Climate Change Canada

 

5.2.2: Projected changes in snow cover

Projections of surface temperatures across Canada for the near-term under a high emission scenario (RCP8.5) show warming in all seasons in the multi-model average, with concurrent decreases in projected SCF across all of Canada during all seasons (Figure 5.4; Mudryk et al., 2018). During winter, projected snow cover reductions will be greatest across southern Canada, where temperature increases result in less snowfall as a proportion of the total precipitation. Temperatures will remain sufficiently cold at higher latitudes that winter (January/February/March) SCF in this region is not projected to change in response to warming. During spring, the region of snow sensitivity to temperature forcing is projected to shift north, as snow cover retreats across the boreal forest, sub-Arctic, and high Arctic. This leads to projected negative SCF trends (loss of snow) across these regions during the April through June period. Important differences in spring snow cover projections between emissions scenarios emerge by the end of the century, with stabilized snow loss under a medium emission scenario (RCP4.5) but continued loss under a high emission scenario (RCP8.5) (Brown et al., 2017).

Projected changes in SWEmax indicate that reductions will be extensive (5% to 10% per decade through 2050, or a cumulative loss of 15% to 30% over the entire 2020–2050 period) over much of southern Canada, with the greatest changes in the Maritimes and British Columbia (see Figure 5.5). The decreases across the prairies, Ontario, Quebec, and the Maritimes are attributable to increasing temperatures that will shift the proportion of total precipitation that currently falls as snow toward rain (Sospedra-Alfonso and Merryfield, 2017). (Note that the greatest near-term reductions in SWEmax, according to the climate model projections, will be just south of the Canadian border.) Projected changes in British Columbia are consistent with projected SWEmax reductions in the Western Cordillera (Fyfe et al., 2017). While SWEmax is projected to increase by mid-century in the Eurasian Arctic (Brown et al., 2017), minimal change is projected across high-latitude land areas of Canada because increased snowfall is expected to be offset by increasing temperatures that shorten the snow accumulation season.

The greatest snow loss across Canada during the 2020–2050 period is projected to occur in the shoulder seasons (October–November and May–June; Thackeray et al., 2016). During mid-winter, there is minimal percentage change in projected snow cover extent because winter temperatures across northern regions of Canada will remain cold enough to sustain snow cover and there is greater climatological snow extent in winter, which results in smaller percentage changes. The projected trends are similar to the rate of change already observed during the historical period. Trends from a large ensemble of simulations from the Canadian Earth System Model (CanESM2) are slightly stronger than the CMIP5 multi-model mean because projected warming is greater in CanESM2 than in the CMIP5 multi-model mean (Thackeray et al., 2016).

 

5.3: Sea ice

Perennial sea ice in the Canadian Arctic is being replaced by thinner seasonal sea ice (very high confidence). Summer sea ice area (particularly multi-year ice area) declined across the Canadian Arctic at a rate of 5% per decade to 20% per decade since 1968 (depending on region); winter sea ice area decreased in eastern Canada by 8% per decade.

It is very likely that increased temperatures under all emissions scenarios will result in continued reduction in sea ice area across the Canadian Arctic in summer and the east coast in winter. Most Canadian Arctic marine regions will be sea ice-free for part of the summer by 2050 (medium confidence), although the region to the north of the Canadian Arctic Archipelago and Greenland will be the last area in the Arctic with multi-year ice present during the summer (very high confidence). Multi-year ice will, therefore, still drift into the Northwest Passage (and present a navigation hazard for shipping) even when the Arctic Ocean is sea ice-free during the summer.

