Regional, ecological, and temporal patterns in Douglas-fir climate-growth relationships
in the British Columbia Interior

Hardy Griesbauer
B.Sc. Augustana University College, 1995

Thesis Submitted In Partial Fulfillment Of
The Requirements For The Degree Of
Master of Science
In
Natural Resources and Environmental Studies
(Forestry)

The University of Northern British Columbia
October 2008

© Hardy Griesbauer, 2008

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 K1A0N4
Canada

Your file Votre reference
ISBN: 978-0-494-48778-5
Our file Notre reference
ISBN: 978-0-494-48778-b

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

AVIS:
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 Plntemet, prefer,
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.

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

While 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
How climate change will affect tree growth across species' ranges remains a critical
knowledge gap. Tree-ring data were analyzed from 33 Douglas-fir stands spanning a wide
climatic and geographic range in the interior of British Columbia to identify regional and
ecological patterns between climate and growth. Populations growing in warm and dry
climates had growth patterns correlated mostly to local annual precipitation, whereas
populations growing in high-elevation wet and cold climates were more correlated to
annual/winter snowfall/winter temperatures and quasi-periodic ocean-atmosphere climate
systems. Populations growing near the climatic margins of the species' range had the
strongest responses to climate variability. Examining these relationships over different
temporal scales revealed that some climate-growth relationships have varied substantially
over time. Our results indicate that in the interior of British Columbia, Douglas-fir growth
responses to climate change will not be uniform over time and space.

11

TABLE OF CONTENTS
ABSTRACT
TABLE OF CONTENTS
LIST OF TABLES
LIST OF FIGURES
ACKNOWLEDGEMENTS

ii
iii
vi
vii
ix

Chapter 1. Introduction
1.1 Climate change in western North America and British Columbia
1.2 Climate change and forests
1.2.1 The effects of climate change on forest growth
1.3 Douglas-fir
1.3.1 Douglas-fir climate-growth relationships
1.3.1.1 Douglas-fir growth responses to precipitation
1.3.1.2 Douglas-fir growth responses to temperature
1.4 Project objectives
1.5 Literature cited

10
14
15
17
19
19
21
22
24

in

Chapter 2. Regional and ecological patterns in Douglas-fir climate-growth
relationships in the British Columbia Interior
Abstract
2.1 Introduction
2.2 Materials and methods
2.2.1 Sampling
2.2.2 Processing and chronology development
2.2.3 Data analysis
2.2.3.1 Chronology analysis
2.2.3.2 Climate-growth relationships
2.2.3.2.1 Climate data
2.2.3.2.2 Regional climate-growth relationships
2.2.3.2.3 Climatic gradient analysis
2.3 Results
2.3.1 Chronologies
2.3.2 Multivariate analysis
2.3.2.1 Northern Interior sites
2.3.2.2 Southern Interior sites
2.3.2.3 Chilcotin Plateau sites
2.3.3 Climate-growth relationships
2.3.3.1 Regional climate-growth relationships
2.3.3.1.1 Northern Interior sites
2.3.3.1.2 Southern Interior sites
2.3.3.1.3 Chilcotin Plateau sites
2.3.3.2 Climatic gradient analysis
2.4 Discussion
2.4.1 Ecological growth patterns
2.4.2 Climate-growth relationships
2.4.2.1 Low- to mid-elevation dry sites
2.4.2.2 High-elevation sites
2.4.2.3 Sites at the northern geographic limits
2.4.3 Conclusion
2.5 Literature cited
Appendix A. Sample site mean annual temperature and precipitation
normals.
Appendix B. Regional growth chronology time series derived from
residual growth chronologies.
Appendix C. Example of climate gradient analysis.

32
33
36
36
37
38
38
40
40
44
44
45
45
46
46
47
48
49
49
49
50
51
51
52
52
53
53
57
59
60
62
87
88
89

IV

Chapter 3. Temporal variability in climate-growth relationships at the climatic
margins of Douglas-fir in British Columbia
Abstract
3.1 Introduction
3.2 Materials and methods
3.2.1 Field sampling
3.2.2 Processing and chronology development
3.2.3 Principal components analysis
3.2.4 Climate data
3.2.5 Analysis of climate-growth relationships and interregional growth patterns
3.3 Results
3.3.1 Temporal variation in high-elevation growth and
climate-growth relationships
3.3.1.1 Inter-regional growth patterns
3.3.1.2 Temperature-growth relationships
3.3.1.3 Snowfall-growth relationships
3.3.1.4 Ocean-atmosphere climate indices-growth
relationships
3.3.2 Temporal variation in dry-site inter-regional growth
patterns and climate-growth relationships
3.3.2.1 Inter-regional growth patterns
3.3.2.2 Temperature-growth relationships
3.3.2.3 Precipitation-growth relationships
3.4 Discussion
3.4.1 High-elevation sites
3.4.2 Dry sites
3.4.3 Conclusions
3.5 Literature cited
Appendix A. Time series of October-March precipitation anomalies for
Vernon, British Columbia, from 1904 to 2004.
Appendix B. Time series of annual (prior July to current June)
precipitation anomalies for Vernon, British Columbia, from 1904 to
2004.
Appendix C. Time series of annual (prior July to current June)
precipitation anomalies for Williams Lake, BC, from 1940 to 2005.

131

Chapter 4. Summary and management implications
4.1 Introduction
4.2 Key study results and management implications
4.2.1 Mid- to low-elevation dry-site Douglas-fir populations
4.2.2 High-elevation Douglas-fir populations
4.3 Research limitations and future directions
4.4 Conclusions
4.5 Literature Cited

132
132
133
137
139
140
142

90
91
93
93
94
96
97
99
102

102
102
103
103

104
104
105
106
107
111
114
116
130

130

v

LIST OF TABLES
Table 2.1. Site descriptions and chronology statistics.

70

Table 2.2. Sample site climate normals.

71

Table 2.3. Summary of meteorological records used in Chapter 2.

72

Table 2.4. Missing precipitation data in meteorological records (between 1940
and 2004).

72

Table 2.5. Monthly climate values used to derive annual and seasonal
variables used for correlation analysis.

73

Table 2.6. Principal components used for climate-growth analysis.

73

Table 2.7. Principal components analysis summary.
Table 2.8. Correlation coefficients between Douglas-fir growth variation and
annual/seasonal climatic variables.

74
75

Table 3.1. Meteorological records used in Chapter 3.

123

Table 3.2. Missing rain, snow, and precipitation data in meteorological
records.

124

Table 3.3. Monthly variables used to create seasonal variables for climategrowth analyses.

124

Table 3.4. Correlation coefficients between high-elevation Douglas-fir growth
and ocean-atmosphere climate system indices.

124

VI

LIST OF FIGURES
Figure 2.1. Sampling region overview map.

76

Figure 2.2. Sampling site locations in Northern Interior study region.

77

Figure 2.3. Sampling site locations in Chilcotin Plateau study region.

78

Figure 2.4. Sampling site locations in Southern Interior study region.

79

Figure 2.5. Chronology statistics against sample site mean annual climate normals.

80

Figure 2.6. Variation in annual precipitation-growth correlation and regression
coefficients over annual heat-moisture index normals.

81

Figure 2.7. Variation in annual precipitation-growth correlation and regression
coefficients over summer heat-moisture index normals.

81

Figure 2.8. Variation in annual precipitation-growth correlation and regression
coefficients over mean annual precipitation normals.

82

Figure 2.9. Variation in annual precipitation-growth correlation and regression
coefficients over (log-transformed) mean annual precipitation normals.

82

Figure 2.10. Variation in winter PNA index-growth correlation and regression
coefficients over annual heat-moisture index.

83

Figure 2.11. Variation in winter PDO index-growth correlation and regression
coefficients over annual heat-moisture index.

83

Figure 2.12. Northern Interior PC loads against geographic variables and climatic
normals.

84

Figure 2.13. Southern Interior PC loads against geographic variables and climate
normals.

85

Figure 2.14. Chilcotin Plateau PC loads against geographic variables and climate
normals.

86

vn

Figure 3.1. Correlations between climate and growth for high-elevation Douglas-fir
populations in two regions of British Columbia over two sub-periods.

125

Figure 3.2. Moving correlation functions between climate and growth for highelevation Douglas-fir populations in two regions of British Columbia.

126

Figure 3.3. Correlations between climate and growth for Douglas-fir populations
growing in dry climates in two regions of British Columbia over two sub-periods.

127

Figure 3.4. Moving correlation functions between climate and growth for Douglas-fir
populations growing in dry climates in two regions of British Columbia.

128

Figure 3.5. Moving correlation functions between temperatures and growth for
Douglas-fir populations growing in dry climates in two regions of British Columbia.

129

Figure 4.1. Projected future climate for Douglas-fir sample sites in the three study
regions.

145

Figure 4.2. Growth-annual precipitation regression coefficients against site annual
heat-moisture index.

146

Figure 4.3. Growth-annual precipitation regression coefficients against site mean
annual precipitation.

146

vm

ACKNOWLEDGEMENTS
I would like to thank Dr. Scott Green for his mentorship, support, and guidance throughout
this project, Dr. Greg O'Neill and Dr. Stephen Dery for guidance, advice, and suggestions as
members of my committee, Emily Muller, Kara Przeczek, Kasia Caputa, and Yumiko
Miyamoto for field and lab assistance, Chris Steeves for GIS data and maps, and of course,
my wife Gretchen Prystawik for everything!

This projected was funded by the Forest Science Program (British Columbia), Natural
Sciences and Engineering Research Council of Canada, and the province of British
Columbia.

IX

Chapter 1. Introduction
1.1 Climate change in western North America and British Columbia
Anthropogenic factors have contributed to an increase in global temperatures of
approximately 0.6 °C over the last century, with warming occurring from 1910 to 1940 and
from 1975 to the present day (Meehl et al. 2005, Christensen et al. 2007, Jones 1988). The
rate of temperature increase has been faster in British Columbia (BC), where annual
temperatures have increased by 0.5 to 1.7°C over the past century (the strongest increases
have been in the northern and interior regions of the province) (BC MWLAP 2002).
Seasonally, spring temperatures in BC show the strongest and most consistent trend over the
past century, although winter and summer temperatures in some regions also show significant
upward trends, especially over the last 50 years (BC MWLAP 2002, BC MoE 2007).
Nighttime minimum temperatures have been increasing more rapidly than daytime maximum
temperatures, thus lowering the diurnal temperature range, reducing growing season frost
events, and lengthening the growing season (Karl et al. 1993, Dai et al. 1999, BC MWLAP
2002, BC MoE 2007). Annual precipitation has also increased significantly over most of
western North America (Zhang et al. 2007, Trenberth 1990), although some regions of BC
have become drier over the past 50 years (BC MoE 2007). Generally, winters in BC have
become drier and the spring and summer seasons have become wetter (BC MoE 2007). The
frequency of droughts, severe storms, and other extreme climatic events may also be
increasing in association with climate change (Easterling et al. 2000).

Temperatures are projected to increase at possibly unprecedented rates during the 21st
century, although future scenarios are uncertain due to a link with future changes in

10

greenhouse gas emissions (Christensen et al. 2007, Moberg et al. 2005). Climate change
models project that BC temperatures could increase by 2-5 °C above 1961-1990 averages by
the middle of the 21st century, with the fastest rate of increase occurring over the winter,
especially in the interior and northern regions of the province (BC MoE 2007, BC MWLAP
2007).

Precipitation trends are also difficult to predict, and model projections vary (Christensen et
al. 2007). Annual precipitation over North America is generally expected to increase,
although some areas may experience net drying due to increased evaporation and
transpiration associated with higher temperatures (Christensen et al. 2007). Although winter
precipitation in BC is expected to increase, snowpacks may actually diminish as a result of
shorter winters, increased winter rains, and earlier spring snowmelt, particularly at low- to
mid-elevations (Christensen et al. 2007, Knowles et al. 2006, Leung et al. 2004, BC MoE
2007, BC MWLAP 2002). Summer precipitation is expected to decrease over most of
western North America (Christensen et al. 2007), although some projections for BC show a
modest increase over most of the province with a possible slight decrease in southern coastal
areas by the mid-21st century (BC MoE 2007).

Embedded in the temperature and precipitation trends over the past century are annual and
decadal patterns of climatic variation that reflect the influence of spatially large quasiperiodic ocean-atmosphere systems on local climate (Cayan et al. 1998, Mantua and Hare
2002, BC MoE 2007). In western North America, the El Nino Southern Oscillation (ENSO),
Pacific Decadal Oscillation (PDO), and Pacific/North American teleconnection index (PNA)
are important drivers of local seasonal temperature and precipitation patterns (Mantua et al.
11

1997, Hare and Mantua 2001, Cayan et al. 1998, Knowles et al. 2006, Leathers et al. 1991,
BC WLAP 2002, BC MoE 2007) and therefore may influence ecosystem functioning (e.g.,
Mantua et al. 1997, Case and Peterson 2005, Fagre et al. 2003, Daniels and Veblen 2004).

ENSO represents a quasi-periodic oscillation in sea surface temperature and atmospheric
pressure anomalies in the tropical-subtropical Pacific and Indian Oceans (Allan 2000) that
generally alternates between two extreme phases (often termed El Nino and La Nina events)
every 2 to 7 years, with individual events lasting anywhere from 8 to 24 months (Allan 2000,
Hare and Mantua 2001). ENSO affects climate variability on an almost global scale (Allan
2000) and is the strongest climate signal over most of the Pacific Ocean after the seasonal
cycle (Hare and Mantua 2001). El Nino (La Nina) events typically bring warmer (cooler)
and drier (wetter) winters to BC (BC MoE 2007, BC MWLAP 2002). Smith and
Sardeshmukh (2000) created a monthly ENSO index time series using an average of
standardized monthly sea surface temperature and air pressure anomalies
(http://www.cdc.noaa.gov/people/cathy.smith/best/enso.ts.lmn.txt).

PDO is a quasi-periodic system of decadal variability in sea surface temperatures, sea level
pressure, and wind patterns over the Pacific Ocean (Mantua et al. 1997, Mantua and Hare
2002, Biondi et al. 2001), with "warm" and "cool" phases lasting approximately 15 to 30
years (Mantua and Hare 2002, Gedalov and Mantua 2002), although PDO reconstructions
based on tree-rings suggest that phase reversals may have occurred at different frequencies
prior to the instrumental record (Gedalov and Smith 2001, Gedalov and Mantua 2002, Biondi
et al. 2001). PDO may also have a 50-70 year cycle, although this has not been extensively
studied to date (Mantua and Hare 2002, Hare and Mantua 2001). Over the past century, PDO
12

warm phases have occurred from approximately 1925 to 1946, and from 1977 to present day
(Mantua et al. 1997, Mantua and Hare 2002, Gedalov and Mantua 2002, Hare and Mantua
2001, Biondi et al. 2001). Cool phases have occurred from approximately 1900 to 1925, and
from 1947 to 1976 (Mantua et al. 1997, Mantua and Hare 2002, Gedalov and Mantua 2002,
Hare and Mantua 2001, Biondi et al. 2001). In western North America, warm (cool) PDO
phases tend to result in warm and dry (cool and wet) anomalous climatic conditions (Hare
and Mantua 2001, BC MoE 2007, BC MWLAP 2002). PDO and ENSO can have strong
combined impacts on climate variability in western North America (Gershunov et al. 2001,
Biondi et al. 2001). A time series of monthly PDO indices, derived from the leading
principal component of sea surface temperature anomalies in the North Pacific Ocean, is
available online at: http://jisao.washington.edu/pdo/PDO.latest.

PNA describes a connection pattern between mid- to upper-tropospheric air flow anomalies
over the North Pacific Ocean and North America, and is a leading mode of atmospheric
variability in the Northern Hemisphere winter (Leathers et al. 1991). PNA variability occurs
at timescales ranging from weeks to decades (Leathers et al. 1991), and is strongly correlated
to surface temperatures and precipitation (mostly winter and early spring) in the United
States and Canada (Shabbar et al. 1997, Leathers et al. 1991). A PNA time series derived
from a rotated (Varimax) principal components analysis of standardized monthly mean 500millibar height anomalies over the Northern Hemisphere is available online at:
http://www.cdc.noaa.gov/Correlation/pna.data.

13

1.2 Climate change and forests
Climate change is expected to have profound effects on forested ecosystems through multiple
pathways (Millar et al. 2007, Parmesan 2006, Loehle 2000, Gates 1990, Walther 2003) and in
conjunction with co-occurring phenomena such as forest fragmentation, exotic species
invasions, habitat degradation, increased atmospheric carbon dioxide concentrations, nitrogen
deposition, and pollution (Noss 2001, D'Arrigo et al. 2007, Millar et al. 2007). This
combination of factors may create historically unprecedented environmental conditions and a
highly uncertain situation for forest managers (Millar et al. 2007).

Some effects of climate change will be dramatic. For example, large-scale forest losses may
result from a higher frequency of extreme events such as droughts, windstorms, and wildfires
(Christensen et al. 2007, Gates 1990, Breshears et al. 2005) as well as climate-changemediated disease and insect outbreaks (Woods et al. 2005, Carroll et al. 2006). Somewhat
more subtle forest adaptation and migration responses to climate change will be associated
with changes in growing season length, growing degree-days, soil moisture availability, and
other climate-related factors important to plant life cycles (Davis and Shaw 2001, Parmesan
2006).

Adaptation will occur through numerous pathways such as physiological responses,
phenological shifts, and evolution that may allow in situ individuals and populations to
persist under new environmental conditions (Kozlowski and Pallardy 2002, Korner 2003,
Davis and Shaw 2001, Walther 2003, Parmesan 2006). Although trees are generally resilient
to climatic variation (Kozlowski and Pallardy 2002, Noss 2001), adaptive capacities may
vary with factors such as species, region, ecological conditions, age, and genetic diversity,
14

and in some cases, may be challenged by the unprecedented rapid rate of current
environmental change (Aitken et al. 2008, Davis and Shaw 2001, Spittlehouse 2005, Noss
2001, Millar etal. 2007).

Over a long period, Northern Hemispheric tree species are generally expected to migrate
northwards and upwards in elevation as temperatures increase (Davis and Shaw 2001, Lenoir
et al. 2008, Parmesan 2006). Tree migration will be a slow and somewhat random process
controlled by seed dispersal, germination, and seedling establishment rates, and may be an
insufficient response to current climate change from a species survival perspective due to
habitat fragmentation, competition from already established species, topographical barriers
such as mountain ranges, and other local factors (Parmesan 2006, Millar et al. 2007, Davis
and Shaw 2001, Hoegh-Guldberg et al. 2008, Aitken et al. 2008). Management interventions
such as facilitated migration may be applicable where species or populations have low
capacity to adapt or migrate in response to climate change, however, many critical
knowledge gaps and concerns need to be addressed (Hoegh-Guldberg et al. 2008, Millar et al.
2007, Aitken et al. 2008).

1.2.1 The effects of climate change on forest growth
As climate is a key factor influencing tree growth (Fritts 1976, Kimmins 1997, Nigh et al.
2004), large patterns of forest productivity shifts may occur as a result of climate change
(Nigh et al. 2004, Littell 2006, D'Arrigo et al. 2007). Elucidating species-, site-, and timespecific growth responses to climate change using dendroecological techniques (the study of
tree-rings and ecology; Fritts 1976) can help address critical knowledge gaps regarding future

15

species adaptation, productivity, abundances, resilience to disturbance, and community
associations (Fritts 1976, Littell and Peterson 2005, Littell 2006, Carrer and Urbinati 2006,
Zhang and Hebda 2004, Case and Peterson 2005, Pfeifer et al. 2005). Most tree-ring studies
have focused on the influence of temperature and precipitation on tree growth, as they are
primary climatic influences (Fritts 1976, Kimmins 1997) and historical data for these
variables are easily obtained from climate station instrumental records.

Many tree-ring studies have shown that climate-growth relationships vary among species
(e.g., Daniels and Watson 2003, Peterson and Peterson 2001, Adams and Kolb 2005), which
may reflect unique adaptive capacities, growth strategies, and resource requirements that
allow species to coexist (Davis and Shaw 2001, Green 2005, He et al. 2005). Speciesspecific climate-growth relationships have also been shown to vary along environmental
gradients such as elevation (Adams and Kolb 2005, Zhang and Hebda 2004, Case and
Peterson 2005), latitude (Watson and Luckman 2002, Feliksek and Wilczynski 2003), and
climate regime (Littell 2006, Littell and Peterson 2005), possibly reflecting the dominant
climatic limitation in a given environment. For example, growth tends to be limited by
temperatures in cooler and wetter environments, and by precipitation in warmer and drier
environments (Adams and Kolb 2005, Littell 2006), although interactions between
temperature and precipitation are also important regulators of forest growth (Littell and
Peterson 2005). Growth responses to a given climatic variable may be especially strong in
populations growing near ecotonal boundaries and may provide an indication of future
broader species responses (Littell and Peterson 2005, Neilson 1993, Loehle 2000, Walther et
al. 2002).

16

Recent studies have shown that climate-growth relationships can vary over time, possibly as
a function of climate change. For example, some historically cold-limited northern forests
have shown a decrease in temperature sensitivity and/or developed a new sensitivity to
precipitation over recent decades, possibly as a result of temperature-induced drought and/or
co-occurring non-climatic anthropogenic forcings such as pollution, increased atmospheric
carbon dioxide concentrations, and global dimming (Jacoby and D'Arrigo 1995, Lloyd and
Fastie 2002, Barber et al. 2000, and see D'Arrigo et al. 2007 for a review). Similar findings
have been reported for many tree species in northern Eurasia and North America (e.g., Pfeifer
et al. 2005, Wilmking and Myers-Smith 2008, Carrer and Urbinati 2006), implying that this
phenomenon may be widespread. Climate-growth relationships have also been shown to
vary during different phases of ocean-atmosphere climate systems such as PDO (e.g., Fagre
et al. 2003, Daniels and Veblen 2004). These types of findings challenge assumptions of
stable ecosystem responses to climate change (Millar et al. 2007, Millar et al. 2006) and have
significant implications for various aspects of forest management such as carbon cycle
modeling, growth and yield, the management of climatically sensitive populations, and the
use of tree-rings as historical climate proxies (Millar et al. 2006, D'Arrigo et al. 2007, Carrer
and Urbinati 2006, Wilson and Elling 2004, Pfeifer et al. 2005).

1.3 Douglas-fir
This project used dendroecological techniques to study the climate-growth relationships of
Douglas-fir (Pseudotsuga menziesii) in the British Columbia Interior. This species was
chosen for study because it typically has a strong growth response to climate, and, in the
interior of British Columbia, covers a wide climatic amplitude and occurs at three

17

geographic/climatic limits (high-elevation, dry, and northern). Douglas-fir is a productive
and commercial conifer found naturally in western North America from Mexico to BC
(Hermann and Lavender 1990), and is also an economically important exotic species in
Europe, Australia, New Zealand, and Chile (St. Clair et al. 2005, Hermann 1987). Two main
varieties of the species are recognized in BC; the coastal variety (P. menziesii (Mirb.) Franco
var. menziesii), which is generally found proximal to the Pacific Coast, and the interior
variety (P. menziesii var. glauca (Beissn.) Franco), which ranges eastward from the
rainshadow of the Coast Mountains. This project focused on the interior variety.

Interior Douglas-fir has a wide ecological amplitude, largely as a function of high intraspecific genetic diversity with a resulting strong capacity to adapt to local conditions (St.
Clair et al. 2005, Arno 1991, Anekonda et al. 2002, Rehfeldt 1991). In the BC Interior, the
species ranges from dry semi-arid forests in the lee of the Coast Mountains to low- to highelevation wetbelt forests on the western slopes of the Rocky Mountains (Arno 1991) to cold
interior forests near Fort St. James (Jull 1999), and inhabits seven biogeoclimatic zones,
including the Interior Douglas-fir, Engelmann Spruce-Subalpine Fir, Sub-Boreal Pine and
Spruce, Montane Spruce, Ponderosa Pine, Interior Cedar-Hemlock, and Sub-Boreal Spruce
zones (Meidinger and Pojar 1991, Hamann and Wang 2006). This species has a competitive
advantage in drier, fire-prone ecosystems, as mature trees have fire-resistant bark and can
grow adventitious roots, allowing trees to persist and reproduce after low-intensity
groundfires (Hermann and Lavender 1990, Delong 1999).

18

1.3.1 Douglas-fir climate-growth relationships
Many population- and regional-level dendroecological studies have shown Douglas-fir radial
growth to be sensitive to climatic (mostly temperature and precipitation) variables (e.g.,
Adams and Kolb 2005, Daniels and Watson 2003, Case and Peterson 2005, Zhang and Hebda
2004, Littel 2006, Littell and Peterson 2005, Watson and Luckman 2002, Nigh et al. 2004).
Growth responses of Douglas-fir to climate have been strong enough to be used in
reconstructing historical temporal precipitation (e.g. Daniels and Watson 2003, Watson and
Luckman 2001), temperature (Wilson and Luckman 2003, Laroque and Smith 2005), and
PDO (D'Arrigo et al. 2001, Biondi et al. 2001) variability.

1.3.1.1 Douglas-fir growth responses to precipitation
Douglas-fir radial growth is generally more sensitive to water deficits than any other
environmental variable (Littel 2006, Lopushinksy 1991, Case and Peterson 2005, Zhang and
Hebda 2004), although this growth response can vary with local ecological conditions (Littel
2006, Case and Peterson 2005, Zhang and Hebda 2004). In warm and dry environments,
Douglas-fir radial growth can be strongly limited by growing season precipitation (Adams
and Kolb 2005, Daniels and Watson 2003, Littell 2006, Case and Peterson 2005), reflecting
relatively poor stomatal control during droughts and low sapwood water storage capacity
(Lopushinksy 1991). In extended droughts, Douglas-fir stomata may remain permanently
closed, resulting in limited gas exchange, photosynthesis, (Lassoie and Salo 1981) and
reduced transpiration (Lopushinski and Klock 1974).

19

Although Douglas-fir growth in cold environments (i.e., high-elevation and northern) is often
primarily limited by cold temperatures (Feliksek and Wilczynski 2003, Case and Peterson
2005, Zhang and Hebda 2004, Littell 2006, Adams and Kolb 2005), significant positive
correlations with growing season precipitation have been found (Case and Peterson 2005,
Watson and Luckman 2002, Laroque and Smith 2005, Feliksek and Wilczynski 2003),
possibly reflecting a sensitivity to drought events that reduce an already short growing season
(Zhang and Hebda 2004). Douglas-fir growth in high-elevation environments can also have
negative correlations with snowfall (Littell 2006, Case and Peterson 2005), as deep
snowpacks may delay soil thaw and the start of the growing season (Littel 2006, Case and
Peterson 2005, Vaganov et al. 1999), decreasing root growth and water uptake (Brubaker
1980).

Douglas-fir annual growth has also been shown to be correlated to precipitation from the
previous year's growing season. Specifically, the earlywood portion of the annual ring can
be correlated to precipitation from the previous growing season, whereas the latewood
portion can be correlated to current summer precipitation (Daniels and Watson 2003, Watson
and Luckman 2002). A correlation between earlywood and previous growing season
precipitation may be explained in that primordial tissues are formed during the end of the
growing season, and water availability during the summer influences carbohydrate allocation
to the following year's formative tissues (Fritts 1976, Littel 2006). As well, Douglas-fir
needles are fully growing and photosynthesizing early in the season (compared to cooccurring species such as ponderosa pine [Pinus ponderosa Douglas ex C. Lawson]), and
therefore, may use soil moisture present from the previous year's precipitation (Watson and
Luckman 2002).
20

1.3.1.2 Douglas-fir growth responses to temperature
Douglas-fir radial growth can be positively correlated to growing season and winter
temperatures, particularly in cold high-elevation (Case and Peterson 2005, Zhang and Hebda
2004, Laroque and Smith 2005) and continental (Feliksek and Wilczynski 2003) sites. The
typically low temperatures found in cold environments limit growth by inhibiting
photosynthesis, glucose uptake, and meristematic activities (Korner 2003, Grace et al. 2002)
as well as creating limiting site conditions such as low soil temperatures, increased
snowpacks, and short growing seasons (Korner 2003, Kimmins 1997, Case and Peterson
2005, Laroque and Smith 2005). Growing season frost events can damage mature Douglasfir needles, inhibit hormonal system functioning, and delay the activity of formation tissue,
resulting in a smaller annual increment the following year (Feliksek and Wilczynski 2003).
Although warmer temperatures associated with recent climate change may alleviate cold
limitations to growth in these types of environments, they may also result in complex climate
interactions and new limitations to tree growth such as drought (Lloyd and Fastie 2002,
Barber et al. 2000). As well, winter temperatures may rise above the bud-chilling
requirement of Douglas-fir, resulting in a delay of bud burst in the spring and lower growth
(McCreary et al. 1990).

