Two pairs of small stagnant ice bodies on the Hazen Plateau of northeastern
Ellesmere Island, the St. Patrick Bay ice caps and the Murray and Simmons
ice caps, are rapidly shrinking, and the remnants of the St. Patrick Bay ice
caps are likely to disappear entirely within the next 5 years. Vertical
aerial photographs of these Little Ice Age relics taken during August of
1959 show that the larger of the St. Patrick Bay ice caps had an area of
7.48 km
The Hazen Plateau of northeastern Ellesmere Island, Nunavut, Canada, is a
rolling upland, with elevations rising from about 300 m above sea level
(a.s.l.) near Lake Hazen to over 1000 m along the northeastern coast of the island. The
plateau is unglaciated with the exception of two pairs of small stagnant ice
caps – the unofficially named St. Patrick Bay ice caps and, 110 km to the
southwest, the Murray and Simmons ice caps (Fig. 1). They are
collectively referred to here as the Hazen Plateau ice caps. As of 2001, the
larger St. Patrick Bay ice cap ranged in elevation between about 880 and
720 m a.s.l., with the smaller one spanning 820 to 700 m. The
Murray and Simmons ice caps lie in higher terrain; in 2001, both fell
between about 1100 and 1000 m a.s.l. The Hazen Plateau ice caps
are interpreted as forming and attaining their maximum extents during the
Little Ice Age (LIA, ca. 1600–1850) (Koerner, 1989). Like much of the Queen
Elizabeth Islands, the Hazen Plateau is presently a polar desert; annual
precipitation is typically only 150–200 mm, with a late summer and early
autumn maximum (Serreze and Barry, 2014). Summer precipitation may be
variously rain or snow. Summers are very cool but variable; assessed as part
of a multiyear glaciological study (Braun et al., 2004), the average 10 m
July air temperatures at the Murray Ice Cap summit (1100 m) measured for the
years 1999 through 2001, respectively, were 4.0, 0.2, and
1.6
Directly measured mass balances (meter water equivalent) of the Hazen Plateau ice caps. Where a value represents a multiyear record, the average annual value is shown in parentheses. Asterisks denote minimum estimates.
Sources:
This paper documents the behavior of the Hazen Plateau ice caps over the
past 55
The location of the St. Patrick Bay (STPBIC), Murray (MIC), and Simmons (SIC) ice caps. The inset map shows Ellesmere Island (EL), Axel Heiberg Island (AHI), Greenland (GL), the Arctic Ocean (AO), and stations Alert (ALR) and Eureka (EUR). Use is made of 850 hPa temperature data from the Alert radiosonde record and precipitation records from Eureka.
Surface areas (km
Table 1 lists all available direct mass balance estimates of the ice caps
(in meters water equivalent, or m w.e.). Table 2 provides all available
estimates of ice cap areas (km
In July of 1972, Canadian scientists H. Serson and J. A. Morrison surveyed
the larger of the two St. Patrick Bay ice caps. They landed by helicopter in
foul weather to find the ice cap totally covered with snow. They installed
eight accumulation stakes along a roughly 2 km transect partway
across the ice cap. The range in elevation along this transect was about 60 m,
which compares to a range for the entire ice cap of about 160 m. Later
that same summer, on 20–21 August, the ice cap was visited by G. Hattersley-Smith
and A. Davidson, who noted a “partial cover of winter snow
all around the ice margins for at least a kilometer” (Hattersley-Smith and
Serson, 1973), in striking contrast to conditions depicted in the August
1947 and 1959 aerial photographs. They concluded that while the ice cap had
been in decline (as suggested from the 1947 and 1959 photographs), by the
early 1970s it had returned to good health, “thickening slightly and
extending its margins” (icy firn was observed atop the dirty melt surface
and a perennial snow cover extended beyond the ice cap margins). This is
consistent with a known shift towards cooler summers and increased
precipitation over the eastern Canadian Arctic (Bradley and Miller, 1972;
Bradley and England, 1978). Hattersley-Smith and Serson estimated a mass
balance for the 1971/1972 season of
In 1982 and 1983, the St. Patrick Bay ice caps were the focus of detailed
energy and mass balance investigations (Serreze, 1985; Bradley and Serreze,
1987; Serreze and Bradley, 1987). The stake network was expanded on the
larger St. Patrick Bay ice cap and several stakes were installed on the
smaller one. At the end of the 1982 field season in early August, the entire
ice cap was bare ice with a well-developed cryoconite surface. Assuming that
the 1982 melt season had largely ended by early August (all visible melt had
stopped by the time that the field camp had been evacuated), the 1981/1982
mass balance for the larger ice cap was estimated at
However, the summer of 1983 was cool, and the snow never completely melted
off the surrounding tundra. The 1982/1983 annual mass balance for the larger
St. Patrick Bay ice cap was estimated at
To our knowledge, there were no further visits to the Hazen Plateau ice caps
until 1999, when C. Braun, D. Hardy, and R. Bradley of the University of
Massachusetts Amherst established a network of 11 accumulation stakes on the
Murray Ice Cap, which they further expanded in the year 2000. A new network
of 15 stakes was established on the Simmons Ice Cap in 2000. Winter snow
accumulation was measured on both ice caps in late May of 1999 through 2001,
and summer ablation was measured in late July and early August from
1999–2002. For the 4 years analyzed, 1999–2002, annual balances of both
ice caps were negative in all years, ranging from
Outlines of the St. Patrick Bay ice caps based on aerial photography from August 1959, GPS surveys conducted during August 2001, and for August of 2014 and 2015 from ASTER. The base image is from August 2015.
