THE INFLUENCE OF HISTORICAL FORESTRY PRACTICES AND
CLIMATE ON THE SEDIMENT RETENTION FUNCTION OF WETLANDS
by
Katrina Caley
B.Sc. (Environmental Science), University of Ottawa, 2008
THESIS SUBMITTED IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
IN
NATURAL RESOURCES AND ENVIRONMENTAL STUDIES
(ENVIRONMENTAL SCIENCE)
THE UNIVERSITY OF NORTHERN BRITISH COLUMBIA
September 21, 2011
© Katrina Caley, 2011
1+1
Library and Archives
Canada
Bibliotheque et
Archives Canada
Published Heritage
Branch
Direction du
Patrimoine de I'edition
395 Wellington Street
Ottawa ON K1A0N4
Canada
395, rue Wellington
Ottawa ON K1A 0N4
Canada
Your file Votre reference
ISBN: 978-0-494-87573-5
Our file Notre reference
ISBN: 978-0-494-87573-5
NOTICE:
AVIS:
The author has granted a non
exclusive license allowing Library and
Archives Canada to reproduce,
publish, archive, preserve, conserve,
communicate to the public by
telecommunication or on the Internet,
loan, distrbute and sell theses
worldwide, for commercial or non
commercial purposes, in microform,
paper, electronic and/or any other
formats.
L'auteur a accorde une licence non exclusive
permettant a la Bibliotheque et Archives
Canada de reproduire, publier, archiver,
sauvegarder, conserver, transmettre au public
par telecommunication ou par I'lnternet, preter,
distribuer et vendre des theses partout dans le
monde, a des fins commerciales ou autres, sur
support microforme, papier, electronique et/ou
autres formats.
The author retains copyright
ownership and moral rights in this
thesis. Neither the thesis nor
substantial extracts from it may be
printed or otherwise reproduced
without the author's permission.
L'auteur conserve la propriete du droit d'auteur
et des droits moraux qui protege cette these. Ni
la these ni des extraits substantiels de celle-ci
ne doivent etre imprimes ou autrement
reproduits sans son autorisation.
In compliance with the Canadian
Privacy Act some supporting forms
may have been removed from this
thesis.
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
ii
Abstract
Wetlands provide beneficial functions and services (e.g. sediment retention, nutrient seques
tration) to downstream aquatic environments. The resiliency of these functions under dis
turbance conditions is, however, not fully understood. Two wetland-lake systems (Boswell
and Viewland) in the central interior of British Columbia whose contributing catchments
have historically been impacted by forestry practices were selected to examine how wetland
sediment retention responds to disturbance. Core chronologies and sedimentation rates were
calculated from unsupported 210Pb measurements using the Constant Rate of Supply (CRS)
model, and sediment source contributions were determined using a multivariate unmixing
model, for both wetlands and their downstream lakes. Sedimentation rates did not signifi
cantly change post-logging in either lake; however, the dominant source to Viewland Lake
changed from channel bank material to subsurface material. The increase in the proportion
subsurface material consistent with increases in dry density and magnetic susceptibility, and
decreases in median grain size and C:N. The bordering wetland was not found to contain any
material other than channel bank material. The ephemeral nature of the wetland channel,
as well as the length of the channel and the significant decrease in median grain size are
thought to have prevented sediment deposition, or increased the potential for resuspension
and further transport. Sedimentation rates were greatest near the inflow of Boswell wetland,
however, the strongest responses to forestry practices were observed near the wetland out
flow. Similarly, significantly lower median grain sizes could have limited deposition in the
upstream areas of the wetland. Increases in precipitation as snow and stream discharge in
addition to effects associated with forestry practices are thought to have been responsible for
driving sedimentation rates in both catchments; however, changes in source contributions
were likely the result of active forestry practices.
Contents
iii
Contents
Abstract
ii
Contents
iii
List of Tables
vi
List of Figures
viii
Acknowledgements
xii
1 Literature Review
1.1 Introduction
1.2 Wetlands
1.2.1 Features and functions
1.2.2 Water flow and sediment storage
1.2.3 Hydrophytic macrophytes
1.3 Sediment transfer
1.4 Sediment deposition and storage
1.5 Disturbance-response regimes
1.6 Forestry practices
1.7 Climate
1.8 Research questions and objectives
1.8.1 Research question 1
1.8.1.1 Wetland and lake sedimentation rates
1.8.1.2 Paleoenvironmental reconstruction
1.8.2 Research question 2
1.8.2.1 Changes in sediment provenance
1.8.2.2 Sediment source tracing
1.9 Thesis organization
1
1
3
3
4
5
6
7
8
11
12
13
14
14
14
16
16
16
18
2 Methodology
2.1 Study area
2.1.1 Boswell Lake catchment
2.1.2 Viewland Lake catchment
2.2 Sample collection and preparation
2.2.1 Wetland coring
2.2.2 Lake coring
2.2.3 Source materials
20
20
20
22
24
24
26
28
Contents
2.3
Radionuclides and core chronology
2.3.1 Origin of lead-210
2.3.2 Lead-210 dating models
2.3.3 Caesium-137
2.3.4 Core chronology
2.4 Proxy measurements
2.4.1 Bulk physical properties
2.4.2 Magnetic susceptibility
2.4.3 Particle size
2.4.4 Total carbon and nitrogen
2.4.5 Geochemistry
2.5 Climate and stream discharge
2.6 Sediment source tracing
2.6.1 Source groups
2.6.2 Multivariate unmixing model
2.7 Statistical analysis
2.7.1 Pre- versus post-logging
2.7.2 Climate and stream discharge trends
2.7.3 Correlations
iv
29
29
29
31
32
33
34
34
35
36
36
37
38
38
39
39
39
41
42
3 Results: Lake and wetland sedimentation rates
3.1 Physical descriptions
3.2 Lead-210 profiles and core chronologies
3.3 Boswell Lake catchment
3.3.1 Total sedimentation rates
3.3.2 Proxy indicators
3.3.3 Long-term changes in bulk physical properties
3.3.4 Hydrometerological influences and trends
3.4 Viewland Lake catchment
3.4.1 Total sedimentation rates
3.4.2 Proxy indicators
3.4.3 Long-term changes in bulk physical properties
3.4.4 Hydrometerological influences and trends
43
43
44
50
50
50
52
56
58
58
60
60
64
4 Results: Sediment source tracing
4.1 Source groups
4.2 Boswell Lake catchment
4.2.1 Composite fingerprint
4.2.2 Sediment source contributions
4.2.3 Correlations
4.3 Viewland Lake catchment
4.3.1 Composite fingerprint
4.3.2 Sediment source contributions
4.3.3 Correlations
66
66
68
68
71
72
75
75
75
78
Contents
v
5 Discussion
5.1 Boswell Lake catchment
5.1.1 Lake sediment
5.1.2 Wetland buffering function
5.2 Viewland Lake catchment
5.2.1 Lake sediment
5.2.2 Wetland buffering function
5.3 Importance of landscape position
5.4 Local versus regional effects
5.5 Study limitations
5.6 Future research directions
79
79
79
81
84
84
85
88
89
90
92
6 Conclusions and management implications
6.1 Conclusions
6.2 Management implications
6.3 Final remarks
94
94
97
98
Bibliography
99
A Bathymetric maps
110
B ClimateBC variables
113
C Microscope image of tephra
114
D Lead-210 dating models
115
List of Tables
vi
List of Tables
2.1
2.2
Morphometric characteristics of Boswell Lake and Viewland Lake. Informa
tion for Boswell Lake was taken from The Angler's Atlas (2010)
Summary of model assumptions. Adapted from Carroll k Lerche (2003). . .
3.1 Summary of the x2 values produced by the Sediment Isotope Tomography
(SIT) model. These values represent the goodness-of-fit between an observed
distribution (measured unsupported 210 activities) and a theoretical distribu
tion (modelled unsupported 210Pb activities). For a sample size of 10, two sam
ple distributions would be considered to be not significantly different (p>0.05)
if the x2 value was <16.9
3.2 Summary of the two-sample t-tests results comparing pre- and post-logging
total sedimentation rates (g cm-2 y_1) in Boswell Lake and wetland cores. In
BL-Pl, BL-D8, and BL-D10 the post-logging periods are above 4 cm, 7 cm,
and 7 cm, respectively. Values in brackets denote sample size
3.3 Summary of two-sample t-tests comparing the means of pre- and post-logging
periods for each proxy indicator measured in the Boswell Lake and wetland
cores. In BL-Pl, BL-D8, and BL-D10 the post-logging periods are above 4 cm,
7 cm, and 7 cm, respectively. Values in brackets denote sample sizes which
are consistent across all proxies
3.4 Summary of the two-sample t-tests results comparing pre- and post-logging
sedimentation rates (g cm-2 y_1) in Viewland Lake and wetland cores. In VLP1 and VL-D1, the post-logging periods are above 2 cm and 7 cm, respectively.
Values in brackets denote sample size
3.5 Summary of the two-sample t-test results comparing the means of pre- and
post-logging periods for each proxy indicator measured for the Viewland Lake
and wetland cores. In VL-P1 and VL-D1, the post-logging periods are above
2 cm and 7 cm, respectively. Values in brackets denote sample sizes. Total
number of samples are given under dry bulk density. Other values are given
where sample size was less than the total
3.6 Final model produced by the stepwise linear regression for Viewland Lake
(VL-P1) sedimentation rates. PAS=precipitation as snow
3.7 Final model produced by the stepwise linear regression for Viewland wetland
(VL-D1) sedimentation rates
4.1
Fuzzy k-means clustering results for Boswell Lake and Viewland Lake source
materials
23
31
48
50
54
58
62
64
64
67
List of Tables
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
Kruskal-Wallis H-test probabilities ( p ) for distinguishing surface, subsurface
and channel bank materials in the Boswell Lake catchment using individual
fingerprint properties. Mean concentration values are also given for each fin
gerprint property for each source type
Kruskal-Wallis H-test probabilities (p) for distinguishing surface, subsurface
and channel bank materials in the Viewland Lake catchment using individ
ual fingerprint properties. Mean concentration values are also given for each
fingerprint property for each source type
Fingerprint properties selected by the stepwise Multivariate Discriminant Func
tion Analysis to distinguish source types in the Boswell Lake catchment. . .
Percent relative errors and standard errors for the unmixing model calculations
for the Boswell Lake (BL-P1) and wetland (BL-D8) cores
Summary of the significant (p<0.05) correlations found between each source
material and sedimentation rates, and proxy indicators
Fingerprint properties selected by the stepwise Multivariate Discriminant Func
tion Analysis to distinguish source types in the Viewland Lake catchment. .
Percent relative errors and standard errors for the unmixing model calculations
for the Viewland Lake (VL-P1) and wetland (VL-D1) cores
Summary of the significant (p<0.05) correlations found between each source
material and each proxy indicator for the Viewland Lake core (VL-P1). ...
vii
69
70
71
73
74
75
76
78
List of Figures
viii
List of Figures
1.1
1.2
1.3
Conceptual diagram of a geomorphological disturbance, where: Ra = reaction
time; and, Rx = relaxation time. The sum of Ra and Rx equals the response
time. Diagram from Viles et al. (2008)
Conceptual diagram of a biogeomorphological response to precipitation. Dia
gram from Viles et al. (2008) (adapted from original by Knox, 1972)
A conceptual diagram outlining the sediment fingerprinting approach. Adapted
from Collins & Walling (2002) to this wetland focus
9
10
17
2.1
Map of the province of British Columbia. The rectangle indicates the approx
imate location of the Quesnel River Basin which is composed of three water
sheds: the Cariboo River Watershed; the Quesnel River Watershed; and, the
Horsefly River Watershed
2.2 Map of the Boswell Lake catchment. Forestry practices were active in the
catchment during two time periods: A) 1960-1975; and B) 1982-2008. Inset:
Outline of the Quesnel River Watershed. The star represents the approximate
location of the Boswell Lake catchment in the watershed
2.3 Map of the Viewland Lake catchment. Forestry practices were active in the
catchment in 1983 (A). Inset: Outline of the Horsefly River Watershed. The
star represents the approximate location of the Viewland Lake catchment in
the watershed
2.4 (a) Boswell Lake and (b) Viewland Lake and wetland coring locations. Codes
containing a 'D' indicate an open barrel core, 'P' refers to a percussion core,
and those with an 'E' denote a core taken with an Ekman dredge. Note: The
location of the stream containing Boswell wetland core BL-D13 was not shown
in original spatial dataset. This line feature was created by extracting point
locations from a Google Earth image of the catchment
2.5 Diagram of the open barrel corer used to retrieve wetland sediment cores. . .
2.6 Comparison of the 1- and 2-sample t-tests used to evaluate post-logging (F)
changes in the sediment profiles against average pre-logging (X) conditions.
For 1-sample t-tests, sedimentation rate and proxy values given by individual
post-logging 1 cm core slices (F,) represent the null hypothesis (//)
40
Unsupported 210Pb and 137Cs activity depth profiles for Boswell Lake and
wetland cores. Unsupported 210Pb error bars represent the sum of the total
210Pb and supported 210Pb errors. Values without errors were measured at
the minimum detectable limit of the gamma assay
46
3.1
21
24
25
27
28
List of Figures
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
Core chronologies for (a) Boswell Lake (BL-P1) and wetland cores (b) BLD8, and (c) BL-D10 produced by the Constant Rate of Supply (CRS) model.
Error bars were also calculated using the CRS model and represent the error
on each of the calculated dates
Unsupported 210Pb and 137Cs activity depth profiles for Viewland Lake and
wetland cores. Unsupported 210Pb error bars represent the sum of the total
210Pb and supported 210Pb errors. Values without errors were measured at
the minimum detectable limit of the gamma assay.
Core chronologies for (a) Viewland Lake (VL-P1) and (b) wetland (VL-D1)
cores produced by the CRS model. Error bars were also calculated using the
CRS model and represent the error on each of the calculated dates
Total sedimentation rates (calculated using the CRS model) for Boswell Lake
and wetland cores. The highlighted areas represent the periods of time that
forestry practices were present in the Boswell Lake catchment. Error bars on
the sedimentation rates represent the standard error calculated using the CRS
model. An error value could not be calculated for the bottom of the BL-D8
profile
The seven proxy indicators (dry bulk density, percent water content, magnetic
susceptibility, median particle size, total C, total N, and C:N) are shown over
time for the dated portion of each of the Boswell Lake and wetland cores.
The highlighted areas represent the years that forestry practices were present
in the Boswell Lake catchment. Core logs and general descriptions of the
sediment are also provided for each core (top left)
Long-term depth profiles of dry bulk density and percent water content for
(a) Boswell Lake (BL-Pl) and wetland cores (b) BL-D8, and (c) BL-D10.
Values are presented over depth as they extend beyond the dated region of
the sediment cores where 210Pb was not present in measurable concentrations.
Highlighted areas represent years that forestry activities were present in the
catchment. The date (2,410 yrs BP) provided at 56 cm is the location of the
Bridge River tephra layer in the lake core
Annual stream discharge (1924-2009) and mean annual precipitation (19012002) values. Stream discharge values are for Quesnel River at Likely, BC and
were taken from the Water Survey of Canada (Environment Canada). Mean
annual precipitation measurements are specific to the Boswell Lake catchment
and were modelled using ClimateBC. The small dashed line represents the
linear regression line for the full time series of stream discharge. The large
dashed lines are linear regression lines in between each set of breakpoints. . .
Total sedimentation rates (calculated using the CRS model) for Viewland
Lake and wetland cores. The horizontal line at 1983 represents the year the
Viewland Lake catchment was logged. Error bars on the sedimentation rates
were also calculated using the CRS model
ix
47
48
49
51
53
55
57
59
List of Figures
3.10 The seven proxy indicators (dry bulk density, percent water content, magnetic
susceptibility, median particle size, total C, total N, and the C:N) are shown
over time for the dated portion of each of the Viewland Lake and wetland
cores. The horizontal line represents the year the Viewland Lake catchment
was logged (1983). Core logs and general descriptions of the sediment are also
provided for each core (top left)
3.11 Long-term depth profiles of dry bulk density and percent water content for
(a) Viewland Lake and (b) wetland cores. Values are presented over depth as
they extend beyond the dated region of the sediment cores where 210Pb is not
present in measurable concentrations. Horizontal lines represent the year that
logging activities were present in the catchment. Although not show here, the
Bridge River tephra layer (2,410 yrs BP) occurred at 67 cm depth in the lake
core (VL-P1)
3.12 Precipitation as snow (mm) and sedimentation rates (g cm-2 y_1) over time
for the Viewland Lake core. Climate data are specific to the Viewland Lake
catchment area and were modelled using ClimateBC. Sedimentation rates were
calculated using the CRS model
Results of the Principle Component Analysis (PCA) of the fingerprint prop
erties for (a) Boswell Lake and (b) Viewland Lake sediment source materials.
F=forest, F_sub=forest subsoil, L=logged, L_sub=logged subsoil, R=road,
CB=channel bank. Biplots represent the first two principle components of
the PCA
4.2 Results of the multivariate unmixing model for the (a) Boswell Lake and (b)
wetland cores. Values on the secondary y-axis represent the dates calculated
using the CRS model for each 1 cm core slice containing detectable concentra
tions of 210Pb. Each date aligns with the bottom of its respective 1 cm core
segment
4.3 Results of the multivariate unmixing model for the (a) Viewland Lake and
(b) wetland cores. Values on the secondary y-axis represent the dates calcu
lated using the constant rate of supply model for each 1 cm core slice contain
ing detectable concentrations of 210Pb. Each date aligns with the bottom of
its respective 1 cm core segment. The asterisk (*) identifies core slices that
were not corrected for particle size due to a lack of material
x
61
63
65
4.1
5.1 The components of catchment connectivity (from Bracken & Croke (2007)). .
A.l Bathymetric map for Boswell Lake. Map was obtained online from the An
glers' Atlas
A.2 Bathymetric map for Viewland Lake. Map was created in ArcGIS using
latitude-longitude coordinates and water depths obtained during a depth sur
vey of the lake
67
72
77
89
Ill
112
List of Figures
xi
C.l Microscope image of the tephra found in both the Boswell Lake and Viewland
Lake cores. Tephra was identified as having originated from the Bridge River
event (ca. 2,410 calendar years BP) based on the glass shard morphology and
tephra colour
114
D.l Comparison of the 210Pb-based depth-to-age models (CIC, CRS, SIT) for (a)
Boswell Lake and wetland cores (b) BL-D8 and (c) BL-D10. Error bars are
not given to enhance the readability of the figure
D.2 Comparison of the 210Pb-based depth-to-age models (CIC, CRS, SIT) for (a)
Viewland Lake and (b) wetland cores. Error bars are not given to enhance
the readability of the figure
116
117
Acknowledgements
xii
Acknowledgements
I would like to express my sincere thanks to my supervisor, Phil Owens. Thank you for
giving me the opportunity to develop my own project, and allowing me to make the most of
this opportunity. Your guidance and support throughout this process has been invaluable,
and I appreciate the time and effort that you devoted to my education and this project.
Many thanks to the members of my committee: Dr. Brian Menounos, Dr. Ellen Petticrew,
and Dr. John Rex. Thank you for sharing your knowledge and expertise with me. Each of
you has provided me with feedback that has both helped me and challenged me, and I would
not have been able to develop the current thesis without it.
There are a number of people who helped me in the field and the lab, and whose assis
tance allowed this project to move forward. A big thank you to Rob Little for enduring the
"mozzies" over many days in soggy wetlands. Without your height and weight I never would
have taken a single sediment core. And to Ty Smith for your willingness to head out onto
frozen lakes with me and haul up tubes of dirt. Thank you to Rick Holmes and Bill Best of
the QRRC for always being so welcoming and supportive. Thanks to Dr. Paul Sanborn for
the use of your sediment corer. Thanks to Dr. Richard Jones and Dr. Klaus Kuhn for your
guidance during field work. Thanks to Dr. John Clague for the use of your lab, and to Dr.
Will Blake for radionuclide analysis. Thanks to Mel Grubb and Lyssa Maurer for taking the
time to teach me all you know about processing sediment cores.
I gratefully acknowledge the financial support I received from the FRBC-QRRC Landscape
Ecology Research Program, the NSERC Alexander Graham Bell Graduate Scholarship, the
Patrick Lloyd Graduate Scholarship, and the UNBC Master's Tuition Scholarship.
Thank you to everyone at UNBC for making Prince George more than just the city where I
go to school. Thanks to COL friends for organizing much needed recesses, and other hilari
ous distractions. I could not have asked for a better group of friends to share these last few
years with.
Many thanks to my family and friends for their continual love and support. Thank you to
my parents Jim, Sue, Betty and Joe, my brothers Stephen and Joey, and my sister Lisa and
her family. And to my grandparents Olympiada, Michael and Rhea for always telling me to
never stop learning.
Finally, I would like to thank two people whose friendships mean the world to me. To Danielle
and Rick, thank you for all of your encouragement, words of wisdom, and wonderful hugs.
You have both made the last stretch of this process more enjoyable, and I am so grateful to
have met you.
1
Chapter 1
Literature Review
1.1
Introduction
Wetlands regulate the flow of materials in many landscapes from terrestrial surfaces to
aquatic systems (Johnston, 1991). This ecosystem service is thought to be disproportionately
large compared to the actual wetland area (Hemond, 1988; Johnston et al., 1990). As a result,
the loss and degradation of wetlands can have significant repercussions for downstream water
quality and wetland habitat quality. For example, the sediment retention function also
regulates the flow of limiting nutrients, such as phosphorus which is typically bound to
the surface of sediment particles. Furthermore, the highest concentrations of phosphorus are
associated with the smallest size fractions of sediment (Owens & Walling, 2002) which require
low energy environments (such as those found in wetlands) to be deposited. When present in
high concentrations, phosphorus has been shown to produce eutrophic conditions (Schindler,
1977), often resulting in depleted oxygen concentrations. Thus, the loss of upstream wetlands
and their functions could result in the increased delivery of phosphorus to lakes, which might
otherwise have been deposited within upstream wetlands.
Wetland losses have been driven primarily by conversions to other land use activities,
such as agriculture, urban development, flooding for hydroelectric production, and drain
ing for pest management. Increasing densities of human settlements have been linked to
increased fragmentation of wetland area and greater distances between individual wetlands
(Gibbs, 2000). With increasing recognition of the important role that they play in the water
shed, wetlands and their functions have been receiving more attention. Attempts to protect
wetlands have included designating wetland areas for conservation purposes, identifying eco
Chapter 1. Literature Review
2
logically significant wetlands, and placing economic value on wetlands and those functions
that benefit society (Davis & Froend, 1999). Canada, for instance, has lost approximately
70% of its total wetland area. In an attempt to preserve the remaining wetland areas, it
has been estimated that based on wetland services (e.g. migratory bird habitat and water
quality improvement), Canada's wetlands have a value of approximately $19,580 per hectare
per year (British Columbia Ministry of Environment, 2010).
Wetlands are commonly defined with respect to their hydrological (i.e. water table loca
tion) and biological features (i.e. vegetation), with the biological features ultimately driven
by the hydrological regime. In a review on sediment storage in fluvial wetlands, Phillips
(1989) redefined wetlands in a geomorphic context. The author stated that "their presence
and extent is both a reflection and a determinant of the magnitude of sediment storage (or
remobilization) within a drainage basin". This definition suggests that any major hydrologic
or geomorphic changes in the drainage basin could have significant impacts on downstream
wetlands, and has been corroborated by the results of other studies. In a forested wetland in
West Tennessee, USA, Hupp &; Bazemore (1993) observed that the channelization of streams
in the upstream drainage basin resulted in less sediment deposition than in the unchannelized
streams. Flow constriction caused by channelization increased stream velocities preventing
sedimentation in the wetland.
