Changes in Temperature and Precipitation Across Canada

The annual average temperature in Canada increased by 1.7°C (likely range 1.1°C –2.3°C⁵) between 1948 and 2016 (updated from Vincent et al., 2015⁷³; Figure 4.3 and Table 4.1), roughly twice the increase observed for the Earth as a whole (0.8°C for 1948–2016 according to the global mean surface temperature dataset produced by the Met Office Hadley Centre and the Climatic Research Unit at the University of East Anglia, UK, HadCRUT4 [Osborn and Jones, 2014]). Warming was not uniform across seasons, with considerably more warming in winter than in summer. The mean temperature increased by 3.3°C in winter, 1.7°C in spring, 1.5°C in summer, and 1.7°C in autumn between 1948 and 2016 (see Figure 4.4 and Table 4.1). The changes in temperatures are significant at the 5% level (i.e., there is only a 5% possibility that such changes are due to chance). As well, warming was unevenly distributed across the country. The largest increases in the annual mean temperature were in the northwest, where it increased by more than 3°C in some areas. Annual mean temperature over northern Canada increased by 2.3°C (likely range 1.7 °C–3.0°C) from 1948 to 2016, or roughly three times the global mean warming rate. Warming was much weaker in the southeast of Canada, where average temperature increased by less than 1°C in some maritime areas. Winter warming was predominant in northern British Columbia and Alberta, Yukon, Northwest Territories, and western Nunavut, ranging from 4°C to 6°C over the 1948–2016 period. Spring had a similar warming pattern, but with smaller magnitude. Summer warming was much weaker than that in winter and spring, but the magnitude of the warming was generally more uniform across the country than during other seasons. During autumn, most of the warming was observed in the northeast regions of Canada (mainly in northern Northwest Territories, Nunavut, and northern Quebec). In addition to higher temperatures, the reduction in snow cover (see Chapter 5) and earlier snowmelt (see Chapter 6) also indicate Canada has warmed.

In southern Canada, annual mean temperature increased by 1.9°C between 1900 and 2016 (updated from Vincent et al., 2015⁷³). This warming is significant at the 5% level. This temperature did not rise steadily over time. Temperature increased until about the 1940s, decreased slightly until 1970, and then increased rapidly through 2016. This long-term behaviour of temperature is consistent with that observed globally (see Chapter 2, Section 2.2.1; Hartmann et al., 2013²⁵), but the magnitude of warming in Canada is larger. Mean temperature in southern Canada increased by 2.8°C in winter, 2.2°C in spring, 1.7°C in summer, and 1.6°C in autumn during the same period.

It is extremely likely that human activities have caused more than half of the observed increase in global mean surface temperature from 1951 to 2010 (Bindoff et al., 2013⁵). This causal effect was established through detection and attribution analysis, comparing the observed changes with the natural internal climate variability and with the expected climate responses to human activities (see Chapter 2, Section 2.3.4). Changes in the climate become detectable if they are large when compared with natural internal climate variability, and the change is attributed to human activity if it is (1) consistent with the expected “fingerprint” of human-caused change, as simulated by climate models (see Chapter 3); and (2) inconsistent with other plausible causes. For Canada and the Arctic, where natural internal variability of temperature is high, attribution of observed warming is more difficult than it is on a global scale. Nevertheless, evidence of anthropogenic influence on Canadian temperature has emerged (Gillett et al., 2004¹⁹; Zhang et al., 2006⁹²; Wan et al., 2018⁷⁶), with a detectable contribution to warming in annual and seasonal temperatures and in extreme temperatures.

Two modes of natural internal climate variability that affect temperatures in Canada are the Pacific Decadal Oscillation (PDO) and the North Atlantic Oscillation (NAO) (see Chapter 2, Box 2.5). About 0.5°C of the observed warming of 1.7°C over the 1948–2012 period can be explained by a linear relationship between the PDO and the NAO. Assuming this is completely due to natural climate variability, roughly 1.1°C (likely range 0.6°C–1.5°C) of the observed 1.7°C increase in annual mean temperature in Canada from 1948–2012 can be attributed to human influence (see Figure 4.5; Wan et al., 2018). There is a 33% probability that anthropogenic influence increased Canadian temperature by at least 0.9°C. It is likely that more than half of the observed warming in Canada is due to human influence. The effects of natural internal climate variability on Canadian temperature trends differ in different parts of Canada, enhancing the warming trend in the western Canada and reducing the warming trend in eastern Canada over the past half of the 20th century (Vincent et al., 2015⁷³). The detection of anthropogenic influence on Canadian temperature is also corroborated by other independent evidence, including the attribution of Arctic temperature change to the influence of GHGs and aerosols (Najafi et al., 2015⁵⁵). The reduction in spring snow pack and the ensuing reduction in summer streamflow in British Columbia have been attributed to anthropogenic climate change (Najafi et al. 2017a⁵⁶, 2017b⁵⁷; see Chapter 6, Section 6.2.1). Anthropogenic warming has also increased fire risk in Alberta (Kirchmeier-Young et al., 2017a³⁷; see Section 4.4.1.2).

