Conference PaperPDF Available

Holocene lake-level recession, permafrost aggradation and lithalsa formation in the Yellowknife area, Great Slave Lowland


Abstract and Figures

The Great Slave Lowland occupies the north shore of Great Slave Lake. After glaciation, it was inundated by Glacial Lake McConnell and ancestral Great Slave Lake. Holocene lake-level recession around Yellowknife is determined from accelerator mass spectrometer ages of peat and detrital organics. In the last 8000 years, recession occurred at about 5 mm/year, and permafrost is youngest near the modern shoreline and older at higher elevations. Silty-clay sediments are abundant, and lithalsas (ice-rich permafrost mounds within mineral soils) occurring within 40 m above the present lake level are less than 6000 years old. They are common on Yellowknife River alluvium deposited within the last 3000 years. Lithalsas on this surface are assumed to have developed as permafrost aggraded into saturated sediments, and ground ice has formed within the last 250 years.
Content may be subject to copyright.
Holocene lake-level recession, permafrost
aggradation and lithalsa formation in the
Yellowknife area, Great Slave Lowland
S.A. Wolfe and P.D. Morse
Geological Survey of Canada, Natural Resources Canada, Ottawa,
Ontario, Canada
The Great Slave Lowland occupies the north shore of Great Slave Lake. After glaciation, it was inundated by Glacial
Lake McConnell and ancestral Great Slave Lake. Holocene lake-level recession around Yellowknife is determined from
accelerator mass spectrometer ages of peat and detrital organics. In the last 8000 years, recession occurred at about 5
mm/year, and permafrost is youngest near the modern shoreline and older at higher elevations. Silty-clay sediments are
abundant, and lithalsas (ice-rich permafrost mounds within mineral soils) occurring within 40 m above the present lake
level are less than 6000 years old. They are common on Yellowknife River alluvium deposited within the last 3000 years.
Lithalsas on this surface are assumed to have developed as permafrost aggraded into saturated sediments, and ground
ice has formed within the last 250 years.
L’écorégion « Great Slave Lowland » se situe sur la rive nord du Grand lac des Esclaves. Suite à la déglaciation, elle a
été inondée par le lac glaciaire McConnell, ainsi qu’un précurseur du Grand lac des Esclaves. La récession holocène du
niveau du lac près de Yellowknife a été déterminée en datant des tourbes et des débris organiques par spectrométrie de
masse par accélérateur (AMS). Pendant les 8000 dernières années, le niveau du lac a baissé d’environ 5 mm par
année, le pergélisol le plus jeune se retrouve donc près du littoral moderne, alors que le plus âgé se retrouve sur les
terres plus élevées. Les argiles limoneuses abondent et l’on retrouve des lithalses (buttes cryogènes riches en glace)
plus jeunes que 6000 ans jusqu’à 40 m au-dessus du niveau du lac moderne. Celles-ci se retrouvent souvent sur les
sédiments alluviaux de la rivière Yellowknife qui furent déposés durant les 3000 dernières années. Les lithalses sur ces
surfaces se sont probablement développées avec l’expansion du pergélisol dans des dépôts saturés, la glace de sol
s’étant formée durant les derniers 250 ans.
The Great Slave Lowland is a low-relief, Precambrian
granitic bedrock plain extending about 250 km along the
north shore of Great Slave Lake (Fig. 1), and includes
Yellowknife, the capital of the Northwest Territories, which
resides near the present-day shoreline at 62.4°N latitude
and 114.4°W longitude. Following retreat of the
Laurentide Ice Sheet, inundation by Glacial Lake
McConnell and ancestral Great Slave Lake resulted in
extensive deposition of silts and clays throughout the
Lowland. In a cold-continental setting, permafrost
aggradation into these sediments has given rise to
widespread discontinuous permafrost with extensive ice-
rich permafrost mounds (i.e. lithalsas) across the region
(Wolfe et al. 2014).
The primary purpose of this paper is to establish a
preliminary estimate of the rate of Holocene lake-level
recession in the Yellowknife area, to provide a baseline
estimate of regional terrestrial emergence and permafrost
aggradation. A secondary purpose is to examine the
distribution of lithalsas in the Yellowknife area in relation
to elevation and surficial sediments in order to estimate
their age and environmental settings of formation.
Permafrost within a lithalsa exposed by thaw slumping
into the Yellowknife River permits direct assessments of
the origin of the ice, and the age and depositional
environment of the enclosing sediment.
The Great Slave Lowland is bordered to the south by
Great Slave Lake, and to the north by the Great Slave
Upland of higher relief. As with much of northern Canada,
the region was glaciated during latest, Wisconsinan,
glaciation. With the retreat of the Laurentide Ice Sheet,
about 13 000 cal BP, it was inundated by Glacial Lake
McConnell that formed along the retreating glacial margin
between about 13 000 and 9500 cal BP (Lemmen et al.
1994; Smith 1994). At its maximum extent at about 10 700
cal BP, the lake included the combined basins of Great
Bear, Great Slave and Athabasca lakes (Fig. 1). Lake
McConnell was largely the result of glacio-isostatic
loading (Lemmen et al. 1994). Evidence of the lake exists
in the form of raised deltas, strandlines, beaches and
wave-washed eskers and drumlins. In the Great Slave
basin, maximum lake elevation was between 320 and 350
m asl along the East Arm of Great Slave Lake (Smith
1994; Kerr et al. 2014), compared to only about 180 m asl
at Fort Smith (Vanderburgh and Smith 1988), the
elevation difference due to differential isostatic uplift.
Fig. 1. Great Slave Lowland, Northwest Territories with
approximate margin of the Laurentide Ice Sheet and
extent of Glacial Lake McConnell at about 10 700 cal BP
During recession, the lake first separated from the Great
Bear basin, and remained in existence until about 9500
cal BP, when the remaining two basins separated
(Lemmen et al. 1994). The resulting ancestral Great Slave
Lake continued to decline due to differential uplift, towards
its present elevation of 156 m asl, constrained by the
Mackenzie River outlet at Fort Providence (Vanderburgh
and Smith 1988).
Glacial Lake McConnell and ancestral Great Slave
Lake left prominent signatures on the Lowland landscape,
with silt and clay covering nearly 70% of the land area
(Wolfe et al. 2014). The Lowland area, being submerged
to depths of 100 to 150 m, was a deepwater basin for fine-
grained glaciolacustrine sediment deposition. During
recession of ancestral Great Slave Lake, waves and rivers
re-worked and re-deposited sediments in nearshore
environments (Fig. 2A). Consequently, glaciolacustrine
and lacustrine deposits are broadly distributed across the
The area presently experiences a continental subarctic
climate, with an annual mean air temperature of -4.1°C,
cold winters (-25.6°C January mean), warm summers
(17.0°C July mean), and an average of 281 mm of
precipitation with 40% falling as snow (Environment
Canada, 2015). The area occupies poorly-drained low
relief terrain characterized by numerous water bodies
separated by fens, peatlands, mixed woodlands, white
birch (Betula papyrifera) and black spruce (Picea
mariana) forests and bedrock outcrops (Fig. 2A and B).
