Content uploaded by Gryphen Goss
Author content
All content in this area was uploaded by Gryphen Goss on Jun 06, 2022
Content may be subject to copyright.
UNIVERSITY OF CALGARY
Glacier retreat and fluvial landscape response
by
Gryphen Amarinda Goss
A THESIS
SUBMITTED TO THE FACULTY OF GRADUATE STUDIES
IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE
DEGREE OF MASTER OF SCIENCE
GRADUATE PROGRAM IN GEOSCIENCE
CALGARY, ALBERTA
AUGUST, 2021
© Gryphen Amarinda Goss 2021
i
Abstract
This thesis investigates the planform response of proglacial fluvial systems to
deglaciation, specifically in northwest North America. Increase glacier retreat is extremely
widespread today, drawing our attention to proglacial environments. Fluvial systems are
being directly impacted by increasing rates of glacier retreat, resulting in changes to
discharge, erosion and deposition, and sediment supply. It is important to understand how
fluvial systems are going to reconfigure their geometries in response to these changing
parameters and to help identify and assess future behaviour. Fluvial response to
deglaciation was studied by two experiments, first focusing on the response to an
instantaneous and dramatic change in discharge without a concomitant change in sediment
supply, and second, focusing on slower changes, where sediment supply may have altered.
In the first experiment, a rapid drop in discharge in Äy Ch (ay-CHEW,
formerly known as Slims River) decreased the overall braiding complexity, while a rapid
increase in discharge to Kaskawulsh River increased the braiding complexity. Further,
alluvial fans in the Kaskawulsh River valley were eroded due to the wider active braid
plain, while fans within Äy Ch underwent minimal changes. In the second experiment,
at 9 of 11 locations, where changes over multiple decades occurred, braiding intensity
markedly decreased as proglacial lake area increased, also increasing the number of wide
channels at the expense of narrower ones. The remaining two locations slightly increased
in braiding intensity without ever forming a proglacial lake (control sites).
This appears to be a ubiquitous trend in glaciated regions, as retreat affects sensitive
proglacial rivers by disrupting the flow of discharge and sediment supply. Glacier retreat
has resulted in drainage reorganization, directly altering levels of flow, and caused the
ii
formation of proglacial lakes, altering discharge and sediment supply to the rivers inducing
geomorphic changes. These alluvial river systems have reconfigured their geometry in
response to changing discharge and sediment supply.
iii
Epigraph
In matters of science curiosity gratified begets not indolence,
but new desires.
-James Hutton
iv
Acknowledgements
I must first acknowledge the endless support from my supervisor, Dan Shugar. During my
thesis, the world was turned upside down due to the coronavirus, making it very difficult to do any
research. After cancelling my fieldwork and scrambling for a new direction, Dan was incredibly
patient and helpful in keeping me on track and headstrong. There wasnt a moment of insecurity
or worrying about not completing my thesis on time. Jim Best was also incredibly supportive and
patient as I attempt to navigate being a first-time graduate student in a time of a pandemic. I am a
better researcher because of him. Together, both Dan and Jim provided the most critical and
uplifting perspective I could have imagined.
My labmate and roommate during my entire masters, Meghan Sharp, was a key component
in the completion of my masters and my sanity. From early mornings biking to campus to early
mornings walking from our beds to our desks down the hallway after going remote, she sustained
my motivation. I would also like to thank my family, Leslee Goss, Glenn Goss, Damon Myers,
and David Bailey, who is my incredibly supportive family and provided me with the positivity I
needed during the COVID lockdown and the many stressful moments that occurred.
Lastly, I would like to thank two external sources of funding, the Arctic Institute of North
Americas Grant-In-Aid award and the Denton, Andrews, Porter Glacial Geology Award from
the Geological Society of America. Both of which greatly aided my masters research and
motivation.
v
Table of Contents
Abstract i
Epigraph iii
Acknowledgements iv
Table of Contents v
List of Figures vii
List of Tables x
List of Symbols xi
Chapter 1: Introduction 1
1.1 Channel form and fluvial processes 1
1.1.2 Discharge and sediment flux 4
1.1.3 Channel bed degradation and aggradation 7
1.1.4 Braided systems 8
1.3 Climate change impact on proglacial systems 11
1.3.1 Proglacial lakes 12
1.4 Objectives 13
1.5 Thesis outline 14
1.6 Statement of contribution 15
Chapter 2: Study areas 16
2.1 Kaskawulsh Glacier 16
2.2 Glaciers of western North America 18
2.2.1 Southeast Alaska 19
2.2.2 Yukon 22
2.2.2 British Columbia and Southeast Alaska 22
Chapter 3: Methods 26
3.1 Data collection and processing 26
3.1.1 Satellite imagery for inter-annual mapping 26
3.1.2 Satellite imagery for multi-decadal mapping 26
3.1.3 Historical air photos 27
3.2 Quantifying braiding intensity 29
3.3 Discharge data from proximal stream gauges 31
3.4 Minimizing error and discounting external factors 33
vi
Chapter 4: Results 35
4.1 Fluvial response on the Äy Ch and Kaskawulsh River from 2013-2020 35
4.1.1 Braiding intensity 35
4.1.2 Downstream wetted channel width 38
4.1.3 Areal change of alluvial fans 39
4.2 Multi-decadal changes in proglacial lake extent and river geometry 41
4.2.1 Braiding complexity and proglacial lake development 41
4.2.2 Total wetted width 43
4.2.4 East Fork Susitna fluvial response to proglacial lake formation 45
Chapter 5: Discussion 47
5.1 The effects of altering discharge 47
5.2 Upstream sediment trap induces downstream changes 50
Chapter 6: Conclusions and future research 54
6.1 Conclusions 54
6.2 Future research 54
Appendices 56
Appendix 1: Data Collection and Processing for Kaskawulsh Glacier 56
Appendix 2: Data Collection for All Study Sites 58
Appendix 3: Missing Meteorological Data 59
References 60
vii
List of Figures
Figure 1.1 Channel form is driven by sediment size and supply, and channel gradient and
stability. From Church (2006). Pg 3
Figure 1.2 Flow velocity mean channel depth and mean channel width, all have a positive
relationship with discharge. From Leopold and Maddock (1953). Pg 5
Figure 1.3 Suspended sediment and discharge have a positive relationship, while total dissolved
solids have the opposite relationship. From Ballantyne and Murton (2018). Pg 7
Figure 1.4. Active braid plain of the Äy Ch system in SW Yukon in 2019. Photo by author.
Pg 9
Figure 1.5 Overview of different Braiding Indices. The Channel Count Index is used in this thesis.
From Egozi and Ashmore (2008). Pg 10
Figure 1.6 A diagram showing hydroclimatic effects of climate change and the potential
sedimentary and geomorphic responses. From East and Sankey (2020) Pg 12
Figure 2.1 Map of the study area in southwest Yukon. Kaskawulsh Glacier terminates at the
drainage divide of Äy Ch and Kaskawulsh River. White arrows indicate flow directions.
Alluvial fans in Äy Ch (white circles) and Kaskawulsh (white squares) rivers, described in the
text, are labelled. White boxes indicate locations shown in figure 2.2. Source: Sentinel-2 (2017-
07-16). Pg 17
Figure 2.2 Segments of the Äy Ch before and after river capture (top row), and Kaskawulsh
River before and after river capture (bottom row). Images of each river at peak melt and changes
in discharge are attributable to the river capture event. Sources: Rapideye, SPOT-5, PlanetScope.
Pg 18
Figure 2.3 Map showing locations of the study sites. Names indicated in the legend. Pg 19
Figure 2.4 Sites 1 and 2, Kukak Bay and Hallo Glacier, respectively. Located in southwest Alaska
on the coast of the Aleutian Islands. Site 1 is a control site. Pg 20
Figure 2.5 Sites 3 5, located in the central Alaska Range. East Fork Susitna Glacier and NW
Robertson, sites 3 and 4, produce a proglacial lake over the past 60-70 years, while Robertson, site
5, does not. Pg 21
Figure 2.6 Äy Ch and Kaskawulsh River, sites 6 and 7, respectively, reside in southwest Yukon.
Pg 22
Figure 2.7 Meade and Tulsequah Glacier, sites 8 and 9, are both tributaries of the Juneau Icefield.
Pg 23
viii
Figure 2.8 Klinaklini, site 10, and Lillooet, site 11, are in southwest British Columbia, the lowest
latitude sites. Lillooet contained a small lake in 1951. Pg 24
Figure 3.1. This workflow outlines the process of collecting and preparing satellite imagery and
historical air photos for processing in ArcGIS Pro (Agisoft Metashape Professional (Software),
n.d.; ArcGIS Pro, Esri Inc., n.d.). Pg 28
Figure 3.2 Flow stage is lower in the earlier period of the summer, compared to peak flows later
in the summer. Pg 30
Figure 3.3 Discharge of Klinaklini River (top) from 1977 to 2019. Discharge of Tulsequah River
via the Taku River (bottom) from 1987 to 2021. This stream gauge is located downstream of the
confluence of the Taku River and the Tulsequah river. Pg 32
Figure 3.4 Wetted channels were digitized within 3 pixels of feature boundary. Pg 33
Figure 3.5 Precipitation and temperature from Kluane Lake Research Station (left) and Haines
Junction (right) weather stations. Provided by Gwenn Flowers at Simon Fraser University and
Environment Canada, respectively. Pg 34
Figure 4.1 Äy Ch (A) and Kaskawulsh River (B) braiding intensity from 2013 to 2020. The
Red dashed line is the time of river capture. Pg 36
Figure 4.2 Seasonal variation in braiding intensity from May-June, and July-August, from 2013-
2020. Äy Ch (top) and Kaskawulsh River (bottom). The Red dashed line is the time of river
capture. Pg 37
Figure 4.3 Braiding intensity and total channel wetted width for Äy Ch (A) and Kaskawulsh
River (B). Pg 38
Figure 4.4 Äy Ch alluvial fans (A) underwent minor changes during the river capture event.
Kaskawulsh fans (B) underwent increased erosion. Pg 39
Figure 4.5 Äy Ch fans (top) underwent minimal changes as discharge decrease, while
Kaskawulsh (bottom) underwent increased erosion as discharge increased. Pg 40
Figure 4.6. Nine sites undergo reduced braiding intensity while the proglacial lake area
increased. At two sites, the braiding intensity increased without any formation of a proglacial
lake. East Fork Susitna Glacier is seen in row one, Meade Glacier in row two, and Klinaklini
Glacier in row three. Each glacier, at time 1, terminates directly in the river, where at time 2,
glacier retreat has resulted in a proglacial lake. Each associated river, at time 1, resembles a
relatively highly braided river compared to time 2, where braiding has greatly reduced with the
formation of a proglacial lake. Pg 42
Figure 4.7 At nine of the eleven sites, the frequency of smaller channels reduced reflecting a
reduction in braiding. At two of the sites where no lake formed, the frequency of channels was
only minorly effected. Pg 44
ix
Figure 4.8 Time series of proglacial lake size and braiding intensity (inset) and total wetted width
for East Fork Susitna River. Pg 46
Figure 5.1. Schematic illustrating planform change of a proglacial river as lake area increases.
