotes a w et period. The blue bars below the x-axis indicate potential frost period.

The January soil temperature (4.57 °C and 3.12“C for the CDFmm and CWHxm,
respectively), air temperature (1.5 m, 3.54 °C and 2.19°C, respectively), and wind speed
(0.23 m/s and 0.06 m/s, respectively) were significantly higher in the CDFmm than in the
CWHxm (Appendix I). The February soil temperature (4.44 °C and 3.28°C) and wind speed
(0.30 m/s and 0.06 m/s) remained significantly higher in the CDFmm. The air temperature
163

(1.5 m, 5.24°C and 4.13°C) in March was again significantly greater in the CDFmm than in
the CWHxm. In April, the relative humidity (85.3% and 92.8 %) was significantly lower in the
CDFmm than in the CWHxm, and remained significantly lower (78.5 % and 90.9 %) through
May when the air temperature (0.3 m, 9.81 °C and 9.01°C) was also significantly higher in
the CDFmm. There were no significant differences in microclimate measurements in the
month of June. The wind speed was significantly higher in the CDFmm for July (0.28 m/s
compared to 0.01 m/s) and August (0.23 m/s vs. 0.01 m/s). The air temperature (0.3 m,
16.44 °C and 15.30°C) and soil moisture (3.44 % and 6.27 %) were significantly different in
August, with warmer and drier conditions in the CDFmm. There were no significant
differences in microclimate in the months of September, October, and November. The soil
temperature (5.73 °C and 4.19°C) and wind speed (0.30 m/s and 0.09 m/s) were significantly
greater in the CDFmm than in the CWHxm in December.
Daily mean time series for soil moisture content, air and soil temperatures, and
precipitation were compared for the CDFmm and CWHxm, as presented in Figures 4-13 and
4-14. The CDFmm soil moisture content increased during the months of September through
to March when precipitation events were frequent and the air and soil temperature were
decreasing (Figure 4-13). Then, in April, the precipitation events became less frequent and
the air and soil temperatures increased, resulting in a decrease in soil moisture over the
growing season months (May - September). Soil temperature was greater than air
temperature during the colder months (October - March). The soil and air temperature are
similar in April, then the air temperature increased above the soil temperature in the
growing season.

164

C
D
F
m
m

*JL
OCT

NOV

DEC

JAN

M AR

FEB

APR

M AY

JUN

JUL

AU G

•10

Soil Moisture {%)

Soil Temperature (C)—

Air Temperature ( C

)

Precipitation (m m /day)

Figure 4-13: Time series of the mean daily soil moisture, soil temperature, precipitation, and air
temperature for the Coastal Douglas-fir moist maritime (CDFmm) mesic study sites (n=3) for the
baseline year (September 15, 2010 - September 14, 2011).
The CWHxm soil moisture content was greatest during the months from September
through May (Figure 4-14). The soil moisture content remained high during June and into
July. By July 15, the soil moisture content began decreasing until mid-September. Daily
precipitation values were greatest during the months of December through March. The
frequency of precipitation events decreased in July, when the decrease in soil moisture
content began and air and soil temperature were increasing.

165

SEP

C
W
H
x
m

OCT

NOV

DEC

FEB

M AR

APR

JUN

M AY

JU l

AUS

SEP

-10

—

•Soil Moisture (%)

Soil Temperature (C)

Air Temperature ( C

)

Precipitation (m m /day)

Figure 4-14: Time series of the mean daily soil moisture, soil temperature, precipitation, and air
temperature for the Coastal Western Hemlock very dry maritime (CWHxm) mesic study sites (n=3)
for the baseline year (September 15, 2010 - September 14, 2011).
Note that the trends for the CDFmm and CWHxm are similar, with more frequent
precipitation, cooler temperatures and increased soil moisture content in the winter, and
then drier and warmer trends in the summer for both subzones. Note the spikes in
precipitation, particularly in the growing season, followed immediately by spikes in the soil
moisture content. When the soil is dry, the rate of precipitation infiltration into the soil
profile is indicated by the subsequent increase in soil moisture content during the growing
season. Also note that as air and soil temperatures increase in spring and precipitation
decreases, the soil moisture content decreases and continues to decrease until the fall rains
and cooler temperatures return, even with the spikes in soil moisture after the more
infrequent precipitation events that occur in the growing season.

166

4.4.3

Growing Season Microclimate Summaries: Comparisons of the Coastal Douglas-fir
moist maritime (CDFmm) and Coastal Western Hemlock very dry maritime (CWHxm)
subzones

The daily means for the 2011 growing season (May 15 - September 14) were tested for
significant differences between the CDFmm and CWHxm (df = 2, p < 0.05). Significant
differences between subzone microclimate data were detected in the daily mean air
temperature (0.3 m), and wind speed (Table 4-7).
Table 4-7: Growing season (May 15 - September 14, 2011) mean and standard error (S.E.) and
paired t-test results testing for a difference in the mean for each microclimate variable (df =2, p <
0.05) in the Coastal Douglas-fir moist maritime (CDFmm, n=3) and Coastal Western Hemlock very
dry maritime (CWHxm, n=3) mesic study sites. Significant results in bold.
CDFmm

CWHxm

Paired t-test

Microclimate variable

Mean

S.E.

Mean

S.E.

t

Soil Moisture (%)

6.02

0.99

8.45

0.48

2.407

P
0.138

Soil Temperature ("C)

12.63

0.24

12.40

0.29

0.442

0.702

Air Temperature (1.5 m, °C)

14.67

0.49

14.28

0.11

0.715

0.549

Air Temperature (0.3 m, °C)

14.39

0.91

13.78

0.94

11.011

0.008

Precipitation (mm/day)

38.5

11.5

58.1

9.1

1.396

0.298

Wind Speed (m/s)

0.26

0.06

0.01

0.01

4.467

0.047

Solar Radiation (W/m2)

22.5

2.6

21.7

5.2

0.267

0.817

Relative Humidity (%)

78.5

2.4

83.6

0.9

3.195

0.086

Any extreme microclimate conditions experienced at the CDFmm and CWHxm mesic
sites during the growing season generally occurred within a day of each other (Table 4-8).
Exceptions include the relative humidity minima (September for CDFmm and July for the
CWHxm), and the soil temperature maxima (September for the CDFmm and August for the
CWHxm), wind speed maxima (May for the CDFmm and June for the CWHxm), and relative
humidity maxima (July for the CDFmm and May for the CWHxm, Table 4-8). The CDFmm
microclimate extremes generally occurred in the early and late months of the growing
167

season (May and September), except for the high air temperatures in August and relative
humidity in July. The microclimate extremes in the CWHxm also occurred in the early and
late months of the growing season except for the soil and air temperatures peaking in
August, wind speed highs in June, and the low relative humidity in July.
Table 4-8: Comparisons of the extremes (minimum and maximum) for the mean daily growing
season microclimate attributes in the Coastal Douglas-fir moist maritime (CDFmm, n=3) and Coastal
Western Hemlock very dry maritime (CWHxm, n=3) mesic study sites, between May 15, 2011 and
September 14, 2011. N/A values indicate multiple occurrences of minimum or maximum values.
CWHxm

CDFmm
Microclimate
Variable

Soil Moisture (%)
Soil Temperature
(°C)
Air Temperature
(1.5 m, °C)
Air Temperature
(0.3 m, °C)
Wind Speed (m/s)
Relative Humidity
(%)
Solar Radiation
(W/m2)
Total daily
precipitation (mm)

Minimum
(Date)
2.65
(14-Sep)
8.72
(18-May)
8.77
(17-May)
8.73
(17-May)
0.03
(15-Sep)
61.22
(08-Sep)
10.68
(13-Sep)
0.00
(n/a)

Maximum
(Date)
10.96
(15-May)
15.20
(12-Sep)
20.44
(21-Aug)
20.25
(21-Aug)
0.64
(16-May)
98.65
(17-Jul)
89.85
(19-May)
15.62
(15-May)

Minimum
(Date)
4.30
(14-Sep)
7.85
(17-May)
7.70
(17-May)
8.11
(17-May)
0.00
(n/a)
68.20
(05-Jul)
4.37
(13-Sep)
0.00
(n/a)

Maximum
(Date)
13.89
(15-May)
14.45
(27-Aug)
19.78
(21-Aug)
16.93
(21-Aug)
0.19
(29-Jun)
100.00
(15-May)
96.97
(19-May)
15.53
(15-May)

Soil moisture persistence within the rooting zone (25 cm - 30 cm depth) was calculated
using sequential autocorrelation function values computed using the autocorrelation
function in the stats package for R statistical software (R Core Development Team 2010) at a
set lag interval of one day (Figure 4-15,4-16).

168

The autocorrelation function (ACF) values range between +1 and -1. Values of +1 indicate
that the soil moisture content is positively correlated with the content at the beginning of
the growing season. A gradual decrease away from the +1 ACF value suggests a decreasing
trend in the relationship, and a value approximating 0 ACF indicates random soil moisture
content (e.g., no trend in the soil moisture content into the growing season). The dotted
lines on Figures 4-15 and 4-16 represent the 95% confidence interval for the autocorrelation
function.
The soil moisture persistence is determined by the number of days the soil moisture
value is significantly autocorrelated with the soil moisture content measure at the beginning
of the growing season (May 15). When the ACF decreases below the 95% confidence
interval (dotted line), the autocorrelation of the soil moisture content on that day is no
longer significantly correlated with the soil moisture content on May 15 (day 0).
The soil moisture ACF scores were significant for 28 days in the CDFmm (Figure 4-15).
The trend in decreasing ACF values suggests that the soil moisture content became less
correlated with the soil moisture content of day 0 over time.

169

o
UL

O

oo

d

<
C

o

<o

•4 -*

d

c

d

<D

CM

O
c
u
<*-

o

3

<

d
o

d
CM

0

10

30

20

40

50

Time Lag (days)
Figure 4-15: Soil moisture persistence in the growing season on the Coastal Douglas-fir moist
maritime (CDFmm) mesic sites (n=3). Blue dotted line is the 95% confidence interval for a 0
(random) ACF score. The x-axis (time lag) denotes the number of days for which the soil moisture
persists.
The soil moisture ACF is significant for 33 days in the CWHxm (Figure 4-16). Similar to the
CDFmm, the trend in decreasing ACF values suggests that the soil moisture content is
becoming less correlated with the soil moisture content of day 0 over time.

170

o
UL

O

00

o

<
C

O <o
d
O
C
3
*•—
-4 -4

C

O
00
2
t_

o

o

o
o
4->
3

o
ci

<

<N
0

30

20

10

40

50

Time Lag (days)
Figure 4-16: Soil moisture persistence in the growing season on the Coastal Western Hemlock very
dry maritime (CWHxm) mesic sites (n=3). Blue dotted line is the 95% confidence interval for a 0
(random) ACF score. The x-axis (time lag) is the number of days for which the soil moisture persists.
The 2011 growing season comparison of soil moisture persistence in the CDFmm and
CWHxm indicates that there is less soil moisture available to vegetation on the CDFmm
mesic sites than that growing on the CWHxm mesic sites.

4.5

DISCUSSION AND APPLICATIONS

4.5.1

Microclimate Data

The CDFmm and CWHxm mesic site microclimate data support current concepts of the
distinction between Coastal Douglas-fir moist maritime and Coastal Western Hemlock very
dry maritime subzones as described by the BEC system (Meidinger and Pojar 1991). Overall,

171

the CDFmm is warmer and drier than the CWHxm all year round (Table 4-6), and particularly
warmer and drier in the summer season (Figure 4-10,4-11) with a significant difference in
soil moisture in August (Appendix I). Interestingly, the wind speed in the CDFmm is greater
than that in the CWHxm. This could be because of its proximity to the ocean and its greater
exposure to maritime winds.
The trends of daily soil moisture, soil temperature, air temperature, and precipitation are
similar in the two subzones when presented in time series (Figure 4-13, 4-14). On the other
hand, the CDFmm soil moisture, wind speed, and relative humidity range is greater and the
CWHxm range in temperature and precipitation is greater than the corresponding ranges in
the CDFmm. The CDFmm demonstrates a more maritime climate, moderated by the Pacific
Ocean, while the CWHxm, with higher elevation and located further inland on Vancouver
Island, experienced greater highs and lows for a number of microclimate attributes
throughout the 2010-2011 baseline year.
The climatic diagrams portray a longer dry season in the CDFmm due to lower
precipitation in June and August, 2011. The precipitation amounts in the CWHxm were
greater than those of the CDFmm. The CDFmm frost-free period was longer in 2011 than
shown in the published climatic diagram for the CDF (Meidinger and Pojar 1991). The dry
season endures through June to August in the historical climatic diagram. The monthly
temperature was similar in 2011 compared to the 1991 published figure, but the overall
precipitation amount was greater in 2011. The published figure (Meidinger and Pojar 1991)
was derived from a multi-year data set from one location (Victoria airport), while the 2010-

172

2011 figure is only one year of data but for three CDFmm locations (Victoria, Duncan, and
Bowser).
The growing season comparisons suggest that there is not a significant difference
between the two subzones for mean soil moisture, soil and air (1.5 m) temperature, solar
radiation, and relative humidity, but there were significant differences observed for air
temperature at 0.3 m above the ground surface and wind speed at ca. 3 m above the
ground surface. The timing of the extreme (minimum and maximum) values in CDFmm and
CWHxm forest microclimates conditions was similar, with such events generally occurring
within one day of each other.
The persistence of soil moisture into the growing season and the amount of precipitation
falling in each subzone is less for the CDFmm than the CWHxm. This suggests that the
vegetation within the CDFmm is more likely to experience drier conditions during the
growing season, again supporting the current classification of the two forest subzones. The
soil moisture content from the winter and spring rains persists for 28 days in the CDFmm
and 33 days in the CWHxm, after which the soil moisture status is more dependent on
summer rains. Note that there were precipitation events within the 28 and 33 days (Figure
4-12,4-13), but they were not sufficient in frequency nor amount to maintain soil moisture
content. Soil moisture conditions in both subzones were drier in the growing season than
during the rest of the year, and continue to get drier as the season develops. The vegetation
in both forest types is reliant on precipitation events to provide water to the rooting zone
during the growing season.

173

The evaluation of one year of microclimatic conditions does not consider the climate,
soil, and vegetation feedbacks. The plant-soil feedback varies across environmental
conditions (Ehrenfeld et al. 2005). Models can suggest potential outcomes for climatemediated shifts in microlimate conditions, but uncertainty around the outcomes based on
the relationship between evapotranspiration rates and plant survival strategies for different
plant functional types, particularly under water stress, requires intensive field data (Detto et
al. 2006). Research on the complexity of hydrological processes and soil moisture
distribution across spatial and temporal scales is continuously evolving at the rate with
which technology (e.g., computer programs and analysis capacity) is developed. Modelled
outputs are more reliable with a well-calibrated forcing data set. Continued field
measurement and monitoring is important for model calibration and is also required to
increase our understanding of hydrological processes, soil moisture dynamics, and the
relationships of climate, soil, and vegetation. Based on even one year of microclimate
comparisons, it is clear that growing conditions for vegetation differ subtly between the two
subzones. Climate envelope projections suggest these ecosystems will be less similar in
future timeslices (Chapter 2). Long-term field sampling of these conditions will improve our
understanding of climate, soil, and vegetation relationships and increase our ability to
evaluate the impacts of future climatic change.

4.5.2

Research and Management Applications

The microclimate data and the network of monitoring sites described here are available
to support other research and management questions in the CDFmm and CWHxm.

174

Currently field studies are being conducted on the study sites by research scientists and
graduate students on the decomposition of fine woody materials, soil and vegetation
nutrient content, growth and vigour of salal (Gaultheria shallon), and the relationships
between climate and Douglas-fir growth (Klassen et al. 2010). Combining the findings for all
research questions on these sites will improve our understanding of climate, microclimate,
soil and vegetation relationships, and ecosystem function. The results will provide the data
for further water balance modelling associated with a range of science and management
questions (e.g., climate change, fire regimes, and soil water balance).

4.5.3

Long-term Data Collection and Monitoring Plan

This chapter describes the first step in quantifying and comparing site types near the
transition zone between the CDFmm and CWHxm. This research project will continue on all
15 study sites (including the mesic, transition, and edaphic sites) to meet both short-term
objectives for evaluating climate, soil, and vegetation relationships and long-term objectives
associated with modelling soil water balance and monitoring sites for impacts of climatic
change (Table 4-9).

175

Table 4-9: Short-term and long-term objectives for the climate change monitoring network
established in the Coastal Douglas-fir moist maritime (CDFmm) and Coastal Western Hemlock very
dry maritime (CWHxm) study area on southeastern Vancouver Island, British Columbia.
Short-term objectives

> validation of the current subzone and site series classification within the BEC
system for the dry coastal ecosystems
> calibration of relative and absolute soil moisture regime determinations
> assessments of the correlation between soil moisture, climate, and vegetation
parameters
>

improved ecosystem mapping

>

investigate the availability of soil moisture across topo-edaphic conditions

>

provide infrastructure for other field research projects to increase our
understanding of climate, soil, and vegetation relationships

Long-term objectives

>

define the relationship between actual soil moisture availability and
microclimate data and terrain and soil features, thereby facilitating projection
of soil moisture regime shifts with changing climate

> explore the potential use of soil moisture sampling as an operational indicator
of climatic change
>

detect and report indicators (microclimate, soil water balance, vegetation) of
changing climate in a transition zone expected to be responsive to such
climate shifts

>

improved climate change modelling

> test hypotheses of change in site water balance with changes in climatic
regime
> testing hypothesized trajectories of change in ecosystem variables through
correlative study

The CDFmm is characterized by a soil moisture deficit during the growing season (Green
and Klinka 1994) and this makes it unique among other forest types on the BC coast. This
chapter describes the establishment of a series of monitoring transects through the
latitudinal gradient of the CDFmm. Each transect starts in the heart of the CDFmm and
crosses over the transition area (ecotone) into the adjacent CWHxm. Subzone differences

176

and the transition from CDFmm to CWHxm is quantified using the mesic sites on each
transect. Mesic sites are where soil moisture conditions are primarily controlled by climate
(Kimmins 1987). These sites are monitored because they are representative of regional
climate (Meidinger and Pojar 1991). Future work will include further monitoring and
analysis across a range of soil moisture conditions in the CDFmm subzone.

4.6

CONCLUSIONS

Forest ecosystems in many parts of the world are experiencing changes in climate, with
global air temperatures increasing by 0.8 °C (0.35 °C per decade in the northern hemisphere
since 1970) and precipitation increasing by 3-5% over the last century (Boisvenue and
Running 2006, Christensen et al. 2007). Uncertainty in feedback mechanisms in the system
suggests that every increase in precipitation does not equate to increased moisture
available for vegetation, as increases in temperature also result in increased evaporation
(Waring and Running 2007). Uncertainty in regional impacts results in uncertainty regarding
rates of change, such as how the increase in temperature could also result in decreased
water storage in the form of snow and ice, further altering soil moisture regimes and
hydrologic processes that impact forest ecosystems.
All the uncertainty points to the need for a better understanding of boundary layer
climate and water balance relationships. Hydrological processes and other physical and
biological feedbacks are driving factors determining the composition, productivity, and
functioning of ecosystems on the Earth's surface. To increase our ability to quantify and
comprehend the complexity of interactions involved in hydrological processes and soil

177

moisture regimes, models are being developed to examine components of the soil-plantatmosphere system and the associated processes that define the relationships between the
components. The uncertainty and assumptions of models are due to the stochastic nature
of weather (particularly precipitation events), gaps in scientific knowledge, and the lack of
field data to quantify spatial and temporal differences in climate, microclimate, soils, and
vegetation. A long-term monitoring network is now in place to facilitate quantification and
monitoring in BC's dry coastal forests.

178

5.

Current and Future Biogeography of the Coastal Douglas-fir Zone on
Vancouver Island: Synthesizing Climatic Envelope Models and
Microclimate Field Study Findings

5.1

INTRODUCTION
Research on climate change suggests that shifts in ecosystem and species distributions

will occur in coastal British Columbia (BC) forest ecosystems. This study focused on the
southwest corner of the province, completing climatic envelope models for the Coastal
Douglas-fir (CDF) biogeoclimatic unit and its diagnostic plant species on Vancouver Island,
small islands of the Salish Sea, and the adjacent mainland BC (Figure 3-1). The potential
impacts of climatic change to the relatively dry coastal forests, as classified in the BC
Biogeoclimatic Ecosystem Classification (BEC) system (Green and Klinka 1994), were
explored using climatic envelopes to investigate how climatic change will impact the
distribution of suitable climate space (SCS). Suitable climate space is defined as the
geographic area that is within the defined range of the climatic envelope for the ecosystem
(Rose and Burton 2011). The climatic envelope is an empirical definition of the climatic
conditions acceptable for an ecosystem based on its current spatial extent and the climate
attributes associated with the ecosystem at its known locations. Climatic envelopes were
constructed using climate and ecosystem models and data inputs that were analysed with
geographic information system (GIS) and statistical software.
The subzone units are the foundation of the BEC system and are delimited by regional
climate (i.e., without the effect of local topography). The Coastal Douglas-fir moist maritime
(CDFmm) and the Coastal Western Hemlock very dry maritime (CWHxm) subzones are
179

distributed within the rainshadow of the Vancouver Island and Olympic mountain range,
resulting in a Mediterranean-type climate regime. The CDFmm climate regime is unique to
the remainder of the BC coast, but is also found south of the BC border, in the Pacific
Northwest United States (PNW US). To facilitate climatic envelope modelling of dry, coastal
ecosystems as classified in BC, the BEC ecosystem unit concept was extended into the PNW
US using ecological plot data.
The range of suitable climate space for the extended CDF and CWHxm was projected for
baseline (1961-1990) timeslice using ClimateWNA (Wang et al. 2010) interpolated climate
normals and future (2010-2039, 2040-2069, 2070-2099) timeslices using a five-model global
climate model (GCM) ensemble (Chapter 2). Climatic envelopes for 18 species which occur
together in mature and old-growth forests and are considered diagnostic of zonal sites in
the CDFmm subzone were also projected for the southwestern BC study area for the four
timeslices (Chapter 3). The species SCSs were overlaid and the area common to all 18
species SCSs was identified as the species overlap SCS area, and compared to the SCS for the
CDF as a whole.
The project also included the establishment of a long-term monitoring network of study
sites on Vancouver Island. Five study sites on three transects were established (Figure 4-3).
The study design includes sites across a soil moisture gradient from the CDFmm through the
transition zone, into the CWHxm. The 2010-2011 baseline microclimate was summarized for
two sites on each transect, the mesic CDFmm and mesic CWHxm, as representatives of
regional climate for each subzone.

180

The objective of this chapter is to synthesize the findings of the three previous thesis
chapters: (1) the ecosystem-based climatic envelopes, (2) the species climatic envelopes,
and (3) the baseline microclimate measures, with recommendations to apply the results to
management and further research.