 

Climate-driven changes to sea ice affect local ecosystems throughout Arctic Canada and influence northern residents through impacts on travelling, hunting, and fishing, with implications for people’s lives, livelihoods, cultural practices, and economic activities. Satellite data show dramatic changes in Arctic sea ice cover during the past 40 years, which are unprecedented over the past 150 years (Walsh et al., 2017). The once-dominant ice that lasts over at least one complete summer melt season (multi-year ice, MYI) in the Arctic Ocean has been largely replaced by ice that melts completely during the summer (seasonal first-year ice, FYI) (Maslanik et al., 2011; Comiso, 2012). This change is important because FYI drifts and melts more readily (Tandon et al., 2018; Stroeve et al., 2012). Average ice thickness over the Arctic Ocean has decreased considerably (Kwok and Rothrock, 2009; Haas et al., 2010; Laxon et al., 2013; Richter-Menge and Farrell, 2013; Kwok and Cunningham, 2015; Tilling et al., 2015). Since 2007, a series of new record-low Arctic sea ice extents have been recorded in the month of September (when sea ice extent reaches the annual minimum), with a loss rate of approximately 13% per decade relative to the 1981–2010 mean (; Stroeve et al. 2012). Concurrent with these changes in ice cover, shipping activity in Canadian Arctic waters has increased over the past decade (Pizzolato et al., 2016; Dawson et al., 2018). Decreases in sea ice extent are no longer confined to the months of low ice cover (August/September/October), but are now also observed during the once-stable winter season (Serreze et al. 2007; Parkinson, 2014).

While the changes in sea ice extent and character across the Arctic are dramatic, there is considerable regional variability. Canadian sea ice areas are composed of portions of the open Arctic Ocean where ice can circulate freely (the western Arctic/Beaufort Sea region), contrasted with the narrow waterways of the Canadian Arctic Archipelago (CAA), where ice is land fast for most of the year. Along the more temperate east coast and in Hudson Bay, the ice melts completely each spring. (Sea ice does not occur on the Pacific coast.)

 

5.3.1: Observed changes in sea ice

Estimates of total ice and MYI area within Canadian Arctic waters are available from the Canadian Ice Service Digital Archive (CISDA), which is an integration of a variety of datasets, including satellite measurements, surface observations, airborne and ship reports, and operational model results (see Canadian Ice Service, 2007 and Tivy et al., 2011a for complete details). This record has been shown to provide more accurate estimates of sea ice concentration (SIC) in Canadian waters compared to satellite passive microwave estimates (Agnew and Howell, 2003). Analysis of seasonally averaged trends in SIC over the 1981–2015 period (selected to match the period of snow datasets described in Section 5.2.1) found reductions over Canadian waters in all seasons (see Figure 5.2). Regions with the strongest SIC declines were eastern Canadian waters in winter and spring, and the CAA and Hudson Bay in summer and fall. SIC trend patterns are closely associated with warming patterns during the seasons of ice onset and growth (from October through March). However, dynamic effects (such as wind, which redistributes sea ice) also influence the observed ice reductions in spring and summer (Mudryk et al., 2018).

The CISDA archive also extends the record of total and MYI back to 1968, almost 10 years earlier than coverage by satellite passive microwave observations. Between 1968 and 2016, sea ice area, averaged over the summer period, has decreased significantly in almost every region of the Canadian Arctic, by up to 20% per decade in some regions (e.g., the Hudson Strait and Labrador Sea; see Figure 5.7). Compared with trends computed over the periods 1968–2008 (Tivy et al., 2011a) and 1968–2010 (Derksen et al., 2012), more regions are now experiencing significant decreases, and the rate of decline is stronger in all regions except Hudson Bay. The largest declines in MYI have occurred in the CAA (approximately 9% per decade) and Beaufort Sea (approximately 7% per decade).