Permafrost is extensively discontinuous (Heginbottom et
al. 1995) and lithalsas, which are permafrost mounds
formed by ice segregation in fine-grained sediment, are
locally widespread with nearly 1800 identified in the region
west of Yellowknife (Wolfe et al. 2014).
Lithalsas in the Yellowknife area are typically circular
to elongate, 25 to 50 m wide and up to 300 m long and 8
m high. They are typically vegetated by birch or birch-
spruce forests and commonly occur along the margins of
small ponds, which in some cases represent thermokarst
ponds related to thawing of former lithalsa terrain (Fig.
2A). The raised topography of lithalsas, combined with the
typical deciduous (birch) forest cover makes these
features readily recognizable in the field and on stereo
aerial photographs, which were used by Wolfe and Kerr
(2014) to map the lithalsa distribution in the area.
Terrestrial emergence accompanied lake-level
recession throughout the Great Slave Lake basin. Given
the low gradient of the Lowland region, in some areas a
few metres of elevation adjustment resulted in
considerable land exposure over a short time period (Fig.
Permafrost aggradation followed terrestrial
emergence, though the regional climate has varied during
the Holocene. Prior to 6000 cal BP, exposed terrain north
of the Lowland was dominated by tundra, shrub tundra
and spruce forest tundra (Huang et al. 2004), implying
conditions colder than present. Between 6000 and 3500
cal BP treeline moved northward of its present position in
response to climate warming (Moser and MacDonald
1990; MacDonald et al. 1993), and it is uncertain if
permafrost was locally sustained within the Lowland at
that time. The regional climate cooled, but was variable,
from about 3000 Cal BP to present (Huang et al. 2004).
Local mean annual air temperatures have recently
warmed at a rate of about 0.3°C per decade since the
1940s (Riseborough et al. 2013) with an accelerated trend
of about 0.6°C per decade since 1970 (Hoeve et al.
2004), consistent with a pan-Arctic warming trend
beginning AD 1966 (IPCC 2013).
Smith (1994) reconstructed the rate of lake-level
lowering of Glacial Lake McConnell and Great Slave Lake
from radiocarbon-dated buried wood in raised deltaic
sediments on the southern side of Great Slave Lake.
Based primarily on woody debris collected along former
delta-front positions of the Slave River delta (Fig. 1),
Vanderburgh and Smith (1988) estimated the Holocene
lake-level recession of Great Slave Lake to be about 2
mm per year from 7000 14C yr BP (about 8000 cal BP) to
present. Although long-term uplift rates for the Yellowknife
area are unknown, present-day uplift is reported to be on
the order of 6.3 +/- 0.06 mm per year based on a time
series of 15 years (Mazzotti et al. 2011).
Given difference in maximum elevation of Glacial Lake
McConnell between the northern and southern sides of
the basin, we hypothesize that the rate of Holocene lake-
level recession for the north shore of Great Slave Lake
and the Yellowknife area, exceeded that for the south
shore of the lake derived from the Slave River Delta by
Smith (1994). Therefore, this study derives a preliminary
estimate of the rate of Holocene lake-level recession for
the Yellowknife area. The implications of lake-level
recession are assessed in the context of local permafrost
landscape evolution by reconstructing shorelines and
examining lithalsa distribution and cryostratigraphy in the
Fig. 2. Lithalsa and ground ice examples in the Great Slave Lowland. A) Yellowknife River extending southward into
Back Bay on Great Slave Lake, with alluvial plain and lithalsas in the foreground; B) Lithalsa terrain at Site 5 along NWT
Highway 3 near Boundary Creek at an elevation of about 10 m above the present level of Great Slave Lake; note Great
Slave Lake at a distance of about 7 km in background; C) Ground ice slump at Site 6 along the Yellowknife River; D)
Slump headwall at Site 6 exposing alluvial and lacustrine sediments with segregated ground ice at depth.
Yellowknife area in relation to elevation and surficial
sediments in order to estimate their ages and
environmental settings of formation.
3.1 Peatland cores and lake-level recession
Accelerator mass spectrometer (AMS) dating of peatland
sites represents a means to establishing the timing of
lake-level recession in the Great Slave Lowland, where
terrestrial organic accumulation occurred following
emergence. Peatlands are relatively rare in the Lowland
representing less than 2% of the total surface terrain
cover (Wolfe and Kerr 2014). Unlike permafrost areas
further south, peat plateaus are uncommon in the Great
Slave Lowland, as the surrounding terrain underlain by
unconsolidated sediments typically also contains
permafrost (Morse et al. in press). Peatland sites selected
for dating were typically near, but above, ponds and did
not appear to have undergone alterations in hydrology
that, especially, would have affected peat initiation.
Cores were obtained from peatlands in the Yellowknife
area, ranging in elevation from 205 m asl near the margin
with Great Slave Upland to 165 m asl within the Lowland.
Elevations were derived from multiple sources, including
topographic base maps, GPS and Lidar data (where
available), and were conservatively estimated the nearest
5 metres (± 2 m). Elevations were derived for both the
sample sites and the nearby surface water levels. Local
surface water levels were used as the estimates of land
surface elevation at the time of terrestrial emergence, as
these discounted the effects of heave due to permafrost
aggradation and ground ice formation and accumulation
of peat above the actual mineral surface, which is variable
and can locally exceed 3 m.
In most instances, coring was undertaken with a
CRREL corer using a 7.6 cm diameter core barrel,
although at one site (Site 4) a peat auger was used to drill
multiple boreholes to confirm stratigraphic contacts. Holes
were cored/augered through the active layer and
underlying frozen peat, and into the underlying sediment.
In all cases, in situ organic samples were collected from
the base of the peat unit and above the mineral sediment
for AMS dating by Beta Analytic and AMS ages were
converted to calendar years using Calib©7.0.4 (2014). In
situ basal ages are considered to represent the minimum-
limiting ages of peat accumulation, and thus for terrestrial
emergence. In instances where two cores were obtained
from the same site (e.g. Sites 3 and 4), the older of the
two AMS ages was used to represent the limiting age. At
one site, detrital organics (wood charcoal) collected from
within the underlying subaqueous alluvial sediments were
also dated to bracket emergence. These detrital ages are
considered to represent maximum-limiting ages of
subaqueous alluvial deposition in relation to former Great
Slave Lake levels.
