Chapter 2: Overview of past changes in climate

Chapter description

This chapter provides a synthesis of past changes in climate across Canada, as well as their causes, drawing on the assessments in this chapter and in chapters 4 to 9. It also includes a detailed assessment of past changes in average surface air temperature, precipitation, and near-surface wind speed.

Chapter key messages

Plain language summary^{Footnote 1}

This chapter and the following one (Chapter 3) perform a unique role in this report. They assess past (Chapter 2) and future (Chapter 3) changes in the climate across Canada, by focusing on changes in surface air temperature and precipitation and drawing on the detailed technical assessments presented elsewhere in this report. In addition, this chapter considers past changes in near-surface wind speed, and pays special attention to the sources of historical climate information required to view today’s climate in the context of past climatic conditions in Canada. The chapter does this by focusing on direct climate observations provided by meteorological instruments and observers, as well as the information derived from those observations with the help of weather and climate models.  

This synthesis of past changes observed in Canada’s climate leads to the inescapable conclusion that Canadians are experiencing changes across the climate system that are very largely driven by a single common factor—the rise in greenhouse gas concentrations in the atmosphere due to human activity. The effects of rising greenhouse gas concentrations can be seen most directly in the warming of the climate, but also in changes to precipitation, snowfall, and river streamflow. Other effects include the melting of glaciers and large ice sheets, thawing of permafrost, losses of sea and lake ice, warming of the oceans around Canada, rising sea levels, the acidification of ocean water as it absorbs more carbon dioxide from the atmosphere, and changes in the frequency and intensity of a wide range of extreme events. While the evidence of human influence is more difficult to discern in some aspects of climate change than others, the overall body of evidence points unavoidably in only one direction, to human influence as the dominant cause of the observed changes. This conclusion is confirmed through the results of climate change detection and attribution studies that rigorously identify the expected effects of human influence on the climate in observations of past temperature and precipitation changes. The understanding of the mechanisms that connect warming to other changes in the climate system adds additional support to this conclusion. Simply put, no other mechanism of change acting on the climate since the pre-industrial era (approximated in the report as 1850 to 1900) could have caused sustained system-wide changes of the magnitude observed.

Canada has warmed rapidly since the 1950s. Canada’s warming rate between 1948 and 2023 was about 70% higher than the global warming rate over the same period. However, most of the warming since 1948 has happened since 1970, both in Canada and globally. Canada’s warming rate between 1970 and 2023 was almost double the global warming rate during this shorter period. In addition, Canada is warming substantially faster than our neighbours to the south in the contiguous United States and faster than most other global land areas. Warming is occurring more rapidly in Canada than in most other places because our northern climate, particularly at higher latitudes, is affected by Arctic amplification (a topic that is covered extensively in Chapter 4, section 4.2; Box 4.2).

Although observations prior to 1948 are much scarcer, sufficient data are available between 1900 and 1948 to also examine how temperatures in southern Canada have changed since the beginning of the 20th century. That assessment clearly shows that the period after 1948 was warmer than the half-century that preceded it, consistent with our understanding of how the global average temperature has changed since 1900. In addition, new methods for combining results from climate models with observations from detection and attribution studies enable us to estimate that, as of 2025, human-caused warming has increased Canada’s annual average temperature by roughly 2.4°C (90% uncertainty range: 1.8−2.9°C) relative to the level that prevailed during the pre-industrial period.

Canada has also experienced considerably higher precipitation since 1949, with annual average precipitation in Canada as a whole being about 10% higher in recent years than at the beginning of this period. While annual average precipitation is low in Canada’s North, the relative rate of increase is more than double the rate of increase in Canada’s South (~19% versus 7.5%, respectively, over the 1949–2023 period). The higher relative rate of increase in precipitation at high latitudes is an expected feature of a warming climate. As with temperature, seasonal and regional variations can also be found in the extent of the change in precipitation amounts, although the details of these changes cannot be assessed as confidently as for temperature. We assess, on the basis of formal detection and attribution studies, that human influence has caused most of the change in precipitation amounts that has occurred in northern Canada and at least part of the change that has occurred in southern Canada.