Land use activities do not necessarily result in the loss of wetland area, but they do disturb
the surrounding area and have been found to alter hydrological conditions. Forestry prac
tices, for instance, have been linked to surface compaction, increased runoff, and increased
sediment production (Church & Eaton, 2001). Little information is currently available re
garding the long-term variability of the sediment storage function of wetlands, how this
function is impacted by forestry practices (which have been linked to increased suspended
sediment concentrations), and the subsequent impact on downstream water quality (Zedler
& Kercher, 2005). Additionally, few studies have focused on the status of wetland functions
pre- and post-disturbance. Long-term studies are therefore needed to address the capacity of
Chapter 1. Literature Review
3
wetlands to act as buffers under upstream disturbance conditions. This requires an adequate
assessment of baseline functions, as well as monitoring during and after land use activities.
Due to high costs and limited resources, it is unreasonable to think that long-term mon
itoring data would be available for all wetlands and lakes residing in logged catchments.
Other techniques are available that enable the development of a long-term historical dataset.
Specifically, paleoenvironmental reconstruction attempts to reconstruct past environmental
conditions using sediment records. This requires the measurement of "proxy indicators"
which allow inferences to be made about historical environmental conditions based on cur
rent knowledge of these indicators, and the processes that control their behaviour in a par
ticular environment (Smol, 2008, 2010). This type of approach provides insight into baseline
environmental conditions, as well as responses to natural and/or anthropogenic disturbances.
The aim of this thesis is to use paleoenvironmental reconstruction techniques to evaluate
the sediment retention function of two wetlands in the central interior of British Columbia,
and how they have responded to changing hydrologic and geomorphic conditions as a result
of historical forestry practices. The following sections will provide an overview of literature
that has been published on wetland characteristics and the sediment storage function of
wetlands, sediment transport and storage, disturbance response regimes, impacts of forestry
practices and climate on sediment yields, as well as key analytical techniques that have been
used to complete the present study. Research questions and objectives are also provided at
the end of the literature review.
1.2
1.2.1
Wetlands
Features and functions
A general definition of a wetland includes three main characteristics: all wetlands are tem
porarily or permanently inundated, possess hydric (i.e. oxygen deprived) soils, and are inhab
ited by rooted vegetation that is adapted to these conditions. Categories of wetlands often
Chapter 1. Literature Review
4
rely on hydrology, vegetation type, and pH to define their boundaries. Marshes are typically
inhabited by herbaceous vegetation, while swamps are capable of supporting woody species.
Peatlands exist where the soil is rich in organic matter and the water table is at or below the
ground surface. Acidic and alkaline conditions belong to bogs and fens, respectively. "True
bogs" are more specifically bogs with no defined inflow or outflow, and receive water only
through precipitation which is lost only by evaporation. Floodplains and riparian wetlands
are situated along stream and river banks where they are able to intercept lateral runoff
from uplands. Floodplains also receive overbank flood water when rivers exceed bankfull
levels (Mitsch & Gosselink, 2000). These definitions provide a brief overview of the diver
sity of wetlands that exist. Sub-types of wetlands lie within each of these categories whose
definitions are a function of dominant vegetation type, ultimately driven by climate.
Wetlands are commonly viewed as transitional environments between terrestrial and
aquatic ecosystems, filtering or buffering downstream ecosystems and improving water qual
ity. Wetlands also act as carbon sinks storing organic matter, support a diverse array of
biota, and mitigate flood events (Hemond, 1988; Zedler & Kercher, 2005). Their ability to
enhance sedimentation, trap nutrients, metals and contaminants, and improve water quality
has also been recognized as an important tool for the management of wastewater (Srivastava
et al., 2008) and agricultural runoff (Owens et al., 2007). However, the function(s) that a
wetland is able to support is dependent on wetland type, and more importantly, its position
in a watershed (Johnston et al., 1990).
1.2.2
Water flow and sediment storage
The movement, or advective-dispersive transport, of particulate matter through wetlands is
largely a function of water flow, and the presence (or absence) of aquatic vegetation (Huang
et al., 2008). Wetlands that are characterized by inundation or ponding typically promote a
sediment trapping or buffering function (Johnston, 1991; Hupp & Bazemore, 1993) as these
low-lying regions significantly reduce water velocity and facilitate sedimentation (Hemond,
Chapter 1. Literature Review
5
1988). An urban wetland studied by Brown (1985) was found to reduce peak discharge at
its outlet by 12-70% compared to the inlet. In the same study, the author also observed that
sedimentation rates were enhanced during peak discharge events, especially those associated
with early-May to late-June storms when 56-70% reductions in peak discharge were recorded.
This is consistent with the findings of Johnston et al. (1990) who observed that wetlands
were associated with higher concentrations of suspended solids during periods of high flow.
1.2.3
Hydrophytic macrophytes
According to Manning's equation, mean flow velocity (in main channels with low slopes) is
inversely proportional to surface roughness (also known as Manning's roughness coefficient),
and is largely controlled by the presence or absence of vegetation, which is also influenced
by slope and substrate type. Hydrophytic macrophytes, or flood-tolerant plants, play an
important role in water flow dynamics and sedimentation in wetlands (Clarke, 2002). Water
flow reduction and sediment accumulation by macrophytes are dependent on vegetation
density and type (Dawson, 1978).
Petticrew k, Kalff (1992) observed that as leaf area
index (LAI) increased, water velocity near the lake bed decreased. Leaf pattern (Clarke,
2002; Huang et al., 2008), plant morphology, as well as shoot movements and flexibility
(Sand-Jensen & Pedersen, 1999) can also result in small-scale velocity variations which can
be accompanied by a decrease in turbulence. In addition to significantly reducing water
velocity (Sand-Jensen & Pedersen, 1999), dense stands of vegetation prevent bed scouring
and sediment resuspension (Sand-Jensen & Mebus, 1996; Braskerud, 2001). Although these
studies focused on lake environments, it is likely that hydrophytic macrophytes would have
the same effect on water velocity, and sediment trapping in wetlands.
Flow modification as a result of wetland vegetation is also linked to the trapping of fine
organic and inorganic particles (Clarke, 2002). LAI was found to explain 74% of the variation
in the percent of clay accumulating below macrophyte stands in lake bottoms (Petticrew &
Kalff, 1992). This suggests that an increase in the density of macrophytes would facilitate the
Chapter 1. Literature Review
6
deposition and retention of fine sediments. The reconstruction of historical sedimentation
rates in several lakes revealed that lakes containing macrophytes had greater accumulations of
sediment over time compared to those which were relatively macrophyte-free (Brenner et al.,
1999). However, the role of macrophyte stands in sediment retention is often temporary, as
their predominant function is to stabilize the sediment bed through the binding effects of
their roots (Sand-Jensen, 1998). Similarly, Phillips (2003) suggested that most wetlands are
temporary storage sites of sediment to buffer the output to downstream waterbodies.
1.3
Sediment transfer
Sediment transfer is a two stage process that involves sediment production or mobilization,
and subsequently sediment transport by a medium capable of entraining the sediment par
ticles. Firstly, sediment production requires material to be eroded from a terrestrial surface.
Surface erosion may occur as a result of bank erosion, raindrop erosion, sheetwash, soil creep,
or rapid mass movement (Pye, 1994; Church & Eaton, 2001). Several local and regional fac
tors moderate hillslope erosion, including lithology, vegetation cover, availability of rock and
soil, slope length, steepness and roughness (Pye, 1994). Second, sediment transport typi
cally occurs during storm events when a sufficient volume of water is present to overcome
the shear stress acting on a particle (Pye, 1994). In an undisturbed forested environment,
surface erosion due to running water is rare because of interception by vegetation and subse
quent infiltration (Lehre, 1982; Swanson et al., 1982). However, the intensity and frequency
of weather events also plays an important role in controlling rates of erosion. Blais et al.
(1998) found that lake sedimentation rates decreased an average of 80% during a year when
annual runoff experienced a 63% reduction.
Mobilized sediment eventually enters either dispersive or channelized pathways (Bracken
& Croke, 2007). The ability of these pathways to transport sediment is a function of stream
power (i.e. stream volume and velocity), and particle size. Dispersive pathways are associated
with overland flow and are characterized by diffuse connectivity. These pathways exhibit a
Chapter 1. Literature Review
7
branching structure and tend to lose volume and power as they travel down a hillslope. As
a result, they lose the ability to transport greater volumes of sediment, and larger particles
such as sand and gravel. Furthermore, due to their systematic branching and limited power,
there is a low probability of these pathways reaching streams via overland flow (Church &
Eaton, 2001). On the other hand, channelized pathways are typically longer (Croke et al.,
2005), accumulate water with distance, and are more likely to be directly connected to the
fluvial system (Church & Eaton, 2001).
The "sediment delivery ratio" relates the amount of sediment delivered to the catchment
outlet or the sediment yield (t km-2 y_1) to the gross erosion in the basin (t km"2 y_1)
(Walling, 1983). In general, sediment yield has a positive relationship with slope angle as
the degree of inclination provides the potential energy for runoff (Pye, 1994). For smaller
catchment sizes the relationship between slope angle and delivery is significant, however, it
tends to change with increasing catchment size. Larger basins provide more opportunities for
temporary sediment storage as compared to smaller, less complex catchments. Furthermore,
hillslopes in larger basins have been found to be decoupled from the fluvial network which
again interrupts sediment delivery to the catchment outlet (Phillips, 1995). Other factors
affecting this relationship include sediment source characteristics, drainage patterns, channel
conditions, vegetation cover, and land use (for a comprehensive review see de Vente, 2007).
1.4
Sediment deposition and storage
The deposition and storage of sediment in aquatic environments depends on the properties
of both the depositional environment and the material being transported. More specifically,
Stokes' Law states that the deposition of a sediment particle strongly depends on the size of
the sediment particle and the viscosity of the transport medium (Pye, 1994); however, this
relationship only holds true for low Reynolds numbers (i.e. low turbulence). As the size of
a sediment particle increases, the energy required to keep it suspended in the water column
also increases. Therefore, as the energy of the system decreases (i.e. velocity decreases), its
Chapter 1. Literature Review
8
ability to support larger particles will diminish resulting in sediment sorting (Powell, 1998).
This relationship assumes that sediment particles are transported as discrete particles and
does not account for the behaviour of aggregated particles in a water column. Aggregated or
flocculated particles (floes) are comprised of organic and inorganic material which are bound
together by surface adhesion. Floes have varied sizes, shapes and densities, and as a result,
different settling velocities from their discrete counterparts (Droppo et al., 1997). Although
the formation and transportation of floes have been found to have a significant impact on
the deposition of both organic and inorganic materials, they will not be considered in the
present study. Compaction and degradation of the material in the sediment cores would
likely not provide an accurate representation of the material at the time of deposition.
With respect to the characteristics of a depositional environment, sediment storage oc
curs when the energy of the system is low enough to facilitate deposition, and where there
is minimal re-suspension as a result of wave action and bed scouring. Lakes typically redis
tribute sediment from shallower areas towards the deepest point of the basin (Davis, 1968,
1973), also known as "sediment focusing". Consequently, lakes often provide excellent en
vironments for reconstructing historical sedimentation rates and sediment yields. Wetlands
have also been recognized as areas of sediment deposition (Johnston, 1991), however, storage
in these systems is often temporary. Wetlands regulate the movement of sediment through
the watershed and buffer downstream environments against environmental change (Phillips,
1989, 2003).
1.5
Disturbance-response regimes
The processes of sediment transport and delivery under natural and disturbance (e.g. forestry
practices) conditions have been reviewed above. As well, the quantity of sediment delivered
to the catchment outlet was considered in terms of the sediment delivery ratio, and catchment
size and complexity. However, the timing of sediment delivery, and the concept of equilibrium
states, have not yet been discussed. Viles et al. (2008) proposed a conceptual model for
9
Chapter 1. Literature Review
a generalized geomorphological disturbance response of a system (Fig. 1.1). The model
illustrates that there is often a lag between the timing of the disturbance, or "forcing", and
the observed response. Lags in the response have been attributed by Viles et al. (2008) to
be the result of stabilizing effects, or characteristics and/or processes in the catchment that
limit erosion and/or increase sediment storage.
Response time
Response time
Rx
Ra
Rx
Ra
Ra
KM
c
1
1
0>
Forcings
Time
Figure 1.1: Conceptual diagram of a geomorphological disturbance, where: Ra = reaction
time; and, Rx = relaxation time. The sum of Ra and Rx equals the response
time. Diagram from Viles et al. (2008).
In the conceptual model by Viles et al. (2008) (Fig. 1.1), forcings are intended to represent
any disturbance, including storm events, climate change, and human activities. Figure 1.2
(originally from Knox, 1972) has been broken down into several components to illustrate a
possible biogeomorphological response to a climate forcing (i.e. fluctuating precipitation).
The increased growth of vegetation, as a result of increased precipitation, has an inverse
relationship with erosion potential as vegetation growth provides a stabilizing effect by main
taining a strong soil structure through the binding effects of the root network. This biological
response thus translates into a negative feedback on sediment delivery by mitigating hillslope
erosion. A similar response has been observed in field studies where despite the presence of
active logging, water yield did not increase, and was attributed to the rapid reestablishment
Chapter 1. Literature Review
10
of vegetation (Paterson et al., 1998).
Precipitation
Erosion potential
x
O
X
/
/•.
Vegetation
\
A.
v
Geomorphic work
(sediment yield)
TIME
Figure 1.2: Conceptual diagram of a biogeomorphological response to precipitation. Di
agram from Viles et al. (2008) (adapted from original by Knox, 1972).
The previous example has shown how stabilizing effects can produce either a lag or
dampen the response to a catchment disturbance. However, their ability to do so also depends
on the climate regime, the magnitude of the disturbance, and the cumulative impacts of
destabilizing effects (i.e. processes that promote erosion). Forcings can also produce alternate
stable states wherein the relaxation phase does not return the system to its previous condition
and a new level of equilibrium is obtained (Owens et al., 2010). State changes in terms of
sediment delivery can relate to either the quantity or quality of the sediment, or both.
While the quality of the sediment (i.e. presence of contaminants or elevated concentrations
of metals) is typically influenced by human activities (e.g. mining), sediment quantity can
be influenced by both human activities and hydroclimatic processes. In a review comparing
the effects of landscape disturbance and climate change on erosion, Slaymaker (2001) argued
that, with the exception of the polar regions, human land use activities have a much greater
impact on global erosion rates than climate change. Bracken & Croke (2007), however,
suggested that local climate conditions and storm events provide the conditions required to
Chapter 1. Literature Review
11
generate runoff, which in turn mobilize and deliver sediments to downstream waterbodies.
1.6
Forestry practices
The goal of the present study is to evaluate the sediment trapping function in wetlands
using historical forestry practices as an indicator of disturbance. Therefore, it is important
to understand how forestry practices impact forest hydrologic and geomorphic processes.
The following provides a brief overview of forestry practices and their effects on water yield
and sediment production.
Undisturbed forests are vital elements in the water cycle as they intercept rainfall, pro
mote infiltration, and contribute water vapour to the atmosphere through evapotranspiration. Spittlehouse (2006a) estimated that 60 to 65% less rainfall reaches river systems when a
tree canopy is effective in intercepting rainfall. It has been well documented that the removal
of forest cover is strongly related to increases in water yield (Harr et al., 1982; Keppeler &;
Ziemer, 1990; Stednick, 1996), although this relationship tends to be seasonal and is strongly
influenced by precipitation (Bosch & Hewlett, 1982).
The term "forestry practices" is used here to include forest harvest as well as other
associated activities. Forestry roads are known to cause soil compaction and reduced water
infiltration (Croke et al., 1999), both of which alter the volume and distribution of overland
flow (Pike & Scherer, 2003), and change the magnitude and timing of peak flows following
storm events. Jones & Grant (1996) found that a 25% patch-cut watershed in the western
Cascades, Oregon, USA, with 6% road cover increased peak flows to the same extent as a
clearcut watershed. Additionally, peak flows remained 25% greater than before logging and
road construction over the following 25 years.
Several studies have identified forestry roads as a major source of fine-grained sediment
from logged catchments (Reid & Dunne, 1984). A detailed geochemical analysis of lake sed
iment in central British Columbia by Christie & Fletcher (1999) revealed that the sediment
did not originate from cut blocks, but instead from forestry roads and culverts. Cut blocks
Chapter 1. Literature Review
12
have also been reported to increase sediment transfer, however, significant contributions of
sediment are more likely to occur as a result of subsequent mass wasting events. Generally,
these occur several years post-logging when root networks and other organic debris have
decomposed. Furthermore, the impact that logging and roads have on fine-grained sediment
production is influenced by landscape position (Tague & Band, 2001), the degree of con
nectedness to the stream network, and the importance of the road within the overall road
network (Sheridan & Noske, 2007).
Ditches and culverts installed under roads are common features in a forest road network
and are used to prevent the flooding and erosion of roadways, and to channel the flow of
existing streams. As a result, drainage systems also provide direct connections between the
road network and fluvial pathways (Croke & Mockler, 2001). Moreover, road construction
coupled with the presence of culverts tends to produce channelized pathways which increases
the drainage density of the catchment (Wemple &; Jones, 1996). This suggests that ditches
and culverts would also increase the amount of fine-grained
sediment mobilized, and the
probability of it reaching the fluvial system.
1.7
Climate
In addition to anthropogenic land disturbances, climate has been shown to influence sedi
ment erosion and delivery (Walling, 1999). Therefore, it is important to consider the effects
of varying climatic conditions on sediment retention in wetlands in addition to the effects
of historical forestry practices. Event-based processes, such as rainfall and spring freshet
can increase the potential for sediment erosion and delivery to occur. During stream mon
itoring of Fitzsimmons Creek in southwestern British Columbia, Canada, Menounos et al.
(2006) found that the highest measured suspended sediment concentration occurred dur
ing a rainfall-driven bank-full discharge event. From a paleoenvironmental reconstruction
perspective, intense rainfall events or a fast spring freshet could increase the deposition of
sediment in downstream wetlands and lakes resulting in greater sedimentation rates for that
Chapter 1. Literature Review
13
time period. Alternatively, large decreases in rainfall can also have an impact on sedimenta
tion rates which have been shown to drastically decline during years of little rainfall (Blais
et al., 1998).
The magnitude and frequency of precipitation events is often the result of larger scale
processes, such as the El Nino/La Nina-Southern Oscillation (ENSO). These processes typ
ically have a longer periodicity and affect a larger area or region. For example, ENSO is
a quasiperiodic climate process that occurs approximately every five years. Throughout
most of North America, El Nino results in warmer and drier winters and summers, while
La Nina produces cooler and wetter winters and summers (Ropelewski & Halpert, 1986).
With respect to sediment erosion and delivery, larger scale climatic processes can enhance
or subdue the intensity and/or the timing of rainfall events or the spring freshet. Dery et al.
(2009) found that a phase change to a cool phase in Arctic Oscillation resulted in increased
stream discharge in North American rivers. An intensification of the hydrological regime
could provide the necessary energy to increase sediment erosion on the landscape, as well as
increase the delivery of sediment to the stream network.
1.8
Research questions and objectives
The following research questions and objectives focus on evaluating the sediment reten
tion function of two wetland buffers in the central interior of British Columbia. To evalu
ate this wetland function, wetland sedimentation rates, as well as the sedimentation rates
of their downstream lakes, were determined using paleoenvironmental reconstruction tech
niques. The goal of this thesis was to establish if the selected wetland buffers provided the
downstream lake with a sediment buffering function, and whether or not that function was
compromised by the disturbance caused by upstream historical forestry activities.
Chapter 1. Literature Review
1.8.1
Research question 1
1.8.1.1
Wetland and lake sedimentation rates
14
The first set of research questions relate to sediment storage in wetland buffers. Do wetland
sedimentation rates increase after the onset of forestry activities? Also, are sedimentation
rates in downstream lakes affected by forestry activities?
These questions address the sediment buffering function of the wetland and its capacity to
function "normally" when its contributing catchment area is disturbed by forestry activities.
Assuming that the majority of sediment-bearing runoff that reaches the lake must pass
through the wetland first, then any increase in lake sedimentation rates would suggest that
either the wetland does not perform a sediment buffering function, or that the capacity of
that function has been exceeded. The objective was then to calculate sedimentation rates
for both the wetland and the lake to characterize baseline sedimentation rates against which
post-harvest rates could be compared.
1.8.1.2
Paleoenvironmental reconstruction
One aim of paleoenvironmental reconstruction is to develop a historical dataset consisting of
several lines of evidence which guide the interpretation of a system and its changes over time.
One of the advantages of using this approach is that a long-term record can be produced
for a system for which no monitoring data exist. The natural variability of the system (e.g.
lake, wetland) along with responses to catchment disturbances can thus be reconstructed.
Oldfield (1977) developed a conceptual model identifying lakes as ideal environments for
reconstructing past conditions. While lakes are not closed environments, they have been
shown to focus sediment toward the deepest areas, continuously creating an archive of the
material delivered from hillslopes, river channels and the atmosphere, and of the material and
organisms produced in situ. Retrieval of an intact sediment profile from the deepest point
of a lake should therefore provide a long-term record of the conditions in the surrounding
Chapter 1. Literature Review
15
catchment area and the lake itself. Other environments, such as wetlands and floodplains,
have since been recognized as being able to provide a similar record of long-term change;
however, other features and processes (e.g. vegetation, water level) affect the temporal and
spatial distribution of sediment in these systems.
The collection and interpretation of the long-term dataset require the measurement of
proxy indicators which, as mentioned earlier, allow inferences to be made about historical
environmental conditions based on current knowledge of these indicators, and the processes
that control their behaviour in a particular environment (Smol, 2008, 2010). The selec
tion of a set of proxy indicators should be driven by the research questions and objectives.
Commonly used proxies include: bulk physical characteristics; mineral magnetic properties;
grain size and shape; geochemical and nutrient concentrations; radionuclides; isotopic trac
ers; and, remnants of organisms which are not susceptible to physical breakage or chemical
dissolution (Smol, 2008). In order to understand the changes of these measurements in a
temporal context, core chronologies must be developed. Core chronologies can be developed
using radionuclide activities found in the sediment; most frequently used is the unsupported
component of 210Pb.
The development of a dataset using paleoenvironmental reconstruction techniques is a
powerful tool that can provide information on long-term environmental change, as well the
impacts of human activities on natural processes (e.g. sediment delivery), and environmental
quality. A long-term perspective is especially important when attempting to establish guide
lines for restoration projects (Foster et al., 2011), as guidelines based on recent monitoring
data may be inaccurate due to recent disturbances, cumulative effects, and the establishment
of alternate stable states (Owens et al., 2010).
Chapter 1. Literature Review
1.8.2
Research question 2
1.8.2.1
Changes in sediment provenance
16
The second research question pertains to the source of the sediment being delivered to the
wetlands and lakes. Do the relative contributions of sediment source materials identified
within each catchment area change as a result of forestry activities?
If sedimentation rates in either the wetland, lake, or both are altered, then it is important
to evaluate a suite of alternate, potential drivers of that change, since it may not be related
to forestry practices. Climatic factors, such as precipitation, are known to play a significant
role in sediment delivery, sometimes having a larger impact on sedimentation rates than
landscape disturbances (Blais et al., 1998). One would suspect that if forestry practices
were driving increased sedimentation rates, then an increase in the relative proportion of
subsurface soil material would occur as a result of erosion of surface soil exposing underlying
subsurface materials (Thompson et al., 1975). Similarly, the construction and use of roads
would increase the delivery of subsurface sediment. By identifying and characterizing the
sediment sources in the catchment, it is then possible to determine the relative contribution
of each to the wetland and lake, and how they change over time relative to the timing of the
disturbance. However, this assumes that the fluxes of sediment from sites of intermediate
storage do not change.
1.8.2.2
Sediment source tracing
Sediment source tracing has been used in many studies to identify the impact of different land
use types on contemporary suspended sediment loads (Walling & Woodward, 1995; Collins
et al., 1998). Long-term studies have also been carried out which used sediment stored within
depositional environments, such as floodplains (Owens et al., 1999), to reconstruct historical
changes in sediment sources over time. Figure 1.3 illustrates the main principles on which
the sediment source tracing process is founded. The diagram has been adapted to match the
Chapter 1. Literature Review
17
research questions raised by this study.