Growing season temperatures are often negatively correlated with Douglas-fir radial growth
on warm and dry sites at mid- to low-elevations (Case and Peterson 2005, Zhang and Hebda
2004, Laroque and Smith 2005, Feliksek and Wilczynski 2003, Daniels and Watson 2003,
Watson and Luckman 2001). An increase in temperature affects respiration and transpiration
processes and may reduce growth by increasing water loss, reducing nutrient storage,
decreasing foliage efficiency (Laroque and Smith 2005, Kimmins 1997, Balatinecz and
21

Anderson 1986), or by exceeding a species-specific upper photosynthesis threshold (which
can be as low as 20-25°C for Douglas-fir; Salo 1974, Doehlert and Walker 1981).

1.4 Project objectives
A key step in preparing for climate change will be addressing the many knowledge gaps
regarding historical and possible future tree species growth responses to climate variability
(Littell and Peterson 2005, Littell 2006). The climate-growth relationships of Douglas-fir
have been extensively studied, however, many knowledge gaps remain. Although Douglasfir climate-growth relationships have been quantified across elevation (e.g., Zhang and
Hebda 2004, Case and Peterson 2005) and latitude (e.g.,Watson and Luckman 2002)
gradients, to our knowledge, there have been no studies examining changes in Douglas-fir
climate-growth relationships over a large climatic gradient (as defined by local climate
normals) in the British Columbia Interior. As well, we know very little about the climategrowth relationships at the climatic margins of the species in the BC Interior, and how these
relationships relate to broader species responses. Further, there appears to be very little
knowledge about the temporal stability of Douglas-fir climate-growth relationships. This
project addressed these knowledge gaps by describing regional, climatic, and temporal
patterns in Douglas-fir climate-growth relationships across a wide climatic range in the BC
Interior, including populations growing at the climatic margins of the species' range. When
describing site characteristics in this study, the term 'climatic' refers to climate normals
(defined as the average from 1961-1990), whereas 'ecological' refers to both climate normals
and associated site characteristics such as elevation and topography.

22

This project focused on the following specific questions:
1) Are there spatially coherent ecological patterns in Interior Douglas-fir climate-growth
relationships?
2) How do climate-growth relationships change over a climatic gradient, including
populations located in climatically extreme environments?
3) Are climate-growth relationships stable over time?
4) Are there ecological patterns in temporally unstable climate-growth relationships?

These questions were addressed in two research chapters. Regional and ecological patterns
in Douglas-fir climate-growth relationships across the BC Interior are described in Chapter 2.
In chapter 3, temporal patterns in climate-growth relationships are examined with a focus on
populations growing at two climatic limits (cold and wet high-elevation sites and dry climate
sites). The management implications of key study results are discussed in Chapter 4.

23

1.5 Literature cited
Adams, H.D. and T.E. Kolb. 2005. Tree growth response to drought and temperature in a
mountain landscape in northern Arizona, USA. Journal of Biogeography 32: 16291640.
Aitken, S.N., S. Yeaman, J.A. Holliday, T. Wang, and S. Curtis-McLane. 2008. Adaptation,
migration or extirpation: climate change outcomes for tree populations. Evolutionary
Applications 1: 95-111.
Allan, R.J. 2000. ENSO and climatic variability in the past 150 years, p. 3-56. In El Nino and
the Southern Oscillation: multiscale variability and global and regional impacts. Diaz,
H. F. and Markgraf, V. (eds). Cambridge University Press, New York.
Anekonda, T.S., M.C. Lomas, W.T. Adams, K.L. Kavanagh, and S.N. Aitken. 2002. Genetic
variation in drought hardiness of coastal Douglas-fir seedlings from British Columbia.
Canadian Journal of Forest Research 32: 1701-1716.
Arno, S.F. 1991. Ecological relationships of interior Douglas-fir, p. 47-51. In Interior
Douglas-fir: The species and its management symposium proceedings. Baumgartner,
D. M. (ed). Department of Natural Resource Sciences, Washington State University,
Pullman, Washington.
Balatinecz, J.J. and H.W. Anderson. 1986. Effects of climatic factors on the formation of
wood and its quality, p. 195-200. In Climate Applications in Forest Renewal and
Forest Production, Proceedings of Forest Climate 1986. Maclver, D. C, Street, R. B.,
and Auclair, A. N. (eds). Canadian Forestry Service, Ottawa, Ontario.
Barber, V., G. Juday, and B. Finney. 2000. Reduced growth of Alaska white spruce in the
twentieth century from temperature-induced drought stress. Nature 405: 668-672.
BC MoE. 2007. Environmental Trends in British Columbia: 2007. State of Environment
Reporting. BC Ministry of Environment, Victoria, B.C.
BC MWLAP. 2002. Indicators of climate change for British Columbia, 2002. Indicators of
climate change for British Columbia, 2002. BC Ministry of Water, Land and Air
Protection, Victoria, B.C.
Biondi, F., A. Gershunov, and D. Cayan. 2001. North Pacific decadal climate variability
since 1661. Journal of Climate 14: 5-10.
Breshears, D.D., N.S. Cobb, P.M. Rich, K.P. Price, CD. Allen, R.G. Balice, W.H. Romme,
J.H. Kastens, M.L. Floyd, J. Belnap, J.J. Anderson, O.B. Myers, and C.W. Meyer.
2005. Regional vegetation die-off in response to global-change-type drought.
Proceedings of the National Academy of Sciences 102: 15144-15148.

24

Brubaker, L.B. 1980. Spatial patterns of tree growth anomalies in the Pacific Northwest.
Ecology 6: 798-807.
Carrer, M. and C. Urbinati. 2006. Long-term change in the sensitivity of tree-ring growth to
climate forcing in Larix decidua. New Phytologist 170: 861-872.
Carroll, A.L., J. Regniere, J.A. Logan, S.W. Taylor, B.J. Bentz, and J.A. Powell. 2006.
Impacts of climate change on range expansion by the mountain pine beetle. Mountain
Pine Beetle Initiative Working Paper 2006-14. Natural Resources Canada, Victoria,
B.C.
Case, M.J. and D.L. Peterson. 2005. Fine-scale variability in growth-climate relationships of
Douglas-fir, North Cascade Range, Washington. Canadian Journal of Forest Research
35: 2743-2755.
Cayan, D.R., M.D. Dettinger, H.F. Diaz, and N.E. Graham. 1998. Decadal variation of
precipitation over Western Northern America. Journal of Climate 11: 3148-3166.
Christensen, J.H., B. Hewitson, A. Busuioc, A. Chen, X. Gao, I. Held, R. Jones, R.K. Kolli,
W.-T. Kwon, R. Laprise, V. Magaha Rueda, L. Mearas, C.G. Menendez, J. Raisanen,
A. Rinke, A. Sarr, and P. Whetton. 2007. Regional Climate Projections, p. 848-940.
In Climate Change 2007: The Physical Science Basis. Contribution of Working
Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate
Change. Solomon, S. D., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K. B.,
Tignor, M., and Miller, H. M. (eds). Cambridge University Press, New York.
D'Arrigo, R., R. Villalba, and G. Wiles. 2001. Tree-ring estimates of Pacific decadal climate
variability. Climate Dynamics 18: 219-224.
D'Arrigo, R., R. Wilson, B. Liepert, and P. Cherubini. 2007. On the 'Divergence Problem' in
Northern Forests: A review of the tree-ring evidence and possible causes. Global and
Planetary Change 289-305.
Dai, A., K.E. Trenberth, and T.R. Karl. 1999. Effects of clouds, soil moisture, precipitation,
and water vapor on diurnal temperature range. Journal of Climate 12: 2451-2473.
Daniels, L.D. and T.T. Veblen. 2004. Spatiotemporal influences of climate on altitudinal
treeline in northern Patagonia. Ecology 85: 1284-1296.
Daniels, L.D. and E. Watson. 2003. Climate-fire-vegetation interactions in the Cariboo
forests: a dendrochronological analysis. Report to Forest Innovation and Investment.
Forest Innovation and Investment Research Program, Vancouver, BC.
Davis, M.B. and Shaw. R.G. 2001. Range shifts and adaptive responses to Quaternary
climate change. Science 292: 673-679.

25

Delong, S.C. 1999. Ecology of Douglas-fir at its northern limits, p. 27-28. In Ecology and
management of interior Douglas-fir (Pseudotsuga menziesii var glauca) at the
northern extreme of its range. Proceedings of a workshop held 7-9 October 1996 in
Fort St. James, British Columbia. Lousier, J. D. and Kessler, W. B. (eds). University
of Northern British Columbia, Prince George, BC.
Doehlert, D.C. and R.B. Walker. 1981. Photosynthesis and photorespiration in Douglas-fir as
influenced by irradiance, C02 concentration, and temperature. Forest Science 27:
641-650.
Easterling, D.R., G.A. Meehl, C. Parmesan, S.A. Changnon, T.R. Karl, and L.O. Mearns.
2000. Climate extremes: observations, modeling, and impacts. Science 289: 20682074.
Fagre, D.B., D.L. Peterson, and A.E. Hessl. 2003. Taking the pulse of mountains: Ecosystem
responses to climatic variability. Climatic Change 59: 263-282.
Feliksek, E. and S. Wilczynski. 2003. Dendroecological characterization of Douglas-fir
(Pseudotsuga menziesii Franco) in the Wielkopolska region. Electronic Journal of
Polish Agricultural Universities 6: Series: Forestry.
Fritts, H.C. 1976. Tree rings and climate. The Blackburn Press, New Jersey.
Gates, D.M. 1990. Climate change and forests. Tree Physiology 7: 1-5.
Gedalov, Z. and N.J. Mantua. 2002. A multi-century perspective of variability in the Pacific
Decadal Oscillation: new insights from tree rings and coral. Geophysical Research
Letters 29: 1-4.
Gedalov, Z. and D.J. Smith. 2001. Interdecadal climate variability and regime-scale shifts in
Pacific North America. Geophysical Research Letters 28: 1515-1518.
Gershunov, A., N. Schneider, and T. Barnett. 2001. Low-frequency modulation of ENSO Indian Monsoon rainfall relationship: signal or noise? Journal of Climate 14: 24862492.
Grace, J., F. Berninger, and N. Laszlo. 2002. Impacts of climate change on the tree line.
Annals of Botany 90: 537-544.
Green, D.S. 2005. Adaptive strategies in seedlings of three co-occurring, ecologically
distinct northern coniferous tree species across and elevational gradient. Canadian
Journal of Forest Research 35: 910-917.
Hamann, A. and T. Wang. 2006. Potential effects of climate change on ecosystem and tree
species distribution in British Columbia. Ecology 87: 2773-2786.

26

Hare, S.R. and N.J. Mantua. 2001. An historical narrative on the Pacific Decadal Oscillation,
interdecadal climate variability and ecosystem impacts. Report of a talk presented at
the 20th NE Pacific Pink and Chum workshop, Seattle, WA, 22 March 2001.
http://www.iphc.washington.edu/Staff/hare/html/papers/pcworkshop/pcworkshop.pdf.
He, J.-S., Q.-B. Zhang, and F.A. Bazzaz. 2005. Differential drought responses between
saplings and adult trees in four co-occurring species of New England. Trees 19: 442450.
Hermann, R.K. 1987. North American tree species in Europe. Journal of Forestry 85: 27-32.
Hermann, R.K. and D.P. Lavender. 1990. Douglas-fir. In Silvics of North America: 1.
Conifers; 2. Hardwoods. Agriculture Handbook 654. Burns, R. M. and Honkala, B. H.
(eds). U.S. Department of Agriculture, Washington, DC.
http://www.na.fs.fed.us/spfo/pubs/silvicsmanual/Volumel/pseudotsuga/menziesii.h
tm.
Hoegh-Guldberg, O., L. Hughes, S. Mclntyre, D.B. Lindenmayer, C. Parmesan, H.P.
Possingham, and C D . Thomas. 2008. Assisted colonization and rapid climate change.
Science 321: 345-346.
Jacoby, G.C. and R. D'Arrigo. 1995. Tree-ring width and density evidence of climatic and
potential forest change in Alaska. Global Biogeochemical Cycles 9: 227-234.
Jones, P.D. 1988. Hemispheric surface air temperature variations: Recent trends and an
update to 1987. Journal of Climate 1: 654-660.
Jull, M.J. 1999. Douglas-fir silviculture "on the edge": silvicultural systems at the northern
range of the species, p. 34-44. In Ecology and management of interior Douglas-fir
(Pseudotsuga menziesii var glauca) at the northern extreme of its range. Proceedings
of a workshop held 7-9 October 1996 in Fort St. James, British Columbia. Lousier, J.
D. and Kessler, W. B. (eds). University of Northern British Columbia, Prince George,
BC.
Karl. T.R., Jones. P.D., R.W. Knight, G. Kukla, N. Plummer, V. Razuvayev, K.P. Gallo, J.
Lindsaey, R.J. Charlson, and T.C. Peterson. 1993. A new perspective on recent global
warming: asymmetric trends of daily maximum and minimum temperature. Bulletin
of the American Meteorological Society 74: 1007-1023.
Kimmins, J. P. 1997. Forest Ecology. Prentice-Hall Inc., New Jersey.
Knowles, N., M.D. Dettinger, and D.R. Cayan. 2006. Trends in snowfall versus rainfall in the
Western United States. Journal of Climate 19: 4545-4559.

27

Kozlowski, T.T. and S.G. Pallardy. 2002. Acclimation and adapative responses of woody
plants to environmental stresses. The Botanical Review 68: 270-334.
Korner, C. 2003. Alpine plant life. Springer-Verlag, New York.
Laroque, S.J. and D.J. Smith. 2005. A dendroclimatological reconstruction of climate since
AD 1700 in the Mt. Waddington area, British Columbia Coast Mountains, Canada.
Dendrochronologia 22: 93-106.
Lassoie, J.P. and D.J. Salo. 1981. Physiological response of large Douglas-fir to natural and
induced soil water deficits. Plant physiology and biochemistry 11: 139-144.
Leathers, D.J., B. Yarnal, and M.A. Palecki. 1991. The Pacific/North American
Teleconnection Pattern and United State Climate. Part I: Regional Temperature and
Precipitation Associations. Journal of Climate 4: 517-528.
Lenoir, J., J.C. Gegout, P.A. Marquet, P. de Huffray, and H. Brisse. 2008. A significant
upward shift in plant species optimum elevation during the 20th century. Science 320:
1768-1771.
Leung, L.R., Y. Qian, X. Bian, W.M. Washington, J. Han, and J.O. Roads . 2004. Midcentury ensemble regional climate change scenarios for the western United States.
Climatic Change 62: 75-113.
Littell, J. S. 2006. Climate impacts to forest ecosystem processes: Douglas-fir growth in
north-western U.S. mountain landscapes and area burned by wildfire in western U.S.
eco-provinces. PhD Dissertation. University of Washington. Seattle, WA.
Littell, J.S. and D.L. Peterson. 2005. A method for estimating vulnerability of Douglas-fir
growth to climate change in the northwestern U.S. The Forestry Chronicle 81: 369374.
Lloyd, A.H. and C.L. Fastie. 2002. Spatial and temporal variability in the growth and climate
response of treeline trees in Alaska. Climatic Change 52: 481-509.
Loehle, C. 2000. Forest ecotone response to climate change: sensitivity to temperature
response functional forms. Canadian Journal of Forest Research 30: 1632-1645.
Lopushinksy, W. 1991. Water relations of interior Douglas-fir, p. 35-46. In Interior Douglasfir: The species and its management symposium proceedings. Baumgartner, D. M.
(ed). Department of Natural Resource Sciences, Washington State University,
Pullman, Washington.
Lopushinksy, W. and G.O. Klock. 1974. Transpiration of conifer seedlings in relation to soil
water potential. Forest Science 20: 181-186.
Mantua, N.J. and S.R. Hare. 2002. The Pacific Decadal Oscillation. Journal of Oceanography
58: 35-44.

28

Mantua, N.J., S.R. Hare, Y. Zhang, J.M. Wallace, and R.C. Francis. 1997. A Pacific
interdecadal climate oscillation with impacts on salmon production. Bulletin of the
American Meteorological Society 78: 1069-1079.
McCreary D.D., D.P. Lavender, and R.K. Hermann. 1990. Predicted global warming and
Douglas-fir chilling requirements. Annals of Forest Science 47: 325-330.
Meehl, G.A., W.M. Washington, W.D. Collins, J.M. Arblaster, A. Hu, L.E. Buja, W.G.
Strand, and H. Teng. 2005. How much more global warming and sea level rise?
Science 307: 1769-1772.
Meidinger, D. and J. Pojar. 1991. Ecosystems of British Columbia. BC Ministry of Forests,
Victoria, B.C.
Millar, C, R. Neilson, D. Bachelet, R. Drapek, and J. Lenihan. 2006. Climate change at
multiple scales, p. 30-60. In Forests, carbon and climate change: a synthesis of
science findings. Oregon Forest Resources Institute, Portland, Oregon.
Millar, C.I., N.L. Stephenson, and S.L. Stephens. 2007. Climate change and forests of the
future: managing in the face of uncertainty. Ecological Applications 17: 2145-2151.
Moberg, A., D.M. Sonechkin, N.M. Holmgren, N.M. Datsenko, and W. Karlen. 2005. Highly
variable Northern Hemisphere temperatures reconstructed from low- and highresolution proxy data. Nature 433: 613-617.
Neilson, R.P. 1993. Transient ecotone response to climatic change: some conceptual and
modelling approaches. Ecological Applications 3: 385-395.
Nigh, G.D., C.C. Ying, and H. Qian. 2004. Climate and productivity of major conifer species
in the interior of British Columbia, Canada. Forest Science 50: 659-671.
Noss, R.F. 2001. Beyond Kyoto: forest management in a time of rapid climate change.
Conservation Biology 15: 578-590.
Parmesan, C. 2006. Ecological and evolutionary responses to recent climate change. Annual
Review of Ecology, Evolution, and Systematics 37: 637-669.
Peterson, D.W. and D.L. Peterson. 2001. Mountain hemlock growth responds to climatic
variability at annual and decadal time scales. Ecology 82: 3330-3345.
Pfeifer, K., W. Kofler, and W. Oberhuber. 2005. Climate related causes of distinct radial
growth reductions in Pinus cembra during the last 200 yr. Vegetation History and
Archaeobotany 14:211-220.
Rehfeldt, J. 1991. The genetic resource of Douglas-fir in the interior northwest, p. 53-62. In
Interior Douglas-fir: The species and its management symposium proceedings.
Baumgartner, D. M. (ed). Department of Natural Resource Sciences, Washington
State University, Pullman, Washington.

29

Salo, D. J. 1974. Factors affecting photosynthesis in Douglas-fir. PhD Dissertation.
University of Washington. Seattle, WA.
Shabbar, A., B. Bonsai, and M. Khandekar. 1997. Canadian Precipitation Patterns Associated
with the Southern Oscillation. Journal of Climate 10: 3016-3027.
Smith, C.A. and P. Sardeshmukh. 2000. The effect of ENSO on the intraseasonal variance of
surface temperature in winter. International Journal of Climatology 20: 1543-1557.
Spittlehouse, D.L. 2005. Integrating climate change adaptation into forest management.
Forestry Chronicle 81: 691-695.
St. Clair, J.B., N.L. Mandel, and K.W. Vance-Borland. 2005. Genecology of Douglas Fir in
Western Oregon and Washington. Annals of Botany 96: 1199-1214.
Trenberth, K.E. 1990. Recent observed interdecadal climate changes in the Northern
Hemisphere. Bulletin of the American Meteorological Society 71: 988-993.
Vaganov, E.A., M.K. Hughes, A.V. Kirdyanov, F.H. Schweingruber, and P.P. Silkin. 1999.
Influence of snowfall and melt timing on tree growth in subarctic Eurasia. Nature
400: 149-151.
Walther, G.-R. 2003. Plants in a warmer world. Perspectives in Plant Ecology, Evolution,
and Systematics 6: 169-185.
Walther, G.-R., E. Post, P. Convey, A. Menzei, C. Parmesan, T.J.C. Beebee, J.-M.
Fromentin, O. Hoegh-Guldberg, and F. Bairlein. 2002. Ecological responses to recent
climate change. Nature 416: 389-395.
Watson, E. and B. Luckman. 2001. Dendroclimatic reconstruction of precipitation for sites in
the southern Canadian Rockies. The Holocene 11: 203-213.
Watson, E. and B.H. Luckman. 2002. The dendroclimatic signal in Douglas-fir and
ponderosa pine tree-ring chronologies from the southern Canadian Cordillera.
Canadian Journal of Forest Research 32: 1858-1874.
Wilmking M. and I. Myers-Smith. 2008. Changing climate sensitivity of black spruce (Picea
mariana Mill.) in a peatland-forest landscape in Interior Alaska. Dendrochronologia
25: 167-175.
Wilson, R. and W. Elling. 2004. Temporal instability in tree growth/climate response in the
Lower Bavarian Forest region: implications for dendroclimatic reconstruction. Trees
18: 19-28.
Wilson, R.J.S. and B.H. Luckman. 2003. Dendroclimatic reconstruction of maximum
summer temperatures from upper treeline sites in Interior British Columbia. The
Holocene 13: 851-861.

30

Woods, A., K.D. Coates, and A. Hamann. 2005. Is an unprecedented Dothistroma needle
blight epidemic related to climate change? BioScience 55: 761-769.
Zhang, Q.-B. and R.J. Hebda. 2004. Variation in radial growth patterns of Pseudotsuga
menziesii on the central coast of British Columbia, Canada. Canadian Journal of
Forest Research 34: 1946-1954.
Zhang, X., F.W. Zwiers, G.C. Hegerl, F.H. Lambert, N.P. Gillett, S. Solomon, P.A. Stott, and
T. Nozawa. 2007. Detection of human influence on twentieth-century precipitation
trends. Nature 448: 461-465.

31

Chapter 2. Regional and ecological patterns in Douglas-fir climate-growth relationships
in the British Columbia Interior
Abstract
How climate change will affect tree growth across species' ranges remains a critical
knowledge gap. Tree-ring data were analyzed from 33 Douglas-fir stands spanning a wide
climatic and geographic range in the interior of British Columbia to identify regional and
ecological patterns between climate and growth. Populations growing in warm and dry
climates had growth patterns correlated mostly to local annual precipitation, whereas
populations growing in high-elevation wet and cold climates were more correlated to
annual/winter snowfall/winter temperatures and quasi-periodic ocean-atmosphere climate
systems. Site annual/summer heat-moisture index and precipitation regime were able to
predict the strength of population growth responses to annual precipitation and PNA/PDO
index, and the strongest responses to both variables were found in populations growing at the
climatic extremes. Across most of its range, Douglas-fir growth appears to be limited by
annual moisture, likely reflecting a high sensitivity to available soil moisture. Future
temperature-induced drought conditions in BC could result in widespread heightened
sensitivity to precipitation and productivity declines.

32

2.1 Introduction
In British Columbia (BC), annual temperatures have increased by 0.5 to 1.7°C over the past
century and are projected to increase by 2 to 5°C above 1961-1990 averages by the middle of
the 21 st century, depending on future greenhouse emissions (BC MWLAP 2002, BC MoE
2007, Christensen et al. 2007, Meehl et al. 2005). Nighttime minimum temperatures have
increased more rapidly than daytime maximum temperatures, thus lowering the diurnal
temperature range, reducing growing season frost events, and lengthening the growing season
(Karl et al. 1993, Dai et al. 1999, BC MWLAP 2002, BC MoE 2007). Annual precipitation
has also increased significantly over most of western North America (Zhang et al. 2007,
Trenberth 1990), although some regions of BC have become drier over the past 50 years (BC
MoE 2007). Projections show a modest precipitation increase over most of BC, with wetter
summers and more winter rains at low- to mid-elevations (BC MoE 2007), however, some
areas may experience net drying as a result of increased temperatures and lower snowpacks
(Christensen et al. 2007, Leung et al. 2004, Knowles et al. 2006). The frequency of droughts,
severe storms, and other extreme climatic events may also increase in association with
climate change (Easterling et al. 2000). In western North America, oscillatory interactions
between the Pacific Ocean and the atmosphere are important sources of low-frequency
climate variation that will additionally influence future regional climate regimes (Mantua et
al. 1997, Hare and Mantua 2001, Cayan et al. 1998, Knowles et al. 2006, Leathers et al.
1991).

Temperature-precipitation changes are expected to significantly affect forests by influencing
growing season length, growing degree-days, soil moisture availability, and other climate-

33

related factors important to plant life cycles (Kimmins 1997, Fritts 1976, Davis and Shaw
2001). Tree responses to climate change will be expressed through changes in reproductive,
physiological, and phenological processes (Aitken et al. 2008, Walther 2003, Walther et al.
2002, Aber et al. 2001, Hansen et al. 2001, Parmesan and Yohe 2003, Gates 1990, Graham et
al. 1990), and will reflect species-specific adaptive capacities as well as the influence of local
factors such as soil nutrients, competition, forest disturbances, and genetic diversity (Aitken
et al. 2008, Hamann and Wang 2006, Graumlich 1991, Spittlehouse and Stewart 2003). Over
a long period, species- and site-specific responses to climate change will result in altered
forest compositions and new species ranges (Hamann and Wang 2006, Davis and Shaw
2001).

Tree growth and productivity may rapidly shift in response to climate change (Littell 2006,
Stephenson and van Mantgem 2005, Hanson and Weltzin 2000) and identifying the climate
mechanisms underlying these shifts may provide important information regarding species
adaptation, survival, productivity, resilience to disturbance, and long-term range shifts
(Peterson and Peterson 2001, D'Arrigo et al. 2007, Carrer and Urbinati 2006, Zhang and
Hebda 2004, Littell and Peterson 2005, Case and Peterson 2005). Climate-growth
relationships should be quantified across the spectrum of the species' ecological range,
including its extremes (Peterson and Peterson 2001, Littell 2006, Littell and Peterson 2005,
Case and Peterson 2005, Zhang and Hebda 2004), as climatic growth limitations may vary
with local environmental conditions. For example, trees growing in cold environments tend
to be limited by low temperatures (Barber et al. 2000, Case and Peterson 2005, Zhang and
Hebda 2004), whereas dry ecosystem forests are often more limited by precipitation (Adams
and Kolb 2005, Lloyd and Fastie 2002). Stands growing near their ecological margins may
34

be especially sensitive to climate variation due to strong environmental limitations and may
provide an indication of future broader species responses (Loehle 2000, Lloyd 2005, Neilson
1993, Case and Peterson 2005).

Interior Douglas-fir {Pseudotsuga menziesii var. glauca (Mirb.) Franco) is a commercially
important conifer that grows over a wide range of ecological conditions and is found at its
high-elevation, northern, and dry limits in the interior of British Columbia (Hermann and
Lavendar 1990, M l 1999, Delong 1999, Arno 1991). Numerous studies have shown this
species has a high sensitivity to climate (e.g., Daniels and Watson 2003, Watson and
Luckman 2002) and that its climate-growth relationships can vary over elevation (e.g., Zhang
and Hebda 2004, Case and Peterson 2005, Adams and Kolb 2005), continentality (e.g.,
Feliksek and Wilczynski 2003), and latitude (e.g.,Watson and Luckman 2002) gradients,
however, relatively little is known about Douglas-fir climate-growth relationships along its
climatic range (as defined by local mean annual temperature and precipitation) (Littell 2006),
particularly in British Columbia (Spittlehouse 2003).

This study used dendroecological methods (Fritts 1976) to quantify radial growth variation
and climate-growth relationships of interior Douglas-fir across a wide climatic and
geographic range in the British Columbia Interior, encompassing sites near the geographic
and climatic margins. The objectives of this study were to: 1) identify climate variables that
influence Douglas-fir growth and 2) identify and examine regional and ecological (i.e. site
climate normals and associated features such as elevation) patterns of interior Douglas-fir
growth variation across its climatic range in the interior of British Columbia.

35

2.2 Materials and methods
2.2.1 Sampling
We sampled 33 mature stands of Douglas-fir in three geographic regions (Figure 2.1) in the
British Columbia Interior: 1) Northern Interior (Figure 2.2), 2) Chilcotin Plateau (2.3), and 3)
Southern Interior (Figure 2.4). The 33 sample stands (sites) were strategically chosen to
encompass a wide range of climatic regimes, including sites located at the extremes in each
region. Potential sampling sites were identified by overlaying spatial mean annual
temperature (MAT) and precipitation (MAP) data (1961-1990 climate normal period,
Spittlehouse 2006) obtained from ClimateBC (v3.1, Wang et al. 2006) onto mature Douglasfir coverage obtained from the BC Ministry of Forest and Range's Vegetation Resource
Inventory. Sample site MAT ranged from 1.5 to 7.4 °C, and MAP ranged from 335 to 2077
mm (Table 2.2). The elevation range was 370 to 1540 m (Table 2.1). The sampled sites
encompassed seven biogeoclimatic (BEC) zones (Meidinger and Pojar 1991), representing a
wide diversity of ecosystems (Table 2.2).

