Skip booklet index and go to page content

Arctic Ecozone+ Status and Trends Assessment

Description of the Condition of the Arctic Ecozone+

Abiotic drivers

Introduction to climate-related drivers

In the Arctic, as elsewhere, ecosystems are determined to a large extent by abiotic drivers such as mean and extreme seasonal temperatures and precipitation, number of degree-days and frost-free days, insolation (amount of sunlight reaching the surface), wind and albedo (amount of sunlight reflected from the surface, rather than being absorbed). These factors, plus soil moisture, determine the extent of permafrost (perennially frozen ground), which limits root penetration and has other profound effects on plants and animals. Terrestrial habitats are further influenced by microclimates determined by such physical features as slope, aspect, soil type, and proximity to large water bodies.

The combined influence of these abiotic drivers creates an environment unfavourable to tree growth and generally defines Arctic ecosystems as “tundra,” a mosaic of wet and dry habitats dominated by varying proportions of mostly perennial grasses, forbs and low-growing shrubs. Because of the variety and dominance of shrubs in the heather family (Ericaceae), shrublands are often called “shrub-heath” or “dwarf shrub-heath” habitats, although other families, such as willows (Salicaceae) and dwarf birches (Betulaceae), are also prominent. Spring growth, flowering, and seed-setting periods are brief, and animal life histories are closely linked to vegetation phenology. Tundra closely follows the limits of continuous permafrost, while discontinuous permafrost characterizes the transition between tundra and taiga.

Large-scale changes in abiotic drivers, principally temperature and precipitation, can have profound effects on distribution of tundra, and on the distribution and composition of habitats within it. For example, as noted in the introduction, a warming period beginning about 8,000 years ago changed the Northern Arctic from polar desert to dry tundra.

The Earth has warmed in recent decades, largely due to greenhouse gas emissions. The degree of warming has been unprecedented since earlier interglacial periods in the Pleistocene Epoch (IPCC, 2007). The Arctic Climate Impact Assessment (ACIA, 2005) and the IPCC Assessment Report 4 (IPCC, 2007) both reported that surface air temperatures in the circumpolar Arctic have increased over the past few decades at almost twice the rate of the global mean, with warming being particularly pronounced since the early 1990s. Figure 12 maps temperature anomalies for 2001–2005 relative to 1951–1980. This illustrates the greater warming in the circumpolar Arctic and also the regional trend of greater warming in the western part of the North American Arctic. Figure 13 shows the trends in mean annual global and Arctic temperatures from 1880 to 2012.

Figure 12. Global increases in annual temperatures for 2001–2005 relative to 1951–1980, showing the greatest warming over land and at high latitudes in the Northern Hemisphere.
Map
Source: Overland et al. (2007), based on Hansen et al. (2006)
Long description for Figure 12

This map shows mean surface temperature anomalies for 2001–2005 relative to 1951–1980 in the Northern Hemisphere. The map illustrates there was greater warming in the circumpolar Arctic (generally 1.2–2.1°C) and also the regional trend of greater warming in the western part of the North American Arctic.

Mean annual global and circumpolar Arctic temperatures (64 to 90°N latitude), 1880 to 2012.
YearGlobal
Degrees Celsius
Arctic
Degrees Celsius
1880-0.47-0.92
1881-0.38-0.77
1882-0.32-1.24
1883-0.31-0.11
1884-0.59-1.12
1885-0.48-1.21
1886-0.59-1.11
1887-0.67-1.49
1888-0.40-1.39
1889-0.18-0.60
1890-0.58-1.02
1891-0.61-1.18
1892-0.52-1.30
1893-0.55-0.87
1894-0.45-1.10
1895-0.37-0.82
1896-0.35-1.08
1897-0.27-0.64
1898-0.46-1.08
1899-0.38-1.02
1900-0.19-0.56
1901-0.19-0.42
1902-0.36-1.56
1903-0.40-0.42
1904-0.53-0.24
1905-0.36-0.01
1906-0.25-0.16
1907-0.55-0.57
1908-0.41-0.29
1909-0.42-0.66
1910-0.35-0.62
1911-0.38-0.25
1912-0.37-0.66
1913-0.37-0.47
1914-0.11-0.48
1915-0.09-0.61
1916-0.32-0.30
1917-0.52-0.85
1918-0.45-1.14
1919-0.19-0.89
1920-0.290.09
1921-0.17-0.07
1922-0.23-0.31
1923-0.260.16
1924-0.230.28
1925-0.26-0.19
1926-0.050.36
1927-0.19-0.18
1928-0.140.50
1929-0.34-0.01
1930-0.170.46
1931-0.110.46
1932-0.150.24
1933-0.29-0.35
1934-0.130.68
1935-0.250.10
1936-0.130.15
1937-0.021.05
19380.051.25
1939-0.140.46
19400.050.75
19410.03-0.31
19420.050.25
1943-0.011.11
19440.040.95
1945-0.080.41
1946-0.07-0.17
19470.050.91
1948-0.080.07
1949-0.120.18
1950-0.220.11
1951-0.070.00
1952-0.010.09
19530.060.85
1954-0.130.64
1955-0.12-0.48
1956-0.23-0.24
19570.040.09
19580.07-0.15
19590.030.56
1960-0.030.37
19610.08-0.15
19620.020.50
19630.01-0.09
1964-0.26-0.64
1965-0.17-0.26
1966-0.07-0.79
1967-0.010.42
1968-0.08-0.29
19690.03-0.02
19700.07-0.17
1971-0.07-0.09
1972-0.04-0.32
19730.210.18
1974-0.04-0.14
19750.020.26
1976-0.180.00
19770.200.23
19780.13-0.13
19790.19-0.58
19800.330.33
19810.431.21
19820.12-0.26
19830.370.31
19840.200.45
19850.180.34
19860.230.02
19870.38-0.32
19880.460.71
19890.350.46
19900.500.69
19910.500.85
19920.21-0.12
19930.240.67
19940.370.45
19950.541.43
19960.470.86
19970.520.84
19980.830.94
19990.560.56
20000.551.17
20010.661.07
20020.761.33
20030.751.50
20040.670.72
20050.842.08
20060.741.75
20070.842.00
20080.621.38
20090.771.17
20100.911.87
20110.772.00
20120.741.67

Top of Page

Figure 13. Mean annual global and circumpolar Arctic temperatures (64 to 90°N latitude), 1880 to 2012.

Land-surface air temperature anomalies. Based on meteorological station measurements, the annual means include the Canadian stations discussed below. Data are anomalies, calculated on the base period 1951 to 1980.

Graph
Source: ecozone+ data provided by authors of Zhang et al. (2011)
Long description for Figure 13

This graphic represents a line graph showing the following information:

Mean annual global and circumpolar Arctic temperatures (64 to 90°N latitude), 1880 to 2012.
YearGlobal

Degrees Celsius
Arctic

Degrees Celsius
1880-0.47-0.92
1881-0.38-0.77
1882-0.32-1.24
1883-0.31-0.11
1884-0.59-1.12
1885-0.48-1.21
1886-0.59-1.11
1887-0.67-1.49
1888-0.40-1.39
1889-0.18-0.60
1890-0.58-1.02
1891-0.61-1.18
1892-0.52-1.30
1893-0.55-0.87
1894-0.45-1.10
1895-0.37-0.82
1896-0.35-1.08
1897-0.27-0.64
1898-0.46-1.08
1899-0.38-1.02
1900-0.19-0.56
1901-0.19-0.42
1902-0.36-1.56
1903-0.40-0.42
1904-0.53-0.24
1905-0.36-0.01
1906-0.25-0.16
1907-0.55-0.57
1908-0.41-0.29
1909-0.42-0.66
1910-0.35-0.62
1911-0.38-0.25
1912-0.37-0.66
1913-0.37-0.47
1914-0.11-0.48
1915-0.09-0.61
1916-0.32-0.30
1917-0.52-0.85
1918-0.45-1.14
1919-0.19-0.89
1920-0.290.09
1921-0.17-0.07
1922-0.23-0.31
1923-0.260.16
1924-0.230.28
1925-0.26-0.19
1926-0.050.36
1927-0.19-0.18
1928-0.140.50
1929-0.34-0.01
1930-0.170.46
1931-0.110.46
1932-0.150.24
1933-0.29-0.35
1934-0.130.68
1935-0.250.10
1936-0.130.15
1937-0.021.05
19380.051.25
1939-0.140.46
19400.050.75
19410.03-0.31
19420.050.25
1943-0.011.11
19440.040.95
1945-0.080.41
1946-0.07-0.17
19470.050.91
1948-0.080.07
1949-0.120.18
1950-0.220.11
1951-0.070.00
1952-0.010.09
19530.060.85
1954-0.130.64
1955-0.12-0.48
1956-0.23-0.24
19570.040.09
19580.07-0.15
19590.030.56
1960-0.030.37
19610.08-0.15
19620.020.50
19630.01-0.09
1964-0.26-0.64
1965-0.17-0.26
1966-0.07-0.79
1967-0.010.42
1968-0.08-0.29
19690.03-0.02
19700.07-0.17
1971-0.07-0.09
1972-0.04-0.32
19730.210.18
1974-0.04-0.14
19750.020.26
1976-0.180.00
19770.200.23
19780.13-0.13
19790.19-0.58
19800.330.33
19810.431.21
19820.12-0.26
19830.370.31
19840.200.45
19850.180.34
19860.230.02
19870.38-0.32
19880.460.71
19890.350.46
19900.500.69
19910.500.85
19920.21-0.12
19930.240.67
19940.370.45
19950.541.43
19960.470.86
19970.520.84
19980.830.94
19990.560.56
20000.551.17
20010.661.07
20020.761.33
20030.751.50
20040.670.72
20050.842.08
20060.741.75
20070.842.00
20080.621.38
20090.771.17
20100.911.87
20110.772.00
20120.741.67

Top of Page

A major consideration for the Arctic Ecozone+ is the role of feedback mechanisms in amplification of climate change in the Arctic--particularly as these mechanisms involve changes in ecosystems. For example, reductions of surface albedo from the retreat of sea ice has led to enhanced warming of the lands peripheral to the Arctic Ocean (Walsh, 2008); shorter snow cover duration has resulted in greater surface warming and increased shrub growth, further reducing surface albedo and leading to additional regional warming (Sturm et al., 2001; Chapin III et al., 2005; Walsh, 2008). Increased shrub growth also increases the ability to retain a deeper snow cover which impacts the ground thermal and hydrological regimes, microbial activity, and nutrient and carbon cycling (Myers-Smith, 2007).

Observations by Canadian Inuit about climate change are documented in many sources, including the Arctic Climate Impact Assessment (ACIA, 2005) and studies based on interviews in Inuit communities (Kassi, 1993; McDonald et al., 1997; e.g., Fox, 1998; Fox, 2004; Huntington et al., 2005). A common theme is the increased variability of the climate. For example, people in Igloolik report that weather is more unpredictable and sudden weather changes have forced hunters to spend extra unplanned nights on the land (Ford et al., 2006). People in Nunavut report experiencing warmer temperatures year-round accompanied by changes in the length and timing of the seasons (Government of Nunavut, Department of Environment, 2005). The communities of the Inuvialuit Settlement Region have been observing changes associated with warming in their region for a longer period than those living in the Eastern Arctic communities and these changes appear more pronounced (Furgal et al., 2002; Nickels et al., 2002; Furgal and Seguin, 2006). Milder winters, warmer summers, a shorter fall and a slower and later freeze-up were among the many weather-related changes observed by Banks Island Inuvialuit. Fluctuations in the seasons were noted, particularly the earlier arrival of spring (Ashford and Castleden, 2001).

Climate variability and change is the unifying theme of this report and the material outlined above is expanded upon in the relevant sections throughout the report.

Top of Page

Climate trends since 1950

Overview

This section is based on records from Canada’s network of climate stations, analysed for the country as a whole, and on an ecozone+ basis for this report. Results of the analysis for Canada are presented in the ESTR thematic report Canadian climate trends, 1950–2007 (Zhang et al., 2011). A base or reference period (1961–1990) was used to construct both a regionally averaged and station-by-station anomaly series of precipitation and temperature trends. The results of these trend analyses for the Arctic Ecozone+ are summarized in Table 2 and presented and discussed below in relation to the Canadian trends from the same analysis and with reference to Canadian and circumpolar Arctic climate trends discussed by other authors. Zhang et al. (2011) also analysed trends related to snow, reported in the section on Ecosystem functions/processes  (page 25).

Temperature and precipitation stations are widely distributed across the Arctic Ecozone+, but station locations are biased to coastal areas and the station density is low. The ecozone+ average values for temperature and precipitation cover at least three distinct climatic regions [Canadian Arctic Archipelago (also known as the Arctic Archipelago), Western Arctic, and Eastern Arctic] and long-term variations of climate in these regions may be different. This means that ecozone+ averages are not necessarily representative of status and trends at different locations within the ecozone+. Maps showing trends at individual stations are included to provide a better indication of regional trends.

See Zhang et al. (2011) for more information on methodology of this analysis.

Table 2. Overview of climate trends since 1950 across the Arctic Ecozone+.
Climate variableTrends from 1950 to 2007
TemperatureSignificant increases in ecozone+ seasonal average temperatures in summer and fall. Significant increases at many climate stations in all seasons (Figure 14). No significant cooling trends at any station in any season.
PrecipitationIncreased significantly in all seasons across the ecozone+, with the greatest increase being in winter. Significant increases at many climate stations in fall, winter and spring (Figure 16). Significant increase at only 1 station in summer.

Top of Page

Temperature

Annual mean temperature increased about 1.4°C between 1950 and 2007 across Canada, with warming trends evident in most parts of the country. Although for the country as a whole the greatest temperature increases occurred during winter and spring, warming in the Arctic Ecozone+ was most consistent in summer and fall, with significant temperature increases of 0.9°C in the summer and 1.7°C in the fall. These means mask the greater changes that have occurred at some locations, as can be seen on the maps showing magnitude and significance of temperature trends for each Arctic climate station included in the data set (Figure 14).

Growing season in this national analysis is defined as the period during which the mean daily temperature exceeds 5°C (based on a running average of 5 days). Significant increases in growing season length were observed in many places in Canada, largely due to earlier starts to the growing season. The selection of 5°C as the minimum temperature is arbitrary and the characteristics of the actual growing season vary among species, with cold-adapted tundra plants starting growth just above the freezing point. As this 5°C threshold for the start of the growing season was not met at many Arctic locations, trends in growing season in the Arctic were not detected through this analysis. The records, however, show significant increases in temperatures at many Arctic stations, across all seasons (Figure 14). Temperature measurements taken continuously since 1995 as part of the International Tundra Experiment (ITEX) site at Alexandra Fiord, Ellesmere Island, show an increase in the growing season length (measured as days above 0°C) of about 1.5 days per year over the 12 years of measurement (Figure 15). Plant productivity has increased across the ecozone+, indicating both an increase in peak productivity and in growing season length (Zhou et al., 2001; Goetz et al., 2007) – see the section on Primary productivity (page 71).

Figure 14. Change in mean temperature, 1950–2007, for: a) spring (March–May), b) summer (June–August), c) fall (September–November), and d) winter (December–February).
Maps
Source: ecozone+ data provided by authors of Zhang et al. (2011)
Long description for Figure 14

This figure is four maps showing the direction, magnitude, and significance of seasonal temperature trends at 20 Arctic climate stations between 1950 and 2007. The maps show significant increases in temperatures at many of the stations, across all seasons. There is only one station that showed a decline greater than 30C (summer and winter) in any season.

Top of Page

Figure 15. Growing season length in control plots of a long-term warming experiment in a Cassiope heath tundra at Alexandra Fiord, Ellesmere Island.

Growing season is defined here as the number of days with mean daily air temperature greater than 0°C. Air temperature was measured at 10 cm above the ground surface. Data are means (n=2) with range bars. The trend line is a linear regression showing an increase of ca. 1.5 days per year over the 12 years.

Graph
Source: G. Henry, unpublished data
Long description for Figure 15

This line graph shows the following information:

Growing season length in control plots of a long-term warming experiment in a Cassiope heath tundra at Alexandra Fiord, Ellesmere Island.
YearAverage growing season length (days)
199579
199662.5
199776
199888
199975
200077
200182
200286
200385.5
200492
200583
200688

Top of Page

Precipitation

Precipitation increased at the majority of stations across Canada between 1950 and 2007, with stations in Arctic Canada experiencing the greatest relative increases. Precipitation increased significantly in all seasons in the Arctic Ecozone+, showing increases from 16 to 73% of the base period mean. Fall, winter, and spring precipitation showed signs of a regime shift in the mid- to late 1970s while summer appeared to be more stable--though still significantly increasing, but by only 16% of the base period mean. The abrupt change in precipitation during the fall, winter, and spring in the late 1970s may be related to a change in the atmospheric circulation, specifically a shift in the Pacific North America teleconnection pattern that occurred in 1976 (Brown and Braaten, 1998). Figure 16 shows the magnitude and significance of changes in precipitation for each season at each station.

While it is not yet clear what is responsible for the precipitation changes in Canada, a study by Min et al. (2008) suggests that precipitation increases over Northern Hemispheric high latitudes (north of 55°N) may have been a result of anthropogenic influences on climate. The observed trend towards more precipitation in Canada, especially in the North, is consistent with climate model projections of future changes in precipitation, and is thus likely to continue in the future (Zhang et al., 2011). As projected increases in precipitation are strongly correlated with projected warming, precipitation increases in the Arctic are likely to be much larger than global mean increases (Kattsov et al., 2007).

The Arctic Climate Impact Assessment (2005) reported an increase in precipitation of 1.4% per decade from 1900 to 2005 for the Arctic (60 to 90°N)--considerably greater magnitude of change than has been observed in this analysis for the Arctic Ecozone+. The larger increases in winter precipitation observed in this analysis are consistent with both model projections and with observations for the Arctic as a whole (ACIA, 2005; Lemke et al., 2007).

Figure 16. Change in the amount of precipitation, 1950–2007, for: a) spring (March–May), b) summer (June–August), c) fall (September–November), and d) winter (December–February).

Expressed as a percentage of the 1961–1990 mean.

Maps
Source: ecozone+ data provided by authors of Zhang et al. (2011)
Long description for Figure 16

This figure is four maps with symbols showing the direction, magnitude, and significance of changes in seasonal precipitation amounts at Arctic climate stations between 1950 and 2007. Precipitation increased significantly in all seasons in the Arctic Ecozone+, showing increases from 16 to 73% of the base period (1961–1990) mean. Fall, winter, and spring precipitation showed signs of a regime shift in the mid- to late 1970s while summer appeared to be more stable--though precipitation still increased significantly.

Top of Page

Climate oscillations

The section is based on the ESTR thematic report Large‐scale climate oscillations influencing Canada, 1900–2008 (Bonsal and Shabbar, 2011) with the addition of information specific to the Arctic, as noted.

Observed trends and variability in Canadian climate are influenced by large-scale atmospheric and oceanic oscillations known as teleconnections. There are several identified teleconnection patterns used to describe various circulation features across the globe (Bonsal and Shabbar, 2011). Relationships between these teleconnections and Canadian climate are strongest during the cold season (late autumn through spring) although some connections with summer conditions have been identified. In addition, they have strongest and more consistent impacts on temperature variables and, to a lesser extent, on precipitation-related factors. The Arctic Ecozone+ is particularly influenced by the Arctic Oscillation (AO) and by North Pacific patterns of oscillation, for example the Pacific Decadal Oscillation (PDO), though other climate oscillations such as the El Niño/Southern Oscillation (ENSO) also affect Arctic temperatures.