Effective PPT event
Erosion of catchment
sediment sources
Mixing process during
sediment delivery
Surface &
sub-surface
Source types
Land use types &
channel banks
Comparison of source material
& wetland/lake core samples
using fingerprint properties
Sediment source
ascription
Suspended sediment
deposition in
wetlands/lakes
Figure 1.3: A conceptual diagram outlining the sediment fingerprinting
Adapted from Collins & Walling (2002) to this wetland focus.
approach.
The process of sediment source tracing requires that sediment source types can be dif
ferentiated according to their fingerprint properties. A simple, and commonly used set of
source groups consists of: surface soil material; subsurface soil material; and, channel bank
material. Where well-defined land use activities are present in the catchment (e.g. agricul
ture, mining, urban areas and roads), many studies rely upon an a priori selection of source
groups. Others have used statistical methods to verify the accuracy of the selected source
groups with respect to the selected fingerprint properties (Hatfield k, Maher, 2009). It is also
important to consider the underlying bedrock, and if source areas extend beyond a single
bedrock type as this could impact the geochemical and mineralogical composition of the
eroded sediment (Walling & Woodward, 1995).
Various fingerprint properties have been used to characterize source groups, including soil
geochemistry (Foster, 1994), radionuclides (Walling et al., 1993), nutrients (Walling et al.,
2008), mineral magnetic properties (Yu & Oldfield, 1989), and colour (Martmez-Carreras
et al., 2010). The selection of a set of properties should be driven by the characteristics
Chapter 1. Literature Review
18
of the catchment(s) being studied, surrounding land use types, and ultimately the research
questions being asked. A composite fingerprint is then statistically selected from the full
set of fingerprint
properties. Ideally, the final composite fingerprint consists of several pa
rameters from more than one property type (Collins & Walling, 2002). Once an appropriate
composite fingerprint has been been identified, a multivariate unmixing model can be used
to calculate relative contributions of each source material.
If a paleoenvironmental approach is being taken then the nature of the properties must
also be considered in concert with the characteristics of the depositional environment. The
challenge with reconstructing sediment source contributions over time is that fingerprint
properties behave differently after deposition than when suspended in a water column (Owens
et al., 1999). Post-depositional changes (or diagenesis) such as decomposition, physical
mixing and bioturbation alter the nature of the fingerprint
properties and their vertical
distribution in the sediment profile. It is therefore necessary to select properties which behave
conservatively not only during transport, but also in a depositional environment (Motha
et al., 2002). Physical, mineral magnetic, radionuclide and geochemical properties tend to be
conservative in a sedimentary environment, and have been widely used in both contemporary
and historical studies. However, due to sorting effects during transport differences in grain
size need to be accounted for by targeting a specific size fraction (e.g. <63 nm) for analysis
(Carter et al., 2003; Foster et al., 2008), and by including a particle size correction in the
unmixing model (Collins et al., 1997).
1.9
Thesis organization
The chapters of this thesis on the effects of historical forestry practices and climate on the sed
iment retention function of two wetlands in the central interior of British Columbia have been
organized according to a traditional thesis style which follows the scientific method. Chap
ter two describes the study sites and methodology used to address the research questions.
Chapters three and four summarize the results from the paleoenvironmental reconstruction
Chapter 1. Literature Review
19
and sediment source tracing procedures, respectively. Chapter five provides a discussion on
each of the study sites by interpreting the results in chapters three and four simultaneously.
Study limitations and future research directions are given at the end of chapter five. Finally,
a conclusion is given in chapter six, along with management implications and final remarks.
Chapter 2. Methodology
20
Chapter 2
Methodology
2.1
Study area
Located in the Cariboo Mountains (mean elevation of 1,375 m above sea level) in the cen
tral interior of British Columbia (Fig. 2.1), the Quesnel River Basin is composed of three
watersheds: the Quesnel River Watershed, the Cariboo River Watershed, and the Horsefly
River Watershed. All three watersheds have a combined area of approximately 12,000 km2.
The land area within the basin has been used historically and at present for various resource
extraction activities including forestry, agriculture, ranching and mining.
The two selected study wetlands were chosen not only because their catchments have a
history of forestry practices, but also because they border another depositional environment
(i.e. a lake). Since the inflows of these lakes are surrounded by wetlands, they may have
been provided with a buffering function which would have influenced the amount of sediment
delivered to them. Since the project aims to establish the ability of the two study wetlands
to promote sedimentation under disturbance conditions, it was necessary to also evaluate
lake sedimentation rates over time.
2.1.1
Boswell Lake catchment
The Boswell Lake catchment (52°32'25"N, 121°27'6"W; see Figure 2.2) is situated in the
Quesnel River Watershed, and has an area of 2.1 km2. According to the Biogeoclimatic
Ecosytem Classification (BEC) System, the catchment is located in an interior cedar hemlock
zone characterized by a wet and cool climate (British Columbia Ministry of Forests and
Range, 2008). Mean annual temperature for this BEC zone ranges from 2 to 8.7°C, mean
Chapter 2. Methodology
21
Cariboo River
Quesnel River
Horsefly River
River
Lake
400 Kilometers
Watershed boundary
40 Klometers
Figure 2.1: Map of the province of British Columbia. The rectangle indicates the ap
proximate location of the Quesnel River Basin which is composed of three
watersheds: the Cariboo River Watershed; the Quesnel River Watershed;
and, the Horsefly River Watershed.
annual precipitation is 500-1200 mm, 25-50% of which falls as snow (Ketcheson et al., n.d.).
The local bedrock geology consists of basaltic volcanic rocks from the Upper TYiassic period
(Massey et al., 2004). The maximum and average slopes of the catchment are 38° and 12°,
respectively.
The wetland bordering Boswell Lake (herein referred to as Boswell wetland) is situated
at the inflow of the lake, and has a surface area of 0.020 km2 and a maximum width of
approximately 85 m (measured from the wetland-land border to the lake edge). Four channels
cross the wetland border flowing in a south to north direction from the deforested areas to
Boswell Lake, and meandering through the wetland is low. Two of the four channels were
identified as "major" channels, and the other two were labelled as "minor". based on their
size, degree of inundation, and connectivity to the logged slopes. The two major channels
were flowing during visits to the lake in the late spring, summer and fall, while flow in the
minor channels was only observed in the spring and not during any other visit to the study
site. Water depths in the major channels ranged from 10 to 65 cm with increasing depths
downstream. The dominant vegetation types in the wetland and the wetland channels are
Chapter 2. Methodology
22
sedges (Carex spp.) and Yellow Water Lilies (Nuphar variegata). Sedges form the bulk of
the vegetation in the wetland and were primarily observed closest to the lake edge where the
wetland channels became diffuse. Based on these characteristics the Boswell wetland has
been identified as a fen (MacKenzie & Moran, 2004).
Forestry activities occurred in the catchment during two separate time periods. During
the first period from 1960 to 1975, approximately 42% (0.873 km2) of the catchment was
clearcut. Prom 1982 until 2008 clearcut logging affected another 15% (0.324 km2) of the
catchment. Currently a 1.4 km active gravel road crosses through the catchment on the
north side of Boswell Lake. A 14.4 km network of deactivated dirt roads associated with
the first logging period (1960-1975) also exists on the south side of Boswell Lake. Field
observations confirmed that these dirt roads are no longer in use as there is substantial
vegetative growth along these roadways.
2.1.2
Viewland Lake catchment
The Viewland Lake catchment (52°25'44"N, 121°6'57"W; see Figure 2.3) is located in the
Horsefly River Watershed, and has an area of 2.5 km2. It is located in an interior cedar
hemlock zone having a wet and cool climate (British Columbia Ministry of Forests and
Range, 2008). Mean annual temperature ranges from 2 to 8.7°C, mean annual precipitation
is 500-1200 mm, 25-50% of which falls as snow (Ketcheson et al., n.d.). The local bedrock
geology is composed of two groups; sedimentary rocks from the mid-to-Upper Triassic period,
and basaltic volcanic rocks from the Upper Triassic period (Massey et al., 2004). Maximum
and average slopes of the catchment are 35° and 7°, respectively. The Viewland Lake
catchment area drains into three lakes all connected by a single channel running between
them from north to south. The top two lakes are each fed by a channel that originates from
the cutblock.
The wetland bordering Viewland Lake (herein referred to as Viewland wetland) runs
along the entire east side of the lake chain, and has a surface area of 0.074 km2 and a
Chapter 2. Methodology
23
width of approximately 30 m (measured from the wetland-land border to the lake edge).
During site visits in late summer and autumn it was observed that the channel through the
Viewland wetland was not flowing.
It is currently unknown whether or not this channel
experiences any degree of flooding during the year. Channel meandering was also found to
be relatively low. Sedges were observed to be the dominant vegetation type in the wetland
channel. Similar to Boswell wetland, sedges were densest near the lake edge. Yellow water
lilies were also present, however, they only occurred near the lake edge where flooding was
present. Based on these characteristics the Viewland wetland has been identified as a fen
(MacKenzie Si Moran, 2004).
Forestry practices in the Viewland Lake catchment occurred only in 1983 resulting in
deforestation of 58% of the catchment. A deactivated unpaved road is present just east of
Viewland Lake which crosses over the channel that flows into the lake. As the road has not
yet been decommissioned, a culvert from the inital construction of the road still remains in
place. Total road length in the Viewland Lake catchment is approximately 6.9 km.
Morphometric characteristics of both Boswell Lake and Viewland Lake can be found in
Table 2.1. Their bathymetric maps are presented in Appendix A.
Table 2.1: Morphometric characteristics of Boswell Lake and Viewland Lake. Information
for Boswell Lake was taken from The Angler's Atlas (2010).
Measurement
Catchment:Lake Area
Surface Area (km2)
Volume (m3)
Mean depth (m)
Maximum depth (m)
Perimeter (m)
Boswell Lake
Viewland Lake
0.06
0.128
148,000
1.2
2.5
1850
0.03
0.073
219,030
3.0
8.2
1468
Chapter 2. Methodology
121*2OTW
24
121#27tTW
121*26'0"W
Study sit©
Contour
Stream
•• Lake
'i%B. Wetland
{
I Catchment boundary
Logged
Figure 2.2: Map of the Boswell Lake catchment. Forestry practices were active in the
catchment during two time periods: A) 1960-1975; and B) 1982-2008. Inset:
Outline of the Quesnel River Watershed. The star represents the approximate
location of the Boswell Lake catchment in the watershed.
2.2
2.2.1
Sample collection and preparation
Wetland coring
Initially Boswell Lake was intended to be the primary study site with Viewland Lake acting
as a secondary study site in the event that the cores taken from the primary site did not
produce a useful core chronology. As a result, a more detailed sampling campaign was
undertaken at Boswell Lake and wetland, and only one of the three lakes in the Viewland
catchment was selected for coring. The middle lake was chosen because its stream drains a
larger area that was impacted by forestry practices (see Fig. 2.3), and it was thought that a
stronger logging signal may be observed in the middle lake.
Six sampling locations were identified for Boswell wetland and one for Viewland wetland
from which a single core was taken (Fig. 2.4(a)). Since wetlands typically do not possess a
single deepest point where sediment focusing will occur (unlike many lakes), it was necessary
Chapter 2. Methodology
121*8*0"W
l21a7VW
121WW
25
121'5D"W
121WW
62"26"0"N-
Studysite
Contour
Stream
h::!^ VSWand
I
I Catchment boundary
$$$ L
°09«'
Figure 2.3: Map of the Viewland Lake catchment. Forestry practices were active in the
catchment in 1983 (A). Inset: Outline of the Horsefly River Watershed. The
star represents the approximate location of the Viewland Lake catchment in
the watershed.
to identify areas in the wetland which, based on physical characteristics such as inundation
and channelization, likely experienced the greatest sedimentation rates. Four channels were
identified in Boswell wetland, and one in Viewland wetland, from which the sediment cores
were taken. As both wetlands were not completely inundated, sediment transport and de
position was assumed to have occurred primarily along these pathways. This assumption
is consistent with the observations of Craft & Casey (2000) who found that "open" (e.g.
riparian and floodplain)
wetlands had greater sediment accumulation rates than "closed"
(e.g. depressional) wetlands; where open and closed refer to the degree of connectivity to the
hillslope and surface water bodies.
At Boswell wetland, sampling areas were selected near the wetland inflow and outflow
to characterize the sedimentation rates along each of the major channels (i.e. four sampling
locations in total). Based on the characteristics of the two minor channels, described above,
these pathways were considered to be relatively less important for sediment delivery and these
Chapter 2. Methodology
26
two sampling locations were not analyzed in the current study. In the case of Viewland
wetland, only one sampling location was selected as only one channel exists between the
deforested area and the middle lake (Fig. 2.4(b)). Core lengths were dependent on the
characteristics of the sediment at each site and ranged between 0.25 and 1.0 m. Refer to
Figure 2.4 for a map of the sampling locations at both sites. All wetland coring took place
during July and August 2009.
As this study is primarily concerned with contemporary sedimentation rates it was crucial
to obtain profiles with intact upper sediment layers. It was decided that the open-barrel
coring method was therefore more appropriate than other methods (i.e. Russian Peat Corer)
as it minimally disturbs the top of the sediment profile (Glew, 2001). Using 2 m lengths of
PVC piping, 7.6 cm in diameter, two 3.2 cm holes were drilled approximately 2.5 cm from
the top of the PVC pipe. A metal rod 2.5 cm in diameter and 50 cm long was fit through
the top holes to provide a handle to assist in core removal (Fig. 2.5). This corer design was
adapted from Reinhardt et al. (2000).
2.2.2
Lake coring
For each of the two study sites, one core was retrieved from the deepest point of the lake
using a percussion corer (Reasoner, 1993). A core catcher constructed from stove pipe metal
was fixed in the bottom of the core tube to prevent captured sediments from being lost
during retrieval. An additional short core was taken using an Ekman dredge to ensure that
an undisturbed sample of the water-surface interface was taken. Coring at Boswell Lake
occurred in October 2009. Since Viewland Lake does not have direct vehicle access, cores
were retrieved in March 2010 when there was sufficient ice cover to provide a stable coring
platform.
All lake and wetland cores were transported back to UNBC where they were stored at
a temperature of approximately 4°C to prevent decomposition of organic matter and any
changes that may be associated with exposing anoxic soil to oxidizing conditions. Cores
Chapter 2. Methodology
(a) Boswell Lake and wetland
•
Processed core
(b) Viewland Lake and wetland
Figure 2.4: (a) Boswell Lake and (b) Viewland Lake and wetland coring locations. Codes
containing a 'D' indicate an open barrel core, 'P' refers to a percussion core,
and those with an 'E' denote a core taken with an Ekman dredge. Note:
The location of the stream containing Boswell wetland core BL-D13 was not
shown in original spatial dataset. This line feature was created by extracting
point locations from a Google Earth image of the catchment.
27
Chapter 2. Methodology
50 cm
|
- -1—
28
7.6cm
1 |
=3
,
O
i—i
3.2 cm
Figure 2.5: Diagram of the open barrel corer used to retrieve wetland sediment cores.
that were selected for further analysis were cut length-wise, photographed and logged prior
to slicing the sediment at 1 cm intervals.
2.2.3
Source materials
To assess if forestry practices resulted in altered sediment source contributions and changes in
the dominant sediment sources, sediment samples were collected from six source types: har
vested surface soil material; harvested subsurface soil material; forested surface soil material;
forested subsurface soil material; road surface soil material; and channel bank material. Ap
proximately 5-8 samples were taken for each source type and a GPS coordinate was recorded
for each sampling location. Samples were taken with a stainless steel trowel which was rinsed
with distilled water and acetone between each sample to minimize cross-contamination. Each
sample was itself a composite of 3-5 subsamples collected within an area of approximately
5 m by 5 m to account for any local spatial variability.
All source materials were air dried prior to laboratory analysis. Samples that contained
moisture after air drying were placed in an oven at 60 °C until dry. Source materials were
then disaggregated and sieved to <63 /im. Analysis of this particle size fraction was intended
to minimize the differences in the particle size composition between core sediment and source
Chapter 2. Methodology
29
material (Carter et al., 2003; Motha, 2003) as most of the lake and wetland sediment was
<63 fim.
2.3
2.3.1
Radionuclides and core chronology
Origin of lead-210
Reconstructing sediment chronologies over the last 100-150 years requires the use of a ra
dionuclide which has a relatively fast decay rate. Lead-210 (210Pb) has a half-life of 22.26
years and is ubquitous in the environment as a result of the natural decay of 238U in bedrock
(Binford, 1990). Following the decay of 238U to 226Ra, 226Ra then decays to 222Rn. 222Rn
forms a gas which escapes to the atmosphere, and through several additional decays, becomes
210Pb.
In order for 210Pb to fall out of the atmosphere it needs to adsorb onto atmospheric
particulates and/or water droplets which are typically delivered to land and water surfaces
via precipitation. In the water column 210Pb binds to fine particles and organic material and
is deposited on the bottom of the water body (e.g. ocean, lake, river, wetland). This fraction
of 210Pb is referred to as unsupported 210Pb.
222Rn is also
produced in the soil which decays through the same decay series result
ing in the in situ production of 210Pb. This is known as supported 210Pb (Binford, 1990;
Noller, 2000). Unsupported 210Pb is calculated as the difference between the total 210Pb and
estimates of the supported component.
2.3.2
Lead-210 dating models
Several models exist that utilize the unsupported fraction of 210Pb to assign chronologies
to sediment core profiles. Most commonly used are the Constant Initial Concentration
(CIC) and Constant Rate of Supply (CRS) models. The CIC model assumes that the
initial concentration of unsupported 210Pb remains constant over the time that unsupported
210Pb
is measurable (Turner & Delorme, 1996). As a result, the log transformation of
Chapter 2. Methodology
30
unsupported 210Pb activities should yield a linear decrease over depth. When the CIC model
is applied to a non-monotonic decay curve, resultant core chronologies include one or more
time inversions. These occur because other processes acting on the sediment profile have
impacted unsupported 210Pb activities leading to an imperfect decay curve. For example,
Appleby et al. (1988) found that due to organic matter degradation and loss over time, the
CIC model was not suitable for peat cores. Sediment dates (t) can be determined with the
CIC model using:
<">
where A is the radioactive decay constant for 210Pb, Cx (Bq kg-1) is the activity of unsup
ported 210Pb at depth x\ and C0 (Bq kg-1) is the activity at the surface.
The CRS model assumes that the absolute flux rate of 210Pb remains constant, regardless
of background sedimentation, such that higher rates of background sedimentation will lead
to lower 210Pb concentrations (Appleby &: Oldfield, 1978). Unlike the CIC model, it is able
to account for fluctuations in unsupported 210Pb sedimentation which may have occurred
either in response to climatic variations or anthropogenic disturbance (Brenner et al., 1999;
Cohen et al., 2005). Futhermore, inversions in the unsupported 210Pb profile may be better
explained by a dilution effect of unsupported 210Pb due to an increase in sedimentation rates.
However, one limitation of the CRS model is that it tends to over-estimate sediment ages
near the bottom of the profile. Sediment ages (t) based on the CRS model can be calculated
by:
(2 - 2)
where Ax is the inventory of unsupported 210Pb (Bq m-2) to depth x; and A0 is the total
inventory of unsupported 210Pb (Bq m~2).
Other more computationally intensive models have been proposed, such as the Sediment
Chapter 2. Methodology
31
Isotope Tomography (SIT) model. Unlike the other dating models, the SIT model allows
both the absolute flux rate of 210Pb and the sedimentation rate to vary (Carroll et al., 1995;
von Gunten et al., 2008). Another difference between the SIT model and other conventional
210Pb dating
models is that it reconstructs the unsupported 210Pb activity profile before
calculating a core chronology. This is accomplished by modelling nonexponential changes
in unsupported 210Pb with a Fourier sine series, while any additonal changes caused by
other processes are modelled with a Fourier cosine series. A 210Pb profile is selected when
a pre-determined measure of fit (x2) is achieved which compares the modelled profile to the
original unsupported 210Pb profile (Carroll & Abraham, 1996). See Table 2.2 for a summary
of all model assumptions.
Table 2.2: Summary of model assumptions. Adapted from Carroll & Lerche (2003).
Model name
Constant Initial Concentration
Constant Rate of Supply
Sediment Isotope Tomography
2.3.3
Specific activity
Accumulation rate
Flux of 210Pb
constant
variable
variable
variable
variable
variable
variable
constant
variable
Caesium-137
Ideally, paleoenvironmental studies should not rely on a single dating model, and would
employ the use of a marker horizon to confirm the constructed chronology. The most com
monly used marker horizon is caesium-137 (137Cs). 137Cs is an artificial fallout product of
atmospheric bomb testing that began in the early 1950s and ended in the early 1970s. Peak
fallout of 137Cs as a result of atmospheric bomb testing occurred in 1963 (Owens et al.,
1996), and is often represented in sediment profiles as a peak in down-core measurements.
The location of the 1963 137Cs peak, as well as the onset of increasing 137Cs concentrations
(1954), can be used paleolimnological studies to verify the accuracy of core chronologies (von
Gunten et al., 2008). Good agreement between these peaks and the location of the modelled
dates in the profile (i.e. using unsupported 210Pb) should indicate that the dating model
Chapter 2. Methodology
32
is appropriate for that environment. A secondary 137Cs peak produced by the explosion of
the Chernobyl nuclear reactor in 1986 has also been used for the same purpose, but is not
detected within western Canada. Post-depositional processes, such as mixing, can impact
the 137Cs profile (He & Walling, 1996; Foster et al., 2006). Other studies have used Ambrosia
pollen (Blais et al., 1995), tephra (Reasoner k. Healy, 1986), stable lead (Blais et al., 1998),
and other metals and contaminants (Cooke & Abbott, 2008) to mark known historical events
and verify core chronologies.
2.3.4
Core chronology
Since several cores were taken from Boswell wetland and not all could be analyzed due to time
and financial constraints, it was necessary to select representative for laboratory analysis.
Cores BL-D8 and BL-D10 from the far west channel were selected as the key wetland cores
(Fig 2.4(a)). Several attempts were required to retrieve cores from the other major stream
which likely disturbed and redistributed the top sediments contaminating other coring sites.
Cores BL-D8 and BL-D10 were successfully removed on the first attempt minimizing the
disturbance and redistribution of top sediments.
Lake cores (BL-Pl and VL-P1) and selected wetland cores (BL-D8, BL-D10 and VL-Dl)
were analyzed for 210Pb and 137Cs.
210Pb,
with a half-life of approximately 22.26 years,
was decided to be the most appropriate radionuclide for constructing a core chronology for
the last 100-150 years1. Core chronologies were calculated using the CRS, CIC, and the
SIT models. Details of all models are found in Section 2.3.2. Software for the SIT model
was provided by Dr. J. Carroll of the Polar Environmental Centre, Norway. The final core
chronology was selected based on the model whose assumptions were satisfied, and produced
the smallest date errors (i.e. error bars). As it is assumed that similar processes are acting
on these systems, a single dating model was selected for all cores. The results of the 210Pb
dating models were also compared to 137Cs activities to verify model accuracy.
1Most equipment for measuring radionuclide activities is only able to detect
half-lives
radionuclides up to 4-5
Chapter 2. Methodology
33
Approximately 1-3 g of sediment from each 1 cm core section was packed into a 4 mL
plastic vial and left for three weeks to allow equilibrium to be reached between 214Pb and
its parent radioisotope 226Ra (Kohler et al., 2000). Measurements of total 210Pb, supported
210Pb and 137Cs activities were undertaken at
the Plymouth University Consolidated Ra
dioisotope Facility, England, UK using a EG&G Ortec well (GWL-170-16-S N-type) HPGe
Gamma spectrometry system over a period of 24 to 48 hours for each sample. Longer mea
surement times were necessary to minimize the higher error associated with lower sample
masses2.
2.4
Proxy measurements
Proxy measurements were used to compile information on the physical and chemical char
acteristics of the sediment trapped by the study wetlands and lakes over the last century.