Sample sites and trees were chosen to maximize the climate signal in growth variation. Sites
without evidence of edaphic limiting factors (e.g., shallow rocky soils, cold air ponding,
steep northerly aspects) or recent abiotic or biotic disturbance were selected in order to
minimize the growth impacts of non-climatic factors. The influence of inter-tree competition
was minimized by sampling trees with a dominant canopy position. We avoided sampling
trees with obvious growth defects or evidence of disease, fire scarring, or insect attacks as
much as possible. At each site, between 18 and 40 mature trees greater than 100 years old
were sampled to build a robust population sample. A single increment core was taken from
each sample tree at approximately breast height (1.3m, Josza 1988). Where sampling

36

occurred on slopes, increment cores were taken perpendicular to the slope direction to avoid
compression and tension wood (Josza 1988).

2.2.2 Processing and chronology development
Increment cores were mounted and sanded to enhance the contrast of tree ring boundaries.
Cores were then visually crossdated to assign calendar years to annual growth rings (Cook
and Kairiukstis 1990). Ring widths were measured using WinDendro™ image analysis
software (Regent Instruments Inc. 2005) to the nearest 0.01mm. Crossdating was verified
using COFECHA software (Holmes 1983). After checking for measurement errors, cores
that did not crossdate with the master chronology were excluded from the final chronology
(Daniels and Watson 2003).

The ARSTAN software (Cook and Krusic 2005) was used to remove the age-related growth
trend from each series, as this trend usually dominates the variance and masks the climate
signal (Fritts 1976, Cook and Kairiukstis 1990, Cook 1985). A cubic smoothing spline with
a 50% frequency response cutoff of 60 years was used to remove very low-frequency (i.e.,
age-related trend) variation (Cook 1985, Cook and Kairiukstis 1990). Cubic smoothing
splines remove unwanted variation by filtering data in the frequency domain; the spline used
in this study removed 50% of the signal amplitude at a frequency of 1 cycle/60 years, and an
increasing (decreasing) percentage of the signal amplitude at frequencies lower (higher) than
60 years (Cook 1985, Cook and Kairiukstis 1990).

37

After detrending, ARSTAN was used to analyze autocorrelation and apply a common
autoregressive model to each series for a given site (Cook and Krusic 2005, Cook 1985).
Residuals from autoregressive modeling were then averaged together using robust mean
calculation to produce a residual growth chronology that represented stand growth variation
(Cook and Krusic 2005). Removing autocorrelation from growth chronologies is a common
procedure in dendroecological studies (e.g., Chhin et al. 2008, Case and Peterson 2005,
Littell 2006, Pfeifer et al. 2005) because autocorrelation has statistical implications (Pfeifer et
al. 2005, Littell 2006, Meko 2007) and may be a function of non-climatic processes such as
stand-level disturbances and morphological influences on growth (Littell 2006, Cook 1985,
Fritts 1976). A preliminary analysis of this dataset revealed that autocorrelation confounded
climate-growth relationships (i.e. autocorrelation may have resulted from non-climatic
sources, Chhin et al. 2008, Biondi and Swetnam 1987). Robust mean calculation is more
resistant to outliers, which may serve to reduce noise from tree-level growth disturbances
(Cook 1985, Cook and Kairiukstis 1990).

2.2.3 Data analysis
2.2.3.1 Chronology analysis
Chronology statistics were calculated for each site, including mean sensitivity, standard
deviation (both indicate annual ring-width variability, Cook and Kairiukstis 1990, Fritts
1976), and average intrasite correlation (average correlation of all trees within a site to the
chronology, analogous to the population signal, Cook and Kairiukstis 1990, Fritts 1976).

The 33 chronologies were analyzed using principal components analysis (PCA, The
Mathworks Inc. 2007) to identify common growth patterns and generate composite

38

chronologies for climate-growth analysis. Analysis was restricted to a common time period
(1900-2005). This time period was also chosen because it matched the longest climate
records in the study and therefore allowed for subsequent analysis of climate-growth
relationships over the past century (see Chapter 2). PCA uses eigenvalues and eigenvectors
to recombine a dataset into a smaller set of latent variables (often termed 'principal
components', hereafter, PCs) that represent the main modes of variability in the original data
(Fritts 1976, Biondi and Waikul 2004, Tabachnick and Fidell 1989). In this case, extracted
PCs were time series that represented growth variation. PCs with a corresponding eigenvalue
greater than 1 were considered significant and retained for analysis (Tabachnick and Fidell
1989). Varimax rotation was used to improve the interpretability of PCA results relative to
the original data set (The Mathworks Inc. 2007, Tabachnick and Fidell 1989).

A preliminary PCA on the entire domain (i.e., using all chronologies) revealed that growth
patterns were dominated by regional variation that masked underlying ecological patterns. In
order to control for regional variation, we treated each sampling region (Northern Interior,
Southern Interior, and Chilcotin Plateau) as a separate group for PCA (i.e., a separate PCA
was completed for each region). This approach also allowed for a comparison between the
main growth patterns in each region and allowed for correlation analysis between PCs and
local climate station data (see below for description of local climate station data). Using
PCA to compare the main modes of variability between groups is rare in dendroecology,
however, this approach has been used in other fields (e.g., ichthyology, Rising and Somers
1989, Douglas 1993, and c.f. Tabachnick and Fidell 1989).

39

The relationships between the regional PCs and the original chronologies were analyzed
using correlation analysis. Correlation coefficients (also referred to as 'loadings') computed
between the original chronologies and the PCs were plotted against site-specific climatic
normals and geographic variables to determine the ecological conditions and spatial patterns
represented by the PCs.

2.2.3.2 Climate-growth relationships
Climate-growth relationships were examined in two separate analyses. The first analysis
(termed 'regional climate-growth relationship analysis', see below) correlated regional
growth variation (as represented by the PCs) with Environment Canada Climate Station Data
and ocean-atmosphere climate system indices to determine the primary climatic limitations in
each study region. The second analysis (termed 'climatic gradient analysis') examined how
population-level climate-growth relationships varied over a climatic gradient spanning the
entire study domain.

Analyses were performed using annual, seasonal, and monthly variables, however, results
involving only annual and seasonal variables are presented because they explained more
growth variation than monthly variables (e.g., Littell 2006, Watson and Luckman 2002).

2.2.3.2.1 Climate data
Environment Canada climate station temperature and precipitation data were used as
predictor variables in the regional climate-growth relationship analysis. Climate stations
corresponding to regional PCs were chosen based on spatial and ecological interpretations of

40

the PCA results, proximity to sample sites, and sufficient time series length. Adjusted
monthly precipitation records (Mekis and Hogg 1999) and homogenized monthly
temperature records (Vincent and Gullett 1999, Vincent et al. 2002) provided by the Climate
Research Division of the Meteorological Service of Canada
(http://www.cccma.bc.ec.gc.ca/hccd/) for the chosen climate stations were used to account
for non-climatic variation in the instrumental records caused by changes in station location,
measurement procedures, instrument changes, and other factors (Mekis and Hogg 1999,
Vincent and Gullett 1999, Vincent et al. 2002).

We made further adjustments to two data records. The Williams Lake temperature record
was not homogenized at the time of this study, therefore we used the observed data.
However, the station moved location (and over 300m up in elevation) in 1961, which likely
introduced inhomogeneities in the climate record. We accounted for this by using
ClimateBC elevation-adjusted and interpolated homogenized historical temperature data
(Wang et al. 2006, Mitchell and Jones 2005) from 1939-1960 (see below for a description of
the ClimateBC model). A visual comparison of Williams Lake observed temperature data
and ClimateBC temperature data from 1939 to 1960 (data not shown) indicated a possible
step change in temperatures around 1960, which presumably could be associated with the
change in station location and elevation.

At the time of this study, adjusted precipitation, rain, and snow data for the Vernon
Coldstream Ranch were only available up to April 1997. To create a time series that matched
the other records used in the analysis, we used observed data from a nearby station (Vernon
Bella Vista) to extend the record to 2005.
41

Climate station data are summarized in Table 2.3. Missing precipitation, rain, and snow data
in the Tatlayoko, Williams Lake, and Vernon stations over the analysis time period (19402004) are listed in Table 2.4. We did not replace missing precipitation data due to a high
potential for errors in estimating precipitation (Mekis and Hogg 1999). Missing data were
omitted pair-wise from analysis.

For the climatic gradient analysis (see below for details on this analysis), site-specific climate
time series data were generated using the ClimateBC model (Wang et al. 2006). ClimateBC
combines bi-linear interpolation and elevation adjustment techniques to downscale gridded
climate data from the PRISM model (Daly et al. 2002) to produce scale-free spatiotemporal
climate data covering western Canada (Spittlehouse 2006). The model produces climate time
series and historical/future normals for a specific geographic point, based on user-provided
latitude, longitude, and elevation inputs. Model outputs include monthly temperature and
precipitation variables, as well as annual derived variables such as frost-free period,
continentality, growing-degree days, and heat-moisture index. Spittlehouse (2006) discusses
ClimateBC model validation and limitations.

Seasonal and annual variables were derived by averaging climate station monthly values (for
the regional climate-growth relationships analysis) and ClimateBC monthly values (for the
climatic gradient analysis) (Table 2.5). Where monthly values were missing in the climate
station record (Table 2.4), the associated derived seasonal variable was computed as a nonnumber and omitted pair-wise from analysis. Growth was compared to climate variables

42

from the previous year's July to the current year's October because climate in the preceding
year affects the current year's growth (Fritts 1976).

Seasonal and annual heat - moisture index variables were calculated from derived
seasonal/annual temperature and precipitation data using the following formula (similar to
Wang et al. 2006):

1) HM = (TMX + 10)/(PPT/1000)
where: HM = seasonal or annual heat-moisture index
TMX = seasonal or annual mean of monthly average maximum temperatures (°C)
PPT = seasonal or annual mean of monthly precipitation totals (mm)

El Nino Southern Oscillation (ENSO) and Pacific Decadal Oscillation (PDO) are two quasiperiodic ocean-atmosphere processes (Allan 2000, Mantua et al. 1997) that influence climate
in British Columbia (Shabbar 2006, Shabbar et al. 1997, BC MoE 2007, BC MWLAP 2002)
and have been linked to Douglas-fir growth in the Pacific Northwest United States (US)
(Case and Peterson 2005). The Pacific-North American index (PNA) is a mid-tropospheric
pressure anomaly influenced by ENSO and strongly correlated with North American winter
temperature and precipitation (Leathers et al. 1991, Shabbar 2006, Shabbar et al. 1997).
Seasonal and annual ENSO, PDO, and PNA indices were derived from monthly indices
obtained from the US Department of Commerce's National Oceanic and Atmospheric
Administration (http://www.cdc.noaa.gov/ClimateIndices/List/) (Table 2.5).

43

2.2.3.2.2 Regional climate-growth relationships
Regional climate-growth relationships were quantified by computing Pearson simple
correlation coefficients between growth variation (as represented by the regional PCs) and
local climate station data (see below for a description of climate data), from 1940 to 2004 (65
years), a time period common to all climate stations. Growth variation was also correlated to
ENSO, PDO, and PNA indices over the full available records (see below). The objective of
this analysis was to identify important regional climate-growth limitations and to compare
these limitations between ecologically similar sites in all three study regions. Bootstrapped
bias-corrected accelerated (BCA) 95% confidence intervals were calculated for each
correlation coefficient using 1000 iterations (Moore and McCabe 2006, The Mathworks Inc.
2007).

Climate and tree-growth (PC) data were screened for normality using the Lilliefors test (The
Mathworks Inc. 2008, Meko 2007, Conover 1980) prior to climate-growth analysis, and
natural logarithm transforms were applied to data failing the Lilliefors test at the 95%
significance level (The Mathworks Inc. 2007). The Lilliefors test was then repeated to
ensure that data transformation was successful (this transformation was successful in all
cases).

2.2.3.2.3 Climatic gradient analysis
Variations in population-level Douglas-fir climate-growth relationships were analyzed along
a climatic gradient across all sites. This was done in two steps. First, regression and
correlation coefficients were calculated between the 33 chronologies and site-specific climate

44

data (e.g. annual precipitation) generated by the ClimateBC model, from 1940 to 2002 (63
years). We chose this time period to match the regional PC climate-growth relationship
analysis (historical ClimateBC data extends only to 2002, thus there was a difference of 2
years between the two analyses). In the second step of this analysis, the 33 regression and
correlation coefficients calculated for a given variable were regressed against site-specific
climate normals (1961-1990 period, Spittlehouse 2006) such as annual heat-moisture index
(AHM), MAT, MAP, and others, produced by the ClimateBC model (Table 2.2). In this
way, it was possible to determine if a site's climatic conditions could predict growth
association (represented by the correlation coefficient) or sensitivity (represented by the
regression coefficient) to local climate variability. A similar analysis was completed using
population-level responses to PDO and PNA winter indices. Chronology statistic variation
along a climatic gradient was also examined using similar methods. An example of how we
performed the climatic gradient analysis is presented in Appendix C.

2.3 Results
2.3.1 Chronologies
Final chronologies were built using from 15 to 31 trees, and chronology lengths ranged from
113 to 373 years (Table 2.1). All sites had relatively high average intrasite correlations
ranging from 0.352 to 0.695. The expressed population signal for all chronologies exceeded
0.85 (data not shown), indicating a sufficient population signal for climate analysis (Wigley
et al. 1984). Mean sensitivity ranged from 0.136 to 0.448 and standard deviation ranged
from 0.126 to 0.375.

45

Chronology mean sensitivity, intrasite correlations, and standard deviation increased linearly
with site AHM normals (Figure 2.5). When examined against the individual components of
AHM, chronology statistics had no relationship with temperature normals (MAT), and an
apparent non-linear relationship with precipitation normals (MAP), however, it should be
noted that MAP data were positively skewed. Similar results (data not shown) were found
when chronology statistics were regressed against summer heat-moisture index (SHM), and
its components (mean warmest month temperature and mean summer precipitation).
2.3.2 Multivariate analysis
2.3.2.1 Northern Interior sites
PCA of Northern Interior chronologies extracted three principal components (PCs) that
represented 75.4% of the total chronology variance (Table 2.7). The three retained PCs
explained 30.8%, 25.9%, and 18.7% of the variance, respectively.

Visual analysis of chronology-PC relationships revealed that PCI (hereinafter referred to as
'NIPC1', see Table 2.6) had the strongest correlations with the northernmost (north of
~54°N) and westernmost (west of -123.5°W) sites in this region (Figure 2.12). This
geographic pattern was confirmed by the high loadings of the Fort St. James sites and two
sites north of Prince George (Kerry and McLeod) on this PC (Table 2.7, Figure 2.2). When
the NIPC1-chronology correlation coefficients were plotted against climate normals, sites
with low mean annual temperature (MAT) and mean coldest month temperature (MCMT)
had a stronger relationship with this PC (Figure 2.12). NIPC1 also represented sites with
lower precipitation and higher heat-moisture indices. Based on these observations, NIPC1
was interpreted to best represent northernmost and westernmost sites with cold and dry

46

conditions; these sites are located near the northern margins of the species' range in the
province (Table 2.6).

The second PC (NIPC2) had a stronger correlation with the southernmost sites of the study
area (Figure 2.12). This was also confirmed by the high loadings of the Greggl 1, Gregg24,
and PG1412 sites on this PC (Table 2.7 and Figure 2.2). NICP2 had higher correlations to
sites with lower MAP, higher MAT, and lower precipitation as snow (PAS) (Figure 2.12).
This PC was interpreted to represent the warmer and drier (southernmost) sites in the
Northern Interior study region (Table 2.6).

The third PC (NIPC3) correlated most with the easternmost, high-elevation sites (Figure
2.12) located on the western slopes of the Rocky and Cariboo Mountains east of Prince
George (Figure 2.2). This PC was correlated to sites with low MAT and high
precipitation/snowfall (Figure 2.12). This was also confirmed by high loadings (Table 2.7)
with the three sites located in subalpine (Engelmann Spruce-Subalpine Fir BEC zone,
Meidinger and Pojar 1991) forests (Table 2.2). NIPC3 was interpreted to represent radial
growth variation for high-elevation sites that are at the wet and snowy extremes for Douglasfir in this region (Table 2.6).

2.3.2.2 Southern Interior sites
The Southern Interior chronologies had two modes of variability that represented 67.6% of
the total variance (Table 2.7). The two retained PCs explained 35.3% and 32.3% of the total
variance, respectively.

47

The first PC (SIPC1, Table 2.6) was best correlated to low-elevation sites in the study area
(Figure 2.13). Sites with high average temperatures and low precipitation correlated strongly
with SIPC1. This PC had the highest loadings with the hottest and driest sites (Figure 2.13),
and therefore was interpreted to represent Douglas-fir growing at its warm and dry margins
in the region (Table 2.6).

The second PC (SIPC2) was correlated mostly to high-elevation sites in the study area
(Figure 2.13). Radial growth variation represented by this PC best represented sites with
lower annual and summer temperatures, and higher precipitation (annual, summer and snow,
Figure 2.13). The wettest and snowiest sites in the region (Wap 1 and Wap 2) had the
strongest correlation with SIPC2 (Table 2.7). This PC seemed to represent similar ecological
conditions as NIPC3 from the Northern Interior study region, and was interpreted to
represent Douglas-fir at its upper elevation and precipitation extremes in this region (Table
2.6).

2.3.2.3 Chilcotin Plateau sites
The Chilcotin chronologies had two modes of variability that represented 79.5% of the
original chronologies' variance (Table 2.7). The first two principal components explained
43.2% and 36.2% of the variance, respectively.

This first PC (CHPC1, Table 2.6) was best correlated to the easternmost sites in the region
(Figures 2.3 and 2.14). The four driest sites of the region (and entire study, Table 2.2) had

48

strong correlations with this PC (Table 2.7 and Figure 2.14). Two of these sites (Hoodoo and
Gang) are located along the forest-grassland transition in the Bunchgrass BEC zone
(Meidinger and Pojar 1991); growth variation represented by this PC was interpreted to
represent Douglas-fir growing at the dry extremes in the region (Table 2.6).

The second PC (CHPC2) best represented the four westernmost sites (Figures 2.3 and 2.14).
This PC was strongly correlated to sites with higher mean annual precipitation totals and
lower annual heat-moisture index than sites represented by CHPC 1 (Figure 2.14).

2.3.3 Climate-growth relationships
Correlation and regression coefficients were calculated for all annual, seasonal, and monthly
variables, however, only annual and seasonal variables are presented because they explained
more growth variation than monthly variables (e.g., Littell 2006, Watson and Luckman
2002). Rather than present results for all seasonal and annual variables, we focused our
results and discussion on the annual and seasonal variables that appeared important in
explaining growth variation.

2.3.3.1 Regional climate-growth relationships
2.3.3.1.1 Northern Interior sites
Douglas-fir growth at cold sites along its northern range margins (as represented by NIPC1)
was positively correlated to spring rainfall and annual PNA index, and was positively
(negatively) correlated to spring rain (heat-moisture index) (Table 2.8). Warmer and drier
sites in the Northern Interior (NIPC2) tended to have high radial growth in years with

49

relatively high annual, prior summer, and spring precipitation and low annual, prior summer,
and spring heat-moisture indices. High-elevation sites in the Northern Interior (NIPC3) had a
climate-growth relationship distinct from other sites in the region. Radial growth in these
sites was positively related to annual and winter temperatures and negatively related to winter
snowfall. Winter PNA index was positively correlated to NIPC3 variation.

2.3.3.1.2 Southern Interior sites
Douglas-fir growth at warm and dry sites in the Southern Interior (SIPC1) had a positive
relationship to annual and spring precipitation. Maximum temperatures in the previous
summer and current spring tended to negatively influence growth on these sites. Heatmoisture index variables for annual, prior summer, and current spring were also negatively
correlated to SIPC1 growth.

Radial growth in high-elevation sites in the Southern Interior (SIPC2) tended to be higher
during warmer years. The strongest relationships to climate were a negative correlation to
winter snowfall and a positive correlation to winter temperatures. Rainfall in the previous
summer had a positive relationship to SIPC2 growth. These sites were the only ones of the
study with a significant relationship (positive) to annual, winter, and previous summer PDO
and ENSO indices. Annual and winter PNA indices were also positively correlated to SIPC2
growth.

50

2.3.3.1.3 Chilcotin Plateau sites
Both Chilcotin PCs (CHPC1 and CHPC2) had similar relationships to local climate. Growth
across the sampling area appeared to be strongly limited by annual precipitation and heatmoisture index. Growth represented by CHPC1 was also correlated to prior summer and
spring rain. CHPC2 growth was correlated to annual and prior winter mean temperatures.

2.3.3.2 Climatic gradient analysis
Correlation coefficients computed between each of the 33 ring-width chronologies and sitespecific annual precipitation (1940-2002) varied linearly along a climatic gradient defined by
site AHM (Figure 2.6) and SHM (Figure 2.7). Linear slopes of natural log-transformed (due
to skewness) regression coefficients vs. site AHM and SHM indicated that growth sensitivity
(represented by the regression coefficient) increased by approximately 16 and 7% per oneunit increase in AHM and SHM, respectively. Correlation and regression coefficients
computed between growth and spring/prior summer precipitation varied similarly but had
weaker linear relationships with heat-moisture indices than annual precipitation (data not
shown).

When plotted against a precipitation (MAP, MSP, and PAS) gradient, non-linear
relationships between the correlation and regression coefficients and site climate were
strongly apparent, with a response threshold evident at approximately 700 millimetres per
year (Figure 2.8, MSP and PAS not shown). MAP, MSP, and PAS data were positively
skewed, and log transformations resulted in significant linear relationships (e.g., Figure 2.9,
MSP and PAS not shown). Results using log-transformed MAP indicated that a 1% decrease

51

in MAP would increase a population's correlation coefficient with annual precipitation by
approximately 0.0032 units and would also increase the regression coefficient with annual
precipitation by approximately 3.3%.

No relationships were found when site-specific temperature normals such as MAT, mean
warmest month, mean coldest month, and continentality were used to define the climatic
gradient (data not shown). As well, no patterns were found with coefficients computed
between the chronologies and derived ClimateBC variables such as frost-free period and
growing-degree days. This was also found for coefficients computed between the
chronologies and temperature variables.

Correlation and regression coefficients computed between the chronologies and OctoberMarch PDO and PNA indices (1901-2005 and 1951-2005, respectively) varied linearly along
an AHM gradient (Figures 2.10 and 2.11). Sites with a low AHM (i.e., the coldest and/or
wettest sites) had the strongest (positive) relationship to both indices. Only three
chronologies were significantly correlated to ENSO, therefore, we did not complete a
climatic gradient analysis using ENSO index.

2.4 Discussion
2.4.1 Ecological growth patterns
Visual analysis of PC-chronology relationships indicated that in all three study regions,
Douglas-fir growth patterns differentiate primarily along precipitation and temperature
gradients, which are somewhat associated with elevation (Case and Peterson 2005, Korner

52

2007). This supports similar findings for Douglas-fir in coastal environments at smaller
spatial scales (Case and Peterson 2005, Zhang and Hebda 2004).

Leading modes of growth variation in each region tend to be dominated by sites located at
the climatic extremes, however, these modes also characterize trees growing in moderate
climatic conditions. For example, the first PC in the Southern Interior (SIPC1) was most
correlated to Douglas-fir growth on a site in the hot and dry Ponderosa Pine xeric-hot BEC
zone/subzone (Meidinger and Pojar 1991), but was also significantly correlated to Douglasfir growth in the Interior Cedar-Hemlock BEC zone (Meidinger and Pojar 1991), a much
wetter ecosystem. This pattern was also found in the visual plots of PC-chronology loadings.
Therefore, climatic factors that are significantly correlated to this PC (e.g., annual
precipitation) are likely most limiting to Douglas-fir growing at the hot and dry climatic
extremes, but may also be limiting (although to a lesser degree) in ecosystems with a more
moderate climate (Littell 2006).

2.4.2 Climate-growth relationships
2.4.2.1 Low- to mid-elevation dry sites
Principal components explaining the majority of variation in all three regions were generally
most correlated to precipitation and integrated precipitation - temperature (heat-moisture
index) variables. This suggests that moisture is likely the most limiting environmental factor
to Douglas-fir growth across most of its range in the BC Interior, and supports similar
findings for this species (e.g., Zhang and Hebda 2004, Case and Peterson 2005, Daniels and
Watson 2003, Littell 2006, Adams and Kolb 2005, Watson and Luckman 2002). Growing

53

season water deficits reduce Douglas-fir photosynthesis, gas exchange, and transpiration,
resulting in less cambial growth (Lopushinsky and Klock 1974, Lassoie and Salo 1981)
(possibly at the expense of root growth; Waring 1991), as this species has relatively poor
stomatal control during droughts (Lassoie and Salo 1981) and low sapwood water storage
capacity (Lopushinksy 1991). Water stress thresholds (and other climate-growth limitations)
may vary with age, microsite conditions, and genetics (Littell 2006).

Annual precipitation amounts generally resulted in higher correlations than seasonal and
monthly precipitation amounts, likely reflecting an integration of previous growing season,
winter, and spring precipitation influence on growth (Watson and Luckman 2002, Littell
2006, Case and Peterson 2005, Daniels and Watson 2003). Numerous physiological
mechanisms may explain this climate-growth relationship. For example, primordial tissues
are formed during the end of the growing season, and water availability during the late
summer influences carbohydrate allocation to the following year's formative tissues (Fritts
1976, Littell 2006, Kozlowski and Pallardy 1997). Needle drop as a result of late growingseason moisture stress may reduce photosynthetic surface area and result in lower growth the
following year (Zhang and Hebda 2004). Trees may allocate more carbohydrates to root
systems during the growing season after a dry summer to compensate for drought-reduced
root growth, resulting in a smaller annual ring (Lassoie and Salo 1981). Winter snow and
spring precipitation can be important because they provide early season soil moisture (Carrer
and Urbinati 2006), especially for Douglas-fir, as its needles tend to be fully grown and
photosynthesizing relatively early in the year (Watson and Luckman 2002).

54

The climatic gradient analysis showed that growth responses to annual precipitation, as well
as chronology statistics (mean sensitivity, standard deviation, and intrasite correlations), are
strongest in Douglas-fir growth in dry areas, and further showed how these responses vary
with site local conditions over the entire study domain. Climatic gradient analysis results
suggest that Douglas-fir growth will become strongly associated with, and more sensitive to,
annual precipitation where local climate becomes warmer and/or drier (i.e. annual heatmoisture index increases). Non-linear increases in growth-precipitation sensitivity along a
gradient defined by MAP suggest that if sites receive, on average, less than approximately
700 millimetres of annual precipitation, growth sensitivity to precipitation will increase very
rapidly. A decrease in available soil moisture in dry-climate sites could conceivably result in
significant productivity declines and potential mortality.

When predicting population growth responses to annual precipitation using a climatic
gradient analysis, using the interactions between site temperature and precipitation (i.e., heatmoisture index) to define the site climatic characteristics explained the most variation in
climate-growth responses and chronology statistics. Climatic gradients defined by logtransformed precipitation-only variables also significantly explained climate-growth response
variation (although less than heat-moisture indices), whereas temperature-only variables did
not significantly explain variation. Heat-moisture indices likely reflect the importance of soil
moisture to Douglas-fir growth by integrating moisture inputs (precipitation) and
evaporation/transpiration forcings (temperatures) (Wang et al. 2006). The integrative effects
of temperature-precipitation changes on forest ecosystems must be considered when adapting
forest management to climate change, as they may impact forest ecosystem functioning and

55

processes more than either temperature or precipitation dynamics alone (Littell and Peterson
2005).

Only the hottest sites in the study (Southern Interior, SIPC1) had negative growth
relationships to temperatures, with the exception of NIPC2 (negatively correlated to spring
maximum temperatures). As negative Douglas-fir growth relationships to temperatures have
been extensively documented in British Columbia and other regions (Case and Peterson
2005, Zhang and Hebda 2004, Laroque and Smith 2005, Feliksek and Wilczynski 2003,
Daniels and Watson 2003, Watson and Luckman 2001), we expected a somewhat broader
negative temperature response. Our results indicate that in hot and dry sites in the Southern
Interior, summer and spring temperatures may exceed species-specific photosynthesis
threshold limits (Brubaker 1980), which can be as low as -20-25 °C for Douglas-fir (Salo
1974, Doehlert and Walker 1981). As a comparison, Hadley (1969) found that ponderosa
pine (Pinus ponderosa Douglas ex C. Lawson) had optimal photosynthesis at temperatures of
25-35 °C. Although these thresholds are specific to their respective study areas, they do
suggest that Douglas-fir may have an upper temperature-growth threshold that is low relative
to other drought-resistant species (Brubaker 1980). An increase in temperature affects
respiration and transpiration processes (Balatinecz and Anderson 1986, Kimmins 1997) and
may reduce growth by increasing water loss, reducing nutrient storage, and decreasing
foliage efficiency (Laroque and Smith 2005, Kimmins 1997, Balatinecz and Anderson 1986).
Projected sustained temperature increases across the province (BC MoE 2007, Christensen et
al. 2007) may result in broader negative temperature-growth responses in Douglas-fir.