Overall braiding and channel width reduce as lake area increases. Sketch-based on East Fork
Susitna Glacier. Pg 51
x
List of Tables
Table 2.1 Each study site, region, and RGI ID. http://www.glims.org/maps/gtng . Pg 25
Table 3.1. River dimensions and number of cross-sections used. Pg 30
Table 4.1 Proglacial Lake area and CCI at time 1 and 2. Grey shaded regions are control sites. Pg
41
Table 4.2 Median Total Wetted Width at time 1 and 2. Pg 43
Table A1. Satellite, resolution, symbology, and dates each satellite supplied. Pg 55
Table A2. Satellite, historical air photos, data source, resolution, symbology, and dates each
satellite supplied. USGS data is collected from Earth Explorer. NAPL (National Air Photo
Library). Pg 57
Table A3. Number of days missing data for weather stations (Haines Junction (HJ) and Kluane
Lake Research Station (KLRS)). Precipitation (P) and Temperature (T). Pg 58
xi
List of Symbols
w = Water surface width (m)
d = Mean depth (m)
v = Mean flow velocity (m/s)
b, f, m = Slope of line relative to discharge, Figure 1.2
a, c, k = Intercepts of line relative to discharge, Figure 1.2
Q = Discharge (m3/s)
𝜏 ∗ = Shear stress (N/m2)
𝜌 = Fluid density (kg/m3)
g = Acceleration of gravity (m/s2)
S = Slope (m/m)
𝜌 = Sediment density (kg/m3)
D = Grain size (mm)
Qs = Sediment Flux (kg/m2/y)
CCI = Channel Count Index
Nwc = Number of wetted channels
Nxs = Number of primary cross-sections
1
Chapter 1: Introduction
Anthropogenic impacts, such as global warming and environmental degradation, have
resulted in an accelerated geomorphic change in many Earth systems, including cryospheric and
fluvial environments (Adler et al., 2019; Wasson, 1996). Climatic changes have resulted in
increased glacier retreat, increased growth of glacial lakes, and drainage reorganization, all
modifying fluvial systems (Huss & Hock, 2018; Marren, 2005; Shugar et al., 2017, 2020). Rivers,
especially those fed by glaciers, are at risk of major changes that may completely alter their
geometry. Understanding how fluvial systems are responding to deglaciation over the past century
will assist future assessment and prediction of channel form and behaviour.
Much of society is built around and utilizes resources from fluvial systems and trusts that
the geometry and channel form remain stable through time. Once these systems begin to change,
society must anticipate and adapt to these changes. Fluvial behaviour and channel form is driven
by several factors including discharge, sediment supply and size, gradient, and width (Leopold et
al., 1964), and is susceptible to alteration as each change due to a variety of factors including
anthropogenic impacts.
1.1 Channel form and fluvial processes
Channel pattern (meandering, braided, anastomosing, and straight) is the planform
configuration of the fluvial system. Each pattern is characterized by the degree of sinuosity, the
length of the centerline of the channel (in single-channel rivers), or the widest channel (in multi-
channel rivers), divided by the overall length of the reach (Friend & Sinha, 1993). Meandering
systems have the greatest sinuosity, while straight have the least. Figure 1.1 illustrates the effects
2
of changing gradient, sediment supply and calibre on channel form. Meandering channels, typical
in low-relief regions like prairies, are primarily characterized by a finer sediment calibre (grain
size), lower gradient, and lower sediment supply, relative to braided systems. If a meandering
system undergoes changes that result in an increased sediment supply or calibre (i.e. hillslope
landslides, faulting), the system must accommodate this change. In a simplified context, this
meandering system must construct mid-channel bars to store this excess sediment. The amount of
sediment exceeds the competence of the river. Alternatively, if a braided system undergoes a
reduction in discharge and sediment supply (i.e. damming), it becomes starved. This system
reduces the number of channels or braids it needs to transport discharge and sediment, transitioning
to a meandering system. Climatic instabilities have amplified changes in these parameters,
influencing the state of fluvial systems. This will be further discussed in section 1.3 Climate
Change Impacts on Proglacial Fluvial Systems.
3
Figure 1.1 Channel form is driven by sediment size and supply, and channel gradient and stability. From
Church (2006)
4
1.1.2 Discharge and sediment flux
Alluvial systems in alpine settings are highly vulnerable to fluctuations in discharge and
sediment. For example, an increase in discharge can occur over several years due to increased
glacier/snow melt or immediately from precipitation, while a decrease can occur due to damming,
lake formation, decreased glacier melt or decreased precipitation. Leopold and Maddock (1953)
describe the relationship of hydraulic geometry, which demonstrates that as fluvial systems change
channel form, for example, due to increases in flow velocity, mean channel depth and mean
channel width, there is an increase in the channels ability to transport a larger sediment supply
and calibre with increased discharge (Fig. 1.2). Discharge (Q) (eq. 1 4), is a function of channel
width (w), depth (d), and flow velocity (v). Where a, b, c, f, m, k are numerical constants: b, f, m,
represent the slope of the three lines in Figure 1.2, and a, c, k, represent the intercepts.
1 𝑤 𝑎𝑄= Channel Width
2 𝑑 𝑐𝑄= Channel Depth
3 𝑣 𝑘𝑄= Flow Velocity
4 𝑄 𝑎𝑄 𝑥 𝑐𝑄 𝑥 𝑘𝑄 𝐷𝑖𝑠𝑐ℎ𝑎𝑟𝑔𝑒
5
Figure 1.2 Flow velocity mean channel depth and mean channel width, all have a positive relationship with
discharge. From Leopold and Maddock (1953)
Alluvial channel geometry is the direct consequence of the sediments being transported
and deposited. The ability of the flow to mobilize sediment of a given size is expressed as stream
competence. This is quantified by Shields number (eq. 5). Shields number is a dimensionless
measurement of shear stress exerted by streamflow on the bed. 𝜌 is the sediment density, 𝜌 is the
fluid density, g is the acceleration of gravity, d is flow depth, D is grain size, and S is the slope.
(5) 𝜏 ∗ 𝜌𝑔𝑑𝑆/𝑔𝜌 𝜌𝐷
Sediment calibre (grain size) (D), plays a major role in channel morphology and is
inversely related to the sediment flux (Qs) and stream power (product of discharge and channel
gradient, S) (eq. 6). Essentially, the capacity of a stream to transport a sediment load and the
6
competence to move a particular grain size is dependent on the material available, bank strength,
and vegetation (Church, 2002).
6 𝑄𝑠 ~ 𝑄𝑆/𝐷
Sediment transport can be broken into several categories, - bedload, suspended load, and
wash material. The bedload consists of coarser material that is supported by the bed. This material
undergoes traction transport by rolling or bouncing along the bed. Suspended load involves fine-
grained material, such as silt, that is entrained in the flow due to turbulence, located in the middle
region of the flow. Wash material, the upper region near the surface includes any fines and
dissolved solids entrained in the flow and carried for a long distance. Saltation is the process by
which bed material projected from the bed via turbulent flow/eddies, can also contribute to the
suspended load (Church, 2006). Suspended load within rivers consists of mostly clay and sand-
size sediment. As discharge increases, suspended load increases, while the concentration of
dissolved solids decreases (Fig. 1.3 (Ballantyne & Murton, 2018). While this is subject to available
sediment, most alpine valleys are not sediment supply limited. A common source of sediment in
braided systems is unconsolidated glaciogenic deposits liberated by glacier retreat (Ballantyne &
Murton, 2018).
7
Figure 1.3 Suspended sediment (dashed line) and discharge have a positive relationship, while total
dissolved solids (solid line) have the opposite relationship. From Ballantyne and Murton (2018)
1.1.3 Channel bed degradation and aggradation
Fluvial aggradation and degradation reflect the sediment supply/budget within a fluvial
system (Törnqvist, 2013). Changes in the base-level of the system are driven by several factors,
such as vertical crustal movements or changes in discharge (Goudie, 2004). Long term changes in
discharge could be due to the formation of upstream reservoirs, such as dams or lakes. These
reservoirs act as sediment traps, reducing the supply of sediment to the river, potentially resulting
in long term impacts on the system (Kondolf et al., 2014). For example, Ahmed and Abdelbary
(2004) found that the Nile Rivers Aswan High Dam resulted in downstream bed degradation as a
result of sediment starvation. The impounded Lake Nasser has an aggregate storage capacity twice
as large as the Nile River, which changes both discharge and sediment supply to the system.
Annual discharge has reduced to less than a third of its pre-dam value, reducing the channels
competence and increasing scouring. Degradation, or scouring, can lead to incision of the channel
8
bed to depths five times the mean depth (Best & Ashworth, 1997). Similar to the impacts of
damming, proglacial lakes interrupt the delivery of meltwater and sediment to proglacial fluvial
systems (Carrivick & Tweed, 2013). Channel degradation and scouring in fluvial systems could
occur due to the formation of proglacial lakes, as discharge and sediment supply is altered. In
particular, proglacial lakes trap the coarser sediment load, but much of the finer-grained suspended
load is still transported downstream. And so, while proglacial braid plains are typically
characterized by abundant sediment with an extremely wide range of grain sizes, coarse material
is withheld from the system as proglacial lakes form and expand. The downstream impacts of this
shift, however, are unknown.
Channel aggradation requires the introduction or redistribution of sediment, potentially
from landslides or drainage reorganization (Petts, 1979). As sediment is redistributed and
deposited, channel beds aggrade, building high stage bedforms (i.e. gravel bars, dunes) (Bristow
& Best, 1993). Aggrading systems involve localized deposition that builds alluvial ridges and
terraces, opposed to downcutting and removal of material. As deposition progresses, lateral
changes involving increased braiding occur. This results in the widening of braid plains within
alluvial valleys (Slingerland & Smith, 2004).
1.1.4 Braided systems
Braided systems typically occur in alluvial valleys with steep and narrow channels, high
erodibility, and sediment supply. Exceptions can involve braided systems that flow from high
topography regions into foreland regions, where channels become more gradual and open. This
environment contains a relatively wide braid plain to accommodate the array of braid channels that
transport the abundance of water and sediment. The active braid plain of the Äy Ch system in
southwest Yukon is seen in Fig. 1.4. The array of channels extends laterally across the plain to
9
efficiently transport meltwater from Kaskawulsh Glacier to Kluane Lake. Braid bars form when
the capacity of flow to transport sediment is lower than the volume of bedload being transported
(Mackin, 1948). Channels undergo zones of bifurcation and convergence, where zones of
diverging and converging flow occur (Carson, 1984). Braided fluvial systems can occur in regions
of both homogeneous and heterogeneous sediment calibre. Gravel-bed braided systems are less
responsive to changes in discharge than sand-bed systems (Bristow & Best, 1993).
Understanding the evolution of braided systems is important for both geological and
engineering purposes. These systems have high rates of sedimentation, channel migration, and
bank erosion, emphasizing awareness of the many braided systems that contain proximal
infrastructure (i.e. roads, bridges, communities). Geologically speaking, these systems deposit
such considerable amounts of sediment, they produce sedimentary sequences ideal for aquifers,
hydrocarbon reservoirs, and accumulation of mineral deposits (Bristow & Best, 1993). It is very
beneficial to observe current landforms and behaviour in todays climate.
Figure 1.4. Active braid plain of the Äy Ch system in SW Yukon in 2019. Photo by author.
10
To evaluate the planform geometry of braided systems, braid indices quantify the level of
braiding at a certain time. Braiding indices measure the braiding intensity of a river reach based
on three main characteristics: bar dimensions; the number of channels in the network; and the total
channel length in the given reach (Egozi & Ashmore, 2008). There are three common indices, the
Bar Index, the Channel Count Index, and the Sinuosity Index (Fig. 1.5). The Bar Index and
Sinuosity Index are less common in practice, where the Bar Index is defined as the sum of twice
the length of all the bars in a reach, divided by the total reach length, and the Sinuosity Index is
the total length of the channels per river reach divided by the total reach length (Brice, 1964; Hong
& Davies, 1979). The Channel Count Index is the number of channels that intersect a series of
cross-sections throughout the river, divided by the total number of cross-sections used (Howard et
al., 1970). More detail of the Channel Count Index and the methods used in this thesis are found
in Chapter 3.