Earth system models or global climate models provide projections of future climate change based on a range of future scenarios incorporating GHGs, aerosols, and land-use change (see Chapter 3, Section 3.3.1). The fifth phase of the Coupled Model Intercomparison Project (CMIP5, see Chapter 3, Box 3.1) was an internationally coordinated effort that produced a multi-model ensemble of climate projections. Results from this ensemble specific to Canada have been generated using output from 29 CMIP5 models, based on three scenarios: a low emission scenario (RCP2.6), a medium emission scenario (RCP4.5), and a high emission scenario (RCP8.5). Results for a fourth scenario that was part of the CMIP5 protocol (RCP6.0) are also available, but from fewer models. These multi-model results are described by Environment and Climate Change Canada (ECCC, 2016¹¹) and are available for download from the Canadian Climate Data and Scenarios website (http://climate-scenarios.canada.ca/?page=download-intro).

In the following, multi-model climate change projections for 2031–2050 and 2081–2100 (relative to a 1986–2005 reference period) are shown for Canada for a low emission scenario (RCP2.6) and a high emission scenario (RCP8.5), spanning the range of available scenarios. The low emission scenario assumes rapid and deep emission reductions and near-zero emissions this century, whereas the high emission scenario assumes continued growth in emissions this century. The two time periods were chosen to provide information for the near term (2031–2050), when differences in emission scenarios are modest, and for the late century (2081–2100), when climatic responses to the low and high emission scenarios will have diverged considerably. This latter difference illustrates the long-term climate benefit associated with aggressive mitigation efforts. The multi-model median change is shown in map form, along with time series of the average from individual models, which is computed for all Canadian land area. The box and whisker symbols at the right side of the time series provide an indication of the spread across models for 2081–2100. Values for different regions in Canada are provided in Table 4.2.

Projected temperature changes for winter (December–February average), summer (June–August average), and annual mean are shown in Figures 4.6, 4.7, and 4.8, respectively. Enhanced warming at higher latitudes is evident in the winter and annual mean. This is a robust feature of climate projections, both for Canada and the Earth, and is due to a combination of factors, including reductions in snow and ice (and thus a reduction in albedo) and increased heat transport from southern latitudes (see Chapter 3). This high-latitude amplification is not apparent in the summer maps because, over the Arctic Ocean, summer temperatures remain near 0°C — the melting temperature of snow and sea ice. In the near term (2031–2050), the differences in pattern and magnitude between the low emission scenario (RCP2.6) and the high emission scenario (RCP8.5) are modest (on the order of 0.5°C to 1°C). However, for the late century (2081–2100), the differences become very large. Under the high emission scenario, projected temperature increases are roughly 4°C higher, when averaged for Canada as a whole, than under the low emission scenario. The differences are even greater in northern Canada and the Arctic in winter. In southern Canada, projected winter temperature change is larger in the east than in the west, with British Columbia projected to warm slightly less than elsewhere in Canada. The projected summer change is more uniform across the country.

The maps in Figures 4.6, 4.7, and 4.8 illustrate the median projection from the CMIP5 multi-model ensemble — some models project larger changes and some project smaller changes. The spread across models provides an indication of the projection uncertainty discussed in Chapter 3, Section 3.3.2. The spread among the CMIP5 ensemble is only an ad hoc measure of uncertainty. Actual uncertainty could be larger, because CMIP5 models may not represent the full spectrum of plausible representations of all relevant physical processes (Kirtman et al., 2013³⁹). The spread across models also includes natural, year-to-year variability, which continues in the future much as it has in the past. Even when averaged for a region as large as Canada, differences in projected temperature among models are on the order of a couple of degrees. Under a low emission scenario (RCP2.6), annual mean warming in Canada stabilizes at about 1.8°C above the 1986–2005 reference period after about 2050, whereas, under a high emission scenario (RCP8.5), annual warming continues throughout the 21st century and beyond, reaching about 6.3°C above the reference period by 2100. Additional values for Canada as a whole and for various regions are presented in Table 4.2.

Temperature change is one of the key indicators of a changing climate, and many other climate variables are directly or indirectly tied to temperature. The changes in mean temperature are the projected response to emissions of GHGs and aerosols from human activities, and natural internal climate variability will continue to be superimposed on these forced changes. Natural internal climate variability is simulated by the climate models used to make projections of future climate change, and this is evident in the year-to-year variability in the Canada-average temperature time series in Figures 4.6, 4.7, and 4.8 (the individual thin lines). Indeed, this year-to-year variability looks much like what has been observed in the past (see Figure 4.2). In contrast, the underlying forced response (approximated by the multi-model average — the thick line in the figures) is a slowly, monotonically changing value that closely tracks the cumulative emissions of GHGs since the pre-industrial era (see Chapter 3, Section 3.4.1). In assessing the impacts of a warming climate, this combination of slow forced change and natural internal variability is important to keep in mind — the future will continue to have extreme warm and cold periods superimposed on a slow warming forced by human activities.