3.2 Shoreline reconstruction and landscape evolution
Relations between lake-level history and landscape
evolution were examined using topographic Canadian
Digital Elevation Data (CDED) available from GeoGratis
(, combined with published
surficial sediment and lithalsa distributions for the
Yellowknife area (Wolfe and Kerr 2014). Topographic data
were used to reconstruct raised shoreline levels. Timing of
sediment deposition or reworking, including beach sands
and alluvial silts, as well as the maximum-limiting ages of
lithalsas, were estimated from their elevations and
discussed in the context of reconstructed lake-level
recession rates.
A slump exposing near-surface sediments and ground
ice, situated on an island in the Yellowknife River about 1
km (Fig. 2C) north of Back Bay, was examined to
determine sediment and ground ice origin and age. Debris
along a portion of the headwall was cleared to expose at
2.2 m high by 6 m wide section (Fig. 2D). This was
measured, sketched and photo-documented, noting
stratigraphic contacts and ice-textures. Sediment samples
were collected for grain size analysis, Atterberg limits, and
moisture content performed by the Geological Survey of
Canada. Ice samples were collected from the lower
portion of the section, where visible ice content exceeded
70%, and waters were submitted to the NWT Taiga Lab
for isotopic analysis. In situ roots within unfrozen and
frozen portions of the exposure were collected for AMS
dating by Beta Analytic.
4.1 Peatland cores and Holocene lake-level recession
Table 1 summarizes AMS ages, locations, depths,
material and depositional environments of organic
samples. Figure 3A illustrates the near-surface
stratigraphy and AMS ages derived from seven peatland
cores at five sites of varying elevation in the Lowland.
Core depths vary from about 1.5 m at Site 2 to 4.5 m at 5
and peat thicknesses range from 93 to 295 cm. Calibrated
AMS ages range from about 8000 to 1100
Sediments immediately underlying peat at these sites
are typically clayey-silts, with layers of sandy-silt and silty-
clay at greater depths (Fig. 3A). Note that at Site 3,
bedrock was encountered beneath a relatively thin layer
of sediment underlying peat. These fine-grained
sediments most likely represent alluvial or lacustrine
sediments pre-dating terrestrial emergence. At Site 5,
near Boundary Creek 30 km west of Yellowknife, detrital
charcoal contained within clayey-silt sediments date to
about 1710 (1810 to 1610) and 1460 (1533-1395) Cal BP,
respectively, with an apparent age reversal relative to
depth of deposition (Fig. 3A). These ages indicate
deposition within a subaqueous environment, possibly
either a shallow bay of ancestral Great Slave Lake or a
former alluvial plain of Boundary Creek. In either case, the
charcoal was most likely derived from a burned, forested
surface at a higher elevation. This is significant, as it
indicates that terrestrial emergence here must have
occurred between about 1460 (the age of the younger
detrital organics) and 1170 (the age of the basal peat) Cal
BP, at a site presently about 10 m above, and more than
6 km inland of, Great Slave Lake (Fig. 2B). Thus, given
the low elevation gradient, considerable terrestrial
emergence occurred over a relatively short time period.
C years BP.
A general relation is apparent between basal peat AMS
ages and elevations across the area (Fig. 3B).
A preliminary Holocene lake-level reconstruction may
be derived for ancestral Great Slave in the Yellowknife
region from the AMS ages obtained beneath the
peatlands in the region. Figure 3B illustrates the derived
lake-level recession rate for the Yellowknife area using
calibrated radiocarbon ages. A rate of approximately
Table 1. AMS dates used to infer late Holocene lake-level recession and terrestrial emergence in the Great Slave
Site Age
2σ range
(Cal BP)
(m asl)
Material Depositional
7130 ± 40
210 / 205
6730 ± 30
210 / 205
6580 ± 40
195 / 195
5880 ± 30
195 / 195
1870 ± 30
185 / 175
1590 ± 30
185 / 175
1240 ± 30
170 / 165
basal peat
1770 ± 30
170 / 165
burned wood
1570 ± 30
170 / 165
burned wood
165 / 156
in situ root
80 ± 30
165 / 156
in situ root
Elevations include collection elevation and local surface water (lake/pond/or stream) elevation. Local water level
elevations, rather than collection elevations, are used in lake-level reconstructions.
Fig. 3. Elevation-age relations in the Great Slave Lowland.
A) Near-surface stratigraphy and AMS ages beneath
terrestrial peatlands at varying elevations. B) Calibrated
AMS ages in relation to elevation. Dashed line represents
approximation of lake-level recession and terrestrial
5 mm/year is inferred. This is roughly two and half times
that of the southern side of Great Slave Lake (Smith and
Vanderburgh 1988). This may be attributed to either a true
difference in the rates of Holocene post-glacial isostatic
adjustment as hypothesized, or to differences in
methodology (i.e. in situ terrestrial peatlands dated in this
study versus allochthonous buried logs used on the Slave
River Delta). Although Smith (1994) also estimated the
rate prior to 8000 Cal BP (18 mm/year) no estimate is
made for Yellowknife, though it may be significantly
4.2 Shoreline reconstruction and landscape evolution
Figure 4 depicts the present-day (Fig. 4A), and + 40 m
reconstructed (Fig. 4B), shorelines in the Yellowknife
area. At present, the Yellowknife River flows along a 6 km
channel from Tartan Rapids (TR in Fig. 4A), with broad
alluvial plains and two main islands along the river, into
Fig. 4. Shoreline reconstructions based on elevation in the
Yellowknife area, with surficial sediment and lithalsa
distributions. A) Present-day shoreline, also showing + 20,
+ 40 and >50 m elevations. B) Shoreline configuration at
+ 40 m, and showing > 50 m elevation relative to present
day lake level. TR = Tartan Rapids
Back Bay on Great Slave Lake where alluvial plains also
occur at the north end of the bay (Fig. 4A). In contrast,
with lake levels at + 40 m relative to present-day elevation
this was part of the larger Yellowknife Bay (Fig. 4B), and
the river entered Great Slave Lake several kilometres
tothe north (not shown in figure). At that time, lake levels
were locally near the dividing limit between the Great
Slave Upland and Lowland, and the surface of sandy
outwash deposits in the Yellowknife area were reworked
as beaches (Fig. 4B). With lake-level recession, the
Yellowknife River became confined to the present-day
channel, with alluvium deposited in shallows along the
margins of the river. Based on the local relief, this
occurred when the lake was less than + 20 m relative to
present-day level (Fig. 4A).
Utilizing the Holocene lake-level recession of about 5
mm per year over the last 8000 years, the + 40 m
shoreline relative to present-day is estimated at about
8000 Cal BP (Fig. 4B). Thus, most of the Great Slave
Lowland was terrestrially exposed within the last 8000
years. More specifically, in the local Yellowknife area,
Back Bay formed about 4000 years ago when lake levels
were about + 20 m relative to present. The Yellowknife
River was not confined to its present channel below
Tartan Rapids until about 3000 years ago, when lake level
was about + 15 m. Tartan Rapids itself is relatively new,
as Prosperous Lake above the rapids is less than + 2 m
relative to Great Slave Lake. Similarly, most of the
alluvium along the Yellowknife River was deposited within
the last 3000 years, and probably more recently.