5.2

SYNTHESIS DISCUSSION

5.2.1

Extending the Definitions of the Coastal Douglas-fir Biogeoclimatic Zone

Impacts of climate change have been projected for ecosystems and vegetation on the
coast of British Columbia (BC) using climatic envelopes for ecosystems (Hamann and Wang
2006, Campbell and Wang in press, Wang et al. in press) and for species (Thuiller 2003,
Rehfeldt et al. 2006, McKenney et al. 2007, Iverson et al. 2008). Preliminary ecosystem
climatic envelope definitions were based on BEC mapping without extending the ecosystem
types into the south beyond the BC border. Early results of that sort of analysis have
consequently produced suitable climate space (SCS) along the southern boundary of BC for
ecosystems with no current analog in the province.
In Chapter 2 of this thesis, the BC ecosystem classification was extended beyond the
southern boundary of the province using ecological mapping from BC and ecological plot
data from the Pacific Northwest (PNW) of the United States (US) to reduce the area of
undefined (no-analog) ecosystem SCS in southern BC. Using BEC criteria and definitions for
the CDF (Green and Klinka 1994), a vegetation assemblage indicating warmer and drier
climate conditions, provisionally named the CDF dry maritime (CDFdm) subzone, was
identified in the US PNW. Collectively, the CDFmm and CDFdm are referred to as the

181

"extended CDF". The projections using the extended CDF tested the hypothesis that the
undefined SCS found in previous studies may be suitable to more southern expressions of
the CDF ecosystem (CDFdm), currently found south of the BC border in Washington and
Oregon. Subsequently, this approach still projected the emergence of an undefined SCS on
the southern tip of Vancouver Island as soon as the 2040-2069, in areas previously within
the extended CDF SCS. When the climate variables within the undefined SCS were explored,
it was concluded that the combination of climate variables in that space is unlike any
climatic regime currently found in coastal BC. This reaffirms the need to complete
international ecosystem classification systems that will facilitate future efforts in climate
change modelling in cross-boundary areas. It also raises the possibility that future
conditions in BC's CDF zone may be unlike conditions currently experienced anywhere on
the PNW Coast.

5.2.2

Climatic Envelope Projections for the Study Area

The five-model global climate model (GCM) ensemble climate projections for the BC
study area (Vancouver Island, adjacent areas of the mainland, and the Salish Sea islands;
Figure 3-1) project increases in air temperature variables such as mean annual temperature
(MAT), mean temperature of the warmest month (MWMT), mean temperature of the
coldest month (MCMT), extreme minimum temperature (EMT), and the temperature
difference between MWMT and MCMT (TD) from historic norms (1961-1990) to the 20702099 timeslice (Table 5-1, Appendix E). Mean annual precipitation (MAP) is projected to
increase over time, yet the mean summer precipitation (MSP) is projected to decrease from

182

the historic baseline to the 2070-2099 timeslice. Such changes suggest a shift in climate that
will result in warmer, drier, and longer growing seasons and warmer and wetter winters in
southwestern BC. The combination of changes projected to emerge in the 16 climate
variables evaluated in this research is beyond the suitable range for biogeoclimatic
subzones currently found in southwestern BC. Those novel conditions are also beyond the
climatic envelope definitions of an extended CDF, including the CDFmm and CDFdm
subzones based on vegetation assemblages found in the US (Chapter 2).

Table 5-1: Global climate model (GCM) five-model ensemble mean annual climate variables and the
overall change in climate variables for the baseline (1961-1990) and projected for future (20102039, 2040-2069, and 2070-2099) timeslices for the southwestern British Columbia study area.
Climate
Variable*
MAT (°C)
MWMT(°C)
MCMT(°C)
TD(°C)
MAP(mm)
MSP(mm)
AHM
SHM
DD<0
DD>5
NFFD
FFP
PAS(mm)
EMT(°C)
Eref
CMD

Baseline
5.6
14.0
-1.7
15.7
2809
575
6.6
30.2
402
1211
211
140
742
-26.4
545
65

Timeslice
2010-2039
2040-2069
6.7
7.8
17.0
15.5
-0.7
-1.5
17.7
16.9
2878
2846
532
553
7.4
7.0
35.1

40.9

350
1471

295
1754

229
158
636
-25.2
583
85

246
179
542
-23.5
618
110

2070-2099
9.3
19.0
0.7
18.3
2967
526
7.7
46.1
214
2127
268
203
405
-20.7
663
135

Overall Change
(Baseline to 2070-2099)
3.7
5.0
2.4
2.6
159
-49
1.2
15.9
-188
916
57
63
-336
5.7
119
70

* Climate variables are defined in Appendix B.
The extended CDF SCS closely matches the current distribution of the CDFmm in the BC
study area. The SCS for the neighbouring (more northerly and higher elevation) very dry
183

maritime subzone of the Coastal Western Hemlock biogeoclimatic zone (CWHxm) expands
beyond the currently mapped CWHxm boundary (BC MFR 2008), suggesting that climate is
not the only factor determining CWHxm distribution. The SCSs for individual plant species
closely match the current species distribution maps for the study area (Pojar and
MacKinnon 1994) but the species overlap area, aimed to represent the CDFmm, extends
beyond the currently mapped CDFmm. Climatic envelopes, taken alone, generally represent
the fundamental climatic niche and overestimate the realized niche of ecosystems and
species (Rehfeldt 2004). These results reflect the fact that the distributions of species and
ecosystems are not solely determined by climate, but represent a snapshot in time,
reflecting the current response of vegetation to terrain, soils, biogeographic history,
competition, and disturbance regime as well. Nonetheless, the climatic envelope technique
used here models the potential full geographic extent of the climatic niche, now and in the
future.
The distribution of the extended CDF SCS is projected to decrease in size and to shift
northward by the end of this century. The CWHxm SCS is also expected to decrease in area
and to shift northward, away from its current locale and the baseline projection of areas
with suitable climate for the CDF and CWHxm (e.g., the CWHxm climatic envelope is no
longer adjacent or overlapping the extended CDF climatic envelope). Model projections
suggest that locations currently occupied by CWHxm could become more similar to
conditions which characterize the extended CDF climatic niche. Projections for both the CDF
and the CWHxm have some occurrence on mainland BC, but are primarily distributed on
Vancouver Island and the smaller islands in the Salish Sea, now and in the future time

184

periods. The overlap area for the CDF and CWHxm SCS is projected to decrease over time. In
the baseline projections, the CWHxm SCS includes a large portion of the CDFmm as
currently mapped (BEC v.7) and the extended CDF SCS, with the exception of the southern
tip of Vancouver Island and a small portion of the small islands of the Salish Sea. By the end
of the century, the overlap is projected to be reduced to 15 km2, with the two SCSs adjacent
to one another as per the current BEC mapping. If such projections are reliable, this
suggests that the ecosystem transition may become more distinct with a steeper climate
gradient between the two ecosystems.

5.2.3

Some Insights into Plant Community Organization and Distribution

Ecologists have long debated the fundamental basis for successional progression and
environmental organization of forest ecosystems. There are theories of plant community
composition developing as cohesive units, strongly determined by regional climate and time
since disturbance (Clements 1907,1916). Opposing theory states that plant associations
result from the independent distributions and environmental tolerances of individual
species (Gleason 1917,1926). The paleo-ecological record unequivocally demonstrates that
tree species have responded individually to past shifts in climate (Davis 1986). The 18 plant
species that are diagnostic of zonal sites in the CDFmm subzone were each individually
modelled using climatic envelopes (Chapter 3). The species overlap area for the 18
individual species were compared to the extended CDF climatic envelope projections
(Chapter 2) in the southwestern BC study area. The species SCS overlap covered a larger
geographic area (ranging from 31374 km2 to 43 156 km2 over the four timeslices) than that

185

modelled for the CDFmm as a whole (projected to have a smaller SCS of only 5 453 km2 to
10 100 km2 over the same time periods). The area of net SCS overlap among 18 individual
species still defines a surprisingly broad climatic envelope, though acknowledged to exclude
biological complexity known to determine the presence of species on the ground (Iverson et
al. 2009). In contrast, the ecosystem climatic envelopes are indirectly representing the net
interactions of all the species (not just the diagnostic individuals within the plant
community) and with a number of unidentified site factors and environmental processes,
not just the climatic tolerances of diagnostic species within the plant community. Therefore,
while the results suggest that ecosystem climatic envelopes do a better job of representing
cumulative biological complexity, there is no guarantee that the same set of (unidentified)
biotic and abiotic constraints determining ecosystem distribution today will extend into the
future. Perhaps the climatic constraints (envelopes) for these various ecosystem-level
processes and functions need to be determined too, with the SCS for such processes then
allocated among the plant species available at the same geographic locations.
Though the area of the species overlap SCS and the extended CDF SCS differ, the
directional trend projected for both is a shift northward by the end of the century, away
from of the southeast end of Vancouver Island, and out of the area currently mapped as the
CDFmm (BC MFR 2008). The northward migration matches other modelled projections in
North America where over 100 species were projected to shift northward, with more
successful tree regeneration at the northern edge of the species range (Hamann and Wang
2006, McKenney et al. 2007, Iverson et al. 2009, Wang et al. in press).

186

The individual species climatic envelopes suggest that species common to the CDFmm
will not be affected equivalently over time. Of the 18 diagnostic species, the SCS for 11
species (Acer macrophyllum, Arbutus menziesii, Pseudotsuga menziesii, Thuja plicata,
Cornus nuttallii, Holodiscus discolor, Mahonia nervosa, Rosa gymnocarpa, Symphoricarpos
albus, Polystichum munitum, and Pteridium aquilinum) is projected to increase over time,
while the SCS is projected to decrease for two species (Abies grandis and Hylocomium
splendens). The SCS for the five remaining species (Achlys triphylla, Gaultheria shallon,
Lonicera ciliosa, Kindbergia oregana, and Rhytidiadelphus triquetrus) is projected to initially
increase, and then decrease in size by the end of the century. Though the overall changes in
SCS distribution within the BC study area varies, the general northward shift in distribution
is demonstrated in all species projections (Figure 3-4). The range of suitable climate space
for Abies grandis, Achlys triphylla, Gaultheria shallon, Hylocomium splendens, Kindbergia
oregana, Lonicera ciliosa and Rhytidiadelphus triquetrus demonstrate the most dramatic
potential for northerly shifts in SCS, resulting in a loss of SCS in the study area.

5.2.4

Climatic Constraints on Ecosystem and Species Distributions

Climatic envelopes provide insight into ecosystem threshold responses and species
sensitivity to changes in climate (Wang et al. in press). The constraint analysis for the
extended CDF SCS boundary on Vancouver Island suggests that relationships between
climatic and geographic factors, such as the interaction of temperature and latitude or
temperature and elevation, are important in delineating ecosystem SCS. Along the 250 km
boundary of the baseline climatic envelope of the CDFmm on Vancouver Island, the climate

187

variables that constrain the CDFmm SCS shifts from moisture- and warmth-related climate
variables (summer heat: moisture index, SHM, climatic moisture deficit, CMD, annual heat:
moisture index, AHM, mean annual temperature, MAT) in the south to variables indicative
of coldness (frost-free period, FFP, extreme minimum temperature, EMT, number of frostfree days, NFFD, mean cold monthly temperature, MCMT) in the north (Figure 2-19).
The sensitivity of individual species SCS to climate variables was assessed using
controlled adjustments of each climate variable. This analysis investigates both the impact
of uncertainty in GCM projections and the impact of projected shifts in climate attributes to
the species SCS. The temperature-related climate variables seem to have less impact on the
SCS for individual species, while the precipitation- and moisture-related variables appear to
have more of an impact on the distribution of species SCS, with CMD adjustments impacting
12 of the 18 species (Table 3-4). However, the precipitation projections stemming from
even the most reliable GCMs are more variable and unreliable than the temperature
projections (Randal et al. 2007). The range of suitable climate space for Cornus nuttallii was
most sensitive of all 18 species, with adjustments of any of the 16 climate variables used in
the climatic envelope definitions resulting in changes in the area of suitable climate space
(Table 3-4). The sensitivity analysis emphasizes that changes in individual climate attributes
impact individual species SCSs differently, further suggesting that a uniform shift of discrete
plant communities is not likely based on climate-species relationships (Table 3-4).

188

5.2.5

Establishment of a Forest Microclimate Monitoring Network

Most climate change studies currently focus on the present conditions and how they may
change in the future. Other approaches investigate correlations between ecosystem
attributes and climate indicators by reconstructing the historical context that is interpreted
from tree rings, fossils, seed banks, glaciers, and soil sediments. In populated areas, there
are climate records (usually limited to temperature and precipitation measures) that date
back 100 to 150 years (usually limited to developed areas and low elevations); such direct
records are lacking for most of the globe. Consequently, most research seeking to
understand and project the climatic relationships for unmanaged ecosystems is requesting
and recommending further field studies to generate climate records representative of a
greater variety of conditions (e.g., forests, alpine, coastal areas, and other remote
geographic areas) to improve ecological models and to validate model outputs. Increases in
the amount and variety of climate data available will also facilitate the future modelling of
meso-scale and micro-scale systems. In response to these needs, a long-term monitoring
network was established in the CDFmm and CWHxm on Vancouver Island (Chapter 4).
Consisting of nine stations located on mesic forest sites in CDFmm, CWHxm, and transitional
locations, it is supplemented by another six stations situated in drier and wetter forest sites
at locations near the CDFmm mesic sites. This network is well positioned to track
microclimate differences and changes across a soil moisture gradient throughout the
current distribution of the CDFmm and into the adjacent CWHxm subzone.
The field data on site conditions and microclimate differences reported in Chapter 4 are
limited to the 2010-2011 baseline year (September 15, 2010 - September 14, 2011).

189

Preliminary summaries indicate that there is a significant difference in several microclimate
attributes between mesic sites in the CDFmm and CWHxm subzones. The mean annual
microclimate data indicates that the CDFmm is warmer and drier both above and below the
ground surface. The mean seasonal data suggest similar trends, although the CWHxm had
higher air temperatures in the fall season (September 15 - December 14, 2010). The soil
moisture content is similar for the mesic sites in the winter season (January 15 - March 14,
2011), and then the CDFmm mesic sites become drier (on average) than the CWHxm mesic
sites. These preliminary results confirm the value of soil moisture as a great integrator of
microclimate (Rodriguez-lturbe 2000), important in explaining ecological differences across
the study area. Soil moisture is, on average, lower in the CDFmm, reflecting higher air and
soil temperatures and lower precipitation than in the CWHxm, resulting in the characteristic
soil moisture deficit of the CDFmm. This suggests that the soil moisture content limits or
strongly influences the CDFmm ecosystem and that soil moisture is an appropriate measure
to assess and monitor impacts of climate change, and may serve as a sensitive indicator of
any forthcoming shifts in climate in this region. The soil moisture content CDFmm differed
significantly from the CWHxm during August 2011, but continued monitoring to re-evaluate
soil moisture hypotheses with multiple year data sets is recommended.
The 2010-2011 baseline field data for mean annual temperature (MAT) and mean annual
precipitation (MAP) values at the mesic sites fit the climate envelope definitions created
using current ecosystem distribution data and current (1961-1990) ClimateWNA (Wang et
al. 2012) climate data interpolated for the CDFmm and CWHxm ecosystem units. The values
are within the lower range of the climatic envelope definitions, but this is expected because

190

the field data represent the forest microclimate where climate is buffered by the forest
canopy. These field sites will be monitored to evaluate the climatic envelope model results
which suggest that the CDF study sites will remain in the CDFmm SCS until the end of the
2040-2069 timeslice, and the CWH study sites will remain in the CWHxm SCS until the end
of the 2010-2039 timeslice. The field data will also be compared to other projections of
warming temperatures and shifts in precipitation regime (Manabe and Wetherald 1986,
U.S. Committee of Global Change 1988, Christensen et al. 2007). Long-term data series such
as those being generated at these monitoring study sites will also be used to fully
characterize microclimate trends and to detail the relationships between macroclimate,
terrain, and soils across sites, and to thereby more fully comprehend growing conditions in
the dry coastal forest ecosystem. Furthermore, it is hoped that a number of other ecological
research and monitoring projects will take advantage of the detailed microclimate data
being collected, in order to track changes in plant, animal, fungal, and microbial
communities at these study sites.

5.3

APPLICATIONS

5.3.1

Management Applications

The uncertainty of future climate conditions challenges current regimes of ecosystem
management. Projections of future distributions of forest ecosystems are important
considerations for the implementation of sustainable resource management planning and
climate change adaptation planning, so long as the limitations and uncertainty embedded in
the modelling techniques are recognized. Adaptive management accepts uncertainty and

191

risk, makes use of innovative strategies and methods, and learns by comparison between
the results of management and predictions (Kimmins 1987). Adaptive management is
important to all aspects of resource and land management including conservation planning,
resource extraction, fire prevention, watershed planning, invasive species control, wildlife
management, agriculture planning, and urban development (Lindenmayer and Franklin
2002 ).

Climatic envelope models developed and applied in this thesis (Chapters 2 and 3) can
provide guidance to management plans by providing information on likely areas of
ecosystem and species persistence and information on areas likely to undergo rapid
changes. The microclimate field monitoring network and preliminary data (Chapter 4) are
also important for refining tools such as the provincial Biogeoclimatic Ecosystem
Classification system, used to guide resource planning, model validation, and trials for
innovative management techniques under rapidly changing conditions.
An understanding of the distribution of species is a significant component of
conservation planning (Franklin 1995, Austin 1998, 2002, Guisan and Zimmermann 2000).
Climate envelope models facilitate conservation planning by providing direction in
identifying areas of least change that may be used as migratory corridors and refugia for
species that are otherwise experiencing shifts in habitat conditions (Rose and Burton 2009).
Conversely, the same approach is valuable in identifying areas where indicators of climate
change impacts are mostly likely to be expressed, guiding the establishment of monitoring
networks (such as the one described in Chapter 4).

192

The CDFmm is an important source of timber that is managed on both private and Crown
land in British Columbia. Douglas-fir (Pseudotsuga menziesii var. menziesii) is the dominant
tree species in this subzone, and is one of the most valued tree species in the study area
and across BC. Therefore, projections regarding climate suitability for its persistence and
future range are directly applicable to forest management, particularly for foresters, seed
collectors, seed orchards, and seedling nurseries (Wang et al. in press). Fire management
planning is also an important aspect of forest management and municipal services in the dry
ecosystems of the BC coast, especially with many ex-urban (rural residential) residences
being established in wooded areas around population centers and throughout the Gulf
Islands. These are multiple land use areas where watershed planning and management to
safeguard water resources is important to sustain the health and vigour of ecological, social,
and economic communities. Potential shifts in fire, wind, and drought conditions as well as
relative competitive abilities associated with climate change may also increase the
immigration of species not currently found within the CDFmm, including invasive species.
Invasive species (such as Scotch broom (Cytisus scoparius), already an exotic invader in
much of the CDFmm) are opportunistic species that will impede the development of
ecosystems after disturbance (Thuiller et al. 2006). Individual shifts in species distributions
will potentially result in an ineffective BEC system and subsequent impacts to resource
management without a functioning classification system. On the other hand, the narrower
and more well defined projections for ecosystems (e.g., the climate prevailing at zonal sites
of the CDFmm) compared to the overlap area of component species (Chapter 3) may be
indicative of self-organizing abilities among species and environmental processes in such a

193

manner that recognized forest ecosystems will persist or reassemble even under climate
change (Hauessler 2011).
Wildlife management requires an understanding of habitat and how it will change in the
future. Results of climate envelope projections suggest that forest ecosystems may become
drier, resulting in more open canopies and shifts in understory species. Such changes would
impact management planning for migratory and resident birds, ungulates, and have indirect
impacts on aquatic species such as salmon that return to fresh water systems to spawn.
Productive lands at low elevations of the CDFmm and CWHxm have long growing
seasons, ideal for agricultural crops. Projections of impacts of climate change will help direct
farming and household gardening to reduce crop failure. Tools such as the plant hardiness
mapping (www.planthardiness.gc.ca) are being developed to identify locations suitable for
the successful growth and survival of plant species across Canada (McKenney et al. 2007).
Urban and ex-urban (rural residential) areas are growing at increasing rates within the
CDFmm and CWHxm. Changes in climate will impact water sources, food availability, power
requirements, construction, road maintenance, and other infrastructure associated with
urban development planning. The likelihood and seasonal persistence of summer moisture
deficits will be an important consideration in planning for all of these developments.
Climate change research is complex, framed with uncertainty and time lags that
confound the relationships between climate, soil, and vegetation. Climate envelope
modelling provides potential scenarios that can be used for risk-based management, future
planning, and further research into ecosystem science. The changes will not be consistent
across the landscape and there is uncertainty around the time lags, dispersal and

194

establishment limitations and refugia that will dictate species survival and their migration
into new suitable habitat. Caution must be exercised in interpreting any projections of
future ecosystem conditions, their underlying relationships, and their inherent limitations
and uncertainty must be recognized (Heikkinen et al. 2006).

5.3.2

Research Applications

The CDFmm and CWHxm are part of two of the BEC zones projected to be least affected
by climate change provincially (Wang et al. in press). However, the CDFmm and CWHxm are
highly valued ecosystems (economically, socially, and environmentally) and therefore
remain a priority area for research.
Climatic envelopes might bound the sometimes-constraining climate-ecosystem
relationships, but there are other important drivers in the system. The climate envelopes
and field data provided by this project will facilitate more complex process modelling.
Future research into the interactions of plant species and other biological components as
well as the abiotic components of the ecosystem is required to define the impacts of
climate change. Many studies have focussed on tree species, but further research is
required to consider the potential impacts of climate on understory species and the
ecosystem services associated with these forest types. Advancements in modelling suggest
new abilities to include multiple environmental parameters and feedbacks in the system by
employing hierarchical models made possible with new software developments (Araujo and
New 2006). Future research on these processes and employing these modelling approaches
will improve the projection of climate change impacts.

195

Vegetation distribution patterns are determined by moisture, elevation, parent
materials, and continentality gradients (Whittaker 1960,1967). Climatic envelope
projections can be enhanced by extending the same approach to include terrain and soil
features to model species and ecosystem distributions at a regional scale, and to explore
differences between macroclimate projections and site-level ecosystem responses. The use
of non-climate variables (e.g., soil, elevation, slope position) would be advantageous in
contributing to the precision of a species model (Iverson et al. 2009).
It would be desirable to monitor the species that are projected to gain or lose suitable
climate space on Vancouver Island, in order to validate the climate envelope projections
and document the speed and extent of any climate-induced changes in vegetation. Species
such as Abies grandis, Achlys triphylla, Gaultheria shallon, Hylocomium splendens,
Kindbergia oregana, Lonicera ciliosa, and Rhytidiadelphus triquetrus are projected to lose
suitable climate space. Other species, such as Acer macrophyllum, Arbutus menziesii, Thuja
plicata, Cornus nuttallii, Holodiscus discolour, Mahonia nervosa, and Symphoricarpos albus
are projected to gain suitable climate space. The range of suitable climate space for the
three moss species (Hylocomium splendens, Kindbergia oregana, and Rhytidiadelphus
triquetrus) is projected to decrease in the near future (2010-2039 timeslice).
Microclimate data generated in association with this thesis will be available for further
study of climate, soil, and vegetation relationships. In 2010, an integrated climate change
project was initiated to investigate the potential impacts of climate change by evaluating
the relationships of macroclimate and microclimate, nutrient regimes, vegetation
productivity and regeneration across site types in the broad transition zone between the

196

CDFmm and CWHxm subzones (Klassen et al. 2010; Klassen and Saunders 2010, 2011). The
microclimate data will also be used in association with dendrochronology studies exploring
tree ring and climate relationships, investigations into climate and nutrient cycling
processes (Kranabetter et al. in press) and the decomposition of litter and fine woody
material, and studies of plant growth and vigour and vegetation dynamics. The data could
also contribute to wildlife habitat, ecohydrology, and fire regime research. There is also the
potential to expand the microclimate monitoring network and field-based climate change
research initiative to include additional slope positions and BEC subzones on Vancouver
Island.
Ongoing research into the principles underlying BEC and refining it to incorporate the
impacts of climate change are required to maintain their relevance for guiding resource
management (Haeussler 2011). Results presented here indicate that climate will change in
the study area and the classification system will have to evolve to capture the changes in
forest composition and structure associated with such changes. The field data collection on
forest microclimates and soil moisture initiated as part of this thesis (Chapter 4) provide a
foundation for water balance modelling to evaluate impacts of climate change across the
soil moisture gradient between the CDFmm and CWHxm subzones.