Because the components of the global climate system are closely interconnected, temperature change in a particular region, such as Canada, is closely related to the change in global mean. This is illustrated in the left panel of Figure 4.9, which shows Canadian mean temperature change versus global mean temperature change. As noted previously, Canadian mean temperature is projected to increase at roughly double the global mean rate, regardless of the forcing scenario. That is, the relationship between Canadian and global temperature change remains constant, as shown by the fact that the results from the different scenarios are all aligned. This connection between global mean and Canadian mean temperature change provides a way of estimating the implications of global change for Canada under alternative forcing scenarios. In other words, impacts estimated under one forcing scenario can be scaled to approximate impacts under another forcing scenario, since the ratio of Canadian to global temperature change is roughly constant. Of course, this assumes that impacts scale directly with temperature (which may not always be the case).

The IPCC Fifth Assessment concluded that “Global mean temperatures will continue to rise over the 21st century if GHG emissions continue unabated” (Collins et al., 2013⁸, p. 1031). Because of the connection between global mean and Canadian mean temperature changes, it is virtually certain that temperature will also continue to increase in Canada as long as GHG emissions continue.

The annual highest daily maximum temperature, averaged across the country, increased by 0.61°C between 1948 and 2016 (updated from Wan et al., 2018⁷⁶). The largest increases were in northern Canada, while decreases were observed in the southern Prairies (see Figure 4.10a). The highest daily maximum temperature that occurs once in 20 years, on average, also increased (Wang et al., 2014⁷⁸). The annual lowest daily minimum temperature, averaged across the country, increased by 3.3°C between 1948 and 2016, with the strongest warming in the west (see Figure 4.10b) (updated from Wan et al., 2018⁷⁶). The lowest daily minimum temperature that occurs once in 20 years, on average, increased more strongly (Wang et al., 2014⁷⁵). Overall, extreme cold temperatures increased much more rapidly than the extreme warm temperatures, consistent with greater warming in winter than in summer, as well as greater warming in night temperatures than in day temperatures.

Indices of high temperature, such as hot days and hot nights, are particularly relevant to public health. Hot days, defined as days with maximum temperature above 30°C, are rarely observed in the regions north of 60° north latitude. In southern Canada, the number of hot days annually increased by about 1 to 3 days at a few stations over the 1948–2016 period (see Figure 4.10c; also see Vincent et al., 2018⁷⁴). Most locations in Canada are not warm enough to have hot nights, defined as nights with daily minimum temperature above 22°C, and the number of hot nights has significantly increased only at a few stations in southern Ontario and Quebec.

Warming in winter and spring has resulted in a significant decrease in the number of frost days (days with daily minimum temperature of 0°C or lower) and ice days (days with daily maximum temperature of 0°C or lower), as well as shortened winter seasons (Vincent et al., 2018⁷⁴). Averaged for the country as a whole, frost days have decreased by more than 15 and ice days by more than 10 days from 1948 to 2016. These changes are consistent across the country. As a result, the frost-free season has been extended by 20 days, starting about 10 days earlier and ending about 10 days later. Heating degree days have decreased while cooling degree-days have increased (see Figure 4.10e and f). The length of growing seasons (see Figure 4.10d) and the number of growing degree days have also increased. The growing season, which starts when there are six consecutive days with daily mean temperature above 5°C in spring or summer and ends when this condition fails to be met late in the year, started earlier and ended later, resulting in an increase in growing season length of about 15 days between 1948 and 2016. With the longer growing season, the number of growing degree days increased.

It is very likely that anthropogenic forcing has contributed to the observed changes in the frequency and intensity of daily temperature extremes on the global scale since the mid-20th century (Bindoff et al., 2013⁵; see also Chapter 2, Section 2.3.4). Several detection studies have shown that the annual lowest daily minimum temperature (Zwiers et al., 2011⁹³; Min et al., 2013⁵³; Kim et al., 2015³⁶) and the annual highest daily maximum temperatures (Wang et al., 2017⁸¹) have been influenced by human activity in three subregions of North America. In Canada, an increase of 3.2°C in the annual lowest daily minimum temperature was observed from 1948 to 2012 (Wan et al., 2018⁷⁶). Only a small fraction (about 0.5°C) of this increase can be related to natural internal climate variability, and anthropogenic influence may have contributed as much as 2.8°C (likely range 1.5° to 4.2°C) to the warming (see Figure 4.5). In addition, much of the observed warming seen in the annual highest daily maximum temperature may also be attributable to anthropogenic influence. Overall, most of the observed increase in the coldest (likely) and warmest (high confidence) daily temperatures of the year in Canada from 1948 to 2012 can be attributed to anthropogenic influence.