56

The climate-growth relationship analysis also revealed relatively uniform climatic limitations
to Douglas-fir growth on the Chilcotin Plateau. The eight sites sampled in the Chilcotin
Plateau were the driest of the entire study, and all appear to be strongly limited by annual
precipitation and heat-moisture index. Although the PCA indicated two significantly
different growth patterns in this study region, these growth patterns may simply reflect
geographic (longitudinal) differences in temporal precipitation patterns, and not different
climate-growth relationships.

2.4.2.2 High-elevation sites
High-elevation Douglas-fir populations have similar growth relationships in the Northern and
Southern Interior regions and appear to be unique from the mid- to low-elevation populations
in that low annual temperatures and winter snowfall are primary limiting factors. This result
is consistent with similar findings in coastal high-elevation Douglas-fir populations in BC
(Zhang and Hebda 2004) and the Pacific Northwest US (Case and Peterson 2005). In cold
environments with ample moisture supply, low temperatures are often assumed to be primary
climatic growth limitations (Littell 2006, Pfeifer et al. 2005), as they reduce glucose
utilization, C0 2 uptake, and photosynthesis (Korner 1998, Kimmins 1997, Grace et al. 2002),
can cause damage tree tissues (Korner 1998, Chhin et al. 2008), and may be associated with
shorter growing seasons (Zhang and Hebda 2004, Littell 2006, Grace et al. 2002). Colder
winter temperatures may also be associated with higher snowfall, which was negatively
correlated to growth in both the Northern and Southern Interior regions. Negative snowgrowth responses have also been found in other high-elevation Douglas-fir populations (Case
and Peterson 2005) as well as other tree species in western North America (Ettl and Peterson

57

1995, Peterson et al. 2002, Laroque and Smith 2005) and Siberia (Vaganov et al. 1999).
Higher snowfall and corresponding snowpacks can limit growth by delaying the start of the
growing season, as cambial and root growth generally only begin when the upper soil layers
have thawed after snowmelt (Vaganov et al. 1999, Emmingham and Waring 1977, Graumlich
and Brubaker 1986, Zhang and Hebda 2004). Wind and snow breakage may limit growth in
high-elevation environments, however, their effects on growth would be non-linear and
therefore likely not reflected in interannual growth patterns (Littell 2006).

Although growth appears to be most limited by temperatures and snowfall in high-elevation
environments, growth in both regions was also correlated to prior summer rain, which is
consistent with other cold-environment Douglas-fir populations (Case and Peterson 2005,
Laroque and Smith 2005, Feliksek and Wilczynski 2003). Trees growing in cold
environments may respond to drought events because these stressful periods further reduce
an already short growing season (Zhang and Hebda 2004).

Annual and seasonal PDO and ENSO indices were strongly correlated to Douglas-fir growth
at high-elevations in the Southern Interior; similar growth relationships have been observed
in Douglas-fir in the Pacific Northwest US (Case and Peterson 2005) and in other tree species
as well (Holman and Peterson 2006, D'Arrigo et al. 2001, Biondi et al. 1999). PDO and
ENSO signals did not appear in high-elevation sites in the Northern Interior, suggesting these
climatic processes may be regional in their influence in the British Columbia Interior. High
PDO and ENSO index values correlate to warm and dry weather patterns (Hare and Mantua
2001, Shabbar et al. 1997, Shabbar 2006) in British Columbia (BC MoE 2007, BC MWLAP
2002), which may result in longer growing seasons through lower snowpack depths and
58

warmer soils (Case and Peterson 2005). Winter PNA index was correlated to high-elevation
growth in both regions. PNA index values are correlated to surface winter and early spring
temperature and precipitation variation over much of North America (Leathers et al. 1991),
and may provide a good indication of high-elevation Douglas-fir growth over a large spatial
range.

The climatic gradient analysis showed that Douglas-fir growing in wet and cold
environments (those with a relatively low annual heat-moisture index) have stronger growth
relationships with PDO and PNA variation, and that these relationships weaken linearly as
sites become warmer and/or drier. Projected temperature increases and snowfall decreases
(Christensen et al. 2007, BC MoE 2007, BC MWLAP 2002) may result in an increase in site
heat-moisture index and a subsequent future weakening of the influence of PDO and PNA on
high-elevation Douglas-fir growth.

2.4.2.3 Sites at the northern geographic limits
The PCA indicated that Douglas-fir populations at the northern geographic margins are
unique from other populations in the Northern Interior region, however, these populations
have a somewhat similar climate-growth relationship as warmer sites in the North in that
growth appears to be most limited by spring precipitation. As low temperatures are assumed
to limit Douglas-fir range along its northern margins (Jull 1999, Delong 1999, Vyse et al.
1991), we expected a potentially strong temperature signal in NIPC1, however, the climate
influence on growth appears to be relatively low.

59

There are several possible reasons for relatively weak climate-growth correlations. Douglasfir along their northern limit may be limited by severe growing season frost events (Jull 1999,
Delong 1999, Vyse et al. 1991), which can damage mature needles and reduce radial growth
by inhibiting hormonal system functioning (Feliksek and Wilczyhski 2003). Growing season
frosts and other short-term climatic extremes may affect tree physiology more than average
climatic conditions in some environments (Graham et al. 1990, Innes 1994), but would not be
reflected in the monthly temperature data used for this study. Climate-growth relationships
in this environment may also be non-linear or vary over time.

2.4.3 Conclusions
As forest managers prepare for climate change, a better understanding of forest species
vulnerability to climate change (BC MoFR 2006) across large regions and climatic gradients
is required (Peterson and Peterson 2001, Case and Peterson 2005). Using high-resolution
spatiotemporal climate data from the ClimateBC model (Wang et al. 2006), we were able to
sample and develop 33 Douglas-fir growth chronologies covering a wide climatic and
geographic range in the interior of British Columbia. Across most of its range in the
province's interior, Douglas-fir annual growth appears to be chiefly limited by precipitation
totals from the previous July to current June, with the strongest sensitivity in population
growth at the dry margins. Even at the northern margins of the species' range near Fort St.
James, Douglas-fir growth appears to be mostly limited by precipitation, and we were unable
to detect a strong cold-temperature limitation on growth in these presumably cold-margin
populations. Future decreases in soil moisture availability as a result of increasing

60

temperatures (Christensen et al. 2007) may result in a widespread increase in precipitation
sensitivity and growth declines, with initial and strong changes in drier areas.
At the high-elevation margins of its range, Douglas-fir growth is more limited by annual and
winter temperatures and snowfall, and strong associations with PDO, ENSO, and PNA
indices may reflect the influence of Pacific Ocean - atmosphere climate systems on winter
conditions in mountainous environments. Warmer temperatures and less snowfall may
increase Douglas-fir productivity in these environments.

61

2.5 Literature cited
Aber, J., R.P. Neilson, S. McNulty, J.M. Lenihan, D. Bachelet, and R.J. Drapek. 2001. Forest
processes and global environmental change: predicting the effects of individual and
multiple stressors. BioScience 51: 735-752.
Adams, H.D. and T.E. Kolb. 2005. Tree growth response to drought and temperature in a
mountain landscape in northern Arizona, USA. Journal of Biogeography 32: 16291640.
Aitken, S.N., S. Yeaman, J.A. Holliday, T. Wang, and S. Curtis-McLane. 2008. Adaptation,
migration or extirpation: climate change outcomes for tree populations. Evolutionary
Applications 1: 95-111.
Allan, R.J. 2000. ENSO and climatic variability in the past 150 years, p. 3-56. In El Nino and
the Southern Oscillation: multiscale variability and global and regional impacts. Diaz,
H. F. and Markgraf, V. (eds). Cambridge University Press, New York.
Arno, S.F. 1991. Ecological relationships of interior Douglas-fir, p. 47-51. In Interior
Douglas-fir: The species and its management symposium proceedings. Baumgartner,
D. M. (ed). Department of Natural Resource Sciences, Washington State University,
Pullman, Washington.
Balatinecz, J.J. and H.W. Anderson. 1986. Effects of climatic factors on the formation of
wood and its quality, p. 195-200. In Climate Applications in Forest Renewal and
Forest Production, Proceedings of Forest Climate 1986. Maclver, D. C, Street, R. B.,
and Auclair, A. N. (eds). Canadian Forestry Service, Ottawa, Ontario.
Barber, V., G. Juday, and B. Finney. 2000. Reduced growth of Alaska white spruce in the
twentieth century from temperature-induced drought stress. Nature 405: 668-672.
BC MoE. 2007. Environmental Trends in British Columbia: 2007. State of Environment
Reporting. BC Ministry of Environment, Victoria, B.C.
BC MoFR. 2006. Preparing for climate change: Adapting to impacts on British Columbia's
forest and range resources. BC Ministry of Forests and Range, Victoria, BC.
BC MWLAP. 2002. Indicators of climate change for British Columbia, 2002. Indicators of
climate change for British Columbia, 2002. BC Ministry of Water, Land and Air
Protection, Victoria, B.C.
Biondi, F., D.R. Cayan, and W.H. Berger.1999. Decadal-scale changes in southern California
tree-ring records, p. pp. 303-306. In Proceedings of the 10th Symposium on Global
Change Studies. Karl, T. R. (eds). American Meteorological Society, Boston, MA.
Biondi, F. and Swetnam, T.W. 1987. Box-Jenkins Models of forest interior tree-ring
chronologies. Tree-Ring Bulletin 47: 71-96.

62

Biondi, F. and K. Waikul. 2004. Dendroclim2002: A C++ program for statistical calibration
of climate signals in tree-ring chronologies. Computers & Geosciences 30: 303-311.
Brubaker, L.B. 1980. Spatial patterns of tree growth anomalies in the Pacific Northwest.
Ecology 6: 798-807.
Carrer, M. and C. Urbinati. 2006. Long-term change in the sensitivity of tree-ring growth to
climate forcing in Larix decidua. New Phytologist 170: 861-872.
Case, M.J. and D.L. Peterson. 2005. Fine-scale variability in growth-climate relationships of
Douglas-fir, North Cascade Range, Washington. Canadian Journal of Forest Research
35: 2743-2755.
Cayan, D.R., M.D. Dettinger, H.F. Diaz, and N.E. Graham. 1998. Decadal variation of
precipitation over Western Northern America. Journal of Climate 11:3148-3166.
Chhin, S., E.H. Hogg, V.J. Lieffers, and S. Huang. 2008. Influences of climate on the radial
growth of lodgepole pine in Alberta. Botany 86: 167-178.
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. Sarr, and P. Whetton. 2007. Regional Climate Projections, p. 848-940.
In Climate Change 2007: The Physical Science Basis. Contribution of Working
Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate
Change. Solomon, S. D., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K. B.,
Tignor, M., and Miller, H. M. (eds). Cambridge University Press, New York.
Conover, W. 1980. Practical Nonparametric Statistics. John Wiley and Sons, New York.
Cook, E. 1985. A time series analsysis approach to tree-ring standardization. PhD
Dissertation. University of Arizona. Tucson, Arizona.
Cook, E. and L. Kairiukstis. 1990. Methods of dendrochronology: Applications in the
environmental sciences. Kluwer Academic Publishers, Dordrecht, Netherlands.
Cook, E.R. and P. Krusic. 2005. Users Manual for Program ARSTAN. Tree-Ring Lab,
Lamont-Doherty Earth Observatory, Columbia University, New York.
http://www.ldeo.columbia.edu/res/fac/trl/public/publicSoftware.html.
D'Arrigo, R., R. Villalba, and G. Wiles. 2001. Tree-ring estimates of Pacific decadal climate
variability. Climate Dynamics 18: 219-224.
D'Arrigo, R., R. Wilson, B. Liepert, and P. Cherubini. 2007. On the 'Divergence Problem' in
Northern Forests: A review of the tree-ring evidence and possible causes. Global and
Planetary Change 289-305.
Dai, A., K.E. Trenberth, and T.R. Karl. 1999. Effects of clouds, soil moisture, precipitation,
and water vapor on diurnal temperature range. Journal of Climate 12: 2451-2473.

63

Daly, C , W.P. Gibson, G.H. Taylor, G.L. Johnson, and P. Pasteris. 2002. A knowledgebased approach to the statistical mapping of climate. Climate Research 22: 99-113.
Daniels, L.D. and E. Watson. 2003. Climate-fire-vegetation interactions in the Cariboo
forests: a dendrochronological analysis. Report to Forest Innovation and Investment.
Forest Innovation and Investment Research Program, Vancouver, BC.
Davis, M.B. and Shaw. R.G. 2001. Range shifts and adaptive responses to Quaternary
climate change. Science 292: 673-679.
Delong, S.C. 1999. Ecology of Douglas-fir at its northern limits, p. 27-28. In Ecology and
management of interior Douglas-fir {Pseudotsuga menziesii var glaucd) at the
northern extreme of its range. Proceedings of a workshop held 7-9 October 1996 in
Fort St. James, British Columbia. Lousier, J. D. and Kessler, W. B. (eds). University
of Northern British Columbia, Prince George, BC.
Doehlert, D.C. and R.B. Walker. 1981. Photosynthesis and photorespiration in Douglas-fir as
influenced by irradiance, CO2 concentration, and temperature. Forest Science 27:
641-650.
Douglas, M.E. 1993. Analysis of sexual dimorphism in an endangered Cyprinid fish {Gila
cypha Miller) using video image technology. Copeia 1993: 334-343.
Easterling, D.R., G.A. Meehl, C. Parmesan, S.A. Changnon, T.R. Karl, and L.O. Mearns.
2000. Climate extremes: observations, modeling, and impacts. Science 289: 20682074.
Emmingham, W.H. and R.H. Waring. 1977. An index of photosynthesis for comparing forest
sites in western Washington. Forest Science 7: 165-174.
Ettl, G.J. and D.L. Peterson. 1995. Growth response of subalpine fir {Abies lasiocarpa) to
climate in the Olympic Mountains, Washington, USA. Global Change Biology 1:
213-230.
Feliksek, E. and S. Wilczyhski. 2003. Dendroecological characterization of Douglas-fir
{Pseudotsuga menziesii Franco) in the Wielkopolska region. Electronic Journal of
Polish Agricultural Universities 6: Series: Forestry.
Fritts, H.C. 1976. Tree rings and climate. The Blackburn Press, New Jersey.
Gates, D.M. 1990. Climate change and forests. Tree Physiology 7: 1-5.
Grace, J., F. Berninger, and N. Laszlo. 2002. Impacts of climate change on the tree line.
Annals of Botany 90: 537-544.
Graham, R.L., M.G. Turner, and V.H. Dale. 1990. How increasing C02and climate change
affect forests. Bioscience 40: 575-587.

64

Graumlich, L.J. 1991. Subalpine tree growth, climate, and increasing CO2: An assessment
of recent growth trends. Ecology 72: 1-11.
Graumlich, L.J. and L.B. Brubaker. 1986. Reconstruction of annual temperature (1590-1979)
for Longmire, WA, derived from tree rings. Quaternary Research 25: 223-234.
Hadley, E.B. 1969. Physiological ecology of Pinusponderosa in southwestern North
Dakota. American Midland Naturalist 81: 289-314.
Hamann, A. and T. Wang. 2006. Potential effects of climate change on ecosystem and tree
species distribution in British Columbia. Ecology 87: 2773-2786.
Hansen, A.J., R.P. Neilson, Dale. V.H., C.H. Flather, L.R. Iverson, D.J. Currie, S. Shafer, R.
Cook, and P.J. Bartlein. 2001. Global change in forests: responses of species,
communities, andbiomes. BioScience 51: 765-779.
Hanson, P.J. and J.F. Weltzin. 2000. Drought disturbance from climate change: response of
United States forests. The Science of the Total Environment 262: 205-220.
Hare, S.R. and N.J. Mantua. 2001. An historical narrative on the Pacific Decadal Oscillation,
interdecadal climate variability and ecosystem impacts. Report of a talk presented at
the 20th NE Pacific Pink and Chum workshop, Seattle, WA, 22 March 2001.
http ://www. iphc .Washington. edu/Staff/hare/html/papers/pcworkshop/pcworkshop .pdf.
Hermann, R.K. and D.P. Lavender. 1990. Douglas-fir. In Silvics of North America: 1.
Conifers; 2. Hardwoods. Agriculture Handbook 654. Burns, R. M. and Honkala, B. H.
(eds). U.S. Department of Agriculture, Washington, DC.
http://www.na.fs.fed.us/spfo/pubs/silvics_manual/Volume_l/pseudotsuga/menziesii.h
tm.
Holman, MX. and D.L. Peterson. 2006. Spatial and temporal variability in forest growth in
the Olympic Mountains, Washington: sensitivity to climatic variability. Canadian
Journal of Forest Research 36: 92-104.
Holmes, R. 1983. Computer-assisted quality control in tree-ring dating and measuring. TreeRing Bulletin 43: 69-78.
Innes, J.L. 1994. Climatic sensitivity of temperate forests. Environmental Pollution 83: 237243.
Josza, L. 1988. Increment core sampling techniques for high quality cores. Special
Publication SP-30. Forintek Canada Corp., Vancouver, BC.

65

Jull, M.J. 1999. Douglas-fir silviculture "on the edge": silvicultural systems at the northern
range of the species, p. 34-44. In Ecology and management of interior Douglas-fir
{Pseudotsuga menziesii var glauca) at the northern extreme of its range. Proceedings
of a workshop held 7-9 October 1996 in Fort St. James, British Columbia. Lousier, J.
D. and Kessler, W. B. (eds). University of Northern British Columbia, Prince George,
BC.
Karl. T.R., Jones. P.D., R.W. Knight, G. Kukla, N. Plummer, V. Razuvayev, K.P. Gallo, J.
Lindsaey, R.J. Charlson, and T.C. Peterson. 1993. A new perspective on recent global
warming: asymmetric trends of daily maximum and minimum temperature. Bulletin
of the American Meteorological Society 74: 1007-1023.
Kimmins, J. P. 1997. Forest Ecology. Prentice-Hall Inc., New Jersey.
Knowles, N., M.D. Dettinger, and D.R. Cayan. 2006. Trends in snowfall versus rainfall in the
Western United States. Journal of Climate 19: 4545-4559.
Kozlowski, T. T. and Pallardy, S. G. 1997. Growth control in woody plants. Academic Press,
San Diego.
Korner, C. 1998. A re-assessment of high elevation treeline positions and their explanation.
Oecologia 155: 445-459.
Korner, C. 2007. The use of'altitude' in ecological research. TRENDS in Ecology and
Evolution 22: 569-574.
Laroque, S.J. and D.J. Smith. 2005. A dendroclimatological reconstruction of climate since
AD 1700 in the Mt. Waddington area, British Columbia Coast Mountains, Canada.
Dendrochronologia 22: 93-106.
Lassoie, J.P. and D.J. Salo. 1981. Physiological response of large Douglas-fir to natural and
induced soil water deficits. Plant physiology and biochemistry 11: 139-144.
Leathers, D.J., B. Yaraal, and M.A. Palecki. 1991. The Pacific/North American
Teleconnection Pattern and United State Climate. Part I: Regional Temperature and
Precipitation Associations. Journal of Climate 4: 517-528.
Leung, L.R., Y. Qian, X. Bian, W.M. Washington, J. Han, and J.O. Roads . 2004. Midcentury ensemble regional climate change scenarios for the western United States.
Climatic Change 62: 75-113.
Littell, J. S. 2006. Climate impacts to forest ecosystem processes: Douglas-fir growth in
north-western U.S. mountain landscapes and area burned by wildfire in western U.S.
eco-provinces. PhD Dissertation. University of Washington. Seattle, WA.
Littell, J.S. and D.L. Peterson. 2005. A method for estimating vulnerability of Douglas-fir
growth to climate change in the northwestern U.S. The Forestry Chronicle 81: 369374.

66

Lloyd, A.H. 2005. Ecological histories from Alaskan tree lines provide insight into future
change. Ecology 86: 1687-1695.
Lloyd, A.H. and C.L. Fastie. 2002. Spatial and temporal variability in the growth and climate
response of treeline trees in Alaska. Climatic Change 52: 481-509.
Loehle, C. 2000. Forest ecotone response to climate change: sensitivity to temperature
response functional forms. Canadian Journal of Forest Research 30: 1632-1645.
Lopushinksy, W. 1991. Water relations of interior Douglas-fir, p. 35-46. In Interior Douglasfir: The species and its management symposium proceedings. Baumgartner, D. M.
(ed). Department of Natural Resource Sciences, Washington State University,
Pullman, Washington.
Lopushinksy, W. and G.O. Klock. 1974. Transpiration of conifer seedlings in relation to soil
water potential. Forest Science 20: 181-186.
Mantua, N.J. and S.R. Hare. 2002. The Pacific Decadal Oscillation. Journal of Oceanography
58: 35-44.
Mantua, N.J., S.R. Hare, Y. Zhang, J.M. Wallace, and R.C. Francis. 1997. A Pacific
interdecadal climate oscillation with impacts on salmon production. Bulletin of the
American Meteorological Society 78: 1069-1079.
Meehl, G.A., W.M. Washington, W.D. Collins, J.M. Arblaster, A. Hu, L.E. Buja, W.G.
Strand, and H. Teng. 2005. How much more global warming and sea level rise?
Science 307: 1769-1772.
Meidinger, D. and J. Pojar. 1991. Ecosystems of British Columbia. BC Ministry of Forests,
Victoria, B.C.
Mekis, E. and W.D. Hogg . 1999. Rehabilitation and analysis of Canadian daily precipitation
time series. Atmosphere-Ocean 37: 53-85.
Meko, D. M. 2007. GEOS 585A: Applied Time Series Analysis. Page accessed: 2007. URL:
http://www.ltrr.arizona.edu/~dmeko/geos585a.html
Mitchell, T.D. and P.D. Jones. 2005. An improved method of constructing a database of
monthly climate observations and associated high-resolution grids. International
Journal of Climatology 25: 693-712.
Moore, D. S. and McCabe, G. P. 2006. Introduction to the Practice of Statistics. W.H.
Freeman and Co., New York, N.Y.
Neilson, R.P. 1993. Transient ecotone response to climatic change: some conceptual and
modelling approaches. Ecological Applications 3: 385-395.

67

Parmesan, C. and G. Yohe. 2003. A globally coherent fingerprint of climate change impacts
across natural systems. Nature 421: 37-42.
Peterson, D.W. and D.L. Peterson. 2001. Mountain hemlock growth responds to climatic
variability at annual and decadal time scales. Ecology 82: 3330-3345.
Peterson, D.W., D.L. Peterson, and G.J. Ettl. 2002. Growth responses of subalpine fir to
climatic variability in the Pacific Northwest. Canadian Journal of Forest Research 32:
1503-1517.
Pfeifer, K., W. Kofler, and W. Oberhuber. 2005. Climate related causes of distinct radial
growth reductions in Pinus cembra during the last 200 yr. Vegetation History and
Archaeobotany 14:211 -220.
Regent Instruments Inc. 2005. WinDENDRO™. An image analysis system for tree-rings
analysis. Regent Instruments Inc. Quebec
Rising, J.D. and K.M. Somers. 1989. The measurement of overall body size in birds. The
Auk 106: 666-674.
Salo, D. J. 1974. Factors affecting photosynthesis in Douglas-fir. PhD Dissertation.
University of Washington. Seattle, WA.
Shabbar, A. 2006. The impact of El Nino-Southern Oscillation on the Canadian climate.
Advances in Geosciences 6: 149-153.
Shabbar, A., B. Bonsai, and M. Khandekar. 1997. Canadian Precipitation Patterns Associated
with the Southern Oscillation. Journal of Climate 10: 3016-3027.
Spittlehouse, D. 2006. ClimateBC: Your access to interpolated climate data for BC.
Streamline Watershed Management Bulletin 9: 16-21.
Spittlehouse, D.L. and R.B. Stewart. 2003. Adaptation to climate change in forest
management. BC Journal of Ecosystem Management 4: 1-11.
Spittlehouse, D. 2003. Water availability, climate change and growth of Douglas-fir in the
Georgia Basin. Canadian Water Resources Journal 28: 673-688.
Stephenson, N.L. and P.J. van Mantgem. 2005. Forest turnover rates follow global and
regional patterns of productivity. Ecology Letters 8: 524-531.
Tabachnick, B. G. and Fidell, L. S. 1989. Using Multivariate Statistics. Harper Collins
Publishers, Inc., New York.
The Mathworks Inc. 2007. Matlab 2007. Version 2007b. The Mathworks Inc.
http://www.mathworks.com
Trenberth, K.E. 1990. Recent observed interdecadal climate changes in the Northern
Hemisphere. Bulletin of the American Meteorological Society 71: 988-993.
68

Vaganov, E.A., M.K. Hughes, A.V. Kirdyanov, F.H. Schweingruber, and P.P. Silkin. 1999.
Influence of snowfall and melt timing on tree growth in subarctic Eurasia. Nature
400: 149-151.
Vincent, L.A. and D.W. Gullett. 1999. Canadian historical and homogenous temperature
datasets for climate change analyses. International Journal of Climatology 19: 13751388.
Vincent, L.A., X. Zhang, B.R. Bonsai, and W.D. Hogg . 2002. Homogenization of daily
temperatures over Canada. Journal of Climate 15: 1322-1334.
Vyse, A., R.A. Smith, and B.G. Bondar. 1991. Management of interior Douglas-fir stands in
British Columbia: past, present, and future, p. 177-185. In Interior Douglas-fir: The
species and its management symposium proceedings.Department of Natural Resource
Sciences, Washington State University, Pullman, Washington.
Walther, G.-R. 2003. Plants in a warmer world. Perspectives in Plant Ecology, Evolution,
and Systematics 6: 169-185.
Walther, G.-R., E. Post, P. Convey, A. Menzei, C. Parmesan, T.J.C. Beebee, J.-M.
Fromentin, O. Hoegh-Guldberg, and F. Bairlein. 2002. Ecological responses to recent
climate change. Nature 416: 389-395.
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.
Waring, R.H. 1991. Responses of evergreen trees to multiple stresses, p. 371-390. In
Response of plants to multiple stresses. Mooney, H. A., Winner, W. E., and Pell, E. J.
(eds). Academic Press, San Diego.
Watson, E. and B.H. Luckman. 2002. The dendroclimatic signal in Douglas-fir and
ponderosa pine tree-ring chronologies from the southern Canadian Cordillera.
Canadian Journal of Forest Research 32: 1858-1874.
Wigley, T.M.L., P.D. Jones, and K.R. Briffa. 1984. On the average value of correlated time
series, with applications in dendroclimatology and hydrometeorology. Journal of
Climate and Applied Meteorology 23: 201-213.
Zhang, Q.-B. and R.J. Hebda. 2004. Variation in radial growth patterns of Pseudotsuga
menziesii on the central coast of British Columbia, Canada. Canadian Journal of
Forest Research 34: 1946-1954.
Zhang, X., F.W. Zwiers, G.C. Hegerl, F.H. Lambert, N.P. Gillett, S. Solomon, P.A. Stott, and
T. Nozawa. 2007. Detection of human influence on twentieth-century precipitation
trends. Nature 448: 461-465.