The data provided multiple lines of evidence for understanding the nature of the material
captured by both wetland and lake environments and, therefore, the type of material being
mobilized from the hillslopes. Dry bulk density and percent water content were also used in
conjunction with magnetic susceptibility to match age-equivalent sediment layers in overlap
ping sections of cores from each of the two lakes (i.e. Ekman and percussion cores) and create
contiguous lake sediment profiles (Snowball & Sandgren, 2001). This process is similar to
that of core correlation which aims to match cores taken from various coring locations so
that chronologies may be extended to undated cores (Foster et al., 1985). The Boswell Lake
core is therefore a combination of Ekman (BL-E1) and percussion (BL-P1) cores, however,
it will be referred to as BL-Pl. Similarly, the Viewland Lake core is a combination of cores
VL-E1 and VL-P1, but will be referred to as VL-P1.
2The ideal mass for gamma spectrometry is 5 g.
However, this mass could not be reached with the
material retrieved from any of the sediment cores as they were highly organic and had low clastic contents.
Chapter 2. Methodology
2.4.1
34
Bulk physical properties
Changes in dry bulk density and percent water content were used to provide information
on the type of material being delivered to the lake and wetland. Increases in dry bulk
density may be indicative of more minerogenic material which has been previously linked
to the mobilization of subsurface soil material (Thompson et al., 1975). Each 1 cm section
of sediment was placed in a pre-weighed plastic WhirlPak bag and re-weighed. Bags of
sediment were frozen at —10 °C and subsequently placed in a freeze drier for approximately
72 hours to remove all moisture. The bags were then re-weighed to determine the mass
of dry sediment. Dry bulk density and percent water content were calculated according to
Equations 2.3 and 2.4, respectively.
2.4.2
.
Dry mass (g)
Dry
bulk densityJ = ——
-—5J
Volume (cm3)
.
(2.3)
v
'
^
Wet mass (g) - Dry mass (g)
_
Percent water content =
—
—
• 100
Wet mass (g)
(2.4)
Magnetic susceptibility
Magnetic susceptibility is a measure of the concentration of magnetic minerals in the sedi
ment, or the clastic content of the sediment. A large positive magnetic susceptibility value
indicates that the materials in the sediment maintain a magnetic charge after the sediment
has been exposed to a magnetic field. Conversely, low or negative magnetic susceptibility
values indicate that the materials in the sediment do not maintain a magnetic charge after
exposure to a magnetic field. For example, iron-bearing minerals have a high magnetic sus
ceptibility values while wood and other plant materials have low values (Nowaczyk, 2001).
Trends in down-core magnetic susceptibility have previously been linked to the timing of de
forestation and erosion of minerogenic soils (Thompson et al., 1975), and were used for the
same purpose in the present study. Magnetic susceptibility was measured in triplicate at each
Chapter 2. Methodology
35
1 cm interval for all wetland and lake cores using a Bartington MS2 Magnetic Susceptibility
System at the University of Northern British Columbia. All magnetic susceptibility mea
surements were normalized by sediment mass to give mass-specific magnetic susceptibility
(Sandgren k. Snowball, 2001).
2.4.3
Particle size
Particle size analysis was completed for all lake and key wetland cores as well as all source
materials of the <63 fxm particle size fraction. Variations in element concentrations may
be related to grain size and must therefore be taken into account in the mixing model
(described further below). Down core changes in particle size distribution have also been
linked to historical changes in land cover and human activities (van Hengstum et al., 2007).
Analysis of pre-logging conditions will provide background particle size distributions and
their natural variations against which periods of forestry practices and post-logging can be
compared. Particle size analysis could not be completed for several slices (6, 7, 11-14 cm) of
the Viewland wetland core (VL-D1) as not enough inorganic material was present in these
1 cm core slices to reach the recommended degree of obscuration3 (Sperazza et al., 2004).
Sediment samples were pre-treated with 30% hydrogen peroxide and heated to approx
imately 70 °C to digest organic material. Approximately 10 mL of a 0.55% sodium hexametaphosphate solution, (NaP03)6, was added to each sample to promote dispersal of the
individual sediment grains and prevent flocculation
(Sperazza et al., 2004). Samples were
stirred for approximately 30 seconds prior to particle size analysis to resuspend particles into
the water column. Particle size distributions were determined using a Mastersizer 2000 laser
diffractometry analyzer in the Department of Earth Sciences laboratory at Simon Fraser
University, Burnaby, BC.
3Obscuration is a measure of the quantity of sediment added to the analyzer.
obscuration has been recommended to minimize variability of results.
Between 15 and 20%
Chapter 2. Methodology
2.4.4
36
Total carbon and nitrogen
As a result of low sample masses for each core section, measurements of total carbon (C)
and total nitrogen (N) were used in lieu of organic matter content. Dry sediment samples
(ca. 0.05 g) from each 1 cm core section were sent to the Forestry and Technical Services
laboratory (Ministry of Forests and Range) in Victoria, BC for analysis of total C and total
N content. A C:N ratio was then calculated from the total C and N percentages for each 1 cm
core section to provide additional information on the source of organic matter. Typically
C:N values between 4 and 10 represent organic matter derived from phytoplankton. Values
greater than or equal to 10 are more indicative of vascular terrestrial vegetation (Meyers &
Teranes, 2001; Kim, 2003).
2.4.5
Geochemistry
A suite of 34 geochemical properties4 were measured for all lake and wetland cores (except
Boswell wetland core BL-D10), as well as all source materials. BL-D10 was not analyzed
for geochemistry due to a miscommunication regarding sample priorities in other analyses
and unavoidable time constraints. The geochemical properties then became the fingerprint
properties used in the sediment source tracing procedure. Dry sediment samples were pre
pared for geochemical analysis by adding concentrated acid (5 mL HN03 and 1 mL HC1)
and further digesting the samples in a microwave digester. Digested samples were diluted
with Milli-Q water such that the total volume equalled 50 mL. Samples were analyzed by
Inductively Coupled Plasma Atomic Emission Spectrometry (ICP-AES) using a Leeman
PS1000-UV to determine element concentrations. Sample preparation and ICP-AES anal
yses were completed in the Central Equipment Laboratory at the University of Northern
British Columbia.
4The 34 geochemical
properties measured were: lithium; beryllium; sodium; magnesium; aluminum;
silicon; phosphorus; potassium; calcium; titanium; vanadium; chromium; manganese; iron; cobalt; nickel;
copper; zinc; arsenic; selenium; strontium; zirconium; molybdenum; silver; cadmium; tin; antimony; barium;
tungsten; mercury; thallium; lead; bismuth; and, uranium.
Chapter 2. Methodology
2.5
37
Climate and stream discharge
Since sediment transport typically occurs during large precipitation events or spring snowmelt,
historical climate data were compiled to determine whether any fluctuations
in lake or
wetland sedimentation rates could be explained by variations in weather patterns. A cli
mate modelling program, ClimateBC, has been developed by British Columbia's Ministry of
Forests and Range to estimate historical and future climate variables. This program requires
a user input of latitude, longitude and elevation for the area of interest (Spittlehouse, 2006b).
It calculates climate variables for that location by interpolating between existing weather
stations. Although this program was created to produce climate data for areas where weather
stations are limited, this in turn has become a limitation for the model itself. ClimateBC
has been found to provide poor climate predictions for areas not well-covered by weather
stations (Spittlehouse, 2006b). Despite this limitation, data calculated by ClimateBC span
the full temporal range of the dated cores, and was thus chosen over the incomplete datasets
from Environment Canada5.
Latitude, longitude and elevation data were extracted from a digital elevation model for
every 30 m grid in each catchment. These data were subsequently run in ClimateBC to
produce climate data from 1901 to 2002. Each climate variable was then averaged over
each catchment for every year that climate data were calculated. A full list of the variables
produced by ClimateBC is provided in Appendix B. ClimateBC was downloaded from the
University of British Columbia's Centre for Forest Gene Conservation (University of British
Columbia, 2010). All mapping and spatial analyses were completed using ArcGIS 9.3.
Although stream discharge data were not collected as a part of this study, data were
available for the Quesnel River at Likely, BC (52°36'56" N, 121°34'16" W). These data do
Environment Canada's National Climate Archive was found to have an incomplete dataset for the Likely,
BC weather station (1974-1993) which is approximately 9 km from Boswell Lake. The next closest weather
stations to Boswell Lake are located in Barkerville (1888-2008) and Williams Lake (1936-2009) which are
both located approximately 60 km from Boswell Lake. A weather station situated at Gruhs Lake (19502009), approximately 16 km from Viewland Lake, was the closest weather station that could be used to
characterize weather conditions around Viewland Lake.
Chapter 2. Methodology
38
not reflect the same magnitude of water flow in the channels draining the Boswell Lake
catchment, however, they provide an estimate of the trends that may have been observed
over a similar time frame (1924-2009). As a nearby stream discharge gauge was not found
for the Viewland Lake catchment, stream discharge are only presented for the Quesnel River
(Boswell Lake catchment). Stream discharge data were downloaded from the Water Survey
of Canada website (Environment Canada, 2010).
2.6
2.6.1
Sediment source tracing
Source groups
Source materials were initially classified according to six source types: harvested surface soil
material; harvested subsurface soil material; forested surface soil material; forested subsur
face soil material; road surface soil material; and, channel bank material. However, since the
source materials were all derived from a single bedrock type, and forestry practices do not
typically alter the geochemical regime of the disturbed area, it was recognized that the source
materials could have similar characteristics and may not be significantly different from one
another. To determine if a different set of groups was more appropriate than the a priori
groupings, a Principle Component Analysis (PCA) was performed on the geochemical data.
Visual examination of the biplot of the first two principle components provided an estimate
of the number of source groups. A fuzzy k-means clustering analysis of the geochemical data
was subsequently used to confirm the number of source groups which informed the multi
variate unmixing model. The aim of this statistical test is to establish the number of groups
that minimizes Dunn's coefficient which measures the "fuzziness" of the resulting group(s)
(Trauwaert, 1988).
Chapter 2. Methodology
2.6.2
39
Multivariate unmixing model
The multivariate unmixing model described by Collins et al. (1997) was used to estimate the
relative proportion of each source material in each 1-cm sediment core slice of all lake and
wetland cores. The unmixing model is composed of a set of linear equations, which is subject
to the conditions: (a) each source contribution must be greater than zero and less than one;
and (b) the sum of the source contributions must equal one. A final solution is found when
the sums of squares of the relative errors have been minimized, and all conditions have been
met. The optimization routine was carried out using Microsoft Excel Solver (version 2003).
Unlike the original model by Collins et al. (1997), Equation 2.5 does not include an
organic matter correction factor. The inclusion of an organic matter correction factor, along
with a particle size correction factor, may result in the over-correction of the fingerprint
properties (Carter et al., 2003). The unmixing model is given in Equation 2.5:
2
(2.5)
where Cj = concentrations of tracer parameters (i) in each 1 cm sediment core slice, Si
= mean concentration of tracer parameter (i) for each source material, Z = particle size
correction factor (ratio of core slice specific surface area to mean specific surface area for
each source type), and Ps = percentage contribution from each source type (s = surface; sub
= subsurface; cb = channel bank).
2.7
2.7.1
Statistical analysis
Pre- versus post-logging
Two-sample t-tests were used to compare post-logging total (clastic and organic sediment)
sedimentation rates to pre-logging rates. The two periods of forestry practices in the Boswell
Lake catchment were combined and analyzed as a single "post-logging" period to improve the
Chapter 2. Methodology
40
statistical power since sample sizes were low in individual post-logging periods. Therefore,
the post-logging period in the Boswell Lake catchment begins at the onset of logging and
includes periods of active logging and the recovery periods (i.e. 1960-2009). A series of
one-sample t-tests were also used to compare total sedimentation rates calculated for each
post-logging 1-cm core slice to pre-logging rates. These comparisons were useful in identifying
peaks or depressions in the post-logging period that represented significant departures from
average pre-logging conditions. A Holm correction (sequential Bonferroni correction) was
applied to the p-values of the 1-sample t-tests for each proxy within a given core to account
for an inflated type I error that may occur as a result of multiple tests on a single group of
data (Holm, 1979; Rice, 1989). See Figure 2.6 for a diagram comparing the 1- and 2-sample
t-tests used on the lake and wetland core measurements.
Y,
Y2
Y3
Y4
Y5
YE
M
X
1-sample
2-sample
Figure 2.6: Comparison of the 1- and 2-sample t-tests used to evaluate post-logging ( Y )
changes in the sediment profiles against average pre-logging (X) conditions.
For 1-sample t-tests, sedimentation rate and proxy values given by individual
post-logging 1 cm core slices (Fj) represent the null hypothesis ( f x ) .
Where the assumption of normally distributed residuals was not met, non-parametric
statistics were applied (e.g. Mann-Whitney U-test, Wilcoxon Rank Sum Test). The same
approach was also taken for the analysis of each proxy indicator. It has also been recog
nized that temporal autocorrelation is often an issue with sediment cores which violates the
Chapter 2. Methodology
41
assumption of independent samples. Temporal autocorrelation can be the result of lags be
tween sediment erosion and delivery, physical mixing of the sediments, and post-depositional
mobility of proxy indicators (e.g. radionuclides). Corrections for temporal autocorrelation
often requires increasing the temporal separation between samples which can be achieved by
combining samples. Not correcting for temporal autocorrelation can result in increased type
I error (a); however, due to low sample sizes a correction was not applied.
2.7.2
Climate and stream discharge trends
To account for any potential effects of climate and stream discharge on lake and wetland
total sedimentation rates stepwise linear regressions were fit. Variable selection was carried
out in both directions (i.e. forward selection and backward elimination) and a final model
was selected when the Akaike's Information Criterion (AIC) value was minimized. Predictor
variables included climate variables produced by ClimateBC (mean annual precipitation,
mean annual temperature, precipitation as snow, beginning and end of frost-free period),
and stream discharge. It was necessary to narrow down the climate variables to a smaller
set such that the total number of variables entered into the model did not exceed the total
degrees of freedom. A factor was also included in the model to identify pre- and post-logging
periods. Climate and stream discharge data were averaged for each core to match the time
intervals represented by each 1 cm core slice.
Although averaging the climate data was necessary to be included in the stepwise linear
regression models, this also reduced the resolution of the climate series and flattened
out
annual variations. Trends in annual climate and discharge data were assessed using the
method outlined by Tome (2004), which identifies breakpoints and linear trends in time
series data. The sign and magnitude of the trends in between sets of breakpoints were
examined with the regression lines produced by the model. A single regression line was also
calculated for each full time series to observe long-term trends.
Chapter 2. Methodology
2.7.3
42
Correlations
Pearson product moment correlations were used to relate changes in the contribution of each
sediment source to down-core variations of total sedimentation rates, as well as those of each
proxy indicator. A change in the dominant sediment source, or the relative proportions of
these sources, may be the result of forestry practices; though, it is also possible that changes
in source materials are not related to fluctuating total sedimentation rates in either the lake
or the wetland.
All statistical results were considered significant at an a level of 0.05. All statistical
analyses were conducted using R 2.10.1 (2009).
Chapter 3. Results: Lake and wetland sedimentation rates
43
Chapter 3
Results: Lake and wetland sedimentation rates
This chapter addresses the research questions: do wetland sedimentation rates increase in
response to forestry activities?; and, are sedimentation rates in downstream lakes affected by
forestry activities? The information presented here adresses the variations in sedimentation
rates, as well as the physical and chemical properties of the sediment, before and during/after
periods of active logging. Historical changes in climate and their impact on sedimentation
rates and sediment properties have also been explored in addition to forestry practices.
Finally, variations in bulk physical properties over the last century were compared to changes
over the deeper undated profile to understand their importance in the context of the longerterm natural variability.
3.1
Physical descriptions
Material in the Boswell Lake core was predominantly light brown, fine-grained organic sedi
ment (gyttja) which was found throughout the top 80 cm of the core (see Figure 3.6). Below
80 cm alternating bands of dark and light brown fine organic sediment were visible. A light
grey tephra layer was found at 56 cm depth. Visual inspection of the tephra under a polar
izing light microscope revealed chunky glass shards containing lineated gas vesicles. Based
on the observed colour and glass shard morphology (Brian Menounos pers. comm.), it was
concluded that the tephra originated from the Bridge River eruption ca. 2,410 calendar years
ago (Clague et al., 1995). See Appendix C for a microscope image of the glass shards found
in the Boswell Lake core. Other than the tephra there were no other obvious changes in
texture along the length of the core.
Chapter 3. Results: Lake and wetland sedimentation rates
The Viewland Lake core was primarily composed of light brown, fine-grained
44
organic
sediment (gyttja) similar to that found in the Boswell Lake core. A 0.5 cm layer of light
grey, fine-grained clastic sediment was found 1.5 cm down-core below which was a thin layer
of dark brown, fine-grained
organic sediment. Alternating light and dark brown layers of
sediment were also seen deeper (ca. 70 cm) in the core. However, the layers of sediment were
not as well-defined as those found in the Boswell Lake core. A light grey tephra layer was
found at 67 cm down-core that, based on colour and shard morphology, was correlative with
the Bridge River event (Brian Menounos pers. comm.).
Wetland cores from both sites consisted of dark brown, unsorted, organic-rich sediments.
Large pieces of woody debris, roots and twigs were observed throughout all wetland cores
in no observable pattern. The top 10 cm of the Viewland wetland core was predominantly
composed of twigs and other woody debris.
3.2
Lead-210 profiles and core chronologies
In Boswell Lake, background concentrations of unsupported 210Pb were reached at 9 cm in
the lake core (BL-P1), 13 cm in one wetland core (BL-D8), and 12 cm in the other wetland
core (BL-D10). Background activities were reached by 19 cm and 16 cm in the Viewland Lake
(VL-P1) and the wetland (VL-D1) cores, respectively. Shallower 210Pb profiles in the Bowell
Lake core versus either of the wetland cores suggests that less material is accumulating in
the lake as compared to the wetland. Alternatively, regular resuspension and transport of
material from the lake bottom could lower 210Pb concentrations. The opposite is observed in
the Viewland Lake scenario where, according to the 210Pb profile, the lake is accumulating a
greater amount of material than the wetland. Since wetland cores were taken from channels,
erosion of the channel bottom due to flowing water may have redistributed sediments creating
unconformities in the depositional record. The implications of this sampling design are
discussed in the study limitations (see Section 5.5).
Before each 210Pb dating model was applied to the unsupported 210Pb activities of each
Chapter 3. Results: Lake and wetland sedimentation rates
45
sediment core, the model assumptions were reviewed against the activity profiles of un
supported 210Pb in Figures 3.1 and 3.3. Non-monotonic decreases in unsupported 210Pb
activities in all cores confirm that the CIC model is not a suitable dating model. Modelled
unsupported 210Pb activities produced by the SIT model were accompanied by relatively
large x2 values (Table 3.1) which suggests that the model does a poor job reconstructing the
original unsupported 210Pb profiles. This is consistent with the limitations of the SIT model
as outlined by Carroll & Abraham (1996) which state that large fluctuations in unsupported
210Pb activities may not be accurately modelled by a Fourier sine series.
Thus, the final core
chronologies were calculated using the CRS model - see Figures 3.2 and 3.4.
With the exception of the VL-Pl core, well-defined 137Cs peaks are not present in any of
the cores which suggests that post-depositional changes (i.e. bioturbation, upward/downward
diffusion) have impacted down-core concentrations of 137Cs (Foster et al., 2006). As a result,
137Cs was not used to verify core chronologies in this study. Despite a strong peak, 137Cs data
were also disregarded for the VL-Pl core. Based on the resultant core chronologies there is
a large discrepancy between the 137Cs peak and the CRS-modelled 1963 date for this core.
The 137Cs peak also coincides with a layer of fine-grained silty-clay material which, through
the binding effects of clay (Ambers, 2001), likely limited the mobility of 137Cs resulting in
increased concentrations. Davis et al. (1984) stated that high mobility of 137Cs in organicrich sediments is the result of the breakdown of organic material. On the other hand, 210Pb
is bound tightly to organic material (Dorr & Miinnich, 2006) suggesting that the use of
210Pb
to date organic-rich sediments is more reliable than 137Cs (Davis et al., 1984). A
similar conclusion was reached by Foster &; Lees (1999) who also found 137Cs profiles to be
unreliable.
Chapter 3. Results: Lake and wetland sedimentation rates
Unsipported 2'5Pb activity (Bq kg"')
Unsupported ^^Pb activity (Bq kg-')
500
1 1/° 1
h^"' \-¥i
-
K'
i
f
o
y
'
rH M
V -
I «f
3
CO -
(M -
'
H"'
-
750
4
eo -
© _
© _
o- Unsupported 21®Pb
-°~ ,37Cs
rw
Unsupported z1Tb
c*
r~
100
100
200
137Cs activity (Bq kg"1)
200
1S7Cs activity (Bq kg"1)
(a) Lake core (BL-Pl)
(b) Wetland core (BL-D8)
Unsupported 2,0Pb activity (Bq kg"1)
250
I
f
S
Unsupported
100
Pb
200
mCs activity (Bq kg"1)
(c) Wetland core (BL-D10)
Figure 3.1: Unsupported 210Pb and 137Cs activity depth profiles for Boswell Lake and
wetland cores. Unsupported 210Pb error bars represent the sum of the total
210Pb and supported 210Pb errors. Values without errors were measured at
the minimum detectable limit of the gamma assay.
46
Chapter 3. Results: Lake and wetland sedimentation rates
(a) Lake core (BL-P1)
(b) Wetland core (BL-D8)
©
-«r
<0
CO
o
2000
1900
1950
1850
Year
(c) Wetland core (BL-D10)
Figure 3.2: Core chronologies for (a) Boswell Lake (BL-P1) and wetland cores (b) BLD8, and (c) BL-D10 produced by the Constant Rate of Supply (CRS) model.
Error bars were also calculated using the CRS model and represent the error
on each of the calculated dates.
47
Chapter 3. Results: Lake and wetland sedimentation rates
Unsipported 210PbadMty(Bqkg'1)
0
250
500
750
Unsupported 210Pb activity (Bq kg-1)
1000
250
500
° J——H3 1
£
IQ
c
-®100
Unsupported J15Pb
mCs
Unsupported 4 "Pb
200
100
137Cs activity (8q kg"')
(a) Lake core (VL-P1)
200
,3TCs acfivrty (Bq kg"1)
(b) Wetland core (VL-D1)
Figure 3.3: Unsupported 210Pb and 137Cs activity depth profiles for Viewland Lake and
wetland cores. Unsupported 210Pb error bars represent the sum of the total
210Pb and supported 210Pb errors. Values without errors were measured at
the minimum detectable limit of the gamma assay.
Table 3.1: Summary of the x2 values produced by the Sediment Isotope Tomography
(SIT) model. These values represent the goodness-of-fit between an observed
distribution (measured unsupported 210 activities) and a theoretical distribu
tion (modelled unsupported 210Pb activities). For a sample size of 10, two sam
ple distributions would be considered to be not significantly different (p>0.05)
if the x2 value was <16.9.
Core
BL-P1
BL-D8
BL-D10
VL-P1
VL-D1
X 2 (Bq kg"1)
924
80.3
649
11800
181
48
Chapter 3. Results: Lake and wetland sedimentation rates
49
£
o%•
1900
(a) Lake core (VL-Pl)
2000
1950
1900
1850
(b) Wetland core (VL-Dl)
Figure 3.4: Core chronologies for (a) Viewland Lake (VL-Pl) and (b) wetland (VL-Dl)
cores produced by the CRS model. Error bars were also calculated using the
CRS model and represent the error on each of the calculated dates.
Chapter 3. Results: Lake and wetland sedimentation rates
3.3
3.3.1
50
Boswell Lake catchment
Total sedimentation rates
Comparison of pre- and post-logging sedimentation rates (Fig. 3.5) using two-sample t-tests
showed that the two periods are not significantly different for any of the cores taken from
Boswell Lake or the wetland (Table 3.2). Individual one-sample t-tests for each of the 1 cm
increments in the post-logging period revealed no significant departures from average prelogging sedimentation rates in Boswell Lake.