The impact on cold-season temperatures over various regions of Canada has led to significant relationships between teleconnection patterns and several spring climate-related variables. These include the duration of lake and river ice in western Canada (Bonsal et al., 2006), the timing of snowmelt and spring peak streamflow across western North America (Stewart et al., 2005), and the timing of spring runoff in the Mackenzie River (related to the PDO and other indices, but not the AO or ENSO) (Burn, 2008).

Indices that represent these large-scale oscillations are correlated with fluctuations and cycles in Arctic ecosystems. The linkages are not always straightforward, with ecosystem variability often being linked with more than one teleconnection, and with the relationship varying from region to region. Some examples of associations with Arctic biotic communities:

  • High values of the AO are associated with periods of reduced plant growth and reduced reindeer populations on Svalbard (Norway) (Aanes et al., 2002).
  • Large-scale climate patterns play a role in caribou population fluctuations, but the associations vary from region to region: in Alaska, the PDO is correlated with changes in western Alaska herds, while herds in northeastern Alaska are more affected by the AO (Joly et al., 2011).
  • Mortality of greater snow goose young during summer on Bylot Island was shown to be lower at both extremes of the AO: in years when the index is very positive (low summer temperatures) and in years when it is very negative (high summer temperatures) (Dickey et al., 2008).
Arctic Oscillation Index

The Arctic Oscillation (AO) represents atmospheric circulation variability over the Northern Hemisphere north of the tropics where sea-level pressure over the polar regions varies in opposition with that over middle latitudes (about 45°N) (Thompson and Wallace, 1998). It is strongly correlated with the North Atlantic Oscillation (NAO) that influences climate in eastern North America and Europe. When the AO is negative, winter temperatures are lower, ocean waters are fresher, sea ice coverage is greater, and sea ice is thicker than when the AO is positive. During 1950–1980, the AO shifted frequently between positive and negative, and remained positive from 1989 to 1995 (Richter-Menge et al., 2006) (Figure 17). Features of the positive AO phase include warmer-than-average conditions over large regions in northern Europe and Asian Russia, with opposite anomalies in the eastern half of North America (Buermann et al., 2003). The AO has been linked with a variety of physical and ecological features in Arctic and subarctic latitudes. For example, positive AO levels are associated with less winter and spring season snow extent, especially in Eurasia (Bamzai, 2003), earlier melt onset, especially in the European Arctic (Tedesco et al., 2009), higher river discharge to Bering Strait, but lower river discharge to Hudson Bay (Déry et al., 2005), and declines in some Alaskan caribou herds, including the Porcupine Herd whose range extends into Canada (Stenseth et al., 2003).

Figure 17. Time series of the annually averaged Arctic Oscillation Index (AO) for the period 1950–2005.
Graph
Source: based on Figure 5 from Richter-Menge et al. (2006) (courtesy of I. Rigos)
Long description for Figure 17

This bar graph shows the annually averaged Arctic Oscillation Index for the period 1950–2005. From 1950–1980, the Arctic Oscillation shifted frequently between positive and negative, with negative AO in the majority of years. It remained strongly positive from 1989 to 1995.  From 1996 to 2006, it shifted between positive and negative again.

Top of Page

The different phases in large-scale teleconnections often lead to larger-amplitude regional responses in observed climate. As a result, a number of the observed 20th-century climate changes can be attributed, at least in part, to changes in these various teleconnection patterns (Solomon et al., 2007). For example, Hurrell (1996) found that the North Atlantic Oscillation, ENSO, and variations in North Pacific circulation collectively explained a significant portion of Northern Hemisphere winter temperature variability during the 20th century. Specifically for North America, the mid-1970s climate shift to positive Pacific Decadal Oscillation and more frequent El Niño events appear to have led to contrasting changes across the continent, with the west warming more than the east for a period thereafter (Trenberth et al., 2007). This shift to a positive PDO has therefore also been associated with the trend toward warmer winter and spring temperatures over western Canada, including in the Arctic.

It is not currently known how increases in greenhouse gas concentrations have affected the occurrence of these large-scale climate oscillations. Furthermore, the effects of projected future climate change on the major teleconnection patterns remain uncertain since there is a lack of agreement among the various climate models concerning the future frequency and structure of large-scale atmospheric and oceanic modes. Climate oscillations do not account for recent warming patterns in the Canadian and circumpolar Arctic (Figure 18).

Figure 18. Northern Hemisphere surface temperature anomalies averaged over three periods with different types of dominating pattern of natural variability.

The North Pacific pattern was dominant in the Arctic from 1977 to 1988, while the Arctic Oscillation dominated the region from 1989 to 1995 (see Figure 17). In recent years (the period from 2000–2005), however, neither of these alternate states is evident and the pattern of warm temperature anomalies is circumpolar in distribution.

Maps
Source: Overland and Wang (2007)
Long description for Figure 18

This figure has three circular heat maps of the Northern Hemisphere showing periods with different types of dominating patterns of natural variability. The North Pacific pattern was dominant in the Arctic from 1977 to 1988, while the Arctic Oscillation dominated the region from 1989 to 1995. In recent years (the period from 2000–2005), however, neither of these alternate states is evident and the pattern of warm temperature anomalies is circumpolar in distribution. These maps demonstrate that climate oscillations do not account for recent warming patterns observed in the Canadian and circumpolar Arctic.

Top of Page

Ecosystem functions/processes

Introduction

In 1973, Dunbar (1973) published a landmark paper, “Stability and fragility in Arctic ecosystems”. He examined the apparent dichotomy that Arctic ecosystems are susceptible to damage, even though species that live there all year are extraordinarily tough. Others migrate huge distances and habitats undergo extreme annual fluctuations in living conditions. Arctic terrestrial animals, such as lemmings (), can crash to extremely low population levels or even be locally extirpated, but can recover as long as there is a vast area available in which to harbour recolonizing stocks. Species are protected from invasion by ecological equivalents, as well as their predators and pathogens, and by climatic limitations that maintain tundra habitats. He pointed out that “stability” has to be defined in terms of coherence and continuity of annual and multi-annual cycles, and that “fragility” relates not so much to low resistance to small-scale physical damage as to restricted growth and turnover rates that limit resilience. He gave this example of vulnerability due to slow turnover:

“If an Arctic lake full of char or lake trout is fished out in a season or two, which has happened many times, the damage is either permanent or will take many years to repair... Lakes, especially small lakes, in the Arctic, are probably the most vulnerable part of the landscape... Many of them harbour only one species of herbivore (a copepod crustacean), and one species of primary carnivore, usually a mysid or an amphipod. These are analogous to the lemming of the tundra, but without the protection of large geographic scale. They represent extreme cases of ecosystem simplicity, in which the removal of one species means ecological disaster.”

Since then, our understanding of resilience has improved, so that, for example, we know that, when populations drop below a minimum viable population level (a threshold), their recovery becomes dependent on conditions that enhance resilience for that species: for example, low mortality from predation or harvest, or sufficient range size and sufficient connectivity to allow other populations to act as sources for recolonization. Stochastic events such as unusual winter storms, industrial disturbance, or introduction of a new predator or pathogen may pose high probabilities of extinction or extirpation. Peary caribou (Rangifer tarandus pearyi) are a case in point, as discussed in the Ecosystem composition - Peary caribou section (page 107). After a decade of intensive research, scientists are beginning to understand how a changing climate may destabilize Arctic ecosystems--and, indeed, have already detected signs of destabilization--but we have much to learn. The science on cumulative impacts, including understanding of interactions between climate change and industrial disturbance, is in its infancy. The following sections focus on processes that may indicate or foretell ecosystem perturbations. Two important characteristics of the Arctic Ecozone+ strongly influence the resilience of its ecosystems to perturbation: 1) the huge scales (temporal and geographical) at which most Arctic ecosystems operate (at least for vertebrates) in the processes of annual movements, population cycles, and recolonization events; and 2) the dominant evolutionary role of winter in driving adaptive traits and habitat structure. Understanding the influence of scale the specific adaptations and habitat structures characteristic of Arctic wildlife, in the light of current change, is essential for conservation and enhancement of Arctic ecosystem resilience.

Top of Page

The Cryosphere

Changes to the cryosphere (permafrost, snow, sea ice, glaciers, and lake and river ice) could also be classified as abiotic drivers (secondary to the primary climatic drivers of temperature and precipitation). We include these trends in the Ecosystem processes section to highlight and discuss their linkages with ecosystem disturbance and other ecosystem processes such as hydrology and primary productivity. Rapid shrinking of all aspects of the cryosphere has been a dominant trend in Arctic ecosystems that began over 30 years ago and is particularly evident over the past decade. The implications of this loss of frozen water and frozen ground, and the interconnections with other ecosystem processes, are profound and, in many cases, poorly understood. A major source of the increasingly high rate of change cryosphere shrinkage is the positive feedback resulting from reduction in surface reflectivity (albedo) due to loss of ice and snow. This reduction in albedo increases the capacity to absorb and store heat in Arctic ecosystems, enabling more melting. In the face of projections for continued global warming, further losses and changes in timing and other characteristics of ice and snow in their various forms are expected (AMAP, 2011; Jeffries et al., 2012).

Top of Page

Permafrost

The sections below on permafrost trends are based on Trends in permafrost conditions and ecology in northern Canada (Smith, 2011), an ESTR national thematic report. Some information has been updated, as noted.

Information on recent trends in permafrost temperatures in the Arctic Ecozone+ comes from a number of monitoring sites from the Western Arctic to the Eastern Arctic and the High Arctic. In general, changes in shallow permafrost temperatures over the last decade are greater in the Arctic Ecozone+ compared to areas below the treeline (taiga and boreal ecozones+). Tundra vegetation, with less of a vertical structure than forest, allows more direct light penetration to the soil and provides less insulation against re-radiation of absorbed heat. The presence of colder permafrost also means that phase change and the presence of unfrozen water do not obscure the climate signal. There is, therefore, a more direct link between changes in air temperature and changes in permafrost temperature.

Western Arctic

Permafrost temperature data collected since the 1980s from the northern Mackenzie Basin (Smith et al., 2005b) indicate that warming of permafrost has occurred since the early 1990s. On the Tuktoyaktuk Peninsula for example, at a depth of 28 m, permafrost temperatures increased between 1990 and 2002 at a rate between 0.02 and 0.06°C per year. Analysis by Burn and Kokelj (2009) indicates that near-surface ground temperatures in the tundra uplands of the Mackenzie Delta region increased by 1 to 2°C from the early 1970s to 2007. Modelling analysis conducted for a permafrost monitoring site at Herschel Island (in the Beaufort Sea off the Yukon coast) indicates that permafrost temperature at a depth of 20 m has increased by 1.9°C over the past 100 years (Burn and Zhang, 2009). Recent field observations at this monitoring site also indicate an increase in active layer thickness since 1985.

Central Southern Arctic

Permafrost temperatures to 3 m depth have been collected since 1997 at Baker Lake, Nunavut. Between 1997 and 2007 a general increase in thaw depth (Figure 19) has been observed, although there is some interannual variability within the short record (Smith et al., 2005b; Throop et al., 2008). The largest increase in thaw depth occurred between 1997 and 1998 and this was related to the longer thaw season in 1998 (Smith et al., 2001b).

Figure 19. Maximum summer thaw depth for a site at Baker Lake.
Graph
Source: Smith (2011), adapted and updated from Smith et al. (2005b) and Throop et al. (2008)
Long description for Figure 19

This bar graph shows the maximum summer thaw depth for a site at Baker Lake for the years 1997 to 2007.  Between 1997 and 2007 a general increase in thaw depth has been observed, although there is some interannual variability within the short record. The largest increase in thaw depth--approximately 0.9 m--occurred between 1997 and 1998.

Top of Page

Eastern Arctic

Permafrost temperature data collected since 1978 at CFS Alert Nunavut can be used to characterize trends in permafrost in the High Arctic. Although air temperatures have increased since the 1980s, distinct warming in shallow permafrost temperatures has only been observed since the mid-1990s. Between 1994 and 2001, an increase in permafrost temperatures of about 0.15°C per year occurred at a depth of 15 m (Smith et al., 2005b). Although some cooling of permafrost was observed between 2000 and 2002, more recent data indicates that warming of permafrost is continuing.

Figure 20 shows trends in annual mean temperatures at two depths and two boreholes up to 2011. Increases in winter temperatures appear to be the main factor in the recent ground temperature increases, especially at High Arctic sites with relatively low snow cover (and thus little insulation of the ground during winter) (Smith et al., 2012).

Figure 20. Mean annual ground temperatures at two boreholes and two depths at Alert, Nunavut, 1978 to 2011.

Data standardized and data gaps addressed following methods in Smith et al. (2012).

Graph
Source: adapted from Romanovsky et al. (2012)
Long description for Figure 20

This line graphs shows trends in the mean annual ground temperatures at two boreholes and two depths at Alert, Nunavut for the years 1978 to 2011. Between 1994 and 2001, an increase in permafrost temperatures of about 0.15°C per year occurred at a depth of 15 m. A smaller increase in permafrost temperatures was also observed at a depth of 24 m.

Top of Page

Shallow permafrost temperatures at Iqaluit also began to warm in 1993, with warming continuing through the 1990s (Figure 21). Temperatures of shallow permafrost (measured at a depth of 5 m) increased at a rate of 0.4°C per year between 1993 and 2000. A similar trend has been observed in northern Quebec, where cooling of about 0.10°C per year was observed between the mid-1980s and mid-1990s at a depth of 10 m (Allard et al., 1995). An increase in air temperatures commencing in 1993 in northern Quebec has been associated with warming of permafrost since 1996 to depths of 20 m (Allard et al., 2002; Ouranos, 2004; Chouinard et al., 2007) and an increase in active layer thickness (Brown et al., 2000).

Figure 21. Monthly ground temperature at a depth of 5 m at Environment Canada’s borehole at Iqaluit, 1988 to 2002.

The 12-month running means for ground and air temperatures are also shown.

Graph
Source: Smith (2011), adapted from Smith et al. (2005b)
Long description for Figure 21

This line graph shows the monthly ground temperature at a depth of 5 m at Environment Canada’s borehole at Iqaluit, 1988 to 2002. The graph also displays the running mean ground temperature and the running mean air temperature.  The graph demonstrates that shallow permafrost temperatures at Iqaluit also began to warm in 1993, with warming continuing through the 1990s. Temperatures of shallow permafrost (measured at a depth of 5 m) increased at a rate of 0.4°C per year between 1993 and 2000.

Top of Page

Recent increases in shallow ground temperatures have also been observed in other Arctic regions such as Scandinavia and Svalbard (Isaksen et al., 2007a; Isaksen et al., 2007b; Harris and Isaksen, 2008). Although snow cover is generally thin at these High Arctic sites, its variability can be an important factor affecting the response of permafrost temperatures to changes in air temperature (Smith et al., 2003). Changes in snow cover may counteract changes in air temperature occurring over the same period such that permafrost temperatures may increase in the High Arctic during periods of higher snow cover but lower air temperature (Taylor et al., 2006).

Ecological consequences of changes in permafrost

The observed trends of warming permafrost and increased thickness of the active layer are projected to continue and accelerate as the Arctic climate warms (ACIA, 2005; Intergovernmental Panel on Climate Change, 2007). The implications for ecosystem processes and ecosystem structure are huge. Melting of permafrost is also implicated in feedback loops that: 1) lead to increases in air temperatures in the region through changes in vegetation communities and albedo (Chapin III et al., 2005); and 2) lead to increases in warming at the global scale through changes in the carbon balance of the Arctic landscape. Research provides clear links between permafrost conditions and many ecosystem characteristics. Though it can be hard to demonstrate and measure trends in these interlinked impacts at the level of change to populations and ecosystem services, the observed trends in permafrost are likely now influencing Arctic ecosystem processes, structure, and composition. Warming and melting of permafrost projected for the decades to come will be accompanied by major ecosystem change. The discussion below outlines some of the ways in which permafrost drives and influences ecosystem processes. See also the section on Permafrost disturbance (page 54).

Patterned ground caused by permafrost provides diversity in microclimates for Arctic plants and arthropods (Strathdee and Bale, 1998; Vonlanthen et al., 2008; Walker et al., 2008a). The annual freeze-thaw cycles of the active layer lead to ground patterns such as frost boils and ice-wedge polygons which are a defining feature of the tundra. Frost boils are patches of barren or sparsely vegetated soil where differential frost heave forms distinct patterns of stones and sediments (Walker et al., 2004). Frost boils cycle organic matter and nutrients vertically within the boil (Walker et al., 2004) and are favourable sites for seed germination (Sutton et al., 2006), contributing to the diversity and amount of vegetation available to herbivores (Walker et al., 2001).

One of the most productive tundra ecosystems is non-acid tussock tundra which is a key to migratory tundra caribou herds and their associated predators (Walker et al., 2001). The productivity is driven by the freeze-thaw cycles of the active layer which induces the migration of cations essential to plants. The near-surface ice-rich zone of permafrost is a sink for cations such as calcium and magnesium which reach the active layer and become available for plant growth. When permafrost melts, more of these nutrients are contributed to the active layer (Kokelj and Burn, 2005).

The active layer is kept moist to water-logged by underlying permafrost. Correspondingly, microbial activity is reduced and the microbial release of nitrogen from decomposing organic matter is slow, limiting plant growth (Weintraub and Schimel, 2005). Net primary productivity is related to nutrient availability and is highest on deeply thawed soil, dominated by nitrogen fixers and disturbed by animals that release nutrients.

The depth of the active layer varies from tens of centimetres to 1 to 2 metres and constrains the activities of burrowing animals in summer. Frost boils and frost cracks are preferred sites for lemming burrows (Potter, 1972), possibly because they are warmer than burrows deeper in the active layer. Lemmings spend most of their time in their burrows in summer to reduce the risk from avian predators (Boonstra et al., 1996). The burrows and associated latrines add to the micro sites of disturbed ground and enhanced nutrient availability for plant growth. Other burrowing mammals, including Arctic ground squirrels (Spermophilus parryii), select sites without permafrost, such as river banks (Batzli and Sobaski, 1980).

Permafrost is the main factor influencing where Arctic foxes (Vulpes lagopus) den. The foxes select south-facing, well-drained sites with early snowmelt, conditions indicative of a deeper active layer (Szor et al., 2008). Pregnant polar bears denning on the Hudson Bay coast select thawed peat banks to dig dens. The re-use of the dens over hundreds of years as measured by annual tree ring chronologies suggest the importance of the thawed peat banks for the bears (Scott and Stirling, 2002).

Recent techniques have revealed bacteria and fungal spores within permafrost (Steven et al., 2008). Permafrost’s frozen sediments and ice have carbon dioxide, methane, oxygen, and nitrogen in pore spaces, as well as a thin film of water, and evidence points to active microbial ecosystems within permafrost (Steven et al., 2006). This has implications for global warming. Steven et al. (2006) drew attention to “the considerable microbial biomass believed to inhabit the significant areas of terrestrial permafrost, [and] this may have important implications on global nutrient cycling and biogeochemical processes, such as the C, N, and S cycles”.

Top of Page

Snow

This section is based on analysis of ground-based snow cover measurements for Canada as a whole and for the Arctic Ecozone+ as analysed for the ESTR report Canadian climate trends, 1950–2007 (Zhang et al., 2011), with reference to additional studies, as noted.

Snow cover duration has significantly decreased almost everywhere in Canada, with the greatest changes occurring in spring over western and northern Canada. This is a consequence of warmer springs observed over these regions, and is part of a Northern Hemispheric trend toward spring warming and earlier melting of snow and ice (Lemke et al., 2007; Derksen and Brown, 2012).