Figure 1.5 Overview of different Braiding Indices. The Channel Count Index is used in this thesis. From
Egozi and Ashmore (Egozi & Ashmore, 2008)
11
1.3 Climate change impact on proglacial systems
Climate change has caused fluctuations in snow and glaciers that have changed the amount
and seasonality of runoff in snow-dominated and glacier-fed rivers with impacts on water
resources (IPCC 2.3.1.1 (2019)) and proglacial regions because meltwater from snow and ice often
dominate runoff response (Kormann et al., 2015). These hydroclimatic changes impact fluvial
systems inducing sedimentary and geomorphic responses (East & Sankey, 2020). Global warming
has catalyzed meteorologic changes, such as shifting precipitation, reducing ice and snowpack,
vegetation disturbances, and droughts (Fig. 1.7). All together these instigate increased erosion,
forest fires, thermal rock stress, and fluvial transport capacity. East and Sankey (2020) explore
potential geomorphic responses, such as changes in slope stability and sediment yield, fluvial
morphology, and aeolian sediment mobilization. Anticipated responses include changes in slope
stability due to increased droughts or more extreme rainfall, increased sediment yields due to
increased erosion from increased precipitation, increased aeolian sediment mobilization due to
increased droughts and vegetation mortality, and lastly changes in fluvial morphology. It is
suggested that fluvial geomorphic responses are potentially due to changes in precipitation and
erosion, and increased frequency of fires and transport capacity. The direct effects of glacier retreat
are not explored in this diagram.
12
Figure 1.6 A diagram showing hydroclimatic effects of climate change and the potential sedimentary and
geomorphic responses. From East and Sankey (2020)
In this thesis, I explore changes in fluvial geomorphology due to changes in discharge and
sediment supply in proglacial rivers due to glacier retreat. Shifting precipitation and increased
global temperatures have increased glacier retreat, increasing runoff, drainage reorganization, and
the number of proglacial lakes (Adler et al., 2019; East & Sankey, 2020; Shugar et al., 2017). It
is presumed that hydroclimatic changes, such as increased discharge, will increase fluvial transport
capacity, the maximum amount of material that can be carried (Holden, 2005), and sediment
supply. But these assumptions may not be ubiquitous in glaciated regions.
1.3.1 Proglacial lakes
Current deglaciation is increasing the number and size of proglacial lakes around the world
and can be linked to climate, glacier dynamics, meltwater, and sediment fluxes (Carrivick &
13
Tweed, 2013; Otto, 2019; Shugar et al., 2020; Tweed & Carrivick, 2015). Fluvial transport
capacity and sediment supply may change as glacier retreat continues.
There are three common types of glacial lakes those dammed by bedrock, moraine, and
ice. Moraine dammed lakes are confined in the proglacial region and are the main interest in this
thesis. These lakes reside in the depression between the terminus of the glacier and a moraine.
Proglacial lakes significantly reduce the transfer of sediment from the glacier to the fluvial system
(Otto, 2019). Sedimentation and bathymetry vary within proglacial lakes and are dependent on
several factors, including flow velocity, turbidity, density, and thermal stratification (Ashley,
2002). Flow velocity and bathymetry entering the lake dictates the trajectory or path of sediment
deposition. Fine sediment is transported in the upper regions of the lake unless redirected
downwards due to turbidity flows. Overall, coarse sediment is deposited in the lake due to reduced
flow velocity, while fine silt and clay remain in suspension until the flow velocity drops below
<0.1 m/s (Ashley, 2002; Bennett & Glasser, 2011). The behaviour of sediment into and out of
proglacial lakes is similar to industrialized dammed reservoirs and likely elicit a similar fluvial
response. As glacier retreat continues, fluctuating discharge and sedimentation, proglacial fluvial
systems will be forced to alter their geometry to accommodate these changes.
1.4 Objectives
To better understand and predict environmental change, it is critical to quantify both the
spatial and temporal response of landscapes to climatically driven fluctuations. Glaciofluvial
landscapes, common in these high latitudes, are likely to respond to variations in streamflow, as
triggered by changes in air temperature and precipitation (Goudie, 2006). However, less is known
about exactly how these landscapes may react to rapid and quasi-permanent changes in river
14
discharge or the timescales over which such changes may occur. The combination of steep terrain,
abundant meltwater and erodible sediment, resulted in glaciated landscapes may be sensitive to
fluctuations in discharge, and hence adjust their morphology to changes in discharge, sediment
supply, and bed gradient as has been shown extensively before e.g. (Church, 2006; Fahnestock,
1964). Previous research (Chew & Ashmore, 2001; Fahnestock, 1964; S. N. Lane et al., 1996)
provides a foundation for understanding glaciofluvial behaviour, suggesting that proglacial stream
morphology changes rapidly in response to changes in glacial melt and upstream sediment supply.
However, what remains unclear is the longevity and magnitude of these responses, and empirical
data are required to constrain fluvial geomorphic models and improve the prediction of alluvial
channel patterns in response to climatically-driven change (Anisimov et al., 2008). This thesis
describes and quantifies the planform geomorphic response of subarctic rivers and alluvial fans to
changes in their discharge and sediment supply.
The goal of this thesis is to evaluate the behaviour of proglacial fluvial systems as
deglaciation progresses due to climate change. Through the use of satellite remote sensing and
historical aerial imagery, alluvial river systems can be examined at a high spatial and temporal
resolution to observe their geometry through time. The first objective is to identify how a fluvial
system responds to a nearly instantaneous and permanent alteration in discharge by quantifying
the braiding complexity and as well as the areal change of alluvial fans. The second objective is to
characterize multi-decadal planform changes due to changes in sediment supply.
1.5 Thesis outline
This thesis is organized as follows. Chapter 2 introduces the study areas explored in this
research, and Chapter 3 explains the methodologies used to identify geomorphic changes, and data
15
collection and processing. Chapter 4 describes the results found in this research, and Chapter 5
discusses and analyzes these results. Chapter 6 provides a summary and describes potential future
work. Appendices 1 and 2 provide more detail of the data collection and processing of satellite
imagery and air photos.
1.6 Statement of contribution
This project began based on the idea that the retreat of Kaskawulsh Glacier and the
diversion of flow altered the braiding complexity of two adjacent subarctic rivers. This idea was
proposed by Dan Shugar and is explored in the first component of this thesis. The second
component originated from the idea of exploring other rivers systems potentially influenced by
glacier retreat. This was proposed by both Dan Shugar and Jim Best. Each of these components is
reviewed in chapter 2. Gryphen Goss collected data (satellite imagery and historical air photos),
processed imagery, and analyzed braiding, alluvial fan area, and proglacial lake area, all of which
can be found in the chapter in 3 and 4. Interpretation and contextualization of results were initially
completed by Gryphen Goss and further defined by Dan Shugar and Jim Best, found in chapter 5.
Both Dan Shugar and Jim Best conducted commenting and suggestions for the manuscript, while
editing was done by Gryphen Goss.
16
Chapter 2: Study areas
2.1 Kaskawulsh Glacier
Kaskawulsh Glacier terminates near the eastern edge of Kluane National Park and Reserve
in the St. Elias Mountains of southwest Yukon (Figure. 2.1). Historically, Äy Ch (ay-CHEW,
formerly known as Slims River) has drained Kaskawulsh Glacier meltwater north into Kluane
Lake and eventually to the Yukon River, while Kaskawulsh River drained meltwater east into the
Dezadeash and Alsek rivers towards the Gulf of Alaska (Bostock, 1969; Rampton, 1981).
Interestingly, the Holocene hydrological record of the Kluane Lake region is more complex
(Clague et al., 2006). In the early Holocene, Kaskawulsh Glacier terminated much farther up the
valley, allowing Kluane Lake to drain to the south, down what is now the Äy Chù valley, and
into Kaskawulsh River (Bostock, 1969). During the Little Ice Age, Kaskawulsh Glacier advanced
to its farthest position down the valley, forming its end moraine while alluvial fans were
constructed within each river valley (Clague et al., 2006). An increase in sediment load to both
rivers impeded southerly flow from Kluane Lake, thereby increasing lake levels, reversing the
flow of the Äy Ch (Bostock, 1969), and forcing the lake to spill over into the Duke River (to
the north). Over the past few decades, two proglacial lakes at the terminus of Kaskawulsh Glacier
have provided the source of meltwater for each river (Shugar et al., 2019), with the proglacial lake
feeding Äy Ch positioned at a higher elevation than that feeding Kaskawulsh River. In May
2016, river capture was induced by the retreat of Kaskawulsh Glacier (Shugar et al., 2017), and
most meltwater from Kaskawulsh Glacier was abruptly redirected to the Kaskawulsh River.
17
Figure 2.1 Map of the study area in southwest Yukon. Kaskawulsh Glacier terminates at the drainage divide of Äy
Ch and Kaskawulsh River. White arrows indicate flow directions. Alluvial fans in Äy Ch (white circles) and
Kaskawulsh (white squares) rivers, described in the text, are labelled. White boxes indicate locations shown in Figure
2.2. Source: Sentinel-2 (07-16-2017)
This river capture event resulted in a dramatic and substantial decrease in discharge to the
Äy Ch, but a concomitant rapid increase in flow within the Kaskawulsh River (Figure. 2.2) that
continues to the present. As a result, the changes at Äy Ch and Kaskawulsh River can be thought
of as representing the future state of many proglacial rivers pre- and post-peak water, respectively,
but occurring over years instead of decades.
4 km
18
Figure 2.2 Segments of the Äy Ch before and after river capture (top row), and Kaskawulsh River before and after
river capture (bottom row). Images of each river at peak melt and changes in discharge are attributable to the river
capture event. Sources: Rapideye, SPOT-5, PlanetScope
2.2 Glaciers of western North America
To further explore the fluvial response to deglaciation over longer time scales, an additional
nine sites were selected throughout western North America (Fig. 2.3). Each site was chosen based
on glacier type, a typical valley glacier that terminates in an alluvial valley with a proglacial river
system and each site is located in North America due to the accessibility of data. Of the 11 sites,
nine saw the birth and growth of a proglacial lake over the study period, while the remaining two
did not (they are considered control sites). Table 1 identifies each glacier, region, and
corresponding Randolph Glacier Inventory (RGI) ID.
Ç
Ç
Ä’äy Chù
Kaskawulsh
19
Figure 2.3 Map showing locations of the study sites. Names indicated in the legend. *Control sites
2.2.1 Southwest Alaska
Sites 1 and 2 are located in southwest Alaska on the southern coast of the Aleutian Islands.
Kukak Bay (unnamed) is site 1 (Fig. 2.4) and the first control site, as it never contains a proglacial
lake. This site has undergone several kilometers of retreat in the past 70 years, evident in historical
photos, nearly doubling the length of its proglacial river. Hallo Glacier, site 2 (Fig. 2.4), contained
a relatively small proglacial lake in 1951 compared to in 2020.
8
20
Figure 2.4 Sites 1 and 2, Kukak Bay and Hallo Glacier, respectively. Located in southwest Alaska on the coast of the
Aleutian Islands. Site 1 is a control site.