69

JO
'o
o
•—

I

C3

CO t-~ SO t—

^ H ^ t © m 0 0 O s © ' * J - O s 0 0 i n s o r - ~ O s 0 0 a s s o r - - C - - C N

so ^ r- © ro ^ —< ^- i n e n t-~ T T
r-oooororororororosororoinossocNcNcninoo
t n t n m N N n N N
O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O

o o ' 0 0 ' o o ' o o ' o o o o o ' o o o o o o ' o o o ' o

00

o

tH

s o c N " s o < 3 _ i n r o i ~ - - > - H r i - r o s © i n r o s o c N ' 3 " © s o r o s o ~ H o o © i n i n © s o o o c N r - r - - r o _o
^ © ^ ^ ^ O s © ^ s o r - - - i > © ' - ^ O s o s o s ^ - ' S D r o r o r - - ' O O i n i n - < 3 - r o - > 3 - r - - - c N i n s o ' r t - o s "o
c
o
O O O O O O O O O O

roinoCTt-^H'—'~-|OsTt-sooom''3"ro©cNoor-->-"©r'-inassooo.-H^rorooosor--©
M t ^ ^ - T r M ^ m o o t ^ i n o i t - oo©cNCN'*t-inr--inininosinsooO'^©inrorooo~H

ooo'o'ooooo'ooooooooooooooooooo'ooooo

so 0 0

00

< — 1

^r

• < * •

0
00
00

00

rs (N
r- ro SO

•*r
OS

r\|

r-

os

CN
00

•

00
<S1 so

0

1

1

1

1

1

1

1

1

r-

00
00

00

i—1

00

• *

• *

00

1

OS r o
OS r o
00

1

SO
00

1

00
SO
00

O

1

•

t-~-

00
1

OO 0 0

i/»

00

so so so so so so so SO SO SO SO i n s o i n i n i n i n i n i n i n i n
0 0 0 0 0 0 0
O O O O O O O O O O
0 0 0 0
0 0 0 0 0 0 0
O O O O O O O O O O
0 0 0 0
( S M ( S ( S N (N (N (N (N (N (N C N C N C N C N C N C N C N C N C N r N

1

SO

1

00

1

OO

00

00

1

1

00

1

1

OO OO OO 0 0

1

• *

00

1

(SM(S(S(N(N(SNN(S(S

O O O O O O

i ^ i n ^ 0 i n v i v i i On "On "O0 O
" AO^ "On

1

i-~ O N 0 0 r o r o OS i n i n r^ 0
r o 0 0 0 0 r o so 0 0 rs i n Os CN r o r o
SO <l)
<L>
,-* r- r o r o r- r- r o (SON rr-o '^
0 0 CN r~
CN r\| so 0 0 0
W) SO
>; i n r o rr- T t r o V~i •*r SO CN
r- ^r ( N ^ CM SO ^^ *
—*
CN CN
>->
^H
r- ^r r o r o i n O r o r o Os OS r o O Os i n 0
•*
1 0 src r o 0 0 i n r o so OO SO 0 0 r o 0 0 OS OO SO i n

00

_o
"o
c
o
43

U

( N t S N t s | ( S N m r ) ( S ( S ( S | ( N ( N N r - i ( N ( S

T f ^ ^ ^ © ^ c N r o r - - - " r o i / 0 ' - * ' - - < o 1o r o r o v i r o s O ( N O - - H S o r - - - ' s i - s D i r > i / o o o s ^ ! - > / o

N ( S " N N M N N N M M N M M r t r s

•^Osoooooooor-~Osc-~oooo^H

O
o
o
o
O O O
O
O
o
o
o m © ^ © •*J- CN -si- O
M O O
O O O 0 0 0
O
i n © © © © © © © © i n © i n
O OO O O CN O
CN ro r- in ro ro CN o
^ o <* ro CN ^r x00 cN
C N S O S O S O ^ O O ' ^ - O O ' - ^ ' ^ - C N 00
"j m « m « rH « Ol h
00 r~
<-> so r -

00

S5

OJ

CN

•*" 1 - "
CN CN

— CN O0 O ,—< so 00 >n in
r-~ CN so i n >—, s o o o o s o o i n O r o s O ' ^ ^ j - r - OS CN OO CN Os O
in 00 10 o •—; rq ro ^ in
i n r o r o r o - ^ ro1 —; ^f "*f CN CN Os
« ^ « so t-~ m > t ^ t i - > t ^ ' M * r i m n 6 —< 00 OS OS OS —» "!f OS OO OS OS OS 00' 00 00 Os
C N C N C N C N C N C N C N C N C N C N C N C N CN CN CN
W
CN CN CN
CN

ion

uuo

OH

2

ill

0 0 3 ;

<u

« — -*_*,<
PQ Q H,

III

PH

CN
• *

at
C/3

3
O

0
HJ

d
f
«
CN
r
o
<"
0>
«)
a
0 a
kH
B" &1 &1 "3
u
O
J5
rt
fl
fl
HH
c/3 0^0
a, 0 H Jsi
3 55 :& ^ E£ ^ J on
LV,

M

u •z Z 2 Z Z Z Z Z Z Z Z Z Z Z Z Z « I M I « K « W O I ! « W

<

s

CN ro CN CN SO ro CN •t ro so C N Tt o in t-~ 00 ^-1 00 O ~H r- r- o in o •* r- in in so so os - H *o
so r- in rR
<-i t-- 00 OS Os O SO
OS 00 Os SO SO O so r-; os in in in SO >-i O S <-< 00ro 00 00 00 CN so
CN «
~
CN ^ in in in in ro TT ro ro ro ^- CO ^" "*' ro -sf ro
^ o o o oO O O O ©
in in in
in in
in in in in in in in in in in in in
in in in in in^ a
in in in in in
II >
<u
so - 0
c -rt
1-,

33 33 3333 33 33 33

CN

ro CN CN

S3

3E

eries c orr

o
-O

a
o

a
o
"5b
o

o

T3

a;

^

e
o
CN

O

CN

a

•J
ca
H

H

I 2; H

ica

o
K
O
•3 00 -a

Que

00
CN
u
c/3

< a 33

X §§

inc

O

Table 2.2. Sample site climate normals.
Site

BEC

MAT

MWMT

MCMT

TD

MAP

MSP

AHM

SHM

PAS

IDF dk4
36.8
Alexis
2.3
13.7
-9.9
201
68.2
106
23.6 335
Arch
SBS dw3
-11.2
19.6
1.8
13.7
24.9 604
255
53.6
266
-11.2
Battleship
SBS dw3
2.1
13.9
512
229
23.7
25.1
60.6
213
Bison
SBS dw3
13.7
-11.1
255
19.7
264
1.9
24.8 603
53.7
Bowron
ICH wk4
14.6
-9.9
804
346
16.2
42.1
3.1
24.5
304
Dolphin
SBSmkl
13.3
-11.3
663
270
17.3
1.5
24.6
49.2
307
SBS wkl
14.8
-8.9
389
15.7
38.1
Ferndale
3.5
23.7 861
298
Gang
BGxw2
16.4
-8.7
211
39.5
4.7
25.1 372
77.9
105
SBS dw3
14.6
-10.1
582
22.4
Gregg 11
3
24.7
255
57.3
226
Gregg24
SBS dw3
14.3
-10.1
24.4
260
21.6
55
2.8
593
235
Hedrick
ESSF wk2
-9.8
8.4
1.7
12.7
22.6
1386 525
637
24.3
BGxw2
Hoodoo
4.2
15.9
-8.9
202
41.8
78.4
94
24.7 341
Ispah
SBS wkl
14.1
-9.5
332
17.6
2.9
23.6 733
42.3
273
-11.2
Kerry
SBS wkl
2.2
14
279
16.1
25.3 755
50.3
339
LaCasa
PPxhl
-4.5
394
7.3
18.7
23.2
173
43.9
108.3 92
-10
Martin
SBPS xc
2
12.9
23
421
178
28.6
72.6
185
-10.3
202
23.4
Martin2
SBPS xc
1.5
12.3
22.7 493
61
226
SBSmkl
2.1
-12
252
17
Mcleod
14.3
26.3 713
56.8
331
ESSF wkl
Narrow
1.9
12.6
-9.6
22.2 992
443
12
28.4
400
IDF dk4
-9.7
192
Newton
2.4
13.7
23.4 338
36.8
71.3
113
ESSF wk2
-10.4
1064 386
35
Parsnip
1.8
13.5
23.9
11.1
477
PG1412
SBS dw3
3.5
14.6
-9.1
309
19.8
47.2
242
23.7 680
-11.2
Pinchi
SBS dw3
2
13.7
240
21.8
24.9 549
57.2
235
Queest
ICH mw3
-6.6
5.6
17
23.6 917
368
17
46.2
273
ICH mkl
-7.3
22.4
Shorts
3.8
14.8
22.1 615
255
58
236
ICH mw2
Sicamous
7.4
18.9
-5
24
848
379
20.5
50
169
Star
ICH mkl
2.4
-8.5
13.5
22
867
325
14.3
41.5
410
Tat
IDFdw
2.8
12.4
-8
20.4 461
183
27.6
68
185
Tat2
IDFdw
174
3
12.9
-7.8
28.9
74.2
179
20.7 452
ICH vkl
-9.4
Wapl
2.5
13.9
23.3 2077 676
6
983
20.5
ICH vkl
14.1
-9.3
Wap2
2.7
23.4 2062 680
6.1
20.7
952
Wap3
ICH wkl
15
-8.3
10.8
3.6
23.3
1255 483
30.9
488
Whiteman IDF mw3
5.8
-5.9
217
31.6
78.4
17
22.9 499
145
Note: Abiotic ecological descriptors are for climate normal period of 1961-1990 (Wang et al. 2006).
BEC=biogeoclimatic zone/subzone/variant (Meidinger and Pojar 1991). MAT=mean annual temperature (°C).
MWMT=mean warmest month temperature (°C). MCMT=mean coldest month temperature (°C).
TD=continentality (°C). MAP=mean annual precipitation (mm). MSP=mean summer precipitation (mm).
AHM=annual heat-moisture index. SHM=summer heat-moisture index. PAS=precipitation as snow (cm).

71

Table 2.3. Summary of meteorological records used in Chapter 2.
Station name
Latitude Longitude Elevation Station
(°W)
(°N)
(m)
ID
Northern Interior
Fort St. James
54.45
124.25
686
1092970
Prince George
53.89
122.68
691
1096450

Data length
(years)
110(1895-2004)
94 (1913-2006)

Southern Interior
Vernon Coldstream
Ranch
Vernon Bella Vista*

50.22
50.26

119.2
119.31

482
427

1128580 98(1900-1997)
1128553 21(1985-2005)

Chilcotin
Williams Lake**
Tatlayoko

52.18
51.67

122.05
124.41

940
870

1098940 69 (1939-2007)
1088010 76 (1930-2005)

* Vemon Bella Vista rain, snow, and precipitation data from April 1997 to December 2005 added to Vernon
Coldstream Ranch data to extend time series.
** ClimateBC elevation-adjusted and interpolated homogenized historical temperature data (Wang et al. 2006,
Mitchell and Jones 2005) from 1939-1960 used to account for possible inhomogeneities in Williams Lake
observed data.

Table 2.4. Missing precipitation data in meteorological records (between 1940 and 2004).
Station name
year
month
Variable(s)
Vernon Coldstream Ranch
1963
August
Rain, snow, PPT
Williams Lake A
1947
September
Rain, snow, PPT
October
Rain, snow, PPT
1960
June
Rain, snow, PPT
1979
October
Rain, snow, PPT
November
Rain, snow, PPT
Tatlayoko Lake
1977
April
Rain, snow, PPT
May
Rain, snow, PPT
1985
June
Rain, PPT
1998
May
Rain, snow, PPT
October
Rain, snow, PPT
November
Rain, snow, PPT
December
Rain, snow, PPT
1999
January
Rain, snow, PPT
February
Rain, snow, PPT
March
Rain, snow, PPT
April
Rain, snow, PPT
May
Rain, snow, PPT
June
Rain, snow, PPT
Rain, snow, PPT
July
August
Rain, snow, PPT

72

Table 2.5. Monthly climate values used to derive annual and seasonal variables used for
correlation analysis.
Variable
Monthly inputs
Annual
Prior July to current June
Prior summer
Prior July to prior September
Winter
Prior October to current March
Spring
April to June
Summer
July to September
Table 2.6. Principal components used for climate-growth analysis.
Region PC Number Identifier
Ecological conditions
Northern
Interior

Southern
Interior

Chilcotin

1

Corresponding
climate station

NIPC1

Northernmost sites in the province;
characterized by low MAT and
relatively low MAP. These are among
the coldest sites of the study.

Fort St. James

NIPC2

Low-elevation sites near Prince
George; characterized by relatively
warm and dry conditions.

Prince George

NIPC3

High-elevation sites near Prince
George on the western slopes of the
Rocky and Cariboo Mountains;
characterized by cold temperatures and
high precipitation/snowfall.

Prince George

SIPC1

Low-elevation sites near Vernon;
characterized by hot and dry
conditions. These are the hottest sites
of the entire study.

Vernon

SIPC2

High-elevation sites near Vernon and
Revelstoke; characterized by cold
temperatures and high
precipitation/snowfall. These are the
snowiest sites of the entire study.

Vernon

1

CHPC1

Eastern sites on the Chilcotin Plateau
near Williams Lake; characterized by
very low precipitation. These are the
driest sites of the entire study.

Williams Lake

2

CHPC2

1

Western sites on the Chilcotin Plateau
Tatlayoko Lake
near Tatlayoko Lake; characterized by
cold temperatures and slightly higher
precipitation than CHPC1. These are
among the coldest sites of the entire
study.
Note: Ecological conditions associated with each PC are based on visual interpretation of chronology-PC
loadings.

73

Table 2.7. Principal components analysis summary.
Northern Interior
Component
Eigenvalues
Variance
(%)
Cumulative
variance(%)

Southern Interior

Chilcotin Plateau

1
9.21

2
1.56

3
1.29

1
3.98

2
2.11

1
4.92

2
1.44

57.6

9.7

8.1

44.2

23.4

61.5

17.9

57.6

67.3

75.4

44.2

67.6

61.5

79.5

Rotation sums of squared loadings
Eigenvalues
Variance
(%)
Cumulative
variance (%)

4.93

4.14

2.99

3.18

2.91

3.46

2.9

30.8

25.9

18.7

35.3

32.3

43.2

36.3

30.8

56.7

75.4

35.3

67.6

43.2

79.5

0.868
0.797
0.899
0.261
0.747
0.327
0.47
0.391
0.186
0.126
0.551
0.597
0.109
0.487
0.299
0.758

0.272
0.302
0.254
0.645
0.285
0.763
0.78
0.806
-0.03
0.517
0.306
0.227
0.548
0.19
0.802
0.421

0.199
0.213
0.149
0.484
0.239
0.216
0.129
0.036
0.823
0.673
0.503
0.423
0.691
0.636
0.273
0.100

0.773
0.661
0.676
0.813
0.059
-0.002
0.036
0.588
0.823

-0.188
0.379
0.562
0.063
0.743
0.889
0.876
0.543
-0.002

0.846
0.858
0.874
0.482
0.468
0.867
0.095
0.172

0.309
0.16
0.15
0.687
0.717
0.325
0.919
0.906

Loadings
Arch
Battle
Bison
Bowron
Dolphin
Ferndale
Gregg 11
Gregg24
Hedrick
Ispah
Kerry
Mcleod
Narrow
Parsnip
PG1412
Pinchi

Lacasa
Queest
Shorts
Sicamous
Star
Wapl
Wap2
Wap3
Whiteman

Alexis
Gang
Hoodoo
Martin
Martin2
Newton
Tat
Tat2

Loading (correlation coefficient) significance levels: 0.191 (95%), 0.249 (99%), and 0.315 (99.9%), based on
N=106.
Bolded values refer to highest loadings.

74

Table 2.8. Correlation coefficients between Douglas-fir growth variation and annual/seasonal
climatic variables.
Northern Interior
NIPC 1 NIPC 2
Annual
Precipitation
Mean temperature
HM
PDO
ENSO
PNA
0.274
Prior winter
Snow
Mean temperature
PDO
ENSO
PNA
Prior summer
Rain
HM
Max. temperature
PDO
ENSO
PNA
Spring
Rain
0.365*
HM
-0.406*
Max. temperature
PDO
ENSO
PNA
Summer
Rain
Heat-moisture
index
Max temp.

NIPC 3

0.557**

Southern Interior
SIPC 1
SIPC 2

Chilcotin
CHPC 1 CHPC 2

0.387*

0.534**

0.352*
-0.498**

0.429**
-0.387*

0.589**
0.261
-0.467** -0.585**

0.456**
0.349**
0.466**
-0.432**
0.397*

-0.523**
0.530**
0.476**
0.342**
0.593**

0.400**
0.324*
-0.271

0.437**
-0.463**
-0.304

0.272
-0.251
-0.403**

0.283

0.324

-0.257
0.324**
0.390**

0.327*
-0.317*
-0.317

0.328*
0.337**

All relationships are significant at p<0.05 (bootstrapped confidence intervals). * = p<0.01 **=p<0.001
All relationships calculated from 1940 to 2004 (n=65), except PDO (1901-2005, n=105), ENSO (1900-2003,
n=104), and PNA (1951-2005, n=55).
HM=heat-moisture index.

75

Paciric
Ocean

175

350 km

Scale: 1:12,042,495

Figure 2.1. Sampling region overview map.

76

Cjf,

S McLood
Dolphin

a
Bison _ B

Arch
B
Kerry

•Battle
vPinchi

Parsnip
Fort St.
James

.#&&'»

. , . , ;, ,

; * f

•.--»• ,
fcfj'
J?~. *W«? ' ' *''
^

—'^»'•

" .-

.""

•'"

<•£„*""»

*'

' v . . ,.*•>'<.-,

*

'

'••'.
* '•*
"

" ^

Nechako River
''

Vanderhoof

H
Ferndale

B

Prince
George
Gregg11„

•

Gregg24

Ispah

•

"•'."..
Narrow * .„ .s.:'
~ *•*

'"'
.
v

"fl~'"~

9 PG1412 •
Scale: 1:1,468,620

Figure 2.2. Sampling site locations in Northern Interior study region. Red squares are
sampling sites, black dots are settlements.

77

'y"1 > *

*

1
VViltianfysLake
* * #

Alexis
-i.tr v

Creek

3*

i

Martin2

Newton

s

> i j

El Martin

J.
•i*

.

t

f

• * ' *

.-^-

/

..

"""-fc,

g V

J'

***>
ii

* 1

.

Hoodoo .
1

ar
#

•

";': -.-4

^ ^

16

32 km.

Scale: 1:1,141,517

Figure 2.3. Sampling site locations in Chilcotin Plateau study region. Red squares are
sampling sites, black dots are settlements.

78

; *- v -

) "• Abji

: • -. ;• '•*'"'"" .".,v ,r

'.

/t&'jS

kit

i-..-:*,V
i Armvai%..
g^i".

, ••.'.:l:\.

ff*** >. •

•

:;

i

.,

:<y

• , : . • ' .

.,

* i » . ' " •••""""*• - ^ " j f ^ 1 •"';.; -,c"*,Ji;'«?l

'%,
B star

Shorts f r * r*'/'

•

/ .

;

>••••-

v v - - ^

20

40 km.

Scale: 1:1,455,598

Figure 2.4. Sampling site locations in Southern Interior study region. Red squares are
sampling sites, black dots are settlements.

79

0.6

0.5

y=0.006112X+0.07963

*

0.4
*

111

CO
a

0.3

CO

0.2

2

*

#

*

*

0.1
fJt*^F**

0.8

R^O.713

y=0.006646x+0.33656

0.7
CO

v.

0.6
*

0.5

*

0.4
* *

*

R^O.537

•

•

0.3
•

'

'

'

•

•

0.5
y=0.0058053x+0.062709

0.4
>
a 0.3
CO

tz
as
CO

0.2

i

*
*

*
*

*

* • ****

*

*

0.1
2

4
6
MAT

8

500 1000 1500 2000
MAP

10

20 30
AHM

40

Figure 2.5. Chronology statistics against sample site mean annual climate normals.
Rbar=average intrasite series correlation with chronology. MAT=mean annual temperature
(°C), MAP=mean annual precipitation (mm). AHM=annual heat-moisture index.
Relationships between chronology statistics and AHM are all significant at p<0.0001.

80

0.6

0.4

-4
a

'o
ID

oo

c
g

1

0.2

o
o

0

ID

y = -0.20389 + 0.017605 * x
R 2 =0.76

-0.2

,

-0.4

y =-9.4425 +0.16104 *x
R 2 = 0.786

-10
20

10

30

40

20

10

50

30

40

50

AHM

AHM

Figure 2.6. Variation in annual precipitation-growth correlation and regression coefficients
over annual heat-moisture index normals. AHM=annual heat-moisture index. Regression
coefficients were log transformed prior to analysis. Relationships are significant at
p<0.0001. Black dashed lines represent 95% significance levels for correlation coefficients.

1
y =-10.1471 + 0.077963 * x
R2 = 0.691

y = -0.28214 + 0.0084529 * x
R 2 = 0.657

0.8
0.6

sion coe ffici

•E

|
o
!E

•
y •
• •• /

8 0.4
o

.8 0.2

««xf»

(Li

«

•

-2

/

-4

•

.
. */

/

• • / •V
t*'

-b

•

•

m

*y*

*y*m
ym

o

y
••
• •

-0.2 [
-0.4

-10
20

40

60
80
SHM

100

120

f

20

40

60
80
SHM

•

100

•

120

Figure 2.7. Variation in annual precipitation-growth correlation and regression coefficients
over summer heat-moisture index normals. SHM=summer heat-moisture index. Regression
coefficients were log transformed prior to analysis. Relationships are significant at
pO.OOOl. Black dashed lines represent 95% significance levels for correlation coefficients.

81

0.6

0.05

0.4

0.04

•£

SB
o
•Li

oo
c
o

0.2

0.03

o
o
c
o

0.02

III

0.01

Hi

'«
w

11

o
o

22
'o
!E

-0.2
-0.01

-0.4

500

1000 1500
MAP

2000

2500

500

1000 1500 2000
MAP

2500

Figure 2.8. Variation in annual precipitation-growth correlation and regression coefficients
over mean annual precipitation normals. MAP=mean annual precipitation (mm). Black
dashed lines represent 95% significance levels for correlation coefficients. MAP data is
positively skewed. Relationships were not tested for significance (see Figure 2.9), and are
only presented to demonstrate non-linear relationships.

0.6

0.4
•E

uoo

*

•

%

v
•_»_•

0.2

•

X ••

*

0

2

•

6

6.5
7
log(MAP)

y = 15.3527 +-3.2742 * x
R 2 = 0.731

-5
-6
-7
-8

I -9

-0.2

-0.4
55

-4
•E
iii

gressio ncoe

relation coeffii

•8

-3
y = 2.2825 + -0.32424 *x
R 2 =0.58

7.5

8

-10
5.5
5

6.5
7
log(MAP)

7.5

Figure 2.9. Variation in annual precipitation-growth correlation and regression coefficients
over (log-transformed) mean annual precipitation normals. MAP=mean annual precipitation
(mm). Regression coefficients and MAP data were log transformed prior to analysis.
Relationships are significant at p<0.001. Black dashed lines represent 95% significance
levels for correlation coefficients.

82

0.6
• V

a)
o
!E
<D
o
o

•

y = 3.65273 + -0.01883 *x
R2 = 0.765
••

y = 0.13032 +-0.0056128 *x
R2= 0.676

0.15

X. • •
• >•

K**

•

0.2

0.2

regress ion coeffic;ient

0.4 •

•

•

c

•

V:
. \

\
-0.2

••

0.1
0.05
0
-0.05
-0.1
-0.15

-0.4
)

10

20

30

40

10

50

20

30

40

50

AHM

AHM

Figure 2.10. Variation in winter PNA index-growth correlation and regression coefficients
over annual heat-moisture index. AHM=annual heat-moisture index. Relationships are
significant at p<0.001. Black dashed lines represent 95% significance levels for correlation
coefficients.

0.6

y = 0.31794 +-0.011138 *x
R2=0.51

0

0.4

I§

•E

1 r

y = 0.059486 + -0.0023409 * x
R2= 0.569

0.05

o
!E

0.2

O

o
c
o
w

!

S1 -0.05
-0.2
-0.4
10

20

30
AHM

40

50

-0.1

0

10

20

30

40

50

AHM

Figure 2.11. Variation in winter PDO index-growth correlation and regression coefficients
over annual heat-moisture index. AHM=annual heat-moisture index. Relationships are
significant at p<0.001. Black dashed lines represent 95% significance levels for correlation
coefficients.

83

1
o

***
>
*

0.5

Q.

0 . **

1
3 0-5

•

** t

A*

D-

*

0
•

1

•

....

_

1

•

i

i

i

•

i

**

i

*

*

**

1

i

1

•

#
I

*

1

—

1

_

•

1
•

**#

*$*

V>

S 0.5
53.5

54 54.5
Latitude

**
* %

55

*

^**

122 123 124 125
Longitude

800 10001200 1400
Elevation

1

***

(J 0.5
Q.

/

v.* .

* * **

V

**

1

****

0.5

*

•

I

I

I

•

i

1
#* *

0.5
0
1

i

i

•
I

v..
I

0

I

I

•

*

s'"
•*.v'

1

-

i

'

*

•

*W

#

•

i

.
200

-10
MCMT

•***
Q_

*
i

16 -12

14
MWMT

4 12

1
o

•

*

2
3
MAT

a.

•

* **v

8 0.5

*V

* * *

#

•

0

**
*>*

•

•

*

* *

*

1

400
PAS

600

•

i.

*****
*
#

.

'

•
i

i

•

i

i

•

30

20

40
SHM

60

1

8 0.5
0

iu0

*

500

1000
MAP

*
1500 200

*

**v

•
400
MSP

600

10

20
AHM

Figure 2.12. Northern Interior PC loads against geographic variables and climatic normals.
Latitude units=°N. Longitude units=°W. Elevation units=metres. MAT=mean annual
temperature (°C). MWMT=mean warmest month temperature (°C). MCMT=mean coldest
month temperature (°C). TD=continentality (°C). MAP=mean annual precipitation (mm).
MSP=mean summer precipitation (mm). AHM=annual heat-moisture index. SHM=summer
heat-moisture index. PAS=precipitation as snow (cm).

84

•*

5

+

+ #

* *

•

JJF

os

t

j |

*

Q.
*

#

•

*

CM 0.5
O

#

•

II

*

*

#

!.

* .
•

Q.

*

•

I

*

*

•

•

.

51

*

118

•

„ , , L . ,.„

*

*

.

#
*

•

*

#
50.S
Latitude

*

•

*

*
50

II

119
Longitude

•

*
120

0

1000
2000
Elevation

***
***
Q_

CN

#*

0

*

1F

9

0.5

*
**
*

0

5

**

0

5
MAT

* *

05

**

10 12 14 16 18 20 22-10
MWMT

* *
***

*

*

*

*
#

#

*

#•

#

-8 -6
MCMT

-4

» *
***

•

*

# *

0

#

*

*

*
500 100015002000
MAP

*

*

•

0

#

*

#
* *
*

500
MSP

500
PAS

1000

*

*

*
0

*

I

*

*
*

1000

*

# *

#
*

•

*
0

•

CM 0.5

%

+ *
*

• *

*
#

*

*

20
40
AHM

*
*

I

*
•

*
50
100
SHM

150

Figure 2.13. Southern Interior PC loads against geographic variables and climate normals.
Latitude units=°N. Longitude units=°W. Elevation units=metres. MAT=mean annual
temperature (°C). MWMT=mean warmest month temperature (°C). MCMT=mean coldest
month temperature (°C). TD=continentality (°C). MAP=mean annual precipitation (mm).
MSP=mean summer precipitation (mm). AHM=annual heat-moisture index. SHM=summer
heat-moisture index. PAS=precipitation as snow (cm).

85

•»

*#

0

*

4*

*

u 0.5

*

*

»

•

*

Q-

1

•

3 0.5

#

•

*•

*

•

*
*

51.5

*•

52
Latitude

1
o

#

52.5 122 123 124 125 500
Longitude

1000
1500
Elevation

**

0

#
<*

0.5
0
1

*

*

*•

* '

/
1

1

I

•

•

i

*
**

8 0.5

•
*

**
•

Q_

*

2
MAT

4

10

15
MCWIT

• *

20

-10

# •

*

#

*

G 0.5
0
* *

400
MAP

*

• *
*

-6 100

• **
*

*•

300

-8
MCMT

* 0 #

*

0.5

Q-

*

**

**

0

u

«*

*

150 200
PAS

* *

0

*

•*

*
*

*

*
*

*

*

* •
# *

500 160 180 200 220 20
MSP

*
*

30
AHM

250

*
*

* *
**
40

*
60

70
SHM

80

Figure 2.14. Chilcotin Plateau PC loads against geographic variables and climate normals.
Latitude units=°N. Longitude units=°W. Elevation units=metres. MAT=mean annual
temperature (°C). MWMT=mean warmest month temperature (°C). MCMT=mean coldest
month temperature (°C). TD=continentality (°C). MAP=mean annual precipitation (mm).
MSP=mean summer precipitation (mm). AHM=annual heat-moisture index. SHM=summer
heat-moisture index. PAS=precipitation as snow (cm).

86

Appendix A. Sample site mean annual temperature and precipitation normals.
MAP (mm)
t

00

o
o

K)
O
O
O

ro
to
o
o

?•

a

+

—)

MAT=mean annual temperature (1961-1990). MAP=mean annual precipitation (1961-1990).

87

Appendix B. Regional growth chronology time series derived from residual growth
chronologies.

NIPC1

7\
'-I

f"

I f

t

'

.

'h

'

*

"

i

'. \

'

I

•

"

.

'•

NIPC2
.)

!.

i

I

Era V^^^lMA^rM\p
J

I

I

L

-i

1

1

r

_i

I

i

L

i

J

'

i.....l.

i

I

i

i

! .....!

!

!