Outlines of the Murray and Simmons ice caps based on aerial photography from August 1959, GPS surveys conducted during August 2001, and for July 2016 from ASTER. The base image is from August 2016.
The use of ASTER in conjunction with the air photographs and GPS surveys enables a fairly detailed assessment of changes in ice cap areas from 1959 through the present. Clear-sky late summer (July–August) scenes of the St. Patrick Bay ice caps showing a strong brightness contrast between the ice and the bare, dark plateau surface, enabled manual mapping of the ice cap perimeters from ASTER for the years 2005, 2009, 2014, 2015, and 2016. For the Murray and Simmons ice caps, ASTER estimates were obtained for 2001, 2007, and 2016. For 2001, areas of the Murray and Simmons were available from both ASTER and the surface-based GPS surveys. Considering the GPS surveys for this year as ground truth, the ASTER areas for this year are accurate to within 1 % for the Murray Ice Cap and 3 % for the Simmons Ice Cap. It is assumed that this is representative of the accuracy of area mapping from ASTER for the other years.
As of July 2016, and Murray and Simmons ice caps cover 39 and 25 % of
the areas in 1959 based on the aerial photographs. By sharp contrast, both
of the St. Patrick Bay ice caps in 2016 cover only 5 % of their former
areas and have been reduced to ice patches, with the smaller ice body now
covering only 0.15 km
Outlines of the St. Patrick Bay ice caps for 1959 from aerial photography, for 2001 from GPS surveys, and for 2014 and 2015 from ASTER are shown in Fig. 2. The reductions in ice cap area are striking. Note the obvious shrinkage even between the years 2014 and 2015. Shrinkage of the Murray and Simmons ice caps is shown in Fig. 3, based on outlines from 1959, 2001, and 2016. The shrinkage of these two ice caps is clearly evident, albeit less pronounced.
Using the area estimates through 2002 and extrapolating forward, Braun et al. (2004) suggested that the Hazen Plateau ice caps would disappear by the middle of the 21st century or soon thereafter and that, given their larger size, the Simmons Ice Cap and the larger of the two St. Patrick Bay ice caps would be the last to go. However, based on data through 2016 and extrapolating forward (Fig. 4), it now appears that both of the St. Patrick Bay ice caps will disappear around the year 2020.
Time history of ice cap areas and projected times of disappearance.
From the analyses described above, and results from other glaciological
investigations for the Canadian Arctic and the Arctic as a whole, the
following conclusions are drawn:
The Hazen Plateau ice caps are unlikely to be relics of the last glacial
maximum but rather likely formed during the LIA (ca. 1600–1850) (Koerner,
1989). They may have retained their LIA extents through the first couple of
decades of the 20th century (Hattersley-Smith, 1969; Sharp et al., 2014) but
have been in overall decline ever since. Braun et al. (2004) speculate on the
basis of a mapped lichen trim line that the Murray Ice Cap may have attained
a maximum LIA extent of about 9.6 km From the 1960s through part of the 1970s, the ice caps may have
experienced a period of reduced loss or occasional growth (1971/1972,
1982/1983) in response to cooling. This basic pattern likely holds for
monitored Canadian Arctic glaciers and ice caps as a whole (Bradley and
Miller, 1972; Hattersley-Smith and Serson, 1973; Ommanney, 1977; Bradley and
England, 1978; Braun et al., 2004; Sharp et al., 2014). Since then, apart from occasional years such as 1982/1983, annual mass
balances of the four ice caps have been persistently negative (Braun et al.,
2004). This is in turn consistent with the broader pattern of reductions in
mass and area of Arctic glaciers and ice caps (Dowdeswell et al., 1997;
Dyurgerov and Meier, 1997; Arendt et al., 2002; Koerner, 2005; Sharp et al.,
2011, 2014; Fisher et al., 2012; Mortimer et al., 2016). It is also
consistent with a negative mass balance of the Greenland Ice Sheet since at
least the 1990s (Shepherd et al., 2012).