Sites 3 - 5 are located in the arid interior of Alaska in the Central Alaska Range. Both East
Fork Susitna Glacier, site 3 and NW Robertson, site 4 (Fig. 2.5), have undergone nearly a kilometer
of retreat in 60-70 years, producing proglacial lakes. Robertson Glacier, site 5 (Fig. 2.5), is the
second control site and resides in a valley adjacent to site 4. This site has undergone comparatively
less retreat in the past 42 years.
21
Figure 2.5 Sites 3 5, located in the central Alaska Range. East Fork Susitna Glacier and NW Robertson, sites 3 and
4, produce a proglacial lake over the past 60-70 years, while Robertson, site 5, does not.
22
2.2.2 Yukon
Sites 6 and 7 (Fig. 2.6), previously described, include two fluvial systems at Kaskawulsh
Glacier in southwest Yukon. These sites, Äy Ch and Kaskawulsh River are evaluated in 2015,
opposed to 2020, due to the river capture event that occurred in 2016. During the past 65 years,
two proglacial lakes reside at the head of each river.
Figure 2.6 Äy Ch and Kaskawulsh River, sites 6 and 7, respectively, reside in southwest Yukon
2.2.2 British Columbia and Southeast Alaska
Sites 8 and 9 are located on the Juneau Icefield, one in southeast Alaska and the other in
northwest British Columbia. Meade Glacier, site 8 (Fig. 2.7), is a tributary of the Juneau Icefield
in the NE corner. Tulsequah Glacier, site 9 (Fig. 2.7), is also a tributary of the Juneau Icefield,
23
flowing into the Taku River. Tulsequah Glacier has an interesting history of periodic glacial
outburst floods occurring up to 3 times a year. With glacier retreat and thinning, outburst floods
have become progressively smaller (Geertsema & Clague, 2005).
Figure 2.7 Meade and Tulsequah Glacier, sites 8 and 9, are both tributaries of the Juneau Icefield
Lastly, sites 10 and 11 are located in southwest-central British Columbia. Klinaklini
Glacier, site 10 and Lillooet Glacier, site 11 (Fig. 2.8), are the lowest latitude study sites, also
residing near the coast.
24
Figure 2.8 Klinaklini, site 10, and Lillooet, site 11, are in southwest British Columbia, the lowest latitude sites. Lillooet
contained a small lake in 1951
Nine of the selected sites produced a proglacial lake over the period investigated. For
control, two sites are included that did not produce a proglacial lake during the past 70 years. Site
1 is the first control site, located in the region of southwest Alaska and site 5, located in the region
of the Central Alaska range. Spatial details of each site can be found in table 2.1.
25
Table 2.1 Each study site, region, and Randolph Glacier Inventory RGI ID. http://www.glims.org/maps/gtng
Glacier
Region - Coordinates
RGI ID
Hallo
SW Alaska -
58°23'29.68"N
154°10'15.35"W
RGI60-01.20277
East Fork Susitna
Central Alaska -
63°23'48.08"N
146°51'21.14"W
RGI60-01.00022
NW of Robertson
(unnamed)
Central Alaska -
63°20'16.81"N
144°38'38.54"W
RGI60-01.00576
Robertson
Central Alaska -
63°17'5.46"N
144°23'44.20"W
RGI60-01.00578
Kaskawulsh -
Äy Ch
SW Yukon -
60°50'51.12"N
138°37'8.37"W
RGI60-01.16201
Kaskawulsh -
Kaskawulsh River
SW Yukon -
60°49'7.98"N
138°31'24.86"W
RGI60-01.16201
Meade
SE Alaska -
59°14'31.31"N
135° 4'40.97"W
RGI60-01.01524
Tulsequah
NW British Columbia -
58°48'9.00"N
133°41'15.59"W
RGI60-01.01521
Klinaklini
SW British Columbia -
51°17'52.11"N
125°46'3.02"W
RGI60-02.05157
Lillooet
SW British Columbia -
50°43'34.02"N
123°40'38.22"W
RGI60-02.02386
Kukak Bay
(unnamed)
SW Alaska -
58°19'30.54"N
154°30'44.18"W
RGI60-01.20319
26
Chapter 3: Methods
3.1 Data collection and processing
3.1.1 Satellite imagery for inter-annual mapping
Satellite imagery was selected for interannual mapping of Äy Ch and Kaskawulsh
River. The imagery was collected from SPOT-5, PlanetScope, Worldview-2, Rapideye, and
Sentinel-2 and processed in ArcGIS Pro. This timeline included data from 2013 - 2020, May to
September, and depending on available coverage, data was selected about every two weeks. To
maintain consistency, all satellite imagery was resampled to 10 m resolution because that was the
lowest resolution accessible. To improve the visibility of the water within the channels, the
symbology was adjusted, by modifying the band combination, stretch and statistics. During
analysis, an RGB band combination was used (imagery shown in black and white in figures 2.2 to
stay consistent with historical air photos). Stretch was consistent for every satellite using Percent
Clip and Dynamic Range Adjustment adjusted the statistics accordingly. Table A1 (Appendix
A), identifies each satellite and corresponding resolution, symbology, and dates used.
3.1.2 Satellite imagery for multi-decadal mapping
Satellite imagery was also selected for multidecadal mapping for all 11 study sites. The
imagery was collected from PlanetScope for 2020, except for Äy Ch and Kaskawulsh River,
which involved imagery from 2015 (pre-river capture). This satellite imagery was processed as
above, although the resolution remained at the native 3 m since only PlanetScope imagery was
27
being used. Table A2 (Appendix A) identifies each location, dates supplied, data source, and
symbology.
3.1.3 Historical air photos
Historical air photos from the U.S. Geological Survey and Government of Canada were
selected for each of the 11 sites between 1949 and 1979 (the oldest date available). Air photos
were processed in Agisoft Metashape to create an orthorectified image of each site (Agisoft
Metashape Professional (Software), n.d.). Raw images were imported into Metashape and aligned
to produce a sparse cloud. Once this sparse cloud was created, GCPs (ground control points) were
placed to georeference the image. GCPs were extracted from ArcGIS Pro using high-resolution
base map imagery and from Arctic DEM. Once the air photos were georeferenced and
appropriately aligned, dense clouds were generated. Once the DEM was created in Metashape,
orthoimagery were generated. The orthophotos were then exported as rasters and imported into
ArcGIS Pro for analysis, as described in section 3.2. Figure 3.1 outlines the workflow for
processing both satellite imagery and historical air photos. Once all the imagery and air photos
were imported into ArcGIS and prepared for the next stage of preprocessing, mapping of the
braiding intensity, alluvial fan, and lake area could commence (ArcGIS Pro, Esri Inc., n.d.).
28
Figure 3.1. This workflow outlines the process of collecting and preparing satellite imagery and historical air photos
for processing in ArcGIS Pro. AWW is the Average Wetted Width. (Agisoft Metashape Professional (Software), n.d.;
ArcGIS Pro, Esri Inc., n.d.)
29
3.2 Quantifying braiding intensity
The braiding intensity of each river was quantified using the dimensionless Channel Count
Index (CCI; (Egozi & Ashmore, 2008)). The CCI is a widely used braid index that involves cross-
sectional characterization and measurements of wetted channels along a river and is given by:
7 𝐶𝐶𝐼 ∑
where 𝑁𝑤𝑐 is the number of all wetted channels and 𝑁𝑥𝑠 is the number of cross-sections. The
process for determining CCI begins with defining a series of cross-sections and the length of the
river reach. The length of the river reach, defined by Egozi and Ashmore (2008), should be at least
ten times the average wetted width (AWW). Average wetted width is the mean width of the active
braid plain, measured through a series of cross-sectional widths downriver. This was measured
using the first timestamp for each site. Once the river reach is defined, CCI cross-section spacing
can be determined. Cross-sections must be of equal separation from the head of the river to the
end of the reach, and being no farther apart than the AWW (Egozi & Ashmore, 2008). Cross-
section separation was chosen by identifying an easily multipliable value as close to the AWW,
without exceeding it. River reach was then divided by the separation value to identify the number
of cross-section implemented (adding one to include the zero mark). Table 4 identifies all the
parameters for each river, including AWW, river reach length, number of cross-sections and the
distance between them.
30
Table 3.1. River dimensions and number of cross-sections used
Site
River Reach
(m)
AWW (m)
(At time 1)
XS Separation
(m)
Number of
XS
Kukak Bay
~2,800
361
350
9
Hallo
~8,800
412
400
22
East Fork
Susitna
~18,000
385
375
49
NW of
Robertson
~21,000
723
700
31
Robertson
~19,200
622
600
33
Ä’äy Chù
~21,750
1550
750
29
Kaskawulsh
~22,500
1600
750
31
Meade
~19,000
880
850
23
Tulsequah
~21,000
1303
1000
22
Klinaklini
~27,000
1283
1000
28
Lillooet
~16,100
368
350
47
The braiding intensity was quantified by digitizing the wetted channels along each cross-
section for every satellite image and air photo. To further quantify how these alluvial valleys
responded to changes, the channel width of each wetted channel throughout the river reach was
measured. This identifies the sum of wetted channels along each cross-section downstream.
The area of 15 alluvial fans within the Äy Ch and Kaskawulsh river valleys, from 2013-
2020, were measured and normalized by their 2015 (pre-capture) extents. For this analysis, satellite
images were used from as close to mid-July as possible, to minimize any potential differences
caused by seasonal differences in flow in these ungauged basins. For example, Figure 3.2,
identifies
Figure 3.2 Flow stage is lower in the earlier period of the summer, compared to peak flows later in the summer.
Low flow
High flow
31
a lower flow stage, earlier in the summer, than at peak flow. Dates were chosen near the same date
for consistency and at a higher flow stage for more define boundaries.
Lastly, the area of proglacial lakes for the glaciers was measured. The area of both the fans
and proglacial lakes was identified by digitizing the outer boundary and quantifying the area in
ArcGIS.
3.3 Discharge data from proximal stream gauges
Only two of 11 sites contained a stream gauge located on or near the river, Klinaklini and
Tulsequah, collected from Environment Canada and the USGS. The stream gauge located on the
Klinaklini River is positioned downstream near the mouth of the river (Fig. 3.3 top). The Tulsequah
stream gauge is position on the Taku River, downstream of the confluence of the Taku and
Tulsequah rivers (Fig. 3.3 bottom). At both sites, discharge does not appear to increase or decrease
through time as proglacial lakes continue to increase in size. Unfortunately, the data lack at the
time when the lakes first formed, however the discharge does not appear to change as upstream
reservoirs grow.
32
Figure 3.3 Discharge of Klinaklini River (top) from 1984 to 2016. Discharge of Tulsequah River via the Taku River
(bottom) from 1992 to 2016. This stream gauge is located downstream of the confluence of the Taku River and the
Tulsequah river
33
3.4 Minimizing error and discounting external factors
To minimize human error, digitizing was kept within approximately three pixels of the
feature of interest (or as close to the defined line as possible) and was consistent during all data
collection (Fig. 3.4). 10-meter and 3-meter resolution can be compared by evaluating
measurements of both Äy Ch and Kaskawulsh rivers since they were analyzed for both
components of this thesis at different resolutions. At 10-meter resolution, the braiding intensity for
both rivers was collectively lower than when measured at 3-meter resolution since braiding
intensity is a relative dimensionless measurement. This variation is to be expected, since, at a high
resolution (3-meters), more features (bars) are visible, producing a more complex system and
higher braiding intensity. The resolution was chosen based on the availability of data and for
consistency.