While there is a lack of studies directly attributing observed changes in other temperature indices, there is high confidence that substantial parts of the observed changes in most of these temperature indices are also due to anthropogenic influence. It is more difficult to detect anthropogenic influence in values such as annual lowest daily minimum temperature, which are sampled once a year, than in other temperature indices that integrate information from many data samples in a year. These indices are less affected by natural internal variability, while nevertheless retaining the climate responses to external forcing.

The models used to make projections of future climate are discussed in Chapter 3, Section 3.3. When using climate model projections for impact studies, it is often important to consider that the model simulated current climate may differ from observed climate — a reflection of model biases (Flato et al. 2013¹⁷). Many temperature indices are connected to absolute thresholds (like the freezing temperature), and, so, mean biases can substantially alter their usefulness. As a result, where absolute values are important, some form of bias correction is needed. This is a method of correcting the model output to remove, to the extent possible, the influence of model biases. The assessments of projected changes in temperature indices discussed in this subsection are, unless otherwise stated, based on statistically downscaled and bias-corrected data (Li et al., 2018⁴³; Murdock et al., 2014⁵⁴; Werner and Cannon, 2016⁸²; see Chapter 3, Section 3.5).

Daily extreme temperatures, hot and cold, are projected to increase substantially (see Figure 4.11). Annual highest daily maximum temperature is projected to track the projected changes in summer mean temperature, but at a slightly higher rate (the largest difference between the two is less than 0.5°C, appearing in 2081–2100 under a high emission scenario [RCP8.5]). Annual lowest daily minimum temperature is projected to warm faster than winter mean temperature over most of Canada, increasing the extreme minimum temperature in southern Canada by about 3°C by the end of the century under a high emission scenario (RCP8.5). Table 4.3 summarizes projected changes in Canada. For example, averaged over the country, the annual highest daily maximum temperatures are projected to increase by 1.4°C over the 2031–2050 period under a low emission scenario (RCP2.6), and by 2°C for the same period under a high emission scenario (RCP8.5) compared with the current climate (1986–2005). The corresponding projected increase in 2081–2100 under the low emission scenario (RCP2.6) is 1.5°C, only slightly higher than the increases in 2031–2050. A much larger increase, of about 6°C, is expected in 2081–2100 under the high emission scenario (RCP8.5).

In addition to changes in magnitude, the frequency of certain temperature extremes is also expected to change. Extreme hot temperatures are expected to become more frequent, while extreme cold temperatures less frequent. For example, under a high emission scenario (RCP8.5), the annual highest daily temperature that would currently be attained once every 10 years, on average, will become a once in two-year event by 2050 — a five-fold increase in frequency. The annual highest daily temperature that occurs once every 50 years in the current climate is projected to become a once in five-year event by 2050 — a 10-fold increase in frequency (see Figure 4.12). These projected changes indicate not only more frequent hot temperature extremes, but also relatively larger increases in frequency for more rare events (e.g., 10-year extreme versus 50-year extreme; see also Kharin et al., 2018³⁴).

The projected increase in the number of hot days is substantial. In regions that currently experience hot days, the increase may be more than 50 days by the late century under RCP8.5 (see Figure 4.13a). Areas with hot days will progressively expand northward, depending on the level of global warming. The number of frost days and ice days is projected to decrease, with projections ranging from about 10 fewer days in 2031–2050 under the low emission scenario (RCP2.6) to more than 40 fewer days in 2081–2100 under the high emission scenario (RCP8.5) (see Table 4.3) The length of the growing season (see Figure 4.13b) and the number of cooling degree days (see Figure 4.13c) are projected to increase, while heating degree days (see Figure 4.13d) are projected to decrease (see Table 4.3).

Changes in temperature indices and extremes are closely related to changes in mean temperature. The linear relationship between the change in growing season length and Canadian mean temperature (the right side of Figure 4.9; Li et al., 2018⁴³) is one example of how impacts (in this case related to agricultural productivity or forest growth) can be related to temperature change, regardless of the specific pathway of future GHG emissions. Such relationships not only assist impact assessments, but also assist plain-language communication regarding global mitigation efforts and their effect on regional climate impacts. A second example is freezing degree days, a measure of winter severity (e.g., Assel, 1980²), which is the annual sum of degrees below freezing for each day, expressed in units of °C-d. A reduction of several hundred °C-d is projected across southern Canada in the late century, with changes of 1000°C-d or more projected for the North (see Figure 4.14). For context, the historical value of freezing degree days at Whitehorse is roughly 1800°C-d, at Edmonton, roughly 1400°C-d, and at Toronto, roughly 375°C-d (based on 1981–2010 data http://climate.weather.gc.ca/climate_normals/index_e.html).