Above 10 cm (ca. 1938), a gradual increase in total sedimentation rates was observed
in wetland core (BL-D8) and reached a maximum of 0.0756 g cm-2 y_1 in the top 1 cm.
The second wetland core (BL-D10) had two distinct peaks in sedimentation rates at 3 and
6 cm down-core. Both peaks corresponded to the two logging periods in the Boswell Lake
catchment, however, neither was found to be statistically greater than average pre-logging
sedimentation rates.
Table 3.2: Summary of the two-sample t-tests results comparing pre- and post-logging
total sedimentation rates (g cm-2 y-1) in Boswell Lake and wetland cores. In
BL-P1, BL-D8, and BL-D10 the post-logging periods are above 4 cm, 7 cm,
and 7 cm, respectively. Values in brackets denote sample size.
BL-D8 (Wetland)
BL-P1 (Lake)
Period
Mean
sd
Pre
Post
0.0057
0.0079
0.0035 (3)
0.0013 (4)
p
Mean
sd
0.386
0.0342
0.0492
0.023 (4)
0.018 (7)
BL-D10 (Wetland)
p
Mean
sd
p
0.308
0.0169
0.0466
0.0109 (3)
0.0228 (7)
0.067t
tp was calculated using non-parametric analysis as the assumption of normality was not
met.
3.3.2
Proxy indicators
Depth profiles of all proxies are presented in Figure 3.6. Two-sample t-test results are
summarized in Table 3.3. Median grain size in Boswell Lake (BL-P1) significantly decreased
post-logging, and was the only proxy indicator for which a significant change occurred. One-
Chapter 3. Results: Lake and wetland sedimentation rates
1 * t —-L*J
Lake(BL-PI)
Wetland (BL-D8)
Wetland (BL-D10)
002
0 04
0 06
0 08
Sedimentation rate {g cm"2 y"1)
Figure 3.5: Total sedimentation rates (calculated using the CRS model) for Boswell Lake
and wetland cores. The highlighted areas represent the periods of time that
forestry practices were present in the Boswell Lake catchment. Error bars
on the sedimentation rates represent the standard error calculated using the
CRS model. An error value could not be calculated for the bottom of the
BL-D8 profile.
51
Chapter 3. Results: Lake and wetland sedimentation rates
52
sample t-tests revealed that significant decreases in median grain size occurred at 1 (4=11.0,
p=0.019), 2 (4=1.14, p=0.020) and 4 cm (4=9.70, p=0.019) down-core, all of which were
within the post-logging periods. A small post-logging increase was observed in dry bulk
density, which was mirrored by a small decrease in percent water content. The start of these
changes, however, occurred approximately 20 years (ca. 1940) before the beginning of the
first logging period. Maximum values of dry bulk density were reached at 3 cm and were
followed by a decrease at 2 cm down-core, with the opposite pattern being observed for water
content. At 6 cm depth, minor increases in magnetic susceptibility and C:N were observed
along with small decreases in total carbon and total nitrogen.
A significant post-logging decrease in median grain size was observed in wetland core BLD8. While generally smaller median grain sizes were observed post-logging, a small increase
occurred during the first logging period. However, the second logging period did not produce
any distinct changes in median grain size. Decreases in total C, total N and C:N occurred at
the end of the first logging period. These three proxies continued to fluctuate throughout the
second period of logging, but only C:N remained significantly lower than pre-logging values.
On average, magnetic susceptibility values were found to be significantly higher during
the post-logging period in wetland core BL-D10; yet, this change only appears at the end of
the second logging period. Aside from a small decrease in dry bulk density at the end of the
first logging period, no other notable changes were seen in the BL-D10 wetland core.
3.3.3
Long-term changes in bulk physical properties
Although 210Pb is only able to date (with any accuracy) approximately the last 100-150 years,
data for the bulk physical properties were still collected beyond the 210Pb-dated region of
the sediment cores. Long-term changes of these proxies allow recent changes to be placed
in a broader context. Dry bulk density and percent water content were measured to various
depths for all three cores (Fig. 3.7). The amount of data available is dependent on the length
of core that could be retrieved from individual sampling locations.
9
•8
0>
CO
BL-P1
BL-D8
BL-D10
T—i—i—i—i—i—r
l—i—i—i—i—i—r
00
65
0.2
04
0.6
Dry Density (gem-*)
75
85
95
Water Content {%)
i
I
r
-1.0 -0.5 00
05
1.0
—&—
Lake (BL-P1)
-•©-
Wetland (BL-D8)
-•A-
Wetland (BL-D10)
&
sc
Magnetic Susceptibility
Fine Organic Matter (Gyttja)
(SI x 10"® m3 kg'1)
Peat / Organic Matter
I
CD
%
et-
K mt
B
CX
Co
CD
p.
S3
»
"e
X3
e
E
£
o
o
o
Year
Figure 3.8: Annual stream discharge (1924-2009) and mean annual precipitation (19012002) values. Stream discharge values are for Quesnel River at Likely, BC and
were taken from the Water Survey of Canada (Environment Canada). Mean
annual precipitation measurements are specific to the Boswell Lake catchment
and were modelled using ClimateBC. The small dashed line represents the
linear regression line for the full time series of stream discharge. The large
dashed lines are linear regression lines in between each set of breakpoints.
Chapter 3. Results: Lake and wetland sedimentation rates
3.4
3.4.1
58
Viewland Lake catchment
Total sedimentation rates
Post-logging sedimentation rates were not found to be significantly differently from prelogging rates in the Viewland Lake core (Table 3.4). The depth profile of Viewland Lake
sedimentation rates (Fig. 3.9) shows a post-logging peak at 2 cm depth, however, a onesample t-test on this core slice revealed it is not significantly greater than pre-logging rates.
On the other hand, the decrease found at the top-most layer (0-1 cm) does reveal that
sedimentation rates dropped below pre-logging rates. This drop below pre-logging rates is
due to the fact that sedimentation rates peak in the late-1940's and do not drop again until
the late-1950's.
On average, post-logging sedimentation rates in the wetland (VL-D1) were significantly
higher than those before logging occurred in the catchment area (Table 3.4). Figure 3.9
shows gradually increasing sedimentation rates over the pre-logging period. Post-logging
sedimentation rates sharply increased at 3 and 6 cm, both of which are significantly greater
than pre-logging rates (£=-13.0, p=<0.001; £=-11.7, p=<0.001).
Table 3.4: Summary of the two-sample t-tests results comparing pre- and post-logging
sedimentation rates (g cm-2 y-1) in Viewland Lake and wetland cores. In VLP1 and VL-D1, the post-logging periods are above 2 cm and 7 cm, respectively.
Values in brackets denote sample size.
VL-P1 (Lake)
Period
Pre
Post
Mean
sd
VL-D1 (Wetland)
p
Mean
sd
p
0.0163 0.0064 (7)
0.0282 0.0259 (15)
0.0206 0.0175 (2) 0.824* 0.0423 0.0101 (7) <0.001*
*Significant at p=0.05
tp was calculated using non-parametric analysis as the assumptions of normality were not
met.
Chapter 3. Results: Lake and wetland sedimentation rates
o
o
o
o
Lake(VL-P1)
Wetland (VL-D1)
in
CD
000
0 02
0.04
006
008
Sedimentation rate (g cm"2 y"')
Figure 3.9: Total sedimentation rates (calculated using the CRS model) for Viewland
Lake and wetland cores. The horizontal line at 1983 represents the year the
Viewland Lake catchment was logged. Error bars on the sedimentation rates
were also calculated using the CRS model.
59
Chapter 3. Results: Lake and wetland sedimentation rates
3.4.2
60
Proxy indicators
Depth profiles of all proxies are found in Figure 3.10. Two-sample t-test results are summa
rized in Table 3.5. With respect to the Viewland Lake core, except for median grain size,
little variation was observed in any of the proxies during the pre-logging period. Median
grain size increased sharply at 13 cm and quickly decreased moving up-core. Apart from C:N,
a significant post-logging change was evident in all proxies. Sharp increases were observed
in dry bulk density and magnetic susceptibility, while percent water content, median grain
size, total C and total N all decreased immediately after the catchment was logged in 1983.
Although the change was not statistically significant, C:N shows evidence of a post-logging
increase. All proxies returned to pre-logging conditions in the 0-1 cm core slice.
Proxy indicators measured for the Viewland wetland core remained relatively consistent
in both the pre- and post-logging periods. Magnetic susceptibility increases significantly
post-logging from an average value of -0.03 to 0.00. At 2 cm, there is a sharp decrease in
dry bulk density and corresponding increase in percent water content, both of which return
to pre-logging conditions at 1 cm.
3.4.3
Long-term changes in bulk physical properties
Deeper profiles (30 cm) of dry bulk density and percent water content were also collected
for the Viewland Lake and wetland cores (Fig. 3.11). The large increase in dry bulk density
and corresponding decrease in percent water content that occurred immediately after the
logging event in 1983 were much higher and lower, respectively, than any other changes that
have taken place over the deeper profile of the lake core. Dry bulk density values calculated
for the Viewland wetland core (VL-D1) increased consistently down-core which may be the
result of sediment compaction over time. However, the post-logging decrease in dry bulk
density observed at 6 cm appears to extend beyond the range of normal variability. The
opposite pattern was observed for percent water content in the wetland core with a strong
increase in percent water content at 6 cm depth.
o
f ^
s. 1
Sn
L sub
K
f°>
a,
F subl sub
F sub
•04
-0.2
00
02
(a) Boswell Lake catchment
(b) Viewland Lake catchment
Figure 4.1: Results of the Principle Component Analysis (PCA) of the fingerprint prop
erties for (a) Boswell Lake and (b) Viewland Lake sediment source materials.
F=forest, F_sub=forest subsoil, L=logged, L_sub=logged subsoil, R=road.
CB=channel bank. Biplots represent the first two principle components of
the PCA.
Table 4.1: Fuzzy k-means clustering results for Boswell Lake and Viewland Lake source
materials.
No. of Groups
1
2
3
4
5
Dunn's coefficient
Boswell Lake Viewland Lake
1.000
0.648
0.513
0.428
0.364
1.000
0.678
0.465
0.399
0.339
Chapter 4. Results: Sediment source tracing
68
surface, subsurface and channel bank samples (i.e. a fairly simple source categorization)
to justify carrying out the final steps of the sediment source tracing procedure. For the
Boswell Lake catchment, a total of 24 out of 34 geochemical properties were found to have
significantly different mean values (corresponding to each of the three source categories; see
Table 4.2). Similar results were found for the Viewland Lake catchment with 29 out of 34
properties having significantly different mean values (Table 4.3). Based on these results it
was determined that three source categories - surface soil material, subsurface soil material,
and channel bank material - would be used in the subsequent sediment source tracing steps.
4.2
Boswell Lake catchment
4.2.1
Composite fingerprint
From the 24 fingerprint properties that were identified by the Kruskal-Wallis H-test to have
at least one pair of significantly different means (Table 4.2), a composite fingerprint
was
developed to correctly label the source group for each sample. Using stepwise multivariate
discriminant function analysis (MDFA)1, Se and A1 were found to be the most appropriate
combination of properties to use in the multivariate unmixing model as they were able to
correctly assign 100% of the source materials to their original groups (Table 4.4). An addi
tional fingerprint property, Ba, was incorporated into the composite fingerprint
to increase
the discriminatory power of the composite fingerprint. The additional property was selected
on the basis of the next property provided by the results of the stepwise MDFA which would
result in overall lower Wilks' lambda values.
'This procedure selects a combination of fingerprint properties that minimizes Wilks' lambda and is able
to distinguish source types within a given catchment. The selected properties then become the composite
fingerprint to be used in the multivariate unmixing model.
Chapter 4. Results: Sediment source tracing
69
Table 4.2: Kruskal-Wallis H-test probabilities (p) for distinguishing surface, subsurface
and channel bank materials in the Boswell Lake catchment using individual
fingerprint properties. Mean concentration values are also given for each fin
gerprint property for each source type.
Fingerprint
property
Surface mean
(mg kg"1)
Li
6.38
Be
0.15
Na
141
Mg
2654
A1
9177
Si
5588
P
955
K
187
Ca
14673
Ti
585
V
34.0
Cr
23.1
Mn
1316
Fe
11274
Co
5.9
Ni
12.3
19.7
Cu
Zn
242.8
As
4.0
Se
0.55
Sr
68.8
Zr
2.8
0.72
Mo
Ag
0.37
Cd
0.58
Sn
0.94
Sb
0.13
Ba
126.7
W
0.06
Hg
0.19
Tl
0.21
Pb
11.62
Bi
0.16
U
0.29
* Significant at p=0.05
Subsurface mean
(mg kg *)
Channel bank mean
(mg kg x)
p
16.42
0.38
265
7227
22085
13315
890
169
10487
1119
84.9
58.7
525
27715
13.3
31.9
52.1
93.5
20.8
1.21
49.5
6.2
0.69
0.63
0.46
0.94
0.47
96.3
0.14
0.07
0.19
7.55
0.26
0.77
3.71
0.15
107
2505
7265
4430
1154
264
39240
346
23.1
20.7
421
9889
3.9
13.8
58.1
50.7
8.4
2.99
129.4
4.4
1.09
0.61
1.13
0.74
0.45
50.0
0.03
0.23
0.16
6.06
0.07
1.01
0.002*
0.010*
0.110
0.018*
0.006*
0.006*
0.044*
0.002*
<0.001*
0.073
0.002*
0.028*
0.008*
0.010*
0.035*
0.078
0.003*
0.203
0.004*
<0.001*
<0.001*
0.039*
0.104
0.039*
<0.001*
0.439
0.163
0.015*
0.592
<0.001*
0.207
0.008*
0.199
0.009*
70
Chapter 4. Results: Sediment source tracing
Table 4.3: Kruskal-Wallis H-test probabilities ( p ) for distinguishing surface, subsurface
and channel bank materials in the Viewland Lake catchment using individ
ual fingerprint properties. Mean concentration values are also given for each
fingerprint property for each source type.
Fingerprint
property
Surface mean
(mg kg x)
5.03
Li
Be
0.18
Na
160
Mg
2943
8181
A1
Si
4926
P
1114
205
K
Ca
23122
Ti
689
V
40.1
Cr
31.4
1085
Mn
Fe
11272
Co
5.0
Ni
13.4
Cu
26.0
Zn
94.7
As
1.6
Se
0.65
Sr
94.7
Zr
3.4
Mo
2.50
Ag
0.54
Cd
0.78
Sn
0.45
Sb
0.10
Ba
125.8
W
0.04
Hg
0.20
0.19
Tl
Pb
10.31
0.03
Bi
U
0.59
* Significant at p=0.05
Subsurface mean
(mgkg x)
Channel bank mean
(mgkg *)
p
13.05
0.43
301
6638
18754
11272
1060
191
11900
1200
99.6
73.4
805
33005
13.1
32.6
55.3
109.0
4.1
1.17
80.9
6.4
1.84
0.54
0.57
0.49
0.14
104.8
0.09
0.08
0.17
6.23
0.06
0.89
5.68
0.25
204
4077
10472
6358
1193
226
13350
486
38.8
34.4
241
12603
5.0
21.3
67.8
45.9
1.5
1.53
71.6
5.3
3.22
0.55
0.42
0.41
0.19
65.4
0.04
0.19
0.15
3.06
0.02
0.72
0.002*
0.010*
0.110
0.018*
0.006*
0.006*
0.044*
0.002*
<0.001*
0.073
0.002*
0.028*
0.008*
0.010*
0.035*
0.078
0.003*
0.203
0.004*
<0.001*
<0.001*
0.039*
0.104
0.039*
<0.001*
0.439
0.163
0.015*
0.592
<0.001*
0.207
0.008*
0.199
0.009*
Chapter 4. Results: Sediment source tracing
71
Table 4.4: Fingerprint properties selected by the stepwise Multivariate Discriminant
Function Analysis to distinguish source types in the Boswell Lake catchment.
Fingerprint property
Se
Al
Ba
4.2.2
Wilks' lambda
0.427
0.241
0.130
Cumulative % source type samples classified
correctly
77.1
100
100
Sediment source contributions
The results of the multivariate unmixing model for Boswell Lake and wetland cores are
shown in Figure 4.2. The dominant sediment sources for both cores were channel bank and
subsurface material. It is important, however, to remember when interpreting these results
that there are errors associated with these percentages. The errors for the Boswell Lake and
wetland source tracing results are summarized in Table 4.5. The percent relative errors are
considerably higher than values that have been documented by other source tracing studies
(Collins et al., 1997; Carter et al., 2003), which report that errors below ±15% provide an
accurate interpretation of the source material proportions. Here, errors range from ca. 5-40%
in the lake core, and ca. 13-50% in the wetland core. These higher error values likely reflect
the high degree of overlap observed among the three source groups with respect to their
geochemical characteristics (see Fig 4.1(a)).
At 6 cm down-core (ca. 1936), the lake core was composed of 100% channel bank material
and gradually received a greater proportion of subsurface material up-core. A maximum of
26% subsurface material was reached at 2 cm depth. A slight decrease to 23% subsurface
material was observed at 1 cm with a corresponding increase in channel bank material. The
increase in subsurface material also coincided with the time periods during which logging
was present (1960-1975, 1982-2008), approximately 3 to 4 cm depth.
The maximum percentage of subsurface material (62%) occurred at 11 cm depth (ca.
1876) in the wetland core (BL-D8). A general decreasing trend of subsurface material was
then observed up-core until a minimum of 16% was reached at 2 cm. Surface material
Chapter 4. Results: Sediment source tracing
72
appeared along with subsurface and channel bank material in the top 1 cm, however, it only
accounted for 6% of the material in the core. The beginning of the post-logging period, which
occurred at approximately 7 cm depth in the wetland core, did not appear to correspond
with any significant changes in sediment source contributions.
2002
2006
1990
2003
1975
1996
1960
1991
1947
1983
1936
1974
1904
1963
1952
1946
1938
1876
40
80
20
Percent contribution
Subsurface
• Surface
(a) Boswell Lake core (BL-P1)
40
60
Percent contribution
Subsurface
• Surface
• Channel bank
(b) Wetland core (BL-D8)
Figure 4.2: Results of the multivariate unmixing model for the (a) Boswell Lake and (b)
wetland cores. Values on the secondary y-axis represent the dates calculated
using the CRS model for each 1 cm core slice containing detectable concen
trations of 210Pb. Each date aligns with the bottom of its respective 1 cm
core segment.
4.2.3
Correlations
Table 4.6 provides a summary of all correlation coefficients and significance values. Sedi
mentation rates were not significantly related to changes in any of the source materials in
the lake core. Changes in source materials were, however, highly correlated with other proxy
indicators (see Table 4.6). In the Boswell Lake core (BL-Pl), median grain size was found
to be inversely related to percent subsurface material. Increasing proportions of subsurface
Chapter 4. Results: Sediment source tracing
73
Table 4.5: Percent relative errors and standard errors for the unmixing model calculations
for the Boswell Lake (BL-P1) and wetland (BL-D8) cores.
BL-Pl
Depth (cm)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
BL-D8
% Error
Standard error
% Error
Standard error
22.0
16.4
18.5
15.9
16.7
20.7
20.2
18.3
15.8
10.2
30.2
39.1
35.7
19.8
16.2
4.5
8.0
13.7
17.4
18.0
0.52
0.33
0.04
0.81
0.01
0.05
0.01
0.07
0.29
0.16
0.04
0.15
0.07
0.02
0.00
0.01
0.05
0.24
0.09
0.37
23.1
24.0
15.6
13.5
13.7
20.4
17.8
25.4
19.1
23.0
21.0
19.4
31.1
44.4
41.9
49.3
41.5
38.3
45.4
41.4
0.01
0.08
0.12
0.11
0.21
0.64
0.54
0.43
1.07
1.78
1.66
0.47
0.92
0.21
0.26
0.40
0.65
0.43
0.68
0.57
Chapter 4. Results: Sediment source tracing
74
material were also associated with increases in total nitrogen. The reverse was true of chan
nel bank material which had a positive relationship with median grain size and a negative
relationship with total nitrogen.
Sedimentation rates in the Boswell wetland (BL-D8) were negatively correlated with
subsurface material, and positively with channel bank material, dry bulk density and median
grain size increased with increasing subsurface material, while percent water content was
found to decrease with greater proportions of subsurface material. Percent channel bank
material had the opposite relationship with these proxy indicators; dry bulk density and
median grain size decreased, and percent water content increased with increasing relative
contributions of channel bank material.
Table 4.6: Summary of the significant (p<0.05) correlations found between each source
material and sedimentation rates, and proxy indicators.
% Surface
Lake (BL-Pl)
Sedimentation rates
dry bulk density
Water content
Magnetic susceptibility
^50
Total carbon
Total nitrogen
C:N
r
0.74
p
% Subsurface
r
% Channel bank
r
NS
NS
NS
NS
NS
-0.77
NS
0.057 0.82
NS
P
NS
NS
NS
NS
0.043
NS
0.023
NS
NS
NS
NS
NS
NS
NS
NS
NS
0.024
0.62
<0.001 -0.85
<0.001 0.86
NS
NS
-0.92
<0.001 -0.93
<0.001 0.83
NS
0.77
-0.83
P
NS
NS
NS
NS
0.042
NS
0.022
NS
Wetland (BL-D8)
Sedimentation rates
dry bulk density
Water content
Magnetic susceptibility
dso
Total carbon
Total nitrogen
C:N
NS = not significant
-0.67
0.88
-0.88
0.92
-0.77
0.041
<0.001
<0.001
NS
<0.001
<0.001
<0.001
NS
Chapter 4. Results: Sediment source tracing
4.3
Viewland Lake catchment
4.3.1
Composite fingerprint
75
According to the Kruskal-Wallis H-test, 29 out of the 34 fingerprint properties were appro
priate to continue on to the stepwise MDFA analysis (Table 4.3). From these 29 properties,
Se and Pb were selected as the composite fingerprint as they were able to correctly identify
the original source group for 100% of the source materials (Table 4.7). To ensure that the
chosen composite fingerprint was a reliable source discriminator, a third fingerprint property,
As, was added. Similar to the selection process for the source materials in the Boswell Lake
catchment, As was chosen because it was the next property selected by the stepwise MDFA
to lower Wilks' lambda values.
Table 4.7: Fingerprint properties selected by the stepwise Multivariate Discriminant
Function Analysis to distinguish source types in the Viewland Lake catch
ment.
Fingerprint property
Se
Pb
As
4.3.2
Wilks' lambda
0.225
0.129
0.082
Cumulative % source type samples classified
correctly
82.0
100
100
Sediment source contributions
Results of the multivariate unmixing model are provided in Figure 4.3. When interpreting
the relative changes in the source type contributions over time it is important to keep in
mind the errors associated with these values. The errors for the Viewland Lake core range
from approximately 39% to 81%. Errors for the wetland core are considerably lower and
range from approximately 12% to 39%. Similar to the unmixing results for the Boswell Lake
catchment, these higher errors are likely due to the high overlap observed for the geochemical
characteristics of the source types (see Fig. 4.1(b)).
Chapter 4. Results: Sediment source tracing
76
Periodic changes in source type contributions are found throughout the Viewland Lake
core. In general, the dominant source contributing to the lake was channel bank material.
Subsurface material is present at 16 and 17 cm down-core constituting approximately 57%
and 40% of the sediment, respectively. Relatively small amounts of surface materials are
present at 11 to 14 cm down-core and do not appear again until 4 cm. Subsurface material
is the dominant source (70%) at 2 cm with surface material making up the other 30%.
Surface and subsurface materials are still present in similar proportions in the top 1 cm,
however, channel bank material is present and accounts for 19% of the material in the lake
core. The timing of these changes in the top 2 cm of the lake core align well with the onset
of forestry practices. Logging in the Viewland Lake catchment area occurred only in 1983
which coincides with 3 cm depth in the lake core.
The wetland core (VL-D1) did not show any variations in source material composition
over time and was composed entirely of channel bank material.