Figure 3.4 Wetted channels were digitized within 3 pixels of feature boundary
To discount external factors that could affect channel form, such as meteorological factors,
data from surrounding weather stations was collected. This data was sparse due to the lack of
34
stations in proximal locations. Two weather stations were in the range of the Kaskawulsh Glacier
region, Kluane Lake Research Station (KLRS), provided by Gwenn Flowers at Simon Fraser
University, and Haines Junction, provided by Environment Canada. Figure 3.5 shows the
precipitation for KLRS (left) and Haines Junction (right), as well as air temperature. There does
not appear to be any major fluctuations in either parameter that may drive a change in channel
form other than discharge due to river capture. This is, however, difficult to justify given that these
stations are ~ 30 km from these river systems and the variable weather that occurs in glacial
valleys. A meteorological effect would induce a temporary and immediate response in channel
form, opposed to a permanent change. Later on, in Chapter 5, this concept is further discussed.
Figure 3.5 Precipitation and temperature from Kluane Lake Research Station (left) and Haines Junction (right) weather
stations. Provided by Gwenn Flowers at Simon Fraser University and Environment Canada, respectively. Dates
missing data can be found in table A3.
35
Chapter 4: Results
The following sections describe the results produced from analyzing the inter-annual and
multi-decadal fluvial response to deglaciation. The first analysis of Äy Ch and Kaskawulsh
fluvial response is evaluated through braiding intensity, total wetted width, and alluvial fan area
from 2013 2020. Second, 11 study sites undergo analysis through braiding intensity, total wetted
width, and lake area from ~ 1950s to 2020.
4.1 Fluvial response on the Äy Ch and Kaskawulsh River from 2013-
2020
Collectively, braiding intensity, total wetted width, and the areal extent of alluvial fans
provide an overview of the fluvial response to changes in discharge due to river capture. Here,
changes in braiding intensity are measured using the CCI to identify spatial and temporal changes
in channel form. Total wetted width defines the sum of active channel width along each cross -
section and lastly, areal extent defines the percent of erosion or growth occurring on channel
margins. Below are the results for the fluvial response on the Äy Ch and Kaskawulsh River.
4.1.1 Braiding intensity
During the melt seasons of 2013 - 2015, Äy Ch was characterized by a median (annual)
CCI of 3.5 - 4.6 (range 2.5 to 5.4) (Fig. 4.1) but in 2016, after the river capture, this decreased to
a median of 3.1. From 2017-2020, the CCI of Äy Ch remained low, with median values between
2.3 and 3.1 (range 1.0 to 3.5). From 2013 - 2015, Kaskawulsh River was characterized by a median
CCI of 2.5 - 2.8 and a narrow range (1.9 to 3.2), but following river capture, CCI nearly doubled,
with a median of 4.8, and remained high thereafter, with median values between 4.3 and 4.7 from
2017-2020 (range 2.0 to 5.8).
36
Figure 4.1 Äy Ch (A) and Kaskawulsh River (B) braiding intensity from 2013 to 2020. The red dashed line is the
time of river capture
B
B
Years
37
Seasonal variation in braiding intensity is seen in Figure 4.2. From May to June, before
river capture, Äy Ch was characterized by a median CCI of 4.1 - 4.9 (range 2.5 to 5.2) (Fig.
4.2) but from July to August, the median was 3.1 5.4 (range 2.7 5). After river capture, the
range from May to June, remained relatively consistent, with medians ranging from 2.1 3.0. The
single outlier is the last timestamp before the switch occurred in May, with a CCI of 4.1. Before
river capture, May June, Kaskawulsh River was characterized by a median CCI of 2.2 - 2.6 and
a range (1.8 to 3.2), and in July and August, median CCI increased less than 1. After river capture,
in May and June, the median increased 3.2 4.5 (ranges 1.9 5.5). In July and August, medians
remained relatively similar.
Figure 4.2 Seasonal variation in braiding intensity from May-June, and July-August, from 2013- 2020. The Red
dashed line is the time of river capture
38
4.1.2 Downstream wetted channel width
Total wetted width (sum of wetted channels) was measured along each cross-section (Fig.
4.3). Between 2013- 2015, Äy Ch filled the entire braid plain and the total wetted width (TWW)
ranged from 1000 - 1375 m, with a peak around 10km downriver. From 2016-2020, TWW showed
a dramatic reduction to a range of 375 - 700 m, with a similar peak at the 10km mark. Throughout
the Äy Ch river reach, in 2013 - 2015, TWW was much more variable than in 2016 - 2020.
Between 2013-15, Kaskawulsh River was characterized by a TWW in the upper 12 km of between
10-100 m, with downstream reaches displaying both a greater TWW from 10 to 300 m and TCWW
variability. From 2016-2020, in the upper 12km, the range in TWW increased to 50 to 575 m while
the downstream reaches increased to ~ 500 m.
Figure 4.3 Braiding intensity and total channel wetted width for Äy Ch (A) and Kaskawulsh River (B)
B
B
c
B
A
39
4.1.3 Areal change of alluvial fans
Alluvial fans in the Äy Ch valley remained within +/- 2% of their 2015 area between
2013 and 2014 (Fig. 4.4). Between 2016 and 2020, they experienced a more variable change in the
area, ranging from +6% to -9%, but without any consistent pattern in space or time as seen in
Figure 4.5. However, alluvial fans in the Kaskawulsh River valley have all undergone erosion
since the 2016 river capture. Between 2013 2014, alluvial fans in Kaskawulsh River valley
similarly remained within 2% of their 2015 area. From 2016 onwards, there was an overall
reduction in the area of individual fans from -1% to -10% relative to their 2015 values, though one
fan increased in area slightly by 1 % in 2016. Kaskawulsh fan reduction can be seen in Figure 4.5,
as discharge increased, erosion of channel margins increased.
Figure 4.4 Äy Ch alluvial fans (A) underwent minor changes during the river capture event. Kaskawulsh fans (B)
underwent increased erosion
B
c
A
B
40
Figure 4.5 Äy Ch fans (top) underwent minimal changes as discharge decrease, while Kaskawulsh (bottom)
underwent increased erosion as discharge increased
41
4.2 Multi-decadal changes in proglacial lake extent and river geometry
The following sections review the responses of 11 proglacial systems to glacier retreat over
the past 50 to 70 years. Over time, nine of these sites either formed or enlarged a proglacial lake
that drained into a proglacial river, whereas the remaining two control sites did not form a lake
during this period.
4.2.1 Braiding complexity and proglacial lake development
All of the nine sites that formed proglacial lakes underwent a reduction in braiding intensity
as the proglacial lake area increased (Fig. 4.6 and Table 4.1). Figure 4.6 also provides imagery of
three sites where lakes formed, as channel width and number of channels reduced. Two sites that
had a lake at T1 followed the same trend as sites without a lake at T1. Two of 11 sites experienced
slight increases in braiding intensity without forming proglacial lakes, including Kukak Bay
Glacier and Robertson Glacier.
Table 4.1 Proglacial Lake area and CCI at time 1 and 2. Grey shaded regions are control sites
Site
Lake Area
(km2) (T1)
Lake Area
(km2) (T2)
CCI
(T1)
CCI
(T2)
Kukak Bay
0
0
3.7
3.8
Hallo
0.16
0.41
5.5
1.9
East Fork
Susitna
0
0.81
4.9
2.6
NW of
Robertson
0
0.16
6.5
3.5
Robertson
0
0
3.9
5.4
Ä’äy Chù
0
0.47
13.4
9.5
Kaskawulsh
0
0.28
5.0
3.6
Meade
0
0.49
11.2
6
Tulsequah
0
0.47
6.1
3.9
Klinaklini
0
0.80
6.4
5.3
Lillooet
0.02
0.35
3.5
2.3
42
Figure 4.6. Nine sites undergo reduced braiding intensity while the proglacial lake area increased. At two sites, the
braiding intensity increased without any formation of a proglacial lake. East Fork Susitna Glacier is seen in row one,
Meade Glacier in row two, and Klinaklini Glacier in row three. Each glacier, at time 1, terminates directly in the river,
where at time 2, glacier retreat has resulted in a proglacial lake. Each associated river, at time 1, resembles a relatively
highly braided river compared to time 2, where braiding has greatly reduced with the formation of a proglacial lake.
43
4.2.2 Total wetted width
In each river, total wetted channel width was measured along the entire study reach. Figure
4.7 and Table 4.2 display the distribution of channel widths before and after the formation of a
proglacial lake. Median channel width increased at every site, except at Klinaklini, where the
median braiding intensity remained the same. There was also a general reduction in the number of
smaller channel widths, shifting the distribution right to larger channel widths.
Table 4.2 Median Total Wetted Width at time 1 and 2
Site
Median
Width (m)
(T1)
Median
Width (m)
(T2)
Kukak Bay
11.1
13.4
Hallo
13.8
22.5
East Fork
Susitna
14.2
16.1
NW of
Robertson
12.2
17.2
Robertson
14.4
17.6
Ä’äy Chù
25.7
29.1
Kaskawulsh
12.8
20.1
Meade
17.5
24.4
Tulsequah
30.4
49.6
Klinaklini
35
35
Lillooet
13.9
19.3
To increase temporal resolution, the East Fork Susitna site was selected to evaluate a time
series of braiding intensity, the lake area, and total wetted width. This site was chosen due to the
availability and coverage of satellite imagery and air photos over time.
44
Figure 4.7 Histograms of channel width for study sites. At nine of the eleven sites, the frequency of smaller channels
reduced reflecting a reduction in braiding. At two of the sites where no lake formed, the frequency of channels was
only minorly effected
45
4.2.4 East Fork Susitna fluvial response to proglacial lake formation
A time series of proglacial changes at East Fork Susitna Glacier is shown in Figure 4.8.
The following section describes the braiding intensity, lake area, and downstream wetted width on
8- 8-1949, 24-7-1977, 19-7-1980, and 17-8-2020. Initially, there was no lake in the proglacial
region (Fig. 4.8, 1949). In 1977, the formation of a lake began, eventually growing to 0.80 km2 in
2020 (Fig. 4.8, 1977, 1980, 2020). During this time, braiding intensity decreased from 4.9 in 1949
to 2.2 in 2020. Total wetted width from 1949 to 1977 produced upstream widths slightly greater
than 250 m and downstream widths slightly greater than 100 m. In 2020, throughout the entire
river reach, widths did not exceed 100 m.
46
Figure 4.8 Time series of proglacial lake size and braiding intensity (inset) and total wetted width for East Fork
Susitna River. Photos of the proglacial zone through time (bottom).
47
Chapter 5: Discussion
Many proglacial rivers around the world are currently experiencing changes in discharge
and sediment supply as a result of glacier retreat (Baewert & Morche, 2014; Marren & Toomath,
2014). How those rivers are responding to those changes is not well known. This thesis discusses
two natural experiments that help address these questions. The first study evaluates the response
of two rivers that experienced very rapid and opposite shifts in discharge, whilst in the second
study, nine sites are investigated that have undergone formation of proglacial lakes, potentially
interrupting the delivery of sediment supply.