For temperature indices and extremes, projections by different models for the near term (2031–2050) under a high emission scenario (RCP8.5) agree on the direction (increasing or decreasing) of changes for almost all regions. The model projections for the late century (2081–2100) also agree on the direction of changes for all temperature indices and extremes for every region under a high emission scenario (RCP8.5). This indicates the robustness of projected changes in temperature indices for the future.

It is virtually certain that, in most places in the world, there will be more hot and fewer cold temperature extremes as global mean temperatures increase (Collins et al. 2013⁸). This will also be the case for Canada.

Averaged over the country, normalized precipitation has increased by about 20% from 1948 to 2012 (Vincent et al., 2015⁷³; Figure 4.15 and Table 4.4). The percentage increase was larger in northern Canada — including Yukon, Northwest Territories, Nunavut, and northern Quebec — than in southern Canada. Nonetheless, significant increases were experienced in parts of southern Canada, including eastern Manitoba, western and southern Ontario, and Atlantic Canada. As mean precipitation is typically higher in southern Canada, the absolute amount of precipitation increase is higher in the south, even though the increase in normalized precipitation is smaller in the region. The regional average of normalized precipitation based on the few available long-term data from stations in northern Canada shows an increase of about 30% from 1948 to 2012 (Vincent et al,.2015⁷³); confidence in the regional average is low, however. As trends from individual locations in northern Canada are all increasing, there is a medium confidence that annual mean precipitation has increased in this region. Taken together, there is medium confidence that annual precipitation has increased for Canada as a whole. Additionally, the percentage increase in normalized precipitation is larger than what might be expected from the warming-induced increase in water-holding capacity of the atmosphere, leading to doubt over the magnitude of historical trends. There is low confidence in the estimate of the magnitude of the trend.

Precipitation has increased in every season in northern Canada. In southern Canada, precipitation has also increased in most seasons but the increase is generally not statistically significant. However, a statistically significant decrease in winter precipitation has been observed in British Columbia, Alberta, and Saskatchewan (Vincent et al., 2015⁷³; Figure 4.16 and Table 4.4).

For long-term observed trends, at the century scale, changes in precipitation can be assessed only for southern Canada, due to the lack of data for northern Canada. An increase was observed over all regions of southern Canada since 1900 and is statistically significant at that spatial scale at the 5% level. Warming has resulted in the proportion of the amount of precipitation falling as snow (i.e., the ratio of snowfall to total precipitation) steadily and significantly decreasing over southern Canada, especially during spring and autumn (Vincent et al., 2015⁷³). This is also the case for the Arctic region. There is a pronounced decline in summer snowfall over the Arctic Ocean and the Canadian Arctic Archipelago, and this decline is almost entirely caused by snowfall being replaced by rain (Screen and Simmonds, 2012⁶⁴). Such a change in the form of precipitation, from snow to rain, has profound impacts in other components of the physical environment, such as river flow, with the spring freshet becoming significantly earlier (Vincent et al., 2015⁷³; see Chapter 6, Section 6.2).

There is medium confidence that there is a human-caused contribution to observed global-scale changes in precipitation over land since 1950 (Bindoff et al., 2013⁵). Much of the evidence of human influence on global-scale precipitation results from precipitation increases in the northern mid- to high latitudes (Min et al., 2008⁵¹; Marvel and Bonfils, 2013⁴⁴; Wan et al., 2014⁷⁵). This pattern of increase is clear in climate model simulations with historical forcing (e.g., Min et al., 2008⁵¹) and in future projections (Collins et al., 2013⁸). Observed precipitation in northern high latitudes, including Canada, has increased and can be attributed — at least in part — to external forcing (Min et al., 2008⁵¹; Wan et al., 2014⁷⁵). Atmospheric moisture increases with warming in both observations and model simulations. Natural internal climate variability from decade to decade contributes little to the observed changes (Vincent et al., 2015⁷³). This evidence, when combined, leads us to conclude that there is medium confidence that the observed increase in Canadian precipitation is at least partly due to human influence.

Multi-model projections of percentage changes (relative to 1986–2005) in winter, summer, and annual precipitation in Canada are shown in Figures 4.17, 4.18, and 4.19. The figures include maps of change for the low emission scenario (RCP2.6) and high emission scenario (RCP8.5) for the near term (2031–2050) and late century (2081–2100), and a national average time series of the normalized local changes for Canada as a whole for the period 1900–2100. Unlike for temperature, which is projected to increase everywhere in every season, precipitation has patterns of increase and decrease. In the near term, a small (generally less than 10%) increase in precipitation is projected in all seasons, with slightly larger values in northeastern Canada. In the late century (2081–2100), under the high emission scenario, the changes are much larger, with extensive areas of increased precipitation in northern Canada (more than 30% of the annual mean in the high Arctic). Since annual mean precipitation is low in the Arctic, even modest changes in absolute amount translate into a large percentage change. In contrast, large areas of southern Canada are projected to see a reduction in precipitation in summer under the high emission scenario (RCP8.5); for example, a median reduction of more than 30% is projected for southwestern British Columbia (see Figure 4.18). The projected decrease in summer precipitation (also projected in other parts of the world) is a consequence of overall surface drying and changes in atmospheric circulation (Collins et al., 2013).