Table 4.8: Percent relative errors and standard errors for the unmixing model calculations
for the Viewland Lake (VL-P1) and wetland (VL-D1) cores.
VL-P1
Depth (cm)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
VL-D1
% Error
Standard error
% Error
Standard error
47.0
44.8
80.5
63.2
62.5
61.3
49.1
51.4
49.8
61.0
57.6
50.7
60.2
56.4
56.2
38.6
38.7
0.05
2.28
1.17
0.85
0.55
0.47
0.48
0.48
0.25
0.65
1.01
1.22
0.86
0.83
0.66
0.60
0.58
14.2
12.0
12.8
13.5
14.0
13.3
23.3
28.5
30.3
31.7
28.7
39.1
45.3
38.1
0.09
0.10
0.09
0.11
0.06
0.11
0.17
0.27
0.34
0.40
0.25
0.62
0.91
0.52
Chapter 4. Results: Sediment source tracing
1996
77
r-
1989
N
1977
f>
1998
1965
1957
1956
V)
1954
!
1953
1951
1948
to
£
SSt
N
ID
1947
0>
1942
O
1933
«-
1923
1904
r*
1938
r>
1888'
1861 #
AO
100
60
40
Percent contribution
• Subsurface
• Surface
60
Percent contribution
H Channel bank
(a) Viewland Lake core (VL-P1)
Subsurface
• Sulace
• Channel bank
(b) Wetland core (VL-D1)
Figure 4.3: Results of the multivariate unmixing model for the (a) Viewland Lake and
(b) wetland cores. Values on the secondary y-axis represent the dates cal
culated using the constant rate of supply model for each 1 cm core slice
containing detectable concentrations of 210Pb. Each date aligns with the
bottom of its respective 1 cm core segment. The asterisk (*) identifies core
slices that were not corrected for particle size due to a lack of material.
Chapter 4. Results: Sediment source tracing
4.3.3
78
Correlations
Correlation analyses were also used to determine if any relationships exist between sediment
sources and sedimentation rates. Source material proportions were also compared to all
proxy indicators (Table 4.9). Changes in sediment source materials were not found to have a
significant relationship with sedimentation rates in Viewland Lake. They were, however, sig
nificantly correlated with several proxy indicators. Increasing dry bulk density and magnetic
susceptibility, as well as decreasing water content, total carbon and total nitrogen were all
associated with increasing percentages of surface materials in the lake core. The same was
true of subsurface material, which was also positively correlated with C:N. Increases in the
percentage of channel bank material were related to decreasing dry bulk density and mag
netic susceptibility, and increasing values of water content, median grain size, total carbon
and total nitrogen.
Variations in sediment source contributions were not compared to sedimentation rates in
the Viewland wetland since sediment sources did not change over time and were composed
entirely of channel bank material.
Table 4.9: Summary of the significant (p<0.05) correlations found between each source
material and each proxy indicator for the Viewland Lake core (VL-P1).
% Surface
Proxy
Sedimentation rates
dry bulk density
Water content
Magnetic susceptibility
dso
Total carbon
Total nitrogen
C:N
NS = not significant
r
0.62
-0.58
0.78
-0.75
-0.61
% Subsurface
r
V
NS
0.007
0.63
0.016 -0.58
<0.001 0.54
NS
<0.001 -0.60
0.009 -0.77
0.57
NS
% Channel bank
r
P
NS
0.007 -0.72
0.66
0.015
0.024 -0.70
NS
0.52
0.011
0.73
<0.001 0.83
0.016
P
NS
0.001
0.004
0.002
0.038
<0.001
<0.001
NS
Chapter 5. Discussion
79
Chapter 5
Discussion
The following chapter discusses the results for each study site independently as the char
acteristics and histories of their catchments are different. Subsequent sections explore the
impacts of local versus regional factors, as well as the importance of landscape position.
Limitations of the study and future research directions are also provided at the end of the
chapter.
5.1
5.1.1
Boswell Lake catchment
Lake sediment
The results of the one and two sample t-tests indicate that lake sedimentation rates did not
change significantly at any point during the post-logging period. A weak or absent logging
signal in downstream lakes has been reported by other paleolimnological studies. Paterson
et al. (1998) found that logging did not have a significant impact on lake chemistry and
species composition, and indicated that site-specific characteristics such as buffer strips, or
rapid re-growth of vegetation can minimize sediment transfers. However, slight increases in
dry bulk density and magnetic susceptibility indicate a possible increase in the delivery of
clastic sediment to the lake from allochthonous sources. Additionally, moderately high (>12)
C:N values are present throughout the entire lake profile which suggests the majority of the
sediment was delivered from a terrestrial source (Meyers & Ishiwatari, 1993), as opposed to
an internal lake source. Furthermore, the underlying bedrock contains limestone (CaC03)
which may cause the sediment to be enriched in inorganic carbon leading to high C:N values,
although further research would be needed to confirm this.
Chapter 5. Discussion
80
Based on the source tracing results, the dominant sediment source to Boswell Lake is
channel bank material, and the secondary source is subsurface soil material. After approx
imately 1947, the proportion of subsurface material delivered to the lake increased. During
this same period, a shift in each of the proxy indicators occurred all providing evidence
for increasing proportions of clastic-rich sediment. This change to increasing proportions of
clastic-rich sediment becomes particularly obvious when dry bulk density and percent water
content are examined over a longer time period (Fig. 3.7). The peak in dry bulk density
at approximately 1976 (3 cm depth) is the highest it has been in over 2,410 years (where
56 cm down-core corresponds to the Bridge River tephra layer). The presence of subsurface
material is, however, not sufficient to account for the unprecedented increase in dry bulk
density since subsurface material has been observed before without similar increases in dry
bulk density (e.g. 9 cm depth).
Based on the location of the Bridge River tephra layer, average sedimentation rates over
the last 2,410 years were found to be approximately 1.9 x 10-3 g cm-2 y-1. Compared to
this value, sedimentation rates increased approximately 4-fold at the beginning of the 1940s
and remained elevated throughout the remainder of the profile. High sedimentation rates
coupled with the erosion of clastic-rich subsurface soil material could account for elevated
dry bulk density values. However, since this shift to increasing proportions of clastic-rich
sediment began in the 1940s, it cannot be attributed solely to forestry practices which began
in 1960.
Statistically, neither variations in Boswell Lake sedimentation rates nor the sediment
characteristics could be explained using changes in climate variables; yet, visible changes in
several hydrometerological variables should be discussed. A break point and subsequent step
increase in average stream discharge for the Quesnel River at Likely, BC occurred in 1944.
This step change was accompanied by increased stream discharge variability, increased mean
annual precipitation, and an increase in the number of frost-free days. Dery et al. (2009)
reported a similar trend in northern Canadian rivers which they described as an intensifica
Chapter 5. Discussion
81
tion of the hydrological cycle. They noted that large-scale climate processes are known to
produce step changes or trend reversals in the hydrological regime of North American rivers.
With respect to the present study, 1945/46 represents a phase shift in the Pacific Decadal
Oscillation (PDO) to a "cool phase". Cool phases in northwestern North America are charcterized by above average October-March precipitation, as well as above average snow pack
and spring stream flow (Mantua & Hare, 2002). This cool phase ended in approximately
1976 and was followed by a warm phase which produced the opposite effect (Woo et al.,
2006).
Surface erosion in an undisturbed catchment is dependent on climate, soil type, vegetation
cover, and water input (Wondzell & King, 2003). Overland flow is not expected to have
caused a significant increase in the erosion and delivery of sediment since it is unlikely
that, based on this climate regime, rainfall intensity would have exceeded the infiltration
capacity of the soil. Therefore, it is reasonable to assume that the majority of sediment
erosion and delivery occurred within the channels. Hooke (1979) found that the amount of
precipitation, antecedent soil moisture conditions, and peak discharge were strong predictors
of channel bank erosion. Significant increases in the amount of precipitation after 1944 may
have therefore driven channel bank erosion. Additionally, the magnitude and timing of spring
snowmelt can strongly impact stream discharge. This biogeoclimatic zone receives up to 50%
of its precipitation as snow. A significant increase in the frost-free period may have resulted
in faster snowmelt thereby increasing channel bank erosion, and possibly the redistribution
of channel bed sediment.
5.1.2
Wetland buffering function
The objective of this study was to characterize the response of the sediment trapping function
of wetlands to a landscape disturbance (i.e. forestry practices). Several cores were taken
throughout Boswell wetland, but only two were analyzed in this study. Cores taken from
the far west channel were selected for analysis because they were successfully retrieved on
Chapter 5. Discussion
82
the first attempt and surface sediments were minimally disturbed. Additionally, cores were
taken from near the wetland inflow and outflow to attempt to characterize the distribution
of sediment along the length of the wetland channel.
In general, the majority of sediment accumulation in the Boswell wetland occurred closest
to the inflow and rapidly decreased moving toward the wetland outflow. This finding is con
sistent with the results of other studies that examined wetland sediment retention (Brueske
& Barrett, 1994; Cahoon, 1994; French et al., 1995; Reed et al., 1997). Forestry activities
did not have a statistically significant impact on sedimentation rates in either wetland core.
However, peaks in the sedimentation rates of the wetland outflow core (BL-D10) correspond
to both periods of forestry activities. The spatial distribution of sediment in a wetland has
previously been attributed to the characteristics of the wetland and channel flow patterns
(Hupp & Bazemore, 1993; Harter & Mitsch, 2003). The only noticeable difference between
the characteristics of the two coring sites in the Boswell wetland channel was the water depth
which increased with increasing distance from the wetland inflow. Harter & Mitsch (2003)
observed that deeper areas within two experimental wetlands had higher sedimentation rates
than shallower areas of the wetlands. Since sedimentation rates were generally higher near
the inflow where the channel is shallower, then the physical characteristics of the sediment
may have influenced its spatial distribution in the wetland.
Median grain size near the wetland inflow experienced a large decrease (i.e. sediment
became finer) after the 1940s and remained significantly low during the active logging periods.
A significant decrease in median grain size may have influenced the conditions necessary
to facilitate sediment deposition. Finer material requires less stream power to remain in
suspension and can therefore be transported over greater distances than coarser material
(Duncan et al., 1987). However, the physical characteristics of the material near the outflow
(i.e. reduced dry bulk density and relatively low magnetic susceptibility values) suggest that
the sediment is not predominantly clastic. Despite the lack of sediment source tracing data
for BL-D10, it could be assumed that a decrease in dry bulk density along with increased
Chapter 5. Discussion
83
sedimentation rates may have been the result of enhanced channel bank erosion (based on
the source tracing results of BL-D8). Nevertheless, deeper areas near the wetland outflow
may have provided more ideal conditions for the deposition and storage of fine sediment.
Sedimentation rates in both wetland cores also possess the same step increase found
in the Boswell Lake core at approximately 1940. Rates of sediment accumulation in the
wetland core near the inflow (BL-D8; 5.2 x 10~2 g cm-2 y_1) are more than double that
of the core near the outflow (BL-D10; 2.3 x 10~2 g cm-2 y_1) which suggests that this
area of the wetland is providing a more effective trapping function. The 1940 step change
also produced small decreases in dry bulk density and magnetic susceptibility. Although
C:N was not greatly affected by the 1940 step change, indicating that the source of the
sediment was still terrestrial, total C and total N experienced large increases shortly after
1940. Total C more than doubled, rising from 11% to 24%. Total N increased from 0.56%
to 1.3%, almost tripling its pre-1940 value. These changes in the physical and chemical
characteristics of the sediment provide strong evidence that this section of the wetland is
trapping primarily allochthonous organic material (Meyers & Ishiwatari, 1993). Moreover,
the strong positive correlation between sedimentation rates and channel bank material along
with a corresponding increase in discharge variability suggests that channel bank erosion
likely intensified during this time.
Although sharp increases in the sedimentation rates near the wetland outflow appear to
correspond with the timing of logging, other drivers of wetland sedimentation rates should
also be recognized. As a result of their sediment and nutrient trapping function, wetlands
tend to be areas of high productivity. Sedimentation rates presented in the current study
reflect total sediment accumulation over time. Trends in mean annual temperature (MAT)
in the Boswell Lake catchment have been found to steadily increase over the last century
by approximately 0.085°C y"1 (£=2.99, p=0.003). Therefore, it is possible that increases in
primary production have contributed to increases in sedimentation rates.
Chapter 5. Discussion
5.2
5.2.1
84
Viewland Lake catchment
Lake sediment
The presence of active logging in the upstream areas of the Viewland Lake catchment did not
have a statistically significant impact on the sedimentation rates found in Viewland Lake.
Although a post-logging increase in sedimentation rates was observed, several much larger
spikes in lake sedimentation rates occurred in the late 1940s to mid 1950s. The results of
the stepwise linear regression for Viewland Lake sedimentation rates indicate that sediment
delivery in this catchment is largely governed by regional climatic processes. Precipitation as
snow experienced a large increase in the early 1900s, peaking in the late 1940s to mid 1950s,
after which it declined until the end of the century. These observations correspond well with
lake sedimentation rates (Fig. 3.12), as well as the phase changes in PDO. The shift to a cool
phase in 1945/46 until approximately 1976/77 would have increased snow pack and spring
freshet leading to greater runoff and sediment delivery (Mantua k Hare, 2002; Woo et al.,
2006).
Despite consistently elevated lake sedimentation rates through the 1940s and 1950s, phys
ical and chemical characteristics of the lake sediment did not change. A high background
C:N suggests that the sediments are largely allochthonous and in-lake productivity does not
make a significant contribution to Viewland Lake. A post-logging decrease in median grain
size is consistent with the results of other studies that found logging activities to increase
the production of fine-grained sediment (Reid &: Dunne, 1984; Tague & Band, 2001). Sharp
increases in both dry bulk density and magnetic susceptibility have previously been associ
ated with periods of land clearance (Thompson et al., 1975; Lott et al., 1994). Thompson
et al. (1975) also noted a decrease in total carbon content, and concluded that these changes
were all indicative of increasing contributions of inorganic allochthonous material. These
changes in the physical properties of the sediment are also supported by the source tracing
results which show evidence of a change from predominantly channel bank material to a mix
Chapter 5. Discussion
85
of subsurface material and surface material (Fig. 4.3(a)).
The possible driver of the two increases in the proportion of subsurface materials found
at the bottom of the core (16 and 17 cm) cannot be identified using the current dataset.
Records of landuse activities are not available for the mid to late 1800s. As well, ClimateBC
is only able to provide modelled climate data beginning in 1901. It has been recognized that
a particle size correction was not available for either of these core slices; however, the inclu
sion of a particle size correction would not drastically alter the source tracing results such
that subsurface material would be excluded from the source tracing results. Based on the
changes that have occurred during the last century it is unlikely that subsurface materials
would have been transported in such high proportions without a significant disturbance. For
example, mining activities would provide the necessary disturbance to increase the erosion
and delivery of subsurface material to downstream waterbodies, and was also a prominent
land use disturbance during this time (late 1800s). However, when considering the phys
ical and chemical changes of the lake sediment associated with the presence of subsurface
material, one would have expected to observe an increase in dry bulk density, magnetic
susceptibility and C:N, none of which appear in the bottom 2 cm.
5.2.2
Wetland buffering function
Statistically, the timing of forestry practices was a significant predictor variable of wetland
sedimentation rates; however, sedimentation rates began to rise in the early 1900s while
logging only occurred in 1983. Although climate variables were not found to explain any
additional variation in wetland sedimentation rates, an overall increase in mean annual pre
cipitation of 0.466 mm y_1 (£=1.58, p=0.117) may have driven increased sediment transport.
A similar line of reasoning was also used by Foster (1995) who could not find a detectable
trend in annual precipitation, but suspected that an increase in precipitation provided the
energy necessary to increase sediment yield. As was discussed for Boswell wetland, possible
increases in primary production should be taken into account as they may have contributed
Chapter 5. Discussion
86
to wetland sedimentation rates. MAT was also found to significantly increase over the last
century by approximately 0.0084°C y"1 (f=3.01, p=0.003). The steady increase of wetland
sedimentation rates could thus be the result of increased MAT driving primary production.
During visual inspection of the core it was found that above 10 cm (mid 1950s) the
dominant material changed from dark brown organic sediment to woody debris. This is
consistent with low magnetic susceptibility values, and decreases in dry bulk density and
high C:N values. Cahoon (1994) found that unmanaged wetlands had higher rates of organic
matter accumulation, especially closer to the wetland inflow. However, he also observed an
increase in the accumulation of minerogenic material which was not observed in the Viewland
wetland. If the wetland was performing a buffering function then evidence of minerogenic
material should have been found in the wetland core. Additionally, such high proportions
of surface and subsurface material should not have been observed in the lake core. Several
explanations for the lack of subsurface material in the wetland core, as well as the presence
of subsurface material in the lake core, have been described below.
When wetland coring was being carried out (late summer) it was noted that water was
not present in the Viewland wetland channel. In general, wetlands that are permanently
inundated provide a more effective sediment trapping function (Johnston, 1991; Hupp &
Bazemore, 1993). As well, the ephemeral nature of flooding in this channel would have pre
vented the growth of wetland vegetation which play an important role in sediment trapping
and stabilization of the wetland bottom. A study by Duncan et al. (1987) tested the ability
of two ephemeral channels to capture various sediment grain sizes derived from logging roads.
They found that throughout periods of active flow, differences in channel length, vegetation
density and the amount of woody debris affected the ability of ephemeral channels to re
duce the delivery of sediment greater than 63 yum to the mouth of the channel, regardless
of stream discharge. However, these characteristics became less important with sediment
less than 63 /mi. Only during extremely low flow conditions was fine sediment retained
in the channels, and slight increases in discharge resulted in the resuspension of previously
Chapter 5. Discussion
87
deposited fine sediment.
The presence of logging roads in the Viewland Lake catchment is thought to be a possible
driver of fine-sediment production. Road density in this catchment is approximately 4-times
that of the Boswell Lake catchment (2.76 km km-1 in the Viewland Lake catchment versus
0.67 km km-1 in the Boswell Lake catchment). In general, surface erosion due to overland
flow in the Viewland Lake catchment would not have contributed a large amount of sediment
to surface waters since it is unlikely that rainfall intensity would not have exceeded infiltration
capacity of the soil. However, surface erosion of roads and subsequent sediment transport
would have been possible since the infiltration rate of compacted road surface is quickly
exceeded during rainfall events. Furthermore, the arrangement of roads relative to streams,
and the number of road-stream crossings can modify the direction and magnitude of water
flow (Jones et al., 2000). The direct connection between the logging road and the Viewland
channel likely increased the rate of sediment delivery from the road surface to the stream
network.
The Viewland Lake catchment was originally included in this study as a secondary site,
and as a result, only one core was collected from the wetland. An upper section of the
lake exists which is also bordered by a wetland with a channelized inflow and outflow (see
Fig. 2.3). This upstream area was also impacted by forestry practices in 1983. Assuming
low water and sediment residence times for the upper lake, and negligible wetland buffering,
sediment could have been transported from another area of the cutblock to Viewland Lake.
Finally, a particle size correction was not available (due to the lack of sediment for
analysis) for the wetland core slice that, based on the timing of logging, would have been
impacted by logging (6 cm). However, the inclusion of a particle size correction should not
have impacted the rank order of the source groups and channel bank material would have
remained the dominant source type. Additionally, changes in the physical and chemical
characteristics do not provide any evidence of an input of subsurface material.
Chapter 5. Discussion
5.3
88
Importance of landscape position
The concept of hydrological connectivity is important when considering the position of a
disturbance in the landscape and its potential to increase sediment yield. At the patch scale,
the factors that have the greatest influence on sediment yield are slope angle, slope length,
and whether runoff will enter a dispersive or channelized pathway (Bracken &; Croke, 2007).
Spicer (1999) used an impact factor in his statistical models to relate likely travel path and
path distance to sedimentation rates in lakes in the central interior of British Columbia.
The impact factor took into consideration slope angle as well as the path of least resistance
down the hillslope. Steeper areas located a shorter distance to a stream or channel were
more likely to deliver sediment to the downstream lake. While, due to sample size, it was
not possible to incorporate an impact factor into the present linear models, the concept of
hydrological connectivity can be used to help explain the spatial distribution of sediment in
Boswell wetland, as well as the presence of subsurface material in Viewland Lake.
Although logging activities had no apparent effect on Boswell Lake sedimentation rates
(see Table 3.2), a disturbance response was observed in the wetland channel from which cores
were collected and analyzed. Notably, the wetland core near the wetland outflow (BL-D10)
exhibited the strongest responses to both logging events. These logging events resulted in
similar areas of deforestation, and both occurred on slopes of moderate inclination. The
main difference between these two events was their locations in the catchment relative to
the channel and the wetland. The first logging event was higher in the catchment while
the second occurred further downstream. By reducing the path length there would have
been less opportunities for in-channel deposition and storage of sediment. Additionally, the
decrease in median grain size observed near the wetland inflow (BL-D8) supports the idea
that an increase in the length of the wetland channel would have been necessary to encourage
sediment deposition, unless extremely low flow conditions were present.
89
Chapter 5. Discussion
5.4
Local versus regional effects
A conceptual framework developed by Bracken & Croke (2007) identified five components
involved in the hydrological connectivity of a catchment.
Figure 5.1 shows four of the
components surrounded by the fifth component, climate. Water and sediment yield, while
they are strongly influenced by landscape position, delivery pathway, runoff potential and
lateral buffering, all are driven by climate variables. The results of this study fit well within
this framework as sedimentation rates in both Boswell Lake and Viewland Lake appear to be
largely driven by regional, medium-term (i.e. decadal) climatic events rather than short-term
localized logging events. In terms of forestry practices, other studies have also found that
lake sedimentation rates (Blais et al., 1998) and resultant lake conditions (Paterson et al.,
1998) were more strongly influenced by regional climatic processes.
Lateral
Buffering
&
Runoff
Potential
Delivery
Pathway
Landscape
Position
&
Climate
Figure 5.1: The components of catchment connectivity (from Bracken h Croke (2007)).
Although forestry practices produced significant responses in sedimentation rates near
the outflow of Boswell wetland (BL-D10), these changes were short-lived and pre-logging
conditions were soon re-established (within approximately four years) . Comparatively, the
step increase in Boswell Lake and wetland sedimentation rates beginning in the 1940s was
sustained throughout the remainder of the sediment profiles and did not return to pre-1940
rates. Similarly, the post-logging change in the dominant sediment source to Viewland
Chapter 5. Discussion
90
Lake from channel bank material to subsurface and surface materials was episodic, although
the system has not yet returned to pre-logging conditions (i.e. dominated by channel bank
material). Ambers (2001) suggested that logging practices "enhance the effect of big storms".
In otherwords, forestry practices prepare the landscape for erosion and sediment transport,
but ultimately suitable hydrological and climatic conditions (i.e. runoff) are required to
mobilize and deliver sediment to downstream waterbodies. This is consistent with the idea
that sediment delivery is limited by the total sediment supply. Therefore, forestry activities
have the potential to increase the amount of sediment on the hillslopes available for transport,
but do not necessarily result in the immediate increase in wetland or lake sedimentation rates.
5.5
Study limitations
The interpretation of the results presented is heavily dependent on the accuracy of the core
chronologies. Conclusions made in this study have relied upon the changes in sedimentation
rates, as well as changes in the physical and chemical characteristics, with respect to the
timing of forestry practices in the catchment. Though, given a different core chronology the
conclusions drawn from this study may have been different. As discussed in Section 2.3.4, the
constant rate of supply (CRS) model was selected because its assumptions were best satisfied
given the unsupported 210Pb profiles. However, the CRS model is not without its flaws. The
calculated ages at the bottom of the profile tend to be over-estimated, and the model does not
take into consideration variable fluxes of 210Pb as in the sediment isotope tomography (SIT)
model. Additionally, the assumption of minimal post-depositional changes to supported and
unsupported 210Pb was made as core chronologies could not be verified using 137Cs activities.