5.1 The effects of changing discharge
Although it has been assumed that climate change and increased glacier retreat and melt
has led to an increase in sediment and discharge in proglacial zones (Goudie, 2006), recent studies
have suggested this may not always be the case as other phenomenon are occurring (proglacial
lakes and re-routing) (Bogen et al., 2015; East & Sankey, 2020; Geilhausen et al., 2013). Proglacial
systems are typically not supply-limited but are usually continually fed by new sediment evacuated
from the glacier (Carrivick & Heckmann, 2017; Gurnell & Clark, 1987), as well as sediment stored
in floodplains, terraces, and alluvial fans. The present study has shown that in Kaskawulsh River,
braiding increased immediately with an abrupt increase in discharge (Fig. 4.1). In the upper reach,
discharge appears to be the main driver altering channel form, as seen in Figure 4.3 where the total
wetted width is larger than before the switch. This trend is consistent throughout the upper reach,
and for four years post river capture. In the lower reach, beyond the most southern fan, total wetted
width and braiding intensity maintains similar behaviour to before the switch. This may be due to
several factors including the proximity of alluvial fans that confine the river, increased erosion of
48
those fans, and also the presence of an active tributary channel (Fig. 4.4). Influx of sediment, from
erosion, can induce channel braid development, as the sediment load exceeds the competence of
the stream, resulting in deposition and bar formation (Smith & Smith, 1984), suggesting that a
secondary driver in changes in channel form could be sediment supply in the lower reach, while
the main driver is due to increased discharge as drainage rerouting occurs. The response seen in
Kaskawulsh is similar to that observed following increases in discharge due to dam
decommissioning, glacier outburst floods, or seasonally during monsoons or increased glacial
melt, but without a plume of sediment feeding the system (Bertoldi et al., 2010; Bishop, 1995; East
& Sankey, 2020; Surian & Rinaldi, 2003).
Alluvial fans within the Kaskawulsh River valley responded to increased discharge by
increased terminal erosion. This valley has incorporated more channels to allow the increased flow
to pass through the channel by widening the overall braid plain. This change in channel width is
again comparable to that of a dam decommissioning, but without the influence of sediment. For
example, in the middle and lower reach of the Elwah River, there was ~5 - 10% widening of the
channel after decommissioning of the Elwah Dam in 2013, due to an increase of the median
discharge from 50 to 200 m3/s (East et al., 2015). The response of Kaskawulsh River to an increase
in discharge is similar, however, it was not only immediate but has continued 4 years after the
switch and will likely remain so until meltwater flows reduce to pre-capture levels as Kaskawulsh
glacier retreats further (Clarke et al., 2015; Menounos et al., 2019; Moore et al., 2020).
In the Äy Ch valley, the opposite scenario was observed, where a dramatic reduction in
braiding intensity occurred immediately following the drop in discharge after flow was diverted
into the Kaskawulsh River. Reduced discharge has resulted in a transition from a wider and more
complex braid plain to a narrower and relatively steady channel form (Li et al., 2018). Before river
49
capture, flow within Äy Ch was high enough at peak flow to produce bank full conditions for
at least certain dates (Fig. 2.2). Äy Chs fluvial response to a decrease in discharge was
immediate, as seen in Figure 4.2, where the single outlier in May 2016 reflects the final timestamp
before the flow is rerouted to Kaskawulsh River that caused a drop in braiding intensity. This not
only identifies the immediacy of the response but also the driving factor of change, being reduced
discharge. As discharge decreases, a decrease in braiding intensity (Fig. 4.4) is seen as a reduction
in the number of smaller channels within the braid plain occurs, leaving larger primary channels
to transport flow and sediment (Egozi, 2006). Over time, this has the potential to greatly reduce
sediment flux throughout the Äy Ch system and into Kluane Lake (Clague, 2006). This
reduction in sediment load then has the potential to induce incision. Chew and Ashmore (2001)
observed incision in the upper reach of the Sunwapta River as sediment is trapped in an upstream
reservoir. This response, reduced discharge and sediment supply, is similar to that of a dammed
river (Best, 2019; Draut et al., 2011; Kondolf et al., 2014), also where the majority of flow and
sediment supply is restricted from the fluvial system. Äy Ch is now prone to the risk of
becoming transport limited, resulting in induced incision and slope reduction, habitat changes, and
ecosystem fragmentation (Best, 2019; Grill et al., 2015).
As a reduction in flow occurred in Äy Ch, the areal change in downstream alluvial fans
did not immediately change; rather both minimal growth and erosion occurred (Fig. 4.5). Fan
evolution, however, has been linked to fluctuations in discharge as a result of hydrological changes
(Bloom et al., 2020; Hansford & Plink-Björklund, 2020; Leier et al., 2005). This lack of change
may be the product of lower flows resulting in reduced erosion and insufficient fan growth (the
location of the fan relative to the flow of water), but also the digitizing error involving difficulty
consistently defining the edge of each fan.
50
For example, Vulcan Fan, an alluvial fan located at the end of the Äy Ch valley to the
south, underwent a major landslide in 2014 that blocked the river, resulting in the lack of sediment
to feed the alluvial fan (Brideau et al., 2019). Overall, the geomorphic response lacked in the Äy
Chù fans with very little change occurring, while a strong response was seen in Kaskawulsh fans
as each fan underwent increased erosion.
5.2 Upstream lakes and downstream changes
Proglacial lakes represent upstream reservoirs that result in a sediment sink and interrupt
flow to proglacial rivers (Otto, 2019). As seen in the lower reach of the Kaskawulsh River, influx
of sediment, from erosion, can induce channel braid development (Smith & Smith, 1984), while
to the contrary, a lessening of sediment can result in braid reduction. The data in the present study
show that the braiding intensity of nine proglacial rivers reduced as a function of reduced sediment
supply as proglacial lake area increased.
Many studies have explored how a fluvial system evolves due to changes in discharge and
sediment supply from climatic forcings, damming/dam removals, landslides, and tectonics (Best,
2019; East et al., 2015; East & Sankey, 2020; Holbrook & Schumm, 1999). Less research has
encompassed the role that the formation of proglacial lakes plays on braided rivers. Chew and
Ashmore (2001), however, evaluated fluvial response to the formation of Sunwapta Lake on the
Columbia Icefield. They found that in the upstream reach of the Sunwapta River, braiding reduced
due to a reduction of gravel input that had become trapped in the proglacial lake. They also found
that in the upper reach, as braiding decreased, there was an increase in incision, as flow was
directed into a narrower braid plain with less sediment. Similarly, it can be seen in Figures 4.6 and
51
4.7, each site where a lake formed, there is a reduction in braiding intensity and the number of
smaller channels, as sediment is withheld in the proglacial lakes (Chew & Ashmore, 2001). Each
of these sites may undergo incision as they become sediment starved and narrower.
Schiefer and Gilbert (2008) provide insight into sediment delivery/interruption at one the
study sites. They confirm that at Lillooet Glacier (site 11), a significant proportion of sediment
supply has been trapped in its proglacial lake (Silt Lake), as seen in lake sediment cores. A varve
chronology of Silt Lake from 1947 to 1962, during active glacier retreat, shows a decrease in
sediment supply, effecting downstream sediment delivery. Reduced discharge is less evident due
to the lack of stream gauge data but has been identified elsewhere (e.g. Kennie and Bogen (2014)).
While only two sites contained stream gauges, neither showed a reduction in flow following the
growth of a proglacial lake, suggesting that the main driver of planform change, in the upper reach,
is likely due to reduced sediment input, inducing incision.
Figure 5.1 schematically illustrates the process of lake formation and decreasing braiding
intensity. The effect of creating an upstream reservoir causes narrowing of the braid plain and
fewer channels, as well as incision in the upper reach, reducing the slope of the braid plain
(Kondolf et al., 2014). In contrast, the two control systems that did not form a proglacial lake have
increased or maintained braiding intensity in their proglacial rivers over this timespan (Fig. 4.6).
These systems are continually supplied sediment and meltwater from the glacier, maintaining or
fluctuating the degree of braiding depending on the rate of retreat and melt occurring.
52
Figure 5.1. Schematic illustrating planform change of a proglacial river as lake area increases. Overall braiding and
channel width reduce as lake area increases. Sketch based on East Fork Susitna Glacier
Bogen et al. (2015) observed sediment starved systems in Norwegian valley glaciers that
formed proglacial lakes during the past century, and found that as glacier retreat continues, the
coarse sediment delivered from sub-glacial tunnels becomes deposited in a proglacial lake. This
reduces the available coarse sediment by trapping it in upstream reservoirs, likely causes a
reduction in downstream braiding, which is observed in every site where a lake has formed in the
53
present study. Bogen et al. (2015) also found that bed topography of present-day valley glaciers
indicates the potential development of large sedimentation basins as glacier retreat continues. If
this trend is consistent with each study site and proglacial lakes continue to grow, an increase in
trapped sediment will maintain a reduced braiding intensity, compared to before the formation of
a lake where braiding intensity values were higher. Geilhausen et al. (2013) have also observed a
significant reduction in downstream sediment flux due to lake formation and retreat of the
Obersulzbachkees Glacier, Austria. They suggest the number of proglacial lakes is expected to
increase simultaneously with glacier retreat, affecting sediment delivery, impacting hydrology,
river ecology, and reservoir management.
The role of proglacial lakes on fluvial systems is similar to the implementation of dams
(Kondolf et al., 2014). Skalak et al. (2013) have produced a conceptual model of how dams may
affect river morphology. Their model suggests with the addition of a dam, morphological
adjustments can include erosion of islands/bars and induce new vegetation growth. This is
observed in our sites where proglacial lakes have formed (Fig. 4.6). Erosion of islands and bars is
reflected in the reduction in braiding and an increase in vegetation is seen within the braid plain as
they become narrower with less flow. These proglacial rivers have seemingly undergone a nearly
permanent reduction in braiding, due to the formation of proglacial lakes and appear to maintain
this geometry as long as a lake resides.
54
Chapter 6: Conclusions and future research
6.1 Conclusions
The present research demonstrates the response of channel morphology to a change in river
discharge and sediment supply that has resulted as a function of anthropogenically accelerated
glacial retreat. This work shows that by increasing discharge, the braiding intensity will increase
as well as cause increases in the erosion of alluvial fans. With decreasing discharge, braiding
intensity remains lower, with nearly unaffected alluvial fans. This work also identifies the
influence of proglacial lakes, where fluvial systems become starved of coarse sediment due to
entrapment within upstream reservoirs whose formation causes the downstream river to undergo
a decrease in braiding and an increase in the median width of channels within the braid plain,
inducing incision in the upper reach.
This study reveals the rapidity and extent of landscape response to changes in the
controlling variables, such as fluid discharge and sediment, which is critical for contemporary
environmental management and the prediction of future change. In addition, such quantification
can also help in deciphering the ancient sedimentary record and interpreting the impacts of auto
and allo cyclic variables that affect landscape change.
6.2 Future research
The present work documents planimetric changes in fluvial systems due to deglaciation,
where changes in discharge and sediment supply can result in bed aggradation and degradation.
Each of the sites herein have undergone changes that may result in increased incision or deposition.
Future research should involve the application of DEM differencing to extract data to identify how
these systems respond to changes in these parameters in the vertical direction. Lane et al. (2010)
55
provide a detailed explanation using coupled photogrammetric and image processing techniques
for creating and utilizing DEMs for understanding the scales and modes of erosion and deposition
of fluvial systems. This analysis, as well as centerline profiles showing gradient changes, could
provide a new perspective of fluvial response to changes in discharge and sediment supply.
Other potential future research opportunities may involve investigating downstream habitat
changes due to changes in sediment supply and discharge. Yarnell et al. (2006) hypothesize that
moderate sediment supply creates maximum spatial heterogeneity in morphology and surface
texture, hence exhibiting more geomorphic diversity, and creating ideal channel conditions for
different ecosystems. In regions where proglacial lakes have reduced sediment supply and bar
deposition, downstream habitats have likely responded to a change in surface textures within the
proglacial region.
56
Appendices
Appendix 1: Data Collection and Processing for Kaskawulsh Glacier
Table A1 describes the preprocessing of satellite imagery used for the first research
component of this thesis. Each satellite image was clipped to the area of interest. This included the
terminus of Kaskawulsh Glacier, and both Äy Ch and Kaskawulsh River.