As was the case for temperature, the national average time series for precipitation in the lower panels of the three figures show relatively small differences between the low emission scenario (RCP2.6) and high emission scenario (RCP8.5) in the near term (2031–2050). The winter season precipitation changes projected under the two scenarios diverge somewhat by the late century, while the summertime changes are near zero over the entire century, regardless of emission scenario. This small change in national average of locally normalized precipitation hides the fact that summertime precipitation changes are projected to be large (and hence impactful) in many areas of Canada. The large percent increases in northern Canada are generally offset by the large percent decreases in southern Canada, so that the average of percent changes for Canada as a whole in the time series plots shows little overall change in summer precipitation. As mean precipitation is much larger in southern Canada than in northern Canada, the absolute amount of precipitation decrease in the southern Canada is larger than the absolute value of precipitation increase in northern Canada. Regional differences are clearly important for impact studies, and quantitative information at the regional level is provided in Table 4.5. In general, changes in precipitation exhibit more temporal and regional variation than changes in temperature, and, so, projection results for precipitation have less confidence than projection results for temperature.

As the climate warms, particularly in northern Canada, there will inevitably be an increased likelihood of precipitation falling as rain rather than snow. This is consistent with the observed changes in the snowfall fraction noted earlier. Although there has not been a systematic analysis for Canada, one analysis projected a decrease in the fraction of precipitation falling as snow, especially in the autumn and spring, for southern Alaska and eastern Quebec (Krasting et al., 2013⁴²). In addition, regional climate model projections show a general increase in rain-on-snow events over the coming century (Jeong and Sushama, 2017³¹).

These results for changes in mean precipitation are consistent with the IPCC Fifth Assessment, in that the high latitudes are projected to experience a large increase in annual mean precipitation by the late of this century under the high emission (RCP8.5) scenario. The projected increase in annual mean precipitation in the high latitudes is a common feature of generations of climate models. It can be explained by the expected warming-induced large increase in atmospheric water vapour (Collins et al. 2013⁸). Over the historical period, an increase in annual total precipitation in the high latitudes has been detected and can be attributed to human influence (Min et al., 2008⁵¹; Wan et al., 2014⁷⁵). There is high confidence in the projected increase in annual mean precipitation. Confidence in projected changes in seasonal mean precipitation is lower. It should be noted that models generally project less summertime precipitation for southern Canada under a high emission scenario.

There do not appear to be detectable trends in short-duration extreme precipitation in Canada for the country as a whole based on available station data. More stations have experienced an increase than a decrease in the highest amount of one-day rainfall each year, but the direction of trends is rather random over space. Some stations show significant trends, but the number of sites that had significant trends is not more than what one would expect from chance (Shephard et al., 2014⁶⁵; Mekis et al., 2015⁴⁶; Vincent et al., 2018⁷⁴). This seems to be inconsistent with global results (Westra et al., 2013⁸³) and the results for the contiguous region of the United States (Barbero et al., 2017³). The number of days with heavy rainfall⁶ has increased by only 2 to 3 days since 1948 at a few locations in southern British Columbia, Ontario, Quebec, and the Atlantic provinces (Vincent et al., 2018⁷⁴). The number of days with one hour total rainfall greater than 10 mm, with 24-hour total rainfall greater than 25 mm, or with 48-hour total rainfall greater than 50 mm also did not show any consistent change across the country (Mekis et al., 2015⁴⁶). Days with heavy snowfall⁷ have decreased by a few days at numerous locations in western Canada (British Columbia to Manitoba), while they have increased at several locations in the North (Yukon, Northwest Territories, and western Nunavut). The highest one-day snowfall amount has decreased by several millimetres (snow water equivalent) at several locations in the southern region of British Columbia and Alberta (Mekis et al., 2015⁴⁶; Vincent et al., 2018⁷⁴).