!

ft

'..._

i

"i

i

i

I

i " ~

J

l_

1970

1980

1990

NIPC3

SIPC1

2h
^
2 -2

r

T—i

\A ^

i

j

I

,

*
h

^

^

^

L

SIPC2
-i

r

, A iA

\AA * /:

Fifty
J

i

i

^
i

^
!

L_i_

CHPC1

!..,,

!

!

Vv^-Vr^H ^J^^VHv^N^V
-••*••••!

i

r'

i

CHPC2

BS3S
1900

_l

L

1910

1920

1930

1940

1950
1960
year

2000

Appendix C. Example of climate gradient analysis.
Mean site
climate

Growth response to
local annual
precipitation (19402002)

AHM

r
P
0.379
0.0203
36.8
0.064
19.6
0.0015
Arch
0.161
23.7
0.0045
Battle
0.0021
19.7
0.101
Bison
-0.024 -0.0005
16.2
Bowron
0.012
17.3
0.0003
Dolphin
1960
2000
1940
1980
-0.043
15.7
-0.001
Ferndale
year
0.574
0.0467
39.5
Gang
b)
22.4
0.144
0.0044
Gregg11
= -0.43277 + 0.046671
0.0077
21.6
0.269
Gregg24
= 0.574
8.4
0.040
0.0004
Hedrick
2
0.0411
0.474
41.8
Hoodoo
0.131
17.6
0.0026
Ispah
-0.018 -0.0003
16.1
Kerry
0.582
0.0323
43.9
Lacasa
0.402
0.0182
28.6
Martin
20
25
30
35
40
45
0.336
0.0141
23.4
Martin2
local annual precipitation (mm)
-0.022
17.0
-0.0003
Mcleod
C)
-0.009 -0.0001
12.0
Narrow
-0.21 + 0.017698 * x
0.402
36.8
0.0256
Newton
= 0.774
-0.065
-0.0006
11.1
Parsnip
o 0.5
0.261
0.0067
19.8
Pg1412
0.176
0.0044
21.8
Pinchi
17.0
0.131
0.0023
Queest
•s
22.4
0.059
0.0014
Shorts
0.259
0.0049
Sicamous 20.5
-0.5
-0.107 -0.0015
14.3
Star
20
50
10
30
40
AHM
0.446
27.6
0.0201
Tat
0.457
28.9
0.0221
Tat2
-0.018 -0.0001
6.0
Wap1
-0.033 -0.0002
6.1
Wap2
10.8
0.035
0.0005
Wap3
0.188
0.0041
Whiteman 31.6
The climatic gradient analysis was completed in two steps. First, correlation and regression coefficients were
computed between each chronology and local annual precipitation from 1940 to 2002. As an example, part a)
of this figure shows a time series of Gang site tree growth (blue) and local annual precipitation (red) from 1940
to 2002. Part b) of this figure shows a regression of Gang site tree growth on local annual precipitation from
1940 to 2002 ((3=0.0467 and r=0.574). This was performed for all 33 chronologies (see table). In the second
step of the climatic gradient analysis, the 33 correlation and regression coefficients (see table) were regressed
against site-specific climatic normals such as mean annual heat-moisture index (AHM, see table) to determine if
coefficients varied with site climate. Part c) of the figure shows the climatic gradient analysis, with the 33
correlation coefficients plotted against mean site AHM. Note that the data point corresponding to the Gang site
is highlighted.
Alexis

I
I

'Li

89

Chapter 3. Temporal variability in climate-growth relationships at the climatic margins
of Douglas-fir in British Columbia
Abstract
A growing body of evidence is challenging assumptions around the temporal stability of
historical climate-tree growth relationships. This study examined the temporal stability of
correlations between regional Douglas-fir growth patterns and climate in three study regions
in British Columbia. Growth patterns describing high-elevation Douglas-fir populations have
become more correlated with annual temperatures and snowfall over the 20th century,
possibly reflecting changing conditions in these environments associated with anthropogenic
climate change. Growth patterns describing dry climate Douglas-fir populations show
variability in the Southern Interior, possibly associated with PDO phases, an extensive
drought in the early 1900s, and the recent development of temperature-growth sensitivity.
Ecologically similar sites in the Chilcotin Plateau show more stability in climate-growth
relationships, possibly reflecting the dominant drought influence in forest-grassland ecotonal
populations. This study adds to the recent studies highlighting historically variable climategrowth relationships and suggests that in BC, Douglas-fir growth responses to climate change
will vary over time.

90

3.1 Introduction
Climate change models project that by 2100, mean annual temperatures in western North
America could be two to seven °C above the mean for the previous 1000 years, with the
greatest increases occurring during the winter months and at higher latitudes and continental
areas (Christensen et al. 2007, BC MoE 2007). Annual (summer) precipitation is also
expected to increase (decrease), although there may be a high degree of spatial variability
associated with topography, continentality, and other factors (Christensen et al. 2007, BC
MoE 2007). The frequency of droughts, severe storms, and other extreme climatic events
may also increase in association with climate change (Easterling et al. 2000).

As climate is a key factor influencing tree growth (Fritts 1976, Kimmins 1997), climate
change is expected to have a profound impact on future forest productivity and will
subsequently challenge many aspects of forest management (Nigh et al. 2004, Littell 2006,
D'Arrigo et al. 2007, Spittlehouse 2005). Tree radial growth can indicate species health,
productivity, and resilience, therefore, identifying links between specific climate parameters
and growth rates can provide information regarding future species viability and resource
sustainability (Littell and Peterson 2005, Littell 2006, Nigh et al. 2004). Many studies have
shown that species-specific radial growth (productivity) responses to climate may vary by
local abiotic conditions (e.g., climate, elevation) (e.g. Littell 2006, Littell and Peterson 2005,
Case and Peterson 2005, Zhang and Hebda 2004, Laroque and Smith 2005), with potentially
strong and unique responses in populations growing near climatically-controlled range
margins (e.g., Barber et al. 2000, Case and Peterson 2005, Zhang and Hebda 2004). Further,
recent studies have shown that critical climatic growth limitations in a given environment can

91

vary over time, possibly as a result of changing local conditions associated with long-term
climate change and/or anthropogenic non-climatic factors such as pollution and increased
atmospheric carbon dioxide concentrations (Biondi 2000, Barber et al. 2000, Lloyd and
Fastie 2002, Carrer and Urbinati 2006, Pfeifer et al. 2005, D'Arrigo et al. 2007). In order to
more fully understand and predict forest ecosystem responses to climate change, ecological
and temporal complexities underlying population-specific climate-growth relationships must
be considered in research strategies and can be elucidated by sampling across a species'
biophysical range (including its extremes) (Littell 2006, Littell and Peterson 2005, Peterson
and Peterson 2001) and examining relationships at different temporal scales (Biondi 2000,
Biondi and Waikul 2004, D'Arrigo et al. 2007, Lloyd and Fastie 2002, Carrer and Urbinati
2006, Pfeifer et al. 2005).

Our study incorporated this concept with dendroecological methods (Fritts 1976) to identify
and describe temporal variations in Interior Douglas-fir {Pseudotsuga menziesii var. glauca
(Mirb.) Franco) climate-growth relationships in the British Columbia Interior. Interior
Douglas-fir is a commercially important conifer with a wide ecological amplitude in BC
(Meidinger and Pojar 1991), ranging from dry semi-arid forests in the lee of the Coast
Mountains to low- to high-elevation wetbelt forests on the western slopes of the Rocky
Mountains (Arno 1991), to cold interior forests near Fort St. James (Ml 1999). Many studies
have shown that Douglas-fir radial growth is sensitive to climate, and that its climate-growth
relationships change with local ecological conditions (e.g., Daniels and Watson 2003, Case
and Peterson 2005, Zhang and Hebda 2004, Littell 2006, Watson and Luckman 2002),
however, to our knowledge, there have been no studies in BC (and only one in the US;
Biondi 2000) examining temporal variation in Douglas-fir climate-growth responses.
92

This study addressed this knowledge gap by examining four regional growth chronologies
generated from a principal components analysis (PCA, The Mathworks Inc. 2007) of 33
population ring-width chronologies collected over a wide geographic and climatic range in
three study regions in the province. Analysis of PCA results (presented in Chapter 2)
indicated that regional growth chronologies significantly described growth patterns across
each region, but best described populations growing at the climatic margins. Thus, this study
assessed temporal stability in regional-scale climate-growth relationships with a focus on
stands growing at the climatic margins. The specific study objectives were to:
1) identify and describe temporal variations in important climate-growth relationships in
Douglas-fir at the cold and wet (associated with high-elevation environments,
hereafter, high-environment) and dry (hereafter, dry site) climatic extremes of its
range in BC, and
2) suggest possible causal mechanisms for these temporal variations.

3.2 Materials and methods
3.2.1 Field sampling
We sampled 33 mature stands of Douglas-fir in three geographic regions (Figure 2.1) in the
British Columbia Interior: 1) Northern Interior (Figure 2.2), 2) Chilcotin Plateau (Figure 2.3),
and 3) Southern Interior (Figure 2.4). The 33 sample stands (sites) were strategically chosen
to encompass a wide range of climatic regimes, including stands located at the extremes in
each region. Potential sampling sites were identified by overlaying spatial mean annual
temperature (MAT) and precipitation (MAP) data (1961-1990 climate normal period,
Spittlehouse 2006) obtained from ClimateBC (v3.1, Wang et al. 2006) onto mature Douglas-

93

fir coverage obtained from the BC Ministry of Forest and Range's Vegetation Resource
Inventory. Sample site MAT ranged from 1.5 to 7.4 °C, and MAP ranged from 335 to 2077
mm (Table 2.2). The elevation range was 370 to 1540 m (Table 2.1). The sampled sites
encompassed seven biogeoclimatic (BEC) zones (Meidinger and Pojar 1991), representing a
wide diversity of ecosystems (Table 2.2).
Sample sites and trees were chosen to maximize the climate signal in growth variation. Sites
without evidence of edaphic limiting factors (e.g., shallow rocky soils, cold air ponding,
steep northerly aspects) or recent abiotic or biotic disturbance were selected in order to
minimize the growth impacts of non-climatic factors. The influence of inter-tree competition
was minimized by sampling trees with a dominant canopy position. We avoided sampling
trees with obvious growth defects or evidence of disease, fire scarring, or insect attacks as
much as possible. At each site, between 18 and 40 mature trees greater than 100 years old
were sampled to build a robust population sample. A single increment core was taken from
each sample tree at approximately breast height (1.3m, Josza 1988). Where sampling
occurred on slopes, increment cores were taken perpendicular to the slope direction to avoid
compression and tension wood (Josza 1988).

3.2.2 Processing and chronology development
Increment cores were mounted and sanded to enhance the contrast of tree ring boundaries.
Cores were then visually crossdated to assign calendar years to annual growth rings (Cook
and Kairiukstis 1990). Ring widths were measured using WinDendro™ image analysis
software (Regent Instruments Inc. 2005) to the nearest 0.01mm. Crossdating was verified
using COFECHA software (Holmes 1983). After checking for measurement errors, cores
94

that did not crossdate with the master chronology were excluded from the final chronology
(Daniels and Watson 2003).

The ARSTAN software (Cook and Krusic 2005) was used to remove the age-related growth
trend from each series, as this trend usually dominates the variance and masks the climate
signal (Fritts 1976, Cook and Kairiukstis 1990, Cook 1985). A cubic smoothing spline with
a 50% frequency response cutoff of 60 years was used to remove very low-frequency (i.e.,
age-related trend) variation (Cook 1985, Cook and Kairiukstis 1990). Cubic smoothing
splines remove unwanted variation by filtering data in the frequency domain; the spline used
in this study removed 50% of the signal amplitude at a frequency of 1 cycle/60 years, and an
increasing (decreasing) percentage of the signal amplitude at frequencies lower (higher) than
60 years (Cook 1985, Cook and Kairiukstis. 1990).

After detrending, ARSTAN was used to analyze autocorrelation and apply a common
autoregressive model to each series for a given site (Cook and Krusic 2005, Cook 1985).
Residuals from autoregressive modeling were then averaged together using robust mean
calculation to produce a residual growth chronology that represented stand growth variation
with autocorrelation removed (Cook and Krusic 2005). Removing autocorrelation from
growth chronologies is a common procedure in dendroecological studies (e.g., Chhin et al.
2008, Case and Peterson 2005, Littell 2006, Pfeifer et al. 2005) because autocorrelation has
statistical implications (Pfeifer et al. 2005, Littell 2006, Meko 2007) and may be a function
of non-climatic processes such as stand-level disturbances and morphological influences on
growth (Littell 2006, Cook 1985, Fritts 1976). A preliminary analysis of this dataset revealed
that autocorrelation confounded climate-growth relationships (i.e. autocorrelation may have
95

resulted from non-climatic sources, Chhin et al. 2008, Biondi and Swetnam 1987). Robust
mean calculation is more resistant to outliers, which may serve to reduce noise from treelevel growth disturbances (Cook 1985, Cook and Kairiukstis 1990).

3.2.3 Principal components analysis
For each of the 3 regions, PCA was used to consolidate the variation in annual radial growth
expressed in the chronologies of each region. Leading principal components (PC)
represented regional growth patterns and were retained as composite chronologies to be used
in determining climate-growth relationships. In order to determine what climatic and
geographic attributes were represented by each regional PC, we calculated correlation
coefficients (also referred to as 'loadings') between the original chronologies and the PCs in
each region, and then visually examined these against corresponding climatic normals (Table
2.2) and geographic variables (Table 2.1) for each chronology (Case and Peterson 2005,
Zhang and Hebda 2004). This visual analysis revealed that sites at climatic extremes were
best represented (had the highest correlations) to regional PCs, however, sites growing in
more moderate conditions also had significant (albeit lower) correlations with each PC.
Thus, we interpreted each PC to represent populations growing at the climatic margins of the
species' range in each region, but also representing populations growing under more
moderate conditions. PCA results, analysis, and discussion are presented in Chapter 2.

For this study, we used four of the regional composite chronologies: 2 chronologies
representing high-elevation interannual growth variation from 1900 to 2005 in the Southern
and Northern Interior study regions, and 2 chronologies representing interannual dry site

96

growth variation from 1900 to 2005 in the Chilcotin Plateau and Southern Interior study
regions. The 1900 to 2005 period was chosen to allow for analysis with complete climate
records where possible (see below).

3.2.4 Climate data
Environment Canada climate station temperature and precipitation data were used as
predictor variables to determine climate-growth relationships with the regional PCs (see
below for details on this analysis). Climate stations corresponding to regional PCs were
chosen based on spatial and ecological interpretations of the PCA results, proximity to
sample sites, and sufficient time series length. Adjusted monthly precipitation records
(Mekis and Hogg 1999) and homogenized monthly temperature records (Vincent and Gullett
1999, Vincent et al. 2002) provided by the Climate Research Division of the Meteorological
Service of Canada (http://www.cccma.bc.ec.gc.ca/hccd/) for the chosen climate stations were
used to account for non-climatic variation in the instrumental records caused by changes in
station location, measurement procedures, instrument changes, and other factors (Mekis and
Hogg 1999, Vincent and Gullett 1999, Vincent et al. 2002).

We made further adjustments to two data records. The Williams Lake temperature record
was not homogenized at the time of this study (Vincent, pers. comm.), therefore we used the
observed data. However, the station moved location (and over 300m up in elevation) in
1961, which likely introduced inhomogeneities in the climate record. We accounted for this
by using ClimateBC elevation-adjusted and interpolated homogenized historical temperature
data (Wang et al. 2006, Mitchell and Jones 2005) from 1939-1960 (see below for a

97

description of the ClimateBC model). A visual comparison of Williams Lake observed
temperature data and ClimateBC temperature data from 1939 to 2002 (data not shown)
indicated a possible step change in recorded temperatures around 1960, which presumably
could be associated with the change in station location and elevation.

At the time of this study, adjusted precipitation, rain, and snow data for the Vernon
Coldstream Ranch were only available up to April 1997. To create a time series that matched
the other records used in the analysis, we used observed precipitation, rain, and snow data
from a nearby station (Vernon Bella Vista) to extend the record to 2005.

Climate station data are summarized in Table 3.1. All climate stations had some missing
precipitation, rain, and snow data, which are summarized in Table 3.2. We did not replace
missing precipitation data due to a high potential for errors in estimating precipitation (Mekis
and Hogg 1999).

Seasonal and annual variables were derived by averaging monthly values (Table 3.3). Where
monthly values were missing in the climate station record (Table 3.2), the associated derived
seasonal variable was computed as a non-number and omitted pair-wise from analysis.
Growth was compared to climate variables from the previous year's July to the current year's
October because climate in the preceding year affects the current year's growth (Fritts 1976).

Seasonal and annual heat - moisture index variables were derived from monthly temperature
and precipitation data using the following formula (similar to Wang et al. 2006):

98

1) HM = (TMX + 10)/(PPT/1000)
where: HM — seasonal or annual heat-moisture index
TMX = seasonal or annual mean of monthly average maximum temperatures (°C)
PPT = seasonal or annual mean of monthly precipitation totals (mm)

El Nino Southern Oscillation (ENSO) and Pacific Decadal Oscillation (PDO) are two quasiperiodic ocean-atmosphere processes (Allan 2000, Mantua et al. 1997) that influence climate
in British Columbia (Shabbar 2006, Shabbar et al. 1997, BC MoE 2007, BC MWLAP 2002)
and have been linked to Douglas-fir growth in the Pacific Northwest United States (US)
(Case and Peterson 2005). The Pacific-North American index (PNA) is a mid-tropospheric
pressure anomaly influenced by ENSO and strongly correlated with North American winter
temperature and precipitation (Leathers et al. 1991, Shabbar 2006, Shabbar et al. 1997).
Index values used for this study were derived by averaging October to March monthly index
values obtained from the US Department of Commerce's National Oceanic and Atmospheric
Administration (http://www.cdc.noaa.gov/ClimateIndices/List/).

3.2.5 Analysis of climate-growth relationships and inter-regional growth patterns
Climate-growth relationships were assessed by 1) examining climate-growth relationships
within the Northern and Southern Interior study regions, and 2) comparing temporal growth
patterns between ecologically similar stands in disparate regions (e.g., comparing highelevation annual growth rates between the Northern and Southern Interior) (hereafter, interregional growth patterns).

99

Analyses were performed using annual, seasonal, and monthly variables, however, results
involving only annual and seasonal variables are presented because they explained more
growth variation than monthly variables (e.g., Littell 2006, Watson and Luckman 2002).

Two approaches were used to examine temporal variation in climate-growth relationships
and inter-regional growth patterns:

1) Sub-period correlation analysis was used to detect changes in climate-growth
relationships and inter-regional growth patterns between the first and second half of the
20th century. Climate and tree growth records were split into two sub-periods using 1950
as a breakpoint (Lloyd and Fastie 2002, Wilmking et al. 2004). Regional climate records
vary in length, therefore, the sample size for the pre-1950 climate-growth analysis
differed for each region. The Williams Lake climate record does not contain pre-1939
data, therefore, no sub-period analysis was computed for the Chilcotin Plateau region.
The post-1950 analysis was computed over the 1951-2005 period.

2) Finer-scale changes in temporal variation were examined using moving correlation
functions (D'Arrigo et al. 2007, Wilson and Elling 2004, Carrer and Urbinati 2006,
Biondi 2000, Biondi and Waikul 2004). A moving correlation function (MCF) calculates
correlation coefficients between two time series using a fixed-length time window that is
incrementally moved over the full series; time-dependent variation in the coefficients
provides a measure of temporal stability in the linear relationship (Biondi and Waikul
2004, Biondi 2000). MCFs were computed using a 31-year window offset in one-year
increments (D'Arrigo et al. 2007), over the full available climate record in each region.

Values presented for a given year refer to the 31-year period ending at that year. Missing
data in the climate records was omitted pair-wise from analysis, therefore, the sample size
was less than 31 over some periods. Bootstrap bias-corrected accelerated 95%
confidence intervals were calculated for each correlation coefficient using 1000 iterations
(Moore and McCabe 2006, The Mathworks Inc. 2007).

MCFs using all climatic variables would have produced too many results for reasonable
interpretation and discussion, therefore, we focused the MCF analysis on a subset of
climatic variables that explained a relatively high proportion of growth variation.
Interpretation of MCF results focused on identifying changes in correlation coefficients
over time and consistent periods of significant climate-growth relationships (i.e., multiple
consecutive 31-year windows with significant correlations). Significant correlation
coefficients over only a few consecutive time windows were treated as potentially
spurious, and moving scatterplots (data not shown) were examined to assess the influence
of potential outliers. Because the MCF analysis focused on identifying periods of
significant correlations, confidence limit adjustments such as Bonferonni (Meko 2007,
Biondi et al. 1997) or False Discovery Rate (Benjamini and Hochberg 1995) techniques
were not employed (Meko 2007).

Moving regression functions were also calculated in a similar fashion as MCFs to detect
changes in the regression slopes. The output from both analyses was virtually identical
(slopes increased with correlation), therefore, only MCF results are presented and
discussed.

101

3.3 Results
3.3.1 Temporal variation in high-elevation growth and climate-growth relationships
3.3.1.1 Inter-regional growth patterns
Annual radial increment patterns in high-elevation populations of Douglas-fir were similar
for the Northern and Southern Interior study regions over the 1900 - 2005 period (r = 0.457,
p<0.001), with a stronger correlation for the 1951 - 2005 period (r=0.564, pO.OOOl) than the
1900 - 1950 period (r=0.304, p<0.05) (data not shown). The MCF confirmed an increasing
common growth signal over time (Figure 3.2a). Since ca. 1940, growth patterns converged
and have remained significantly correlated over subsequent 31-year periods.

3.3.1.2 Temperature-growth relationships
Prior to ca. 1950, growth in the Northern Interior study region was not correlated to any
seasonal or annual climatic variables, and growth in the Southern Interior was correlated to
winter and annual temperatures (Figure 3.1). Since ca. 1950, growth in both regions was
positively correlated to fall, winter, and annual temperatures and heat-moisture index (growth
tended to be higher during warmer and drier years). Annual minimum temperatures
generally had stronger correlations with growth than maximum temperatures.

MCF using annual average temperatures confirmed an increasing association with growth
since ca. 1950 and revealed a similar pattern in temporal coefficient variation between both
regions (Figure 3.2b). Growth became increasingly associated with temperatures after the
1920-1950 period (r ~ 0), developing the strongest association over the ca. 1968-1998 period

102

(r > 0.4) in both regions. MCFs using average maximum and minimum temperatures as well
as winter season temperatures produced similar results (data not shown).

3.3.1.3 Snowfall-growth relationships
Prior to 1950, growth at high-elevation sites in both regions was not correlated with annual or
winter snowfall (Figure 3.1), although Southern Interior populations had a positive growth
correlation with spring snowfall. After ca. 1950, growth in both regions became negatively
correlated with snowfall and positively correlated with winter heat-moisture index (radial
growth tended to be higher after warmer and drier winters). Post-1950 growth had stronger
correlations with annual vs. winter snowfall in both regions. MCF using annual snowfall
indicated that temporal variations in correlation functions differed between the regions
(Figure 3.2c). Douglas-fir in the Northern Interior maintained a relatively stable negative
growth relationship to snowfall after ca. 1950, whereas in the Southern Interior, growth
correlations to annual snowfall were more dynamic. After the 1926-1956 period (r ~ 0),
snowfall became increasingly associated (negative) with growth. The 31-year period with
the strongest association was 1968 to 1998; annual snowfall explained approximately 50% of
growth variation during this period.

3.3.1.4 Ocean-atmosphere climate indices-growth relationships
At high elevation, Southern Interior growth was significantly correlated to ENSO, PDO, and
PNA indices over the full available record (Table 3.3), while Northern Interior growth was
correlated with PNA and PDO. MCFs indicated that growth at high-elevation sites in the
Southern Interior was consistently correlated to ENSO and PDO over most 31-year windows

103

in the latter half of the 20 century, while in the Northern Interior, growth correlations with
ENSO and PDO were insignificant over most time windows (Figures 3.2d and 3.2e). Growth
in both regions was significantly correlated to PNA over most 31-year periods (Figure 3.2f).

3.3.2 Temporal variation in dry-site inter-regional growth patterns and climate-growth
relationships
3.3.2.1 Inter-regional growth patterns
Dry sites in the Southern Interior and Chilcotin Plateau had correlated growth patterns from
1900 to 2005 (r = 0.359, p<0.001, not shown). Growth patterns were significantly correlated
over both sub-periods, with a slightly weaker correlation for the 1951 - 2005 period (r=0.327,
p<0.05, not shown) than the 1900 - 1950 period (r=0.390, pO.Ol, not shown). However,
MCF results revealed this relationship was dynamic, particularly in the latter half of the
century (Figure 3.4a). Correlations decreased after the 1913-1943 period and growth patterns
are insignificant for many 31 -year periods in the latter portion of the 20l century.

3.3.2.2 Temperature-growth relationships
In the Southern Interior, correlation analysis using the two sub-periods indicated that prior
summer, spring, and annual temperature became negatively associated with growth over the
last half of the century, whereas prior to ca. 1950, only prior summer average maximum
temperatures were negatively correlated with growth (Figure 3.3). MCFs revealed that prior
summer maximum and average temperatures and spring maximum temperatures became
negatively correlated to growth in recent decades, although prior summer temperatures were

104

more strongly correlated to growth for a period around the middle of the 20 century
(Figures 3.5a to 3.5c).

Chilcotin Plateau Douglas-fir growth from 1951 - 2005 was not correlated to any temperature
variables (Figure 3.3), although MCFs revealed that growth and spring maximum
temperatures became significantly negatively correlated in recent decades (Figure 3.5c).

3.3.2.3 Precipitation-growth relationships
Southern Interior growth relationships to annual and spring precipitation variables remained
significant (positive) over the two sub-periods, with generally higher correlations over the
1900-1951 period (Figure 3.3). MCFs revealed a dynamic signal strength with annual,
spring, and prior summer precipitation variables over the past century (Figures 3.4b to 3.4f).
After ca. 1950, significant growth responses to annual precipitation, rain, and heat-moisture
index and spring/prior summer precipitation gradually began to weaken; associations were
insignificant over the ca. 1950 - 1980 period. After this period, correlations between growth
and annual, spring, and prior summer precipitation rapidly increased and became significant
again. A moving scatterplot analysis confirmed that this rapid correlation increase was due
to post-1980 values (data not shown).

A sub-period analysis was not computed for the Chilcotin Plateau region, as the Williams
Lake climate record does not contain pre-193 9 data. MCFs calculated for Chilcotin Plateau
growth and annual precipitation, rain, and heat-moisture index from 1940 to 2005 revealed a
relatively stable response (Figures 3.4b to 3.4d). Correlation strengths varied somewhat over

105

time, with the strongest relationship occurring over the ca. 1950-1980 period. Relationships
with prior summer and spring precipitation weakened and became insignificant in recent
decades (Figures 3.4e and 3.4f).

3.4 Discussion
This study detected temporally variable climate-growth relationships in Douglas-fir
populations in the British Columbia Interior, thus broadening the spatial, species, and
ecological extent of this recently described phenomenon (e.g., D'Arrigo et al. 2007,
Wilmking and Myers-Smith 2008, Carrer and Urbinati 2006, Pfeifer et al. 2005). Temporally
variable climate-growth relationships will impact various management areas that rely on
ecological process data such as carbon cycle modeling, growth and yield, the management of
climatically sensitive populations, species distribution models, and the use of tree-rings as
historical climate proxies because they challenge current paradigms based on assumptions of
"stable" ecosystem responses (D'Arrigo et al. 2007, Carrer and Urbinati 2006, Barber et al.
2000, Millar et al. 2006). The adaptation of forest management to climate change will
require a greater understanding of the temporal and ecological complexities of speciesspecific climate-growth relationships that may result in unexpected responses and conditions
(D'Arrigo et al. 2007, Carrer and Urbinati 2006, Wilmking et al. 2004, Pfeifer et al. 2005,
Zhang and Hebda 2004, Littell and Peterson 2005, Peterson and Peterson 2001).

Isolated mechanisms behind unstable climate-growth relationships are difficult to identify as
many climatic and non-climatic (e.g., increased CO2, global dimming, pollution, and nitrogen
deposition) factors potentially responsible for this phenomenon covary (D'Arrigo et al. 2007,

106

Carrer and Urbinati 2006). If non-climatic factors were primary contributors to the
instability in the climate-growth relationships studied here, it would be reasonable to expect
an overall loss of sensitivity to climate variables over time (Lloyd and Fastie 2002), however,
this was not the case. The Douglas-fir populations examined in this study developed new
climatic sensitivities over the last half of the 20th century, implying that the underlying
mechanism is at least partially climatic in nature (Lloyd and Fastie 2002, D'Ariggo et al
2007, Carrer and Urbinati 2006). Multi-decadal climatic variation as a result of 20th century
linear temperature/precipitation trends and/or quasi-periodic ocean-atmosphere climate
systems such as PDO has been suggested as a causal factor in other studies (e.g., Fagre et al.
2003, Daniels and Veblen 2004, Barber et al. 2000, Lloyd and Fastie 2002) and is discussed
below as a potential primary contributor to the unstable climate-growth relationships
observed in this study.