Across the Arctic Ecozone+, snow cover duration decreased by an average of 9 days in the fall and 8.5 days in the spring from the 1950/51 to the 2006/07 snow seasons. Figure 22 shows the fall and spring trends in snow cover duration at each station. Annual maximum snow depth also decreased: by an average of 13.2 cm over the Arctic Ecozone+ from 1950 to 2007 (Figure 23). This is part of a general tendency to lower maximum snow depths over Canada, but is less significant and less consistent from place to place than the decreases in snow cover duration. In southern Canada, the decrease in annual maximum snow depth is linked to a trend toward a smaller fraction of annual precipitation falling as snow. However, there were no significant trends in this ratio when averaged across the Arctic Ecozone+--in fact, increases in the ratio were recorded at many stations (Figure 24), in line with the greater increase in precipitation during the snow season than during summer (Figure 16). The shorter snow season (later onset of snow cover and earlier melt) is likely offsetting the effects of this increase in cold-season precipitation. This is consistent with analysis by Brown and Mote (2009), which shows that, under a scenario of warming and increasing precipitation, maximum snow accumulation still decreases.

Figure 22. Change in snow cover duration ( the number of days with greater than or equal to 2 cm of snow on the ground), 1950–2007.

Map a (the first half of the snow season) indicates changes in the start date of snow cover. Map b (the second half of the snow season) indicates changes in the end date of snow cover.

Maps
Source: ecozone+ data provided by authors of Zhang et al. (2011)
Long description for Figure 22

This figure consists of two maps, one for each half of the snow season with icons indicating direction, magnitude, and significance of changes in snow cover duration at nine Arctic monitoring stations between 1950 and 2007.  Across the Arctic Ecozone+, snow cover duration decreased by an average of 9 days in the fall and 8.5 days in the spring from the 1950/51 to the 2006/07 snow seasons.

Top of Page

Figure 23. Change in the maximum annual snow depth, 1950–2007.
Map
Source: ecozone+ data provided by authors of Zhang et al. (2011)
Long description for Figure 23

This map shows icons indicating the direction, magnitude, and significance of changes in the maximum annual snow depth at Arctic monitoring stations between 1950 and 2007.  Across the Arctic Ecozone+, snow depth decreased by an average of 13.2 cm from 1950 to 2007.

Top of Page

Figure 24. Absolute change in the ratio of snow to total precipitation in Canada, 1950–2007.

Increasing trends indicate an increase in the proportion of precipitation falling as snow.

Map
Source: ecozone+ data provided by authors of Zhang et al. (2011)
Long description for Figure 24

This map shows icons indicating the absolute change in the ratio of snow to total precipitation in Canada for the years 1950 to 2007.  The proportion of precipitation falling as snow increased or showed increasing tendencies at the majority of sites across the ecozone+.

Top of Page

The Arctic snow depth trend results include greater uncertainty than snow cover duration because: 1) snow depths exhibit much stronger local-scale variability than snow cover duration (Brown et al., 2007b); and 2) the snow depth observing network in the Canadian Arctic is sparse and biased to coastal locations. In addition, snow depth measurements are made at open sites near airports that may not be representative of surrounding terrain. With increasing shrubbiness in the Arctic, one would expect to see more snow being retained on the ground in areas of shrub expansion. Snow cover duration, however, is less affected by station conditions.

Circumpolar snow trends

The greatest and most rapid decreases in snow water equivalent and snow cover duration have occurred in high-precipitation maritime regions of the Arctic (Callaghan et al., 2011a). There are also differences between Eurasia and North America: declines in snow depth and snow cover tend to be more recent in Eurasia, mainly after about 1980, while declines are evident since the 1950s in North America. Snow depth, however, is increasing in parts of Eurasia (Callaghan et al., 2011a). Arctic spring snow cover in both Eurasia and North America has decreased greatly, especially over the past decade (Derksen and Brown, 2012) (Figure 25).

Figure 25. Trends in terrestrial spring snow cover for the Eurasian and the North American Arctic, 1967–2012.

Data are June averages for all Northern Hemisphere snow cover, which is virtually all in the Arctic. Values are standardized anomalies with respect to the 1988–2007 mean. Solid lines are five-year running means..

Graph
Source: Conservation of Arctic Flora and Fauna (2013) based on Derksen and Brown (2012), data from NOAA CDR (satellite monitoring)
Long description for Figure 25

This scatter plot and line graph shows the trends in terrestrial spring snow cover for the Eurasian and the North American Arctic for the years 1967 through 2012. The graph shows the June snow cover extent by year.  Arctic spring snow cover has decreased greatly over the period, especially from 2002 to 2012.

Top of Page

Ecosystem consequences of changes in snow

A reduction in duration of snow cover has been shown to have significant positive feedbacks to the Arctic climate system through reduced albedo (Chapin III et al., 2005; Screen and Simmonds, 2012). This mechanism is stronger over tundra than over taiga because of the shallow snowpack, high amounts of incoming solar radiation in the spring melt period, and the characteristics of the vegetation. Tundra vegetation has little vertical structure and therefore little capacity to shade the ground. Chapin et al. (2005) concluded that reductions in the length of the snow-cover period in Alaska over the past few decades have contributed substantially to local atmospheric heating in the summer--increasing heating by about 3 Watts per m2 per decade. Across the entire Arctic region, feedback from changes in snow cover over the period 1970–2000 was simulated to have increased atmospheric heating by 0.9 Watts per m2 per decade (Euskirchen et al., 2007). Natural and anthropogenic impurities in snow are also contributing to reduced albedo. Flanner et al. (2008) show that the surface deposition of black carbon (soot particles) reduces snow surface albedo (a process termed “darkening”), which contributes to melting of snow and ice and increases atmospheric heating.

Snow characteristics, including extent, depth, duration, and timing--but also the characteristics of the snow pack determined by events such as freeze-thaw cycles--strongly influence Arctic terrestrial and freshwater ecosystem processes including permafrost dynamics, hydrology, and primary production. Snow characteristics also directly affect biotic communities and many Arctic species. Observed and predicted effects from changes in snow are assessed in recent Arctic Council projects (AMAP, 2011; Callaghan et al., 2011b; CAFF, 2013) and discussed in several sections of this report.

Top of Page

Sea ice

Status and trends of sea ice are discussed more fully in the Arctic Marine report (Niemi et al., 2010). A summary is provided here because of the importance of sea ice to some of the animals discussed in this report (notably polar bears) and because of the implications to Arctic terrestrial ecosystems of the major changes now occurring in extent and quality of sea ice.

Sea ice extent throughout the year has decreased significantly over the period of record, as measured by remote sensing (Figure 26). The summer minimum ice cover (occurring in September) has declined particularly rapidly in the past few years (Figure 27). Extending the record further back, using ice observation networks and proxy data, indicates that the recent and current rate and extent of decline of summer ice extent is unprecedented at least over the past 1,450 years (Kinnard et al., 2011). Increased summer ice melting has led to a loss of multi-year ice. The percentage of the winter (March) ice cover composed of first-year ice increased from 58% in 1988 to 75% in 2012, while the percentage of ice aged four or more years decreased from 26% in 1988 to 7% in 2012 (Perovich et al., 2012).

Climate modellers project ice-free summers in the Arctic within 30 years (Wang and Overland, 2012). During an ice-free summer, there would still be remnants of land-fast ice and likely ice in sheltered waters between islands of the Canadian Arctic Archipelago (Wang and Overland, 2012). The spatial distribution of sea ice in winter and summer 2012 is shown in Figure 28.

Figure 26. Extent of Arctic sea ice over the annual cycle, recent years compared to 1979–2000 average.
Graph
Source: updated from Fetterer et al. (2010) with data from National Snow and Ice Data Center (2013)
Long description for Figure 26

This line graph shows the following information:

Extent of Arctic sea ice over the annual cycle, recent years compared to the 1979–2000 average. Extent (million km2)
Month1979-2000 Avg20072008200920122013
Jan.14.813.7714.0414.0713.7313.78
Feb.15.614.5314.9814.8414.5614.66
Mar.15.714.6515.2115.1615.2115.04
Apr.1513.8714.4214.5714.7314.37
May13.612.8913.1813.3913.1313.1
Jun.12.211.4911.3611.4610.9211.58
Jul.10.18.138.998.87.938.45
Aug.7.75.366.056.284.716.09
Sep.7.044.34.735.393.635.35
Oct.9.36.778.427.527.07-
Nov.11.310.0510.6210.279.92-
Dec.13.412.3912.5212.5112.2-

Top of Page

Figure 27. Trends in mean September sea ice cover, 1979–2013, showing the recent rapid decline.

The September 2012 minimum ice extent was 48% below the 1979–2000 average. The cooler summer of 2013 resulted in a slightly greater ice minimum extent, similar to 2009 and 24% below the 1979–2000 average.

Graph
Source: updated from Fetterer et al. (2010) with data from National Snow and Ice Data Center (2013)
Long description for Figure 27

This line graph shows the following information:

Trends in mean September sea ice cover, 1979–2008, showing the recent rapid decline.
YearExtent (million km2)
19797.2
19807.85
19817.25
19827.45
19837.52
19847.17
19856.93
19867.54
19877.48
19887.49
19897.04
19906.24
19916.55
19927.55
19936.5
19947.18
19956.13
19967.88
19976.74
19986.56
19996.24
20006.32
20016.75
20025.96
20036.15
20046.05
20055.57
20065.92
20074.3
20084.73
20095.39
20104.9
20114.63
20123.63
20135.35

Top of Page

Figure 28. Distribution of winter and summer sea ice, 2012, and median ice edge for both seasons.
Maps
Source: Eamer et al. (2013) based on data from National Snow and Ice Data Center (2013)
Long description for Figure 28

This figure shows two maps of the northern polar region depicting the winter and summer sea ice distribution in 2012. The maps have an outline comparing the 2012 ice distribution to the median ice edge for 1979–2000.  The winter sea ice distribution is smaller in area than the median ice edge for 1979–2000 in the Barents Sea, and shows smaller fluctuations in other areas.  The summer sea ice distribution is substantially smaller than the 1979-2000 median.

Top of Page

Ecological consequences of rapidly changing sea ice

Major consequences of this regime shift in Arctic sea ice for Canada’s Arctic marine ecosystems are covered in the ESTR report Ecosystem status and trends report: Arctic marine ecozones (Niemi et al., 2010). Consequences at the circumpolar and global scale were assessed through recent Arctic Council projects (AMAP, 2011; CAFF, 2013; Eamer et al., 2013).

The extent and quality of sea ice cover has major implications for terrestrial ecosystems and wildlife, as well as for people. Open water has a much lower albedo (reflectivity) than ice, and consequently absorbs more sunlight. Reduction of sea ice has led to warming in adjacent coastal areas (Serreze and Barry, 2011) with consequences for tundra vegetation (Bhatt et al., 2010), as well as increased coastal erosion, affecting coastal habitats and coastal communities, especially along the Beaufort Sea coast (Lantuit and Pollard, 2008; AMAP, 2011). Sea ice is a major platform for travel and hunting for Inuit, and changes over the past decade have affected communities (Niemi et al., 2010; Eamer et al., 2013).

While the polar bear is the species (of those covered in this report) most clearly at risk from the decline of sea ice, terrestrial animals are also vulnerable. This discussion below on sea ice and tundra ecosystems is based partly on the Conservation of Arctic Flora and Fauna (CAFF) report Life linked to ice: A guide to sea-ice-associated biodiversity in this time of rapid change (Eamer et al., 2013).

Winter sea ice provides foxes and wolves with a means to access and colonize remote islands. A study of the genetics of Arctic foxes throughout their circumpolar range showed that the occurrence of sea ice is the main factor in determining how similar fox populations are to one another: the less sea ice, the more genetically distinct are the populations (Geffen et al., 2007). If island populations become isolated due to loss of winter ice, some populations will be at risk of decline or extinction due to loss of genetic diversity and inbreeding (Geffen et al., 2007; Noren et al., 2011). Wolf (Canis lupis) populations on Banks, Ellesmere and Devon islands have declined and been reestablished in the past through colonization over the sea ice from other islands (Carmichael et al., 2008). As with foxes, wolf populations are at risk of reduction if winter ice conditions alter sufficiently to compromise movement over ice. As winter sea ice is expected to persist in the region of the Canadian Arctic Archipelago, this potential impact is more likely to occur in other circumpolar regions, including parts of the European and Russian Arctic.

Changes in sea ice also affect the movements of some caribou herds. Peary caribou have been observed moving from island to island over the sea ice (Jenkins et al., 2011). The Dolphin and Union Herd migrates annually across the Dolphin and Union Strait between Victoria Island and the mainland, one of the areas where the expansion of open water has been most extensive (COSEWIC, 2004). Late freeze-up of the strait will delay migration to the winter ranges.

Both Arctic foxes and wolves feed on marine resources during winter, travelling long distances to hunt and scavenge (Tarroux et al., 2010). Arctic foxes scavenge seals killed by polar bears (Roth, 2003; Geffen et al., 2007) and may also prey on ringed seal pups, as observed on sea ice off the coast of northern Labrador (Andriashek et al., 1985). They may rely heavily on marine foods during years of low lemming abundance (Tarroux et al., 2012). This suggests that the loss of sea ice could affect coastal populations of Arctic foxes during years of low terrestrial food abundance (Pamperin et al., 2008). Wolves are also known to forage on sea ice in winter, likely scavenging seal carcasses, based on an Alaskan study (Watts et al., 2010). The main changes in sea ice and their potential impacts on these primarily terrestrial carnivores are: 1) delayed onset of ice cover in autumn, leading to the carnivores having to spend longer on land in what could be a relatively food-poor situation (for example, when rodents are at low densities and birds have left); and 2) reduced summer ice cover leading to reduced polar bear populations and possibly reduced seal populations, and therefore reduced density of scavenging or hunting opportunities for the terrestrial carnivores when they are on the ice.

Recent studies using satellite tracking show that both gyrfalcons (Burnham and Newton, 2011) and snowy owls (Therrien et al., 2011) spend substantial periods of time far offshore in winter, presumably preying on marine birds and roosting on ice floes. Snowy owls have been observed hunting seabirds that congregate in patches of open water in Hudson Bay in winter (Gilchrist and Robertson, 2000). Satellite tracking over two winters (2007–2009) of nine adult female snowy owls that had been fitted with transmitters while on their summer range on Bylot Island showed that most spent several weeks on the sea ice between December and April. Analysis of high-resolution satellite images showed that they spent most of their time on the ice around open-water patches frequented by seabirds (Therrien et al., 2011). Changes in winter sea ice extent and polynya formation that affect seabirds could also affect the food intake of snowy owls. As owls are small-mammal predators during breeding season, changes in snowy owl populations would affect the tundra food web (Therrien et al., 2011).

Top of Page

Glaciers and ice caps

During the past 10,000 years, glaciers and ice caps--the remnants of the great ice sheets of the last ice age--have been melting, with intermittent periods of re-advance. The last major advance was during the Little Ice Age from the 14th to 19th centuries. Since then, glaciers around the globe have been shrinking, with increasing rates of ice loss since the mid-1980s. On a time-scale of decades, glaciers in some regions (including the Arctic) have shown intermittent advances, but overwhelmingly the trend is to increasingly rapid melt. The outlook, based on climate change scenarios, is for deglaciation of many mountain regions within the 21st century (Gerland et al., 2007).

About 300,000 km2 of glaciers and ice caps (excluding Greenland) drain to the Arctic Ocean, most directly (Dyurgerov and Carter, 2004). Over half of this areal extent of land ice is in Canada’s Arctic Archipelago (Table 3). These glaciers and ice caps fluctuate in mass and extent with temperature and snowfall (UNEP and WGMS, 2008; Colgan and Sharp, 2008), with the strongest linkage being between mass balance and air temperature (Koerner, 2005). A second important control on extent and volume changes of land ice has been iceberg calving that occurs as flux across the grounding line and retreat/advance of tidewater margins. Iceberg calving accounts for a large portion (up to 40%) of mass loss from ice caps in the Canadian Arctic (Burgess et al., 2005; Mair et al., 2009; Mair, 2012).

Glaciers in the Canadian Arctic Archipelago reached their last maximum extent relatively late: in the latter part of the 19th century. Time series of glacier fluctuations are sparse, with the only long-term mass balance series being for the White and Baby glaciers and the Devon, Meighen, and Melville ice caps and Agassiz Ice Field, all in the Queen Elizabeth Islands. Based on these time series and other evidence, there has been a general melting trend for Canadian Arctic Island glaciers since the late 19th century, with the trend slowing down for a period in the mid-20th century (UNEP and WGMS, 2008), and with accelerated loss in the past 25 years. Analysis of ice cores from the Devon Ice Cap (Devon Island), Penny Ice Cap (Baffin Island), and the Agassiz and Prince of Wales ice fields (Ellesmere Island) indicates that the melt rates since the mid-1990s are the highest in at least the last two millennia (Fisher et al., 2012). The longer record for the Agassiz Ice Field shows that this recent melt rate is the highest in 4,200 years and resembles the melt rate during the Holocene thermal maximum over 9,000 years ago (Fisher et al., 2012).

Table 3. Surface area of glaciers and ice caps, Canadian Arctic Archipelago, mid-1970s. Total surface glacier area for Arctic Archipelago: 151,800 km2
Islandkm2Islandkm2
Ellesmere80,000Coburg225
Ellesmere Ice Shelf500Meighen85
Axel Heiberg11,700Melville160
Devon16,200North Kent152
Bylot5,000Others736
Baffin37,000--

Top of Page

Figure 29. Annual net surface mass balance of four glaciers in the Queen Elizabeth Islands, 1960 to 2011.
Graph
Source: Sharp et al. (2012) based on data from the World Glacier Monitoring Service
Long description for Figure 29

This line graph shows the annual net surface mass balance of four glaciers in the Queen Elizabeth Islands for the years 1960 to 2011.  The four glaciers are the Devon Ice Cap, the Meighen Ice Cap, the Melville Ice Cap and White Glacier.  The graph shows no trend up until 1985 followed by a period of decreasing annual net surface mass balance for all four glaciers from 1985 to 2011.

Top of Page

On Baffin Island, changes in the extent of a few small plateau ice caps (currently less than 50 km2 in size) have been measured since 1975. This record has been extended back in time through the use of chemical and vegetation studies (Miller et al., 1975; Anderson et al., 2008). Some of these ice caps have disappeared already and researchers predict that all small ice caps will have disappeared completely by 2070. Figure 30 shows the measured and inferred rate of decrease in extent since 1940.

Figure 30. Location map and ice cap retreat on Baffin Island, 1949 to 2005.

a) Landsat ETM+ satellite image, 10 August 2000, showing present-day ice caps. The white line delineates approximate extent of vegetation kill zone, corresponding to maximum ice-cap cover during the Little Ice Age. The black circle identifies the location of 14C dates of 5.6 ka on aquatic moss in the basal sediment of a lake.

(b) Ice cap areas between 1949 and 2005. Corresponding ice caps are marked by symbols in (a). Linear regressions project approximate the dates of disappearance of the ice caps.

(c) Areal decrease of the Tiger Ice Cap, 1949–1988, which disappeared in summer 1992.

Graph and Map
Source: Anderson et al. (2008)
Long description for Figure 30

This figure shows a map of an area on Baffin Island, a linear regression graph for ice cap area for five different ice caps on the island, and a small map of ice cap retreat for the Tiger Ice Cap.  The map shows the present-day ice caps and shows an outline which delineates approximate extent of vegetation kill zone, corresponding to maximum ice-cap cover during the Little Ice Age. The linear regression graph shows the measured and inferred rates of decrease in extent since 1940.  The linear regressions show that icecap size declined over the time period for all five ice caps.