While direct observations of snowfall amount and frequency are limited, the amount of precipitation that occurs on days when temperatures are near 0°C serves as an effective proxy for these metrics. An analysis of these proxy data leads to the conclusion that, consistent with warming, both snowfall frequency and amount have decreased across southern Canada and increased in northern Canada.

In contrast to the strong evidence that human influence on Canada’s climate has caused it to warm and the amount of precipitation to increase, a clear signal of change in average surface wind speeds in Canada has yet to appear. This is partly because climate models do not yet give a clear indication of how human influence on the climate may affect near-surface winds, and partly because long-term high-quality wind speed observations are scarce in Canada. One phenomenon that has been observed in the last few decades in many parts of Canada and elsewhere in the world is a gradual reduction in average wind speeds, termed atmospheric stilling. The climate science community has devoted considerable attention to this phenomenon, but a clear consensus on its cause has not emerged.

The ability to observe and understand long-term changes in Canada’s climate depends crucially on the quality and availability of long-term observations and the data products derived from those observations. Therefore, we also include in this chapter a dedicated discussion of the various sources of data that are available and their suitability for use in climate change studies. This discussion includes information on the improvements made since the first edition of CCCR (CCCR2019) in the methods used to process observing station data and make it suitable for climate analyses, as well as in the methods used to create gridded datasets from those data.

2.1: Introduction

This chapter provides a synthesis of the changes observed in the Canadian climate system, based on its own detailed assessment of past changes in average surface air temperature, precipitation, and near-surface wind conditions, as well as the assessments in chapters 4 to 9 (Figure 2.1). This chapter complements Chapter 3, which presents a corresponding synthesis of future changes in the Canadian climate system. Together, chapters 2 and 3 are designed to act as a stand-alone overview of findings from across the entire report and of changes in our climate.

Temperature, precipitation, and wind are fundamental aspects of the weather we experience every single day, resulting in an accumulated experience over time that we perceive as our climate. We refer here to temperature, precipitation, and wind collectively as the basic surface climate variables. These variables largely define what crops we can grow, the nature of the ecosystems that inhabit our landscape, how much water is available for human use, and when and where we can produce renewable energy from water and wind resources. Changes in these basic climate variables affect us and the natural and managed systems on which we rely. They also influence other aspects of the climate system (assessed in chapters 4 to 9) that, in turn, affect how the basic surface climate variables evolve.

Detailed assessments of past changes in Canada’s basic surface climate are presented later in this chapter, but first, we start in section 2.2 with a synthesis of the assessments of these changes in Canada’s climate, drawing on material presented later in this chapter. We also present the assessments of changes in the atmosphere and ocean, the cryosphere (snow, ice, and frozen ground), the water and carbon cycles, and the intensity and frequency of climate extremes, which are covered in greater detail in chapters 4 to 9.

The supporting detailed assessments of past changes in surface air temperature, precipitation, and wind speed are presented in sections 2.4, 2.5, and 2.6 of this chapter, respectively. These assessments rely primarily on long-term observations from Canadian weather and climate stations that have been very carefully quality controlled and homogenized. Homogenization is a process carried out to ensure that observed changes are not affected by non-climatic factors such as the replacement of older instruments with new technology or changes to the landscape around an observing station that affect the measurements recorded. Section 2.3 describes different types and sources of climate data, discusses why some are more suitable than others for monitoring climate change in Canada, and describes the data used in this chapter.