Table A1. Below includes satellite source, resolution, band combination, and dates each satellite
supplied. *Resampled to 10 m. DRA (Dynamic Range Adjustment)
Satellite
Spot-5
PlanetScope
Worldview-2
Rapideye
Sentinel-2
Resolution
10
3*
1.84*
5*
10
Band
Combination
(RGB)
4,3,2
4,3,2
4,3,2
4,3,2
4,3,2
Dates
(2013-2020)
05-06-2015
05-16-2015
05-21-2015
05-26-2015
06-20-2015
06-25-2015
06-30-2015
07-05-2015
07-25-2015
08-04-2015
08-14-2015
09-08-2015
09-13-2015
05-06-2017
06-10-2017
06-27-2017
07-23-2017
07-25-2017
08-04-2017
08-08-2017
08-18-2017
08-30-2017
09-12-2017
09-20-2017
09-28-2017
05-06-2018
05-12-2018
05-15-2018
05-31-2018
06-19-2018
06-22-2018
06-27-2018
07-04-2018
07-08-2018
07-23-2018
07-26-2018
07-31-2018
08-05-2018
08-11-2018
08-17-2018
08-23-2018
08-30-2018
09-03-2018
09-17-2018
09-21-2018
08-11-2014
05-07-2013
05-21-2013
05-23-2013
06-16-2013
06-23-2013
07-16-2013
06-18-2014
06-27-2014
07-01-2014
09-15-2014
05-12-2013
05-10-2016
05-15-2016
06-07-2016
06-17-2016
06-24-2016
07-14-2016
08-16-2016
09-25-2016
06-09-2017
07-19-2017
09-10-2018
09-30-2018
05-20-2019
05-25-2019
06-14-2019
06-24-2019
09-02-2019
09-10-2019
09-17-2019
09-27-2019
05-12-2020
57
05-02-2019
05-13-2019
05-30-2019
06-08-2019
06-29-2019
07-01-2019
07-14-2019
07-26-2019
07-31-2019
08-04-2019
08-12-2019
08-30-2019
09-05-2019
05-07-2020
05-21-2020
05-23-2020
06-16-2020
06-23-2020
07-16-2020
58
Appendix 2: Data Collection for All Study Sites
Table A2. Satellite, historical air photos, data source, resolution, symbology, and dates each
satellite supplied. USGS data is collected from Earth Explorer. NAPL (National Air Photo
Library).
Site
Site #
Time 1
(Air Photo)
Time 2
(Satellite)
Air Photo
Source
Air Photo
Resolution
(m/pix)
Satellite
Imagery
Source
Satellite
Resolution
(m)
Hallo
1
06-11-1951
06-15-2020
USGS
1.08
Planet-
Scope
3
Kukak Bay
2
07-07-1951
07-09-2020
USGS
0.91
Planet-
Scope
3
East Fork
Susitna
3
08-08-1949
08-17-2020
USGS
1.29
Planet-
Scope
3
NW of
Robertson
4
07-02-1954
06-25-2020
USGS
1.14
Planet-
Scope
3
Robertson
5
07-09-1978
07-03-2020
USGS
1.44
Planet-
Scope
3
KG -
Äy Ch
6
07-18-1950
08-15-2015
NAPL
0.40
Planet-
Scope
3
KG -
Kaskawulsh
7
07-18-1950
08-15-2015
NAPL
0.40
Planet-
Slope
3
Meade
8
08-11-1979
08-19-2020
USGS
1.69
Planet-
Scope
3
Tulsequah
9
06-18-1958
07-1-2020
NAPL
0.69
Planet-
Scope
3
Klinaklini
10
07-30-1949
08-4-2020
NAPL
0.43
Planet-
Scope
3
Lillooet
11
08-20-1951
08-13-2020
NAPL
0.75
Planet-
Scope
3
59
Appendix 3: Missing Meteorological Data
Table A3. Number of days missing data for weather stations (Haines Junction (HJ) and Kluane
Lake Research Station (KLRS)). Precipitation (P) and Temperature (T).
Site - Month
2013
2014
2015
2016
2017
2018
2019
HJ - May
P:7
T:0
P:30
T:2
P:0
T:0
P:0
T:0
P:0
T:0
P:31
T:2
P:7
T:7
HJ - June
P:1
T:1
P:30
T:1
P:1
T:0
P:0
T:0
P:1
T:1
P:0
T:0
P:27
T:0
HJ - July
P:6
T:5
P:31
T:0
P:1
T:0
P:0
T:0
P:2
T:2
P:2
T:2
P:31
T:0
HJ - August
P:6
T:7
P:0
T:0
P:0
T:0
P:4
T:3
P:0
T:0
P:2
T:0
P:31
T:0
HJ - September
P:19
T:7
P:0
T:0
P:0
T:0
P:0
T:0
P:1
T:1
P:0
T:0
P:30
T:0
KLRS - May
P:25
T:25
P:0
T:0
P:0
T:0
P:31
T:31
P:0
T:0
P:0
T:0
P:0
T:0
KLRS - June
P:0
T:0
P:0
T:0
P:0
T:0
P:30
T:30
P:0
T:0
P:0
T:0
P:0
T:0
KLRS - July
P:0
T:0
P:0
T:0
P:0
T:0
P:21
T:21
P:0
T:0
P:0
T:0
P:0
T:0
KLRS - August
P:0
T:0
P:0
T:0
P:3
T:3
P:0
T:0
P:0
T:0
P:0
T:0
P:4
T:4
KLRS -
September
P:0
T:0
P:0
T:0
P:30
T:30
P:0
T:0
P:0
T:0
P:0
T:0
P:30
T:30
60
References
Adler, C., Huggel, C., Orlove, B., & Nolin, A. (2019). Climate change in the mountain cryosphere:
Impacts and responses. Regional Environmental Change, 19, 12251228.
https://doi.org/10.1007/s10113-019-01507-6
Agisoft Metashape Professional (Software) ((Version 1.7.3)). (n.d.). [Computer software].
Ahmed, A., & Abdelbary, M. (2004). Effect of Aswan high dam operation on river channel capacity to
convey discharges. Proceedings of the Ninth International Symposium on River Sedimentation
(Volume).
Anisimov, O., Vandenberghe, J., Lobanov, V., & Kondratiev, A. (2008). Predicting changes in alluvial
channel patterns in North-European Russia under conditions of global warming. Human and
Climatic Impacts on Fluvial and Hillslope Morphology, 98, 262274.
https://doi.org/10.1016/j.geomorph.2006.12.029
ArcGIS Pro, Esri Inc. (Version 2.8). (n.d.). [Computer software].
Ashley, G. (2002). Glaciolacustrine environments. In J. Menzies (Ed.), Modern and Past Glacial
Environments (pp. 335359). Butterworth-Heinemann. https://doi.org/10.1016/B978-075064226-
2/50014-3
Baewert, H., & Morche, D. (2014). Coarse sediment dynamics in a proglacial fluvial system (Fagge
River, Tyrol). Geomorphology, 218, 8897. https://doi.org/10.1016/j.geomorph.2013.10.021
Ballantyne, C., & Murton, J. (2018). Periglacial Geomorphology. John Wiley & Sons, Incorporated.
Bennett, M., & Glasser, N. (2011). Glacial geology: Ice sheets and landforms. John Wiley & Sons.
Bertoldi, W., Zanoni, L., & Tubino, M. (2010). Assessment of morphological changes induced by flow
and flood pulses in a gravel bed braided river: The Tagliamento River (Italy). Geomorphology,
114, 348360. https://doi.org/10.1016/j.geomorph.2009.07.017
Best, J. (2019). Anthropogenic stresses on the worlds big rivers. Nature Geoscience, 12, 721.
https://doi.org/10.1038/s41561-018-0262-x
Best, J., & Ashworth, P. (1997). Scour in large braided rivers and the recognition of sequence
stratigraphic boundaries. Nature, 387, 275277. https://doi.org/10.1038/387275a0
Bishop, P. (1995). Drainage rearrangement by river capture, beheading and diversion. Progress in
Physical Geography: Earth and Environment, 19, 449473.
https://doi.org/10.1177/030913339501900402
Bloom, C., MacInnes, B., Higman, B., Shugar, D., Venditti, J., Richmond, B., & Bilderback, E. (2020).
Catastrophic landscape modification from a massive landslide tsunami in Taan Fiord, Alaska.
Geomorphology, 353, 107029. https://doi.org/10.1016/j.geomorph.2019.107029
Bogen, J., Xu, M., & Kennie, P. (2015). The impact of pro-glacial lakes on downstream sediment delivery
in Norway. Earth Surface Processes and Landforms, 40, 942952.
https://doi.org/10.1002/esp.3669
Bostock, H. (1969). Kluane Lake, Yukon Territory, its drainage and allied problems (115G, and 115F E).
Geological Survey of Canada.
Brice, J. C. (1964). Channel patterns and terraces of the Loup Rivers in Nebraska (Report No. 422D;
Professional Paper). USGS Publications Warehouse. https://doi.org/10.3133/pp422D
Brideau, M., Shugar, D., Bevington, A., Willis, M., & Wong, C. (2019). Evolution of the 2014 Vulcan
Creek landslide-dammed lake, Yukon, Canada, using field and remote survey techniques.
Landslides, 16, 18231840. https://doi.org/10.1007/s10346-019-01199-3
Bristow, C., & Best, J. (1993). Braided rivers: Perspectives and problems. Geological Society, London,
Special Publications, 75, 1. https://doi.org/10.1144/GSL.SP.1993.075.01.01
Carrivick, J., & Heckmann, T. (2017). Short-term geomorphological evolution of proglacial systems.
Geomorphology, 287, 328. https://doi.org/10.1016/j.geomorph.2017.01.037
Carrivick, J., & Tweed, F. (2013). Proglacial lakes: Character, behaviour and geological importance.
Quaternary Science Reviews, 78, 3452. https://doi.org/10.1016/j.quascirev.2013.07.028
61
Carson, M. (1984). The meandering-braided river threshold: A reappraisal. Journal of Hydrology, 73,
315334. https://doi.org/10.1016/0022-1694(84)90006-4
Chew, L., & Ashmore, P. (2001). Channel adjustment and a test of rational regime theory in a proglacial
braided stream. Geomorphology, 37, 4363. https://doi.org/10.1016/S0169-555X(00)00062-3
Church, M. (2002). Geomorphic thresholds in riverine landscapes. Freshwater Biology, 47, 541557.
https://doi.org/10.1046/j.1365-2427.2002.00919.x
Church, M. (2006). Bed material transport and the morphology of alluvial river channels. Annual Review
of Earth and Planetary Sciences, 34, 325354.
https://doi.org/10.1146/annurev.earth.33.092203.122721
Clague, J. (2006). Rapid changes in the level of Kluane Lake in Yukon Territory over the last millennium.