Top of Page

Mass loss of glaciers and ice caps due to surface melt and iceberg calving contributes significantly to sea level rise. As the Canadian Arctic Archipelago is the largest glaciated area in the world outside of Greenland and Antarctica, the changes in Arctic Ecozone+ glaciers and ice caps are of global significance.

A study published in 2009 estimates that glaciers and ice caps around the world will contribute 373 ± 21 mm of sea-level rise over the next 100 years, nearly half of the projected rise of 800 mm for the period (Bahr et al., 2009). An analysis of glacier mass balance based on both ground studies and remote sensing (Gardner et al., 2013) estimated that, between 2003 and 2009, glaciers and ice caps in northern Arctic Canada experienced a mass balance loss of a 33 ± 4 Gt/year and southern Arctic Canada glaciers lost 27 ± 4 Gt/year. The total for Arctic Canada is about 28% of the global glacier mass loss (excluding Antarctica and Greenland).

About half of sea-level rise is due to thermal expansion of the ocean. The next biggest source is meltwater from the Greenland and Antarctic ice sheets, followed by meltwater from the world’s smaller glaciers and ice caps (IPCC, 2007). Although the contribution from the smaller glaciers and ice caps will, in the long run, be far less than from the Greenland and Antarctic ice sheets, they react faster to changes in climate and have more immediate effects (Bahr et al., 2009). From 2003 to 2009, global glacier mass loss, excluding Antarctica and Greenland, would have contributed 0.71 ± 0.08 mm/year to sea-level rise, which is about 29% of the observed sea level rise over this period (Gardner et al., 2013).

One of the major ecosystem changes will be the increase in land available for vegetation and the development of tundra ecosystems. This is a process that has been occurring since the late 19th century. The actual area of newly exposed land is not currently large, but could become increasingly important as ice masses disappear from the High Arctic islands (see Table 3 for areas covered). Areas exposed since the glacial maximum at the end of the Little Ice Age (about 1870) are visible as lichen-free zones near the glaciers. Satellite imagery of the Queen Elizabeth Islands (which includes the glaciers shown in Figure 29) shows that the area of terrestrial ice in the islands has decreased by 37% (62,387 km2) since the last glacial maximum (Wolken et al., 2008). A study of ecosystem development in front of a glacier at Alexandra Fiord that has retreated over 300 m since 1980 described patterns of fairly rapid vegetation succession, with diverse plant communities being established, but with sparse plant cover after over 40 years (Jones and Henry, 2003). Areas achieving over 80% plant cover had been ice-free for over 300 years.

Top of Page

Lake and river ice

The duration of Arctic lake and river ice is determined by the dates of autumn freeze-up and spring break-up. In general, rivers experience much more dynamic events than those on lakes, as the formation, growth, and ablation of lake ice primarily occurs in situ, while that on rivers is controlled by the hydraulic effects of flowing water. Overall, however, the timing of such events on both systems is strongly controlled by climate (Prowse et al., 2011).

Analysis of changes in the remains of algae and invertebrates buried in lake sediments has been used by many researchers to identify warming trends and the historical presence/absence of ice cover on northern lakes (Douglas et al., 1994; Korhola et al., 2002; Sorvari et al., 2002; Michelutti et al., 2003; Ruhland et al., 2003; Smol et al., 2005), based on the relationships between ice-cover duration, growing season length, and annual primary production (Rouse et al., 1997; Douglas and Smol, 1999; Keatley, 2007). In general, around the circumpolar Arctic, this evidence points towards warming temperatures and shorter ice durations since the end of the Little Ice Age (around 1850), with greater changes observed in northernmost areas when compared to more temperate locations (Smol et al., 2005). (See section on Changes in algal and invertebrate species assemblages in lakes and ponds on page 96.)

In-situ monitoring datasets are not long enough to evaluate trends in lake and river ice in the Arctic Ecozone+ and remote sensing provides an alternative. Latifovic and Pouliot (Latifovic and Pouliot, 2007) used ground-based monitoring datasets and remote sensing records to test comparability of the methodologies and examine trends in lake ice for 36 lakes across Canada (Latifovic and Pouliot, 2007). In doing so they developed records based just on remote sensing for six Arctic lakes. Break-up trends are shown in Figure 31 A and freeze-up trends in Figure 31 B. They concluded that, based on this limited set of lakes examined, the ice-free period increased significantly for lakes across all of Canada, but that the changes were of greater magnitude in the Arctic, at least in the past 20 years. Rates of change are shown in Figure 31 C.

Figure 31. Ice break-up and freeze-up trends for six Arctic lakes, 1985–2004, and average trends compared with average trends for Canadian lakes south of the Arctic Ecozone+.

Break-up and freeze-up dates are based on remote sensing as part of a larger study for all of Canada that found good agreement between ground-based records and the remote sensing methodology used. There were no ground-based records for the analysis for the Arctic lakes. Lakes are (west to east): Sitidgi, Washburn, Stanwell-Fletcher, Hazen, Angajurjualuk, and Amadjuak.

A. Break-up: only the trend for Lake Hazen was statistically significant (p<0.1)

B. Freeze-up: trends for 3 of the 6 lakes were statistically significant (p<0.1)

C. Comparison with lakes south of the ecozone+: rates for the 30 lakes south of the Arctic were based on ground-based monitoring and remote sensing. Note that, because of the different time periods, the difference in rates of change reflects both the influence of greater warming in the past 2 decades and the effect of higher latitude.

Graph and Maps
Source: adapted from Latifovic and Pouliot (2007)
Long description for Figure 31

This figure shows two maps of the Arctic Ecozone+ with icons representing the timing of ice break-up and ice freeze-up for six Arctic lakes. The figure also includes a bar graph that shows the following information:

Ice break-up and freeze-up trends for six Arctic lakes, 1985–2004, and average trends compared with average trends for Canadian lakes south of the Arctic Ecozone+. (Days per year)
Ice break-upAll Canada
(1970-2004)
Arctic
(1985-2004)
Earlier break-up0.230.99
Later freeze-up0.160.76
Longer ice-free period0.391.75

Top of Page

Rivers and lakes

Trends in river discharge

Changing river discharge is a consequence of changing precipitation patterns, as well as changing temperature patterns that affect water storage features such as soil moisture, permafrost depth, and glacier melt. Arctic residents have noticed changes in river flows and water levels that may reflect climate change. For example, residents in Chesterfield Inlet reported a decline in water levels (rivers, streams, ponds, and lakes) by as much as a metre (Nunavut Research Institute, 2004). They observed that spawning charr had more scratches and bruises caused by moving through shallow waters.

An assessment of Canada’s lake and river systems undertaken for this report, reported in the ESTR thematic technical report Biodiversity in Canadian lakes and rivers (Monk and Baird, 2011) and in Monk et al. (2011), examines status and trends across the country and by ecozone+ where there are sufficient long-term datasets. Only two hydrometric stations in the Northern Arctic and three in the Southern Arctic were of sufficient consistency and length to be included in the analysis of river flows. No stations were included from the Arctic Cordillera. Figure 32 shows the distribution of stations across all ecozones+ and trends in one-day minimum runoff across Canada. Trends are generally grouped regionally and increases in minimum runoff are evident in two out of the five stations within the Arctic Ecozone+. The same analysis was conducted for one-day maximum runoff, resulting in less clear regional trends across the country and no consistency in trends in the few Arctic stations (figure not shown).

Figure 32. Map showing trends in 1- day minimum river flow in natural rivers across Canada, using data for hydrological years 1970–2005.
Map
Source: Monk and Baird (2011)
Long description for Figure 32

This figure is a map which shows trends in 1-day minimum river flow in natural rivers across Canada, using hydrological data for the years 1970–2005.  The map has icons representing monitoring stations and showing whether they recorded a significant increasing trend, increasing tendency, significant decreasing trend, decreasing tendency, or no trend.  Only five hydrometric stations in the Arctic Ecozone+ were of sufficient consistency and length to be included in the analysis, two in the Northern Arctic and three in the Southern Arctic. Trends are generally grouped regionally and increases in minimum runoff are evident in two out of the five stations within the Arctic Ecozone+, both in the Southern Arctic west of Hudson Bay.

Top of Page

With few long-term data sets for the smaller rivers that are wholly or mainly within the Arctic Ecozone+,  little is known about trends in discharge, especially in the Arctic Archipelago. An analysis conducted as part of International Polar Year research extended the sparse hydrometric records for the Arctic Archipelago through extrapolation of observed streamflow attributes to provide estimates of current discharge. The uncertainty associated with the estimates is high because of the sparse data coverage. The authors concluded that the current freshwater flux for the region, and if or how it is changing, are unknown (Spence and Burke, 2008). They highlight the importance of this information for domestic water management decisions and for meeting international obligations to quantify freshwater inputs to the oceans.

An evaluation of hydrometric monitoring for 76 river systems across the Canadian Arctic (1950 to 2008) was undertaken by Mlynowski et al. (2011). The peak in Arctic land area that was monitored through streamflow gauges was 64% in 1990, decreasing to 56% in 2008. Larger river systems typically had the most data available. The peak in record length was in 1998, when 22 river systems had more than 30 years of continuous records. The authors concluded that major improvements were needed in Arctic hydrometric monitoring, especially for the relatively small rivers.

About three-quarters of the landmass of Canada is drained by rivers discharging into the Arctic Ocean, Bering Strait, Hudson/James Bay, and the Labrador Sea (Déry and Wood, 2005). This flow involves almost half (47.9%) of the total discharge of Canadian rivers (Monk and Baird, 2011). Thus hydrological processes and aquatic ecosystem health of rivers that cross briefly through the Arctic Ecozone+ at the far northern end of their courses are influenced most strongly by climatic conditions, terrain, and stressors in ecozones+ to the south. A prime example is Canada’s largest river basin, the Mackenzie River, which drains a total area of 1,787,000 km2, or 20% of the nation’s area (Culp et al., 2005). The river collects drainage from a number of other important rivers, including the Athabasca, Peace, Liard, Slave, Arctic Red and Peel, before draining through the Mackenzie Delta in the Southern Arctic (The Atlas of Canada, 2008a; The Atlas of Canada, 2008b).

Results of trend analyses are very dependent on the subset of stations and the time period used. Long-term trends are further masked by decadal-scale variations related to climate oscillations. Déry and Wood (2005), for example, found significant links between the Arctic Oscillation, El Niño/Southern Oscillation, and Pacific Decadal Oscillation with total annual freshwater discharge of rivers in northern Canada. Results from recent analyses of trends in discharge of Canadian rivers draining to various combinations of high-latitude oceans are shown in Table 4. Three studies that examined trends to the early 2000s (Déry and Wood, 2005; Déry et al., 2005; McClelland et al., 2006a) showed significant decreases in streamflow. Examination of these trends by regional river basins, however, shows that the decreases were more pronounced in the Hudson Bay and Labrador Sea drainages and that there was no significant trend from the 1960s to the early 2000s for rivers draining directly to the Arctic Ocean (Déry and Wood, 2005). Analyses that include data into the 2000s show a reversal of earlier declines in discharge, including in Hudson Bay, and a significant increase in annual average flows since 1989 (Déry et al., 2009; Déry et al., 2011).

Table 4. Summary of recent research articles exploring statistical trends in streamflow and runoff in Canadian rivers draining to northern latitude oceans.
Study catchmentAnalysis periodResults / direction of trendsReference
64 hydrometric sites in Canada draining to high-latitude oceans1964–2003
  • Significant decreased trend in total annual freshwater discharge leading to a 10% decrease in the total annual discharge to the Arctic and North Atlantic Oceans.
  • Broken down by river basin, the declining trend was strongly influenced by reduced flows to Hudson Bay and the Labrador Sea.
Déry and Wood (2005)
45 rivers in northern Canada1964–2007
  • Last two decades (1989–2007) showed significant (15.5%) increase in annual average flows.
  • Increasing variability in streamflow in most drainage regions
Déry et al. (2009)
56 rivers across North America (14 flow into Arctic Ocean and 42 flow into Hudson, Ungava and James bays)1964–2000
  • Discharge to Arctic Ocean decreased from sites in North America.
  • Discharge from sites draining Hudson, Ungava and James Bays decreased by about 2.5km3/y/y during 1964-2000.
  • Suggest concomitant decreases in precipitation and river discharge.
McClelland et al. (2006a)
42 rivers draining into Hudson, James and Ungava bays1964–2000
  • Trends of decreasing annual discharge for 36 out of 42 rivers.
  • Total annual freshwater discharge in 2000 into Hudson, James and Ungava Bays decreased by 13% from its value in 1964.
  • Peak discharge rate associated with snowmelt advanced by 8 days and diminished in intensity.
  • Spring freshet varied by 5 days for each degree of latitude.
Déry et al. (2005)
23 rivers draining to Hudson and James bays1964–2008
  • No detectable overall trend in total discharge to Hudson Bay, but a decline up to the mid-1980s, followed by a period of relatively high flows, then an upward trend to the end of the study period.
  • A shift in seasonality, with a detectable increase in winter discharge and a detectable decrease in summer discharge from 1964 to 2008.
Déry et al.
(2011)

Top of Page

Circumpolar perspective

Freshwater discharge to these northern seas is significant not just in terms of the terrestrial/freshwater ecosystems in Canada’s North but also to ocean processes, including sea ice dynamics and thermohaline circulation, which are strong influences on regional and global climates and of increasing interest as key components in understanding climate change. Piecing together the trends around the Arctic Ocean and subarctic seas puts both the Canadian trends and the importance of Canadian monitoring into perspective. An analysis of freshwater sources to all high latitude oceans (Peterson et al., 2006) concluded that there was a 5.3% increase in river discharge to the Arctic Ocean and an 8.0% decrease in river discharge to Hudson Bay in the 1990s relative to a 1936–1955 baseline. There was a shift in the late 1960s to early 1970s, marking the start of increasing pan-Arctic river discharge to the Arctic Ocean and decreasing river discharge to Hudson Bay. This increase in discharge to the Arctic Ocean is strongly influenced by annual discharge from the six largest Eurasian rivers, which increased by 7% from 1936 to 1999 (Peterson et al., 2002) and reached record highs in 2002 (Peterson et al., 2006) and 2008 (Figure 33).

These analyses include rivers affected by hydroelectric projects, including rivers discharging to James and Hudson Bay. Hydroelectric dams and diversions have profound impacts on annual discharge patterns and water quality parameters like temperature and salinity, but do not have major effects on trends in total annual discharge. Changes in annual discharge related to periods when reservoirs were filling were accounted for in the analyses (Déry and Wood, 2005; McClelland et al., 2006b).

Figure 33. Overall trends in river discharge for six major Eurasian rivers (1936 to 2009) and four major North American rivers (1970 to 2010) rivers flowing to the Arctic Ocean.

North American rivers included in the analysis are the Mackenzie, Yukon, Back, and Peel rivers. The Peel River is a tributary of the Mackenzie River.

Graph
Source: Shiklomanov and Lammers (2011)
Long description for Figure 33

This line graph shows the overall trends in river discharge for six major Eurasian rivers for the years 1936 to 2009 and four major North American rivers for the years 1970 to 2010 for rivers flowing to the Arctic Ocean.  The graph shows that annual discharge from the six largest Eurasian rivers increased by 7% from 1936 to 1999 and reached record highs in 2002 and 2008.  North American rivers flowing into the Arctic Ocean also showed an increase in discharge.

 

Top of Page

Trends in lake area

While remote sensing allows a broad view of change in lake areas, data for such analyses are limited to recent years. Carroll et al. (2011) examined the change in lake area across Canada from 60° to 70°N latitude from 2000 to 2009, based on analysis of MODIS satellite data. They concluded that, overall, lake area declined over the decade. They interpret this net loss of lake area to be related to longer ice-free periods and increased evapotranspiration, an effect that is particularly noticeable in the small, shallow lakes in Nunavut. Working counter to this trend, climate warming also can increase lake area as a result of melting permafrost causing collapsing and flooding, as has been occurring, for example, in the Ungava Peninsula and James Bay area, south of the treeline (Taiga Shield Ecozone+) (Payette et al., 2004b; Vallée and Payette, 2007; Thibault and Payette, 2009).

Top of Page

Ecological processes in lakes and rivers and linkages with climate change

The Arctic climate directly affects physical, chemical, and biological processes in aquatic systems. It also indirectly affects ecological processes through the control of terrestrial hydrologic systems and processes, particularly those associated with cryospheric components such as permafrost, freshwater ice, and snow accumulation and ablation (Prowse et al., 2006). In moist areas in the spring, thaw lakes and ponds have increased in amount and extent, evidently as a result of permafrost melting and increased precipitation (Stow et al., 2004). However, in drier regions of the Northern Arctic and Arctic Cordillera ecozones, ponds have been reduced in extent and some have disappeared.

Ponds on Ellesmere Island that had been permanent water bodies for millennia (Douglas et al., 1994) dried up completely in the warm summers of 2005 and 2006. Surrounding moss and grass wetlands also dried, with the loss of seasonal standing water, and the vegetation became drier and more subject to fire. This drying of the landscape is attributed to increased evaporation related to high summer temperatures and decreased ice cover, based on analysis of specific conductance measurements taken in the ponds since 1983 (Smol and Douglas, 2007a). More permanent ponds and seasonal ponds and wetlands can be expected to be lost through desiccation as the climate warms further. Certainly many aspects of the physical, chemical, and biological characteristics of these sites will be affected.

Changes in hydrology and water temperature greatly affect the life cycles of aquatic organisms either directly, through temperature, day length, or photoperiod changes, or indirectly, through changes in water quality and available nutrients. Many studies have shown recent changes in aquatic ecosystems that are linked with, or portend, climate change. Effects include regime shifts, or widespread species changes and ecological reorganizations (Smol et al., 2005) (See section on Changes in algal and invertebrate species assemblages in lakes and ponds on page 96). Shifts projected for fish populations will range from positive to negative in overall effect, differ among species and also among populations within species depending upon their biology and tolerances, and will be integrated by the fish within their local habitats (Reist et al., 2006b).

Shorter periods of ice cover on lakes and rivers bring with them many ecosystem changes. For example, less ice leads to warmer water, changing lake mixing regimes, and the distribution of nutrients and oxygen. Changing river ice conditions alter hydrological events like the ice-jam driven spring floods that are a critical part of maintaining wetland and riparian habitat (Peters et al., 2006). Changes in ice duration have impacts at all trophic levels, as outlined in several assessments and reviews (Walsh et al., 2005; Wrona et al., 2005; Vincent et al., 2008). These changes are complex--for example, the increased abundance of food available for fish in river systems (from greater productivity), and the increased habitat availability with less ice may cause otherwise anadromous species to remain in rivers year round. Feeding at sea has been linked to larger sizes in fish and larger populations --thus the increased productivity from less lake ice may ultimately lead to decreased fish yields in lakes and rivers (Reist et al., 2006a).

Top of Page

Natural disturbance

Extreme weather events

In the Arctic Ecozone+ severe weather events that influence the timing, amount, or quality of snow can have major ecological impacts on vegetation, small mammals, and ungulates. Deep snow and frozen snow layers block access to vegetation and prey or make food sources difficult to reach, leading to poor body condition and poor reproduction, as well as direct mortality (Miller and Barry, 2009; Hansen et al., 2013). While these conditions are a normal part of the Arctic winter, extreme weather events can mean that the conditions are more intense or last for longer than normal periods. In this section we have not attempted to define when an event is considered “extreme”, but rather we consider the continuum from normal variability to events that, due to their severity, pose risks of major mortality or reproductive failure to wildlife populations. For tundra ecosystem organisms, the major extreme events tend to occur in the shoulder seasons, autumn and spring, when the combination of temperatures and the timing of snow arrival and melt can result in major disturbances to their populations.