5.4

CONCLUSIONS

Climatic envelope models suggest northward shifts for the extended CDF and CWHxm
ecosystems and for the plant species diagnostic of the CDFmm. The climatic envelopes also
generally project a reduction in the geographic extent of ecosystem SCSs over time,

197

suggesting that the climate may not be suitable for the CDFmm on southern Vancouver
Island at the end of the century. A long-term monitoring network of field research sites has
been established to quantify climate, soil, and vegetation relationships, to validate climatic
envelope models, and to monitor change overtime. Though the CDFmm is projected to
experience less severe impacts of climate change when compared to other forest systems in
BC, the introduction of unprecedented (novel) climate could have implications for
ecosystems and species habitats in the study area.
Overall, it is concluded that species climate envelopes are useful to learn what species
could occur in a local area, to better understand which climate variables constrain the
distribution of species, and to predict which species will experience stress or might be
introduced in the future (Iverson et al. 2009). The network of field sites established in
conjunction with this thesis provides a foundation of instrumented study sites on which to
build a climate change research programme. This research is important for management
applications and for improving the understanding of climate, soil, and vegetation
relationships. Collectively, results from this thesis suggest it will be worthwhile to monitor a
number of ecosystem changes in the near future. Specifically, it is recommended that
monitoring be undertaken in the CDFmm-CWHxm transition zone to track climate and
microclimate conditions and the vigour, growth, and regeneration of selected species. In
particular, it would be worthwhile to monitor vigour, mortality, growth, and regeneration in
Abies grandis, Achlys triphylla, Gaultheria shallon, Hylocomium splendens, Kindbergia
oregano, Lonicera ciliosa, and Rhytidiadelphus triquetrus, as these species are projected to
lose SCS in the CDFmm as currently mapped. Further research is needed into the utility of

198

soil and terrain features to modify and refine climatic envelope models and to better inform
complex species and ecosystem distribution models.

199

6.

References

Araujo, M.B. and M. New. 2006. Ensemble forecasting of species distributions. Trends in
Ecology and Evolution 22:42-47.
Austin, M.P. 1987. Models for the analysis of species' response to environmental gradients.
Vegetatio 69: 35-45.
Austin, M.P. 1998. An ecological perspective on biodiversity investigations: examples from
Australian eucalypt forests. Annual Monographs Botanical Gardens 85: 2-17.
Austin, M.P., A.O. Nicholls, and C.R. Margules. 1990. Measurement of the realized
qualitative niche: environmental niches of five Eucalyptus species. Ecological
Monographs 60:161-177.
Austin, M.P. 2002. Spatial prediction of species distribution: an interface between ecological
theory and statistical modelling. Ecological Modelling 157:101-118.
Bailey, R.G. 1997. Ecoregions of North America. U.S. Department of Agriculture, Forest
Service, Washington, DC.
Bates, B.C., A.W. Kundzewicz, S. Wu, and J.P. Palutikof. (Editors). 2008. Climate change and
water, Technical Paper for the Intergovernmental Panel of Climate Change. IPCC
Secretariat, Geneva.
Berry, P.M., T.P. Dawson, P.A. Harrison, and R.G. Pearson. 2002. Modelling potential
impacts of climate change on the bioclimatic envelope of species in Britain and
Ireland. Global Ecology and Biogeography 11:453-462.
Boisvenue, C. and S.W. Running. 2006. Impacts of climate change on natural forest
productivity: Evidence since the middle of the 20th century. Global Change Biology 12:
862-882.
Bonhomme, R. and P. Chartier. 1972. The interpretation and automatic measurement of
hemispherical photographs to obtain sunlit foliage area and gap frequency. Israel
Journal of Agricultural Research 22: 53-61.
Box, E .0.1995. Factors determining distributions of tree species and plant functional types.
Vegetatio 121:101-116.
Brady, N.C. and R.R. Weil. 2008. The nature and properties of soils. [14th edition], Pearson
Education Inc., Upper Saddle River, New Jersey, pp. 173-217.
200

British Columbia Ecological Reports Catalogue (BC EcoCat). 2010. Ministry of Environment.
Available online [October 12, 2009] http://www.env.gov.bc.ca/ecocat/index.htm
British Columbia Land and Resource Data Warehouse (BC LRDW). 2011. Data compiled by
Dan Sirk, Ministry of Forests, Lands, and Natural Resource Operations. [April 27,
2011].
BC Ministry of Environment, Lands, and Parks and BC Ministry of Forests (BC MOELP and BC
MOF). 1998. Field Manual for Describing Terrestrial Ecosystems. Land Management
Handbook No. 25. Victoria, BC.
British Columbia Ministry of Forests (BC MOF). 1995. Forest Practices Code of British
Columbia: Biodiversity Guidebook. Queens Printer. Victoria, BC.
British Columbia Ministry of Forests and Range (BC MFR). 2007. Vegetation and
Environment NexUS Professional (VPro). Victoria, British Columbia, Canada. Available
online [September 30, 2009]:
http://www.for.gov.bc.ca/hre/becweb/resources/software/index.html
British Columbia Ministry of Forests and Range (BC MFR). 2008. BC MFR BEC WEB GIS data
ftp site. Biogeodimatic ecosystem classification coverage version 7. Victoria, British
Columbia, Canada. Available online [September 30, 2009]:
http://www.for.gov.bc.ca/hre/becweb/resources/maps/map_download.html
British Columbia Ministry of Forests and Range (BC MFR). 2009. Biogeodimatic Ecosystem
Classification database (BECdb). Research Branch, Victoria, Canada. [Accessed
November 15, 2009]
Brown, K.J. and R.J. Hebda. 2002. Origin, development, and dynamics of coastal temperate
conifer rainforests of southern Vancouver Island, Canada. Canadian Journal of Forest
Research 32: 353-372.
Bunnell, F.L. 1989. Alchemy or Uncertainty: What Good are Models? General Technical
Report GTR-PNW-232. USDA Forest Service, Portland, Oregon, pp. 27.
Burger, D. 1972. Forest site classification in Canada. Mitt. Ver. Forstl. Standortskunde
Forstpflanzenzuchtung 21:4-36.
Burton, P.J. and S.G. Cumming. 1995. Potential effects of climatic change on some western
Canadian forests, based on phonological enhancements to a patch model of forest
succession. Water, Air, and Soil Pollution 82:401-414.
201

Campbell, E. and T. Wang. 2011. Projected distribution of biogeodimatic (BGC) zones by
2050 for selected forest management regions of British Columbia. In Press.
Campbell, G. 2008. Modelling Available Soil Moisture. Decagon Devices Application Note
13401-01. Available online [January 13, 2010]:
http://www.decagon.com/education/modeling-available-soil-moisture-13401-01-an/
Charney, J. 1975. The dynamics of deserts and droughts. Quarterly Journal of the Royal
Meteorological Society 101:193-202.
Chen, l-C, J.K. Hill, R. Ohlemuller, D.B. Roy, and C.D. Thomas. 2011. Rapid range shifts of
species associated with high levels of climate warming. Science 333:1024-1026.
Chmura, DJ., P.D. Anderson, G.T. Howe, C.A. Harrington, J.E. Halofsky, D.L. Peterson, D.C.
Shaw, and J.B. St.Clair. 2011. Forest responses to climate change in northwestern
United States: Ecophysiological foundations for adaptive management. Forest Ecology
and Management 261:1121-1142.
Christensen, J.H., B. Hewitson, A. Busuioc, A. Chen, X. Gao, I. Held, R. Jones, R.K. Kolli, W.-T.
Kwon, R. Laprise, V. Magana Rueda, L. Mearns, C.G. Menendez, J. Raisanen, A. Rinke,
A. Sarrand P. Whetton. 2007. Regional Climate Projections. In: Climate Change 2007:
The Physical Science Basis. Contribution o f Working Group I to the Fourth Assessment
Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M.
Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (Editors)).
Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.
Clements, F.E. 1907. Plant physiology and ecology. Henry Holt & Co. New York, NY.
Clements, F.E. 1916. Plant succession: an analysis of the development of vegetation.
Carnegie Institute of Washington. Washington, DC.
Clements, F.E. 1936. Nature and structure of the climax. The Journal of Ecology 24: 252-284.
Conservation Data Center (CDC). 2009. BC Species and Ecosystems Explorer. Available
online [September 25, 2009]: http://alOO.gov.bc.ca/pub/eswp/
Crawley, M.J. 2007. The R Book. John Wiley & Sons, Ltd. West Sussex, England, pp. 701-704.
Crins, W.J., P.A. Gray, and P.W.C. Uhlig. 2007. The Ecosystems of Ontario, Part 2:
Ecodistricts. Ontario Ministry of Natural Resources, Peterborough.

202

Cumming, S.G. and P.J. Burton. 1996. Phenology-mediated effects of climatic change on
some simulated British Columbia forests. Climatic Change 34: 213-222.
Curtis, J.T. 1955. A prairie continuum in Wisconsin. Ecology 36: 558-566.
Daly, C., W.P. Gibson, G.H. Taylor, G.L. Johnson, and P. Pasteris. 2002. A knowledge-based
approach to the statistical mapping of climate. Climate Research 22: 99-113.
Daubenmire, R.F. 1976. The use of vegetation in assessing the productivity of forest lands.
Botanical Review 42:115-143.
Davis, M.B. 1986. Climatic instability, time lags, and community disequilibrium, in
Community Ecology. J.M. Diamond and T.J. Case (Editors). Harper & Row, New York,
pp. 269-284.
Decagon Devices. 2009. Frequently asked questions about the Decagon's soil moisture
probes and accessories. Application Note. Available online [January 23, 2010]:
http://www.decagon.com/assets/Uploads/FrequentlyAskedQuestionsabouttheECH20
Soil2.pdf
Dekker, S.C., M. Rietkerk, and M.F.P. Bierkens. 2007. Coupling microscale vegetation-soil
water and macroscale vegetation-precipitation feedbacks in semiarid ecosystems.
Global Change Biology 13: 671-678.
Delcourt, H.R. and P.A. Delcourt. 1988. Quaternary landscape ecology: relevant scales in
space and time. Landscape Ecology 2: 23-44.
DeLong, S.C., H. Griesbauer, W. Mackenzie, and V. Foord. 2010. Corroboration of
biogeodimatic ecosystem classification climate zonation by spatially modelled climate
data. BC Journal of Ecosystems and Management 10: 49-64.
Detto, M., N. Montaldo, J.D. Albertson, M. Mancini, and G. Katul. 2006. Soil moisture and
vegetation controls on evapotranspiration in a heterogeneous Mediterranean
ecosystem on Sardinia, Italy. Water Resources Research 42. W08419 doi:
10.1029/2005WR004693
Dingman, S.L. 2002. Physical Hydrology [2nd edition]. Waveland Press, Inc. Long Grove, IL.
pp. 576-577.

203

Drew, C.A., Y.F. Wiersma, and F. Huettmann. 2011. Predictive species and habitat modeling
in landscape ecology: concepts and applications. Springer Science and Business Media.
New York, NY, USA. pp. 310.
EcoShare: Interagency Clearinghouse of Ecological Information. 2011. USFS. Available online
[February 9, 2010]: http://ecoshare.info/
Ehrenfeld, J.G., B. Ravit, and K. Elgersma. 2005. Feedback in the plant-soil system. Annual
Review of Environmental Resources 30: 75-115.
El Maayar, M., D.T. Price, T.A. Black, E.R. Humphreys, and E.-M. Jork. 2002. Sensitivity tests
of the integrated biosphere simulator to soil and vegetation characteristics in a Pacific
Coastal Coniferous Forest. Atmosphere-Ocean 40: 313-332.
Eng, M. and D. Meidinger. 1999. A method for large-scale biogeodimatic mapping in British
Columbia. Research Branch, BC Ministry of Forests and Range.
Entin, J.K., A. Robock, K.Y. Vinnikov, S.E. Hollinger, S. Liu, and A. Namkhai. 2000. Temporal
and spatial scales of observed soil moisture variations in the extratropics. Journal of
Geophysical Research 105:11865-11877.
Environmental Systems Research Institute (ESRI). 2009. ArcGIS Desktop: ESRI® ArcMap™
9.3.1. Redlands, CA, U.S.A.
Fernow, B.E. 1912. Forest resources and problems of Canada. Proceedings of the Society of
American Foresters 7:133-144.
Forman, R.T.T. 1964. Growth under controlled conditions to explain the hierarchical
distribution of a moss, Tetraphis pellucida. Ecological Monographs 34:1-25.
Franklin, J. 1995. Predictive vegetation mapping: geographic modelling of biospatial
patterns in relation to environmental gradients. Progress in Physical Geography 19:
474-499.
Franklin, J. 2009. Mapping species distributions: spatial inference and prediction. Cambridge
University Press, New York. pp. 5.
Franklin, J.F. and C.T. Dyrness. 1988. Natural Vegetation of Oregon and Washington. Oregon
State University Press. Corvallis, OR. pp. 452.
Gleason, H.A. 1917. The structure and development of the plant association. Bulletin of the
Torrey Botanical Club 43:463-481.
204

Gleason, H.A. 1926. The individualistic concept of the plant association. Bulletin of the
Torrey Botanical Club 53: 7-26.
Gleckler, P.J., K.E. Taylor, and C. Doutriaux. 2008. Performance metrics for climate models.
Journal of Geophysical Research 113. D06104 doi: 10.1029/2007JD008972.
Gotelli, NJ. and A.M. Ellison. 2004. A primer of ecological statistics. Sinauer Associates Inc.
Sunderland, MA, USA. pp. 510.
Government of Ontario Ministry of Natural Resources (GO MNR). 2007. Ecological Land
Classification Primer. Ontario Ministry of Natural Resources, Peterborough. Available
online [April 5, 2011]:
http://www.mnr.gov.on.ca/en/Business/LUEPS/Publication/264779.html
Green, R.N. and K. Klinka. 1994. A field guide for site identification and interpretation for
the Vancouver forest region. Ministry of Forests Land Management Handbook No.28
ISSN 0229-1622.
Green, R.N., R.L. Trowbridge, and K. Klinka. 1993. Towards a taxonomic classification of
humus forms. (Supplement to Forest Science 39) Forest Science Monograph 29 ISSN
0015-749X.
Guijarro, J.A. 2011. Climatol: Some tools for climatology: series homogenization, plus
windrose and Walter & Lieth diagrams. R: package version 2.1. Available online
[January 10, 2011]: http://CRAN.R-project.org/packagesclimatol.
Guisan, A. and N.E. Zimmermann. 2000. Predictive habitat distribution models in ecology.
Ecological Modelling 135:147-186.
Guisan, A. and W. Thuiller. 2005. Predicting species distributions: offering more than simple
habitat models. Ecology Letters 8: 993-1009.
Guswa, A.J. 2002. Models of soil moisture dynamics in Ecohydrology: a comparative study.
Water Resources Research 38:1-15.
Haeussler, S., D. Coates, and J. Mather. 1990. Autecology of common plants in British
Columbia: A literature review. FRDA Report 158 ISSN 0835 0752.
Haeussler, S. 2011. Rethinking biogeodimatic ecosystem classification for a changing world.
Environmental Reviews 19: 254-277.

205

Hamann, A., and T.L. Wang. 2005. Models of climatic normals for genecology and climate
change studies in British Columbia. Agricultural and Forest Meteorology 128: 211-221.
Hamann, A., and T.L. Wang. 2006. Potential effects of climate change on ecosystem and tree
species distribution in British Columbia. Ecology 87: 2773-2786.
Heikkinen, R.K., M. Luoto, M.B. Araujo, R. Virkkala, W. Thuiller, and M.T. Sykes. 2006.
Methods and uncertainties in bioclimatic envelope modelling under climate change.
Progress in Physical Geography 30: 751-777.
Hijmans, R.J. and C.H. Graham. 2006. The ability of climate envelope models to predict the
effect of climate on species distributions. Global Change Biology 12: 2272-2281.
Hirzel, A. and A. Guisan. 2002. Which is the optimal sampling strategy for habitat suitability
modelling. Ecological Modelling 157: 331-341.
Hutchinson, G.E. 1957. Concluding remarks. Cold Spring Habor Symposium in Quantative
Biology 22: 415-427.
Intergovernmental Panel on Climate Change (IPCC). 2007. Climate Change 2007: Synthesis
Report. Contribution of Working Groups I, II and III to the Fourth Assessment Report
of the Intergovernmental Panel on Climate Change [Core Writing Team, Pachauri, R.K
and Reisinger, A. (Editors)]. IPCC, Geneva, Switzerland, pp. 104.
Iverson, L.R. and A.M. Prasad. 2001. Potential changes in tree species richness and forest
community types following climate change. Ecosystems 4:186:199.
Iverson, L.R., A.M. Prasad, and S.N. Matthews. 2008. Modelling potential climate change
impacts on the trees of northeastern United States. Mitigation and Adaptation
Strategies for Global Change 13: 487-516.
Iverson, L.R., A. Prasad, S. Mattherws, M.Pteres, and C. Hoover. 2009. Potential changes in
habitat suitability under climate change: Lessons learned from 15 years of species
modelling. XIII World Forestry Congress. Buenos Aries, Argentina, pp. 1-10.
Johns, T. C., R. E. Carnell, J. F. Crossley, J. M. Gregory, J. F. B. Mitchell, C. A. Senior, S. F. B.
Tett, and R. A. Wood. 1997. The second Hadley Center coupled ocean atmosphere
GCM: model description, spinup and validation. Climate Dynamics 13:103-134.

206

Jones, C. and C. Brown. 2008. Future ecosystem climate change mapping and statistical
modelling of the Campbell River Forest District. Prepared for Campbell River Forest
District, Ministry of Forests and Range (unpublished report).
Kimmins, J.P. 1987. Forest Ecology: A foundation for sustainable forest management and
environmental ethics in forestry. [3rd Edition]. Prentice Hall, New Jersey, pp. 611.
Klassen, H.A., J.M. Kranabetter, J.A. MacKinnon, and S.C. Saunders. 2010. Integrated
Coastal Douglas-fir - Coastal Western Hemlock Climate Change Project. Ministry of
Forests, Lands, and Natural Resource Operations (MFLNR) EP 1388 Project Work Plan.
Klassen, H.A. and S.C. Saunders. 2010. Coastal Forest Ecosystems Case Study: Soil moisture
as an indicator of shifts in ecosystem processes and species composition with climate
change. Ministry of Forests and Range Future Forests Ecosystem Initiative Year-end
Report.
Klassen, H.A. and S.C. Saunders. 2011. Coastal Forest Ecosystems Case Study: Soil moisture
as an indicator of shifts in ecosystem processes and species composition with climate
change. Ministry of Forests and Range Future Forests Ecosystem Initiative Year-end
Report.
Klinka, K., V.J. Krajina, A. Ceska, and A.M. Scagel. 1989. Indicator plants of Coastal British
Columbia. Ministry of Forests Research Program. University of British Columbia Press.
Vancouver, BC. pp. 288.
Knutti, R. 2008. Should we believe model predictions of future climate change?
Philosophical Transactions of the Royal Society 366: 4647-4664.
Koppen, W. 1923. Die Klimate der Erde: Grundriss der Klimakunde. DeGruyter, Berlin, pp.
369.
Kranabetter, J.M., S. Saunders, J.A. MacKinnon, H. Klassen, and D.L. Spittlehouse. 2012. An
assessment of contemporary and historic nitrogen availability in contrasting coastal
Douglas-fir forests through 615N of tree rings. Submitted to Global Change Biology.
Lambers, H., F.S. Chapin, and T.L. Pons. 2008. Plant Physiological Ecology [2nd Edition).
Springer, New York. pp. 604.
Landsberg, J. 2003. Modelling forest ecosystems: state of the art, challenges, and future
conditions. Canadian Journal of Forest Research 33: 385-397.

207

Lemon, E.R., A.H. Glaser, and L.E. Satterwhite. 1957. Some aspects of the relationship of
soil, plant, and meteorological factors to evapotranspiration. Soil Science Society
Proceedings 21: 464-468.
Lemon, J. 2006. Plotrix: a package in the red light district in R. R-News 6:8-12.
Lindenmayer, D.B. and J.F. Franklin. 2002. Conserving Forest Biodiversity: a comprehensive
multiscaled approach. Island Press, Washington, USA. pp. 350.
Lo, Y.-H., J.A. Blanco, and J.P. Kimmins. 2010. A word of caution when planning forest
management using projections of tree species range shifts. The Forestry Chronicle 86:
312-316.
Lortie, C.J., R.W. Brooker, P. Choler, Z. Kikvidze, R. Michalet, R.l. Pugnaire, and R.M.
Calaway. 2004. Rethinking plant community theory. Oikos 107: 433-438.
MacKinnon, A., D. Meidinger, and K. Klinka. 1992. Use of the biogeodimatic ecosystem
classification system in British Columbia. The Forestry Chronicle 68:100-121.
Manabe, S. and R.T. Wetherald. 1986. Reduction in summer soil wetness induced by an
increase in atmospheric carbon dioxide. Science 232: 626-628.
Maxim Integrated Products. 2011. Maxim Thermochron iButton: General description.
Report 19-5101 Revision 5. Available online [January 15, 2010]:
http://datasheets.maxim-ic.com/en/ds/DS1921G.pdf
McCune, B. and M. J. Mefford. 2006. PC-ORD. Multivariate Analysis of Ecological Data.
Version 5.10 MjM Software, Gleneden Beach, Oregon, U.S.A.
McIntosh, R.P. 1967. The continuum concept of vegetation. Botanical Review 33:130-187.
McKenney, D.W., J.H. Pedlar, K. Lawrence, K. Campbell, and M.F. Hutchinson. 2007.
Potential impacts of climate change on the distribution of North American trees.
BioScience 57: 939-948.
Meidinger, D.V., and J. Pojar. 1991. Ecosystems of British Columbia. Research Branch, British
Columbia Ministry of Forests, Victoria, British Columbia, Canada.
Montaldo, N., J.D. Albertson, and M. Mancini. 2008. Vegetation dynamics and soil water
balance in a water-limited Mediterranean ecosystem on Sardinia, Italy. Hydrology and
Earth System Science 12:1257-1271.

208

Morin, X. and W. Thuiller. 2009. Comparing niche- and process-based models to reduce
prediction uncertainty in species range shifts under climate change. Ecology 90:13011313.
Mueller-Dombois, D. and H. Ellenberg 1974. Aims and Methods of Vegetation Ecology.
Wiley, New York. pp. 547.
Nuszdorfer, F.C., K. Klinka, and D.A. Demarchi. 1991. Coastal Douglas-fir Zone in Ecosystems
of British Columbia. British Columbia Ministry of Forests Special Report Series 6. pp.
81-95.
Odum, E.P. 1969. The strategy of ecosystem development. Science, New Series 164: 262270.
Odum, E.P. 1971. Fundamentals of Ecology. [3rd Edition]. W.B. Saunders Co., Toronto, ON.
pp. 574.
Oke, T.R. 1987. Boundary Layer Climates. [2nd Edition]. Methuen Inc., New York, NY. pp. 170,110-190.
Onset. 2009. Silicon Pyranometer Smart Sensor. Onset Computer Corporation Manual, pp.
6.