Lithalsa distribution in relation to elevation and alluvial
deposits in the Yellowknife area provides further insight
into potential timing of permafrost aggradation. Figure 5
shows the frequency distribution of lithalsas mapped in
Figure 4, relative to their elevation and, based on lake-
level recession rates, their maximum age. Whereas most
lithalsas reside within + 30 m of present lake level, and
are thus no older than 6000 years, more than half reside
with + 15 m, and thus have formed potentially within the
last 3000 years. Notably, nearly all lithalsas formed within
+ 10 m of present lake level elevation (the last 2000
years) have developed on alluvium of the Yellowknife
River and Back Bay. This suggests that these saturated
clayey-silt deposits, in close proximity to a water source,
are conducive to lithalsa formation during permafrost
aggradation. Noting the location of other lithalsas in
Figure 4, it is likely that these have similarly formed where
alluvium or nearshore lacustrine sediments were
Fig. 5. Lithalsa occurrence by elevation in the Yellowknife
area also showing occurrence of lithalsas on Yellowknife
River alluvium and maximum-limiting age approximations
based on Holocene lake-level recession rates.
4.3 Recent ground ice formation
Regarding the abundance of lithalsas along the
Yellowknife River, a retrogressive thaw slump (Site 6) on
an island in the river, 1 km north of Back Bay (Fig. 4A),
Fig. 6. Cryofacies description, analytical results (including
AMS age and δ18O values) and environmental
interpretation of ground ice exposure in slump at Site 6 on
the Yellowknife River.
was examined to assess the origin of the ice, and the age
and depositional environment of the enclosing sediment.
The slump has regressed landward from the river by
about 40 m (Fig. 2 C) and the elevation of the ground
surface at the headwall (Fig. 2D) is about 8 m above river
elevation. Figure 6 shows the stratigraphic and analytical
results, and interpretation of the section and Table 1
includes results from two AMS ages at this site. An
organic cover (Unit 1), comprised primarily of un-
decomposed moss, extended to 20 cm depth, below
which an unfrozen clayey-silt (Unit 2) with a blocky texture
extended to a maximum depth of about 90 cm, with
abundant woody roots in the upper 20 to 45 cm. This unit
contained nearly equal amounts of clay (< 4 µm) and silt
(< 63 µm to 4 µm), and a small fraction of fine-grained
sand, and was a low-plasticity clay in which the existing
water content of 28% dry weight was slightly above the
plastic limit of 23%. A thawed, undulating clayey silt and
sandy silt unit (3A and 3B) extended across the section
from about 83 to 100 cm depth. The clayey-silt sub-unit
contained 70% silt, 27% clay and minor sand, and was a
low plasticity clay with a water content of 19% at about the
plastic limit. An AMS age from a root contained within the
sediments was modern. The sandy-silt sub-unit contained
71% silt, 10% clay and 19% fine sand, with a moisture
content of 16%. A silty-clay (Unit 4), with an erosional
upper contact with Unit 3, extended to the base of the
exposure. This was divided into two sub-units based on
the thaw depth at time of observation. The upper thawed
sub-unit (4a) extended from 100 to 120 cm depth, and
contained 51% clay, 46% silt and 3% sand, and was an
intermediate plasticity clay with a moisture content of
30%, which was well above the plastic limit of 20%, and
approaching the liquid limit of 38%. The lower frozen sub-
unit (4b) extended from 120 cm to the base of the
exposure at 220 cm. It contained 58 to 64% clay, 35 to
40% silt and a small fraction of fine-grained sand, and an
intermediate plasticity clay. Ice content within the frozen
sub-unit increased substantively from 40 to 70% visible
ice by volume, and ranged from wavy lenticular to layered
sub-horizontal to inclined lenses to reticulate ice with
suspended ataxtic soil clasts up to 20 cm in diameter.
Moisture content of the combined soil and ice in the upper
portion of the sub-unit was 45%, equivalent to the liquid
limit. In contrast, the moisture content of the soil
component (without excess ice) obtained from a clay clast
in the lower portion of the sub-unit was 36%, still
exceeding the plastic limit of 23%. Oxygen isotopic values
of the excess ice from between 160 and 200 cm depth
ranged from -16.0 to -17.5‰ and the δ18O value of a
single ice crystal was -16.6. Co-isotopic ratios of δ18O/
δD fall within the range of natural lakes in the area
(Gaanderse 2015), confirming a modern water source for
the ice. An AMS age of an in situ root along a sediment-
ice margin in the section was 80 ± 30
Underlying silty-clay sediments from Unit 4 are
interpreted as lacustrine deposits, based on their high
combined silt and clay content. The clay content within
this unit is lower, however, than that typical of Glacial
Lake McConnell sediments, which usually exceed 70%
(Gaanderse 2015), suggesting that these are not
glaciolacustrine sediments. More likely, they represent
ancestral Great Slave Lake deposits, likely in relation to
the expanded Yellowknife Bay as depicted in Fig. 4B or
are more recent. The overlying sandy-silt and clayey-silt
Unit 3, with an erosional unconformity at its base, is
interpreted as fluvial channel deposits, with the sandy-silt
representing Yellowknife River channel deposits and the
clayey-silt representing side deposits in shallower water.
This would correspond to a time when the Yellowknife
River was confined below Tartan Rapids and entering into
the north end of Back Bay as it does today. Unit 2
sediments are interpreted as alluvial plain deposits,
representing a shallow-water or seasonally-flooded
environment as presently observed along the Yellowknife
River shoreline.
The nature of the ground ice from 120 cm to the base
of the exposure is indicative of segregated ice that formed
epigenetically as permafrost aggraded into alluvial
sediments following terrestrial exposure. The significant
increase in volume of ice below 160 cm depth suggests
freezing rates were slow with unlimited available water.
Significantly, the in situ roots within these the sediments
indicate the ground was unfrozen until about 80 ± 30
Atterberg limits indicate clay-like properties of these
fine-grained lacustrine and alluvial sediments, being
conducive to both ice segregation and sensitivity to thaw.