In the autumn, if the snow arrives too late or is too thin to insulate the ground, temperatures at the surface can be extreme for small mammals and invertebrates that depend on this insulation. On the other hand, rapid increases in snow depth in the autumn can make it difficult for large mammals to access forage plants or prey animals. It is for protection and insulation that small mammals seek deeper snow areas in tundra as overwintering habitat (Reid et al., 2012). Heavy snowfall events in August of 1997 and again in August 2000 in the Eureka region of Ellesmere Island covered vegetation, resulting in snow-free seasons of about half the normal length (Mech, 2004). Coincident population crashes of muskoxen, hares, and wolves were documented in the area and are attributed to increased energetic costs to the herbivores (Mech, 2004). In spring, a rapid snowmelt followed by a cold period can have strong effects on small mammals unable to get under the snow for insulation.

Extreme events falling within critical seasons are of particular ecological significance. In a classic paper on winter ecology, Pruitt (1957) described the temperature environment of a central boreal forest site in Alaska and related the changes to small mammal populations. He identified a fall critical period from the date that mean daily air temperatures went below the surface temperature to the date when a depth of about 20 cm of snow dampened daily surface temperature fluctuations. Similarly, the period from when snow depths become less than 20 cm to when mean daily air temperatures rise above surface temperatures is known as the spring critical period (Pruitt, 1957).

Another type of extreme weather event important in tundra systems is icing caused by rain on snow during the freeze period. A short period of temperatures above 0 °C with rain will result in ice layers either on the tundra surface or in the snow. These icing events can be devastating to wildlife populations if they seal the tundra surface making it very difficult for the animals to access forage. Miller and Barry (2009) noted that Peary caribou populations south-central Queen Elizabeth Islands were devastated (over 60% reduction) by exceptionally heavy snow and icing events in the four winters of 1973/74, 1994/95, 1995/96, and 1996/97. These events are projected to become more frequent and more widely distributed with climate change (Putkonen and Roe, 2003). The current distribution and frequency of icing events in the Canadian Arctic are not well-known and are difficult to reconstruct from data from the sparse network of climate stations (Grenfell and Putkonen, 2008). A rain-on-snow event in October 2003 on Banks Island contributed to a reduction of the muskoxen population by 20,000 over a three-year period. This event was analysed by satellite passive microwave imagery, developing a technique to map areas that have internal water and ice layers in the snowpack (Grenfell and Putkonen, 2008).

For tundra vegetation, icing events and extreme weather during the critical periods identified for small mammals are not as important--prolonged changes in temperature and snow regimes are of more significance. The distribution of vegetation along exposure gradients is based on the responses of plant species to the long-term site conditions, including snow depth and summer moisture availability (e.g., Walker, 2000). Sustained changes in soil moisture regimes as a result of flooding or permafrost disturbance (e.g., thermokarst; see section below) can greatly impact tundra vegetation (Vincent et al., 2011).

Additional types of extreme weather events with potential impacts on terrestrial ecosystems are likely increasing in frequency and/or intensity. Extreme heat events, for example, could result in increased fire frequency, loss of vegetation cover, and heat stress effects on tundra fauna, including outbreaks of disease. Intense storms in late summer and fall in the Arctic Ocean are known to alter ocean mixing and to play a role in reduction of sea ice (Jeffries et al., 2012), resulting in impacts on climate of the adjacent land, as well as increased coastal erosion.

Top of Page

Fire

Fire does not currently have a significant impact on Arctic ecosystems, as fires are rare and those that occur tend to be small, due to lack of fire-prone fuels and discontinuity of fuels (as patches of tundra are broken by barren lands, lakes, ponds, and rivers) (Krezek-Hanes et al., 2011). Records and research related to tundra fire ecology in Canada are sparse. Monitoring and analysis of satellite imagery indicates that there were only five large fires in the Arctic Ecozone+ from 1960 to 2007 (Figure 34), though this low frequency may be partly a function of poor monitoring capacity in remote regions (Krezek, 2009, pers. comm.). A 1990 study of fire records, satellite imagery, and observations collected through a questionnaire concluded that fire was more frequent in the western Southern Arctic, rarer to the north and east, and rare to absent in the Arctic Archipelago, related to trends in biomass, human settlement, and climatologic conditions (Wein, 1990).

Most reported fires were small, with larger ones occurring nearer the forest-tundra zone, such as the 1968 fire north of Inuvik (the large fire shown in Figure 34), which spread 15 km from the forest into the tundra (Wein, 1990). The site of this fire was monitored for changes in active layer depth and vegetation until 1993, showing quite variable responses among monitoring locations (Mackay, 1995). At all locations where the active layer deepened (due to the burn or to increased summer air temperatures), the underlying ice-rich permafrost thawed to produce thaw settlement (see next section on Permafrost disturbance). In general, tundra fires remove the vegetation cover and result in a short-term deepening of the active layer (Mackay, 1995).

Figure 34. Area burned by large fires in the Arctic Ecozone+, 1960–2007.

Based on Canadian Large Fire Database from 1960 to 1996 and on analysis of remote sensing imagery from 1997 to 2007.

Graph
Source: data provided by Canadian Forest Service; methodology presented in Stocks et al. (2003) and Krezek-Hanes et al. (2011).
Long description for Figure 34

This bar graph shows the following information:

Area burned by large fires in the Arctic Ecozone+, 1960–2007. Area (ha)
19661968197319761991
2592.0135147.531259.73900450

Top of Page

In northern Alaska a large tundra fire occurred in the exceptionally dry summer of 2007, burning 103,900 ha over a period of two months (Mack et al., 2011). This was the largest tundra fire recorded anywhere. Prior to this fire, researchers at the nearby Toolik Lake Long Term Ecological Research Network site had observed only two very small fires in the region over a 33-year period (LTER Network, 2008). The tundra fire was significant in terms of carbon balance: it was estimated to have released 2.1 teragrams of carbon to the atmosphere, an amount similar to the average annual carbon sink for the entire Alaska tundra biome (Mack et al., 2011).

Although tundra fires occur infrequently in modern times, analysis of charcoal and pollen in lake sediments in Arctic Alaska shows that the shrub tundra between 14,000 and 10,000 years ago burned with a comparable frequency to modern boreal forests (Higuera et al., 2008). These records, along with climate and vegetation reconstructions and analysis of tundra fires in Alaska since 1950 indicate that low effective moisture in summer (from a combination of higher temperatures and/or lower precipitation) and shrub-dominated vegetation cover are key factors in the fire regime. Several analyses now suggest that increases in tundra fires will likely accompany climate warming and the temperature-related increase of shrub vegetation cover that is now being observed in parts of the circumpolar Arctic, including Canada (Steven et al., 2006; Higuera et al., 2008; Hu et al., 2010; Myers-Smith et al., 2011a; Rocha et al., 2012; Elmendorf et al., 2012a).

Top of Page

Permafrost disturbance

Higher temperatures are thawing permafrost resulting in increased frequency and magnitude of slope failures and areas with thermokarst ponding (Lantz and Kokelj, 2008; Schuur et al., 2008; Lantz et al., 2009; Lamoureux and Lafreniere, 2009). These disturbances are restricted to areas with ice-rich permafrost in fine sediment. In the Mackenzie Delta region, retrogressive thaw slumps (Burn and Lewkowicz, 1990) have increased in numbers and size since the early 1970s (Lantz and Kokelj, 2008). Increases in these disturbances have also been observed in the High Arctic on the Fosheim Peninsula, Ellesmere Island (G. Henry, pers. observation, and Table 5) and on Melville Island (Lamoureux and Lafreniere, 2009) (Figure 35). The slumps result in areas with bare sediment, which can provide areas for different combinations of species to become established leading to greater landscape diversity. Lantz et al. (2009) found that green alder (Alnus viridis) had greater growth and reproductive effort on slumps than in undisturbed areas.

Table 5. Frequency of active-layer detachment failure at monitoring sites on the Fosheim Peninsula, Ellesmere Island.
PeriodDuration (years)Number of detachment failuresAnnual rate at which detachments are initiated
Pre-197537–87Note * of Table 52302.6 to 6.2
1975–19871316412.6
1988–20001319014.6

Source: Lewkowicz and Harris (2005)

Notes of Table 5

Note * of Table 5

assumes slope failures had taken place in the past 50 to 100 years, based on the length of time they are estimated to remain visible on the landscape

Return to note * referrer of table 5

Top of Page

In the Low Arctic, these disturbances can expose previously frozen carbon to oxidation and change the tundra sites from carbon sinks to sources (Schuur et al., 2009). Many of the slumps in the Mackenzie Delta region occur on slopes above small lakes and the increased sediment load can significantly change the biogeochemistry of the aquatic environment and alter the biotic communities in the lakes (Kokelj et al., 2009; Thienpont et al., 2013). Lakes affected by slumping arising from permafrost thawing experience shifts in nutrient, light, and phytoplankton relationships, and slump-affected lakes generally have lower nutrients and lower productivity than lakes unaffected by slumps (Thompson et al., 2012).

Figure 35. Active layer detachments on sloped tundra, Melville Island, Nunavut.
Photos
Source: Rudy et al. (2013)
Long description for Figure 35

The photos show the upper zones of two types of active layer detachments: (A) an elongate detachment that is mainly stripped of its vegetative cover, and (B) a compact detachment that has kept more of its vegetation.

Top of Page

In addition to the increases in slope failures, thawing of ice-rich permafrost can lead to subsidence and development of thermokarst ponds (Laurion et al., 2010). The conversion of terrestrial systems to ponds is a profound landscape change, with important implications for the carbon balance of these regions. Thermokarst ponds have been found to be important sources of methane and carbon dioxide (Laurion et al., 2010; Abnizova et al., 2012).

Deeper active layers also contribute to the drying of tundra soils, as water percolates further into the soil and may become unavailable for plants. Warming temperatures also increase evaporation from soils and water bodies, leading to further drying if the loss in moisture is not compensated for by increased precipitation and/or increased moisture from melting permafrost. One of the manifestations of this effect is seen in the drying of Arctic lakes and ponds (Smith et al., 2005a). Smol and Douglas (2007a) describe the drainage of ponds on eastern Ellesmere Island in 2006 which had been permanent water bodies through the Holocene, the ultimate threshold change for these systems. Loss of lakes can result both from increased evaporation and from drainage through the weakened melting permafrost (see also the section on Ecological processes in lakes and rivers and linkages with climate change on page 50). Drying tundra will have important effects on the system response to climate change and will depend on the initial conditions (Shaver et al., 2000).

Permafrost thaw has caused significant disturbances to Arctic landscapes and the combination of climate change and increasing industrial development has the potential to greatly increase rates and magnitudes of permafrost disturbance. These changes have important implications for the structure and function of Arctic terrestrial ecosystems, especially for carbon balance and feedbacks to the atmosphere.

Top of Page

Community and population dynamics

Arctic foods webs are characterized by short food chains, a few dominant animal species that often have cyclic population fluctuations, the central role of small mammals in driving populations at higher trophic levels, and a high degree of responsiveness to regional climatic oscillations and climatic variables such as precipitation, snow depth, and temperature (Elton, 1924; Norrdahl, 1995; Hudson and Bjornstad, 2003; Krebs et al., 2003; Gunn, 2003a; Hörnfeldt et al., 2005; Van Bogaert et al., 2007). They are influenced by a multitude of large-scale disturbances and are particularly vulnerable to global changes that may affect the sustained production of plants and animals on which Arctic cultures depend (Elmqvist et al., 2004). Arctic people are accustomed to large annual fluctuations in the wildlife on which they depend. However, a long-term decline in harvested species, such as caribou, could significantly affect their ability to maintain traditional lifestyles (see the section on Ecosystem goods and services on page 167).

Understanding trends in community and population dynamics requires examining ecosystem processes and integration of trends at all trophic levels, along with trends in drivers and stressors. There are few long-term monitoring programs in Arctic Canada that allow for this integration, though this was a focus of several Canadian-led International Polar Year research programs. For example, the International Tundra Experiment (ITEX) and the CircumArctic Rangifer Monitoring and Assessment (CARMA) networks work to consolidate existing research and monitoring and enhance ongoing observations for the Canadian and the circumpolar Arctic. At the broad-picture level, the Circumpolar Biodiversity Monitoring Program, a project of Conservation of Arctic Flora and Fauna (CAFF) (a working group of the Arctic Council), is working to improve coordination and integration of ecosystem monitoring.

Integration of monitoring and research through intensive study at research sites improves understanding of status and trends in community and population dynamics. These programs track and investigate trends and linkages to understand ecosystems. Examples for the circumpolar Arctic are the research and monitoring programs at Toolik Lake in Alaska, Zackenberg in Greenland, and Abisko in northern Sweden. Canada has not invested significantly in this approach to ecological monitoring and research in the Arctic and consequently understanding of Canadian trends in community and population dynamics is poor. Results from one Canadian site with long-term integrated research are summarized in the “Case study on ecosystem functions and processes: Bylot Island” (page 81). The general discussion on trophic dynamics below is developed further through specific examples in the case study. The bottom trophic level, primary producers, is addressed in the section on Primary productivity (page 71).

Top of Page

Predator-prey cycles

Dynamics of predators in Arctic tundra ecosystems are based mainly on lemmings and other small rodents (Callaghan et al., 2005b). Some predators, including snowy owls and weasels, specialize in rodents and their reproductive success is closely linked to small mammal cycles. Other predators, including foxes, wolverines (Gulo gulo), wolves, and grizzly bears (Ursus arctos), prey on a wider range of species and are able to switch to alternative prey in years of low abundance of small mammals. This means that alternative prey species, such as geese and shorebirds, become indirectly linked to small mammal cycles. Research in the eastern and central Canadian Arctic shows that fluctuations in lemming populations affect their main predators, snowy owls and Arctic foxes, which in turn indirectly affect snow geese, and possibly shorebirds, through shared-predator interactions (Kerbes et al., 2006).

Compared with boreal ecosystems, large predators are not abundant in the Arctic tundra and predation impacts on ungulates are usually low unless the ungulates are at low densities (Callaghan et al., 2005b; Legagneux et al., 2012). In a study across 12 sites in Arctic Canada, 79% of production of small to mid-sized herbivores, including voles, lemmings, and Arctic hares, was consumed by predators, while predators consumed only 9% of combined caribou and muskox production (Krebs et al., 2003).

Wolves and tundra grizzly bears depend on caribou, although the regulatory role of predation for caribou dynamics is uncertain. In the mid-1990s in the Southern Arctic, the Bathurst Herd of 350,000 caribou was estimated to support some 1,500 wolves (Cluff, 2004, pers. comm.) that likely annually killed 40,000 caribou. In addition the herd was estimated to support 500 to 1,000 grizzly bears, with caribou making up 80% of their diet. A grizzly bear may eat between 9 and 18 adult caribou in a year (Gau et al., 2002). For the Northern Arctic Ecozone, Krebs et al. (2003) argue that the dominant ecosystem type is more one driven by variance in weather (bottom-up driven) rather than one controlled by predation (top-down driven). Lagagneux et al. (2012) contrast the predator-driven Bylot Island ecosystem, which lacks large ungulates, with the stronger controls exerted by bottom-up effects on systems in which caribou play a large functional role.

Top of Page

Herbivores and decomposers

Small and mid-sized herbivores are critical in Arctic food webs. The three- to five-year population cycles of small herbivorous mammals are well-known in the ecological literature. The cycles of the various species of lemmings [brown lemmings (Lemmus trimucronatus)andcollared lemmings (Dicrostonys spp.)] (Krebs, 1996; Predavec et al., 2001; Gilg et al., 2003) and voles [tundra voles (Microtus oeconomus)and northern red-backed voles (Myodes rutilus)] (Krebs et al., 2002; Krebs, 2011) have been shown to have a high degree of synchrony across the Canadian Arctic (Krebs et al., 2002). Larger herbivores such as caribou may also cycle with a periodicity of 40 to 70 years (Gunn, 2003a) in the Arctic, but this is not well-documented because of the length of the cycles. Traditional knowledge does, however, provide ample evidence for the cycles in several regions. For example, one account of traditional knowledge passed along through generations of Chipewayan hunters indicates that the Bathurst Herd has been fluctuating in size with a cycle length about 30 years over the past 120 years (Nesbitt and Adamczewski, 2009).

Small mammals

Many small mammal populations throughout the circumpolar Arctic exhibit population cycles of more or less regularity. Some populations do not cycle but instead fluctuate without a pattern or remain at low population levels, with regular outbreaks of high densities being inhibited by conditions like patchy habitat, high predation rates, or particularly harsh winter conditions (for example, lemmings and collared voles showed no signs of cyclic abundance in a study on the Yukon North Slope (Krebs et al., 1995)). Scandinavian small mammal cycles have been well-studied. However, as most Norwegian lemmings (Lemus lemus) live in alpine tundra, the mechanisms driving their population cycles may not be broadly applicable to the Canadian Arctic. Data from northern Scandinavia indicate that lemming cycles may be dampened or lost for decades, while voles maintain regular cycles. Lemming populations in northern Scandinavia remained at low densities in recent decades, followed by recent population outbreaks, while vole cycles have weakened--observations attributed by researchers to the greater sensitivity of lemmings to variations in climate and predation (Ims et al., 2011). Of more relevance to the Canadian Arctic, 20-year studies of small mammals and predator-prey interactions in eastern Greenland indicate that climate change has the effect of increasing the length of the lemming cycle and decreasing maximum population densities, an impact that is detrimental to populations of predators reliant on lemmings (Gilg et al., 2009).

Using an indirect method of studying collared lemmings at 17 sites in the Canadian High Arctic and one site in Alaska, Predavec et al. (2001) were unable to detect a regular periodicity in fluctuations of lemming populations, although variability from year to year was high. Trends in small mammal densities from all monitored sites under the Northwest Territories/Nunavut Small Mammal Survey show, however, that population fluctuations can be synchronized over a large region (Figure 52), at least during the short period over which they have been recorded. Long-term monitoring covering approximately 10 cycles is necessary for a rigorous analysis of cyclic populations (Predavec et al., 2001).

There are no clear trends of long-term changes in the cycles in Canada. At Bylot Island, brown lemmings show cyclic, large-amplitude fluctuations in density with a periodicity of three to four years, though not collared lemmings. There is little evidence of a temporal trend in lemming density, but trend detection is made difficult by the large inter-annual variability. To smooth out those variations, Gauthier et al. (2013) looked at the three-year running mean of density. This analysis suggested a possible decreasing trend in summer lemming density. Lemming populations were especially low during the period 2002 to 2009, but the most recent peak (2011) was relatively high (Gauthier et al., 2013). Continuous, site-specific long-term monitoring of lemmings is still rare in the Canadian Arctic (Krebs et al., 2002), so it remains uncertain to what degree findings from Bylot Island are particular to the site. Only three peaks with similar abundance levels were detected between 1994 and 2012 at the only other site where long-term data on lemmings and voles exist, the Daring Lake Research Station in the central barrenlands, a site that is representative of the tundra-taiga ecotone (Figure 36).

Figure 36. Trends in small mammals (lemmings, voles, and shrews) at sites grouped according to apparent synchrony and ecological regions, 1990–2012.