This chapter generally references changes against the average conditions that prevailed during a specified baseline period. Different baseline periods are used in different contexts (section 2.4.2). One such period that deserves particular attention is the 1850–1900 period, which the Intergovernmental Panel on Climate Change (IPCC) uses as a reference period to assess more recent changes that are attributable to human influences on the climate (IPCC, 2021b). This period, which is loosely termed the pre-industrial period, is used for pragmatic reasons. Insufficient instrumental surface temperature data exist prior to 1850 to estimate the global mean difference from average conditions for earlier periods, and climate models running climate change simulations to reproduce historical conditions generally start the simulations in 1850. The use of the term “pre-industrial” in this report therefore makes very loose reference to the widespread second Industrial Revolution, which started in the mid-19th century and continued into the early 20th century. The effects of cumulative emissions—from the consumption of fossil fuels over the preceding 150 years (which includes both the first and second Industrial Revolutions)—on global average surface temperatures only began to become discernable in the 20th century (e.g., see Figure SPM.1 in IPCC, 2021c). This is the rationale for using 1850–1900 as the pre-industrial reference period for assessing the subsequent changes in climate that can be attributed to the effects of human activities on the global climate.

Given the availability of climate observations, trends and other measures of change are generally assessed in this chapter for the century-long periods of 1900–2023 or 1916–2023 when considering Canada’s South, or 1948–2023 or 1949–2023 when considering Canada as a whole (including Canada’s North). Assessments of past changes in climate variables synthesized from chapters 4 to 9 use periods of various lengths, depending on when systematic observation of those elements began. However, they always consider changes in recent decades, when the global climate system has warmed rapidly (IPCC, 2021b). Geographically, changes in climate are sometimes summarized by region based on provincial and territorial boundaries, as outlined in Figure 2.2. Most often, however, the regions referred to are determined by data availability and the geographic scale and scope relevant to a particular climate variable.

Figure take-away: A visual snapshot of the contents of this chapter and of important cross-chapter linkages.

Figure title: Visual guide to the content of Chapter 2 and key cross-chapter linkages

Figure take-away: This report summarizes changes in some climate variables and climate extremes by region.

Figure title: The six regions of Canada referred to in this report

2.2: Synthesis of past changes

Observed climate warming in Canada and its causes

Key Message 2.1: Canada is warming in all regions and seasons. Canada warmed by 2.0°C during the 1948–2023 period (very likely^{Footnote 2}, ^{Footnote 3} 0.9–3.1°C), with northern regions experiencing warming of 2.6°C (very likely 1.4–4.1°C) and generally more warming occurring in winter than in summer.

Key Message 2.2: Human influence has warmed Canadian average 2015−2024 temperatures to a level that is 2.2°C (likely 1.8–2.6°C) above that in the pre-industrial era, which is double the amount of warming in the global average temperature that is attributable to human influence when comparing 2010−2019 with global average temperatures in the pre-industrial era (high confidence). The observed warming in Canada of 2.0°C from 1948 to 2023 is very strongly dominated by human influences and indistinguishable from the warming that can be attributed to external influences on the climate (very high confidence).

Observed climatic effects of warming in Canada

Key Message 2.3: Most observed climate changes across Canada—on land, in the oceans around Canada, and in the atmosphere—are consistent with a warming climate. Emissions of greenhouse gases from human activity provide the only plausible explanation for this collective change (very high confidence).

Key Message 2.4: The observed occurrence in summer of longer growing seasons, increased building cooling requirements, and higher hot temperature extremes are all consistent with a warming climate, as is the occurrence in winter of reduced building heating requirements and less severe cold temperature extremes. Fire seasons have become longer (high confidence), and the area burned in wildfires has increased (medium confidence). The warming climate has also led to substantial changes in the water cycle and cryosphere (very high confidence), as indicated by increased annual precipitation and intensified precipitation extremes, snow cover and glacier mass reductions, earlier spring peak streamflow in rivers, decreasing lake and river ice, and thawing permafrost.

Key Message 2.5: The effects of the warming climate are also clearly seen in Canada’s ocean areas, through increases in ocean water temperature, more frequent marine heatwaves, freshening of near-surface ocean waters, and decreasing sea ice cover, which has resulted in increasing ocean wave heights in the Arctic (high confidence). Some of Canada’s coastal areas are also experiencing the effects of sea-level rise (very high confidence) and more frequent extreme sea-level events (high confidence), which can contribute to coastal flooding. The oceans around Canada have also absorbed anthropogenic carbon dioxide from the atmosphere, which has resulted in ocean acidification in the ocean’s near-surface layers (very high confidence).