Field water contents, exceeding the plastic limit at the
base of the active layer and within underlying frozen
sediments, enable ongoing thaw slumping by promoting
sliding and flowing of liquefied sediments. It is possible
that slumping is further promoted by the increase in slope
from surface heave through ice segregation in permafrost
at depth, facilitating soil failure at the base of the active
BP. This age equates to between 253 and 25 calendar
years ago (Calib.©7.0.4, 2014), suggesting that
permafrost aggradation into the sediment likely occurred
within the last two hundred years at most. Furthermore, as
the ground surface of the headwall is presently 8 m above
river level, this suggests that as much as 8 m of ground
ice has formed within this time period, heaving the ground
The ground ice exposure on the Yellowknife River
(Site 6 shown in Fig. 4B) provides an example of
comparatively recent historical permafrost aggradation
into alluvial sediments, accompanied by epigenetic
ground ice formation, with ongoing lake-level recession.
This suggests that the processes of permafrost
aggradation and ground ice (i.e. lithalsa) formation have
continued under historically modern climate conditions.
The Great Slave Lowland was initially inundated by
Glacial Lake McConnell, and subsequently by ancestral
Great Slave Lake at about 8000 cal BP. Concerning
regional terrestrial emergence and permafrost
aggradation, the main conclusions of this study are:
1. In the Yellowknife area, Holocene lake-level
recession of ancestral Great Slave Lake occurred at a
rate of about 5 mm/year over the last 8000 years.
2. As lake-level recession exposed land within the
Great Slave Lowland, permafrost aggraded into fine-
grained lacustrine, alluvial and underlying glaciolacustrine
sediments following terrestrial emergence. Ice-rich
permafrost mounds (lithalsas) formed in some areas as
permafrost aggraded into saturated sediments.
3. In the local Yellowknife area, most lithalsas
formed within the last 6000 years, and many are locally
associated with alluvial deposits along the Yellowknife
River, which were deposited within the last 3000 years as
the river became confined to a channel below Tartan
4. The process of permafrost aggradation and
ground ice formation has continued in recent historical
times along alluvial floodplains the present-day shoreline
of the Yellowknife River.
This paper is a contribution to the Climate Change
Geoscience Program of Natural Resource Canada. We
gratefully acknowledge the field assistance of Adrian
Gaanderse and Mike Palmer and review of a draft version
of the paper by Phil Bonnaventure.
Calib.©7.0.4. 2014. Calib Manual. Stuiver, M., Reimer,
P.J. and Reimer, R.. Quaternary Isotope Lab,
University of Washington.
Environment Canada. 2015. Online Climate Data. [accessed 14 May 2015].
Gaanderse, A.J. 2015. Geomorphic origin of a lithalsa in
the Great Slave Lowlands, Northwest Territories,
Canada. MSc. Thesis. Department of Geography and
Environmental Studies. Carleton University. 155 pp.
Heginbottom, J.A., Dubreuil, M.A. and Harker, P.A. 1995.
Canada-Permafrost. In National Atlas of Canada, 5th
Edition. National Atlas Information Service: Natural
Resources Canada, Ottawa, ON; Plate 2.1. MCR
Hoeve, T.E., Seto, J.T.C. and Hayley, D.P. 2004.
Permafrost response following reconstruction of the
Yellowknife Highway. International Conference on
Cold Regions Engineering. Edmonton.
Huang, C., MacDonald G.M. and Cwynar, L.C. 2004.
Holocene landscape development and climate change
in the Low Arctic, Northwest Territories, Canada.
Palaeogeography Palaeoclimatology Palaeoecology
IPCC. 2013. Climate Change 2013: The Physical Science
Basis. Contribution of Working Group I to the Fifth
Assessment Report of the Intergovernmental Panel on
Climate Change. Stocker TF, Qin D, Plattner G-K,
Tignor M, Allen SK, Boschung J, Nauels A, Xia Y, Bex
V, Midgley PM (eds.). Cambridge University Press,
Cambridge, United Kingdom and New York, NY, USA,
1535 pp.
Kerr, D.E., Knight, R.D., Sharpe, D.R., Cummings, D.I.
and Kjarsgaard, B.A. 2014. Surficial Geology,
Snowdrift, NTS 75-L, Northwest Territories; Geological
Survey of Canada, Canadian Geoscience Map-137,
(preliminary), scale 1:125 000.
Lemmen, D.S., Duk-Rodkin, A. and Bednarski, J.M. 1994.
Late glacial drainage systems along the northwest
margin of the Laurentide ice sheet. Quaternary
Science Reviews 13: 805-828.
MacDonald G.M., Edwards, T.W.D., Moser, K.A., Pienitz,
R. and Smol, J.P. 1993. Rapid response of treeline
vegetation and lakes to past climate warming. Nature
Mazzotti, S., Lambert, A., Henton, J., James, T.S., and
Courtier, N. 2011. Absolute gravity calibration of GPS
velocities and glacial isostatic adjustment in mid-
continent North America. Geophysical Research
Letters, 38: L24311.
Moser, K.A. and MacDonald, G.M. 1990. Holocene
vegetation change at treeline Northwest Territories,
Canada. Quaternary Research 34: 227-239.
Morse, P.D., Wolfe, S.A., Kokelj, S.V. and Gaanderse,
A.J.R. In Press. The occurrence and thermal
disequilibrium state of permafrost in forest ecotopes of
the Great Slave region, Northwest Territories, Canada.
Permafrost and Periglacial Processes.
Riseborough, D.W., Wolfe, S.A. and Duchesne, C. 2013.
Permafrost modelling in northern Slave region
Northwest Territories, Phase 1: Climate data
evaluation and 1-d sensitivity analysis; Geological
Survey of Canada, Open File 7333, 50pp.
Smith, D.G. 1994. Glacial Lake McConnell
paleogeography, age, duration, and associate river
deltas, Mackenzie River Basin, western Canada.
Quaternary Science Reviews 13: 829-843.
Vanderburgh, S. and Smith, D.G. 1988. Slave River delta:
geomorphology, sedimentology, and Holocene
reconstruction. Canadian Journal of Earth Sciences
25: 1990-2004.
Wolfe, S.A., Stevens, C.W., Gaanderse, A.J. and
Oldenborger, G. 2014. Lithalsa distribution,
morphology and landscape associations in the Great
Slave Lowland, Northwest Territories, Canada.
Geomorphology 204: 302-313.
Wolfe, S.A. and Kerr, D.E. 2014. Surficial geology,
Yellowknife area, Northwest Territories, parts of NTS
85-J/7, NTS 85-J/8, NTS 85-J/9 and NTS 85-J/10,
Geological Survey of Canada, CGM 183.
... We utilize ERI results over a lithalsa (ice-rich permafrost mound) associated with a retrogressive thaw slump on an island in the Yellowknife River, Northwest Territories (Fig. 1A). The lithalsa has formed by permafrost aggradation within silty-clay of alluvial and lacustrine origins [Wolfe and Morse, 2015]. Observations along the headwall of the scarp reveal up to 70 % visible ice content within permafrost (Fig. 1A inset). ...