Abundance indices are numbers per 100 trap-nights corrected based on Beauvais and Buskirk (1999), modifying estimates of sampling effort to account for sprung traps. Species: northern red-backed vole (Clethrionomys rutilus), tundra vole (Microtus oeconomus), Meadow vole (Microtus pennsylvanicus), shrew (Sorex sp.), brown lemming (Lemmus sibiricus), and collared lemming (Dicrostonyx groenlandicus). Not all species are found at all sites.

Graphs and map
Sources: data coordinated by the NWT-Nunavut Small Mammal Survey, Government of the Northwest Territories (GNWT) and Government of Nunavut (GNU).
A Byron Bay: C. Krebs (UBC, pers. comm.), One-eye Jack Lake: R. Bromley (GNWT, goose research, unpublished data). B Bylot Island: G. Gauthier (CEN, ULaval, updated from Gruyer (2007), Pond Inlet: GNU (unpublished data). C and D GNU (unpublished data). E Baker Lake: GNU (unpublished data), Daring Lake: GNWT (Arctic Research Station, S. Matthews, unpublished data). F Melville Sound Islands (Breakwater Lake, Jameson Lake, Hurd Lake, Wilmot Lake): C. Krebs (UBC, pers. comm.), Hope Bay: C. Shank (raptor research, GNWT, unpublished data), Walker Bay: B. Bromley (GNWT, goose research, unpublished data)
Long description for Figure 36

This figure is a map of the Arctic Ecozone+ showing monitoring points where small mammal data was collected.  Points are grouped by geographic area and each group of points has an associated line graph that shows the abundance index for small mammals in that area.  The graphs show that population fluctuations can be synchronized over a large region.

Top of Page

Caribou

Food webs in the boreal and Arctic ecozones+ have relatively few links. The links, however, can be complex in terms of the dynamics of their inter-relationships. These northern ecosystems are nutrient-limited as so much carbon is inaccessible because only a shallow active layer of the soil thaws each year. Caribou, through their forage intake and output (faecal pellets), have complex and cascading effects, strongly patterned over time and space (Kielland et al., 2006). As well, caribou support a diverse group of other species, including external parasites such as blood-feeding mosquitoes. Mosquitoes, in turn, through the filter-feeding of their larvae, are a key element in nutrient cycling in aquatic systems. Further up the food webs, caribou support large-bodied and medium-sized predators and scavengers. Earlier debates about top-down (predator) or bottom-up (forage) regulation for caribou are now replaced by an appreciation of how nutrition and predation interact (Brown et al., 2007a).

Relationships between plants and caribou include the plants’ responses to caribou’s selective foraging. Caribou strongly select for individual plant species and forage for buds and unfolding leaves to maximize nutritional value (White and Trudell, 1980; Russell et al., 1993). The gregarious and migratory behaviours of migratory tundra caribou force their role in ecosystem structure and functioning to be strongly scale dependent (Griffith et al., 2002). Caribou convert plant tissue into body mass and faecal pellets. Through their local foraging movements and their seasonal migrations, they re-distribute nutrients within and across ecozones+. In the taiga ecozones+, the effects of caribou herbivory lag a season, as caribou are foraging during winter when most plant growth and nutrient cycling are quiescent due to sub-zero temperatures. Over the time scale of decades, caribou winter ranges expand and contract and the herds cycle from high to low abundance. Abundance can vary three-fold, with cascading effects on plants and nutrient cycling as the plant communities shift from one state to another. Succession of plant communities as a response to density of foraging may include, for example, lichen-dominated tundra shifting to greater dominance of moss, and then to increased dominance of grass (Van der Wal, 2006).

Nitrogen is a limiting factor for plant growth. Caribou summer browsing can increase the rate of soil nitrogen cycling through modifying the amount of plant litter, which changes the soil microclimate for decomposition and mineralization processes, and through adding soluble nitrogen from faecal pellets and urine (Olofsson et al., 2004). The changes vary with season and time, and with intensity of browsing (Kielland et al., 2006).

Decomposers

Decomposition rates are low in tundra ecosystems (see section on Permafrost on page 27 and section on Nutrient cycling on page 67). Most primary production enters the detrital food chain (Smith and Smith, 2001). A study on Devon Island showed that lemmings consumed about 3–4% of the standing crop most years, with most of the rest of the vegetation eventually being channelled to a variety of microbes and soil organisms, dominated by bacteria and fungi (Smith and Smith, 2001).

Top of Page

Wildlife diseases and parasites
Ungulates

This section, which draws material from the ESTR technical thematic report Wildlife pathogens and diseases in Canada (Leighton, 2011) and the ESTR technical thematic report Northern caribou population trends in Canada (Gunn et al., 2011c), covers disease and parasites related to Arctic ungulates and implications of climate change. Arctic ungulates, caribou in particular, are key components of Arctic terrestrial ecosystems and are vulnerable to changes in disease and parasite infestation. These changes could be related to disease type and frequency and also changes in the ranges of ungulates and of other parasite hosts. Warm weather can promote outbreaks of disease in cold-adapted animals and this may occur more frequently as the climate changes. For example, an introduced muskox population in Norway was struck by an outbreak of pasteurellosis in 2006, killing about 20% of the population. The outbreak was attributed to unusually hot and humid conditions (Ytrehus et al., 2008). An unusual outbreak of yersiniosis, a bacterial disease, in muskoxen on Banks Island in 1986 may also have been related to a period of unusually hot summer weather (Blake et al., 1991).

Brucellosis

Brucellosis is the name given to all diseases caused by infection with any of the several different species of the bacterial genus Brucella. The clinical manifestations of brucellosis are many, but the most common are infection and inflammation of the female and male reproductive tracts with resulting abortion and male infertility, and infection of joints and tendon sheaths resulting in progressive lameness. Infection persists, often for the lifetime of the animal. People are similarly susceptible to infection with Brucella sp., and brucellosis in animals with which people have contact is a public health risk (Chan et al., 1989; Forbes, 1991; Thorne, 2001).

In the Arctic Ecozone+ infection with Brucella sp. is widespread and of potential ecological and public health significance in barren ground caribou populations (and one introduced herd of reindeer near Tuktoyaktuk, Northwest Territories). Brucella suis biotype 4 infects caribou across the Arctic, Taiga Cordillera, Taiga Plains, Taiga Shield, and the northern edges of the Boreal Plains, Boreal Shield, and Hudson Plains ecozones+ (Forbes, 1991). Brucellosis is widespread in Arctic caribou, with 20–50% of animals in some herds infected (Leighton, unpublished data; Koller-Jones, 2006, pers. comm.). However, its ecological impact, if any, on infected populations is not known. Infection of northern people with this bacterium occurs and is associated with consumption of caribou (Chan et al., 1989; Forbes, 1991). Whether or not B. suis biotype 4 is a naturally-occurring pathogen in North America or a pathogen introduced from Europe in imported reindeer also is not known. There are no records of this infection in woodland caribou (Rangifer tarandus caribou), including in the George River herd of northern Quebec.

A serological survey of a large herd of reindeer in the western edge of the Arctic Ecozone+ and of a barren-ground caribou (Rangifer tarandus groenlandicus) herd (Qamanurjuaq) in the Taiga Shield and adjacent Arctic regions of Manitoba and Nunavut in the 1960s found only 9% of reindeer and 4% of caribou infected (Broughton et al., 1970). The more recent infection rates of 20 to 50% may represent a trend of increasing prevalence and/or increasing surveillance. Any environmental changes that increase the overlap of barren-ground caribou with boreal caribou carry the risk that Brucella suis biotype 4 may become established in boreal caribou populations.

Brucellosis is considered to be responsible at least in part for the recent decline of the Southampton Island Caribou Herd. Since 1968, susceptibility to disease and parasites due to low genetic differences has been a likely catalyst for the widespread infection with brucellosis first detected in the Southampton Herd in 2000. Prevalence of the disease rose to 58.8% in 2011 and may be responsible for the drop in pregnancy rates in Southampton caribou since 2000, with the greatest decline in pregnancy rates being since 2008 (Department of Environment, 2013a).

Parasites

Host–parasite systems are particularly sensitive to climate change because many macroparasites have life cycles with free-living stages whose development and survival are strongly dependent on temperature and moisture conditions. Small changes in climate may have a large impact (Kutz et al., 2004) by influencing the development and survival rates of these free-living life stages as well as the abundance and activity of their arthropod vectors and mollusc intermediate hosts. Projected impacts and some empirical observations predict a longer season for development and transmission of parasites, increased infection levels in host populations, and increased outbreaks of disease (Kutz et al., 2005; Kutz et al., 2008).

Umingmakstrongylus pallikuukensis is a parasitic nematode worm prevalent in muskoxen in the western mainland parts of the Arctic Ecozone+. This lungworm requires a gastropod intermediate host to develop to the infective stage, which historically took two years. An empirically based model showed that temperatures have been warm enough since the early 1990s for the parasite to develop in one season. This could lead to potentially increased infection pressure on muskoxen, with possible impacts on fecundity and survival rates of muskoxen (Kutz et al., 2005). The extent to which this has affected muskoxen populations, however, is not known.

In Canada, Besnoitia tarandi infects caribou and reindeer, and probably muskoxen, across their ranges. It has been documented in caribou since 1922 in both Arctic and subarctic Canada (Ducrocq et al., 2013). Infection is common in barren-ground caribou and has been described in woodland caribou. Infection rates in muskoxen are not known (Choquette et al., 1967; Wobeser, 1976; Gunn et al., 1991a; Ayroud et al., 1995). Although occasional severe manifestations of infection on the skin have been seen, most infections appear to have little or no health consequences for these species. Hunter observations, confirmed by veterinary investigations, suggest that the protozoan Besnoitia has recently emerged as a disease-causing agent in caribou herds in Quebec and Labrador (Kutz et al., 2009; Ducrocq et al., 2013). In 2005, focus group discussions in Inuvialuit, Gwich’in and Dene communities in the Southern Arctic and taiga ecozones, some hunters reported increasing lesions (Kutz, 2007), possibly associated with Besnoitia cysts or even warble larvae (Kutz et al., 2009).

Climate change can also lead to changes in ranges of pathogens, both through increasing the area over which temperatures are warm enough for the pathogens to thrive, and through range changes in the host animals. Current knowledge regarding range extents of four ungulate protostrongylid nematodes is shown in Figure 37 . Model projections indicate that, as temperatures rise, U. pallikuuensis will likely expand its range to the north and east (Kutz et al., 2005). Recent observations suggest that this parasite may have expanded its geographic range significantly in the last decade (Kutz et al., 2009) although interpretation of the data is complicated by increased surveillance for parasites. Protostrongylids have not been detected in ungulates from the Arctic islands and Greenland and may be excluded from high latitudes under current climate conditions but could invade these regions under warmed conditions if their host populations of moose, muskoxen, or caribou were to expand northward (Hoberg et al., 2008). Parelaphostrongylus odocoilei, a parasite with a range south of the Arctic Circle, may expand further north. In 1999, following a year with the warmest recorded annual temperature for the Mackenzie District, fatal pneumonia associated with P. odocoilei was first detected in Dall’s sheep (Ovis dalli) in the Mackenzie Mountains (just south of the ecozone+) (Hoberg et al., 2008).

Figure 37. Geographic ranges for protostrongylid parasites in northern ungulates.

Distributions are depicted for Parelaphostrongylus andersoni in caribou; P. odocoilei in wild thinhorn sheep, mountain goat, woodland caribou, black-tailed deer, and mule deer; Umingmakstrongylus pallikuukensis in muskoxen; and a putative new species of Protostrongylidae in moose, caribou, and muskoxen. The range for P. andersoni in the North is presumed to coincide with caribou, although records substantiated by survey are few.

Map
Source: Hoberg et al. (2008)
Long description for Figure 37

This map shows the geographic ranges for protostrongylid parasites in northern ungulates. Distribution of P. andersoni is spread across Alaska and northern Canada, including Yukon, the Northwest Territories, Nunavut, British Columbia, Alberta, Saskatchewan, Manitoba, Ontario, Quebec, and Newfoundland and Labrador. P. odocoilei is distributed along the west coast of North America, from Alaska to California. U. pallikuukensis is found in the northwestern section of the North West Territories. A putative new species of Protostrongylidae is found scattered across Alaska, Yukon, the Northwest Territories, Nunavut, northern Quebec, and Newfoundland and Labrador.

Top of Page

Emergence of disease may follow climate change, but for macroparasites in particular, there are likely to be lag times determined by the life history of the parasites, including the period of development in the hosts (Hoberg et al., 2008). Cascading and cumulative long-term effects of  climate change, including shifts in host–pathogen relationships, may be among the factors contributing to large-scale changes in abundance and distribution observed in keystone wildlife (e.g., woodland and barren-ground caribou) in northern North America.

Carnivores

Rabies persists in Arctic foxes and red foxes (Vulpes vulpes) in the circumpolar Arctic (Mork and Prestrud, 2004). It is difficult to assess trends in northern Canada, partly because samples are not routinely sent for testing once it has been established that rabies is present in the area during a particular winter (J. Kush, Rabies Lab, CFIA-Lethbridge, pers. comm.). Samples are submitted for rabies testing if a potentially infected animal is suspected of having been in contact with a person or a domestic dog. Although rare, rabies has also been documented in polar bears (P. Hale, Government of Nunavut, pers. comm.).

Little information exists on parasite communities of terrestrial Arctic carnivores. Without empirical baseline data on commonly occurring parasites in these animals, few studies will be able to assess impacts and changes of hosts and parasites as a result of climate change (Brooks and Hoberg, 2007).

Toxoplasma gondii

Using mathematical models to predict change, it has been postulated that the prevalence of Toxoplasma gondii infection in humans will increase due to environmental as well as anthropogenic factors related to climate change (Meerburg and Kijlstra, 2009). With increasing temperatures favouring survival of T. gondii in the environment, as well as changes in ecological distributions of vertebrate hosts, (more animals moving into Arctic areas), climate change will likely increase the prevalence of T. gondii in the Arctic (Hueffer et al., 2013).

Infection with Toxoplasma gondii is common in wildlife, domestic animals, and humans worldwide, including those in the Canadian Arctic (Dubey and Beattie, 1988; Mcdonald et al., 1990; Philippa et al., 2004). While some information exists on the occurrence of T. gondii in wild animals in the Arctic, the prevalence of the parasite in areas with a cold climate is low compared to humid, temperate, and tropical areas (Tenter et al., 2000). Infection of, or antibodies specific to, T. gondii have been found in some wild animals with ranges that include or are close to Arctic areas of North America, including moose (Kocan et al., 1986; Zarnke et al., 2001), wolves (Zarnke et al., 2001), wolverines (Reichard et al., 2008; Dubey et al., 2010) , Canadian lynx (Lynx canadensis) (Labelle et al., 2001; Zarnke et al., 2001; Philippa et al., 2004) , muskoxen (Kutz et al., 2000) , and caribou (Kutz et al., 2001).

In the Arctic, Canadian lynx are the most likely wild definitive hosts for T. gondii and are responsible for contaminating the environment with oocysts. Studies conducted on Canadian lynx reported 1 of 5 (20%) from Nova Scotia, 47 of 106 (44%) from Quebec (Labelle et al., 2001), and 39 of 255 (15%) from Alaska (Zarnke et al., 2001) had antibodies to T. gondii. Canadian lynx generally prefer to stay below the treeline, but are occasionally observed and harvested north of treeline.

Close proximity to the treeline was shown to be an important variable for exposure to T. gondii. Surveys conducted by Kutz et al. (2000) to determine the seroprevalence of antibodies to T. gondii in muskox showed a higher proportion of T. gondii exposure in a mainland herd (closer to the treeline) compared to island populations (farther from the tree line) as well as a higher prevalence in adults and juveniles compared to calves. A significant difference was not found in the seroprevalence of T. gondii antibodies between male and female muskoxen (Kutz et al., 2000). In caribou, Kutz et al. (2001) showed that seroprevalence of T. gondii antibodies was significantly higher in mainland populations compared to island populations. However, there were no statistical differences in the seroprevalence of T. gondii antibodies among age classes or between sexes of caribou (Kutz et al., 2001). Reichard et al. (2008) showed that exposure to T. gondii was common in wolverines harvested in the Kitikmeot Region of Nunavut. Prevalence of T. gondii antibodies did not differ significantly between sex and age of wolverines, nor was exposure associated with proximate to the treeline at time of harvest (Reichard et al., 2008).

Top of Page

Phenology

The start and the duration of the growing season in northern terrestrial systems are largely determined by snow cover. Changes in the depth and duration of snow cover over the past 50 years are discussed above in the section on Snow (page 32). The length of the snow-free season has increased significantly in most regions of the circumpolar Arctic, including Canada. In the Canadian Arctic, snowmelt dates have shifted earlier by an average of about 8.5 days from 1950 to 2007 (Zhang et al., 2011).

Tundra plants can be divided into two groups based on their flowering phenology: early or late flowering (Molau, 1993). The flowering phenology is related to their reproductive strategies as either risking pollination (early-flowering species) or seed production (late-flowering species) (Molau, 1993). Early-flowering species dependent on insects for pollination may flower earlier or later than the peak of the insect populations, and any disconnect in timing may be exacerbated by climate change (Molau, 1997). Earlier snowmelt has resulted in earlier flowering by many species (Callaghan et al., 2011b; Oberbauer et al., 2013), and experimental warming has been shown to advance flowering (Arft et al., 1999; Aerts et al., 2006). Warmer temperatures in eastern Greenland have resulted in earlier flowering and a shorter flowering period, with a concomitant decrease in the number of potential insect pollinators (Høye et al., 2013). However, this is the only long-term monitoring study of the effects of warming on both flowering and insect phenology, and there is a need for more systematic studies on effects of earlier snow melt on pollinators in tundra systems. Late-flowering species risk lower seed production because of the unpredictability of late-season weather, especially in the High Arctic. In general, the end of the growing season and the onset of snow cover have not changed as much as snowmelt (Callaghan et al., 2011a), although experimental warming has been shown to delay fall senescence in some tundra species (Marchand et al., 2004). In many coastal locations, at least in the western Canadian Arctic, however, the predominant change in growing season length is in the autumn, with later onset of snow (due to the warming effect of the ocean) (Dye, 2002).

Leaf phenology also varies by species in deciduous shrubs and forbs, although leaf bud break occurs a few days after snowmelt in most species (Shaver and Kummerow, 1992). Experimental warming has shown that leaf bud break is occurring earlier in most species (Figure 38) (Henry and Molau, 1997; Arft et al., 1999; Oberbauer et al., 2013), and remote sensing has shown that the Arctic tundra regions are snow free and green-up earlier than 20 years ago (Zhou et al., 2001; Parmesan, 2007; Bhatt et al., 2010). A recent synthesis of phenological responses at the plot level over the past 30 years has shown that tundra plants are flowering and leafing-out earlier and that this corresponds with the general warming over the same time period (Oberbauer et al., 2013).

Figure 38. Meta-analysis of the effect of experimental warming on leaf bud break at all International Tundra Experiment (ITEX) sites, circumpolar Arctic.

Significant effects are shown with * (P≤0.05). Mean effect size is shown with a coloured circle; the bars are 95% confidence intervals. The “effect size” (y-axis) is an index, with negative values indicating earlier leaf bud break. The experimental plots were warmed using open-top greenhouses.

Graphs
Source: adapted from Arft et al. (1999)
Long description for Figure 38

This figure shows three graphs which depict a meta-analysis of the effect of experimental warming on leaf bud break at all International Tundra Experiment sites in the circumpolar Arctic.  The graphs indicate that experimental warming has shown that leaf bud break is occurring earlier in most species.