The lack of a detectable change in extreme precipitation in Canada is not necessarily evidence of a lack of change. On one hand, this is inconsistent with the observed increase in mean precipitation. As the variance of precipitation is proportional to the mean, and as there is a significant increase in mean precipitation, one would expect to see an increase in extreme precipitation. On the other hand, the expected change in response to warming may be small when compared with natural internal variability. Warming has resulted in an increase in atmospheric moisture, which is expected to lead to an increase in extreme precipitation if other conditions, such as atmospheric circulation, do not change. On the global scale, observations indicate an increase in extreme precipitation associated with warming. Moreover, the increase can be attributed to human influence (Min et al., 2011⁵²; Zhang et al., 2013⁹⁰). The median increase in extreme precipitation is about 7% per 1°C increase in global mean temperature, consistent with the increase in the water-holding capacity of the atmosphere due to warming (Westra et al., 2013⁸³). Compared with the natural internal variability of precipitation, this amount of increase would be too small to be detectable at individual locations. Only about 8.5% of all stations over global land areas with more than 30 years of data show an increase in extreme precipitation at the 5% significance level, which is slightly higher than the rate of stations showing an increase (5%) that could be expected from chance (Westra et al., 2013⁸³). The detection of the increasing intensity of extreme precipitation over lands on Earth is possible because of the vast amount of data available. On the regional scale, there is much less information, which is the case for Canada, where long-term observations are very limited, and detection becomes more difficult.

In the future, extreme precipitation is projected to increase in Canada. Averaged for Canada, extreme precipitation with a return period⁸ of 20 years in the late century climate is projected to become a once in about 15-year event, under a low emission scenario (RCP2.6) and a once in about 10-year event in 2031–2050 under a high emission scenario (RCP8.5) (see Figure 4.20). Beyond mid-century, these changes are projected to stabilize under the low emission scenario (RCP2.6), but to continue under the high emission scenarios (RCP8.5). An extreme event that currently occurs once in 20 years is projected to become about a once in five-year event by late century under the high emission scenario (RCP8.5). In other words, extreme precipitation of a given magnitude is projected to become more frequent. Moreover, the relative change in event frequency is larger for more extreme and rarer events. For example, an event that currently occurs once in 50 years is projected to occur once in 10 years by late 21st century under a high emission scenario (RCP8.5). The amount of precipitation with a certain recurrence interval is projected to increase. The amount of 24-hour extreme precipitation that occurs once in 20 years on average is projected to increase by about 5% under a low emission scenario (RCP2.6) and by 12% under a high emission scenario (RCP8.5) by 2031–2050, and to increase as much as 25% by 2081–2100 under a high emission scenario (RCP8.5). Differences in the projected percentage changes in annual maximum 24-hour precipitation among regions of Canada for the same emission scenario and time period are notably small. The median value for every region is in general within the range of 25th to 75th percentiles of other regions, except under the high emission scenario toward the late century (see Table 4.6).

While results from global climate model projections (such as those above) are useful for impact assessment and adaptation planning, there is an important caveat, especially for extreme precipitation. It is difficult to interpret these projections at local scales. The spatial resolution of global climate models is coarse (typically 100–250 km). The precipitation extremes in a model therefore represent averages over an area of several thousand square kilometres, and so convey different information than may be required for practical applications. More importantly, climate models may not include all of the physical processes that produce local intense rainstorms. This affects the confidence we can have in statistical downscaling products that transform precipitation from coarse resolution models to smaller areas. While regional climate models may operate at much smaller scales, conventional regional climate models, which are used to conduct most of the dynamical downscaling, do not accurately simulate important processes such as convection. These limitations must be kept in mind when using projections for the purpose of regional and local adaptation; in particular, the projected values given by these global or regional climate models should not be interpreted literally as the measured amount of precipitation at a point location.

Estimating changes in short-duration extreme precipitation at a point location is complex because of the lack of observations in many places and the discontinuous nature of precipitation at small scales. Projection of such extreme precipitation is also difficult because of the shortage of simulations by models with a very high resolution that resolve the physical processes that produce those extreme events (Zhang et al., 2017⁹¹). Nevertheless, multiple lines of evidence support high confidence in projecting an increase in extreme precipitation globally. These lines of evidence include attribution of an observed increase in high-latitude total precipitation to human influence, consistency in projected future increases in extreme precipitation by multiple models, and the physical understanding that warming would result in an increase in atmospheric moisture. It is likely that extreme precipitation will increase in Canada in the future, although the magnitude of the increase is much more uncertain

In June 2013, an extreme flood event in southern Alberta became Canada’s costliest natural disaster to that date, with significant damage to property and infrastructure throughout the region, including in the City of Calgary. The flood displaced almost 100,000 people and resulted in $6 billion in damage, including $2 billion in insured losses (ECCC, 2017¹²).

A storm producing heavy rainfall over the region triggered the flooding event in the Bow River basin, but a combination of both meteorological and hydrological factors led to the extreme flooding. A recent study (Teufel et al., 2017⁶⁹) assessed the contributions of several of these factors, including anthropogenic GHG emissions.

The study used the Canadian Regional Climate Model (CRCM5) to run large ensembles of high-resolution simulations for North America. To assess the contribution of human climate change, simulations of the model were run using present-day levels of GHGs and also using pre-industrial levels to represent the time before humans had a discernable impact on the climate.