Mass balance summaries for four monitored glaciers and ice caps in the
Canadian Arctic (Devon Ice Cap, Meighen Ice Cap, Melville South Ice Cap, and
the White Glacier) are provided as part of the American Meteorological
Society (AMS) State of the Climate reports. As assessed over the period 1980
through 2010, all four have had negative average annual mass balances,
ranging from
The annual mass balance of low-accumulation ice caps and glaciers in the Canadian High Arctic is known to be primarily governed by summer warmth rather than winter accumulation (e.g., Bradley and England, 1978; Koerner, 2005). To place the behavior of the Hazen Plateau ice caps in a climate context, use is made of summer-averaged (June through August) 850 hPa temperature anomalies from the radiosonde record at Alert, located on the northeastern coast of Ellesmere Island (Fig. 1) along with estimated summer temperature anomalies for the LIA. The Alert radiosonde record extends back to 1957. We use monthly mean records contained in the Integrated Global Radiosonde Archive (IGRA; Durre et al., 2006), based on daily 00:00 and 12:00 UTC soundings. Summer averages (JJA) were eliminated if based on fewer than 70 values. The 850 hPa level is about 1400 m a.s.l. for a standard atmosphere, hence roughly 600–700 m above the surface in the vicinity of the St. Patrick Bay ice caps and 300–400 m above the surface in the vicinity of the Murray and Simmons ice caps. While arguably it might be better to look at the 925 hPa level as it is closer to the plateau surface, this level has many missing values in the IGRA record. The time series of anomalies, computed with respect to the standard averaging period 1981–2010, follows in Fig. 5.
Temperature anomalies at the 850 hPa level (
Kaufman et al. (2009) took advantage of a variety of proxy sources (e.g.,
tree rings, ice cores, lake cores) to assemble a record of Arctic summer
surface temperature anomalies that extend back 2000 years. From their
analysis, LIA summer Arctic temperatures anomalies averaged around
If it is also accepted that the ice caps were broadly in equilibrium with average LIA summer temperatures, Fig. 5 suggests generally strong negative annual balances from the beginning of the record through the early 1960s. This was followed by smaller negative and occasionally positive annual balances from the middle of the 1960s through about 2000 and a preponderance of strong negative balances from the beginning of the century through the present. For comparison with the radiosonde record, we also examined 850 hPa summer temperatures over the Hazen Plateau from the National Centers for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) reanalysis (Kalnay et al., 1996) which extend back to 1948. Given that the Alert radiosonde data are assimilated into the reanalysis, it follows that the radiosonde and NCEP/NCAR time series look similar for the period of overlap. The NCEP/NCAR records suggest that the period 1948–1956 not covered by the IGRA record was warm overall with mostly positive anomalies relative to the 1981–2000 baseline. The cooling between the late 1940s through the middle 1960s broadly corresponds to the cooling over the eastern Canadian Arctic such as discussed by Bradley and Miller (1972), Bradley and England (1978), and other studies. The time series of decadal mean summer temperatures at the 700 hPa level for the major glaciated regions of the Canadian Arctic presented by Sharp et al. (2011) based on the NCEP/NCAR reanalysis (their Fig. 9.3) is also consistent with the pattern shown in Fig. 5.
An examination of selected individual years is instructive. Based on summer
1957 temperatures, the 1956/1957 annual balance must have been strongly
negative. The same can be said for 1959/1960 and 1962/1963. By sharp
contrast, the summer of 1972, when Hattersley-Smith and Serson (1973) visited
the ice caps and remarked upon the extensive August snow cover over the
plateau and estimated a positive balance of
Regarding the summer of 2013, the obvious exception to the pattern of recent
warm years, the ASTER data and daily images from the Moderate Resolution
Imaging Spectroradiometer (MODIS) show extensive cloud cover through the
summer, making it difficult to determine whether the snow cover ever entirely
cleared off the plateau. It is likely, however, that the 2012/2013 balance
year was positive for the Hazen Plateau ice caps – the Devon Ice Cap,
Meighen Ice Cap, and the White Glacier all gained mass. Only the Melville
South Ice Cap, lying well to the west, had a negative balance (AMS, 2014).