Top of Page

Top of Page

Nutrient cycling

Tundra ecosystems are nutrient limited, as shown by relatively strong responses to low and moderate levels of fertilization (Henry et al., 1986; Chapin et al., 1995; Shaver and Chapin, 1995). This ultimately stems from the low temperature and low resource environment of the Arctic, with low rates of net primary production, decomposition, and mineralization (Hobbie, 1996; Nadelhoffer et al., 1997; Cornelissen et al., 2007). The uptake by plants of organic forms of nitrogen, for example amino acids, is important in tundra soils and is undoubtedly a response to the low rates of decomposition and mineralization (Kielland and Chapin III, 1992; Schimel and Chapin III, 1996). Research on the effects of climate change on nutrient cycling in tundra systems has shown that availability of inorganic and organic nitrogen can increase in experimentally warmed plots (Schmidt et al., 2002; Rolph, 2003; Aerts et al., 2006).

The soil microbial communities and their functioning are likely to be altered by climate change, both directly by changes in temperature and moisture, and indirectly through changes in vegetation, such as the increase in shrub abundance in the Arctic (Sturm et al., 2001; Bigelow et al., 2003; Stow et al., 2004; Deslippe et al., 2005; Tape et al., 2006; Myers-Smith et al., 2011a). The soil temperature regime has major consequences for Arctic ecosystems. In Alaska, winter biological processes are contributing to the conversion from tundra to shrub communities through a positive feedback that involves the snow-holding capacity of shrubs, the insulating properties of snow, a soil layer that has high water content because it overlies nearly impermeable permafrost, and hardy microbes that can maintain metabolic activity at low temperatures (Sturm et al., 2005; Myers-Smith et al., 2011a). Increasing shrub abundance leads to deeper snow, which promotes higher winter soil temperatures, greater microbial activity, and more plant-available nitrogen (Schimel et al., 2004). High levels of soil nitrogen favour shrub growth the following summer.

Experimental warming has been shown to alter microbial community structure (Walker et al., 2008b; Deslippe et al., 2012) and snow depth changes have been shown to alter microbial processes in some cases in the Low Arctic (e.g., Schimel et al., 2004). While experimental warming caused differences within sites in microbial composition, based on frequency and abundance of genotypes involved in nitrogen transformations, there were greater differences between sites. The greatest effects of warming were found to occur in wet sedge tundra communities (Walker et al., 2008b). However, resistance to change in microbial communities has been found in studies of warming and combined warming and fertilization of tundra communities in the High Arctic (Deslippe et al., 2005; Lamb et al., 2011).

In wet sedge meadow soils, microbial biomass and nutrient availability peaked early in the spring freeze-thaw phase, but then crashed after soil temperatures rose above 0°C, implying that competition for nutrients from roots results in the collapse of the microbial pool (Edwards et al., 2006). Earlier spring snowmelt and warmer temperatures would obviously alter the timing of these changes, but the vascular plant growth would still depend on available nutrients and soil moisture (Nadelhoffer et al., 1997).

Nutrient uptake in many northern plant species depends on mycorrhizal associations. Deslippe et al. (2011) have shown that long-term experimental warming changes the diversity of ectomycorrhizal fungal families on root tips of dwarf birch (Betula nana), a deciduous shrub. Similarly, Fujimura et al. (2008) found a change in the density of genotypes of the fungal community associated with the roots of arctic willow (Salix arctica) in warmed plots of three High Arctic tundra communities, although there was no shift in relative abundance. The effects of these changes in mycorrhizal communities are unknown, but they will likely affect the ability of shrubs and other plant species to absorb nutrients in the warmer soils.

Arctic animals play a major role in nutrient cycling. Caribou, the main large herbivore in the Arctic, have the net effect of forage removal, production of greenhouse gas, and return of nutrients through faecal pellets. Based on energetics modelling (Russell et al., 2005; Gunn et al., 2011c), a caribou, annually, removes 900 kg of food (2.5 kg/day), produces 20 kg of methane (55 g/day) and returns to the ecosystem, nutrients in the form of faecal pellets, 270 kg (30 g x 25 times a day). At the herd scale, annually, 170,000–350,000 caribou remove 140–320 million kg of forage, produce 3–7 million kg of methane, and return 38–77 million kg of faecal pellets spread over the annual range (150–300 kg/km2). As caribou travel and rest on frozen waterways, the nutrient return from faecal pellets is to aquatic as well as terrestrial ecosystems (Gunn et al., 2011c).

Geese, as well as cycling nutrients through their role as tundra grazers, import energy and nutrients into Arctic ecosystems when they return to their breeding grounds after a winter of feeding in fields and wetlands far to the south. These imported resources move through the food chain when the geese are preyed on, for example, by Arctic foxes (Giroux et al., 2012a). Migratory birds may also export nutrients and energy, for example through removing nitrogen from tundra ecosystems when young produced locally migrate south and die (Gauthier et al., 2011). While locally significant, the importance overall of these resource exchanges to tundra ecology is not known (Gauthier et al., 2011; Giroux et al., 2012a).

Top of Page

Carbon storage and release

Permafrost soils are estimated to contain 1,400 to 1,859 Pg of carbon in frozen and seasonally unfrozen surface layers, which amounts to 50% of the soil carbon in the world (McGuire et al., 2009; Tarnocai et al., 2009). In Canada, Arctic tundra systems were estimated in 2008 to contain about 76 Gt of soil organic carbon in the upper metre of the soil profile (Ping et al., 2008), but, given more recent research (Tarnocai et al., 2009; Grosse et al., 2011; Kuhry et al., 2013), this figure may be low. Soil carbon (C) might react to near-term climate change (Clein et al., 2000). Tundra ecosystems have been sinks for carbon over tens of thousands of years (Ping et al., 2008). This “service” was possible because of the cold soils, which restricted decomposition rates and the accretion of permafrost, which helped to freeze the carbon stored in the soil. There is concern that the warming climate will result in a switch in tundra ecosystems from sink to source of carbon for the atmosphere (Oechel and Vourlitis, 1997; Ping et al., 2008; Grosse et al., 2011; Lund et al., 2012).

Large amounts of C and nitrogen (N) could be released in inorganic forms as Arctic soils warm, the active layer deepens, decomposition rates increase, and the growing season lengthens (Nadelhoffer et al., 1997; Grosse et al., 2011). A large release of carbon dioxide (CO2) from these soils would increase atmospheric CO2, enhancing the rate and magnitude of climate change (a positive feedback). The release of other greenhouse gases, especially methane and nitrous oxide, could be increased from the permafrost soils as they warm, which would contribute to the positive feedback (Vincent et al., 2011).

A coordinated study of the net ecosystem production of Arctic tundra sites in the Canadian Arctic found that all sites were sinks for CO2 during the summer, and the differences in rates followed expected patterns due to latitude and proximity to cold coastal conditions (Lafleur et al., 2012). Annual variability in net ecosystem production was considerable at both a wet and a mesic (moderately moist) tundra site in the Low Arctic, and was strongly related to differences in climate during the growing season (Humphreys and Lafleur, 2011).

Warming experiments to date have shown that wet tundra systems remain as sinks when warmed, as both photosynthesis and respiration are increased; however, warming greatly increased the loss of carbon from dry tundra ecosystems (Oberbauer et al., 2007) (Figure 39). In addition, there was a gradient in net carbon exchanges, with High Arctic systems remaining as sinks but with increasing loss of carbon from Low Arctic systems. There are very few studies of carbon flux in response to experimental warming and other environmental perturbations in Canadian tundra ecosystems.

Figure 39. Effect of experimental warming on net ecosystem exchange of CO2 at four ITEX sites in North America, including Alexandra Fiord, Ellesmere Island.
Graph
Source: modified from Oberbauer et al. (2007)
Long description for Figure 39

This bar graph shows the effect of experimental warming on net ecosystem exchange of CO2 at four ITEX sites in North America, including Alexandra Fiord, Ellesmere Island for the years 1998 to 2000 for both wet and dry general site moisture classes. Warming experiments have shown that wet tundra systems remain as sinks when warmed, as both photosynthesis and respiration are increased; however, warming greatly increased the loss of carbon from dry tundra ecosystems.

Top of Page

Pollination

Phenological changes and changes in animals that act as pollinators could affect vegetation. This may not be as great a concern for the Arctic Ecozone+ as for other ecozones+, as many Arctic flowering plants can self-pollinate or are pollinated by the wind (Callaghan et al., 2005b). Nonetheless, most tundra flowering plants are dependent on insect pollination to set seeds (Kevan, 1972), and outcrossing through insect pollination is important to maintain genetic diversity. Insect pollination is an important part of the reproductive success of flowering Arctic plants because it provides more seed genetic variability than self-pollination or asexual reproduction. Insect pollination is more successful at pollen transfer than wind pollination or transfer by animals (Kevan, 1972; Crawford, 2008).

The main pollinators at the few Arctic locations where pollinators have been studied are flies (Diptera), with bumblebees and butterflies also being important (Ellesmere Island: Kevan, 1972; Zackenberg Greenland: Klein et al., 2008; Høye et al., 2013). Studies in Greenland indicate that most insect pollinators are generalists, though some preferences for certain plants are exhibited (Klein et al., 2008). It is not known to what extent changes in plant and insect species distributions, arrival of new species to the Arctic, as well as changes in phenology, might affect pollination (Klein et al., 2008). However, Høye et al. (2013) have found that the flowering season has become shorter and insect flower visitors have declined as temperatures increased over the past 15 years at Zackenberg, Greenland.

There has been little study on pollination relationships in the Canadian Arctic and little is known about the status and current trends related to the changing climate. The synchronization between insects and flowering Arctic plants is complex and it is hard to predict how these relationships will be affected by climate change, but it is expected that impacts will vary with species (Danks, 1992; Danks, 2004). Insects are highly dependent on the microclimates where they live and in the Arctic they are already at the limits of their ability to adapt in relation to cold hardiness, solar dependence, and sensitivity to desiccation (Danks, 2004). The northward movement of shrubs into new areas creates new microhabitats for insects, in particular pollinators that may be adapted to invade Arctic environments (Klein et al., 2008). Insects have the ability to adapt quite readily to environmental change by moving to areas with more favourable conditions and they have previously shown rapid movement into new habitats created by the retreat of glaciers at the end of the last ice age (Danks, 2004; Klein et al., 2008).

Top of Page

Primary productivity

Primary productivity is low in the Arctic compared with other ecosystems (Figure 40). Primary production ranged from a low of 3 kg dry mass/ha/yr to a high of 334 kg dry mass/ha/yr at 12 sites across the Canadian Arctic (Krebs et al., 2003). Within the Arctic, the standing crop of vascular plants was largest in the Western Arctic, while the standing crop of mosses was largest at High Arctic sites such as Melville Island and Ellef Ringnes Island (Krebs et al., 2003).

Figure 40. Relative net primary productivity of major world ecosystems.
Graph
Source: adapted from Schultz (2005)
Long description for Figure 40

This bar graph showing the following information:

Relative net primary productivity of major world ecosystems.
Major world ecosystemsRelative productivity
Polar tundra1
Boreal2
Temperate mid-latitudes3
Wet sub-tropics4

Top of Page

Evidence is accumulating that the Arctic is getting greener and that the net productivity of Arctic ecosystems is increasing.

Top of Page

In lakes and ponds

In Arctic lakes and ponds, primary production has been shown in several studies to have increased, accompanied by changes in algal species assemblages (Smol et al., 2005; Antoniades et al., 2007). In a study designed to evaluate long-term trends in Arctic lake primary production (Michelutti et al., 2005), researchers used reflective spectroscopy to infer chlorophyll a concentrations in sediment cores. Results from the survey of six Baffin Island lakes indicate that there have been pronounced 20th century increases in primary production (see text box below and accompanying Figure 41 showing long-term changes in an Arctic lake). The changes appear to be synchronized with the record of recent climate change. See also discussion in section on Changes in algal and invertebrate species assemblages in lakes and ponds (page 96).

Primary production in Arctic lakes
Lost Pack Lake, Nunavut
Figure 41. Primary production in Lost Pack Lake, as measured by variations in chlorophyll a concentration in lake sediment.

Note: chlorophyll a is the main pigment found in plants and algae and is a measure of primary production. Values are inferred through spectral analysis of lake sediment cores. The mean chlorophyll a value is the mean of the full data set.

Graph and map
Source: adapted from Michelutti et al. (2005)
Long description for Figure 41

This line graph shows chlorophyll a in mg/g difference from the mean value reconstructions from Lost Pack Lake on Baffin Island for the past 5000 years. The graph shows that primary production rose significantly in the late 19th century.

Top of Page

The figure shows chlorophyll a reconstructions from Lost Pack Lake, one of six Baffin Island lakes examined for long-term trends. All lakes show dramatic increases of inferred primary production within the most recently deposited sediment, following prolonged periods of comparatively low values (Michelutti et al., 2005). Dating of the sediment cores indicates that these rapid increases started in the late 19th century and continue to the present. The increases are a departure, in most lakes, from relatively stable levels of primary production that persisted for millennia. A widespread increase in freshwater production over much of northern Canada is also inferred from major shifts in species composition of algae in ponds and small lakes in many areas (also detected from studies of sediment cores) (Smol et al., 2005; Antoniades et al., 2007).

The best explanation for this change in algae is climatic warming leading to longer ice-free growing seasons and associated changes in lake ecosystems (Antoniades et al., 2005; Smol and Douglas, 2007b). The changes are most pronounced in the High Arctic, but similar shifts in algal species are found in many locations in the Northern Hemisphere--with changes being more recent in temperate latitudes (Rühland et al., 2008).

Source: Federal, Provincial and Territorial Governments of Canada (2010)

Top of Page

On land

On land, changing primary production and changing physical structure of biomass, represented by increases in distribution and height or vigour of shrubs (Elmendorf et al., 2012b), will affect herbivores and, consequently, predators. Increasing productivity and advancing tree and shrub cover have the potential to alter predator-prey relationships and to facilitate range extensions for some biota, such as certain perching birds that cannot otherwise occupy the tundra, as well as their predators, diseases, and parasites. Ecological changes such as these may be involved in some of the declining trends noted in the section on Ecosystem composition (page 102).

Several measures of primary productivity show marked and widespread increases, and ground observations and experimental studies provide confirmation and understanding of the nature of this change. This section examines trends in productivity from several studies, and the section on Ecosystem structure (page 88) discusses related changes in the tundra biome.

A study undertaken for this report (Ahern et al., 2011) used analysis of the Normalized-Difference Vegetation Index (NDVI) to measure trends in productivity for all of Canada’s ecozones+. NDVI is a measure of the photosynthetic capacity of plant cover from space-based observation. Dense plant canopies have positive values of NDVI while snow and ice have negative NDVI values. The resulting map (Figure 42) shows significant trends in NDVI for the Arctic from 1985 to 2006. Only a small percentage of Canada showed a negative NDVI trend, while 22% of the nation’s land area showed a positive trend, with the largest positive trends being in regions of Arctic tundra, alpine tundra, the Pacific coast, and the eastern prairies.

Figure 42. Trends in Normalized-Difference Vegetation Index (NDVI) for the Arctic Ecozone+, 1985–2006.

Produced by TerraVista Earth Imaging from data supplied by CCRS.

Map
Source: Ahern et al. (2011)
Long description for Figure 42

This map  shows the spatial distribution of trends in the Normalized-Difference Vegetation Index (NDVI) for the Arctic Ecozone+ for the years 1985–2006.  The vast majority of areas showed a positive trend in NDVI, indicating a general increase in primary productivity.

Top of Page

Trends for the Arctic Ecozone+

Arctic Cordillera: increasing trend in 9.2% of area, decreasing in 0%
The only area that has significant amounts of vegetation is the Labrador Peninsula. The entire peninsula exhibits a strong positive trend in NDVI, particularly the lower elevations bordering Ungava Bay. The vegetation in this area is low vegetation characteristic of tundra regions.
Northern Arctic: increasing trend in 6.8% of area, decreasing in 0.1%
Areas that are particularly notable for increases in NDVI are the northern portion of Banks Island, the Dundas and Sabine peninsulas of Melville Island, the south shore of Bowman Bay on Baffin Island, and the area along the northwestern shore of Hudson Bay. All of these areas are dominated by tundra vegetation.
Southern Arctic: increasing trend in 23.8% of area, decreasing in 0.3%
The Southern Arctic shows extensive areas of increasing NDVI, most notably along the northwestern shore of Hudson Bay, northeast of Great Bear Lake, and the southern portion of the Ungava Peninsula, where the trend is particularly pronounced. All of these increases are in areas of tundra vegetation.

These changes can be seen in a circumpolar perspective in Figure 43. This figure shows the regional trends in increases in tundra and taiga ecosystem NDVI--with major increases being in the Canadian Western Arctic, Alaska, and Siberia.

Figure 43. Trends in primary productivity, based on remote sensing, 1982–2008.

Grey areas showed no significant trend (P≤0.05); white areas were excluded from the analysis. Based on GIMMS (Global Inventory Modeling and Mapping Studies) NDVI (Normalized-Difference Vegetation Index). Note that most of the NDVI increases shown in Figure 42 above can also be seen on this presentation.

Map
Source: Beck and Goetz (2011) (corrected version, 2012)
Long description for Figure 43

This map shows the trends in primary productivity, based on remote sensing between 1982 and 2008, for the European and North American polar area.  This figure shows the regional trends in increases in the Normalized-Difference Vegetation Index for tundra and taiga ecosystems, with major increases observed in the Canadian Western Arctic, Alaska, and Siberia.

Top of Page

Primary production (biomass) in tundra ecosystems

Primary production is typically low in tundra ecosystems and depends on latitude and local topographic position (Bliss and Matveyeva, 1992). Gould et al. (2003) mapped the vegetation of the Canadian Arctic (Figure 44) and provided estimates of the biomass (Figure 45) and net production of tundra ecosystems (Figure 46). They estimated that net primary production ranges from less than 20 g/m2/yr in graminoid/forb barrens in the polar desert of the High Arctic, to as much as 1,000 g/m2/yr in the lowland riparian areas of the Low Arctic (Gould et al., 2003).

Figure 44. Vegetation map of dominant plant functional types of the Canadian Arctic.

Based on Advanced Very High Resolution Radiometer (AVHRR) data.

Map
Source: Gould et al. (2003)
Long description for Figure 44

This map of the Arctic Ecozone+ shows the spatial location of dominant plant functional types of the Canadian Arctic based on Advanced Very High Resolution Radiometer (AVHRR) data.  In general, more erect forms of vegetation (e.g., trees, tall shrubs) are located in the southern parts of the ecozone+, low and dwarf shrub–non tussock graminoids are spread across the central part of the ecozone+, and the northern part is largely dominated by prostate dwarf-shrub-cryptogams.

Top of Page

Figure 45. Above-ground plant biomass in the Canadian Arctic.

Biomass estimated by determining the range of published biomass values for the dominant vegetation type of each polygon in Figure 44.

Map
Source: Gould et al. (2003)
Long description for Figure 45

This map of the Arctic Ecozone+ shows the spatial distribution of above-ground biomass.  The general trend is toward more above-ground biomass (2000-4000 g/m2) in the southern parts of the ecozone+.

Top of Page

Figure 46. Net primary productivity in the Canadian Arctic.
Map
Source: Gould et al. (2003)
Long description for Figure 46

This map of the Arctic Ecozone+ shows the spatial distribution of net primary productivity.  The general trend is toward higher levels of combined above and below-ground primary productivity (250-1000 g/m2/yr) in the southern parts of the ecozone+.