Pacific Climate Impacts Consortium (PCIC). 2011. Regional Analysis Tool. Available online
[October 12, 2009]: http://pacificclimate.org/tools-and-data/regional-analysis-tool
Pearson, R.G., T.P. Dawson, P.M. Berry, and P.A. Harrison. 2002. SPECIES: A spatial
evaluation of climate impact on the envelope of species. Ecological modelling 154:
289-300.
Pearson, R.G. and T.P. Dawson. 2003. Predicting the impacts of climate change on the
distribution of species: are bioclimate envelope models useful? Global Ecology and
Biogeography 12: 361-371.
Pearson, R.G., W. Thuiller, M.B. Araujo, E. Marinez-Meyer, L. Brotons, C. McClean, L. Miles,
P. Segurado, T.P. Dawson, and D.C. Lees. 2006. Model-based uncertainty in species
range prediction. Journal of Biogeography 33:1704-1711.
Pico Technology. 2011. Thermocouple Application Note. Available online [March 25, 2010]:
http://www.picotech.com/applications/thermocouple.html

209

Pierce, D.W., T.P. Barnett, B.D. Santer, and PJ. Gleckler. 2009. Selecting global climate
models for regional climate change studies. Proceedings of the National Academy of
Sciences 106: 8441- 8446.
Pike, J.G. 1964. The estimation of annual runoff from meteorological data in a tropical
climate. Journal of Hydrology 2:116-123.
Pojar, J., K. Klinka, and D.V. Meidinger. 1987. Biogeoclimatic Ecosystem Classification in
British Columbia. Forest Ecology and Management 22:119-154.
Pojar, J. and A. MacKinnon. 1994. Plants of Coastal British Columbia including Washington,
Oregon, & Alaska. Lone Pine Publishing, British Columbia, pp. 528.
R Core Development Team. 2010. R: A language and environment for statistical computing.
R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0. Available
online [January 13, 2010]: http://www.R-project.org/
Radic, V. and G.K.C. Clarke. 2011. Evaluation of IPCC models performance in simulating late
20th century climatologies and weather patterns over North America. Journal of
Climate, doi: 10.1175/JCLI-D-11-00011.
Randall, D.A., R.A. Wood, S. Bony, R. Colman, T. Fichefet, J. Fyfe, V. Kattsov, A. Pitman, J.
Shukla, J. Srinivasan, RJ. Stouffer, A. Sumi, and K.E. Taylor. 2007. Climate Models and
Their Evaluation. In: Climate Change 2007: The Physical Science Basis. Contribution o f
Working Group I to the Fourth Assessment Report o f the Intergovernmental Panel on
Climate Change. Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt,
M.Tignor and H.L. Miller (Editors). Cambridge University Press, Cambridge, United
Kingdom and New York, NY, USA.
Regent Instruments Inc. 2006. WinSCANOPY: canopy analysis with fish-eye and conventional
imaging.
Rehfeldt, G.E. 2004. Interspecific and intraspecific variation in Picea engelmannii and its
congeneric cohorts: biosystematics, genecology, and dimate-change. USDA Forest
Service Rocky Mountain Research Station General Technical Report 134. Fort Collins,
CO.
Rehfeldt, G.E., N.L. Crookston, M.V. Warwell, and J.S. Evans. 2006. Empirical analyses of
plant-climate relationships for the western United States. International Journal of
Plant Science 167:1123-1150.

210

Resources Inventory Committee (RIC). 1998. Standard for Terrestrial Ecosystem Mapping in
British Columbia. ISBN 0-7726-3552-8.
Rodriguez-lturbe, I. 2000. Ecohydrology: A hydrologic perspective of climate-soil-vegetation
dynamics. Water Resources Research 36: 3-9.
Rodriguez-lturbe, I. and A. Porporato. 2004. Ecohydrology of Water-Controlled Ecosystems:
soil moisture and plant dynamics. Cambridge University Press, New York. pp. 438.
Rollenbeck, R. and D. Anhuf. 2007. Characteristics of the water and energy balance in an
Amazonian lowland rainforest in Venezuela and the impact of the ENSO-cycle. Journal
of Hydrology 337: 377-390.
Rose, N.A. and P.J. Burton. 2009. Using bioclimatic envelopes to identify temporal corridors
in support of conservation planning in a changing climate. Forest Ecology and
Management 258S: S64-S74.
Rose, N.A. and P.J. Burton. 2011. Persistent climate corridors: the identification of refugia in
British Columbia's Central interior for the selection of candidate areas for
conservation. BC Journal of Ecosystems and Management 12:101-117.
Rouse, W.R. and R.G. Wilson. 1972. A test of the potential accuracy of the water-budget
approach to estimating evapotranspiration. Agricultural Meteorology 9:421-446.
Schofield, W.B. 1992. Some common mosses of British Columbia. Royal British Columbia
Museum handbook ISSN 1188-5114. pp. 394.
Shimwell, D.W. 1971. The description and classification of vegetation. University of
Washington Press, Seattle, pp. 322.
Shipley, B. and P.A. Keddy. 1987. The individualistic and community-unit concepts as
falsifiable hypotheses. Vegetatio 69:47-55.
Skov, F. and J.-C. Svenning. 2004. Potential impact of climate change on the distribution of
forest herbs in Europe. Ecography 27: 366-380.
Soetaert, K. and P.M.J. Herman. 2009. A practical guide to ecological modelling: using R as a
simulation platform. Springer Science and Business Media. New York, NY, USA. pp.
365.

211

Spittlehouse, D.L. 1989. Estimating evapotranspiration from land surfaces in British
Columbia, in Estimation of Areal Evapotranspiration, International Association of
Hydrological Sciences Publication 177:245-253.
Spittlehouse, D. 2003. Water availability, climate change and growth of Douglas-fir in the
Georgia Basin. Canadian Water Resources Journal 28: 673-688.
Teuling, A.J., R. Uijlenhoet, F. Hupet, E.E. vanloon, and P.A. Troch. 2006. Estimating spatial
mean root-zone soil moisture from point-scale observations. Hydrology and Earth
System Sciences 10: 755-767.
Thuiller, W. 2003. BIOMOD-Optomizing predictions and projecting potential future shifts
under global change. Global Change Biology 9:1353-1362.
Thuiller, W., D.M. Richardson, M. Rouget, S. Proches, and J.R.U. Wilson. 2006. Interactions
between environment, species traits, and human uses describe patterns of plant
invasions. Ecology 87:1755-1769.
U.S. Committee of Global Change. 1988. Toward an understanding of global change.
National Academy Press, Washington, D.C. pp. 213.
U.S. Department of Agriculture (USDA). 1965. Silvics of forest trees of the United States.
Compiled and revised by H.A. Fowells. Agriculture Handbook No. 271.
Varallyay, G.Y. 1990. Influence of climatic change on soil moisture regime, texture,
structure, and erosion. Chapter 4 in Soil on a Warmer Earth: effects of expected
climate change on soil processes, with emphasis on the tropics and sub-tropics.
Elsevier Science Publishers, New York, N.Y. pp. 39-51.
Ventura, F., 0. Fancini, S. Piana, and P. Rossi Pisa. 2010. Soil Moisture Measurements:
Comparison of instrumentation performances. Journal of Irrigation and Drainage
Engineering 136: 81-89.
Wang, T., A. Hamann, D.L. Spittlehouse, and S.N. Aitken. 2006. Development of scale-free
climate data for western Canada for use in resource management. International
Journal of Climatology 26: 383-397.
Wang, T., A. Hamann, and D.L. Spittlehouse. 2010. ClimateWNA v4.52. A program to
generate climate normal, annual, seasonal and monthly data for genecology and
climate change studies in Western North America (WNA) region. Available online

212

[May 12, 2010]:
http://www.genetics.forestry.ubc.ca/cfcg/ClimateWNA/ClimateWNA.html
Wang, T., E.M. Campbell, G.A. O'Neill, and S.N. Aitken. 2012. Projecting future ecosystem
distributions: uncertainties and management applications. In Press.
Wang, T., A. Hamann, D.L. Splittlehouse, and T.Q. Murdock. 2012. ClimateWNA- Highresolution spatial climate data for Western North America. Journal of Applied
Meteorology and Climatology, doi 10.1175/JAMC-D-11-043.1
Waring, R.H. and S.W. Running. 2007. Forest Ecosystems: analysis at multiple scales. [3rd
Edition]. Elsevier Academic Press. San Diego, CA. pp. 420.
Webb, T. 1987. The appearance and disappearance of major vegetational assemblages:
long-term vegetational dynamics in eastern North America. Vegetatio 69:177-187.
Wertz, W.A. and J.F. Arnold. 1975. Land stratification for land-use planning, in B.Berniew
and C.H. Winget (Editors). Forest Soils and Forest Land Management. Laval University
Press, Quebec, pp. 617-629.
Whittaker, R.H. 1951. A criticism of the plant association and climatic climax concepts.
Northwest Science 25:17-31.
Whittaker, R.H. 1960. Vegetation of the Siskiyou Mountains, Oregon and California.
Ecological Monographs 30: 279-338.
Whittaker, R.H. 1967. Gradient analysis of vegetation. Biological Review 42: 207-264.
Whittaker, R.H. 1973. Ordination and classification of communities. Part 5, R.Tuxen (Editor).
Handbook of Vegetation Science. Dr. W. Junk, b.v., Publishers, The Hague.
Wildi, O. 2010. Data Analysis in Vegetation Ecology. John Wiley & Sons, Ltd. Hoboken, NJ.
pp. 21-30.
Williams, P. L. Hannah, S. Andelman, G. Midgely, M. Araujo, G. Hughes, L. Manne, E.
Martinez-Meyer, and R. Pearson. 2005. Planning for climate change: identifying
minimum-dispersal corridors for the Cape Proteaceae. Conservation Biology 19:10631074.
Willoughby, M.G., M.J. Alexander, and B.W. Adams. 2005. Range plant community types
and carrying capacity for the montane subregion. Alberta Sustainable Resource
Development, Public Lands Division. Publication no. T/071.
213

Wong, C., B. Dorner, and H. Sandmann. 2003. Estimating historical variability of natural
disturbances in B.C. Forest Science Program, B.C. Ministry of Forests and Ecosystem
Conservation, B.C. Ministry of Sustainable Resource Management, Victoria, B.C. Land
Management Handbook No. 53.
Woodward, F.l. 1987. Climate and plant distribution. Cambridge University Press,
Cambridge, pp. 174.
World Meteorological Organization (WMO). 1989. Calculations of monthly and annual 30year standard normals, WCDP No.10, WMOTD No.341. World Meteorological
Organization, Geneva.
Zimmermann, N.E. and F. Kienast. 1999. Predictive mapping of alpine grasslands in
Switzerland: Species versus community approach. Journal of Vegetation Science 10:
469-482.

214

Appendix A: Global climate models (GCMs) and carbon emission scenarios available for future
projections using the ClimateWNA interpolation tool (ClimateWNA v4.52 2010).
Carbon
Center and Location
Vintage
IPCC Model ID
Emission
Scenarios
Hadley Center for Climate Prediction and
A2
AR4
UKMO HadCM3
Research/Met Office, UK
Bjerknes Center for Climate Research,
A2
AR4
BCCR BCM2.0
Norway
A1B; A2; B1
AR4
CCCMA CGCM3
Canadian Center for Climate Modelling
and Analysis, Canada

Runs

Runl
Runl
Mean; runl,
run4

Run2
Center for Climate System Research
A2
(University of Tokyo), National Institute for
Environmental Studies, and Frontier
Research Center for Global Change
(JAMSTEC), Japan
Runl
Hadley Center for Climate Prediction and
A1B
AR4
UKMO HadGEMl
Research/Met Office, UK
A1FI; A2; B2 N/A
Commonwealth Scientific and Industrial
TAR
CSIR02
Research Organisation (CSIRO)
Atmospheric Research, Australia
N/A
TAR
ECHAM4
Max Planck Institute for Meteorology,
A2; B2
Germany
Hadley Center for Climate Prediction and
N/A
TAR
A2; Bl; B2
HADCM3
Research/Met Office, UK
B1
Runl
AR4
Center for Climate System Research
MIROC32 HIRES
(University of Tokyo), National Institute for
Environmental Studies, and Frontier
Research Center for Global Change
(JAMSTEC), Japan
TAR
PCM
National Center for Atmospheric
A1FI; A2; Bl; N/A
Research, USA
B2
* Model ID, Vintage, Sponsors, and Country from IPCC AR4 Chapter 8 (Randall et al. 2007)
AR4

MIROC32 MEDRES

215

Appendix B: Annual, seasonal, and monthly climate variables available from ClimateWNA for current
and future projections (ClimateWNA v4.52 2010). Annual climate variables used for climatic
envelope definitions in bold.
Annual Climate Variables
MAT
MWMT
MCMT
TD
MAP
MSP
AHM
SHM
DD<0
DD>5

Mean annual temperature (°C)
Mean temperature of the warmest month (°C)
Mean temperature of the coldest month (°C)
Difference between MWMT and MCMT (°C)
Mean annual precipitation (mm)
Mean May to September precipitation (mm)
Annual heat: moisture index (MAT+10)/(MAP/1000)
Summer (May to Sept) heat: moisture index (MWMT)/(MSP/1000)
Degree-days below 0°C, (chilling degree-days)
Degree-days above 5°C, (growing degree-days)

Degree-days below 18°C, (heating degree-days)
Degree-days above 18°C, (cooling degree-days)
Number of frost-free days
Frost-free period (days)
Julian date on which FFP begins
Julian date on which FFP ends
Precipitation as snow (mm water equivalent)
Extreme minimum daily temperature over 30 years (°C)
Hargreaves reference evaporation (Hargreaves equation with a
latitude correction)
Hargreaves climatic moisture deficit (sum of the monthly difference
CMD
between Eref and precipitation)
Seasonal (Winter, Spring, Summer, Autumn)*
mean temperature (°C) x 4
TAV_wt, TAV_sp,
TAV_sm, TAV_at
TMAX_wt, TMAX_sp,
mean maximum temperature (°C) x 4
TMAX_sm, TMAX_at
mean minimum temperature ('C) x 4
TMIN_wt, TMIN_sp,
TMIN_sm, TMIN_at
PPT_wt, PPT_sp, PPT_sm, precipitation (mm) x 4
PPT at
Monthly (January to December)
mean temperature (°C) x 12
TAV01- TAV12
TMX01 - TMX12
mean maximum temperature (°C) x 12
TMN01 - TMN12
mean minimum temperature (°C) x 12
PPT01 - PPT12
mean precipitation (mm) x 12
^Seasonal: Winter (December-February); Spring (March-May); Summer (June-August);
DD<18
DD>18
NFFD
FFP
bFFP
eFFP
PAS
EMT
Eref

Autumn (September-November)

216

Appendix C: Total area (km2) and climate summary statistics (mean, standard deviation (SD),
minimum, and maximum) for baseline and future extended Coastal Douglas-fir (CDF) suitable
climate space (SCS) in southwestern British Columbia. The climatic envelope definition values
(minimum and maximum) appear in parentheses below the climate variable.
Baseline
Climate Variable
MAT(°C)
(8.2-13)

MWMT('C)
(15.6-21.3)

MCMT(°C)
(1.6-7.7)

TD(°C)
(11.1-18.4)

MAP(mm)
(438-3473)

MSP(mm)
(81-429)

AHM
(6.2-45.9)

SHM
(43.3-248.1)

DD<0
(4-102)

Statistic
Total Area (km2)
Mean
SD
Minimum
Maximum
Mean
SD
Minimum
Maximum
Mean
SD
Minimum
Maximum
Mean
SD
Minimum
Maximum
Mean
SD
Minimum
Maximum
Mean
SD
Minimum
Maximum
Mean
SD
Minimum
Maximum
Mean
SD
Minimum
Maximum
Mean
SD
Minimum

1961-1990
5453
9.4
0.4
8.2
10.3
16.8
0.4
15.6
17.7
2.7
0.6
1.6
5.2
14.2
0.7
11.3
15.8
1274.1
245.9
719.0
2443.0
234.7
64.3
121.0
380.0
15.8
3.4
7.7
27.8
77.1
20.2
44.1
131.4
64.5
17.1
16.0
217

Future Timeslices
2010-2039
7597
10.5
0.5
8.8
11.6
18.5
0.6
16.7
20.6
2.9
0.6
1.6
5.4
15.7
0.8
12.7
18.4
1404.5
317.2
722.8
2888.0
249.4
73.3
113.8
415.4
15.4
3.7
6.8
29.3
81.7
24.9
43.4
156.8
60.6
17.9
13.8

2040-2069
10100
11.2
0.7
9.4
12.5
19.8
0.6
18.0
21.3
3.3
0.8
1.7
5.5
16.5
0.7
13.5
18.4
1542.8
405.6
766.6
3326.0
250.2
67.3
114.6
427.6
14.7
4.0
6.2
29.0
87.7
27.2
44.8
174.9
54.1
21.6
14.4

2070-2099
5558
11.5
0.6
9.8
13.0
20.7
0.5
19.0
21.3
3.4
0.7
1.7
6.2
17.3
0.6
14.2
18.4
2060.9
528.0
1347.0
3464.0
300.5
54.2
182.4
428.8
11.1
2.5
6.2
16.2
72.8
13.7
49.5
119.4
48.6
14.8
10.6

Appendix C (continued).

Baseline
Climate Variable
DD>5
(1574-3005)

NFFD
(256-346)

FFP
(162-303)

PAS(mm)
(11-132)

EMT(°C)
(-21. 7--7.4)

Eref
(571-1001)

CMD
(152-641)

Statistic
Maximum
Mean
SD
Minimum
Maximum
Mean
SD
Minimum
Maximum
Mean
SD
Minimum
Maximum
Mean
SD
Minimum
Maximum
Mean
SD
Minimum
Maximum
Mean
SD
Minimum
Maximum
Mean
SD
Minimum
Maximum

1961-1990
102.0
1872.2
102.1
1599.0
2078.0
292.2
13.2
256.0
333.0
208.2
18.3
162.0
284.0
61.4
23.0
16.0
132.0
-17.6
1.5
-20.9
-10.7
655.8
25.0
571.0
735.0
240.1
58.2
152.0
350.0

218

Future Timeslices
2010-2039
102.0
2215.2
150.0
1810.6
2620.8
305.6
12.4
267.2
340.6
229.0
19.7
176.4
302.4
52.5
22.5
11.8
132.0
-17.1
1.5
-20.8
-10.3
692.4
31.9
597.6
782.6
255.6
67.5
152.0
393.0

2040-2069
102.0
2468.4
204.8
1995.4
2852.4
312.4
14.3
274.6
342.2
242.7
23.7
188.0
302.6
52.9
29.0
11.0
132.0
-16.2
1.7
-20.8
-11.2
717.5
31.6
630.4
825.8
275.2
69.9
152.0
428.8

2070-2099
100.8
2579.4
159.6
2123.6
3004.6
314.2
10.3
282.2
343.6
243.5
17.5
195.6
302.8
67.2
28.9
17.0
132.0
-15.4
1.4
-19.6
-10.8
730.2
25.2
644.8
790.6
242.0
51.2
152.0
365.4

Appendix D: Paired t-test results summary table for the 5-km buffer within and outside the extended
Coastal Oouglas-fir (CDF) baseline suitable climate space (SCS) on Vancouver Island. Climate
variables ranked from most significant (highest t-value) to least significant difference between the
means. Sample number includes 30 paired random samples; latitude (°N) is the mid-point of the
sample window.
Appendix D continued).

Sample
No. and
Latitude

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

Sample No.

SHM

81.1

52.4

28.6

13.58

4.23E-14

1

CMD

254.0

172.6

81.4

11.99

9.19E-13

Latitude

MSP

207.6

314.5

-106.9

11.93

1.04E-12

48.71

AHM

13.2

8.3

4.9

11.83

1.27E-12

MAP

1466.0

2266.8

-800.8

11.39

3.16E-12

DD>5

1754.5

1563.9

190.6

8.20

4.89E-09

PAS

92.2

210.7

-118.4

8.05

7.12E-09

FFP

198.3

180.4

17.9

7.83

1.23E-08

NFFD

283.3

264.1

19.2

7.75

1.50E-08

MWMT

16.6

16.0

0.6

7.73

1.58E-08

MAT

8.9

8.0

0.9

7.52

2.74E-08

EMT

-18.5

-19.8

1.4

5.94

1.87E-06

MCMT

2.2

1.6

0.6

4.33

1.61E-04

DD<0

79.7

102.2

-22.6

4.21

2.26E-04

Eref

645.0

627.1

17.9

3.57

1.26E-03

TD

14.4

14.4

0.1

0.94

3.54E-01

Sample No.

SHM

83.6

53.3

30.3

13.96

2.10E-14

2

AHM

13.7

8.4

5.3

13.05

1.16E-13

Latitude

CMD

261.4

174.7

86.7

12.75

2.03E-13

48.73

MSP

200.0

308.1

-108.0

11.57

2.20E-12

MAP

1400.1

2220.2

-820.1

11.31

3.73E-12

DD>5

1743.6

1552.4

191.1

10.27

3.59E-11

FFP

196.5

179.1

17.4

9.90

8.20E-11

7.9

0.9

9.49

2.13E-10

Sample No.

MAT

8.9

NFFD

281.5

262.3

19.2

9.47

2.21E-10

PAS

90.7

217.0

-126.3

8.44

2.64E-09

MW M T

16.6

16.0

0.6

7.74

1.54E-08

EMT

-18.7

-20.0

1.3

6.70

2.41E-07

MCMT

2.1

1.5

0.6

5.45

7.19E-06

DD<0

82.8

106.6

-23.8

5.12

1.82E-05

Eref

645.2

624.5

20.7

4.80

4.46E-05

TD

14.5

14.5

0.0

0.63

5.35E-01

SHM

82.8

53.1

29.7

15.35

1.86E-15

219

Appendix D (continued)

Sample
No. and
Latitude

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

14.18

1.42E-14

3

CMD

260.3

173.3

87.1

Latitude

AHM

13.6

8.4

5.2

13.30

7.22E-14

48.75

MSP

203.1

308.2

-105.1

13.01

1.24E-13

MAP

1415.4

2208.6

-793.2

12.48

3.48E-13

DD>5

1747.4

1548.5

198.9

10.10

5.24E-11

FFP

195.9

178.3

17.6

10.00

6.61E-11

NFFD

281.4

261.4

20.0

9.56

1.84E-10

MAT

8.9

7.9

1.0

9.25

3.79E-10

PAS

91.6

218.3

-126.7

8.95

7.68E-10

MWMT

16.6

16.0

0.6

8.66

1.56E-09

EMT

-18.7

-20.0

1.3

6.70

2.36E-07

MCMT

2.1

1.5

0.6

5.71

3.57E-06

DD<0

82.1

108.3

-26.1

5.36

9.36E-06

Eref

646.5

623.3

23.2

5.05

2.19E-05

TD

14.5

14.5

0.0

0.18

8.60E-01

Sample No.

SHM

82.8

54.5

28.3

13.53

4.67E-14

4

CMD

262.0

177.1

84.9

12.67

2.40E-13

Latitude

AHM

13.7

8.6

5.1

12.37

4.32E-13

48.77

MSP

203.3

298.1

-94.9

11.44

2.85E-12

MAP

1408.8

2132.2

-723.4

11.26

4.19E-12

DD>5

1750.4

1546.5

203.9

10.42

2.56E-11

16.5

10.04

6.01E-11

FFP

194.8

178.4

MAT

8.9

7.9

1.0

9.93

7.73E-11

NFFD

280.7

261.7

19.0

9.75

1.17E-10

MW M T

16.6

16.0

0.7

9.55

1.84E-10

PAS

91.4

211.1

-119.7

9.33

3.09E-10

EMT

-18.7

-20.0

1.3

6.91

1.34E-07

Eref

649.9

621.9

28.0

6.62

2.95E-07

MCMT

2.1

1.5

0.6

5.89

2.13E-06

DD<0

82.1

109.0

-26.9

5.80

2.74E-06

TD

14.5

14.5

0.0

0.73

4.73E-01

Sample No.