To estimate the probability of the event, return periods were calculated for three-day rainfall totals during May and June exceeding the observed amount. The return period for the observed event in the present-day climate in the Bow River was estimated to be about 60 years. Using climate projections, the return period is estimated to be reduced to about 20 years by the late 21st century (under both an intermediate emission scenario [RCP4.5] and a high emission scenario [RCP8.5]), implying that the type of extreme rainfall that led to the southern Alberta flooding event will become much more common in the future.

Estimated return periods were compared between the present-day and pre-industrial climates, in order to determine the human contribution (see Figure 4.22). Including human GHG emissions resulted in shorter return periods (the event is more likely) for three-day maximum precipitation over the entire southern Alberta region than for pre-industrial levels (see Figure 4.22a). Weather and climate variability tends to be larger for smaller regions, resulting in a smaller ratio between the anthropogenic influence and natural internal variability. Therefore, the anthropogenic influence on the rainfall events over the smaller Bow River basin region was less pronounced (see Figure 4.22b). There is no discernable difference for maximum three-day surface runoff, and thus no anthropogenic influence is detected for this variable (see Figure 4.22c). The authors acknowledge the uncertainties in modelling complex surface hydrological processes and suggest that any increase in rainfall could be offset by decreases in snowpack or frozen ground conditions.

In summary, human influence was detected for the flood-producing rainfall, particularly over the larger region, but human influence could not be detected for the flood itself. A flood event is the result of many factors in addition to the amount of rainfall, including the ground conditions, characteristics of the snowpack, and size and orientation of the storm. As a result, two events with the same rainfall amounts do not necessarily produce floods of the same magnitude. The complex hydrological processes that occur after the rainfall reaches the ground add additional uncertainty which decreases the ability to detect human influence.

Increased GHG emissions, largely due to human activities, result in increased temperatures. Increased temperatures allow more moisture to be available in the atmosphere for precipitation, leading to increased intensity of extreme rainfall events. The 2013 southern Alberta flood was the result of a combination of many factors, and this study demonstrated that human-induced emission of GHGs had increased the likelihood of an extreme amount of precipitation in southern Alberta, at least as large as the amount observed during this event.

In early May 2016, a large wildfire burned almost 600,000 ha (a land area covering 6000 square kilometres) in northern Alberta. This fire resulted in the evacuation of all of the residents of Fort McMurray (over 80,000 people) and halted production in the oil sands (Government of Alberta, 2016²³). Insured losses are estimated at $3.5 billion (IBC, 2016²⁸). The total cost of the event is still being determined, but it is expected to be considerably higher.

The fire ignited near the Horse River amid very dry fuel conditions. High winds a few days later resulted in rapid spread and fire growth. A study has used event attribution to assess the influence of human-induced climate change on several measures of wildfire risk (see Box 4.2), albeit not extreme fire itself, in this region (Kirchmeier-Young et al., 2017a³⁷).

Like the previous example, the study used large ensembles of model simulations, in this case employing the Canadian Earth System Model (CanESM2). To assess human influence, the model was run with only natural forcings (solar and volcanic effects) and also with a combination of natural and anthropogenic forcings. The anthropogenic component includes GHG emissions, aerosols, atmospheric ozone changes, and land-use change.

Fire weather (see Box 4.2), fire behaviour, and fire season measures were calculated to characterize fire risk from climate model output. To quantify the anthropogenic contribution, a risk ratio (NASEM, 2016⁵⁸) was calculated as the ratio of two probabilities: one for the event’s occurrence when the human component is included, and one for the occurrence of the same event with only natural factors. The risk ratio can be interpreted as how many times as likely the event is as a result of anthropogenic factors. For example, a risk ratio of 1 implies no change in the probability of occurrence, and a risk ratio of 2 implies the event is twice as likely, or that there has been a 100% increase in the probability of the event compared with the unperturbed climate.

Results of the analysis show that three of the fire risk indices — extreme values of the Fire Weather Index (FWI; see Box 4.2), high number of spread days, and long fire seasons — all show risk ratio values greater than 1 (see Figure 4.23), indicating extreme values of each measure of wildfire risk are more likely when anthropogenic warming is included. Risk ratios vary among the different fire risk indices analyzed. However, extreme values of all measures describing wildfire risk are more likely with anthropogenic forcing.

Increasing temperatures, like those observed across Canada (see Section 4.2), will lead to drier fuels, and thus increased fire potential, as well as longer fire seasons. It would require increases in precipitation well beyond what is expected with climate change to offset increasing temperatures in terms of the FWI indices (Flannigan et al., 2016¹⁶). The study demonstrated that the extreme Alberta wildfire of 2016 occurred in a world where anthropogenic warming has increased fire risk, fire spread potential, and the length of fire seasons across parts of Alberta and Saskatchewan.