Conference Paper
Full-text available
Günther, F. and Morgenstern, A. (Eds.) (2016): XI. International Conference On Permafrost – Book of Abstracts, 20 – 24 June 2016, Potsdam, Germany. Bibliothek Wissenschaftspark Albert Einstein, doi:10.2312/GFZ.LIS.2016.001
... When connected to Yellowknife Bay, Frame Lake received sediment from the nearby Yellowknife and Cameron rivers. However, as Frame Lake became progressively more shallow and isolated, the sedimentation levels would have dropped (Wolfe & Morse, 2015). It seems that the initiation of urbanization around Frame Lake ''turned on'' the depositional system within the lake, allowing it to transition from a non-depositional environment, to a depositional catchment characterized by high sedimentation rate (x = 0.52 cm/year). ...
Full-text available
Frame Lake, located within the city of Yellowknife, Northwest Territories, Canada, has been identified as requiring significant remediation due to its steadily declining water quality and inability to support fish by the 1970s. Former gold mining operations and urbanization around the lake have been suspected as probable causes for the decline in water quality. While these land-use activities are well documented, little information is available regarding their impact on the lake itself. For this reason, Arcellinida, a group of shelled protozoans known to be reliable bioindicators of land-use change, were used to develop a hydroecological history of the lake. The purpose of this study was to use Arcellinida to: (1) document the contamination history of the lake, particularly related to arsenic (As) associated with aerial deposition from mine roaster stacks; (2) track the progress of water quality deterioration in Frame Lake related to mining, urbanization and other activities; and (3) identify any evidence of natural remediation within the lake. Arcellinida assemblages were assessed at 1-cm intervals through the upper 30 cm of a freeze core obtained from Frame Lake. The assemblages were statistically compared to geochemical and loss-on-ignition results from the core to document the contamination and degradation of conditions in the lake. The chronology of limnological changes recorded in the lake sediments were derived from ²¹⁰ Pb, ¹⁴ C dating and known stratigraphic events. The progress of urbanization near the lake was tracked using aerial photography. Using Spearman correlations, the five most significant environmental variables impacting Arcellinida distribution were identified as minerogenics, organics, As, iron and mercury ( p < 0.05; n = 30). Based on CONISS and ANOSIM analysis, three Arcellinida assemblages are identified. These include the Baseline Limnological Conditions Assemblage (BLCA), ranging from 17–30 cm and deposited in the early Holocene >7,000 years before present; the As Contamination Assemblage (ACA), ranging from 7–16 cm, deposited after ∼1962 when sedimentation began in the lake again following a long hiatus that spanned to the early Holocene; and the Eutrophication Assemblage (EA), ranging from 1–6 cm, comprised of sediments deposited after 1990 following the cessation of As and other metal contaminations. The EA developed in response to nutrient-rich waters entering the lake derived from the urbanization of the lake catchment and a reduction in lake circulation associated with the development at the lake outlet of a major road, later replaced by a causeway with rarely open sluiceways. The eutrophic condition currently charactering the lake—as evidenced by a population explosion of eutrophication indicator taxa Cucurbitella tricuspis —likely led to a massive increase in macrophyte growth and winter fish-kills. This ecological shift ultimately led to a system dominated by Hirudinea (leeches) and cessation of the lake as a recreational area.
... The study area lies in the Great Slave Uplands and Great Slave Lowlands of the Taiga Shield High Boreal Ecoregion (Ecosystem Classification Group 2009), which is characterized by elevated bedrock terrain interspersed with small areas of forest and peat. During the last period of glaciation (8000 to 12 000 years ago), Glacial Lake McConnell covered most of the study region and deposited silts and clays, which now infill many of the topographic depressions throughout the Great Slave Lowlands (Wolfe and Morse 2015). ...
Technical Report
Full-text available
The objective of this NWT Open File is to report the concentration of arsenic and other elements in 479 near-surface soil samples collected within 30 km of Yellowknife in 2015, 2016, and 2017. For this research, the soil samples were cored from locations that were undisturbed by buildings, roads, mining or other visible human activities to minimize the influence of recent post-mining activities and examine the effect of natural processes and the legacy of airborne emissions from former ore roasting. Sampling targeted four distinct terrain units: outcrop soils, forest canopy soils, forest canopy outcrop soils, and peatland soils. Most of the analyses have been done on the Public Health Layer, which is defined as the top 5 cm of material. The arsenic concentration in outcrop soils ranged from 3.5 mg/kg to 3000 mg/kg arsenic with a median of 165 mg/kg. The arsenic concentration in the forest canopy soils ranged from 1.0 mg/kg to 1300 mg/kg with a median value of 38 mg/kg. The arsenic concentration in the forest canopy outcrop soils ranged from 2.1 mg/kg to 4700 mg/kg with a median of 150 mg/kg. Finally, the arsenic concentration in peat soils ranged from 2.9 mg/kg to 3400 mg/kg of arsenic with a median of 95.5 mg/kg. Forty soil samples collected from below the Public Health Layer (approximately 10 cm to 40 cm below the surface) were analysed and are lower in total concentration of arsenic than the corresponding samples from the Public Health Layer, with the exception of two samples. Arsenic concentrations are highly variable at the local and regional scale, and even vary between field duplicates, likely due to the uneven distribution of arsenic-rich minerals in the soil samples, which were unsieved and unground. Statistical analysis indicate that the distance and direction from the former ore roasters, soil depth, elevation and terrain type influence total arsenic concentration.
Lithalsas of the Great Slave Lowland, Northwest Territories, occur within fine-grained glaciolacustrine, lacustrine, and alluvial deposits. Detailed investigations of a lithalsa revealed that it is composed of ice-rich sediments with ice lenses up to 0.2 m thick below 4 m depth. The observed ice accounted for about 2 m of the 4 m between the top of the lithalsa and adjacent terrain. The ice is isotopically similar to modern surface water, but enriched in δ¹⁸O relative to local precipitation. Total soluble cation concentrations are low in the basal, Shield-derived and unweathered glaciolacustrine sediments of the lithalsa. Higher concentrations in the overlying Holocene-aged lacustrine and alluvial deposits may be due to greater ion availability in Holocene surface waters. Increasing Cl⁻ and Na⁺ concentrations in clays at depth likely relate to exclusion and migration of these dissolved ions in pore water during ice lens formation though total soluble cations remain comparatively low. The lithalsa developed 700 to 300 cal yr BP. A conceptual model of lithalsa formation and landscape evolution illustrates that this feature and more than 1800 other lithalsas in the region have developed in association with Holocene terrestrial emergence following lake-level recession.