SHM

85.0

52.4

32.6

17.48

< 2.2e-16

5

CMD

267.6

168.8

98.8

16.06

5.68E-16

Latitude

AHM

14.1

8.2

5.9

15.72

1.00E-15

48.79

MSP

197.9

309.7

-111.8

13.77

2.99E-14

MAP

1362.7

2230.6

-867.8

13.52

4.73E-14

PAS

86.7

231.0

-144.4

10.90

9.06E-12

DD>5

1765.0

1528.5

236.5

9.82

9.99E-11

220

Appendix D (continued).

Sample
No. and
Latitude

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

MAT

8.9

7.8

1.1

9.41

2.55E-10

MWMT

16.7

15.9

0.8

9.12

5.14E-10

NFFD

281.3

259.5

21.8

9.03

6.31E-10

FFP

195.5

177.3

18.2

7.98

8.53E-09

Eref

652.1

617.7

34.4

6.95

1.21E-07

EMT

-18.6

-20.1

1.5

6.68

2.53E-07

MCMT

2.1

1.4

0.8

5.76

3.08E-06

DD<0

80.0

112.3

-32.2

5.76

3.09E-06

TD

14.5

14.5

0.0

0.15

8.82E-01

Sample No.

SHM

84.8

52.7

32.0

14.45

8.72E-15

6

CMD

268.4

170.5

97.9

13.42

5.77E-14

Latitude

AHM

14.2

8.4

5.8

13.32

6.91E-14

12.45

3.67E-13

11.80

1.37E-12

48.81

MSP

199.0

308.5

-109.5

MAP

1358.0

2206.1

-848.1

NFFD

280.5

260.1

20.4

10.95

8.10E-12

DD>5

1761.3

1538.0

223.3

10.86

9.84E-12

PAS

86.6

222.8

-136.2

10.85

1.01E-11

MAT

8.9

7.8

1.1

10.57

1.86E-11

FFP

194.0

177.2

16.7

10.33

3.16E-11

MW M T

16.7

15.9

0.8

10.12

5.03E-11

EMT

-18.7

-20.0

1.3

7.63

2.08E-08

MCMT

2.1

1.4

0.7

6.61

3.04E-07

DD<0

80.4

110.0

-29.6

6.35

6.10E-07

Eref

652.9

620.8

32.1

6.27

7.67E-07

TD

14.6

14.5

0.1

0.83

4.14E-01

SHM

83.2

55.3

27.9

13.08

1.09E-13

7

CMD

265.0

178.4

86.6

12.85

1.68E-13

Latitude

AHM

14.0

8.8

5.2

12.21

5.97E-13

48.83

MSP

202.5

294.7

-92.3

10.66

1.54E-11

Sample No.

MAP

1377.5

2088.9

-711.4

10.39

2.76E-11

MW M T

16.7

15.9

0.8

9.19

4.29E-10

DD>5

1759.4

1543.0

216.4

8.84

1.01E-09

MAT

8.9

7.9

1.0

8.64

1.64E-09

PAS

89.1

211.1

-122.0

8.46

2.53E-09

NFFD

279.8

260.3

19.5

8.21

4.77E-09

FFP

193.0

176.6

16.3

7.92

9.83E-09

Eref

653.5

621.6

31.9

7.21

6.16E-08

EMT

-18.7

-20.0

1.4

6.17

9.97E-07

221

A ppendix D (continued).

Sample
No. and
Latitude

Mean in

Mean out

Mean
difference

t

P

MCMT

2.2

1.4

0.8

5.93

1.94E-06

DD<0

80.1

110.9

-30.8

5.73

3.36E-06

TD

14.5

14.5

0.0

0.00

1.00E+00

Climate
Variable*

Sample No.

AHM

13.9

8.9

4.9

13.67

3.59E-14

8

CMD

262.4

179.8

82.6

13.50

4.96E-14

Latitude

SHM

81.9

55.6

26.3

13.43

5.61E-14

48.85

MAP

1385.2

2056.2

-671.0

11.36

3.37E-12

MSP

205.4

292.0

-86.6

11.31

3.73E-12

PAS

89.9

205.2

-115.3

10.23

3.92E-11

MAT

8.9

7.9

1.0

9.85

9.29E-11

DD>5

1752.3

1545.4

206.9

9.73

1.22E-10

NFFD

279.0

260.2

18.8

9.55

1.86E-10

MWMT

16.7

16.0

0.7

9.32

3.16E-10

FFP

191.6

176.7

14.9

7.53

2.66E-08

EMT

-18.7

-20.0

1.3

7.49

2.94E-08

Eref

653.8

622.2

31.6

7.39

3.82E-08

MCMT

2.2

1.4

0.8

6.92

1.33E-07

DD<0

79.7

110.4

-30.6

6.74

2.12E-07

TD

14.5

14.6

0.0

0.78

4.43E-01

Sample No.

SHM

82.7

57.4

25.3

12.73

2.14E-13

9

CMD

265.0

185.1

79.9

11.57

2.17E-12

Latitude

AHM

14.2

9.3

4.9

11.15

5.27E-12

48.87

MSP

204.9

284.5

-79.6

9.97

7.12E-11

MAP

1369.0

1988.0

-619.0

9.65

1.48E-10

MWMT

16.7

15.9

0.8

8.50

2.32E-09

DD>5

1769.0

1544.1

224.8

8.36

3.24E-09

MAT

9.0

7.9

1.1

8.22

4.61E-09

NFFD

280.0

259.9

20.1

8.10

6.18E-09

PAS

86.7

201.8

-115.1

8.02

7.67E-09

FFP

192.6

176.2

16.4

7.15

7.16E-08

Eref

656.6

622.4

34.2

6.99

1.10E-07

MCMT

2.2

1.4

0.9

6.34

6.22E-07

DD<0

77.3

112.4

-35.1

6.32

6.63E-07

EMT

-18.6

-20.0

1.4

6.23

8.46E-07

TD

14.5

14.6

-0.1

1.17

2.50E-01

Sample No.

AHM

14.5

8.8

5.7

15.32

1.95E-15

10

CMD

268.2

173.6

94.6

15.02

3.22E-15

Latitude

SHM

83.4

54.0

29.4

14.40

9.54E-15

222

A ppendix D (continued).

Sample
No. and
Latitude

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

48.89

DD>5

1755.8

1521.2

234.6

11.03

6.86E-12

MAT

8.9

7.7

1.2

10.91

8.81E-12

NFFD

278.4

257.6

20.8

10.58

1.80E-11

MSP

201.5

301.9

-100.4

10.52

2.05E-11

PAS

84.3

225.0

-140.8

10.51

2.10E-11

MAP

1319.9

2113.4

-793.5

10.47

2.30E-11

MWMT

16.7

15.9

0.8

9.98

6.92E-11

FFP

190.0

175.1

14.9

8.80

1.11E-09

Eref

656.5

617.9

38.6

8.69

1.45E-09

EMT

-18.7

-20.2

1.5

8.64

1.63E-09

MCMT

2.2

1.3

0.9

7.53

2.65E-08

DD<0

79.8

116.2

-36.4

7.42

3.57E-08

TD

14.5

14.6

0.0

0.96

3.45E-01
4.05E-14
6.19E-14

Sample No.

AHM

14.7

8.9

5.8

13.60

11

SHM

83.5

54.7

28.8

13.38

Latitude

CMD

268.4

175.6

92.8

13.07

1.11E-13

48.91

MSP

201.4

296.2

-94.7

10.21

4.12E-11

MAP

1309.5

2065.3

-755.8

9.99

6.68E-11

MAT

8.9

7.7

1.2

9.94

7.57E-11

PAS

82.8

219.8

-137.0

9.62

1.57E-10

DD>5

1761.2

1521.7

239.5

9.57

1.75E-10

NFFD

278.7

257.2

21.5

9.45

2.32E-10

FFP

190.3

174.3

16.0

8.46

2.52E-09

EMT

-18.7

-20.3

1.6

8.20

4.87E-09

MW M T

16.7

15.9

0.8

8.13

5.74E-09

Eref

656.8

618.5

38.3

7.81

1.30E-08

MCMT

2.2

1.3

1.0

7.30

4.86E-08

DD<0

78.3

117.3

-39.0

7.26

5.38E-08

TD

14.5

14.6

-0.1

2.88

7.36E-03

Sample No.

AHM

14.8

9.3

5.6

11.37

3.34E-12

12

CMD

267.8

181.5

86.3

11.00

7.34E-12

Latitude

SHM

83.3

56.0

27.4

10.80

1.13E-11

48.93

MAT

9.0

7.8

1.2

10.01

6.44E-11

DD>5

1775.4

1542.0

233.4

9.94

7.50E-11

PAS

79.7

204.5

-124.8

9.63

1.56E-10

NFFD

280.4

258.8

21.6

9.62

1.57E-10

MWMT

16.7

16.0

0.8

9.23

3.90E-10

MSP

203.6

291.7

-88.1

9.13

5.04E-10

223

A ppendix D (continued).

Sample
No. and
Latitude

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

MAP

1307.4

2009.8

-702.4

8.90

8.64E-10

FFP

192.1

175.0

17.1

8.37

3.16E-09

EMT

-18.5

-20.1

1.6

8.13

5.78E-09

MCMT

2.3

1.3

1.0

8.12

5.85E-09

DD<0

75.3

113.3

-38.1

7.90

1.02E-08

Eref

658.4

623.1

35.3

7.59

2.27E-08

TD

14.5

14.6

-0.2

3.69

9.13E-04

10.54

1.99E-11

Sample No.

AHM

14.8

9.6

5.2

13

SHM

82.6

57.5

25.1

9.34

3.02E-10

Latitude

CMD

265.9

186.6

79.4

8.94

7.79E-10

MAT

8.9

7.8

1.2

8.71

1.36E-09

DD>5

1758.3

1527.9

230.4

8.54

2.06E-09

NFFD

278.7

257.1

21.6

8.49

2.37E-09

PAS

82.1

207.1

-125.0

7.97

8.74E-09

MWMT

16.7

15.9

0.8

7.94

9.26E-09

MAP

1294.8

1950.2

-655.3

7.86

1.14E-08

MSP

203.8

284.0

-80.2

7.60

2.22E-08

MCMT

2.2

1.2

1.0

7.55

2.52E-08

48.95

EMT

-18.6

-20.3

1.7

7.40

3.77E-08

DD<0

78.3

117.4

-39.1

7.37

4.07E-08

FFP

190.2

173.9

16.3

7.26

5.41E-08

Eref

656.1

621.6

34.5

7.22

6.00E-08

TD

14.5

14.7

-0.2

3.82

6.56E-04

AHM

14.4

9.4

5.1

12.08

7.67E-13

14

SHM

80.3

56.3

24.0

10.53

2.02E-11

Latitude

MWMT

16.6

15.9

0.8

10.20

4.18E-11

48.97

MAT

8.8

7.7

1.1

9.97

7.13E-11

DD>5

1739.9

1517.7

222.2

9.79

1.07E-10

CMD

260.6

181.4

79.2

9.76

1.15E-10

MAP

1324.0

1966.6

-642.6

9.13

4.98E-10

MSP

208.8

287.9

-79.1

8.55

2.04E-09

PAS

86.5

214.0

-127.5

8.54

2.10E-09

MCMT

2.2

1.2

1.0

8.25

4.23E-09

Sample No.

NFFD

276.8

256.3

20.5

8.24

4.37E-09

DD<0

80.4

120.3

-39.9

8.17

5.16E-09

Eref

654.6

619.7

34.9

7.39

3.88E-08

EMT

-18.8

-20.3

1.5

6.59

3.23E-07

FFP

188.1

172.7

15.5

5.96

1.79E-06

224

A ppendix D (continued)

Sample
No. and
Latitude

Climate
Variable4'

Mean in

Mean out

Mean
difference

t

P

TD

14.5

14.7

-0 . 2

3.71

8.84E-04

Sample No.

AHM

15.1

9.6

5.5

14.74

5.23E-15

15

MAT

9.0

7.8

1 .2

11.45

2.81E-12

Latitude

SHM

82.8

56.8

26.1

11.25

4.24E-12

48.99

DD>5

1766.2

1526.1

240.0

1 1 .2 1

4.62E-12

MWMT

16.7

15.9

0.9

1 1 .0 0

7.26E-12

CMD

270.5

185.3

85.3

10.99

7.38E-12

NFFD

278.5

256.2

22.3

10.80

1.13E-11

MCMT

2.3

1 .2

1 .0

10.05

5.88E-11

PAS

77.1

2 1 1 .0

-133.9

9.52

2.00E-10

DD<0

75.7

117.9

-42.2

9.44

2.38E-10

FFP

189.0

172.4

16.6

8.96

7.47E-10

EMT

-18.6

-20.3

1.7

8.92

8.34E-10

Eref

661.8

623.8

38.0

8.58

1.88E-09

MAP

1269.6

1984.4

-714.8

8.55

2.05E-09

MSP

203.1

290.2

-87.1

8 .1 2

5.95E-09

TD

14.5

14.7

-0 . 2

3.70

9.04E-04

Sample No.

AHM

14.8

9.5

5.4

9.82

1.00E-10

16

NFFD

275.5

253.4

2 2 .1

8.94

7.82E-10

Latitude

MAT

8 .8

7.6

1 .2

8.85

9.67E-10

49.01

DD>5

1743.2

1505.7

237.5

8.77

1.18E-09

MCMT

2 .2

1 .1

1 .1

8.49

2.38E-09

SHM

81.0

56.4

24.6

8.47

2.45E-09

MWMT

16.7

15.8

0 .8

8.38

3.10E-09

DD<0

79.0

124.5

-45.5

8 .2 2

4.66E-09

CMD

266.4

182.1

84.3

8.08

6.47E-09

FFP

186.0

169.9

16.1

8 .0 0

8.00E-09

EMT

-18.8

-2 0 . 6

1.7

7.65

1.97E-08

MAP

1296.0

1974.8

-678.8

7.16

6.98E-08

PAS

83.4

225.9

-142.6

7.10

8.19E-08

MSP

207.4

290.4

-83.0

6.75

2.08E-07

Eref

660.3

620.5

39.8

6.73

2.20E-07

TD

14.5

14.7

-0 . 2

4.78

4.66E-05

Sample No.

AHM

15.3

9.6

5.7

10.53

2.01E-11

17

MWMT

16.8

15.9

0.9

10.23

3.96E-11

Latitude

MAT

9.0

7.7

1 .2

9.99

6.76E-11

49.03

DD>5

1766.7

1521.6

245.1

9.91

8.15E-11

SHM

82.5

57.1

25.4

9.33

3.14E-10

225

A ppendix D (continued).

Sample
No. and
Latitude

Sample No.

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

NFFD

277.6

254.4

23.2

9.28

3.50E-10

MCMT

2.3

1 .2

1 .1

9.26

3.66E-10

DD<0

75.9

1 2 1 .2

-45.3

8.95

7.70E-10

CMD

270.7

186.8

83.8

8.58

1.89E-09

MAP

1261.7

1958.6

-697.0

7.80

1.34E-08

PAS

77.4

214.4

-137.0

7.64

2.01E-08

FFP

187.9

169.8

18.1

7.50

2.87E-08

MSP

204.6

286.9

-82.3

7.30

4.82E-08

EMT

-18.7

-20.5

1 .8

7.29

4.98E-08
4.35E-07

Eref

663.2

626.1

37.2

6.48

TD

14.5

14.7

-0 . 2

4.07

3.35E-04

AHM

15.1

9.7

5.4

12.49

3.37E-13

18

SHM

82.0

57.9

24.1

10.07

5.67E-11

Latitude

MAP

1258.6

1888.8

-630.2

1 0 .0 1

6.43E-11

49.05

CMD

269.4

190.8

78.6

9.81

1.01E-10

PAS

79.8

206.8

-127.0

9.39

2.71E-10

MAT

8.9

7.7

1 .2

9.08

5.58E-10

NFFD

275.9

253.2

22.7

8.96

7.49E-10

MSP

204.6

277.9

-73.3

8.77

1.19E-09

DD>5

1753.0

1512.0

241.0

8 .6 6

1.55E-09

DD<0

78.1

123.8

-45.8

8.38

3.09E-09

MW M T

16.7

15.8

0.9

8.32

3.62E-09

MCMT

2 .2

1 .1

1 .1

8.29

3.82E-09

EMT

-18.8

-2 0 . 6

1.9

8.16

5.40E-09

FFP

186.0

168.9

17.1

7.69

1.77E-08

Eref

661.4

625.8

35.6

6.73

2.22E-07

TD

14.5

14.7

-0 . 2

4.04

3.62E-04

Sample No.

AHM

15.4

9.9

5.5

14.68

5.83E-15

19

MAT

8.9

7.8

1 .2

1 2 .6 6

2.43E-13

Latitude

SHM

82.8

58.8

24.0

12.47

3.56E-13

49.07

NFFD

276.3

254.3

2 2 .0

1 2 .2 0

6.08E-13

DD>5

1761.4

1533.0

228.4

12.14

6.84E-13

CMD

273.1

194.7

78.4

12.08

7.70E-13

MCMT

2.3

1 .2

1 .1

11.32

3.69E-12

MWMT

16.7

15.9

0.9

11.06

6.34E-12

DD<0

75.5

121.3

-45.7

10.98

7.60E-12

FFP

185.7

168.7

17.0

10.23

3.93E-11

EMT

-18.7

-2 0 . 6

1.9

9.88

8.58E-11

226

A ppendix D (continued)

Sample
No. and
Latitude

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

PAS

76.0

198.7

-122.7

9.44

2.38E-10

MAP

1237.4

1878.5

-641.0

8.90

8.69E-10

MSP

202.9

278.0

-75.0

8 .2 2

4.60E-09

Eref

664.6

630.1

34.5

8.19

4.98E-09

TD

14.5

14.7

-0.3

4.44

1.20E-04

Sample No.

AHM

15.4

9.8

5.6

12.90

1.52E-13

2 0

MCMT

2.3

1 .1

1 .2

11.17

5.04E-12

Latitude

NFFD

277.1

252.8

24.3

11.07

6.32E-12

49.09

MAT

8.9

7.7

1 .2

11.00

7.33E-12

SHM

82.7

58.2

24.5

10.83

1.06E-11

DD< 0

75.0

123.1

-48.1

10.78

1.18E-11

DD>5

1766.6

1526.5

240.1

10.53

2.03E-11

MW M T

16.7

15.9

0.9

9.82

1.00E-10

PAS

75.4

203.6

-128.2

9.59

1.70E-10

CMD

271.5

194.4

77.0

9.46

2.27E-10

EMT

-18.7

-20.7

2 .0

9.18

4.46E-10

FFP

187.0

167.1

19.8

9.03

6.32 E-10

MAP

1240.8

1894.7

-653.9

9.01

6.67E-10

MSP

203.5

279.0

-75.4

8.33

3.52E-09

TD

14.4

14.7

-0.3

5.58

5.05E-06

Eref

662.9

632.5

30.5

5.37

8.95E-06

Sample No.

AHM

15.2

9.9

5.3

1 1 .8 6

1.22E-12

2 1

NFFD

275.3

251.3

24.0

10.14

4.82E-11

Latitude

SHM

81.6

58.7

2 2 .8

9.95

7.31E-11

9.72

1.26E-10
2.68E-10

49.11

Sample No.

PAS

80.0

2 0 1 .6

-1 2 1 . 6

MAP

1254.3

1851.6

-597.3

9.39

FFP

185.1

165.8

19.3

9.35

2.98E-10

CMD

269.2

197.3

71.9

9.34

3.05E-10

DD<0

77.8

126.3

-48.5

9.32

3.18E-10

MAT

8.9

7.7

1 .2

9.24

3.89E-10

MCMT

2 .2

1 .1

1 .2

9.23

3.90E-10

DD>5

1751.2

1517.9

233.3

8 .8 6

9.51E-10

EMT

-18.8

-20.9

2 .1

8.81

1.08E-09

MWMT

16.7

15.8

0.9

8.48

2.43E-09

MSP

205.9

274.1

-6 8 . 2

8.42

2.83E-09

Eref

662.3

632.3

30.0

5.72

3.42E-06

TD

14.5

14.8

-0.3

4.58

8.05E-05

AHM

15.3

9.8

5.5

14.50

8.02E-15

227

Appendix D (continued)

Sample
No. and
Latitude

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P
1.83E-13

2 2

DD<0

75.0

129.3

-54.3

12.81

Latitude

MCMT

2.3

1 .0

1.3

12.59

2.82E-13

49.13

NFFD

276.7

250.3

26.4

11.67

1.78E-12

MAT

9.0

7.6

1.3

11.63

1.93E-12

SHM

82.2

58.3

23.9

11.49

2.59E-12

MW M T

16.7

15.8

1 .0

11.25

4.24E-12

MAP

1252.9

1856.1

-603.2

11.07

6.23E-12

DD>5

1773.6

1503.5

270.0

11.05

6.51E-12

PAS

76.4

209.0

-132.6

10.47

2.31E-11

EMT

-18.8

-2 1 . 0

2 .2

10.36

2.92E-11

FFP

186.3

164.9

21.4

9.63

1.53E-10

CMD

271.3

193.0

78.3

9.60

1.66E-10

MSP

205.1

274.8

-69.7

9.51

2.03E-10

Eref

665.6

627.7

37.9

6.34

6.26E-07

TD

14.4

14.8

-0.3

5.57

5.24E-06

Sample No.

NFFD

275.8

249.9

25.9

16.09

1.12E-15

23

MCMT

2.3

1 .0

1.3

15.28

4.11E-15

Latitude

DD<0

75.6

131.2

-55.6

14.66

1.17E-14
1.39E-14

49.15

Sample No.

MAT

8.9

7.6

1.3

14.55

AHM

15.4

9.9

5.5

14.04

3.40E-14

DD>5

1762.8

1507.0

255.8

13.32

1.22E-13

FFP

184.4

164.0

20.4

13.24

1.42E-13

MWMT

16.7

15.8

1 .0

12.39

6.92E-13

EMT

-18.8

-2 1 . 1

2.3

12.35

7.52E-13

SHM

81.5

58.7

2 2 .8

11.96

1.61E-12

MAP

1246.9

1842.1

-595.2

9.12

7.02E-10

PAS

75.7

208.8

-133.1

8.89

1.21E-09

CMD

271.7

195.4

76.2

8.57

2.59E-09

MSP

206.0

273.4

-67.5

8 .0 2

9.89E-09

TD

14.5

14.8

-0.3

6.42

6.03E-07

Eref

6 6 6 .2

629.6

36.6

5.83

2.93E-06

AHM

15.2

9.5

5.7

14.17

1.43E-14

24

MCMT

2.3

0 .8

1.5

12.49

3.43E-13

Latitude

MW M T

16.8

15.7

1 .1

12.35

4.52E-13

49.17

MAT

9.0

7.5

1.4

1 2 .1 0

7.44E-13

DD<0

74.7

136.9

-62.2

12.07

7.90E-13

SHM

81.4

56.7

24.7

12.03

8.53E-13

DD>5

1768.0

1486.3

281.7

11.79

1.40E-12

228

A ppendix D (continued).