Consistent with this view, Fig. 6 shows that summer (JJA) averaged 850 hPa
temperature anomalies over the Queen Elizabeth Islands from the NCEP/NCAR
reanalysis were about 2
Summer (JJA) 2013 air temperature anomalies at the 850 hPa level from the NCEP/NCAR reanalysis relative to a 1981–2010 baseline.
July 2015 air temperature anomalies at the 850 hPa level from the NCEP/NCAR reanalysis relative to a 1981–2010 baseline.
Regarding accelerating wastage of the St. Patrick Bay ice caps since the dawn
of the 21st century, the outsized warming of the Arctic in recent decades
compared to the rest of the Northern Hemisphere (termed Arctic
amplification) is overall most strongly expressed during the cold season,
and is not nearly as prominent in summer (Serreze and Barry, 2011).
Nevertheless, from the NASA Goddard Institute for Space Sciences (GISS)
analysis (
Rapid wastage of the St. Patrick Bay ice caps over the past 15 years likely also reflects a reduction in summer albedo, as dirt layers become progressively exposed and accumulate at the surface. During the 1982 and 1983 field campaigns, it was observed that summer precipitation over the ice caps was typically in the form of snow, temporarily increasing the surface albedo and adding some mass. The frequency of summer snowfall has likely declined in the (generally) sharply warming climate over the past 15 years. Also, as suggested from the prominent decline in the area of the larger St. Patrick Bay ice cap between 2014 and 2015, when there is an especially warm summer, the thin collar of ice at the ice cap margins (a feature evident in field observations) will be prone to completely melting. The less pronounced area reduction of the Murray and Simmons ice caps must partly be due to their higher elevation and relatively cooler summer conditions. However, the elevation difference is only about 200–300 m, which argues that the stronger response of the St. Patrick Bay ice caps to warming may also be related to ice thickness. Regional differences in the temperature lapse rate (notably the temperature inversion structure) could also be involved.
It is possible that the Hazen Plateau caps could see some temporary recovery given the large natural variability in the Arctic. However, as noted by Alt (1978) and Bradley and England (1978), for stagnant ice caps such as these, all it takes is one warm summer to erase any accumulated mass gains of a previous decade. Assessing variability and trends in Arctic precipitation is notoriously difficult but, as evaluated over the period 1950–2007, annual precipitation has generally increased across Canada and especially across Northern Canada. For example, at station Eureka in central Ellesmere Island (see Fig. 1), annual precipitation appears to have increased by at least 40 % (Zhang et al., 2011). Trends over the plateau are not known, but this suggests that, if anything, precipitation changes are helping to buffer the ice caps from summer mass loss.
Paradoxically, perhaps, loss of the Hazen Plateau ice caps may open new
research opportunities. As they recede, plant remains are exposed that can be
dated and used to better understand the past climate history of the region.
From radiocarbon dates on rooted tundra plants exposed by receding cold-based
ice caps on Baffin Island – given that the plants are killed when the
snow line drops below the collection sites – Miller et al. (2013) were able
to construct a record of summer temperatures over Arctic Canada for the past
5000 years. La Farge et al. (2013) discovered that ice loss in Sverdrup Pass,
Ellesmere Island, has exposed nearly intact plant communities for which
radiocarbon dates point to entombment during the LIA. They also found that
these recently exposed, subglacial bryophytes can regenerate, which may have
important implications for recolonization of polar landscapes. The area
surrounding the receding Hazen Plateau ice caps provides a unique opportunity
to examine this process of recolonization in the High Arctic, as the rates of
ice recession are now well-documented for the last 55
Radiosonde data for station Alert are available at the Integrated Global
Radiosonde Archive (
M. Serreze led the overall effort. C. Braun, D. Hardy, and R. S. Bradley provided GPS data and historical documents. B. Raup analyzed the ASTER data. All authors contributed to the writing.
The authors declare that they have no conflict of interest.
This study was supported by the University of Colorado Boulder, the NASA Snow and Ice DAAC award NNG13HQ033 to the University of Colorado, and NSF Award 9819362 to the University of Massachusetts. Edited by: C. Haas Reviewed by: D. Burgess and M. Sharp