While there are high year-to-year differences due to natural variability, time series of sea ice area clearly show negative trends. The Beaufort Sea experienced record-low sea ice area in 2012, becoming virtually ice free near the end of the melt season (Figure 5.8a; Babb et al., 2016). This was nearly repeated in 2016. The CAA had record-low ice years in 2011 and 2012, eclipsing the previous record set in 1998 (Howell et al., 2013). Baffin Bay has experienced consistently low sea ice area since 1999, while Hudson Bay sea ice area has declined since the mid-1990s (see Figure 5.8d; Tivy et al., 2011b; Hochheim and Barber, 2014). Modelling has demonstrated that the recent extreme lows in Arctic SIC would not have occurred without anthropogenic climate change (see Box 5.1)

The decline of sea ice across the Canadian Arctic is driven by increasing spring air temperature and resulting increases in the length of the melt season. This results in more open water, increased absorption of solar radiation (which further contributes to ice melt), increased water temperature, and delayed fall freeze-up (Howell et al., 2009a; Tivy et al., 2011a; Stroeve et al., 2014; Parkinson, 2014). Changes in sea ice cover are also driven by atmospheric circulation. The Beaufort Sea was once a region where ice would thicken and age before being transported to the Chukchi Sea and recirculated in the Arctic (Tucker et al., 2001; Rigor et al., 2002), but now the region has become a considerable contributor to the Arctic’s MYI loss (Kwok and Cunningham, 2010; Maslanik et al., 2011; Krishfield et al., 2014; Galley et al., 2016). Ice is still being sequestered from the Canada Basin (one of the two ocean basins in the Arctic Ocean) and transported through the Beaufort Sea during the summer months, but this ice is now younger and thinner and melts en route to the Chukchi Sea (Howell et al., 2016a). The CAA was also a region with historically heavy MYI throughout the melt season, but MYI conditions have become lighter in recent years (see Figure 5.7; Howell et al., 2015).

Arctic sea ice thickness has declined in recent years, largely associated with a reduction and thinning of the MYI fraction (e.g., Kwok and Rothrock, 2009; Haas et al., 2010; Laxon et al., 2013; Richter-Menge and Farrell, 2013; Kwok and Cunningham, 2015; Tilling et al., 2015). These studies indicate thickness declines are greater in the Beaufort Sea compared with the north-facing coast of the CAA, which still contains some of the thickest sea ice in the world (Haas and Howell, 2015). Unfortunately, the spaceborne sensors used to obtain sea ice thickness information over the Arctic Ocean are not of sufficient spatial resolution to provide thickness estimates within the narrow channels of the CAA. Although there are only four locations with consistent records and point measurements may not capture regionally representative conditions, the Canadian Ice Service record of in situ land fast ice thickness represents one of the longest datasets in the Arctic, spanning over five decades. Maximum ice thickness has declined significantly at three sites in the CAA (Cambridge Bay, Eureka, and Alert), with decreases ranging from 3.6 to 5.1 cm (± 1.7 cm) per decade from the late 1950s to 2016 (Howell et al., 2016b). No significant trend was found at Resolute, a result that differs from an earlier study by Brown and Cote (1992), which reported a significant increase in maximum ice thickness at Resolute over the 1950–1989 period.

Sea ice along the east coast of Canada is seasonal, with ice melting completely each spring. A robust indicator of change is winter season sea ice area, defined as the average from January through March of each year. The rate of decline between 1969 and 2016, determined from the CISDA for the entire east coast region, is 7.5% per decade (statistically significant at the 1% level; there is only a 1% possibility that the decline is due to chance; see Figure 5.10). This is consistent with the passive microwave time series, which indicates a decline of 9.5% per decade over the 1979–2015 period (Peng and Meier, 2017). There is regional variability within the east coast region, as the rate of decline for the Gulf of St. Lawrence (8.3% per decade) is less than that for eastern Newfoundland waters (10.6% per decade), while the decline for the southern Labrador Sea is not statistically significant at the 5% level (there is a possibility of more than 5% that the decline is due to chance; see Figure 5.10). Years with extensive ice cover are more prominent before 1995, but the region has experienced recent heavy ice years as well (in 2014 and 2015). Sea ice variability in this region is driven largely by temperature and atmospheric circulation (i.e., winds) associated with the Arctic Oscillation (also called the Northern Annular Mode [see Chapter 2, Box 2.5]; Deser and Tang, 2008; Peterson et al., 2015).