Sample
No. and
Latitude

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

NFFD

276.1

247.7

28.4

10.93

8.47E-12

CMD

270.8

188.2

82.6

10.67

1.49E-11

MAP

1257.9

1907.5

-649.6

10.32

3.21E-11

EMT

-18.8

-2 1 . 2

2.5

9.69

1.35E-10

FFP

184.9

162.4

22.5

9.56

1.83E-10

MSP

206.8

282.6

-75.8

9.53

1.95E-10

PAS

76.4

230.2

-153.8

9.47

2.25E-10

Eref

6 6 6 .8

626.1

40.7

7.44

3.41E-08

TD

14.4

14.9

-0.4

5.80

2.77E-06

Sample No.

MCMT

2.3

0.9

1.4

14.21

1.34E-14

25

DD<0

73.5

134.0

-60.5

13.35

6.51E-14

Latitude

AHM

15.3

9.7

5.6

13.28

7.37E-14

49.19

MAT

9.0

7.6

1.4

12.71

2.21E-13

DD>5

1776.3

1499.5

276.8

12.49

3.39E-13

MWMT

16.8

15.7

1 .0

12.46

3.60E-13

NFFD

277.0

248.5

28.5

12.27

5.26E-13

SHM

81.6

57.5

24.1

11.71

1.63E-12

EMT

-18.7

-2 1 . 2

2 .6

10.84

1.02E-11

FFP

186.0

162.5

23.5

10.37

2.89E-11

MAP

1254.9

1896.6

-641.7

8.60

1.78E-09

CMD

271.6

191.0

80.6

8.46

2.52E-09

7.76

1.46E-08

7.65

1.96E-08

MSP

206.2

281.2

-75.0

PAS

74.9

224.7

-149.7

TD

14.4

14.9

-0.4

7.17

6.91E-08

Eref

666.9

627.7

39.3

5.79

2.83E-06

Sample No.

MCMT

2.3

1 .0

1.3

14.10

1.64E-14

26

DD<0

76.0

130.7

-54.7

13.75

3.08E-14

Latitude

NFFD

275.9

249.6

26.3

13.46

5.28E-14

49.21

AHM

15.0

1 0 .1

5.0

12.08

7.67E-13

FFP

184.9

163.1

2 1 .8

1 2 .0 2

8.74E-13

MAT

8.9

7.7

1.3

11.71

1.62E-12

EMT

-18.8

-2 1 . 1

2.3

11.30

3.84E-12

DD>5

1768.0

1515.5

252.5

11.17

5.02E-12

MW M T

16.7

15.8

0.9

10.74

1.27E-11

SHM

80.6

59.4

2 1 .1

10.57

1.84E-11

MAP

1283.5

1806.8

-523.3

10.04

5.99E-11

PAS

78.3

200.9

-1 2 2 . 6

9.91

8.08E-11

MSP

208.7

270.0

-61.2

8.85

9.72E-10

229

A ppendix D (continued)

Sample
No. and
Latitude

Sample No.

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

CMD

268.3

202.5

65.8

7.60

2.21E-08

TD

14.5

14.8

-0.4

6.42

5.01E-07

Eref

665.9

635.6

30.3

4.62

7.20E-05

MCMT

2.3

0.7

1 .6

14.35

1.04E-14

5.9

13.08

1.09E-13

-69.2

13.02

1.21E-13

27

AHM

15.1

9.2

Latitude

DD<0

75.4

144.6

49.23

MAT

8.9

7.4

1.5

12.99

1.28E-13

12.93

1.46E-13

12.78

1.94E-13

DD>5

1769.5

1463.0

306.4

MWMT

16.8

15.6

1 .2

NFFD

275.8

243.3

32.5

11.92

1.06E-12

SHM

80.6

54.3

26.3

11.60

2.06E-12

EMT

-18.8

-21.7

2.9

10.58

1.81E-11

FFP

184.7

158.0

26.7

10.33

3.16E-11

MAP

1271.0

1993.1

-722.1

8.93

8.12E-10

CMD

268.9

180.9

8 8 .0

8.81

1.08E-09

PAS

77.7

254.2

-176.4

8.56

1.97E-09

MSP

209.2

295.9

-86.7

8.26

4.18E-09

TD

14.5

14.9

-0.4

7.02

1.01E-07

Eref

667.0

625.5

41.5

6.27

7.56E-07

Sample No.

MCMT

2 .2

0 .8

1.4

16.92

< 2.2e-16

28

MAT

8.9

7.6

1.3

16.90

< 2.2e-16

Latitude

DD>5

1752.9

1495.2

257.7

16.08

5.56E-16

49.25

MW M T

16.7

15.7

1 .0

15.34

1.89E-15

DD<0

78.7

139.5

-60.8

15.19

2.42E-15

NFFD

273.3

246.4

26.9

14.36

1.04E-14

AHM

14.7

9.6

5.1

13.79

2.91E-14

SHM

78.8

56.5

22.3

13.41

5.87E-14

EMT

-19.0

-21.5

2.4

11.36

3.41E-12

FFP

181.7

160.1

2 1 .6

10.59

1.78E-11

PAS

83.0

225.4

-142.4

9.62

1.59E-10

CMD

266.4

192.6

73.8

9.40

2.64E-10

MAP

1304.8

1915.3

-610.5

8 .8 6

9.52E-10

MSP

213.1

284.0

-70.9

8.78

1.17E-09

Eref

6 6 8 .2

634.4

33.8

5.96

1.79E-06

TD

14.5

14.9

-0.4

5.58

5.11E-06

Sample No.

MCMT

2 .1

0 .8

1.4

16.59

2.41E-16

29

DD<0

80.3

141.5

-61.3

15.41

1.69E-15

Latitude

NFFD

272.9

245.8

27.1

15.15

2.59E-15

230

A ppendix D (continued).

Sample
No. and
Latitude

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

49.27

MAT

8 .8

7.5

1.3

14.83

4.55E-15

DD>5

1747.7

1491.0

256.7

14.33

1.09E-14

MWMT

16.7

15.7

1 .0

13.35

6.55E-14

FFP

181.3

159.4

21.9

11.75

1.51E-12

EMT

-19.1

-2 1 . 6

2.5

11.04

6.70E-12

AHM

14.7

9.5

5.2

1 0 .8 6

9.81E-12

SHM

78.5

56.1

22.4

10.65

1.56E-11

PAS

83.5

227.6

-144.1

9.46

2.30E-10

MAP

1307.0

1919.9

-613.0

8.84

9.91E-10
2.54E-09

MSP

213.9

284.9

-70.9

8.46

CMD

265.5

190.3

75.2

8.30

3.82E-09

Eref

667.4

632.4

34.9

5.52

6.03E-06

TD

14.6

14.9

-0.4

5.31

1.08E-05

Sample No.

MCMT

2 .2

0.9

1.3

15.66

1.10E-15

30

DD<0

78.2

133.3

-55.1

14.75

5.19E-15

Latitude

NFFD

273.5

248.5

25.0

13.91

2.30E-14

49.29

MAT

8.9

7.7

1 .2

13.31

7.01E-14

DD>5

1757.8

1522.6

235.2

1 2 .1 2

7.07E-13

AHM

14.7

1 0 .0

4.7

12.06

8.03E-13

FFP

181.6

161.1

20.5

11.45

2.79E-12

SHM

78.5

58.5

2 0 .1

11.44

2.88E-12

EMT

-19.0

-21.3

2.3

11.35

3.47E-12

MW M T

16.7

15.8

0.9

11.31

3.78E-12

PAS

81.3

200.3

-119.0

9.84

9.53E-11

MAP

1304.4

1827.3

-522.9

9.07

5.79E-10

MSP

213.9

274.5

-60.6

8.89

8.81E-10

CMD

266.4

2 0 1 .6

64.8

8.09

6.42E-09

TD

14.5

14.9

-0.3

5.96

1.78E-06

Eref

669.4

640.4

29.0

4.77

4.78E-05

Sample No.

NFFD

274.2

250.1

24.1

16.11

5.29E-16

31

MCMT

2 .2

1 .0

1 .2

15.49

1.45E-15

Latitude

DD<0

78.0

129.4

-51.4

14.73

5.34E-1S

49.31

MAT

8.9

7.8

1 .1

14.28

1.18E-14

DD>5

1760.1

1547.7

212.4

13.30

7.12E-14

FFP

182.2

161.6

2 0 .6

13.30

7.20E-14

PAS

80.5

178.2

-97.8

12.83

1.77E-13

SHM

78.4

60.4

18.0

12.26

5.34E-13

MW M T

16.8

15.9

0 .8

12.04

8.29E-13

231

A ppendix D (continued).

Sample
No. and
Latitude

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

EMT

-19.0

-2 1 . 2

2.3

11.87

1.17E-12

AHM

14.8

10.4

4.4

1 1 .1 2

5.67E-12

MSP

214.9

265.9

-51.0

10.83

1.05E-11

MAP

1300.5

1743.9

-443.4

10.05

5.85E-11

CMD

265.1

210.4

54.6

8.48

2.42E-09

TD

14.5

14.9

-0.4

6.91

1.36E-07

Eref

668.4

647.1

21.3

4.16

2.56E-04

Sample No.

MCMT

2 .2

0.9

1 .2

12.75

2.07E-13

32

MAT

8.9

7.7

1 .1

1 2 .0 2

8.69E-13

Latitude

DD<0

79.9

132.5

-52.6

11.78

1.42E-12

NFFD

273.7

247.9

25.8

11.35

3.48E-12

DD>5

1750.4

1530.4

2 2 0 .0

11.27

4.07E-12

AHM

14.5

10.4

4.1

10.82

1.07E-11

49.33

SHM

77.4

59.5

17.8

10.57

1.86E-11

MWMT

16.7

15.8

0.9

1 0 .1 1

5.20E-11

FFP

182.1

159.6

22.5

1 0 .1 0

5.24E-11

1 0 .0 1

6.39E-11

EMT

-19.0

-21.4

2.4

MAP

1320.5

1756.3

-435.8

9.36

2.89E-10

PAS

83.1

187.5

-104.4

9.30

3.36E-10

MSP

216.8

269.2

-52.4

9.12

5.15E-10

CMD

261.0

207.9

53.1

7.49

2.94E-08

TD

14.6

14.9

-0.3

6.36

5.90E-07

Eref

665.1

646.6

18.5

3.51

1.49E-03

EMT

-19.1

-21.5

2.4

13.70

3.38E-14

33

NFFD

273.0

248.9

24.1

13.15

9.45E-14

Latitude

MCMT

2 .1

0.9

1 .2

13.08

1.08E-13

49.35

DD<0

80.7

132.3

-51.6

12.80

1.85E-13

Sample No.

FFP

181.1

160.5

2 0 .6

12.42

3.91E-13

SHM

76.0

60.1

15.9

10.90

8.99E-12

MAT

8 .8

7.8

1 .0

10.46

2.37E-11

DD>5

1749.9

1543.0

206.9

10.05

5.90E-11

MWMT

16.7

15.9

0 .8

9.93

7.67E-11

AHM

14.2

10.3

3.9

9.76

1.14E-10

MSP

220.5

266.9

-46.3

9.41

2.57E-10

PAS

8 6 .0

185.1

-99.2

9.32

3.19E-10

MAP

1349.1

1763.2

-414.1

8.97

7.32E-10

TD

14.6

14.9

-0.4

7.35

4.31E-08

CMD

258.5

2 1 1 .8

46.7

7.23

5.86E-08

232

A ppendix D (continued).

Sample
No. and
Latitude

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

Eref

6 6 6 .8

649.5

17.3

3.37

2.12E-03

Sample No.

MCMT

2 .2

0 .8

1.4

15.06

3.03E-15

34

EMT

-19.1

-21.7

2 .6

14.61

6.63E-15

14.36

1.03E-14

Latitude

DD<0

78.9

138.1

-59.2

49.37

NFFD

274.2

246.2

28.1

14.05

1.81E-14

FFP

182.1

158.6

23.6

13.08

1.09E-13

MAT

8.9

7.7

1 .2

11.87

1.18E-12

SHM

76.6

58.5

18.2

11.77

1.44E-12

AHM

14.4

1 0 .1

4.3

11.35

3.45E-12

DD>5

1761.7

1520.6

241.1

11.14

5.34E-12

MW M T

16.7

15.8

1 .0

10.58

1.81E-11

MAP

1330.5

1792.3

-461.8

9.43

2.44E-10
3.89E-10
1.01E-09

PAS

81.4

197.5

-116.0

9.23

MSP

219.2

273.8

-54.5

8.84

CMD

260.7

205.6

55.1

8 .0 2

7.56E-09

TD

14.6

15.0

-0.4

7.11

8.05E-08

Eref

6 6 8 .8

645.8

23.0

4.02

3.76E-04

Sample No.

MCMT

2 .2

1 .0

1.3

13.25

7.91E-14

35

EMT

-18.9

-21.5

2 .6

12.97

1.33E-13

Latitude

DD<0

77.2

131.2

-54.0

12.71

2.21E-13

49.39

NFFD

275.7

248.9

26.8

12.54

3.11E-13

FFP

184.2

159.8

24.4

11.76

1.49E-12

MAT

9.0

7.9

1 .1

11.67

1.77E-12

SHM

76.8

60.4

16.4

11.62

1.96E-12

DD>5

1773.0

1558.7

214.3

11.08

6.07E-12

MWMT

16.8

15.9

0.9

10.77

1.21E-11

AHM

14.5

1 0 .6

3.9

10.26

3.66E-11

MSP

219.4

267.1

-47.6

9.06

5.86E-10

PAS

80.2

176.4

-96.3

8.56

2.00E-09

MAP

1326.9

1731.2

-404.3

8.28

4.01E-09

CMD

260.1

215.9

44.2

8.09

6.45E-09

TD

14.6

15.0

-0.4

7.27

5.30E-08

Eref

667.5

656.1

11.4

2.59

1.48E-02

Sample No.

NFFD

273.8

247.4

26.4

11.94

1.03E-12

36

MCMT

2 .1

0.9

1 .2

11.77

1.44E-12

Latitude

EMT

-19.1

-21.7

2 .6

11.44

2.85E-12

49.41

FFP

182.1

158.7

23.4

11.38

3.28E-12

DD<0

80.9

135.4

-54.5

11.31

3.80E-12

233

A ppendix D (continued).

Sample
No. and
Latitude

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

MAT

8.9

7.8

1 .1

10.91

8.89E-12

DD>5

1753.9

1543.3

2 1 0 .6

10.60

1.73E-11

SHM

75.6

59.2

16.4

10.30

3.37E-11

MW M T

16.7

15.9

0.9

10.23

3.95E-11

AHM

14.3

10.4

3.9

9.91

8.08E-11

MSP

2 2 2 .1

271.6

-49.5

8.65

1.58E-09

MAP

1341.2

1765.1

-423.9

8.05

7.00E-09

PAS

84.0

187.6

-103.6

8 .0 2

7.71E-09

CMD

257.6

2 1 0 .8

46.7

6.95

1.23E-07

TD

14.6

15.0

-0.4

6.40

5.36E-07

Eref

666.9

653.6

13.2

2.79

9.14E-03

Sample No.

MCMT

2 .2

0.9

1.3

14.99

3.40E-15

37

NFFD

275.2

248.8

26.5

14.87

4.23E-15
4.72E-15

Latitude

EMT

-19.0

-2 1 . 6

2 .6

14.80

49.43

DD<0

78.7

133.4

-54.7

14.39

9.74E-15

FFP

183.9

160.0

23.9

13.23

8.15E-14

SHM

76.0

60.1

15.9

13.05

1.16E-13

MAT

8.9

7.9

1 .1

12.44

3.74E-13

DD>5

1765.7

1558.9

206.8

12.09

7.54E-13

MW M T

16.8

15.9

0.9

1 1 .2 2

4.58E-12

AHM

14.4

10.5

3.9

10.79

1.16E-11

PAS

82.3

177.5

-95.2

10.70

1.40E-11

MSP

221.7

267.6

-45.8

10.53

2.02E-11

MAP

1338.1

1733.7

-395.6

9.49

2.15E-10

CMD

257.9

215.7

42.2

8.65

1.58E-09

TD

14.6

15.0

-0.4

7.05

9.45E-08

Eref

666.7

656.8

9.9

2.34

2.62E-02

Sample No.

EMT

-18.8

-21.9

3.1

17.50

< 2.2e-16

38

MCMT

2.3

0 .8

1.5

15.96

6.70E-16

Latitude

DD<0

76.3

141.0

-64.7

15.76

9.34E-16

49.45

NFFD

277.3

245.4

31.9

14.55

7.41E-15

SHM

76.7

57.7

18.9

13.84

2.61E-14

FFP

186.2

157.5

28.7

13.80

2.82E-14

MAT

9.0

7.7

1.3

12.15

6.70E-13

DD>5

1788.2

1531.2

257.0

1 1 .6 6

1.83E-12

MWMT

16.9

15.8

1 .0

11.46

2.76E-12

MSP

220.9

276.7

-55.8

11.41

3.07E-12

PAS

78.9

196.9

-118.0

10.81

1.10E-11

234

A ppendix D (continued)

Sample
No. and
Latitude

Sample No.

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

AHM

14.6

1 0 .0

4.6

10.72

1.32E-11

TD

14.6

15.1

-0.5

10.39

2.72E-11

MAP

1332.6

1807.3

-474.8

10.16

4.64E-11

CMD

259.5

206.0

53.5

8.16

5.40E-09

Eref

6 6 8 .0

652.5

15.4

3.02

5.20E-03

EMT

-18.9

-2 1 . 8

2.9

14.19

1.40E-14
5.43E-14
9.85E-14

39

FFP

186.5

158.5

28.0

13.45

Latitude

NFFD

277.3

247.6

29.7

13.13

49.47

MCMT

2 .2

0.9

1.4

13.11

1.02E-13

DD<0

77.5

135.6

-58.1

12.83

1.75E-13

MWMT

16.8

16.0

0.9

11.60

2.04E-12

MAT

9.0

7.9

1 .1

11.57

2.18E-12

DD>5

1781.5

1564.2

217.3

10.97

7.78E-12

SHM

76.1

59.2

16.9

10.48

2.24E-11

AHM

14.5

10.4

4.1

8.78

1.15E-09

MSP

222.4

272.9

-50.4

8.60

1.79E-09

PAS

79.3

181.6

-102.3

8.47

2.49E-09

TD

14.6

15.1

-0.5

8.17

5.19E-09

CMD

257.4

215.2

42.1

7.53

2.64E-08

MAP

1335.0

1765.1

-430.0

7.34

4.39E-08

Eref

666.5

661.7

4.8

1.73

9.42E-02

Sample No.

EMT

-19.1

-2 2 . 0

2 .8

14.84

4.41E-15

40

MCMT

2 .1

0 .8

1.3

13.39

6.06E-14

Latitude

DD<0

80.9

137.8

-56.9

13.12

1.01E-13

49.49

FFP

183.5

157.5

26.1

12.85

1.69E-13

NFFD

274.6

246.4

28.2

12.78

1.92E-13

SHM

74.4

58.2

16.2

11.91

1.08E-12

MWMT

16.8

15.9

0.9

11.36

3.40E-12

MAT

8.9

7.8

1 .1

11.33

3.63E-12

DD>5

1759.4

1548.5

210.9

11.15

5.31E-12

AHM

13.9

1 0 .2

3.7

10.78

1.16E-11

Sample No.

TD

14.6

15.1

-0.4

10.34

3.05E-11

MSP

226.0

275.5

-49.5

10.29

3.43E-11

MAP

1370.7

1775.1

-404.4

9.42

2.54E-10

PAS

84.2

185.8

-101.7

8.90

8.56E-10

CMD

253.8

209.3

44.4

8 .2 1

4.74E-09

Eref

6 6 6 .6

657.7

8.9

2.51

1.78E-02

EMT

-19.1

-2 1 . 8

2.7

11.84

1.27E-12

235

A ppendix D (continued)

Sample
No. and
Latitude

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

41

MCMT

2 .1

0.9

1 .2

11.38

3.28E-12

Latitude

FFP

183.5

158.3

25.2

11.36

3.39E-12

11.03

6.82E-12

49.51

NFFD

274.4

247.6

26.9

DD<0

81.2

135.8

-54.6

10.49

2.22E-11

MAT

8.9

7.9

1 .0

10.04

6.04E-11

DD>5

1759.2

1561.6

197.6

9.70

1.32E-10

SHM

74.0

58.3

15.7

9.44

2.41E-10

MWMT

16.7

15.9

0 .8

8.97

7.36E-10

MSP

227.4

277.1

-49.7

8.79

1.14E-09

TD

14.6

15.1

-0.4

8.54

2.10E-09

AHM

13.8

10.3

3.5

8.31

3.64E-09

PAS

85.5

181.9

-96.5

8 .2 1

4.71E-09

MAP

1379.3

1776.1

-396.8

7.20

6.28E-08

CMD

252.7

209.4

43.3

6.73

2.20E-07

Eref

666.7

659.7

7.0

2.03

5.18E-02

Sample No.

EMT

-19.2

-21.9

2.7

14.99

3.44E-15

42

MCMT

2 .1

0.9

1 .2

14.61

6.61E-15
8.43E-15

Latitude

FFP

183.1

158.3

24.8

14.47

49.53

NFFD

274.1

247.5

26.6

13.94

2.18E-14

DD<0

82.1

135.2

-53.1

13.39

5.99E-14

Sample No.

MAT

8.9

7.9

1 .0

12.45

3.68E-13

DD>5

1757.5

1568.1

189.4

11.64

1.90E-12

SHM

73.3

58.8

14.5

10.98

7.58E-12

AHM

13.6

10.4

3.2

10.13

4.87E-11

MW M T

16.7

16.0

0 .8

1 0 .1 1

5.14E-11

MSP

229.6

275.3

-45.7

9.21

4.14E-10

PAS

86.9

178.4

-91.5

8.90

8.60E-10

TD

14.7

15.1

-0.4

8.76

1.20E-09

CMD

250.9

2 1 2 .8

38.1

8.15

5.45E-09

MAP

1399.4

1752.7

-353.3

7.86

1.14E-08

Eref

667.0

661.6

5.4

1.99

5.63E-02

EMT

-19.2

-21.9

2.7

13.41

5.81E-14

43

FFP

183.5

159.1

24.3

12.35

4.48E-13

Latitude

NFFD

274.6

248.3

26.3

11.69

1.72E-12

49.55

MCMT

2 .1

0.9

1 .2

11.65

1.86E-12

DD<0

81.9

135.8

-53.9

10.65

1.57E-11

MAT

8.9

7.9

1 .0

9.63

1.55E-10

DD>5

1759.7

1570.9

188.8

9.33

3.08E-10

236

A ppendix D (continued).

Sample
No. and
Latitude

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P
8.98E-10

SHM

72.8

57.7

15.1

8.89

MW M T

16.7

16.0

0 .8

8.47

2.45E-09

TD

14.7

15.1

-0.5

8.30

3.76E-09

AHM

13.5

10.3

3.3

8.19

4.91E-09

MSP

231.4

281.5

-50.1

6.80

1.81E-07

MAP

1409.2

1795.1

-385.9

6 .1 2

1.15E-06

PAS

8 6 .8

187.9

-1 0 1 . 1

5.89

2.14E-06

CMD

248.9

206.1

42.8

5.78

2.91E-06

Eref

666.7

658.3

8.5

1.62

1.17E-01

Sample No.