Post-depositional changes to the sediment record are important to consider as they affect
the reconstruction of the core chronology and the interpretation of past environmental condi
tions. Wetland cores were taken from the channels flowing through the wetlands since it was
suspected that the channels are the major delivery pathways for water and sediment moving
down the hillslopes. Water flow in these channels was observed to be negligible and would
Chapter 5. Discussion
91
have facilitated sediment deposition. However, any increases in flow may have resulted in the
erosion of the channel bottom and the subsequent redistribution of the deposited sediment.
Figure 3.3(b) shows the 210Pb and 137Cs profiles for the Viewland wetland core. Compared to
the Viewland Lake core (Fig. 3.3(a)), the activities of these radionuclides experience greater
fluctuations over time. Erosion of the wetland bottom due to increases in channel flow may
have caused these variations. Unconformities or hiatuses in the sediment record as a result
of erosion would produce under-estimates of the total sedimentation rates in the wetland
channels.
Two important aspects of the wetland were not considered in this study which may
have played a critical role in sediment trapping. Firstly, with changing discharge patterns
over the last century, the water level in the lake and wetland would have been affected.
Dead Black Spruce trees (Picea mariana) were found scattered throughout the wetland
suggesting that the water level would have been sufficiently low at one time to allow for
tree seed germination and tree growth. If channel length and depth are in fact critical for
sediment trapping in wetlands, then water level fluctuations
would have impacted where
sediment deposition would have occurred in the wetland. As seen in Boswell wetland, a
sediment trapping response to deforestation was more prominent in the coring site closest
to the wetland outflow, presumably due to a fining of the sediment. With respect to sample
collection, coring locations were selected based on the present-day wetland boundaries which
were determined using spatial data obtained from the British Columbia Ministry of Forests
and Range, as well as observed vegetation type. If locations of sediment deposition changed
with changing water level, then the cores taken may not be representative of an overall
wetland buffering function. Therefore it is possible that the changes in sedimentation rate
noted for the wetland core collected at the wetland-lake interface (which coincided with the
timing of forestry activities) may be due to changes in lake level caused by increased runoff to
the lake as a result of forest harvest activities (i.e. tree removal and increased road network).
Secondly, and also related to the first point, wetland vegetation type was not an integral
Chapter 5. Discussion
92
component of the study. During site investigation and coring it was noted that the domi
nant species in both wetlands were sedges (Carex spp.) and the Yellow Water Lily ( Nuphar
variegata), however, the presence of these species and their spatial distributions could have
changed depending on the extent of the water level fluctuations.
This second point is not of
major concern as it was pointed out by Duncan et al. (1987) that wetland vegetation did not
have an impact on the deposition of the finest grain sizes (<63 /im) in channels, and stream
discharge was a more important factor for sediment deposition. However, it is important to
note that under certain conditions (i.e. ponded wetlands), vegetation type and density are
strong determinants of sediment trapping and resuspension.
5.6
Future research directions
Certain aspects of the study design have limited the conclusions on the sediment trapping
function to be extended to the full areas of Boswell wetland and Viewland wetland. Only
two cores from Boswell wetland, and one from Viewland wetland, were analyzed. Although
the assumption of negligible overland flow in the wetland area has been made, an additional
three channels exist in each of the study catchments. A better overall assessment of the wet
lands' buffering functions could have been attempted had cores from each of those channels
been analyzed. Additionally, one of the channels in the Viewland Lake catchment was not
affected by forestry practices. Natural temporal variability of these systems was established
using a temporal control (i.e. pre-logging conditions), however, analysis of a core from the
unaffected channel would have provided an appropriate control throughout both the pre- and
post-logging periods. Therefore, future studies attempting to evaluate the sediment buffer
ing function of wetlands should consider selecting a study site which has a combination of
impacted and unimpacted areas.
This study, like many others, has selected small catchments to address the research ques
tions and objectives. The advantage of studying a small catchment is that they are generally
less complex and offer fewer opportunities for terrestrial sediment storage which increases
Chapter 5. Discussion
93
the lag time between sediment mobilization and delivery to the catchment outlet. Addi
tionally, these catchments were selected because the logged hillslopes were moderately steep
and directly connected to the channels and thus the downstream wetlands. However, many
catchments do not have such a simple topography, possess only one lake and one wetland,
and have a history of only a single land use type at one point in time. These spatial and
temporal complexities make cumulative effects of disturbance on sediment delivery and wa
ter quality difficult to understand, especially when the disturbance(s) is a non-point source.
Similarly, understanding the cumulative effects of wetland functions on sediment quantity
and water quality is not a straightforward task, and has been largely unstudied.
One study on the cumulative effects of wetlands on sediment and water quality was con
ducted by Johnston et al. (1990) who analyzed 33 watershed variables extracted from aerial
photographs, along with water quality data, for 15 watersheds in the Minneapolis-St. Paul,
Minnesota, USA metropolitan area. This study aimed to identify wetland characteristics
which are, statistically, more likely to impact stream water quality (i.e. improve or degrade)
and quantity. Johnston et al. (1990) found that wetland proximity was an important factor
in determining the water quality of downstream adjacent lakes and streams, and the effect
of wetland functions on water quality was not detectable downstream of the wetland. It
was recognized that further work needs to be done to determine the distance relationships
between wetlands and downstream water quality. Moreover, this thesis has demonstrated
that when landscape disturbances (e.g. forestry practices), which change the physical char
acteristics of the eroded sediment (e.g. grain size), are coupled with hydroclimatic processes
which increase sediment delivery (e.g. runoff), sediment retention in wetland buffers may
either: a) not occur, or b) be limited due to sediment redistribution from wetland channel
bottoms. Therefore, future studies on the cumulative effects of wetlands on water quality
need to also consider how landscape disturbances have altered the hydrological connectivity
of the watershed and the physical characteristics of the sediment, all in the context of local
and regional climatic processes which are in a constant state of change.
Chapter 6. Conclusions and management implications
94
Chapter 6
Conclusions and management implications
6.1
Conclusions
The results of this thesis provide a description of how the sediment trapping function of
two central interior British Columbia wetlands has changed over time in response to forestry
activities and climate processes, in particular precipitation and snowfall. Previous studies
have identified the need to understand wetland functions and their contributions to water
quality over a long time frame as much of the literature contains primarily contemporary
studies. This study has addressed this gap and has also provided information on the origin
of the sediment deposited in the wetlands and the lakes that they buffer, once again in a
temporal context of the last century. More specifically, this thesis aimed to determine if
sedimentation rates and sediment source proportions in two wetlands and their downstream
lakes were impacted by upstream forestry practices.
It was demonstrated that forestry practices produced a strong increase in Boswell wetland
sedimentation rates which was not observed in Boswell Lake. This suggests that Boswell
wetland provides a buffering function which has not been compromised by an increase in
sediment delivery. Nonetheless, differences in sediment deposition between the two wetland
sampling sites (BL-D8 and BL-D10) suggest that certain areas of the wetland provide a more
effective sediment buffering function than others. The effectiveness of the sediment retention
function of wetlands has previously been related to several factors intrinsic to the wetland,
including percentage of wetland coverage, vegetation type and density, and channel depth
and length. Channel depth and length may have been an important factor in the Boswell
wetland as stronger post-logging responses were observed near the Boswell wetland outflow
Chapter 6. Conclusions and management implications
95
where the channel was deeper. However, sedimentation rates were generally higher near
the wetland inflow which is consistent with the spatial distribution of sediment deposition
reported for previously studied wetlands.
A significant decrease in median grain size near the wetland inflow suggests that dif
ferences in the sedimentation rates between these two sampling sites and their reponses to
forestry practices may also be related to the characteristics of the sediment. It is well-known
that, in addition to the water flow conditions, the properties of the sediment also affect
settling rates. Additionally, forestry practices have been reported to increase the production
of fine-grained sediment. A fining of the sediment delivered from the hillslopes could have
increased the distance necessary for sediment deposition to take place. Similarly, in the
Viewland Lake catchment, there was no evidence of wetland storage of minerogenic subsur
face material, however, lake sediment was predominantly composed of subsurface material
and also experienced a significant decrease in median grain size post-logging. It has been
suggested that the ephemeral nature of the wetland channel, and the smaller width of the
wetland (ca. 30 m) limited the buffering function of the wetland. Therefore, it is possible
that the Viewland wetland did not provide a sufficient "buffering distance" to capture the
fine-grained subsurface material observed in the lake.
This study also showed evidence of a strong climatic control on wetland and lake sedi
mentation rates. An intensification of the hydrological regime, which produced an increase in
both the variability and the magnitude of mean annual precipitation and stream discharge
in the Boswell Lake catchment, may have been the result of the 1944/45 shift to a cool
phase in the Pacific Decadal Oscillation. Likewise, sedimentation rates in Viewland Lake
were found to be strongly influenced by snowmelt. While it is unclear as to why these two
catchments, which are relatively close in proximity, are impacted by two different climate
forcings, these findings are ultimately consistent with those of others who have found that
lake sedimentation rates were largely controlled by the amount of runoff generated.
Since Boswell Lake was not significantly impacted by the forestry practices, it was not
Chapter 6. Conclusions and management implications
96
possible to identify a recovery phase. The sedimentation rates at the outflow of the Boswell
wetland have returned to pre-logging rates, however, they appear to have remained within the
new climatic regime that began in the early 1940s. A consistent increase in the sedimentation
rates near the wetland inflow suggests that sediment production from the hillslopes has not
returned to pre-logging rates. Alternatively, increases in sedimentation rates were strongly
related to increasing contributions of channel bank material which may have been eroded
during periods of increased discharge.
Based on the physical and chemical characteristics of the sediment, recovery to prelogging conditions has already occurred in Viewland Lake, however, the results of the un
mixing model do not entirely support this finding. Subsurface material continued to be the
dominant source material in the top 1 cm of the lake core. Since the dominant driver of sub
surface sediment mobilization in this catchment has been assumed to be road construction,
recovery of this system will depend strongly on the amount of use the road receives and the
amount of sediment mobilized during road deactivation (i.e. culvert decommissioning).
Phillips (1989) stated that wetlands offer sites primarily for temporary sediment storage.
Johnston (1991) argued that wetlands are more likely to provide permanent storage, but also
recognized that the importance of a wetland as a "storage compartment" depends on the
flux into the wetland and the duration of retention. Forestry practices in both catchments,
as well as local and regional climatic influences, were shown to impact both the amount
of sediment, and the dominant sediment sources. Boswell wetland offered a more effective
sediment buffering function than Viewland wetland which has been attributed to the depth
and length of its channel, and thus the length of the wetland buffer. However, differences in
the spatial distribution of sediment along the channel were likely influenced by the position
of logging in the catchment relative to the wetland and lake, as well as the characteristics of
the sediment produced by natural erosive processes and forestry practices.
Finally, Smol (1991) recognized that the flow of information between "neolimnologists"1
1Smol (1991)
systems.
used the term "neolimnologist" to refer to a limnologist working with present-day aquatic
Chapter 6. Conclusions and management implications
97
and paleolimnologists needs to be bidirectional. Paleolimnology requires the understanding of
present-day processes to interpret long-term findings, while conversely neolimnology should
assess present-day processes in a long-term context. The purpose of this exercise would
be to understand the importance of "unusual" events on a broader temporal scale. Based
on the findings of this research I would extend this recommendation to paleolimnologists
whose research focuses on the "medium-term". Changes in sedimentation rates and the
characteristics of the sediment which appeared insignificant over the last century, proved to
be meaningful over a longer time frame. Furthermore, larger scale processes such as climate
forcings have been shown here and in other studies to have a greater and more prolonged
influence on hydrological regimes and sediment delivery than forestry practices.
6.2
Management implications
According to the Forest Planning and Practices Regulations under the Forest and Range
Practices Act set out by the British Columbia Ministry of Forests and Range, riparian
reserve zones2 are not required for fish-bearing
streams with a bank-full width less than
1.5 m or non-fish-bearing streams (FPPR s.47(4)). Riparian reserve zones for wetlands are
not required by this same legislation where wetlands are less than 5 ha in size, and not in dry
or wetland sensitive bigeoclimatic zones. Similarly, wetland complexes that are not located
in dry or wetland sensitive biogeoclimatic environments also do not require reserve zones if
they have an aggreagate area less than 5 ha (FPPR s.48).
A Riparian Management Area Guidebook has been established under the regulations to
better define the purpose of these areas and how they should be applied to streams and
wetlands. The objectives of the riparian management areas fail to recognize the importance
of small streams as sediment delivery pathways, and the implications of increased sediment
loading via these pathways as a result of forestry practices. This thesis has demonstrated
2Riparian reserve zones are defined as zones within the riparian management area that are intended to
protect fish, wildlife habitat, biodiversity and the water values of the riparian reserve zone.
Chapter 6. Conclusions and management implications
98
that while wetlands can provide a sediment trapping function, this function can be impacted
by the both the wetland and stream/channel characteristics, as well as the nature of the
mobilized sediment.
Furthermore, one must consider the prevalence of small streams and wetlands in the land
scape versus that of larger streams and wetlands. In the case of the Quesnel River Basin,
only 12% of the wetlands are greater than or equal to 5 ha while 88% are less than 5 ha.
Watershed activities should take into account the cumulative effects on sediment yield at the
basin scale and the consequences of not adequately protecting these smaller areas which are
much more numerous. However, this then raises the point of, what is "adequate protection" ?
Foster et al. (2011) recognized that paleolimnology offers the potential for watershed man
agers to develop site-specific baseline data which can inform water quality guidelines and
therefore management decisions. When regarded in association with wetland and stream
characteristics and their potential for sediment deposition under applicable climatic and
disturbance regimes, a more appropriate and comprehensive management strategy may be
achieved.
6.3
Final remarks
Wetlands perform important hydrologic and geomorphic functions including buffering down
stream waterbodies from accelerated hillslope erosion. However, landscape disturbances such
as forestry practices have been shown to alter the physical characteristics of the sediment
which, under certain hydrological conditions (i.e. high discharge) can cause the sediment
trapping function of wetlands to be impaired. Therefore, additional research is needed to
improve our understanding of wetland functions and their contributions to water quality,
and the conditions under which landscape disturbances diminish those functions.
Bibliography
99
Bibliography
Ambers, R. (2001). Using the sediment record in a western Oregon flood-control reservoir
to assess the influence of storm history and logging on sediment yield. Journal of Hydrology,
244 (3-4), 181-200.
Appleby, P. G., Nolan, P. J., Oldfield, F., Richardson, N., & Higgitt, S. R. (1988). 210Pb
dating of lake sediments and ombrotrophic peats by gamma assay. The Science of the Total
Environment, 69, 157-177.
Appleby, P. G., & Oldfield, F. (1978). The calculation of lead-210 dates assuming a constant
rate of supply of unsupported Pb-210 to the sediment. Catena, 5, 1-8.
Binford, M. W. (1990). Calculation and uncertainty analysis of 210Pb dates for PIRLA
project lake sediment cores. Journal of Paleolimnology, 3, 253-267.
Blais, J., France, R. L., Kimpe, L. E., & Cornett, R. J. (1998). Climatic changes in
northwestern Ontario have had a greater effect on erosion and sediment accumulation than
logging and fire: Evidence from 210 Pb chronology in lake sediments. Biogeochemistry, 43(3),
235-252.
Blais, J. M., Kalff, J., Cornett, R. J., & Evans, R. D. (1995). Evaluation of 210Pb dating
in lake sediments using stable Pb, Ambrosia pollen, and 137Cs. Journal of Paleolimnology,
13, 169-178.
Bosch, J. M., & Hewlett, J. D. (1982). A review of catchment experiments to determine the
effect of vegetation changes on water yield and evapotranspiration. Journal of Hydrology,
55(1-4), 3-23.
Bracken, L. J., & Croke, J. (2007). The concept of hydrological connectivity and its con
tribution to understanding runoff-dominated geomorphic systems. Hydrological Processes,
21, 1749- 1763.
Braskerud, B. C. (2001). The influence of vegetation on sedimentation and resuspension
of soil particles in small constructed wetlands. Journal of Environmental Quality, 30(4),
1447-1457.
Brenner, M., Keenan, L. W., Miller, S. J., & Schelske, C. L. (1999). Spatial and temporal
patterns of sediment and nutrient accumulation in shallow lakes of the Upper St. Johns
River Basin, Florida. Wetlands Ecology and Management, 6, 221-240.
British Columbia Ministry of Environment (2010). Wetlands in B.C. http://www.env.
gov.bc.ca/wld/wetlands.html. [Online; accessed March 6, 2011].
Bibliography
100
British Columbia Ministry of Forests and Range (2008). Biogeoclimatic Ecosystem Clas
sification Subzone/Variant Map for the Central Cariboo Forest District, Horsefly Subunit,
Southern Interior Forest Region. ftp://ftp.for.gov.bc.ca/HRE/external/lpublish/
becmaps/PaperMaps/wall/DCC\_CentralCariboo\_Wall\_Horsefly.pdf.
[Online; ac
cessed December 28, 2010].
Brown, R. G. (1985). Effects of an urban wetland on sediment and nutrient loads in runoff.
Wetlands, 4, 147-158.
Brueske, C., & Barrett, G. (1994). Effects of vegetation and hydrologic load on sedimenta
tion patterns in experimental wetland ecosystems. Ecological Engineering, 3(4), 429-447.
Cahoon, D. R. (1994). Recent accretion in two managed marsh impoundments in coastal
Louisiana. Ecological Applications, 4(1), 166-176.
Carroll, J. L., & Abraham, J. D. (1996). Sediment Isotope Tomography (SIT) Model
Version 1. Tech. rep., U.S. Department of Energy, USA.
Carroll, J. L., & Lerche, I. (2003). Sedimentary processes: quantification using radionu
clides. San Francisco: Elsevier Science Ltd.
Carroll, J. L., Lerche, I., Abraham, J. D., & Cisar, D. J. (1995). Model-determined sediment
ages from 210Pb profiles in un-mixed sediments. Nuclear Geophysics, 9(6), 553-565.
Carter, J., Owens, P. N., Walling, D. E., & Leeks, G. J. L. (2003). Fingerprinting suspended
sediment sources in a large urban river system. The Science of The Total Environment,
314-316, 513-534.
Christie, T., & Fletcher, W. K. (1999). Contamination from forestry activities: implications
for stream sediment exploration programmes. Journal of Geochemical Exploration, 67, 201210.
Church, M., & Eaton, B. (2001). Technical Report #3: Hydrological Effects of Forest Har
vest in the Pacific Northwest. Tech. rep., The University of British Columbia. Vancouver,
Canada.
Clague, J. J., Evans, S. G., Rampton, V. N., Sc Woodsworth, G. (1995). Improved age
estimates for the White River and Bridge River tephras, western Canada. Canadian Journal
of Earth Sciences, 32(8), 1172-1179.
Clarke, S. J. (2002). Vegetation growth in rivers: influences upon sediment and nutrient
dynamics. Progress in Physical Geography, 26(2), 159-172.
Cohen, A. S., Palacios-Fest, M. R., Msaky, E. S., Alin, S. R., McKee, B., O'Reilly, C. M.,
Dettman, D. L., Nkotagu, H., & Lezzar, K. E. (2005). Paleolimnological investigations of
anthropogenic environmental change in Lake Tanganyika: IX. Summary of paleorecords of
environmental change and catchment deforestation at Lake Tanganyika and impacts on the
Lake Tanganyika ecosystem. Journal of Paleolimnology, 34(1), 125-145.
Bibliography
101
Collins, A. L., h Walling, D. E. (2002). Selecting fingerprint properties for discriminating
potential suspended sediment sources in river basins. Journal of Hydrology, 261 (1-4), 218—
244.
Collins, A. L., Walling, D. E., & Leeks, G. J. L. (1997). Source type ascription for fluvial
suspended sediment based on a quantitative composite fingerprinting technique. Catena,
29, 1-27.
Collins, A. L., Walling, D. E., & Leeks, G. J. L. (1998). Use of composite fingerprints to
determine the provenance of the contemporary suspended sediment load transported by
rivers. Earth Surface Processes and Landforms, 23, 31-52.
Cooke, C. A., & Abbott, M. B. (2008). A paleolimnological perspective on industrial-era
metal pollution in the central Andes, Peru. Science of the Total Environment, 393 (2-3),
262-272.
Craft, C. B., & Casey, W. P. (2000). Sediment and nutrient accumulation in floodplain and
depressional freshwater wetlands of Georgia, USA. Wetlands, 20(2), 323-332.
Croke, J., Hairsine, P., & Fogarty, P. (1999). Sediment transport, redistribution and storage
on logged forest hillslopes in south-eastern Australia. Hydrological Processes, 13(17), 2705 2720.
Croke, J., & Mockler, S. (2001). Gully initiation and road-to-stream linkage in a forested
catchment, southeastern Australia. Earth Surface Processes and Landforms, 26, 205-217.
Croke, J., Mockler, S., Fogarty, P., & Takken, I. (2005). Sediment concentration changes in
runoff pathways from a forest road network and the resultant spatial pattern of catchment
connectivity. Geomorphology, 68 (3-4), 257-268.
Davis, J., & Froend, R. (1999). Loss and degradation of wetlands in southwestern australia:
underlying causes, consequences and solutions. Wetlands Ecology and Management, 7(1),
13-23.
Davis, M. B. (1968). Pollen grains in lake sediments: redeposition caused by seasonal water
circulation. Science, 162(3855), 796-799.
Davis, M. B. (1973). Redeposition of pollen grains in lake sediment. Limnology and Oceanog
raphy, 18(1), 44-52.
Davis, R. B., Hess, C. T., Norton, S. A., Hanson, D. W., Hoagland, K. D., &; Anderson,
D. S. (1984). 137cs and 210pb dating of sediments from soft-water lakes in new england (usa)
and scandinavia, a failure of 137cs dating. Chemical Geology, 44 (1-3), 151-185.
Dawson, F. H. (1978). The seasonal effects of aquatic plant growth on the flow of water in
a stream. In Proceedings of the International Symposium on Aquatic Weeds, (pp. 71-78).
Amsterdam, Netherlands: EWRS.
Bibliography
102
de Vente, J. (2007). The sediment delivery problem revisited. Progress in physical geography,
31(2), 155.
Dery, S. J., Hernandez-Henriquez, M. A., Burford, J. E., & Wood, E. F. (2009). Obser
vational evidence of an intensifying hydrological cycle in northern Canada. Geophysical
Research Letters, 55(13), 1-5.
Dorr, H., & Miinnich, K. O. (2006). Downward movement of soil organic matter and its
influence on trace-element transport (210pb, 137cs) in the soil. Radiocarbon, 31(3), 655-663.
Droppo, I. G., Leppard, G., Flannigan, D., & Liss, S. (1997). The freshwater floe: a func
tional relationship of water and organic and inorganic floe constituents affecting suspended
sediment properties. Water, Air and Soil Pollution, 99, 43-54.
Duncan, S. H., Bilby, R. E., Ward, J. W., & Heffner, J. T. (1987). Transport of road-surface
sediment through ephemeral stream channels. Water Resources Bulletin, 23(1), 113-119.
Environment Canada (2010). Water Survey of Canada, http://www.ec.gc.ca/rhc-wsc/
default.asp?lang=En&n=4EED50Fl-l. [Online; accessed February 27, 2010].
Foster, I. D. L. (1994). Using reservoir deposits to reconstruct changing sediment yields
and sources in the catchment of the Old Mill Reservoir, South Devon, UK, over the past
50 years. Hydrological Sciences Journal, 39(4), 347.
Foster, I. D. L. (1995). Lake and reservoir bottom sediments as a source of soil erosion and
sediment transport data in the U.K. In I. D. L. Foster, A. M. Gurnell, Sc B. Webb (Eds.)
Sediment and water quality in river catchments, chap. 15, (pp. 265-283). Michigan, USA:
Wiley.
Foster, I. D. L., Collins, A. L., Naden, P. S., Sear, D. A., Jones, J. I., & Zhang, Y. (2011).
The potential for paleolimnology to determine historic sediment delivery to rivers. Journal
of Paleolimnology, ^5(2), 287-306.