3.4.1 High-elevation sites
Our results indicate that high-elevation Douglas-fir growth patterns and climate-growth
relationships in both the Northern and Southern Interior study regions have become more
synchronized over the 20th century, possibly reflecting an increasing macro-regional
influence of linear temperature/precipitation increases. We expected that positive
temperature-growth correlations in high-elevation populations would decrease over the 20l
century as temperatures rose and became less limiting in these cold environments (Korner
1998, 2003, Kimmins 1997, Littell 2006), however, growth in both regions actually became
more responsive to annual, prior fall, and winter temperatures after ca. 1950. This
unexpected result may be related to a co-occurring increase in annual/summer precipitation

107

(BC MWLAP 2002, BC MoE 2007, Cayan et al. 1998, Zhang et al. 2007) and cloudiness
(Dai et al. 1999, Henderson-Sellers 1992) in both regions that may have introduced new
environmental conditions such as excessive soil moisture, lower daytime maximum
temperatures and/or reduced direct sunlight. In high-elevation environments, these integrated
factors can negatively influence photosynthesis and growth (Helms 1965, Peterson and
Peterson 2001, DeLucia and Smith 1987) or at least shift critical growth limitations from soil
moisture to thermal energy (Littell 2006) and correlated factors such as sunlight, possibly
explaining the strengthening growth-temperatures correlations.

We note that annual and winter temperatures are highly correlated (e.g., r > 0.9, p<0.00001
for 1900-2005 Southern Interior and 1912-2005 Northern Interior annual and winter average
temperatures), therefore, it is difficult to identify the important variable (annual vs. winter)
that may be influencing growth. Warmer annual temperatures may simply reflect more days
with optimum temperatures for photosynthesis (Helms 1965, Doehlert and Walker 1981),
whereas warmer winter temperatures may reflect an earlier snowmelt date and a
correspondingly longer growing season (Vaganov et al. 1999).

Although snowfall can be a critical forest growth limitation in high-elevation environments
(Case and Peterson 2005, Peterson et al. 2002, Laroque and Smith 2005), our results indicate
that snow-growth responses in high-elevation Douglas-fir populations can vary over time and
by region. In both regions, growth did not become correlated to snowfall until the latter half
of the 20th century, and MCFs highlighted regional differences in snow-growth responses that
may reflect unique climatic limitations in each region. For example, the high-elevation
populations in the Southern Interior had the highest snowfall normals of the entire study

108

(Table 2.2), and may be particularly sensitive to snowfall increases. Growth in these
populations has become increasingly negatively associated with annual snowfall since ca.
1950, and the period of strongest correlations (ca. 1968 to 1998) corresponds with a period of
anomalously high October to March precipitation recorded at Vernon climate stations
(ca.1970 to 2000, Appendix A). Much of this winter precipitation would likely have been
snowfall at higher elevations, therefore, snowpacks during this period may have been
anomalously high. The period of high precipitation ended ca. 2000 and MCF results indicate
that growth correlations with snowfall began to weaken around this time.

While high-elevation Douglas-fir populations in the Northern Interior also developed a
negative growth response to snowfall after ca. 1950, MCF results indicate a relatively stable
response to this variable over the latter half of the 20th century. An analysis of Prince George
climate did not reveal any strong possible climatic mechanisms. Projected significant
decreases in winter snowfall and corresponding snowpacks over much of the province (BC
MoE 2007, BC MWLAP 2002, Christensen et al. 2007) may result in a decrease in snowgrowth responses in both regions as this climatic limitation becomes alleviated.

After ca. 1950, Southern Interior high-elevation growth developed a positive (negative)
response to prior summer rain (temperatures). With a potentially shorter growing season as a
result of deep snowpacks over the ca. 1970 to 2000 period, tree growth may have become
increasingly sensitive to growing season drought events (Zhang and Hebda 2004). Highelevation sites are usually assumed to have ample moisture, however, high summer
temperatures in these environments can cause drought-like soil conditions (Anfodillo et al.
1998) and shorten the growing season (Zhang and Hebda 2004), negatively affecting growth.
109

We note that higher snowpacks in the Southern Interior may have also resulted in greater soil
moisture availability during the subsequent growing season, although this would presumably
result in a reduced growth response to prior growing season variables, which was not
observed.

ENSO and PDO indices strongly influenced Southern Interior high-elevation growth
(particularly over the latter half of the 20th century), but were generally not reflected in
northern high-elevation tree growth. Other studies have shown these indices influence the
growth of Douglas-fir (Case and Peterson 2005) and other tree species (Holman and Peterson
2006, D'Arrigo et al. 2001, Biondi et al. 1999) in other (mostly high-elevation coastal)
regions of North America. High index values are correlated with warm and dry weather
patterns (Mantua et al. 1997, Hare and Mantua 2001) in British Columbia (BC MWLAP
2002), which may result in longer growing seasons through lower snowpack depths (Case
and Peterson 2005). While PDO and ENSO indices were regional in their influence on highelevation Douglas-fir growth, the PNA index appears to have a broad growth influence, as it
was correlated to growth in both regions. PNA may represent a different mode of variability
in surface temperatures and precipitation than PDO and ENSO; research into differences
between these climate systems and their possible effects on forest growth may provide useful
information about future productivity (Case and Peterson 2005, Fagre et al. 2003, Daniels
and Veblen 2004).

Climate-growth relationships may have recently changed in both regions, as MCFs indicated
that common growth patterns and growth responses to annual temperatures, snowfall, and
PNA weakened after ca. 1998. As well, Southern Interior growth responses to PDO and
110

ENSO have weakened somewhat since ca. 1998. Moving scatter plots (data not shown)
showed that the gradual weakening in correlation values after 1998 were a result of post1998 values. Although preliminary, these changes may signal a weakening macroregional
climate influence or changing climatic growth limitations in high-elevation environments.
Over time, growth may become more limited by local climate processes and/or different
climate variables than those identified in this study, resulting in a greater divergence in
growth patterns and climate-growth relationships between northern and southern populations.

3.4.2 Dry sites
Synchronized growth patterns in the first half of the 20th century in Southern Interior and
Chilcotin Plateau dry sites suggest a similar large-scale climatic process influenced growth in
both regions during this period (Zhang and Hebda 2004, Carrer and Urbinati 2006). Treering precipitation reconstructions for many areas in western North America, including
Williams Lake, indicate that a drought covered a large area of western North America from
the early 1900s until the mid-1940s (Watson and Luckman 2004, Daniels and Watson 2003).
This drought may have been the longest and most extensive in the past three centuries in
western North America (Watson and Luckman 2004). Annual precipitation recorded in
Vernon also indicates a prolonged period of below-average precipitation in the first half of
the 20th century, particularly before 1940 (Appendix B). As climatic events such as droughts
may cause widespread and synchronized climatic sensitivities (and possible mortality) in
forests (van Mantgem and Stephenson 2007, Breshears et al. 2005), a prolonged
macroregional drought may be a primary causal factor behind the correlated inter-regional
growth patterns observed in dry sites in both regions.

Ill

A divergence between the regional growth patterns after ca. 1943 may be related to the end
of the large-scale drought as well as a ca. post-1950 negative temperature-growth response
that developed in the Southern Interior but not the Chilcotin Plateau. Warmer spring and
summer temperatures in the Southern Interior may now exceed a species- and region-specific
optimal photosynthesis limit (which can be as low as 20-25 °C for Douglas-fir; Salo 1974,
Doehlert and Walker 1981) and/or causing physiological stress (Kimmins 1997, Balatinecz
and Anderson 1986, Kozlowski and Pallardy 2002). A general lack of temperature response
in Chilcotin Plateau populations may reflect cooler temperatures and/or the dominant growth
influence of drought in this region, however, growth in recent decades has developed a
negative growth correlation with spring maximum temperatures. This may indicate that
temperatures are beginning to exceed tolerances in the Chilcotin Plateau region; projected
future temperature increases (BC MoE 2007, BC MWLAP 2002) could result in more
negative temperature-growth responses developing, as well as a possible convergence in
growth patterns with dry sites in other regions.

Chilcotin Plateau population growth responses to annual precipitation variables were
relatively strong and stable over time, possibly reflecting persistent local drought conditions
associated with forests located at the forest-grassland ecotonal boundary (these sites were the
driest of the entire study). We note that correlations were somewhat stronger over a period of
anomalously dry conditions in the Williams Lake climate record (ca. 1940 to 1975, Appendix
C); in these types of ecotonal populations, a higher drought frequency associated with
climate change (Easterling et al. 2000) may increase growth sensitivity to precipitation, and

112

eventually cause pronounced and widespread forest growth declines and possible mortality
(van Mantgem and Stephenson 2007, Breshears et al. 2005).

Unlike the Chilcotin Plateau populations, Southern Interior growth responses to annual and
seasonal precipitation variables fluctuated substantially over the 20th century. The period of
strong correlations during the early portion of the century concurs with the prolonged period
of low precipitation discussed above (Appendix B), whereas the strengthening of correlations
between growth and precipitation in recent decades may reflect a temperature-induced
drought signal as a result of warming temperatures (Wilmking et al 2004, Littell 2006,
Oberhuber et al. 2007). Shifts in Southern Interior population precipitation-growth responses
may also be related to PDO phases. The two periods of high correlation with annual
precipitation concur approximately with positive (warm/dry) PDO phases (ca. 1925 to 1945,
and ca. 1977 to present day; Mantua et al. 1997, D'Arrigo et al. 2001, Mantua and Hare
2002, Gedalov and Mantua 2002), and the period of low growth response to this variable
occurs during a negative (cool/wet) PDO phase (ca. 1946 to 1976; Mantua et al. 1997,
D'Arrigo et al. 2001, Mantua and Hare 2002, Gedalov and Mantua 2002). This relationship
was also observed in drought-sensitive ponderosa pine in the Pacific Northwest US (Fagre et
al. 2003) and may reflect PDO influence on factors related to tree growth such as
precipitation, temperatures, growing season length, and snowfall. PDO-related climategrowth relationship shifts were not observed in the Chilcotin Plateau populations, possibly
indicating the limited spatial extent of PDO influence on local climate and/or the persistent
drought influence in the Chilcotin Plateau.

113

3.4.3 Conclusions
Future broad-scale forest productivity shifts will have significant ecological and
socioeconomic impacts and will create a challenge for forest managers (Millar et al. 2007).
Long-term shifts in local climate regimes will result in altered temperature-precipitation
interactions and subsequently, new climatic growth limitations may develop. By examining
climate-growth relationships over different temporal scales, it is possible to analyze how
climate-growth relationships have historically changed as well as detect potential recent
changes related to climate change. These changes can provide important information
regarding future responses.

In the British Columbia Interior, Douglas-fir growth in high-elevation and dry environments
appears to be changing, possibly in response to climate change, and these climatically
sensitive populations may serve as indicators of broader future responses. High-elevation
populations are becoming more sensitive to annual temperatures and snowfall, possibly
reflecting limitations associated with co-occurring increases in precipitation and cloudiness.
The influence of PDO and ENSO has become stronger in Southern Interior high-elevation
populations over the 20th century, whereas PNA influence on growth in both regions has been
relatively stable since 1950.

Dry site populations of Douglas-fir in the Southern Interior and the Chilcotin Plateau may
have been similarly affected by an extensive drought in the early portion of the 1900s. Over
the 20 th century, Southern Interior populations have had a variable relationship with
precipitation and temperatures, whereas Chilcotin Plateau populations have generally been

114

consistently limited by the very low annual precipitation in that region. These differences
may reflect unique local site limitations.

Although the use of these temporal variations in climate-growth relationships may be limited
in terms of predictive abilities, the variations themselves highlight the dynamic nature of
climate-growth relationships in Douglas-fir forests and must be considered when adapting
future forest management to climate change. These and other findings challenge assumptions
of stable ecosystem processes and highlight the considerable uncertainty facing forest
managers regarding future forest responses to current rapid environmental changes (Millar et
al. 2007, Spittlehouse and Stewart 2003, Spittlehouse 2005).

115

3.5 Literature cited
Allan, R.J. 2000. ENSO and climatic variability in the past 150 years, p. 3-56. In El Nino and
the Southern Oscillation: multiscale variability and global and regional impacts. Diaz,
H. F. and Markgraf, V. (eds). Cambridge University Press, New York.
Anfodillo, T., S. Rento, V. Carraro, L. Furlanetto, C. Urbinati, and M. Carrer. 1998. Tree
water relations and climatic variations at the alpine timberline: seasonal changes of
sap flux and xylem water potential in Larix decidua Miller, Picea abies (L.) Karst. and
Pinus cembra L. Annales des Sciences Forestieres 55: 159-172.
Arno, S.F. 1991. Ecological relationships of interior Douglas-fir, p. 47-51. In Interior
Douglas-fir: The species and its management symposium proceedings. Baumgartner,
D. M. (eds). Department of Natural Resource Sciences, Washington State University,
Pullman, Washington.
Balatinecz, J.J. and H.W. Anderson. 1986. Effects of climatic factors on the formation of
wood and its quality, p. 195-200. In Climate Applications in Forest Renewal and
Forest Production, Proceedings of Forest Climate 1986. Maclver, D. C , Street, R. B.,
and Auclair, A. N. (eds). Canadian Forestry Service, Ottawa, Ontario.
Barber, V., G. Juday, and B. Finney. 2000. Reduced growth of Alaska white spruce in the
twentieth century from temperature-induced drought stress. Nature 405: 668-672.
BC MoE. 2007. Environmental Trends in British Columbia: 2007. State of Environment
Reporting. BC Ministry of Environment, Victoria, B.C.
BC MoFR. 2006. Preparing for climate change: Adapting to impacts on British Columbia's
forest and range resources. BC Ministry of Forests and Range, Victoria, BC.
BC MWLAP. 2002. Indicators of climate change for British Columbia, 2002. Indicators of
climate change for British Columbia, 2002. BC Ministry of Water, Land and Air
Protection, Victoria, B.C.
Benjamini, Y. and Y. Hochberg. 1995. Controlling the false discovery rate: a practical and
powerful approach to multiple testing. Journal of the Royal Statistical Society. Series
B 57: 289-300.
Biondi, F. 2000. Are climate-tree growth relationships changing in north-central Idaho,
U.S.A.? Arctic, Antarctic, and Alpine Research 32: 111-116.
Biondi, F. 1997. Evolutionary and moving response functions in dendroclimatology.
Dendrochronologia 15: 139-150.

116

Biondi, F., D.R. Cayan, and W.H. Berger.1999. Decadal-scale changes in southern California
tree-ring records, p. pp. 303-306. In Proceedings of the 10th Symposium on Global
Change Studies. Karl, T. R. (eds). American Meteorological Society, Boston, MA.
Biondi, F. and Swetnam, T.W. 1987. Box-Jenkins Models of forest interior tree-ring
chronologies. Tree-Ring Bulletin 47: 71-96.
Biondi, F. and K. Waikul. 2004. Dendroclim2002: A C++ program for statistical calibration
of climate signals in tree-ring chronologies. Computers & Geosciences 30: 303-311.
Breshears, D.D., N.S. Cobb, P.M. Rich, K.P. Price, C D . Allen, R.G. Balice, W.H. Romme,
J.H. Kastens, M.L. Floyd, J. Belnap, J.J. Anderson, O.B. Myers, and C.W. Meyer.
2005. Regional vegetation die-off in response to global-change-type drought.
Proceedings of the National Academy of Sciences 102: 15144-15148.
Carrer, M. and C. Urbinati. 2006. Long-term change in the sensitivity of tree-ring growth to
climate forcing in Larix decidua. New Phytologist 170: 861-872.
Case, M.J. and D.L. Peterson. 2005. Fine-scale variability in growth-climate relationships of
Douglas-fir, North Cascade Range, Washington. Canadian Journal of Forest Research
35: 2743-2755.
Cayan, D.R., M.D. Dettinger, H.F. Diaz, and N.E. Graham. 1998. Decadal variation of
precipitation over Western Northern America. Journal of Climate 11: 3148-3166.
Chhin, S., E.H. Hogg, V.J. Lieffers, and S. Huang. 2008. Influences of climate on the radial
growth of lodgepole pine in Alberta. Botany 86: 167-178.
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. Sarr, and P. Whetton. 2007. Regional Climate Projections, p. 848-940.
In Climate Change 2007: The Physical Science Basis. Contribution of Working
Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate
Change. Solomon, S. D., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K. B.,
Tignor, M., and Miller, H. M. (eds). Cambridge University Press, New York.
Cook, E. 1985. A time series analsysis approach to tree-ring standardization. PhD
Dissertation. University of Arizona. Tucson, Arizona.
Cook, E. and L. Kairiukstis. 1990. Methods of dendrochronology: Applications in the
environmental sciences. Kluwer Academic Publishers, Dordrecht, Netherlands.
Cook, E.R. and P. Krusic. 2005. Users Manual for Program ARSTAN. Tree-Ring Lab,
Lamont-Doherty Earth Observatory, Columbia University, New York.
http://www.ldeo.columbia.edu/res/fac/trl/public/publicSoftware.html.
D'Arrigo, R., R. Villalba, and G. Wiles. 2001. Tree-ring estimates of Pacific decadal climate
variability. Climate Dynamics 18: 219-224.

117

D'Arrigo, R., R. Wilson, B. Liepert, and P. Cherabini. 2007. On the 'Divergence Problem' in
Northern Forests: A review of the tree-ring evidence and possible causes. Global and
Planetary Change 289-305.
Dai, A., K.E. Trenberth, and T.R. Karl. 1999. Effects of clouds, soil moisture, precipitation,
and water vapor on diurnal temperature range. Journal of Climate 12: 2451-2473.
Daniels, L.D. and T.T. Veblen. 2004. Spatiotemporal influences of climate on altitudinal
treeline in northern Patagonia. Ecology 85: 1284-1296.
Daniels, L.D. and E. Watson. 2003. Climate-fire-vegetation interactions in the Cariboo
forests: a dendrochronological analysis. Report to Forest Innovation and Investment.
Forest Innovation and Investment Research Program, Vancouver, BC.
DeLucia, E.H. and W.K. Smith. 1987. Air and soil temperature limitations on photosynthesis
in Engelmann spruce during summer. Canadian Journal of Forest Research 17: 527533.
Doehlert, D.C. and R.B. Walker. 1981. Photosynthesis and photorespiration in Douglas-fir as
influenced by irradiance, C02 concentration, and temperature. Forest Science 27:
641-650.
Easterling, D.R., G.A. Meehl, C. Parmesan, S.A. Changnon, T.R. Karl, and L.O. Mearns.
2000. Climate extremes: observations, modeling, and impacts. Science 289: 20682074.
Fagre, D.B., D.L. Peterson, and A.E. Hessl. 2003. Taking the pulse of mountains: Ecosystem
responses to climatic variability. Climatic Change 59: 263-282.
Fritts, H.C. 1976. Tree rings and climate. The Blackburn Press, New Jersey.
Gedalov, Z. and N.J. Mantua. 2002. A multi-century perspective of variability in the Pacific
Decadal Oscillation: new insights from tree rings and coral. Geophysical Research
Letters 29: 1-4.
Hare, S.R. and N.J. Mantua. 2001. An historical narrative on the Pacific Decadal Oscillation,
interdecadal climate variability and ecosystem impacts. Report of a talk presented at
the 20th NE Pacific Pink and Chum workshop, Seattle, WA, 22 March 2001.
http://www.iphc.washington.edu/Staff/hare/html/papers/pcworkshop/pcworkshop.pdf.
Helms, J. A. 1965. Diurnal and seasonal patterns of net assimilation in Douglas-fir,
Pseudotsuga menziesii (Mirb.) Franco, as influenced by environment. Ecology 46:
698-708.
Henderson-Sellers, A. 1992. Continental cloudiness changes this century. GeoJournal 27:
255-262.

118

Holman, M.L. and D.L. Peterson. 2006. Spatial and temporal variability in forest growth in
the Olympic Mountains, Washington: sensitivity to climatic variability. Canadian
Journal of Forest Research 36: 92-104.
Holmes, R. 1983. Computer-assisted quality control in tree-ring dating and measuring. TreeRing Bulletin 43: 69-78.
Josza, L. 1988. Increment core sampling techniques for high quality cores. Special
Publication SP-30. Forintek Canada Corp., Vancouver, BC.
Jull, M.J. 1999. Douglas-fir silviculture "on the edge": silvicultural systems at the northern
range of the species, p. 34-44. In Ecology and management of interior Douglas-fir
(Pseudotsuga menziesii var glauca) at the northern extreme of its range. Proceedings
of a workshop held 7-9 October 1996 in Fort St. James, British Columbia. Lousier, J.
D. and Kessler, W. B. (eds). University of Northern British Columbia, Prince George,
BC.
Kimmins, J. P. 1997. Forest Ecology. Prentice-Hall Inc., New Jersey.
Kozlowski, T.T. and S.G. Pallardy. 2002. Acclimation and adapative responses of woody
plants to environmental stresses. The Botanical Review 68: 270-334.
Korner, C. 1998. A re-assessment of high elevation treeline positions and their explanation.
Oecologia 155: 445-459.
Laroque, C.P. and D.J. Smith. 2005. Predicted short-term radial-growth changes of trees
based on past climate on Vancouver Island, British Columbia. Dendrochronologia 22:
163-168.
Leathers, D.J., B. Yarnal, and M.A. Palecki. 1991. The Pacific/North American
Teleconnection Pattern and United State Climate. Part I: Regional Temperature and
Precipitation Associations. Journal of Climate 4: 517-528.
Littell, J. S. 2006. Climate impacts to forest ecosystem processes: Douglas-fir growth in
north-western U.S. mountain landscapes and area burned by wildfire in western U.S.
eco-provinces. PhD Dissertation. University of Washington. Seattle, WA.
Littell, J.S. and D.L. Peterson. 2005. A method for estimating vulnerability of Douglas-fir
growth to climate change in the northwestern U.S. The Forestry Chronicle 81: 369374.
Lloyd, A.H. and C.L. Fastie. 2002. Spatial and temporal variability in the growth and climate
response of treeline trees in Alaska. Climatic Change 52: 481-509.
Mantua, N.J., S.R. Hare, Y. Zhang, J.M. Wallace, and R.C. Francis. 1997. A Pacific
interdecadal climate oscillation with impacts on salmon production. Bulletin of the
American Meteorological Society 78: 1069-1079.

119

Meidinger, D. and J. Pojar. 1991. Ecosystems of British Columbia. BC Ministry of Forests,
Victoria, B.C.
Mekis, E. and W.D. Hogg . 1999. Rehabilitation and analysis of Canadian daily precipitation
time series. Atmosphere-Ocean 37: 53-85.
Meko, D. M. 2007. GEOS 585A: Applied Time Series Analysis. Page accessed: 2007. URL:
http ://www. ltrr. arizona. edu/~dmeko/geos5 8 5 a.html
Millar, C, R. Neilson, D. Bachelet, R. Drapek, and J. Lenihan. 2006. Climate change at
multiple scales, p. 30-60. In Forests, carbon and climate change: a synthesis of
science findings.Oregon Forest Resources Institute, Portland, Oregon.
Millar, C.I., N.L. Stephenson, and S.L. Stephens. 2007. Climate change and forests of the
future: managing in the face of uncertainty. Ecological Applications 17: 2145-2151.
Mitchell, T.D. and P.D. Jones. 2005. An improved method of constructing a database of
monthly climate observations and associated high-resolution grids. International
Journal of Climatology 25: 693-712.
Moore, D. S. and McCabe, G. P. 2006. Introduction to the Practice of Statistics. W.H.
Freeman and Co., New York, N.Y.
Nigh, G.D., C.C. Ying, and H. Qian. 2004. Climate and productivity of major conifer species
in the interior of British Columbia, Canada. Forest Science 50: 659-671.
Oberhuber, W., W. Kofler, K. Pfiefer, A. Seeber, A. Gruber, and G. Wieser. 2007. Long-term
changes in tree-ring-climate relationships at Mt. Patscherkofel (Tyrol, Austria) since
the mid-1980s. Trees.
Peterson, D.W. and D.L. Peterson. 2001. Mountain hemlock growth responds to climatic
variability at annual and decadal time scales. Ecology 82: 3330-3345.
Pfeifer, K., W. Kofler, and W. Oberhuber. 2005. Climate related causes of distinct radial
growth reductions in Pinus cembra during the last 200 yr. Vegetation History and
Archaeobotany 14:211-220.
Regent Instruments Inc. 2005. WinDENDRO™. An image analysis system for tree-rings
analysis. Regent Instruments Inc. Quebec
Salo, D. J. 1974. Factors affecting photosynthesis in Douglas-fir. PhD Dissertation.
University of Washington. Seattle, WA.
Shabbar, A. 2006. The impact of El Nino-Southern Oscillation on the Canadian climate.
Advances in Geosciences 6: 149-153.
Shabbar, A., B. Bonsai, and M. Khandekar. 1997. Canadian Precipitation Patterns Associated
with the Southern Oscillation. Journal of Climate 10: 3016-3027.

120

Spittlehouse, D. 2006. ClimateBC: Your access to interpolated climate data for BC.
Streamline Watershed Management Bulletin 9: 16-21.
Spittlehouse, D.L. 2005. Integrating climate change adaptation into forest management.
Forestry Chronicle 81: 691-695.
Spittlehouse, D.L. and R.B. Stewart. 2003. Adaptation to climate change in forest
management. BC Journal of Ecosystem Management 4: 1-11.
The Mathworks Inc. 2007. Matlab 2007. Version 2007b. The Mathworks Inc.
http://www.mathworks.com
Vaganov, E.A., M.K. Hughes, A.V. Kirdyanov, F.H. Schweingruber, and P.P. Silkin. 1999.
Influence of snowfall and melt timing on tree growth in subarctic Eurasia. Nature
400: 149-151.
van Mantgem, P.J. and N.L. Stephenson. 2007. Apparent climatically induced increase of
tree mortality rates in a temperate forest. Ecology Letters 10: 909-916.
Vincent, L.A. and D.W. Gullett. 1999. Canadian historical and homogenous temperature
datasets for climate change analyses. International Journal of Climatology 19: 13751388.
Vincent, L.A., X. Zhang, B.R. Bonsai, and W.D. Hogg . 2002. Homogenization of daily
temperatures over Canada. Journal of Climate 15: 1322-1334.
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.
Watson, E. and B. Luckman. 2004. Tree-ring based reconstructions of precipitation for the
Southern Canadian Cordillera. Climatic Change 65: 209-241.
Watson, E. and B.H. Luckman. 2002. The dendroclimatic signal in Douglas-fir and
ponderosa pine tree-ring chronologies from the southern Canadian Cordillera.
Canadian Journal of Forest Research 32: 1858-1874.
Wilmking, M., G.P. Juday, V.A. Barber, and H.S.J. Zald. 2004. Recent climate warming
forces contrasting growth responses of white spruce at treeline in Alaska through
temperature thresholds. Global Change Biology 10: 1724-1736.
Wilmking M. and I. Myers-Smith. 2008. Changing climate sensitivity of black spruce (Picea
mariana Mill.) in a peatland-forest landscape in Interior Alaska. Dendrochronologia
25: 167-175.
Wilson, R. and W. Elling. 2004. Temporal instability in tree growth/climate response in the
Lower Bavarian Forest region: implications for dendroclimatic reconstruction. Trees
18: 19-28.

121

Zhang, Q.-B. and R.J. Hebda. 2004. Variation in radial growth patterns of Pseudotsuga
menziesii on the central coast of British Columbia, Canada. Canadian Journal of
Forest Research 34: 1946-1954.
Zhang, X., F.W. Zwiers, G.C. Hegerl, F.H. Lambert, N.P. Gillett, S. Solomon, P.A. Stott, and
T. Nozawa. 2007. Detection of human influence on twentieth-century precipitation
trends. Nature 448: 461-465.

122

Tabic 3.1. Meteorological records used in Chapter 3.

Station name

Latitude Longitude Elevation Station
(°N)
(°W)
(m)
ID

Data length
(years)

Northern Interior
Prince George

53.89

122.68

691

1096450 95 (1912-2006)

Southern Interior
Vernon Coldstream
Ranch
Vernon Bella Vista*

50.22
50.26

119.2
119.31

482
427

1128580 98 (1900-1997)
1128553 21(1985-2005)

Chilcotin
Williams Lake**

52.18

122.05

940

1098940 69 (1939-2007)

* Vernon Bella Vista rain, snow, and precipitation data from April 1997 to December 2005 added to Vernon Coldstream Ranch data to
extend time series.
** ClimateBC elevation-adjusted and interpolated homogenized historical temperature data (Wang et al. 2006, Mitchell and Jones 2005)
from 1939-1960 used to account for possible inhomogeneities in Williams Lake observed data.