EMT

-19.1

-2 1 . 8

2.7

13.58

4.26E-14

44

FFP

184.9

160.0

24.9

12.59

2.82E-13

Latitude

MCMT

2 .1

0.9

1 .2

12.37

4.31E-13

49.57

NFFD

275.8

249.2

26.6

11.39

3.19E-12

DD<0

80.5

134.2

-53.7

11.34

3.57E-12

SHM

72.9

57.8

15.2

9.73

1.24E-10

MAT

8.9

7.9

1 .0

9.66

1.45E-10

DD>5

1768.4

1577.9

190.4

9.31

3.24E-10

TD

14.7

15.1

-0.5

9.30

3.30E-10
7.93E-10

AHM

13.6

10.4

3.2

8.94

MSP

231.5

280.4

-49.0

8.72

1.33E-09

MWMT

16.8

16.0

0 .8

8.40

2.93E-09

MAP

1407.9

1773.4

-365.4

7.36

4.15E-08

CMD

247.8

206.3

41.5

7.14

7.36E-08

PAS

85.5

180.3

-94.8

6.98

1.13E-07

Eref

665.4

659.0

6.4

1.69

1.01E-01

Sample No.

FFP

185.0

160.2

24.8

12.46

3.62E-13

45

EMT

-19.1

-2 1 . 8

2.7

12.27

5.27E-13

Latitude

NFFD

275.7

249.6

26.1

11.62

1.97E-12

49.59

MCMT

2 .1

0.9

1 .2

11.24

4.35 E-12

DD<0

81.9

134.9

-53.0

11.08

6.18E-12

MAT

8.9

8 .0

0.9

10.05

5.92E-11

DD>5

1767.6

1584.1

183.5

9.74

1.20E-10

SHM

72.0

57.5

14.5

9.12

5.09E-10

MW M T

16.8

16.0

0 .8

8.81

1.07E-09

MSP

234.7

280.8

-46.1

8 .2 2

4.61E-09

TD

14.7

15.1

-0.5

7.97

8.57E-09

AHM

13.5

10.3

3.1

7.68

1.83E-08

PAS

8 6 .8

176.2

-89.4

7.43

3.44E-08

237

Appendix D (continued).

Sample
No. and
Latitude

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

CMD

244.3

205.9

38.5

6.97

1.16E-07

MAP

1418.1

1773.6

-355.5

6.54

3.64E-07

Eref

664.7

660.6

4.1

1.34

1.91E-01

Sample No.

FFP

186.9

159.6

27.3

15.86

7.93E-16

46

EMT

-19.0

-21.9

2.9

15.73

9.82E-16

Latitude

MCMT

2 .1

0 .8

1.3

14.40

9.54E-15

14.34

1.06E-14
5.52E-14

49.61

NFFD

277.4

248.7

28.7

DD<0

80.4

137.6

-57.2

13.44

MAT

9.0

7.9

1 .0

1 2 .2 1

5.92E-13

DD>5

1776.4

1577.7

198.7

11.30

3.81E-12

SHM

71.9

56.6

15.4

10.41

2.65E-11

TD

14.6

15.2

-0.5

9.82

9.97E-11

MWMT

16.8

16.0

0 .8

9.81

1.01E-10

MSP

235.1

284.6

-49.5

9.62

1.58E-10

PAS

84.7

179.5

-94.8

8.75

1.24E-09

AHM

13.5

1 0 .2

3.3

8.48

2.43E-09

CMD

242.3

202.4

39.9

7.95

9.14E-09

MAP

1415.9

1781.0

-365.2

6.96

1.18E-07

Eref

662.5

659.8

2 .8

1.04

3.07E-01

EMT

-19.1

-2 1 . 8

2.7

15.44

1.58E-15

47

FFP

186.7

160.5

26.2

14.62

6.51E-15

Latitude

MCMT

2 .1

0.9

1 .2

13.75

3.09E-14

49.63

NFFD

276.8

249.7

27.1

13.56

4.43E-14

DD<0

83.1

136.7

-53.6

12.61

2.71E-13

MAT

8.9

7.9

1 .0

11.28

3.99E-12

SHM

70.5

56.7

13.8

1 0 .8 6

9.84E-12

DD>5

1768.8

1585.5

183.3

10.41

2.66E-11

MSP

238.8

285.1

-46.3

1 0 .2 0

4.16E-11

AHM

13.1

10.4

2.7

9.84

9.50E-11

Sample No.

TD

14.7

15.2

-0.5

9.81

1.02E-10

MWMT

16.7

16.0

0.7

8.57

1.92E-09

PAS

88.9

176.2

-87.4

8.55

2.03E-09

CMD

236.7

2 0 2 .2

34.5

8 .1 2

5.94E-09

MAP

1451.5

1766.7

-315.2

7.57

2.40E-08

Eref

659.9

660.5

-0 . 6

0.25

8.01E-01

Sample No.

FFP

187.8

158.7

29.1

15.61

1.19E-15

48

EMT

-19.0

-2 2 . 0

3.0

14.92

3.88E-15

Latitude

NFFD

277.8

247.1

30.7

13.56

4.41E-14

238

A ppendix D (continued).

Sample
No. and
Latitude

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

49.65

MCMT

2 .1

0.7

1.3

13.04

1.17E-13

DD<0

82.3

141.9

-59.7

12.57

2.93E-13

MAT

8.9

7.9

1 .1

10.51

2.11E-11

DD>5

1775.0

1566.3

208.7

1 0 .0 0

6.55E-11

SHM

70.3

55.4

15.0

9.93

7.78E-11

AHM

13.2

1 0 .2

3.0

9.52

1.97E-10

TD

14.7

15.2

-0.5

9.39

2.72E-10

MSP

240.1

291.0

-50.9

9.01

6.67E-10

MWMT

16.8

16.0

0 .8

8.58

1.89E-09

PAS

8 6 .6

190.1

-103.5

7.06

9.25E-08

CMD

234.7

196.6

38.1

6.94

1.27E-07

MAP

1437.6

1786.6

-348.9

6.82

1.73E-07

Eref

658.9

657.1

1.7

0.47

6.42E-01

FFP

188.4

162.6

25.9

14.60

6.69E-15

49

EMT

-19.0

-21.7

2.7

14.11

1.60E-14

Latitude

MCMT

2 .1

0.9

1 .1

13.04

1.18E-13

49.67

NFFD

278.2

252.0

26.2

12.99

1.28E-13

DD<0

82.4

131.7

-49.3

12.36

4.42E-13

TD

14.7

15.2

-0.5

10.93

8.50E-12

MAT

9.0

8 .1

0.9

10.42

2.57E-11

Sample No.

AHM

13.1

1 0 .8

2.3

9.95

7.36E-11

SHM

70.0

57.8

1 2 .2

9.55

1.87E-10

DD>5

1778.2

1608.6

169.6

9.41

2.58E-10

MSP

241.7

280.9

-39.2

8.91

8.37E-10

MWMT

16.8

16.1

0.7

7.88

1.10E-08

MAP

1447.3

1700.3

-253.0

7.11

8.10E-08

PAS

86.9

162.4

-75.5

6.43

4.93E-07

CMD

232.9

205.7

27.1

6.08

1.29E-06

Eref

658.3

661.7

-3.4

1.14

2.62E-01

Sample No.

EMT

-19.2

-2 1 . 8

2 .6

16.84

< 2.2e-16

50

FFP

186.9

161.1

25.8

15.59

1.25E-15

Latitude

NFFD

276.8

249.9

26.9

13.92

2.26E-14

49.69

MCMT

2 .0

0 .8

1 .2

13.90

2.34E-14

DD<0

85.4

138.4

-53.1

12.73

2.14E-13

MSP

245.1

287.8

-42.7

10.52

2.04E-11

SHM

6 8 .8

56.2

1 2 .6

10.26

3.70E-11

MAT

8.9

8 .0

0.9

10.03

6.19E-11

AHM

13.0

1 0 .6

2.4

9.73

1.24E-10

239

A ppendix D (continued).

Sample
No. and
Latitude

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

DD>5

1768.9

1587.8

181.1

9.49

2.14E-10

TD

14.8

15.2

-0.5

9.21

4.16E-10

PAS

89.7

173.4

-83.7

8.26

4.18E-09

MAP

1458.4

1724.1

-265.6

7.86

1.14E-08

MWMT

16.8

16.1

0.7

7.78

1.40E-08

CMD

230.3

198.0

32.3

6.80

1.81E-07

Eref

658.8

658.7

0 .1

0.04

9.66E-01

Sample No.

FFP

188.2

161.5

26.7

13.42

5.77E-14

51

EMT

-19.1

-2 1 . 8

2.7

12.96

1.37E-13

Latitude

MCMT

2 .0

0 .8

1 .2

11.79

1.40E-12

49.71

NFFD

277.7

250.4

27.3

11.47

2.68E-12

DD<0

84.0

136.4

-52.4

10.99

7.44E-12

TD

14.8

15.2

-0.5

9.95

7.47E-11

AHM

13.1

10.7

2.4

9.54

1.89E-10

MAT

8.9

8 .0

1 .0

8.96

7.45E-10

SHM

68.4

56.1

12.3

8.40

2.98E-09

DD>5

1775.1

1594.7

180.4

8.16

5.33E-09

MSP

247.0

288.8

-41.8

7.95

8.98E-09

MW M T

16.8

16.1

0.7

6.97

1.16E-07

MAP

1452.4

1715.8

-263.4

6.76

2.04E-07

PAS

8 8 .0

172.8

-84.8

5.96

1.80E-06

CMD

227.6

197.9

29.7

5.42

7.87E-06

Eref

657.3

658.9

-1 . 6

0.43

6.70E-01

Sample No.

FFP

188.9

164.2

24.8

14.46

8.67E-15

52

EMT

-19.1

-21.5

2.5

14.09

1.68E-14

Latitude

NFFD

278.0

253.6

24.4

13.03

1.19E-13

49.73

MCMT

2 .0

1 .0

1 .1

12.69

2.30E-13

DD<0

84.7

131.1

-46.4

11.85

1.23E-12

TD

14.8

15.3

-0.5

1 0 .2 2

4.06E-11

AHM

13.0

1 1 .0

2 .0

1 0 .1 0

5.29E-11

MSP

249.4

284.8

-35.4

9.89

8.42E-11

MAT

8.9

8 .1

0 .8

9.67

1.42E-10

SHM

67.8

57.4

10.5

9.59

1.71E-10

DD>5

1777.6

1622.2

155.3

9.04

6.17E-10

MAP

1459.9

1657.4

-197.5

8.84

1.01E-09

PAS

89.1

152.5

-63.4

8.31

3.68E-09

MWMT

16.8

16.2

0 .6

6.91

1.37E-07

CMD

225.0

2 0 1 .6

23.4

5.74

3.26E-06

240

A ppendix D (continued).

Sample
No. and
Latitude

Sample No.

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

Eref

656.1

661.8

-5.7

2 .1 0

4.50E-02

EMT

-19.2

-2 1 . 6

2.4

1 1 .2 0

4.74E-12

11.18

4.94E-12

53

FFP

188.2

164.9

23.3

Latitude

TD

14.8

15.3

-0.5

9.94

7.54E-11

49.75

MCMT

2 .0

0.9

1 .0

9.77

1.13E-10

NFFD

277.4

253.6

23.8

9.41

2.58E-10

DD<0

85.8

132.0

-46.2

9.20

4.23E-10

MSP

252.7

289.7

-37.0

8 .1 0

6.29E-09

AHM

13.0

1 1 .0

2 .0

7.98

8.43E-09

SHM

66.9

56.4

10.5

7.92

9.79E-09

MAT

8.9

8 .1

0 .8

7.22

5.96E-08

DD>5

1773.9

1620.8

153.1

6.62

2.95E-07

MAP

1461.2

1672.9

-211.7

6 .2 1

8.98E-07

MWMT

16.8

16.2

0 .6

5.42

7.92E-06

CMD

223.2

197.0

26.1

5.37

8.95E-06

PAS

89.3

159.4

-70.1

5.12

1.84E-05

Eref

656.9

659.3

-2.4

0.74

4.63E-01

Sample No.

EMT

-18.9

-21.4

2.5

14.11

1.62E-14

54

FFP

191.0

166.4

24.6

14.01

1.93E-14

Latitude

MCMT

2 .1

1 .0

1 .1

12.26

5.43E-13

49.77

NFFD

279.6

255.3

24.3

1 2 .0 1

8.88E-13

DD<0

83.1

129.5

-46.4

11.24

4.34E-12

TD

14.8

15.3

-0.5

10.44

2.47E-11

MSP

254.1

288.0

-33.9

9.41

2.60E-10

AHM

13.1

1 1 .2

1.9

8.77

1.18E-09

MAT

9.0

8 .2

0 .8

8.77

1.19E-09

SHM

6 6 .8

56.8

1 0 .0

8.46

2.57E-09

DD>5

1785.3

1632.7

152.5

7.98

8.37E-09

MAP

1453.5

1637.7

-184.2

7.48

3.00E-08

PAS

8 6 .1

148.0

-61.9

7.11

8.11E-08

MW M T

16.8

16.3

0 .6

6.03

1.47E-06

CMD

220.3

198.2

2 2 .2

5.11

1.86E-05

Eref

654.2

660.1

-5.9

2.06

4.84E-02

Sample No.

FFP

191.1

166.8

24.3

15.25

2.18E-15

55

EMT

-19.0

-21.4

2.4

14.55

7.31E-15

Latitude

NFFD

279.5

255.5

24.1

1 2 .6 8

2.33E-13

49.79

MCMT

2 .0

1 .0

1 .0

12.45

3.69E-13

DD<0

83.9

130.0

-46.1

11.28

4.01E-12

241

A ppendix D (continued).

Sample
No. and
Latitude

Climate
Variable*

Mean in

Mean out

Mean
difference

t

P

TD

14.8

15.3

-0.5

9.65

1.47E-10

MSP

257.1

291.2

-34.2

9.15

4.76E-10

MAT

9.0

8 .2

0 .8

8.85

9.77E-10

SHM

6 6 .0

56.2

9.7

8.83

1.02E-09

AHM

13.1

1 1 .2

1.9

8 .6 6

1.57E-09

DD>5

1783.9

1634.1

149.9

7.92

9.86E-09

MAP

1454.6

1641.9

-187.2

7.33

4.53E-08

PAS

86.3

149.1

-62.8

6.85

1.61E-07

MW M T

16.8

16.3

0 .6

6 .1 1

1.17E-06

CMD

217.6

194.9

22.7

5.12

1.84E-05

Eref

653.6

659.4

-5.8

1.91

6.57E-02

Sample No.

FFP

193.4

170.6

22.7

13.79

2.87E-14

56

NFFD

281.4

259.8

2 1 .6

12.44

3.73E-13

Latitude

EMT

-18.8

-2 1 . 1

2.3

12.24

5.62E-13

49.81

MCMT

2 .1

1 .2

0.9

11.30

3.89E-12

DD<0

82.2

1 2 1 .2

-38.9

11.06

6.39E-12

TD

14.8

15.3

-0.5

9.86

9.10E-11

MAT

9.0

8.4

0.7

9.41

2.55E-10

AHM

13.1

11.5

1 .6

8.70

1.40E-09

DD>5

1794.0

1670.0

123.9

8.67

1.52E-09

PAS

84.2

130.6

-46.4

8.26

4.12E-09

MSP

260.1

286.8

-26.7

8 .0 2

7.68E-09

SHM

65.6

57.6

8 .0

7.93

9.51E-09

MAP

1451.5

1597.8

-146.2

7.62

2.10E-08

MWMT

16.9

16.4

0.4

6.51

3.96E-07

Eref

651.1

662.2

-1 1 .1

5.67

3.96E-06

14.6

4.57

8.45E-05

CMD

214.4

199.9

"Climate variables defined in Appendix B

242

Appendix E: Summary statistics (mean and standard deviation, SD, n= 298 077) for each of the five
global climate models (GCM) used for the five-model GCM ensemble projecting climate for three
future timeslices (2010-2039, 2040-2069, 2070-2099) for the Pacific Northwest (southwest British
Columbia, western Washington, and Oregon).
Climate
Variable*

HADCM3_A2
mean

SD

UKMO_HadGEMl_
A2
mean

SD

MIROC32_
MED_A2
mean

HadGEMl.AlB
SD

mean

SD

HADCM3.B2
mean

SD

20 10-2039 Future Timeslice
8 .8

3.0

8.7

3.0

8.7

2.9

9.1

2.9

8.9

3.0

MWMT(°C)

18.1

2.9

18.0

2.9

17.7

2 .6

18.1

2.5

18.4

3.0

MCMT(°C)

1 .0

3.7

0 .0

4.0

0.7

3.7

1.3

3.8

0.9

3.7

TD(°C)

17.1

2.9

17.9

3.0

17.0

2 .8

16.8

2.9

17.5

3.0

MAP(mm)

1988.9

1147.1

1971.6

1118.4

1971.7

1120.9

1910.6

1 1 0 1 .2

1963.7

1149.9

MSP(mm)

343.9

225.1

361.0

236.0

360.9

236.6

307.3

208.4

335.9

2 2 1 .8

AHM

MAT(°C)

15.7

15.5

15.6

15.5

15.6

15.3

16.6

16.6

16.2

16.1

SHM

90.0

81.4

85.1

78.0

83.4

75.1

1 0 1 .1

8 8 .6

94.1

85.7

DD<0

2 0 2 .1

248.5

232.6

259.3

2 1 0 .6

248.2

196.0

247.5

201.5

243.3

DD>5

1957.8

625.8

1945.7

625.6

1928.6

586.1

2051.8

595.0

1999.3

638.2

NFFD

249.6

62.5

249.8

62.1

248.4

60.5

257.7

60.4

249.2

61.7

FFP

173.9

56.9

176.7

56.3

168.1

54.1

188.6

55.8

174.6

55.7

PAS(mm)

321.9

467.9

328.5

462.4

320.0

449.9

288.6

426.4

311.1

460.0

EMT(°C)

-2 1 . 8

7.3

-23.4

7.1

-22.4

7.0

-21.4

7.7

-22.3

6 .8

Eref

772.7

178.7

748.9

173.1

757.3

169.9

761.6

167.4

790.9

186.0

CMD

312.7

239.3

285.6

231.4

293.1

232.5

322.3

230.9

331.1

247.6

20 40 -206 9 Future Timeslice

MAT(°C)

1 0 .0

3.0

9.9

3.1

9.4

2.9

1 1 .2

2 .8

9.8

3.0

MWMT(°C)

20.4

3.4

20.4

3.3

18.5

2 .6

2 0 .8

2.5

2 0 .0

3.4

MCMT(°C)

1 .8

3.7

0.5

4.0

1.5

3.6

3.3

3.8

1.3

3.7

TD(°C)

18.6

3.3

19.8

3.0

17.0

2 .8

17.4

3.2

18.7

3.2

MAP(mm)

2023.8

1186.9

1988.4

1126.6

1984.0

1127.3

1880.2

1081.8

1968.5

1178.1

MSP(mm)

330.4

215.6

366.9

239.3

364.0

238.8

263.1

175.7

321.6

213.4

AHM

16.7

16.5

16.5

16.4

16.1

15.8

18.7

18.5

17.2

17.1

SHM

105.4

95.6

95.3

86.9

8 6 .0

77.1

130.7

115.7

107.3

98.3

DD<0

160.9

212.7

202.9

237.9

177.0

225.6

115.1

188.6

174.9

217.7

DD>5

2310.6

705.1

2320.4

698.0

2116.9

612.3

2647.2

668.9

2251.6

696.2

NFFD

265.4

59.9

265.0

58.1

259.1

58.2

287.1

55.7

259.2

59.5

FFP

192.1

60.4

198.5

56.3

184.8

53.8

228.2

59.9

186.5

57.0

PAS(mm)

267.6

425.1

281.9

418.5

277.2

410.9

193.2

335.3

270.2

427.2

EMT(°C)

-2 0 . 0

7.0

-2 2 . 1

6.9

-2 1 . 1

6.9

-17.0

8 .1

-21.5

6.4

Eref

837.7

195.1

792.3

183.6

779.9

173.3

828.0

173.6

844.1

2 0 1 .8

CMD

365.7

261.1

310.7

247.6

302.7

238.8

394.4

230.9

378.0

267.2

243

A p p e n d ix E (c o n tin u e d ).

Climate
Variable*

HADCM3.A2
mean

SD

UKMO_ HadGEMl_
A2
mean

SD

MIROC32_
MED_A2
mean

HadGEMl.AlB
SD

mean

SD

HADCM3.B2
mean

SD

20 70-2099 Future Timeslice

MAT(°C)

1 1 .6

3.1

1 2 .8

2 .8

11.3

2 .8

1 2 .8

2 .8

10.7

3.1

MWMT(°C)

23.4

3.6

22.7

2 .6

2 0 .8

2.5

22.7

2 .6

21.5

3.8

MCMT('C)

2.5

3.8

4.8

3.8

3.3

3.5

4.8

3.8

1.7

3.7

TD(°C)

2 1 .0

3.2

17.9

3.3

17.5

2 .6

17.9

3.3

19.9

3.5

1963.5

1149.4

1973.2

1207.0
205.6

MAP(mm)

2017.0

1143.4

1963.5

1149.4

2018.1

1146.7

MSP(mm)

374.0

243.6

276.3

174.1

370.5

243.2

276.3

174.1

308.3

AHM

17.7

17.5

19.3

18.9

17.3

17.0

19.3

18.9

18.2

18.2

SHM

106.8

96.0

128.9

110.4

94.6

83.4

128.9

110.4

121.3

1 1 1 .6

DD<0

131.9

187.1

70.6

134.6

100.3

154.9

70.6

134.6

153.1

195.7

DD>5

2810.7

778.2

3112.9

748.8

2636.8

699.7

3112.9

748.8

2503.0

752.0

50.8

268.2

57.1

NFFD

286.2

54.5

309.6

50.8

286.5

53.0

309.6

FFP

225.4

60.4

254.7

65.3

215.8

60.2

254.7

65.3

197.7

57.7

PAS(mm)

197.3

332.8

118.8

245.7

171.4

296.3

118.8

245.7

235.1

394.9

8.4

EMT(°C)

-18.4

7.6

-13.7

8.4

-17.4

6.9

-13.7

-2 0 . 8

6 .0

Eref

847.5

190.2

873.6

174.7

844.6

180.8

873.6

174.7

896.5

218.5

CMD

3 3 8 .5

2 6 3 .2

4 1 2 .5

2 2 4 .0

3 3 3 .1

2 5 5 .7

4 1 2 .5

2 2 4 .0

4 2 5 .2

2 8 6 .4

•Climate variables are defined in Appendix B.

244

Appendix F: Scientific and common names for 18 species common to the Coastal Douglas-fir moist
maritime (CDFmm) subzone listed by tree, shrub, herb, and moss strata.
Common Name

Scientific Name (Genus species)
Tree Stratum
Abies grandis
Acer macrophyllum

Grand fir

Arbutus menziesii
Pseudotsuga menziesii
Thuja plicata

Arbutus
Douglas-fir

Bigleaf maple

Western redcedar

Shrub Stratum
Western flowering dogwood
Salal
Ocean spray
Western trumpet honeysuckle

Cornus nuttallii
Gaultheria shallon
Holodiscus discolour
Lonicera ciliosa
Mahonia nervosa
Rosa gymnocarpa
Symphoricarpos spp.