Top of Page

Experimental warming research at Alexandra Fiord, Ellesmere Island, has shown that a slight increase in summer temperature (about 1°C) results in significant increases in plant growth (Jones et al., 1997; Jones et al., 1999). Significant increases in growth were also noted in similar studies throughout the tundra biome as part of the International Tundra Experiment (ITEX) (Henry and Molau, 1997; Arft et al., 1999; Hollister et al., 2005; Jónsdóttir et al., 2005; Wahren et al., 2005; Walker et al., 2006; Elmendorf et al., 2012a). A meta-analysis of plant community change across the ITEX network showed that up to eight years of experimental warming caused significant increases in growth of shrubs and graminoids (Walker et al., 2006). After nearly 20 years of experimental warming at some sites, the same general patterns were found by Elmendorf et al. (2012a) (Figure 47).

Figure 47. Results of meta-analysis of plant community changes at circumpolar ITEX sites after 1 to 6 years of experimental warming.

Species abundance and canopy height were measured using the standardized ITEX protocols (100 point counts of plants within a 1 m2 plot). On average, daytime air temperature above ground in the experimental plots was raised by 1.5°C over the ambient temperatures in the control plots. The median percent change is for experimental plots in comparison with control plots. Numbers in parentheses are the number of studies included in calculations for each category. A total of 61 Arctic and alpine studies were used in the meta-analysis, five of which are in Canada. Experimental warming had been applied for fewer than six years for the majority of studies.

Graph
Source: adapted from Walker et al. (2006)
Long description for Figure 47

This bar graph shows the median percent change for six categories related to vegetation and graminoid growth. The categories are: mean canopy height (35 studies); shrubs (55 studies); grasses, rushes and sedges (59 studies); forbs (58 studies); lichens (54 studies); and mosses (56 studies).  The graph shows that up to eight years of experimental warming caused significant increases in growth of shrubs and graminoids.

Top of Page

Recent results from long-term studies at Alexandra Fiord, Ellesmere Island, show that there has been a significant increase in biomass (net production) in Canadian High Arctic tundra over the past 20-plus years in response to climate change (Hill, 2006; Hudson and Henry, 2009; Hill and Henry, 2011):

  • Wet sedge tundra biomass increased both above and below ground, with increases from the early 1980s to 2005 of 158% in above-ground vegetation biomass, 67% in root biomass, and 139% in rhizome biomass  (Figure 48) (Hill and Henry, 2011).
  • Snow-bed heath community biomass (at the same site) increased in biomass by 160%, from 33 g/m2 in 1981 to 87 g/m2 in 2008 (Hudson and Henry, 2009). Figure 49 shows the changes in major plant groups in this heath community since 1995.

The increase in biomass in the heath community was in contrast to the effects of long-term experimental warming at the same site, where no increases in overall biomass were found in response to summer season warming of about 1°C (Hudson and Henry, 2010). The increase in average air temperature in the vicinity of the study area over the same time period has been nearly 2°C, which may indicate that certain tundra plant communities will be resistant to relatively small increases in temperature (Figure 50, and see also Figure 14 in the section on Climate trends since 1950 on page 18).

Figure 48. Change in above-ground and below-ground wet sedge meadow biomass over 25 years at Alexandra Fiord, Ellesmere Island.
Graph
Source: after Hill and Henry (2011)
Long description for Figure 48

This bar graph shows the change in above-ground and below-ground wet sedge meadow biomass over 25 years at Alexandra Fiord, Ellesmere Island. Wet sedge tundra biomass increased both above and below ground, with increases from the early 1980s to 2005 of 158% in above-ground vegetation biomass, 67% in root biomass, and 139% in rhizome biomass.

Top of Page

Figure 49. Changes in point cover (biomass index) of major plant functional groups in a snow bed heath community at Alexandra Fiord, Ellesmere Island, 1995 to 2007.

Cover was measured in permanent plots established in 1992 using the ITEX protocols for point frames.

Graph
Source: Hudson and Henry (2009)
Long description for Figure 49

This stacked bar graph shows the changes in point cover (biomass index) of major plant functional groups in a snow bed heath community at Alexandra Fiord, Ellesmere Island for the years 1995 to 2007.  The functional groups graphed are bryophytes, lichens, forbs, graminoids, deciduous shrubs, and evergreen shrubs. The graph shows that snow-bed heath community biomass increased from a total biomass index of around 80 in 1995 to about 100 in 2000 and increased to around 120 by 2007.     Increases were most pronounced for evergreen shrubs and bryophytes.

Top of Page

Figure 50. Mean annual air temperature at four sites in the Canadian High Arctic, 1970–2005.

AF is Alexandra Fiord. Mean annual air temperature change is approaching 1°C per decade.

Graph
Source: Hill (2006)
Long description for Figure 50

Figure 50. Mean annual air temperature at four sites in the Canadian High Arctic, 1970–2005.

Top of Page

Case study on ecosystem functions and processes: Bylot Island

This case study was prepared by G. Gauthier and D. Berteaux for the Arctic biodiversity assessment (CAFF, 2013) and has been revised based on Gauthier et al. (2013).

Site description

Bylot Island is located in the Canadian Arctic Archipelago, at 73°N, 80°W. The 11,067-km2 island is covered by mountains culminating at 1,905 m, an icecap, and several glaciers. The southern part of Bylot is a 1,600-km2 plain gently sloping from about 400 m elevation near the mountains to sea level at the coast, and is bisected by several glacial rivers and creeks (Figure 51). Wet polygons abound in low-lying areas. This plain is covered by tundra vegetation that is rich for its latitude, owing to its southern exposure and protection from northerly winds by high mountains. It is dominated by prostrate dwarf-shrubs, graminoids, forb tundra, and, in polygonal areas, sedge/grass moss wetlands.

Small to intermediate body size species dominate the wildlife assemblage as caribou (Rangifer tarandus) and muskox (Ovibos moschatus) are absent. The bird fauna is especially rich (71 recorded species, 45 as breeders) and comprises a snow goose colony (Chen caerulescens) of about 20,000 pairs. Geese,  brown lemmings (Lemmus trimucronatus), and collared lemmings (Dicrostonyx groenlandicus) are the most abundant herbivores. Several species of avian predators, Arctic fox (Vulpes lagopus), and ermine (Mustela erminea) are present.

Average annual temperature is –14.5°C: 4.5°C in summer and -32.8°C in winter (Gauthier et al., 2011). From 1976 to 2010, the area experienced a strong warming trend in the autumn (September to November), increasing 4.3°C over a 35-year period, and in spring and summer, increasing 2.8°C over the same period, but not in winter (December to February) (Gauthier et al., 2011). The Intergovernmental Panel on Climate Change (IPCC) (2007) projects a 3 to 6°C increase in annual average surface air temperatures from 1980–1999 to 2080–2099 for the area.

Figure 51. Bylot Island and research station location.

The southern part of Bylot Island, where continuous monitoring and intensive studies in the fields of climatology, animal and plant ecology, geomorphology, and limnology have taken place since 1989, is a gently sloping plain bisected by several glacial rivers and creeks. The arrow points to the main research station

Photo
Photo: D. Berteaux
Long description for Figure 51

he southern part of Bylot Island, where continuous monitoring and intensive studies in the fields of climatology, animal and plant ecology, geomorphology, and limnology have taken place since 1989, is a gently sloping plain bisected by several glacial rivers and creeks. The arrow points to the main research station

Top of Page

Monitoring and research

Two 100 km2 study areas centered on glacial valleys have been the focus of continuous monitoring and intensive observational and experimental studies in the fields of climatology, animal and plant ecology, geomorphology, and limnology since 1989. Snow geese, lemmings, Arctic foxes, snowy owls (Bubo scandiacus), long-tailed jaegers (Stercorarius longicaudus), Lapland longspurs (Calcarius laponicus) and, more recently, several species of shorebirds, are the most intensively studied species. Annual plant production in wetlands, plant phenology, and seasonal insect abundance are also monitored. Traditional knowledge on foxes and geese has been collected from members of the community of Pond Inlet located south of Bylot (Gagnon and Berteaux, 2006; Gagnon and Berteaux, 2009).

Key results

The trophic dynamic on Bylot Island is dominated by regular, three- to four-year cycles in lemming abundance (Figure 52). Brown lemming populations show strong fluctuations (greater than forty-fold) but collared lemmings show relatively weak fluctuations (about four-fold) (Gruyer et al., 2008). Predators like foxes, ermines, owls, and jaegers quickly track these fluctuations and their number, reproductive activity, and lemming consumption rate increase dramatically in peak lemming years (Therrien, 2012; Tarroux et al., 2012) . This, in turn, has indirect effects on other species like snow geese due, in part, to prey switching by shared predators (Bety et al., 2002). Goose grazing has a significant impact on the wetland vegetation during the summer, but lemming grazing appears to have little impact on plant production, even in years of high abundance (Gauthier et al., 2004). A trophic balance model showed that less than 10 % of the total annual primary production is consumed by herbivores, but 20 to 100% of the herbivore production is consumed by predators (Legagneux et al., 2012). This suggests that predation plays a key role in the functioning of this ecosystem.

Figure 52. Temporal trends in lemming abundance and Arctic fox and snowy owl reproductive activity.

Figure shows: lemming trends (combined brown and collared lemmings), Arctic fox reproductive activity, and snowy owl reproductive activity on Bylot Island.

Graph
Source: Provided by G. Gauthier and D. Berteaux
Long description for Figure 52

This bar and line graph shows the temporal trends in lemming abundance and Arctic fox and snowy owl reproductive activity for the years 1992 to 2012. Using lemming density, known Arctic fox dens, Arctic fox dens with pups, and snowy owl nests, the graph shows that the trophic dynamic on Bylot Island is dominated by regular, three-to four-year cycles in lemming abundance.  While total known Arctic fox dens increased steadily over the period, Arctic fox dens with pups fluctuated, following trends of lemming density. The abundance of snowy owl nests also followed lemming density trends.

Top of Page

Allochthonous subsidies (energy and resources brought in from outside of the ecosystem) may be important to maintain high predator populations. For instance, high goose populations, which are in part fuelled by an agricultural food subsidy obtained in winter, may help to sustain fox populations, especially in low lemming years (Giroux et al., 2012a). In winter, predators like snowy owls and Arctic foxes use the sea ice for extensive periods, though this may be variable among years for foxes (Therrien et al., 2011; Tarroux et al., 2012), and thus they may depend upon the marine environment for their survival.

The Bylot Island environment has experienced significant changes in temperature and snow characteristics over the study period (Gauthier et al., 2013). Cumulative annual thawing degree-days increased by 37% (Figure 53) from 1989 to 2011 and average temperatures increased in spring and summer, including an increase in June of 3°C per decade, over the same time period. Snowmelt date advanced by 4 to 7 days from 1989 to 2012 and snow depth increased by 48% from 1994 to 2010. Mean annual ground temperature at 10 cm depth showed no trend.

Figure 53. Cumulative thawing degree-days (above 0°C) in the Qarlikturvik Valley, Bylot Island, 1989 to 2011.

Dotted lines show the mean for the whole period.

Graph
Source: Gauthier et al. (2013)
Long description for Figure 53

This line graph shows the cumulative thawing degree-days (above 0°C) in the Qarlikturvik Valley, Bylot Island, for the years 1989 to 2011. The graph shows that cumulative annual thawing degree-days increased by 37% over this time period.          

Top of Page

The strongest temporal trend detected in the Bylot Island ecosystem is an almost doubling (87% increase) of annual above-ground graminoid production (mostly Dupontia fisheri and Eriophorum scheuchzeri) in wetlands over a 20-year period (Figure 54). This is largely due to an increase in summer temperature--the sum of thawing degree-days explains a significant proportion of the annual variation in plant growth (Gauthier et al., 2011). The proportion of the primary production consumed by herbivores also showed a decreasing trend over time. Annual climatic variation is the most important driver of the annual production of young in several migratory birds, including snow geese, as warm spring temperatures increase their breeding effort and advance their phenology (Dickey et al., 2008; Morrissette et al., 2010).

Figure 54. Temporal trend in above-ground live dry biomass of graminoid plants in wetland of Bylot Island, 1990–2010.

Mean ± SE (n = 12 samples per year; trend equation based on individual samples).

Graph
Source: Gauthier et al. (2011)
Long description for Figure 54

This line graph shows the temporal trend in above-ground live dry biomass measured in grams per meter squared of graminoid plants in a wetland of Bylot Island from 1990 to 2010. The graph shows an almost doubling (87% increase) of annual above-ground live dry biomass of graminoid plants in the wetland over a 20-year period, reaching its highest level in 2000.

Top of Page

Despite these strong links with temperature and the observed warming trend, we have yet to see climate-induced trends in most wildlife populations (Gauthier et al., 2013). We have no evidence that lemming cycles have dampened or disappeared in recent years, as has occurred at other sites. Recent analyses suggest, however, that snow depth and quality (density) can affect winter abundance of lemmings and the amplitude of cyclic fluctuations (Duchesne et al., 2011; Bilodeau et al., 2013a; Bilodeau et al., 2013b). The snow goose population has increased significantly over the past 25 years, but this is due to events unrelated to what is happening in the Arctic, such as change in agricultural practices on their wintering ground (Gauthier et al., 2005).

Nonetheless, the more rapid response of lower trophic levels (plants) than higher levels (herbivores and predators) to climate warming may lead to a trophic mismatch. We have evidence that this is already occurring in geese. In years with an early spring, gosling growth is reduced because plants mature too rapidly and the young hatch too late, after the peak in plant nutritive quality (Dickey et al., 2008; Gauthier et al., 2013) (M. Doiron, unpublished data). The median date that geese lay their eggs is related to snowmelt: the earlier the snow melts, the earlier the geese lay their eggs (Figure 55) (Gauthier et al., 2013). The geese undercompensate, however--for an advance in snowmelt date of 10 days, they advanced their laying date by only 3.8 days. Plant growth tracked changes in snowmelt much more closely, indicating a potential for mismatch in timing between hatching and the supply of high-nutrient food in years of early snowmelt.

Figure 55. Annual mean goose laying date and relationship between goose laying date and date of 50% snow cover, Qalikturvik Valley, Bylot Island, 1989 to 2012.

In Figure (a), bars are 25% and 75% percentiles. The regression line in Figure (a) is not significant.

Graphs
Source: Gauthier et al. (2013)
Long description for Figure 55

This figure is comprised of two line graphs. The first graph shows the median goose laying date for the years 1989 to 2012.  The second graph is a scatter plot showing the date of 50% snow cover.  Together these graphs indicate that the median date that geese lay their eggs is related to snowmelt: the earlier the snow melts, the earlier the geese lay their eggs.

Top of Page

Other aspects of ecosystem change include changes in species distribution and in wetlands. The red fox (Vulpes vulpes) invaded the Bylot area in the late 1940s (Gagnon and Berteaux, 2009) and is now reproducing regularly. Drainage and loss of some productive wetlands due to the rapid thermal erosion of ice wedges forming polygons during periods of high spring run-off is a source of concern in this area because such events are likely to increase with climate warming (Fortier et al., 2007). Most of Bylot has been part of Sirmilik National Park since 2001, and this study was a main source of data for evaluating potential changes to the ecological integrity of the park (McLennan et al., 2012). However, the Nunavut Field Unit of Parks Canada has now switched to a much less detailed protocol of ecological integrity monitoring that does not include wildlife monitoring on Bylot Island (D. Berteaux, pers. comm.).

Top of Page

Human stressors on ecosystem functions and processes

Climate change

Climate change is a major stressor for tundra ecosystems in general (ACIA, 2005). At local scales, effects of roads and other potential human disturbances are also important.

While climate change is expected to affect the entire globe, general circulation models consistently indicate that the most severe climate warming will occur in polar latitudes (ACIA, 2005; IPCC, 2007). The Arctic is currently warming at about double the rate of the global average, and some of the most marked changes have occurred in parts of the Canadian Arctic (IPCC, 2007; AMAP, 2011). Paleoecological data confirm that Arctic climate warming during the last century has been well outside the normal range of the previous 400 years and, though there is a natural component to climate change, including at regional and temporal scales related to oscillations, it is clear that the rise in greenhouse gases in the atmosphere due to human activities is a significant human stressor on Arctic ecosystem functions and processes. The impacts of this stressor are woven throughout the discussion in this and other sections of this report. See also the section on Climate trends since 1950 (page 18).

Climate change is a result of anthropogenic increases in greenhouse gases (IPCC, 2007). Trends in the main greenhouse gas, carbon dioxide, are shown in Figure 56 . Air samples from Alert, Nunavut, distant from interference by local sources, are representative of global atmospheric carbon dioxide levels. Annual fluctuations are due to seasonal release of carbon dioxide by plant growth in the Northern Hemisphere. Ice core gas samples from Antarctica show the rapid increase of previously stable carbon dioxide concentrations starting in the late 19th century.

Figure 56. Atmospheric carbon dioxide trends A) measured at Alert, Nunavut (1975–2010) and B) thousand-year trends based on Antarctic ice cores.

Alert measurements are based on weekly air samples. Long-term trends are based on analysis of gas in ice cores from three locations in Antarctica.

Graphs
Source: (A) Environment Canada (2012a); (B) Ahn et al. (2012)
Long description for Figure 56

This figure is comprised of two line graphs.  The first graph shows atmospheric carbon dioxide trends measured at Alert, Nunavut, between 1975 and 2010. The trend line shows that atmospheric carbon dioxide steadily increased over this time frame from approximately 335 ppm to just under 390 ppm. The second graph shows the atmospheric carbon dioxide at three Antartic locations, WAIS Divide (OSU), Law Dome (CMAR,2006) and Dronning Maud Land (Bern, 2005), over the past millennium.  Carbon dioxide increased rapidly at those locations after the year 1800.

Top of Page

Other global-scale stressors

There are other human stressors related to large-scale change that have known or potential impacts on biodiversity in the Arctic, often by interacting with climate change. A full discussion of these is beyond the scope of this report. Changes to the atmosphere that potentially affect biodiversity include ozone depletion and consequent rise of UV-B levels (ACIA, 2005), and Arctic haze (see box below). Of recent concern is the impact of black carbon (soot) from natural and industrial sources. The net effect of black carbon (both in the atmosphere and settling on snow) on the Arctic climate is complex and not well understood, with most studies being based on modeling of interacting factors. Model-based analyses show that the effect of atmospheric black carbon varies considerably depending on the altitude at which it is present in the atmosphere (Flanner, 2013).

Arctic haze

By the 1990s, an unusual haze was noticed in the Arctic that was subsequently linked to emissions of sulphurous and nitrous aerosols. The principal source was air emissions from iron smelting, particularly in Russia. With improved international emission standards, sulphur sources have declined. Non-ferrous metal production remains the dominant source of emissions of acidifying gases to the atmosphere within the Arctic (AMAP, 2006). Other significant anthropogenic sources of sulphur emissions within or close to the Arctic include energy production plants and mining industries. Sources of nitrogen emissions within the Arctic include transportation, in particular shipping and oil and gas activities. Monitoring at Alert showed that, although sulphur aerosols have declined, levels of nitrate aerosol are increasing during the haze season. Forest fires produce soot, which also contributes to haze.

The causes and the effects of acidifying air pollutants and Arctic haze are closely linked to other environmental problems, though these relationships are not well understood. It is not clear, for example, how climate change will influence future acidification and Arctic haze pollution. The effects of haze-producing aerosols on the Arctic climate are complicated by feedbacks between aerosols, clouds, radiation, snow and ice cover, and vertical and horizontal transport processes. Whether the pollutant aerosols cause an overall warming or an overall cooling, or whether the haze itself may affect Arctic ecosystems, is not yet known (AMAP, 2006).

Top of Page