123

Table 3.2. Missing rain, snow, and precipitation data in meteorological records.
Year Month(s)
Station name
1900 January to April
Vernon Coldstream
Ranch
October
1901 February
March
November
1902 February to May
October
1903 February
December
1905 January
1912 March
1924 January
1963 August
Williams Lake A

1947 September
October
1960 June
1979 October
November

Prince George

1912 January to December
1913 December
1914 January
February
1915 January
February
October to December
1916 January
1917 June

Table 3.3. Monthly variables used to create seasonal variables for climate-growth analyses.
Season
Monthly variables
Prior July to September
Prior summer
October to March
Winter
April to June
Spring
July to September
Summer
Prior July to June
Annual
Table 3.4. Correlation coefficients between high-elevation Douglas-fir growth and oceanatmosphere climate system indices.
PDO (1900 - 2005)
ENSO (1900 - 2003) PNA (1951 - 2005)
Northern Interior
0.198, p<0.05
0.400, p<0.01
0.342, pO.QQl
Southern Interior
0.593, pO.OOOl
0.476, p<0.0001
PDO = Pacific Decadal Oscillation; ENSO = El Nino Southern Oscillation; PNA = Pacific/North American teleconnection index

Northern Interior

Post-1950

Pre-1950
1

1~

' I

1

1

1—

—i

r-

prior summer
winter
spring
summer
annual •

Southern Interior
prior summer
winter
spring

• ..vw**..-.'.'
. . . . :. :-..^.;.-.>:.v.-.^o. >»

•

•

'•-r*ri-r , K«;-

B«
summer

™

'•^SR

:&*•«
annual
•

..'l.A

tmn

tav

I—^

tmx

1

1

rain snow

1

1

ppt

hm

tmn

tav

tmx

rain snow

ppt

Correlation Coefficient
-0.6

-0.4

-0.2

0.2

0.4

0.6

Figure 3.1. Correlations between climate and growth for high-elevation Douglas-fir
populations in two regions of British Columbia over two sub-periods. Due to different
climate station record lengths, pre-1950 climate-growth relationships were calculated over
different time periods (Northern Interior: 1912 to 1950, Southern Interior: 1900-1950). Post1950 climate-growth relationships were calculated from 1951 to 2005. Climate variables are
seasonal averages of monthly values. tmn=monthly mean minimum temperature,
tav=monthly mean temperature, tmx=monthly mean maximum temperature,
ppt=precipitation, hm=heat-moisture index. Correlations shown are significant (p<0.05).

125

a) Inter-regional growth patterns

d) ENSO
Northern Interior
0.6
0.4
0.2

jjjWi>jj--";jL TnfL -n j r fl ^Jl|J|^

Southern Interior
0.6

J^J*»

0.4
0.2

•£

b) Annual TAV

ijllil

«

e)PD0
Northern Interior!

m
o
O
tz
o

0.2

-4—'

JO
a*

1 Lflrinhniifir
III_ _ |Mj_I_M 11 ijiMJTUJ 1LJIUJ LU1UJ1UJ lUJU. !_U_ 1UJIUJ LUILUl UJI

•

o
O
0.6
0.4
0.2

f)PNA

c) Annual Snowfall

Northern Interior
0.6
0.4
0.2
Southern Interior
0.6

1MB
^^^1
^^^1

0.4
0.2
1940

1960

1980

2000

BHBB(W

1940

1960

1980

2000

Last year of 31-year window

Figure 3.2. Moving correlation functions between climate and growth for high-elevation
Douglas-fir populations in two regions of British Columbia. TAV=monthly mean
temperatures, ENS0=E1 Nino Southern Oscillation index, PDO=Pacific Decadal Oscillation
index, PNA=Pacific/North American teleconnection index. Significant correlation
coefficients (p<0.05) are shaded.
126

Southern Interior
prior summer
winter

•

Post-1950

Pre-1950

spring
summer
annual

Chilcotin Plateau
prior summer
MSS"***"^

winter
'f _

spring
summer
annual
tmn

tav

tmx

m •
rain snow

hm

Correlation Coefficient

-0.6

-0.4

-0.2

; ; mm
0.2

0.4

0.6

Figure 3.3. Correlations between climate and growth for Douglas-fir populations growing in
dry climates in two regions of British Columbia over two sub-periods. No pre-1950 climategrowth relationships were calculated for Chilcotin Plateau growth because the Williams Lake
climate station record length is too short. Post-1950 climate-growth relationships calculated
from 1951 to 2005. Climate variables are seasonal averages of monthly values.
tmn=monthly mean minimum temperature, tav=monthly mean temperature, tmx=monthly
mean maximum temperature, ppt=precipitation, hm=heat-moisture index. Correlations
shown are significant (p<0.05).

127

a) Inter-regional growth patterns

d) Annual Rain

b)AHM

e) Spring PPT
Southern Interior
0.6

o
O -0.2

0.4

o -0.4

0.2 |

-0.6
o
O

-0.2
-0.4
-0.6

""I^HH

Chilqotin Plateau:
0.6

jjt_....

0.4

L..hnfl^^^Hhltmll Ihrr

! Ill•in

0.2
Chilqotin Plateau \

^ ^ ^

• |

f) Prior summer PPT

c) Annual PPT

1940

1960

1960

2000

1940

1960

1980

2000

Last year of 31-year window

Figure 3.4. Moving correlation functions between climate and growth for Douglas-fir
populations growing in dry climates in two regions of British Columbia. AHM=annual heatmoisture index, PPT=precipitation. Shaded values represent significant (p<0.05) correlation
coefficients.

128

a) Prior summer TAV

1j) Prior summer TMX

• w^m®m
Southern Interior

o
-0.2

0
0.2

^ottffiern l n t V | ^ V ^ H B

-0.4

0.4

-0.6

0.6

^ P

•

Chilcotin Plateau:
0

^

-0.4

:

•

•

•

Chilcotin Plateau:

:

hll

;....Pip t.

-0.2

^

0

i. ptigi YW^

0.2

: ^ TlliU

0.4
0.6

-0.6
1

1

1

L _

1940

1960

1980

2000

c) Spring TMX
0
-0.2
-0.4
-0.6

0

Chilcotin Plateau:

ryJHt^

-0.2
-0.4
-0.6
1940

1960

1980

2000

Last year of 31-year window

Figure 3.5. Moving correlation functions between temperatures and growth for Douglas-fir
populations growing in dry climates in two regions of British Columbia. TAV=mean
monthly temperatures, TMX= mean monthly maximum temperatures. Significant correlation
coefficients (p<0.05) are shaded.

129

Appendix A. Time series of October-March precipitation anomalies for Vernon, British
Columbia, from 1904 to 2004.

1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

year

Decadal variation highlighted with moving 11 -year average filter.

Appendix B. Time series of annual (prior July to current June) precipitation anomalies
for Vernon, British Columbia, from 1904 to 2004.

1900

1910

1920

1930

1940

1950 1960 1970
year
Decadal variation highlighted with moving 11 -year average filter.

1980

1990

2000

Appendix C. Time series of annual (prior July to current June) precipitation anomalies
for Williams Lake, BC, from 1940 to 2005.
I

c
•2

1

0

1-1
en

1

1

i

ImuJ
4
V
\\ ' V '

i A JV

™

t'^M

v

TTj ^V V

v '

Vi '

-2

-3 1
1940

i

1950

1960

i

1

1970
1980
year
Decadal variation highlighted with moving 11-year average filter.

1

1

1990

2000

131

Chapter 4. Summary and management implications
4.1 Introduction
This research project addressed a knowledge gap regarding Douglas-fir radial growth
responses to climatic variability across its climatic and geographic range in the interior of
British Columbia (BC). We sampled 33 chronologies in three large study regions spanning a
wide range of climatic conditions, including stands located at the ecological and geographic
margins of the species' range in the province, in order to 1) identify key historical climatic
sensitivities, 2) link these sensitivities to local climate regimes, and 3) examine how
sensitivities change over a climatic gradient. As well, we examined some climate-growth
relationships at different temporal scales in order to assess their temporal stability. Below,
we briefly summarize the key findings of this study and discuss potential management
implications.

4.2 Key study results and management implications
In all three study regions, Douglas-fir growth patterns tended to be highly correlated between
sites and are differentiated mainly by site climate conditions. For example, mid- to lowelevation dry sites in a given region tended to have correlated growth patterns whereas highelevation wetter sites tended to be more correlated. Highly correlated growth patterns
between ecologically similar stands separated by considerable distances (400+ km) suggest
that across the species' range in BC, Douglas-fir growth is regulated mostly by regional-scale
precipitation and temperature patterns and corresponding responses to changes in these
climate patterns will be widespread.

132

4.2.1 Mid- to low-elevation dry-site Douglas-fir populations
Regional growth patterns describing Douglas-fir growth in relatively dry and/or warm
environments at mid- to low-elevations in all study regions were strongly correlated to annual
(prior July to current June) precipitation, which likely represents the dominant influence of
soil moisture availability across most of the species' range. When examined at a smaller
(population-level) scale, the strength of this response appeared to be largely a function of
local site climate conditions, especially precipitation. When a site's annual precipitation
regime was considered, we found that Douglas-fir became exponentially more sensitive to
annual precipitation with dryness. Temperature (and variables derived solely from
temperature such as frost-free period and degree-days) did not explain any variation in
population precipitation-growth relationships, however, integrating temperature and
precipitation together into a heat-moisture index explained the most variation and may be the
most accurate way to describe a site's local climate conditions (and corresponding growth
sensitivity to annual precipitation). By reflecting both precipitation inputs as well as
evaporation/transpiration forcings (Wang et al. 2006), heat-moisture indices may provide a
good indication of available soil moisture.

Predictive models are integral to forest management in BC (BC MoFR 2006) and the
incorporation of climate variables and related biological responses into growth and yield,
carbon cycle, and species distribution models has been recommended as an important step in
improving future management decisions (Spittlehouse 2005, Ogden and Innes 2007, Aitken
et al. 2008, O'Neill et al. 2008). For example, species distribution models that predict
changes in the fundamental and realized niche of a species based on the concept of a
bioclimate envelope (Pearson and Dawson 2003, Hamann and Wang 2006) may be improved
133

if the climatic envelope used is biologically relevant (e.g., with respect to this study, if annual
heat-moisture index is used instead of more typical variables such as mean annual
temperatures) to the species in question and the model considers specific biological responses
(e.g., growth responses to annual precipitation) (Aitken et al. 2008). Further, these types of
models may improve their predictive ability if they consider response variation among
populations (e.g., with respect to this study, increasing sensitivity to annual precipitation with
site heat-moisture index) across the species' climatic envelope instead of assuming a more
uniform species response (O'Neill et al. 2008).

We can examine our results against future projections of annual heat-moisture index (AHM)
and annual precipitation in BC to suggest possible outcomes in terms of Douglas-fir
productivity. In all three study regions, temperatures are projected to increase at a faster rate
than precipitation, resulting in increased AHM (Figure 4.1). Based on these projections, our
results suggest that growth in all populations will become 1) more variable (as measured by
mean sensitivity), 2) more correlated among individuals within the same population (as
measured by rbar, thus reflecting a stronger climate signal), 3) more correlated with annual
precipitation, and 4) sharply more sensitive (as measured by regression slope) to annual
precipitation where local AHM increases above approximately 20 (Figure 4.2) or where local
MAP decreases below approximately 700mm per year (Figure 4.3). Although annual
precipitation is expected to modestly increase over most of BC (BC MoE 2007), including
the three study regions (Figure 4.1), accelerating temperature increases may result in net
drying (Christensen et al. 2007) and create temperature-induced drought conditions (Barber
et al. 2000). Projected declines in winter snowfall and summer precipitation in some areas
(Figure 4.1) could further reduce the availability of soil moisture during the growing season.
134

Under this scenario, long-term Douglas-fir growth rates could be negatively affected across
the species' range, with pronounced and initial responses likely in populations growing near
the warm and/or dry climatic margins. Projected AHM increases are highest for the
Chilcotin Plateau region (Figure 4.1), where some of the most sensitive populations already
exist; continued exponential increases in growth sensitivity in these populations could
conceivably result in eventual mortality. Douglas-fir stands near the dry and/or warm
margins of the species' range in the province (e.g., relatively high AHM), especially those at
the forest-grassland transition (e.g. in the Chilcotin Plateau and elsewhere in the province in
the Bunchgrass BEC Zone, Meidinger and Pojar 1991), should be highlighted as especially
vulnerable to climate change and future drought events.

Declining Douglas-fir stands can be targeted for harvesting to allow for conversion to more
suitable genotypes or species (such as lodgepole pine, Pinus contorta Douglas ex Louden,
Case and Peterson 2005), or can be thinned to increase resource availability for remaining
trees (Ogden and Innes 2007), however, dry margin sites tend to have limited productivity
due to the severe environmental conditions and may therefore not be economically suitable
for intensive management. At the very least, these populations should be carefully monitored
as early indicators of potential forest dieback stemming from climate change.

Historical data are often used to form baselines that are incorporated into models and
decision-making frameworks (Millar et al. 2007), however, the temporal variations in
climate-growth relationships found in this study support similar findings elsewhere (e.g.,
Lloyd and Fastie 2002, Carrer and Urbinati 2006, D'Arrigo et al. 2007) that highlight the
potential limitations of extrapolations from historical relationships. Our results suggest that

in certain environments, Douglas-fir climate-growth relationships have varied in the past
(possibly associated with drought events and/or Pacific Decadal Oscillation phases) and may
also be changing now as a result of anthropogenic climate change. In other environments,
such as dry sites in the Chilcotin Plateau, climate-growth relationships appear to be more
stable over time, perhaps reflecting the chronic strong drought limitations in forest-grassland
ecotonal populations. The use of moving correlation functions and other techniques designed
to detect temporal variation in climate-growth relationships should be incorporated into
future research to improve assessments of ecosystem responses and to detect possible early
responses as a result of climate change. Although temporal variability in climate-growth
relationships will be difficult to incorporate into models or decision-making frameworks, it
must be considered when making future management decisions based on historical ecological
relationships.

Our study did not explicitly consider the influence of within-species genetic diversity in
climate-growth relationships. Douglas-fir growth-related traits have been shown to vary
steeply along environmental clines, and this variation has been attributed to a relatively high
species genetic diversity that allows populations to adapt to a broad spectrum of
environmental conditions across the species' geographic range (St. Clair et al. 2005, Arno
1991, Rehfeldt 1991). Populations that are closely adapted to local conditions may have
narrow climatic tolerances and could become maladapted if local climate regimes move
outside of those tolerances; therefore, from a species perspective, maladaptation could be
expressed across a wide geographic range (Aitken et al. 2008, O'Neill et al. 2008). Results
from our regional climate-growth analysis support this suggestion, as they showed that
leading growth patterns in all three regions were sensitive to annual precipitation anomalies.
136

However, our results from the population climate-growth analyses also suggest that across
the species' range in BC, populations vary in their adaptation to local climate (assuming that
growth sensitivity can be used as one measure of local adaptation) along a precipitation and
heat-moisture gradient. Thus, sensitivity to annual precipitation may be a species-wide
genetic trait that is expressed more strongly as a site becomes warmer and/or drier.

4.2.2 High-elevation Douglas-fir populations
High-elevation Douglas-fir populations appear to be unique from others in this study in that
cold winter/annual temperatures and snowfall tend to limit growth more than annual
precipitation, especially over the latter half of the past century. The increase in growth
sensitivity to temperatures over that time was somewhat unexpected, as we hypothesized this
signal would weaken in the growth records as temperatures increased and became less
limiting in these cold environments. This type of unexpected growth response may represent
the difficulty of predicting ecosystem responses to changes in a single variable (e.g.,
temperature), as growth may be controlled more by the integrated influence of temperature
and precipitation changes rather than temperature alone (Littell and Peterson 2005). For
example, in high-elevation environments, growth benefits from warmer temperatures over
the past 50 years may have been buffered by limitations imposed by higher precipitation and
cloudiness (BC MoE 2007, Henderson-Sellers 1992, Dai et al. 1999, Cayan et al. 1998,
Zhang et al. 2007) over the same period, resulting in new (and unexpected) growth
sensitivities. When assessing species vulnerability in complex habitats such as highelevation environments where multiple climatic variables are shown to limit growth, the

137

integrative effect of temperature and precipitation changes on growth should be considered as
much as possible.

Forest management must also consider the growth influence of climate processes at various
spatial scales (Littell and Peterson 2005, Tessier, 1989). The strong synchronization between
high-elevation growth patterns and climate-growth relationships in the Northern and
Southern Interior (400+km apart) over the 20 th century likely reflects strengthening
macroregional climatic processes (e.g., anthropogenic temperature/precipitation trends,
ocean-atmosphere climate systems) that can override the influence of local factors such as
topography, soils, and aspect on growth (Littell and Peterson 2005, Tessier 1989). When
large-scale climatic processes weaken, growth patterns across large areas will diverge and
instead reflect local climate conditions (Littell and Peterson 2005, Tessier 1989). Shifts
between macroregional and local climatic influences are important drivers of forest
productivity and should be examined further.

The influence of quasi-periodic ocean-atmosphere climate systems on BC local winter
conditions appears to be strongly reflected in high-elevation Douglas-fir growth. The
Pacific/North American teleconnection index (PNA) was reflected similarly in both regions,
whereas the Pacific Decadal Oscillation (PDO) and El Nino/Southern Oscillation (ENSO)
indices had strong influences on southern populations but not northern. These indices may
be useful in predicting high-elevation growth rates and climate-growth relationships as well
as increasing our understanding of low-frequency large-scale climatic influences on
ecosystem processes. A better understanding of how PNA differs from PDO and ENSO in

138

its influence on local conditions may help elucidate specific climatic parameters that limit
growth in both regions.

High-elevation Douglas-fir productivity may benefit from warmer temperatures, less
snowfall, and longer growing seasons, and populations from other locations may provide
seed source for new forests that are suited to future conditions in these environments. More
research would be required, however, before this species is artificially regenerated
extensively in such areas. Douglas-fir is especially vulnerable to harsh conditions such as
snow press and frost events during early developmental stages (Jull 1999), and unique
silviculture strategies may be required if it is to regenerate successfully and form an
important timber species at high-elevations.

4.3 Research limitations and future directions
The recent development of high-resolution spatiotemporal climate normals and historical
data models such as ClimateBC is an important step in assessing species vulnerability to
climate change (BC MoFR 2006) across wide biophysical ranges (Littell 2006, Peterson and
Peterson 2001). To our knowledge, this is the first study of tree radial growth responses
along such a broad climatic gradient in British Columbia, and we recommend this approach
be applied to other tree species in the province. Species such as western red cedar {Thuja
plicata Donn ex D. Don), western hemlock (Tsuga heterophylla (Raf.) Sarg.), and larches
(Larix spp.), whose range (and possible climatic) limits are found in the province of BC, as
well as other species identified as particularly vulnerable to climate change (Hamann and

139

Wang 2006) could be targeted to provide additional information regarding responses to
climate change both in marginal populations and across climatic gradients.

One of the limitations of the standard dendroecological approaches used in this study is that
only relative growth rates are analyzed and compared between sites (i.e., measurements are
converted into a unitless index to form the growth chronology). Estimates of productivity
impacts from climate change would be potentially improved by using absolute growth rates
(e.g., millimetres per year) and converting them to a productivity measure common to forest
management, such as basal area increment or volume, which could be then used as part of a
plot-based sampling strategy to determine annual productivity increment per unit area (e.g.,
hectare). This approach would also allow for stand-level biological factors such as species,
age, crown-class and abiotic factors such as soil moisture/nutrient regime to be considered
when assessing climate-growth relationships, thus providing information at a finer scale
(both spatial and biological) that could complement data from larger scale studies such as this
one. The main drawback to this approach would be the increased time and resources
required for data collection, which may make this approach more feasible for local standlevel assessments of productivity impacts from climate variability.

4.4 Conclusions
This study revealed important Douglas-fir climate-growth relationships across the species'
range in the British Columbia Interior, complementing other studies examining similar
questions in different regions and at different spatial scales (e.g., Littell and Peterson 2005,
Case and Peterson 2005, Zhang and Hebda 2004). Although this study provides

140

deterministic information that may guide forest management, it is important to acknowledge
the high uncertainty regarding forest responses to climate change. Environmental changes
without historical precedent are expected to occur resulting in complex and unpredictable
ecosystem responses (Millar et al. 2007, Spittlehouse 2005). Forest managers will need a
wide range of tools to accommodate these changes (Millar et al. 2007, Spittlehouse and
Stewart 2003, Littell and Peterson 2005); flexible and adaptive strategies with careful
monitoring will be required to deal with complex and unpredictable future challenges.

141

4.5 Literature Cited
Aitken, S.N., S. Yeaman, J.A. Holliday, T. Wang, and S. Curtis-McLane. 2008. Adaptation,
migration or extirpation: climate change outcomes for tree populations. Evolutionary
Applications 1: 95-111.
Arno, S.F. 1991. Ecological relationships of interior Douglas-fir, p. 47-51. In Interior
Douglas-fir: The species and its management symposium proceedings. Baumgartner,
D. M. (eds). Department of Natural Resource Sciences, Washington State University,
Pullman, Washington.
Barber, V., G. Juday, and B. Finney. 2000. Reduced growth of Alaska white spruce in the
twentieth century from temperature-induced drought stress. Nature 405: 668-672.
BC MoE. 2007. Environmental Trends in British Columbia: 2007. State of Environment
Reporting. BC Ministry of Environment, Victoria, B.C.
BC MoFR. 2006. Preparing for climate change: Adapting to impacts on British Columbia's
forest and range resources. BC Ministry of Forests and Range, Victoria, BC.
Carrer, M. and C. Urbinati. 2006. Long-term change in the sensitivity of tree-ring growth to
climate forcing in Larix decidua. New Phytologist 170: 861-872.
Case, M.J. and D.L. Peterson. 2005. Fine-scale variability in growth-climate relationships of
Douglas-fir, North Cascade Range, Washington. Canadian Journal of Forest Research
35: 2743-2755.
Cayan, D.R., M.D. Dettinger, H.F. Diaz, and N.E. Graham. 1998. Decadal variation of
precipitation over Western Northern America. Journal of Climate 11: 3148-3166.
Christensen, J.H., B. Hewitson, A. Busuioc, A. Chen, X. Gao, I. Held, R. Jones, R.K. Kolli,
W.-T. Kwon, R. Laprise, V. Magaha Rueda, L. Mearns, C.G. Menendez, J. Raisanen,
A. Rinke, A. Sarr, and P. Whetton. 2007. Regional Climate Projections, p. 848-940.
In Climate Change 2007: The Physical Science Basis. Contribution of Working
Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate
Change. Solomon, S. D., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K. B.,
Tignor, M., and Miller, H. M. (eds). Cambridge University Press, New York.
D'Arrigo, R., R. Wilson, B. Liepert, and P. Cherubini. 2007. On the 'Divergence Problem' in
Northern Forests: A review of the tree-ring evidence and possible causes. Global and
Planetary Change 289-305.
Dai, A., K.E. Trenberth, and T.R. Karl. 1999. Effects of clouds, soil moisture, precipitation,
and water vapor on diurnal temperature range. Journal of Climate 12: 2451-2473.

142

Flato, G.M., G.J. Boer, W.G. Lee, N.A. McFarlane, D. Ramsden, M.C. Reader, and A.J.
Weaver. 2000. The Canadian centre for climate modelling and analysis global
coupled model and its climate. Climate Dynamics 16: 451-467.
Hamann, A. and T. Wang. 2006. Potential effects of climate change on ecosystem and tree
species distribution in British Columbia. Ecology 87: 2773-2786.
Henderson-Sellers, A. 1992. Continental cloudiness changes this century. GeoJournal 27:
255-262.
Jull, M.J. 1999. Douglas-fir silviculture "on the edge": silvicultural systems at the northern
range of the species, p. 34-44. In Ecology and management of interior Douglas-fir
{Pseudotsuga menziesii var glaucd) at the northern extreme of its range. Proceedings
of a workshop held 7-9 October 1996 in Fort St. James, British Columbia. Lousier, J.
D. and Kessler, W. B. (eds). University of Northern British Columbia, Prince George,
BC.
Littell, J.S. and D.L. Peterson. 2005. A method for estimating vulnerability of Douglas-fir
growth to climate change in the northwestern U.S. The Forestry Chronicle 81: 369374.
Lloyd, A.H. and C.L. Fastie. 2002. Spatial and temporal variability in the growth and climate
response of treeline trees in Alaska. Climatic Change 52: 481-509.
Meidinger, D. and J. Pojar. 1991. Ecosystems of British Columbia. BC Ministry of Forests,
Victoria, B.C.
Millar, C.L, N.L. Stephenson, and S.L. Stephens. 2007. Climate change and forests of the
future: managing in the face of uncertainty. Ecological Applications 17: 2145-2151.
O'Neill, G.A., A. Hamann, and T. Wang. 2008. Accounting for population variation improves
estimates of the impact of climate change on species' growth and distribution. Journal
of Applied Ecology 45: 1040-1049.
Ogden, A.E. and J. Innes. 2007. Incorporating climate change adaptation considerations into
forest management planning in the boreal forest. International Forestry Review 9:
713-733.
Pearson, R.G. and T.P. Dawson. 2003. Predicting the impacts of climate change on the
distribution of species: are bioclimate models useful? Global Ecology and
Biogeography 12: 361-371.
Peterson, D.W. and D.L. Peterson. 2001. Mountain hemlock growth responds to climatic
variability at annual and decadal time scales. Ecology 82: 3330-3345.

143

Rehfeldt, J. 1991. The genetic resource of Douglas-fir in the interior northwest, p. 53-62. In
Interior Douglas-fir: The species and its management symposium proceedings.
Baumgartner, D. M. (ed). Department of Natural Resource Sciences, Washington
State University, Pullman, Washington.
Spittlehouse, D.L. 2005. Integrating climate change adaptation into forest management.
Forestry Chronicle 81: 691-695.
Spittlehouse, D.L. and R.B. Stewart. 2003. Adaptation to climate change in forest
management. BC Journal of Ecosystem Management 4: 1-11.
St. Clair, J.B., N.L. Mandel, and K.W. Vance-Borland. 2005. Genecology of Douglas Fir in
Western Oregon and Washington. Annals of Botany 96: 1199-1214.
Tessier, L. 1989. Spatio-temporal analysis of cliamte-tree ring relationships. New
Phytologist 517-529.
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.
Zhang, Q.-B. and R.J. Hebda. 2004. Variation in radial growth patterns of Pseudotsuga
menziesii on the central coast of British Columbia, Canada. Canadian Journal of
Forest Research 34: 1946-1954.
Zhang, X., F.W. Zwiers, G.C. Hegerl, F.H. Lambert, N.P. Gillett, S. Solomon, P.A. Stott, and
T. Nozawa. 2007. Detection of human influence on twentieth-century precipitation
trends. Nature 448: 461-465.

144

a) MAT

b)AHM
S 10

^ 4
c
as

€ 2

ilAA
SI

CH

c)MAP

•nl - 1 H
SI

CH

d)MSP

Nl

SI

CH

e)PAS

Figure 4.1. Projected future climate for Douglas-fir sample sites in the three study
regions.Changes are an average of all sample sites in each study region, and represent the
difference between the 1961-1990 reference period and the 2020s, 2050s, and 2080s. The
ClimateBC model was used to generate future climate conditions, based on the CGCM2-A2x
scenario of the Canadian Centre for Climate Modeling and Analysis (Wang et al. 2006, Flato
et al. 2000). NI=Northern Interior, SI=Southern Interior, CH=Chilcotin Plateau, MAT=mean
annual temperature, AHM=annual heat-moisture index, MAP=mean annual precipitation,
MSP=mean summer precipitation, PAS=precipitation as snow.

145

U.U3

0.04

•
•
•

•

0.03 o
o
a
tz
o
w

•
•

•
•
•

0.02

•

V)

0.01

•••

•

0
n ni

i

i

i

10

20

30

—

i

i

40

50

AHM

Figure 4.2. Growth-annual precipitation regression coefficients against site annual heatmoisture index. Note: this relationship was quantified using log-transformed regression
coefficients (due to skewness) in Chapter 2 (Figure 2.6). Figure is presented to demonstrate
non-linear relationship and possible response threshold at AHM = -20.
0.05

0.04

o
ID
O
O

c

0.03

0.02

o
'»
v>
h, 0.01

^ . - .
-0.01

500

1000

1500

2000

2500

MAP

Figure 4.3. Growth-annual precipitation regression coefficients against site mean annual
precipitation. MAP=mean annual precipitation (mm). Note: this relationship was quantified
using log-transformed regression coefficients and MAP data (due to skewness) in Chapter 2
(Figure 2.9). Figure is presented to demonstrate non-linear relationship and possible
response threshold at MAP = ~700mm.

146