Dull Oregon-grape
Baldhip rose
Snowberry

Herb Stratum
Vanilla leaf
Sword fern
Bracken fern

Achlys triphylla
Polystichum munitum
Pteridium aquilinum
Moss Stratum
Hylocomium splendens

Step moss
Oregon beaked moss
Electrified cat's tail moss

Kindbergia oregana
Rhytidiadelphus triquetrus

245

Appendix G: Current and future species suitable climate space (SCS) area (km2) for the full extent
(minimum-maximum), the confidence percentiles ( lst-99th and 5th-95th), and core percentiles (25th75th) for the southwestern British Columbia study area.
Species

Envelope
Parameter

Baseline
(1961-1990)

Abies grandis

Min-Max
2st-99th

77768
17988

5th-95th

6397

2040-2069

2070-2099

78547

76689

59811

21510

20961

10181

6261

4795

1668

96

183

0

0

Min-Max
l st-99th

57020

67782

75607

80706

11903

16956

20567

22261

5th-95th

2271

7421

8005

4244

25th-75,h

0

0

0

0

Min-Max
l st-99th

46373

51903

56229

59565

3120

5081

7229

8164

5,h-95th
25,h-75th

256

1952

2890

1647

0

0

0

0

Min-Max
l s,_99th

95813

96164

96413

96146

29177

35063

36448

2 1 1 2 1

5th-95th

12357

13906

10160

3984

25th-75,h
Min-Max

631
90348

906
94131

6 6

27

96801

97190

25,h-75«h

Acer
macrophyllum

Arbutus menziesii

Pseudotsuga
menziesii

Thuja plicata

Cornus nuttallii

Gaultheria shallon

1^-99^

59056

64145

59510

34883

5th-95th

32406

33052

24821

9452

25th-75,h
Min-Max

2641

1444

185

3

56276

64611

69186

l st-991^

45805
11079

15406

15751

11257

5th-95th

1 1

215

745

8 6 6

25,h-75,h
Min-Max
l st.99th

0

0

0

0

94178

96133

96951

90568

42163

46115

43993

24022

20869

22482

14818

4833

2 1

2 0

0

0

Min-Max
l 5t-99th

61162

67975

75351

77227

15047

19695

22860

21355

5th-95th

4544

6 8 8 6

5936

2213

0

0

0

0

Min-Max

44109

48754

52534

49074

5th-95th
25th 7 5 th
Holodiscus discolor

25th 75.h
Lonicera ciliosa

M ahonia nervosa

Rosa gymnocarpa

2010-2039

^st_ggth

15818

18550

17398

12075

5th-95th

6131

6300

4433

1316

25,h-75,h
Min-Max
l st-99th

119
83477

2 0 1

0

0

87697

90331

90596

18447

22431

8224

9244

21816
6064

9657

5th-95th
25th-75,h
Min-Max

0

0

0

0

90603

92044

93196

93117

l st_99th

20476

25764

28464

22338

5th-95th

8124

9554

7394

3329

0

41

0

0

25^ 75*

246

351

A p p e n d ix G (c o n tin u e d ).

Species

Envelope
Parameter

Baseline
(1961-1990)

2010-2039

2040-2069

2070-2099

Symphoricarpos

Min-Max
l*'_99th

84313

87333

90158

93186

22187

27276

28670

21830

5,h-95th

4024

3398

1842

1366

albus

Achlys triphylla

Polystichum
munitum

25th-75,h

0

0

0

0

Min-Max

91274

93218

93824

84631

l st-99th
5th-95th

25598

30310

26158

10577

10436

9229

6141

560

25th-75th

5

1

0

0

Min-Max
l st-99th

96548

97103

97343

97321

41132

47291

47077

26894

15269

17595

13711

3549

5lh-95th
Pteridium
aquilinum

Hylocomium
splendens

Kindbergia
oregana

Rhytidiadelphus
triquetrus

OVERLAP AREA

25th-75th

5

0

0

0

Min-Max
l st_ggth

96505

96745

96744

96330

42570

48936

54344

53195

5,h-95,h

18720

22522

18992

7961

25,h-75*h

548

1525

29

0

Min-Max
l s,-99th

97077

97343

93011

73879

61091

53542

32372

11913

5th-95th

29206

20834

8775

1379

25th-75th
Min-Max
l s,-99th

15
85183
54206

1

0

0

90686
51510

93939
33295

82984
10030
8 8 8

5,h-95th
th 7 5 th

29260
2319

22511
1716

9423

2 5

215

0

Min-Max
l st-99th

89450

92614

91512

75223

36701

34087

22565

8739

5,h-95th

10918

7206

3653

1597

25th-75th

0

0

0

0

Min-Max
l st-99th

37601
1902

42990
557

43156

31374

2

0

5th-95,h

0

0

0

0

25,h-75th

0

0

0

0

247

Appendix H: Seasonal (Fall, September 15 - December 14; Winter, December 15 - March 14; Spring,
March 15 - June 14; and Summer, June 15 - September 14) microclimate summary for the Coastal
Douglas-fir moist maritime (CDFmm) and Coastal Western Flemlock very dry maritime (CWHxm)
mesic study sites on the south, central, and north transects (n=3) for the baseline year (September
15, 2010 - September 14, 2011) with the mean, standard error (S.E.), minimum (Min.) and maximum
(Max.) values and the results of the paired t-test (df=2, p<0.05).
CDFmm

Microclimate Variable
Mean

CWHxm

Paired t-test

S.E.

Min.

Max.

Mean

S.E.

Min.

Max.

t

P

Fall (September IS - December 14, 2010)

Soil Moisture (%)

9.36

1.84

5.71

11.63

11.97

0.50

10.98

12.47

1.928

0.194

Soil Temperature (°C)

9.37

0.29

8.91

9.90

8.89

0.27

8.46

9.40

0.985

0.428

Air Temperature (1.5 m, "C)

6.26

0.74

5.11

7.64

6.79

0.17

6.58

7.12

0.696

0.559

Air Temperature (0.3 m, °C)

6.58

0.97

4.66

7.79

7.37

0.67

6.52

8.69

0.484

0.676

Total Precipitation (mm)

186.4

21.5

159.6

229.0

300.2

84.6

148.8

441.4

1.081

0.393

Wind Speed (m/s)

0.17

0.02

0.13

0.20

0.13

0.07

0.00

0.22

0.922

0.454

Relative Humidity (%)

94.3

1.3

91.7

95.7

97.4

0.7

94.0

99.5

6.146

0.025

Solar Radiation (W /m 2)

4.3

0.4

3.6

5.0

4.7

1.0

3.6

6.7

0.425

0.713

Winter (December IS - March 14, 2010-2011)

Soil Moisture (%)

10.97

1.86

7.26

12.93

12.38

0.80

10.88

13.60

1.222

0.346

Soil Temperature (*C)

4.67

0.08

4.53

4.82

3.33

0.16

3.14

3.65

15.48

0.004

Air Temperature (1.5 m, *C)

3.18

0.16

2.97

3.37

1.95

0.18

1.62

2.22

4.928

0.039

3.53

2.47

0.83

1.44

4.11

0.220

0.846

467.5

519.5

8.5

502.6

529.6

2.656

0.120

Air Temperature (0.3 m, °C)

2.79

0.62

1.55

Total Precipitation (mm)

347.3

66.6

237.4

Wind Speed (m/s)

0.29

0.08

0.15

0.42

0.07

0.05

0.00

0.16

5.953

0.027

Relative Humidity (%)

96.2

1.9

92.3

96.2

98.3

0.72

97.1

99.6

2.888

0.107

Solar Radiation (W /m 2)

4.1

0.1

3.8

4.3

4.1

0.8

3.0

5.6

0.093

0.934

Soil Moisture (%)

9.16

0.91

7.35

10.19

11.20

0.82

10.17

12.83

2.946

0.098

Soil Temperature ("C)

8.00

0.24

7.61

8.43

6.53

0.53

5.71

7.52

1.931

0.193

Air Temperature (1.5 m, °C)

8.42

0.03

8.35

8.64

7.53

0.34

6.98

8.15

2.399

0.139

Air Temperature (0.3 m, °C)

8.99

0.94

7.36

10.61

7.45

0.09

7.31

7.62

1.695

0.232

Total Precipitation (mm)

103.6

12.3

82.4

124.8

152.1

36.1

81.4

200.2

1.721

0.227

Wind Speed (m/s)

0.30

0.10

0.12

0.45

0.04

0.03

0.00

0.09

3.489

0.073

Relative Humidity {%)

81.7

2.7

76.3

85.1

92.0

1.7

88.7

94.1

10.554

0.009

Solar Radiation (W /m 2)

19.4

1.4

17.8

22.3

20.3

4.6

13.8

29.2

0.258

0.820

4.83

1.36

2.25

6.85

7.56

0.60

6.82

8.75

2.541

0.126

Spring (March 1 5 -Ju n e 14, 2011)

Summer (June 15 - September 14, 2011)

Soil Moisture (%)
Soil Temperature (°C)

13.51

0.23

13.08

13.87

13.25

0.24

12.81

13.65

0.559

0.633

Air Temperature (1.5 m, °C)

15.72

0.49

14.86

16.56

15.36

0.11

15.21

15.58

0.887

0.469

Air Temperature (0.3 m, °C)

15.38

0.91

13.74

16.88

14.70

0.98

12.96

16.36

8.48

0.014

Total Precipitation (mm)

16.3

10.8

0.0

36.8

30.3

26.6

33.2

1.410

0.294

Wind Speed (m/s)

0.25

0.05

0.18

0.35

0.01

2.0
0.01

0.00

0.02

5.020

0.038

Relative Humidity (%)

78.7

3.7

74.7

86.1

81.9

1.9

78.7

85.4

1.482

0.277

Solar Radiation (W /m 2)

20.8

1.9

18.3

24.5

19.5

4.1

13.1

27.0

0.486

0.675

248

Appendix I: Monthly microclimate summary for the Coastal Douglas-fir moist maritime (CDFmm) and
Coastal Western Hemlock very dry maritime (CWHxm) mesic study sites on the south, central, and
north transects (n=3) for the baseline year (September 15, 2010 - September 14, 2011) with the
mean, standard error (S.E.), minimum (Min.) and maximum (Max.) values and the results of the
paired t-test (df=2, p<0.05).
CWHxm

CDFmm

Month/

Paired t-test

Microclimate
Variable

Mean

S.E.

Min.

Max.

Mean

S.E.

Min.

Max.

t

P

6.24

1.30

4.93

8.83

8.13

0.54

7.05

8.69

1.663

0.238

13.69

0.37

12.94

14.09

12.92

0.16

12.72

13.23

1.451

0.284

16.18

1.04

14.67

18.19

14.41

0.19

14.12

14.76

1.976

0.187

15.00

0.52

14.28

16.01

13.85

0.34

13.16

14.20

3.072

0.092

22.7

13.1

0.0

45.4

60.7

3.4

54.6

66.4

3.918

0.059

0.11

0.05

0.01

0.19

0.07

0.04

0.00

0.12

0.605

0.607

82.2

1.63

80.0

85.4

87.9

2.9

82.5

92.5

2.054

0.176

9.8*

3.5

6.3

13.3

9.3

2.6

5.8

14.4

0.453

0.729

7.74

1.53

4.71

9.66

11.28

0.53

10.22

11.86

3.468

0.074

11.13

0.23

10.71

11.51

10.56

0.13

10.36

10.79

1.278

0.226

9.07

0.20

8.68

9.29

9.00

0.22

8.65

9.41

0.258

0.821

9.24

0.68

7.94

10.25

8.97

0,11

8.81

9.17

0.337

0.768

55.5

18.4

19.3

79.0

52.9

28.0

2.9

98.8

0.070

0.950

0.16

0.02

0.14

0.19

0.13

0.07

0.00

0.25

0.407

0.724

90.6

3.8

83.0

95.3

95.5

2.9

89.7

99.1

3.925

0.059

5.7

0.8

4.1

6.6

5.5

1.2

4.1

7.9

0.205

0.856

10.19

1.90

6.51

12.87

12.55

0.44

11.67

13.09

1.619

0.247

7.52

0.25

7.19

8.01

7.03

0.19

6.69

7.34

1.861

0.204

3.70

0.28

3.17

4.14

3.00

0.15

2.73

3.25

3.344

0.079

September (2010-2011)

Soil Moisture (%)
Soil Temperature

CO
Air Temperature
(1.5 m, *C)
Air Temperature
(0.3 m, *C)
Total Precipitation
(mm)
Wind Speed (m/s)
Relative Humidity
(%)
Solar Radiation
(W /m 2)
October (2010)

Soil Moisture (%)
Soil Temperature

ra
Air Temperature
(1.5 m, °C)
Air Temperature
(0.3 m, °C)
Total Precipitation
(mm)
Wind Speed (m/s)
Relative Humidity
(%)
Solar Radiation
(W /m 2)
November (2010)

Soil Moisture (%)
Soil Temperature
CC)
Air Temperature
(1.5 m, "C)

249

Appendix I (continued).
CWHxm

CDFmm

Month/

Paired t-test

Microclimate
Mean

S.E.

Min.

Max.

Mean

S.E.

Min.

Max.

t

P

3.31

0.56

2.41

4.33

3.69

0.60

2.94

4.87

0.359

0.754

62.8

26.1

28.5

114.1

93.3

22.4

56.8

134.2

0.931

0.450

0.17

0.02

0.14

0.20

0.13

0.06

0.00

0.21

0.703

0.555

96.3

0.49

95.5

97.2

98.4

0.99

96.6

99.89

3.978

0.058

2.7

0.4

2.1

3.3

2.7

0.5

2.1

3.8

0.102

0.928

11.26

2.11

7.05

13.42

12.98

0.52

11.97

13.74

1.066

0.398

5.73

0.11

5.51

5.86

4.19

0.19

3.98

4.57

9.273

0.011

4.00

0.13

3.76

4.19

2.56

0.23

2.24

3.00

4.234

0.052

3.47

0.59

2.29

4.11

3.28

0.97

2.03

5.20

0.119

0.916

105.1

27.9

49.2

133.3

219.1

23.8

177.0

259.4

3.044

0.093

0.30

0.09

0.15

0.46

0.09

0.06

0.00

0.19

5.984

0.027

95.3*

2.4

92.9

97.6

98.7

0.7

97.3

99.8

2.843

0.215

2.3

0.1

2.0

2.4

2.2

0.3

1.8

2.9

0.134

0.906

11.02

1.74

7.54

13.0

12.3

0.8

11.0

13.7

1.161

0.366

4.57

0.06

4.45

4.66

3.12

0.18

2.85

3.45

9.586

0.011

3.54

0.12

3.32

3.74

2.19

0.17

1.85

2.44

5.672

0.030

2.94

0.56

1.83

3.56

2.53

0.72

1.52

3.91

0.328

0.774

134.3

19.0

97.2

160.2

149.2

24.2

102.8

184.4

0.482

0.677

0.23

0.05

0.15

0.32

0.06

0.04

0.01

0.15

8.652

0.013

96.0*

1.2

94.9

97.2

99.0

0.5

98.1

99.8

9.275

0.084

3.0

0.1

2.9

3.1

3.1

0.6

2.2

4.3

0.272

0.811

Soil Moisture (%)

10.66

1.72

7.24

12.77

11.86

0.92

10.11

13.21

1.227

0.345

Soil Temperature

4.44

0.12

4.31

4.67

3.28

0.23

2.89

3.68

8.242

0.014

Variable
Air Temperature
(0.3 m, -C)
Total Precipitation
(mm)
Wind Speed (m/s)
Relative Humidity
(%)
Solar Radiation
(W /m 2)
December (2010)

Soil Moisture (%)

Soil Temperature
CC)
Air Temperature
(1.5 m, °C)
Air Temperature
(0.3 m, °C)
Total Precipitation
(mm)

Wind Speed (m/s)
Relative Humidity
(%)
Solar Radiation
(W /m 2)
January (2011)

Soil Moisture (%)

Soil Temperature

CC)
Air Temperature
(1.5 m, *C)
Air Temperature
(0.3 m, X )
Total Precipitation
(mm)

Wind Speed (m/s)
Relative Humidity
(%)
Solar Radiation
(W /m 2)
February (2011)

250

Appendix I (continued).
CWHxm

CDFmm

Month/

Paired t-test

Microclimate
Variable

Mean

S.E.

Min.

Max.

Mean

S.E.

Min.

Max.

t

P

2.15

0.11

1.96

2.32

1.20

0.17

0.86

1.41

3.558

0.071

2.32

0.96

0.61

3.92

2.04

1.06

0.85

4.15

0.141

0.901

87.3

19.5

52.6

120.0

113.9

7.1

100.0

123.6

2.097

0.171

0.30

0.09

0.16

0.46

0.06

0.04

0.00

0.15

S.026

0.037

91.1*

3.4

87.7

94.6

96.8

1.2

95.4

99.2

4.083

0.153

5.0

0.3

4.5

5.5

5.1

1.0

3.8

7.0

0.133

0.906

13.09*

0.17

12.92

13.27

12.31

0.92

10.54

13.64

0.164

0.896

5.06

0.07

4.93

5.17

3.88

0.24

3.46

4.30

4.098

0.055

5.24

0.21

4.99

5.66

4.13

0.26

3.61

4.47

4.862

0.040

4.52*

0.74

3.78

5.25

4.00

0.28

3.57

4.52

0.394

0.761

93.2

22.1

66.6

137.0

176.7

9.8

162.8

195.6

2.865

0.103

0.30

0.10

0.11

0.40

0.09

0.07

0.00

0.23

2.892

0.102

94.3*

0.02

94.2

94.3

98.5

0.6

97.6

99.6

4.294

0.146

9.3

0.3

8.9

9.9

9.4

1.5

7.1

12.2

0.087

0.939

8.55

1.65

5.24

10.29

11.39

0.89

10.07

13.09

2.412

0.137

6.37

0.13

6.12

6.54

5.15

0.25

4.74

5.60

3.380

0.078

5.75

0.05

5.66

5.84

4.83

0.38

4.10

5.39

2.103

0.170

6.19

0.88

4.61

7.64

4.96

0.50

4.35

5.94

0.934

0.449

35.5

12.9

18.4

60.8

57.1

16.4

24.4

76.6

1.957

0.189

0.33

0.11

0.12

0.49

0.06

0.04

0.00

0.13

3.506

0.073

85.3*

1.8

83.4

87.1

92.8

2.4

88.7

97.1

20.399

0.031

17.0

1.4

14.4

19.0

16.6

2.8

13.3

22.1

0.182

0.873

(*C)
Air Temperature
(1.5 m, “0
Air Temperature
(0.3 m, *C)
Total Precipitation
(mm)

Wind Speed (m/s)
Relative Humidity
{%)

Solar Radiation
(W /m 2)
M arch (2011)

Soil Moisture (%)
Soil Temperature
CC)

Air Temperature
(1.5 m, *C)
Air Temperature
(0.3 m, * 0
Total Precipitation
(mm)
Wind Speed (m/s)
Relative Humidity
(%)
Solar Radiation
(W /m 2)
April (2011)

Soil Moisture (%)
Soil Temperature

CC)
Air Temperature
(1.5 m, ° 0
Air Temperature
(0.3 m, ”0
Total Precipitation
(mm)
Wind Speed (m/s)

Relative Humidity
(%)
Solar Radiation
(W /m 2)

251

Appendix I (continued).
CWHxm

CDFmm

Month/

Paired t-test

Microclimate
Variable

Mean

S.E.

Min.

Max.

Mean

S.E.

Min.

Max.

t

P

8.63

1.67

6.40

10.27

11.07

0.84

10.19

12.74

3.024

0.094

8.72

0.13

8.57

8.98

7.77

0.52

7.15

8.81

1.622

0.246

9.94

0.07

9.80

10.04

9.06

0.45

8.59

9.96

1.990

0.185

9.81

0.47

9.05

10.65

9.01

0.41

8.27

9.70

8.782

0.013

38.4

7.1

27.7

51.8

45.3

14.6

20.4

70.8

0.898

0.464

0.29

0.10

0.10

0.45

0.01

0.01

0.00

0.03

2.790

0.108

78.5

3.3

71.9

82.0

90.9

2.5

86.1

94.7

9.955

0.010

24.6

2.7

20.7

29.7

23.8

5.3

14.9

33.3

0.263

0.817

7.15

0.99

5.96

9.11

10.05

0.93

8.32

11.49

2.187

0.160

0.57

10.78

11.93

0.354

0.784

M a y (2011)

Soil Moisture (%)
Soil Temperature
(°C)
Air Temperature
(1.5 m, " 0

Air Temperature
(0.3 m, *C)
Total Precipitation
(mm)
Wind Speed (m/s)

Relative Humidity
(%)
Solar Radiation
(W /m 2)
June (2011)

Soil Moisture (%)

11.61

0.13

11.42

11.85

11.35
*

13.82

0.06

13.71

13.92

13.07

0.03

12.75

13.68

2.098

0.171

13.53

0.60

12.77

14.71

12.85

1.05

11.00

14.62

1.247

0.339

6.3

3.1

0.0

9.4

6.0

0.8

4.4

7.2

0.105

0.926

0.31

0.07

0.22

0.45

0.02

0.02

0.00

0.05

3.557

0.071

79.1

2.5

75.6

83.9

85.5

0.3

84.8

86.0

2.882

0.149

23.4

3.2

19.6

29.8

26.4

9.1

15.1

44.4

0.484

0.676

5.24

1.26

2.89

7.22

9.08

0.45

8.20

9.66

2.605

0.121

13.20

0.14

13.02

13.47

13.03

0.22

12.66

13.43

0.470

0.685

15.66

0.25

15.22

16.10

15.04

0.05

14.94

15.11

2.283

0.150

15.33

0.70

14.05

16.46

14.71

1.17

12.67

16.72

1.294

0.325

12.2

9.7

0.0

31.4

12.3

4.7

8.0

22.8

0.703

0.555

Wind Speed (m/s)

0.28

0.05

0.20

0.36

0.01

0.00

0.01

6.132

0.026

Relative Humidity

78.4

3.8

72.4

85.6

83.7

0.00
1.6

81.5

86.9

2.015

0.182

Soil Temperature
(°C)
Air Temperature
(1.5 m, *C)
Air Temperature
(0.3 m, °C)
Total Precipitation
(mm)
Wind Speed (m/s)
Relative Humidity
(%)
Solar Radiation
(W /m 2)
July (2011)

Soil Moisture (%)
Soil Temperature
(°C)
Air Temperature
(1.5 m, “0
Air Temperature
(0.3 m, ”C)
Total Precipitation
(mm)

252

Appendix I (continued).
CDFmm

Month/

CWHxm

Paired t-test

Microclimate
Variable

Mean

S.E.

Min.

Max.

Mean

S.E.

Min.

Max.

t

P

24.6

6.1

18.0

36.8

20.9

5.5

13.2

31.6

1.986

0.185

3.44

1.42

1.18

6.07

6.27

0.96

4.83

8.08

6.006

0.027

14.43

0.20

14.04

14.71

13.78

0.34

13.36

14.44

1.237

0.342

16.75

0.41

16.10

17.52

16.29

0.16

16.10

16.61

1.765

0.220

16.44

0.79

15.01

17.74

15.30

0.90

13.77

16.88

7.919

0.016

0.4

0.4

0.0

1.2

10.7

3.7

4.2

17.0

2.789

0.108

0.23

0.04

0.18

0.31

0.01

0.01

0.00

0.03

6.080

0.026

78.0

3.2

74.7

84.4

80.9

1.6

78.3

83.7

1.527

0.266

22.7

2.1

18.5

25.0

20.8

2.9

15.2

24.8

1.954

0.190

(%)
Solar Radiation
(W /m 2)
August (2011)

Soil Moisture (%)
Soil Temperature
CC)
Air Temperature
(1.5 m, °C)

Air Temperature
(0.3 m, *C)
Total Precipitation
(mm)

Wind Speed (m/s)
Relative Humidity
(%)
Solar Radiation
(W /m 2)

* missing data for one site, n=2

253