Quaternary Research, 66, 342355. Cambridge Core. https://doi.org/10.1016/j.yqres.2006.06.005
Clague, J., Luckman, B., Van Dorp, R., Gilbert, R., Froese, D., Jensen, B., & Reyes, A. (2006). Rapid
changes in the level of Kluane Lake in Yukon Territory over the last millennium. Quaternary
Research. https://doi.org/10.1016/j.yqres.2006.06.005
Clarke, G., Jarosch, A., Anslow, F., Radi, V., & Menounos, B. (2015). Projected deglaciation of western
Canada in the twenty-first century. Nature Geoscience, 8, 372377.
https://doi.org/10.1038/ngeo2407
Draut, A., Logan, J., & Mastin, M. (2011). Channel evolution on the dammed Elwha River, Washington,
USA. Geomorphology, 127, 7187. https://doi.org/10.1016/j.geomorph.2010.12.008
East, A., Pess, G., Bountry, J., Magirl, C., Ritchie, A., Logan, J., Randle, T., Mastin, M., Minear, J.,
Duda, J., Liermann, M., McHenry, M., Beechie, T., & Shafroth, P. (2015). Large-scale dam
removal on the Elwha River, Washington, USA: River channel and floodplain geomorphic
change. Geomorphology, 228, 765786. https://doi.org/10.1016/j.geomorph.2014.08.028
East, A., & Sankey, J. (2020). Geomorphic and sedimentary effects of modern climate change: Current
and anticipated future conditions in the western United States. Reviews of Geophysics, 58, 6092.
https://doi.org/10.1029/2019RG000692
Egozi, R. (2006). Channel pattern variation in gravel-bed rivers [Ph.D]. University of Western Ontario.
Egozi, R., & Ashmore, P. (2008). Defining and measuring braiding intensity. Earth Surface Processes
and Landforms, 33, 21212138. https://doi.org/10.1002/esp.1658
Fahnestock, R. (1964). Morphology and hydrology of a glacial streamWhite River, Mount Rainier,
Washington. (Vol. 1Vol. 422). US Government Printing Office.
Friend, P., & Sinha, R. (1993). Braiding and meandering parameters. Geological Society, London, Special
Publications, 75, 105. https://doi.org/10.1144/GSL.SP.1993.075.01.05
Geertsema, M., & Clague, J. (2005). Jökulhlaups at Tulsequah Glacier, northwestern British Columbia,
Canada. The Holocene, 15, 310316. https://doi.org/10.1191/0959683605hl812rr
Geilhausen, M., Morche, D., Otto, J., & Schrott, L. (2013). Sediment discharge from the proglacial zone
of a retreating Alpine glacier. Zeitschrift Für Geomorphologie, 57, 2953.
Goudie. (2004). Encyclopedia of geomorphology. Psychology Press.
Goudie. (2006). Global warming and fluvial geomorphology. 37th Binghamton Geomorphology
Symposium, 79(3), 384394. https://doi.org/10.1016/j.geomorph.2006.06.023
Grill, G., Lehner, B., Lumsdon, A., MacDonald, G., Zarfl, C., & Reidy Liermann, C. (2015). An index-
based framework for assessing patterns and trends in river fragmentation and flow regulation by
global dams at multiple scales. Environmental Research Letters, 10, 015001.
https://doi.org/10.1088/1748-9326/10/1/015001
Gurnell, A., & Clark, M. (1987). Glacio-Fluvial Sediment Transfer. Wiley, 524.
Hansford, M., & Plink-Björklund, P. (2020). River discharge variability as the link between climate and
fluvial fan formation. Geology, 48, 952956. https://doi.org/10.1130/G47471.1
Holbrook, J., & Schumm, S. (1999). Geomorphic and sedimentary response of rivers to tectonic
deformation: A brief review and critique of a tool for recognizing subtle epeirogenic deformation
in modern and ancient settings. Tectonophysics, 305, 287306. https://doi.org/10.1016/S0040-
1951(99)00011-6
62
Holden, J. (2005). An introduction to physical geography and the environment. Pearson Education.
Hong, L., & Davies, T. (1979). A study of stream braiding: Summary. Bulletin of the Geological Society
of America, 90, 10941095.
Howard, A., Keetch, M., & Vincent, C. (1970). Topological and Geometrical Properties of Braided
Streams. Water Resources Research, 6, 16741688. https://doi.org/10.1029/WR006i006p01674
Huss, M., & Hock, R. (2018). Global-scale hydrological response to future glacier mass loss. Nature
Climate Change, 8, 135140. https://doi.org/10.1038/s41558-017-0049-x
IPCC. (2019). IPCC, 2019: Climate Change and Land: An IPCC special report on climate change,
desertification, land degradation, sustainable land management, food security, and greenhouse
gas fluxes in terrestrial ecosystems. https://www.ipcc.ch/srccl/
Kennie, P., & Bogen, J. (2014). Changes in sedimentation patterns in proglacial lake Engabrevatn as a
consequence of Svartisen hydropower operations, Nordland, Norway. EGU General Assembly
Conference Abstracts, 11786.
Kondolf, G., Gao, Y., Annandale, G., Morris, G., Jiang, E., Zhang, J., Cao, Y., Carling, P., Fu, K., Guo,
Q., Hotchkiss, R., Peteuil, C., Sumi, T., Wang, H.-W., Wang, Z., Wei, Z., Wu, B., Wu, C., &
Yang, C. T. (2014). Sustainable sediment management in reservoirs and regulated rivers:
Experiences from five continents. Earhs Fre, 2, 256280.
https://doi.org/10.1002/2013EF000184
Kormann, C., Francke, T., Renner, M., & Bronstert, A. (2015). Attribution of high resolution streamflow
trends in Western Austriaan approach based on climate and discharge station data. Hydrology
and Earth System Sciences, 19, 12251245. https://doi.org/10.5194/hess-19-1225-2015
Lane, S. N., Richards, K. S., & Chandler, J. H. (1996). Discharge and sediment supply controls on erosion
and deposition in a dynamic alluvial channel. Geomorphology, 15(1), 115.
https://doi.org/10.1016/0169-555X(95)00113-J
Lane, S., Widdison, P., Thomas, R., Ashworth, P., Best, J., Lunt, I., Sambrook, & Simpson, C. (2010).
Quantification of braided river channel change using archival digital image analysis. Earth
Surface Processes and Landforms, 3, 971-985. https://doi.org/10.1002/esp.2015
Leier, A., DeCelles, P., & Pelletier, J. (2005). Mountains, monsoons, and megafans. Geology, 33, 289
292. https://doi.org/10.1130/G21228.1
Leopold, L., & Maddock, T. (1953). The hydraulic geometry of stream channels and some physiographic
implications. USGS US Government Printing Office.
Leopold, L., Wolman, G., & Miller, J. (1964). Fluvial processes in geomorphology. Journal of
Hydrology, 3, 342342. https://doi.org/10.1016/0022-1694(65)90101-0
Li, Y., Wang, H., & Ma, Q. (2018). Responses of the braided channel to reduced discharge and lateral
inputs of aeolian sand in the Ulan Buh Desert Reach of the Upper Yellow River. Environmental
Earth Sciences, 77, 379. https://doi.org/10.1007/s12665-018-7569-1
Mackin, J. (1948). Concept of the graded river. Bulletin of the Geological Society of America, 59, 463
512.
Marren, P. (2005). Magnitude and frequency in proglacial rivers: A geomorphological and
sedimentological perspective. Earth-Science Reviews, 70, 203251.
https://doi.org/10.1016/j.earscirev.2004.12.002
Marren, P., & Toomath, S. (2014). Channel pattern of proglacial rivers: Topographic forcing due to
glacier retreat. Earth Surface Processes and Landforms, 39, 943951.
https://doi.org/10.1002/esp.3545
Menounos, B., Hugonnet, R., Shean, D., Gardner, A., Howat, I., Berthier, E., Pelto, B., Tennant, C., Shea,
J., Noh, M., Brun, F., & Dehecq, A. (2019). Heterogeneous changes in western North American
glaciers linked to decadal variability in zonal wind strength. Geophysical Research Letters, 46,
200209. https://doi.org/10.1029/2018GL080942
Moore, R., Pelto, B., Menounos, B., & Hutchinson, D. (2020). Detecting the effects of sustained glacier
wastage on streamflow in variably glacierized catchments. Frontiers in Earth Science, 8, 136.
https://doi.org/10.3389/feart.2020.00136
63
Otto, J. (2019). Proglacial lakes in high mountain environments. In T. Heckmann & D. Morche (Eds.),
Geomorphology of Proglacial Systems: Landform and Sediment Dynamics in Recently
Deglaciated Alpine Landscapes. Springer International Publishing. https://doi.org/10.1007/978-3-
319-94184-4_14
Petts, G. (1979). Complex response of river channel morphology subsequent to reservoir construction.
Progress in Physical Geography: Earth and Environment, 3, 329362.
https://doi.org/10.1177/030913337900300302
Rampton, V. (1981). Surficial materials and landforms of Kluane National Park, Yukon Territory.
Geological Survey of Canada.
Schiefer, E., & Gilbert, R. (2008). Proglacial sediment trapping in recently formed Silt Lake, upper
Lillooet Valley, Coast Mountains, British Columbia. Earth Surface Processes and Landforms:
The Journal of the British Geomorphological Research Group, 33, 15421556.
Shugar, D., Burr, A., Haritashya, U., Kargel, J., Watson, C., Kennedy, M., Bevington, A., Betts, R.,
Harrison, S., & Strattman, K. (2020). Rapid worldwide growth of glacial lakes since 1990. Nature
Climate Change, 10, 939945. https://doi.org/10.1038/s41558-020-0855-4
Shugar, D., Clague, J., Best, J., Schoof, C., Willis, M., Copland, L., & Roe, G. (2017). River piracy and
drainage basin reorganization led by climate-driven glacier retreat. Nature Geoscience, 10, 370
375. https://doi.org/10.1038/ngeo2932
Shugar, D., Colorado, K., Clague, J., Willis, M., & Best, J. (2019). Boundary: Mapping and visualizing
climatically changed landscapes at Kaskawulsh Glacier and Kluane Lake, Yukon. Journal of
Maps, 15, 1930. https://doi.org/10.1080/17445647.2018.1467349
Skalak, K., Benthem, A., Schenk, E., Hupp, C., Galloway, J., Nustad, R., & Wiche, G. (2013). Large
dams and alluvial rivers in the Anthropocene: The impacts of the Garrison and Oahe Dams on the
Upper Missouri River. Geomorphology of the Anthropocene: Understanding The Surficial
Legacy of Past and Present Human Activities, 2, 5164.
https://doi.org/10.1016/j.ancene.2013.10.002
Slingerland, R., & Smith, N. (2004). River avulsions and their deposits. Annual Review of Earth and
Planetary Sciences, 32, 257285. https://doi.org/10.1146/annurev.earth.32.101802.120201
Smith, N., & Smith, D. (1984). William River: An outstanding example of channel widening and braiding
caused by bed-load addition. Geology, 12, 7882.
Surian, N., & Rinaldi, M. (2003). Morphological response to river engineering and management in
alluvial channels in Italy. Geomorphology (Amsterdam, Netherlands), 50, 307326.
https://doi.org/10.1016/S0169-555X(02)00219-2
Törnqvist, T. (2013). Fluvial environments/Responses to rapid environmental change. In Encyclopedia of
Quaternary Science (pp. 686694). Elsevier. http://dx.doi.org/10.1016/B0-44-452747-8/00118-6
Tweed, F., & Carrivick, J. (2015). Deglaciation and proglacial lakes. Geology Today, 31, 96102.
https://doi.org/10.1111/gto.12094
Wasson, R. (1996). Land use and climate impacts on fluvial systems during the period of agriculture.
'PAGES Workshop Report', (Series 96-2). http://pastglobalchanges.org/downloads/docs/meeting-
products/wkshp-reps/1996-LUC-clim-imp-fluvial-syst.pdf
Yarnell, S., Mount, J., & Larsen, E. (2006). The influence of relative sediment supply on riverine habitat
heterogeneity. Geomorphology, 80, 310324. https://doi.org/10.1016/j.geomorph.2006.03.005