Foster, I. D. L., Dearing, J. A., Simpson, A., Carter, A. D., & Appleby, P. G. (1985). Lake
catchment based studies of erosion and denudation in the Merevale catchment, Warwick
shire, UK. Earth Surface Processes and Landforms, 10( 1), 45-68.
Foster, I. D. L., & Lees, J. A. (1999). Changing headwater suspended sediment yields in
the LOIS catchments over the last century: a paleolimnological approach. Hydrological
Processes, 13(7), 1137-1153.
Foster, I. D. L., Mighall, T. M., Proffitt, H., Walling, D. E., &: Owens, P. N. (2006). Postdepositional 137Cs mobility in the sediments of three shallow coastal lagoons, SW England.
Journal of Paleolimnology, 35(4), 881-895.
Foster, I. D. L., Oldfield, F., Flower, R. J., & Keatings, K. (2008). Mineral magnetic
signatures in a long core from Lake Qarun, Middle Egypt. Journal of Paleolimnology,
40(3), 835-849.
Bibliography
103
French, J. R., Spencer, T., Murray, A. L., & Arnold, N. S. (1995). Geostatistical analysis of
sediment deposition in two small tidal wetlands, Norfolk, U.K. Journal of Coastal Research,
11(2), 308-321.
Gibbs, J. P. (2000). Wetland loss and biodiversity conservation. Conservation Biology,
14(1), 314-317.
Glew, J. R. (2001). Sediment core collection and extrusion. In W. M. Last, & J. P.
Smol (Eds.) Tracking environmental change using lake sediments: Basin analysis, coring,
and. chronological techniques, vol. 1, chap. 5, (pp. 73-106). Dordrecht, The Netherlands:
Kluwer Academic Publishers.
Harr, R. D., Levno, A., & Mercereau, R. (1982). Streamflow changes after logging 130year-old Douglas fir in two small watersheds. Water Resources Research, 18(3), 637-644.
Harter, S., & Mitsch, W. (2003). Patterns of short-term sedimentation in a freshwater
created marsh. Journal of Environmental Quality, 32(1), 325-334.
Hatfield, R., & Maher, B. (2009). Fingerprinting upland sediment sources: particle sizespecific magnetic linkages between soils, lake sediments and suspended sediments. Earth
Surface Processes and Landforms, 34, 1359-1373.
He, Q., & Walling, D. E. (1996). Use of fallout Pb-210 measurements to investigate longerterm rates and patterns of overbank sediment deposition on the floodplains of lowland
rivers. Earth Surface Processes and Landforms, 21(2), 141-154.
Hemond, H. F. (1988). Cumulative impacts on water quality functions of wetlands. Envi
ronmental management, 12(5), 639.
Holm, S. (1979). A simple sequentially rejective multiple test procedure. Scandinavian
Journal of Statistics, 6(2), 65-70.
Hooke, J. M. (1979). An analysis of the processes of river bank erosion. Journal of Hydrol
ogy, 42(1-2), 39-62.
Huang, Y. H., Saiers, J. E., Harvey, J. W., Noe, G. B., Mylon, S., Team, C., &; Valk, V. D.
(2008). Advection, dispersion, and filtration of fine particles within emergent vegetation of
the Florida Everglades. Water Resources Research, 44 •> 1-13.
Hupp, C. R., & Bazemore, D. E. (1993). Temporal and spatial patterns of wetland sedi
mentation, West Tennessee. Journal of Hydrology, 141, 179-196.
Johnston, C. A. (1991). Sediment and nutrient retention by freshwater wetlands: effects on
surface water quality. Critical Reviews in Environmental Control, 21 (5,6), 491-565.
Johnston, C. A., Detenbeck, N. E., & Niemi, G. J. (1990). The cumulative effect of wetlands
on stream water quality and quantity. A landscape approach. Biogeochemistry, 10(2), 105141.
Bibliography
104
Jones, J., Swanson, F., Wemple, B., Sz Snyder, K. (2000). Effects of roads on hydrology,
geomorphology, and disturbance patches in stream networks. Conservation Biology, 14 (1),
76-85.
Jones, J. A., & Grant, G. E. (1996). Peak flow responses to clear-cutting and roads in small
and large basins, Western Cascades, Oregon. Water Resources Research, 32(4), 959.
Keppeler, E. T., & Ziemer, R. R. (1990). Logging effects on streamflow: Water yield and
summer low flows at Caspar Creek in northwestern California. Water Resources Research,
26 (7), 1669-1679.
Ketcheson, M. V., Braumandl, T. F., Meidinger, D., Utzig, G., Demarchi, D. A., & Wikeem,
B. M. (n.d.). Interior Cedar-Hemlock Zone, chap. 11, (pp. 167-181). Forests, Lands &
Natural Resource Operations.
Kim, J. G. (2003). Response of sediment chemistry and accumulation rates to recent
environmental changes in the Clear Lake watershed, California, USA. Wetlands, 23(1),
95-103.
Knox, J. C. (1972). Valley alluviation in southwestern Wisconsin. Annals of the Association
of American Geographers, 62(3), 401-410.
Kohler, M., Gleisberg, B., Sz Niese, S. (2000). Investigation of the soil-plant transfer of pri
mordial radionuclides in tomatoes by low-level gamma-ray spectrometry. Applied Radiation
and Isotopes, 53, 203-208.
Lehre, A. K. (1982). Sediment budget of a small Coast Range drainage basin in north-central
California. In F. J. Swanson (Ed.) Sediment budgets and routing in forested drainage basins,
(pp. 67-77). Portland, Oregon, USA: U.S. Dept. of Agriculture, Forest Service, Pacific
Northwest Forest and Range Experiment Station.
Lott, A.-M., Siver, P. A., Marsicano, L. J., Kodama, K. P., k Moeller, R. E. (1994).
The paleolimnology of a small waterbody in the Pocono Mountains of Pennsylvania, USA:
reconstructing 19th-20th century specific conductivity trends in relation to changing land
use. Journal of Paleolimnology, 12(2), 75-86.
MacKenzie, W. H., & Moran, J. R. (2004). Wetlands of British Columbia: A Guide to Iden
tification. Victoria, BC, Canada: British Columbia Ministry of Forests. Land Management
Handbook No. 52.
Macklin, M. G., & Lewin, J. (2003). River sediments, great floods and centennial-scale
Holocene climate change. Journal of Quaternary Science, 18(2), 101-105.
Mantua, N. J., & Hare, S. R. (2002). The Pacific Decadal Oscillation. Journal of Oceanog
raphy, 58, 35-44.
Martmez-Carreras, N., Krein, A., Udelhoven, T., Gallart, F., Iffly, J. F., Hoffmann, L.,
Pfister, L., & Walling, D. E. (2010). A rapid spectral-reflectance-based fingerprinting ap
proach for documenting suspended sediment sources during storm runoff events. Journal
of Soils and Sediments, 10(3), 400-413.
Bibliography
105
Massey, N., Maclntyre, D., Desjardins, P., & Cooney, P. (2004). GeoFile 2005-6: Digital
Geology Map of British Columbia - Tile NN10 Central B.C. http://www.em.gov.be.
ca/Mining/Geoscience/PublicationsCatalogue/GeoFiles/Pages/2005-6.aspx. [On
line; accessed August 19, 2010].
Menounos, B., Schiefer, E., & Slaymaker, O. (2006). Nested temporal suspended sediment
yields, Green Lake Basin, British Columbia, Canada. Geomorphology, 79(1-2), 114-129.
URL http://linkinghub.elsevier.com/retrieve/pii/S0169555X06000729
Meyers, P. A., & Ishiwatari, R. (1993). Lacustrine organic geochemistry-an overview of in
dicators of organic matter sources and diagenesis in lake sediments. Organic Geochemistry,
20(7), 867-900.
Meyers, P. A., k, Teranes, J. L. (2001). Sediment organic matter. In W. M. Last, & J. P.
Smol (Eds.) Tracking environmental change using lake sediments: Physical and geochemical
methods, vol. 2, chap. 11, (pp. 239-246). Dordrecht, The Netherlands: Kluwer Academic
Publishers.
Mitsch, W. J., & Gosselink, J. G. (2000). Wetlands. New Jersey: John Wiley & Sons Inc.
Motha, J. A. (2003). Determining the sources of suspended sediment in a forested catchment
in southeastern Australia. Water Resources Research, 59(3).
Motha, J. A., Wallbrink, P. J., Hairsine, P. B., & Grayson, R. B. (2002). Tracer properties
of eroded sediment and source material. Hydrological Processes, i£>(10), 1983-2000.
Noller, J. S. (2000). Lead-210 geochronology. In J. Noller, J. Sowers, & W. Lettis (Eds.)
Quaternary geochronology: methods and applications, (pp. 115-120). American Geophysical
Union.
Nowaczyk, N. R. (2001). Logging of magnetic susceptibility. In W. M. Last, & J. P.
Smol (Eds.) Tracking environmental change using lake sediments: Basin analysis, coring,
and chronological techniques, vol. 1, chap. 8, (pp. 155-170). Dordrecht, The Netherlands:
Kluwer Academic Publisher.
Oldfield, F. (1977). Lakes and their drainage basins as units of sediment-based ecological
study. Progress in Physical Geography, 1(3), 460-504.
Owens, P. N., Duzant, J. H., Deeks, L. K., Wood, G. A., Morgan, R. P. C., &; Collins,
A. J. (2007). Evaluation of contrasting buffer features within an agricultural landscape for
reducing sediment and sediment-associated phosphorus delivery to surface waters. Soil Use
and Management, 23(Suppl. 1), 165-175.
Owens, P. N., Petticrew, E. L., & van der Perk, M. (2010). Sediment response to catchment
disturbances. Journal of Soils and Sediments, 10(A), 591-596.
Owens, P. N., & Walling, D. E. (2002). The phosphorus content of fluvial sediment in rural
and industrialized river basins. Water Research, 36, 685-701.
Bibliography
106
Owens, P. N., Walling, D. E., k He, Q. (1996). The behaviour of bomb-derived caesium-137
fallout in catchment soils. Journal of Environmental Radioactivity, 32(3), 169-191.
Owens, P. N., Walling, D. E., k Leeks, G. J. L. (1999). Use of floodplain sediment cores to
investigate recent historical changes in overbank sedimentation rates and sediment sources
in the catchment of the River Ouse, Yorkshire, UK. Catena, 36, 21-47.
Paterson, A. M., Cumming, B. F., Smol, J. P., Blais, J. M., &; France, R. L. (1998).
Assessment of the effects of logging, forest fires and drought on lakes in northwestern
Ontario: a 30-year paleolimnological perspective. Canadian Journal of Forest ResearchRevue Canadienne De Recherche Forestiere, 28 (10), 1546-1556.
Petticrew, E. L., & Kalff, J. (1992). Water flow and clay retention in submerged macrophyte
beds. Canadian Journal of Fisheries and Aquatic Sciences, 49, 2483-2489.
Phillips, J. (2003). Alluvial storage and the long-term stability of sediment yields. Basin
Research, 15, 153-163.
Phillips, J. D. (1989). Fluvial sediment storage in wetlands. Water Resources Bulletin,
25(4), 867-873.
Phillips, J. D. (1995). Decoupling of sediment sources in large river basins. In Effects of
Scale on Interpretation and Management of Sediment and Water Quality, vol. 226, (pp.
11-16). International Association of Hydrological Sciences.
Pike, R. G., & Scherer, R. (2003). Overview of the potential effects of forest management
on low flows in snowmelt-dominated hydrologic regimes. BC Journal of Ecosystems and
Management, 3( 1), 1-17.
Powell, M. (1998). Patterns and processes of sediment sorting in gravel-bed rivers. Progress
in Physical Geography, 22(1), 1-32.
Pye, K. (1994). Sediment transport and depositional processes. University of Reading PRIS
contribution. Blackwell Scientific Publications.
Reasoner, M. (1993). Equipment and procedure improvements for a lightweight, inexpen
sive, percussion core sampling system. Journal of Paleolimnology, 8(3), 273-281.
Reasoner, M. A., & Healy, R. E. (1986). Identification and significance of tephras encoun
tered in a core from Mary Lake, Yoho National Park, British Columbia. Canadian Journal
of Earth Sciences, 23, 1991-1999.
Reed, D. J., de Luca, N., k Foote, A. L. (1997). Effect of hydrologic management on marsh
surface sediment deposition in Coastal Louisiana. Estuaries, 20(2), 301-311.
Reid, L. M., & Dunne, T. (1984). Sediment production from forest road surfaces. Water
Resources Research, 20(11), 1753-1761.
Reinhardt, C. H., Cole, C. A., k Stover, L. R. (2000). A method for coring inland, freshwater
wetland soils. Wetlands, 20(2), 422-426.
Bibliography
107
Rice, W. R. (1989). Analyzing tables of statistical tests. Evolution, 43( 1), 223-225.
Ropelewski, C., & Halpert, M. (1986). North american precipitation and temperature
patterns associated with the el nino/southern oscillation (enso). Mon. Weather Rev.;(United
States), 114(12), 2352-2362.
Sand-Jensen, I<. (1998). Influence of submerged macrophytes on sediment composition and
near-bed flow in lowland streams. Freshwater Biology, 39, 663-679.
Sand-Jensen, K., & Mebus, J. R. (1996). Fine-scale patterns of water velocity within
macrophyte patches in streams. Oikos, 76(1), 169-180.
Sand-Jensen, K., & Pedersen, O. (1999). Velocity gradients and turbulence around macro
phyte stands in streams. Freshwater Biology, 42, 315-328.
Sandgren, P., h Snowball, I. (2001). Application of mineral magnetic techniques to paleolimnology. In W. M. Last, h J. P. Smol (Eds.) Tracking Environmental Change Using
Lake Sediments: Physical and Geochemical Methods, vol. 2, (pp. 217-236). Dordrecht, The
Netherlands: Kluwer Academic Publishers.
Schindler, D. W. (1977). Evolution of phosphorus limitation in lakes. Science, 195(4275),
260-262.
Sheridan, G. J., & Noske, P. J. (2007). A quantitative study of sediment delivery and
stream pollution from different forest road types. Hydrological Processes, 21(3), 387-398.
Slaymaker, O. (2001). Why so much concern about climate change and so little attention
to land use change? Canadian Geographer, 45(1), 71-78.
Smol, J. P. (1991). Are we building enough bridges between paleolimnology and aquatic
ecology? Hydrobiologia, 214(1), 201-206.
Smol, J. P. (2008). Pollution of Lakes and Rivers: a Paleoenvironmental Perspective.
Oxford: Blackwell Publishing, 2nd ed.
Smol, J. P. (2010). The power of the past: using sediments to track the effects of multiple
stressors on lake ecosystems. Freshwater Biology, 55((Suppl. 1)), 43-59.
Snowball, I., & Sandgren, P. (2001). Application of mineral magnetic techniques to pa
leolimnology. In W. M. Last, & J. P. Smol (Eds.) Tracking environmental change using
lake sediments: Physical and geochemical methods, vol. 2, (pp. 217-237). Kluwer Academic
Publishers.
Sperazza, M., Moore, J. N., & Hendrix, M. S. (2004). High-resolution particle size analysis
of naturally occurring very fine-grained sediment through laser diffractometry. Journal of
Sedimentary Research, 74 (5), 736-743.
Spicer, C. P. (1999). Evaluating the impacts of forest harvesting and natural disturbance
events on sediment yields in small watersheds throughout British Columbia, Canada. Msc,
University of Northern British Columbia, Prince George.
Bibliography
108
Spittlehouse, D. (2006a). Annual water balance of high elevation forest and clearcuts.
In Proceedings of the 27th Conference on Agricultural and Forest Meterology. San Diego,
California, USA: American Meterological Society.
Spittlehouse, D. (2006b). ClimateBC: Your access to interpolated climate data for BC.
Streamline Watershed Management Bulletin, 9(2), 16-21.
Srivastava, J., Gupta, A., &; Chandra, H. (2008). Managing water quality with aquatic
macrophytes. Reviews in Environmental Science and Biotechnology, 7(3), 255-266.
Stednick, J. (1996). Monitoring the effects of timber harvest on annual water yield. Journal
of Hydrology, 176(1-4), 79-95.
Swanson, F. J., Janda, R. J., Dunne, T,, & Swanston, D. N. (1982). Sediment Budgets and
Routing in Forested Drainage Basins. USDA Forest Service. Pacific Northwest Forest and
Range Experiment Station, General Technical Report PNW-141.
Tague, C., & Band, L. (2001). Simulating the impact of road construction and forest
harvesting on hydrologic response. Earth Surface Processes and Landforms, 26, 135-151.
The Angler's Atlas (2010). Boswell Lake, http://www.cinglersatlas.com/lakes/190/.
[Online; accessed June 1, 2009].
Thompson, R., Battarbee, R.. O'Sullivan, P., k Oldfield, F. (1975). Magnetic susceptibility
of lake sediments. Limnology and Oceanography, 20(5), 687-698.
Tome, A. R. (2004). Piecewise linear fitting and trend changing points of climate parame
ters. Geophysical Research Letters, 31(2), 2-5.
Trauwaert, E. (1988). On the meaning of Dunn's partition coefficient for fuzzy clusters.
Fuzzy Sets and Systems, 25(2), 217-242.
Turner, L. J., & Delorme, L. D. (1996). Assessment of 210Pb data from Canadian lakes
using the CIC and CRS models. Environmental Geology, 28(2), 78-87.
University of British Columbia (2010). ClimateBC. http://www.genetics.forestry,
ubc.ca/cfcg/climate-models.html. [Online; accessed September 1, 2010].
van Hengstum, P. J., Reinhardt, E. G., Boyce, J. I., & Clark, C. (2007). Changing sedimen
tation patterns due to historical land-use change in Frenchmans Bay, Pickering, Canada:
evidence from high-resolution textural analysis. Journal of Paleolimnology, 37, 603-618.
Viles, H. A., Naylor, L. A., Carter, N. E. A., & Chaput, D. (2008). Biogeomorphological
disturbance regimes: progress in linking ecological and geomorphological systems. Earth
Surface Processes and Landforms, 33, 1419-1435.
von Gunten, L., Grosjean, M., Beer, J., Grob, P., Morales, A., & Urrutia, R. (2008). Age
modeling of young non-varved lake sediments: methods and limits. Examples from two
lakes in Central Chile. Journal of Paleolimnology, 42(3), 401-412.
Bibliography
109
Walling, D. E. (1983). The sediment delivery problem. Journal of Hydrology, 65(1-3), 209.
Walling, D. E. (1999). Linking land use, erosion and sediment yields in river basins. Hydrobiologia, 410, 223.
Walling, D. E., Collins, A. L., & Stroud, R. W. (2008). Tracing suspended sediment and
particulate phosphorus sources in catchments. Journal of Hydrology, 350 (3-4), 274-289.
Walling, D. E., & Woodward, J. C. (1995). Tracing sources of suspended sediment in river
basins: A case study of the River Culm, Devon, UK. Marine and Freshwater Research, 46,
327-336.
Walling, D. E., Woodward, J. C., Sz Nicholas, A. P. (1993). A multi-parameter approach
to fingerprinting suspended-sediment sources. In Proceedings of the Yokohama Symposium,
vol. 215, (pp. 329-338). International Association of Hydrological Sciences.
Wemple, B. C., & Jones, J. A. (1996). Channel network extension by logging roads in two
basins, western Cascades, Oregon. Water Resources Bulletin, 32(6), 1195-1207.
Wondzell, S. M., &: King, J. G. (2003). Postfire erosional processes in the pacific northwest
and rocky mountain regions. Forest Ecology and Management, 178(1-2), 75-87.
Woo, M.-K., Thome, R., & Szeto, K. K. (2006). Reinterpretation of streamflow trends based
on shifts in large-scale atmospheric circulation. Hydrological Processes, 20, 3995-4003.
Yu, L., & Oldfield, F. (1989). Mixing model for identifying sediment source from magnetic
measurements. Quaternary Research, 32, 168-181.
Zedler, J. B., k Kercher, S. (2005). Wetland resources: Status, trends, ecosystem services,
and restorability. Annual Review of Environment and Resources, 30 , 39-74.
Appendix A. Bathymetric maps
Appendix A
Bathymetric maps
110
100
200
300
400
600
Metres
CAUTION: DO NOT USE THIS HAP FOR NAVIGATIONAL PURPOSES
This map may not reflect current conditions. Uncharted hazards may exist
NOTE S OENOTES BENCH MARK
INSET MAP
SURVEYE0 BY' T WEBBER
OATE-JUNE 19,1979
SHORE OUTLINE FROM' AIR PHOTO
STATISTICS AT TIME OF SURVEY
1 ELEVATION
975 m .
2 SURFACE AREA
126,000 *m.
3. VOLUME
148,000 cu.m.
4 EST ANNUAL FLUCTUATION
5 MEAN DEPTH
1.2 m.
6. MAX OEPTH
2.5 m.
7. PERIMETER
1850 m
8. AREA, 6M CONTOUR
9. HEIGHT OF BENCH MARK
ABOVE WATER LEVEL
2m
FISH AND WILDLIFE BRANCH
MINISTRY OF RECREATION AND CONSERVATION
INVENTORY SECTION
BOSWELL LAKE
u.T.it. co-owowart
BATE' J** 9, 1999
CALCULATIONS' T k •
M.0TT1M • U. *•
rm owe. K.fnomt
Figure A.l: Bathymetric map for Boswell Lake. Map was obtained online from the Anglers' Atlas.
/0JSU39J/f
CMfCK 4M.BM.KWHi.
1-3700
93 A/II
Appendix A. Bathymetric maps
112
N
Contour (meters)
100 Meters
Figure A.2: Bathymetric map for Viewland Lake. Map was created in ArcGIS using
latitude-longitude coordinates and water depths obtained during a depth
survey of the lake.
Appendix B. ClimateBC variables
113
Appendix B
ClimateBC variables
Directly calculated variables:
MAT mean annual temperature (°C)
MWMT mean warmest month temperature(°C)
MCMT mean coldest month temperature (°C)
TD temperature difference between MWMT and MCMT (°C)
MAP mean annual precipitation (mm)
MSP mean annual summer (May to September) precipitation (mm)
AH:M annual heat:moisture index ((MAT+10)/(MAP/1000))
SH:M summer heat:moisture index ((MWMT)/(MSP/100))
Derived variables:
DD<0 degree-days below 0°C, chilling degree-days
DD>5 degree-days above 5°C, growing degree-days
DD510o the Julian date on which DD>5 reaches 100, the date of budburst for most plants
DD<18 degree-days below 18 °C, heating degree-days
DD>18 degree-days above 18 °C, cooling degree-days
NFFD the number of frost-free days
FFP frost-free period
bFFP the Julian date on which FFP begins
eFFP the Julian date on which FFP ends
PAS precipiation as snow (mm)
EMT extreme minimum temperature over 30 years
Appendix C. Microscope image of tephra
Appendix C
Microscope image of tephra
250 pm
Figure C.l: Microscope image of the tephra found in both the Boswell Lake and Viewland
Lake cores. Tephra was identified as having originated from the Bridge River
event (ca. 2,410 calendar years BP) based on the glass shard morphology and
tephra colour.
114
Appendix D. Lead-210 dating models
Appendix D
Lead-210 dating models
115
Appendix D. Lead-210 dating models
2000
(a) Boswell Lake core (BL-P1)
(b) Wetland core (BL-D8)
(c) Wetland core (BL-D10)
Figure D.l: Comparison of the 210Pb-based depth-to-age models (CIC, CRS, SIT) for
(a) Boswell Lake and wetland cores (b) BL-D8 and (c) BL-D10. Error bars
are not given to enhance the readability of the figure.
116
Appendix D. Lead-210 dating models
(a) Viewland Lake core (VL-P1)
(b) Wetland core (VL-D1)
Figure D.2: Comparison of the 210Pb-based depth-to-age models (CIC, CRS, SIT) for (a)
Viewland Lake and (b) wetland cores. Error bars are not given to enhance
the readability of the figure.
117