This paper presents the history and cryostratigraphy of the upper permafrost in the High-Arctic Adventdalen Valley, central Svalbard. Nineteen frozen sediment cores, up to 10.7 m long, obtained at five periglacial landforms, were analysed for cryostructures, ice, carbon and solute contents, and grain-size distribution, and were ¹⁴C- and OSL-dated. Spatial variability in ice and carbon contents is closely related to the sedimentary history and mode of permafrost aggradation. In the valley bottom, saline epigenetic permafrost with pore ice down to depths of 10.7 m depth formed in deltaic sediments since the mid-Holocene; cryopegs were encountered below 6 m. In the top 1 to 5 m, syngenetic and quasi-syngenetic permafrost with microlenticular, lenticular, suspended and organic-matrix cryostructures developed due to loess and alluvial sedimentation since the colder late Holocene, which resulted in the burial of organic material. At the transition between deltaic sediments and loess, massive ice bodies occurred. A pingo developed where the deltaic sediments reached the surface. On hillslopes, suspended cryostructure on solifluction sheets indicates quasi-syngenetic permafrost aggradation; lobes, in contrast, were ice-poor. Suspended cryostructure in eluvial deposits reflects epigenetic or quasi-syngenetic permafrost formation on a weathered bedrock plateau. Landform-scale spatial variations in ground ice and carbon relate to variations in slope, sedimentation rate, moisture conditions and stratigraphy. Although the study reveals close links between Holocene landscape evolution and permafrost history, our results emphasize a large uncertainty in using terrain surface indicators to infer ground-ice contents and upscale from core to landform scale in mountainous permafrost landscapes.
Lithalsas (ice-cored permafrost mounds) are common within silty clay sediments of the Great Slave Lowland, a low-relief bedrock plain extending to about 50m above Great Slave Lake, Northwest Territories. Following retreat of the Laurentide Ice Sheet, sediment deposition in the lowland accompanied inundation by glacial Lake McConnell between about 12 700 and 9300cal BP, and continued subsequently in ancestral Great Slave Lake. Lake-level recession has occurred locally at about 5mm·a-1 for the last 8000years, due primarily to isostatic rebound. Maximum limiting ages of permafrost and lithalsas in the lowland are elevation-dependent, being least near the modern shoreline and greater at higher elevations. Many lithalsas, which are up to 8m high and several hundred metres wide, are less than 3000years old. They are abundant in alluvium of the Yellowknife River deposited within the last 2000years, with permafrost aggradation and lithalsa formation continuing in historical time.
Technical Report
Full-text available
Ninety-eight lakes were sampled within a 30 km radius of the City of Yellowknife to document elemental concentrations in surface waters in an area exposed to 50 years of emissions from gold ore processing. Concentrations of As, Sb, and SO4 are elevated in lakes within 17.5 km of Giant Mine relative to lakes beyond this distance. Arsenic concentrations were highest in small lakes (< 100 ha) that were downwind and proximal to the historic stacks, suggesting a gradient in impact from historic roaster operations at Giant Mine consistent with the predominant wind direction in the region. Concentrations of As exceeded the federal drinking water guideline of 10 µg/L for many of the lakes sampled within 12 km of the roaster stacks, and in some lakes were more than 60 times this limit. This study provides an extensive survey of elemental concentrations in regional lakes surrounding the City of Yellowknife and should be supported by future work to investigate drivers of variation in As concentration in surface waters, interannual variability in water chemistry, and the long-term fate of As and other elements of potential concern in these lakes.
Full-text available
We examine absolute gravity (AG) and vertical Global Positioning System (GPS) time series between 1995 and 2010 at eight collocated sites in mid-continent North America. The comparison of AG and GPS rates aligned to ITRF2005 yields a gravity/uplift ratio of -0.17 ± 0.01 μGal mm-1 (1 μGal = 10 nm s-2) and an intercept of -0.1 ± 0.5 mm yr-1. In contrast, aligning the GPS velocities to ITRF2000 results in a gravity/uplift intercept of -1.3 ± 0.5 mm yr-1. The near-zero gravity/uplift offset for the ITRF2005 (or ITRF2008) results shows a good alignment of the GPS vertical velocities to Earth's center of mass, and confirms that GPS velocities in this reference frame can be compared to predictions of geodynamic processes such as glacial isostatic adjustment (GIA) or sea-level rise. The observed gravity/uplift ratio is consistent with GIA model predictions. The ratio remains constant in regions of fast and slow uplift, indicating that GIA is the primary driving process and that additional processes such as local hydrology have a limited impact on a decadal time-scale. Combining AG and GPS measurements can provide significant constraints for geodetic, geodynamic, and hydrological studies.
Full-text available
FUTURE greenhouse warming is expected to be particularly pronounced in boreal regions1, and consequent changes in vegetation in these regions may in turn affect global climate2–4. It is therefore important to establish how boreal ecosystems might respond to rapid changes in climate. Here we present palaeoecological evidence for changes in terrestrial vegetation and lake characteristics during an episode of climate warming that occurred between 5,000 and 4,000 years ago at the boreal treeline in central Canada. The initial transformation — from tundra to forest-tundra on land, which coincided with increases in lake productivity, pH and ratio of inflow to evaporation — took only 150 years, which is roughly equivalent to the time period often used in modelling the response of boreal forests to climate warming5,6. The timing of the treeline advance did not coincide with the maximum in high-latitude summer insolation predicted by Milankovitch theory7, suggesting that northern Canada experienced regionally asynchronous middle-to-late Holocene shifts in the summer position of the Arctic front. Such Holocene climate events may provide a better analogue for the impact of future global change on northern ecosystems than the transition from glacial to nonglacial conditions.
Permafrost underlies peatlands of the Great Slave region, Northwest Territories, Canada, but permafrost relations beneath other ecotopes of black spruce (Picea mariana), white birch (Betula papyrifera) and mixed forests remain unknown. Permafrost-ecotope relations examined over a 3 year period (2010–13) establish the occurrence and thermal state of permafrost under these different types of forest. Air temperatures and snow depths are regionally consistent. Ground temperature variation primarily reflects latent heat effects during the freezing season, with the duration of season-normalised active-layer freezeback explaining 76% of 1 m ground temperature variation among all sites except xeric peatland. Low apparent thermal diffusivities from substantial latent heat effects strongly attenuate ground temperature variation with depth, and yield zero annual amplitude depths of 7 m or less where annual mean ground temperatures range among sites from-1.4 °C to 0.0 °C. Extensive discontinuous permafrost conditions, related to the extent of forested ecotopes, are commonly in thermal disequilibrium. Whereas permafrost in peatlands may be ecosystem-protected, this represents only about 2% of the area of the region. Permafrost in other forested ecotopes, occurring in ice-rich unconsolidated sediments, is climate-driven and ecosystem-protected because of latent heat effects. Though the rate of permafrost degradation may be reduced, an eventual transition to isolated permafrost retained primarily within ecosystem-driven peatlands implies substantial reductions of permafrost extent in this region.
Lithalsas are permafrost mounds formed by ice segregation in mineral-rich soil that occur within the zone of discontinuous permafrost. Nearly 1800 lithalsas were mapped using archival aerial photographs within the Great Slave Lowland, Northwest Territories, Canada. These are up to 8 m high and several hundred meters in length and width. One lithalsa, examined by electrical resistivity and boreholes, rises 4 to 6 m above an adjacent peatland, shows clear evidence of ice-segregation at depth and ground heave of between 2.5 and 4.0 m, and is estimated to have formed within the past 700 years. Regionally, lithalsas are typically located adjacent to ponds and streams with mature forms vegetated by deciduous (birch) forest or mixed (birch and spruce) forest with a herb–shrub understory and include circular, crescentic and linear forms. They are abundant within the Lowland region, which contains widespread glaciolacustrine, lacustrine and alluvial fine-grained sediments. The lithalsas are most common within the first few tens of meters above the present level of Great Slave Lake, indicating that many are late Holocene, and some less than 1000 years, in age. A comparison with lithalsas in contemporary environments reveals that comparatively warm but extensive discontinuous permafrost, fine-grained sediments (alluvial, lacustrine, marine or glaciomarine), and available groundwater supply provide the climatic and hydro-geological parameters for the development of lithalsas in permafrost terrain. The identification of lithalsas in this region is important given their sensitivity to climate change and potential hazards to northern infrastructure upon thawing.
Two radiocarbon-dated cores from small lakes located approximately 25 km north of the mapped boundary between forest-tundra and tundra provide records of postglacial vegetation change at the treeline near Yellowknife, NWT, Canada. Basal radiocarbon dates of 6180 and 7470 yr B.P. were obtained from the cores. The fossil pollen evidence suggests that the initial vegetation was Betula tundra with a peatland component. Alnus became an important constituent of the pollen assemblages between 6900 and 5500 yr B.P. Both lakes record sharp increases in Picea cf. mariana pollen at approximately 5000 yr B.P., suggesting the establishment of forest-tundra. By 3500 yr B.P. Picea mariana forest-tundra had withdrawn. The proportion of organic to inorganic sediment in the cores was at a maximum between 5000 and 3500 yr B.P. Tundra has dominated the region since 3500 yr B.P. In northwestern Canada, the maximum northward advance of treeline occurred between 9000 and 5000 yr B.P. The asynchrony in treeline advance in central and northwestern Canada may reflect that glacial ice persisted in the interior NWT longer than previously believed. Alternatively, the asynchronous history of the treeline may be a result of the geometric properties of the long-wave westerly disturbance that is manifest in the median summer position of the arctic front and ultimately controls the geographic location of the treeline.
Holocene landscape development in the low arctic tundra zone was studied through fossil pollen analysis and sedimentary studies of a radiocarbon-dated core of lake sediment. The results indicate that Holocene landscape development in this area consists of three main stages. The first stage occurred between 8500 and 6300 14C yr B.P. when the local vegetation shifted from a herb and graminoid tundra to Betula shrub tundra, and finally to spruce forest tundra following the regional deglaciation. Soil erosion decreased greatly in response to vegetation development. The second stage is between 6300 and 3000 14C yr B.P. during which spruce forest tundra was developed and catchment soil erosion decreased due to increased vegetative cover. The chronological difference of several hundred years in the rise of black spruce population among the treeline sites may have been caused by edaphic conditions. During the third stage from 3000 14C yr B.P. to the present, vegetation degraded to the modern Betula-dominated shrub tundra. Peatlands expanded to its modern extent in the treeline zone from 1300 14C yr B.P. It seems treeline advanced northerly ca. 50 km beyond its present position between 6300 and 3000 14C yr B.P., reflecting likely a northward retreat of the polar front following the demise of the ice sheet in the middle Holocene. The retreat of the treeline at ca. 3000 14C yr B.P. may have been responded to an expansion of the north polar vortex or an intensification of the meridional airflow.
The Holocene Slave River delta (8300 km2) is a long (170 km), narrow (42 km average width) alluvial sand body, which extends north from the Slave River rapids at Fort Smith to Great Slave Lake, Northwest Territories. The delta is flanked by the Talston and Tethul rivers and Canadian Shield to the east and by the Little Buffalo River to the west. Wave-associated sedimentary structures in lithostratigraphic logs from river cutbanks indicate that the sandy delta was wave influenced. Most of the logs (34) consist of three facies: basal laminated mud (unknown thickness), interbedded mud and sand (2.5 m), and planar-tabular ripple sets interbedded with cross-laminated to flat-bedded sand (3.0–14.5 m).Eleven radiocarbon-dated wood samples from river cutbanks were used to reconstruct the delta paleoshoreface and to calculate the rate of progradation, which averaged 20.7 m/year from 8070 BP to the present. In the same period isostatic rebound of the delta region relative to the Liard River delta averaged 12 cm/km (a total rebound of 48 m). The data were calculated normal to the retreating Laurentide ice front.From the surface to depths of 59 m, the subaerial and subaqueous delta front exhibits barrier islands, lagoons, offshore bars or sand waves, tensional cracks, slumps and pressure ridges. The barriers and offshore bars consist of medium grain-sized sand, whereas the slumps and pressure ridges are interpreted as mud.
The evolution of drainage systems along the retreating northwestern Laurentide Ice Sheet was complex. The interaction of ice-margin configuration, topography and glacioisostasy resulted in a network of meltwater rivers that variably overflowed to the Arctic and Pacific Oceans and to the Gulf of Mexico. Glacial lakes also changed dramatically in size and location during the period of deglaciation. At the last (and all time) glacial maximum, the ice sheet extended into the eastern Cordillera, blocking northward and eastward drainage to the Arctic Ocean. Some meltwater and most non-glacial runoff were diverted through the mountains to the Yukon River basin, into Alaska and the Pacific Ocean. Retreat from the glacial maximum prior to 21 ka BP allowed proglacial drainage from the western margin of the ice sheet to flow into the Beaufort Sea/Arctic Ocean. Deglaciation was rapid after about 13 ka BP, with the present route of the lower Mackenzie River established between 13 and 11.5 ka BP. Continued ice retreat led to significant southward expansion of the Mackenzie/Beaufort drainage basin at about 11.5 ka BP through drainage capture of glacial Lake Peace, which previously had drained southeastward into the Missouri River and to the Gulf of Mexico. Very rapid ice retreat between 10.5 and 10 ka BP allowed glacial lake McConnell to expand down-slope in contact with the ice margin.
Geomorphic origin of a lithalsa in the Great Slave Lowlands
  • A J Gaanderse
Gaanderse, A.J. 2015. Geomorphic origin of a lithalsa in the Great Slave Lowlands, Northwest Territories, Canada. MSc. Thesis. Department of Geography and Environmental Studies. Carleton University. 155 pp.