This section provides a synthesis of observed climate changes and their causes across Canada’s land areas and in its surrounding oceans. The changes that are measured, documented, and analyzed by Western science are lived and experienced daily throughout the country by Indigenous people in Canada, and thus this section begins with poignant testimony from two different corners of Canada. Case Story 2.1 is written by a group of Inuvialuit youth leaders who bear witness to the systemic changes in the western Arctic climate through the lens of unipkait, the basis for Inuit ways of knowing (Case Story 2.1), while Case Story 2.2 describes changes observed by a Chisasibi Eeyou Elder in Eeyou Istchee (northern Quebec). These stories are followed by an overview of observed temperature changes in Canada in section 2.2.1, which connects those changes with the changes assessed in chapters 4 to 9. The chapter then considers water-related changes (section 2.2.2), climate extremes (section 2.2.3), the oceans around Canada (section 2.2.4), and changes in the terrestrial and ocean carbon cycle (section 2.2.5).

In the synthesis subsections that follow, the chapters and section numbers that support each conclusion are provided in brackets (e.g., 5.4 would refer to Chapter 5, section 5.4).

2.2.1: Canada’s warming climate

Canada is warming. Annual and seasonal average temperatures across Canada have increased, with the greatest warming occurring in winter and in northwestern Canada (2.4). Annual average surface air temperatures for Canada as a whole increased by 2.0°C between 1948 and 2023, and by 2.6°C for Canada’s North (2.4). Canada’s rate of warming in 1948–2023 was substantially higher than that in the contiguous United States and higher than that for the global land area (very high confidence). In addition, Canada warmed about 65% faster than the global land and ocean area combined during this period (high confidence) (2.4). Arctic amplification, the phenomenon of greater warming in Arctic regions caused mainly by processes related to the loss of sea ice (Box 4.2), contributes to the greater rate of warming in Canada relative to the global average, and to the greater rate of warming in Canada’s North relative to Canada’s South (high confidence) (2.4, 4.2). There is high confidence that Arctic amplification has become detectable and is primarily attributable to human influence since the late 20th century (4.2). 

While both human activities and natural climate variations have contributed to the warming observed in Canada, the human factor is dominant (very high confidence) (2.4). The observed warming of 2.0°C between 1948 and 2023 is statistically indistinguishable from the warming that can be attributed to external influences on Canada’s climate (very high confidence).

Canada has insufficient historical observations to estimate the average temperature that prevailed during the pre-industrial period, making it impossible to provide a reliable estimate of the total warming observed in Canada since that period. Nevertheless, the amount of warming specifically caused by humans relative to the pre-industrial period can be estimated by combining observational information from more recent periods with the results of climate models using detection and attribution techniques (2.4). These methods show that human factors had very little influence on Canada’s average temperature between the pre-industrial period and the middle of the 20th century (very high confidence), but by the decade of 2015−2024, Canadian average temperatures were likely 1.8–2.6°C warmer than during the 1850–1900 pre-industrial period because of the human influence on our climate, with 2.2°C being the best estimate of human-caused warming. This is approximately double the amount of warming in the global average temperature due to human influence that was assessed by the IPCC (2021b) (high confidence).

The effects of a warming climate have been observed throughout the climate system in Canada (Figure 2.3). The growing season has become longer; heating degree days (indicating building heating requirements) have decreased and cooling degree days (indicating building cooling requirements) have increased (2.4). Hot extremes have become more frequent and intense, while cold extremes have decreased in frequency and become warmer (8.2). Annual precipitation amounts and the frequency and intensity of extreme precipitation events have increased in Canada as a whole (2.5, 8.3), as would be expected in a warming climate. The coverage, duration, and amount of snow have decreased in some regions (6.2). Earlier spring freshets, which occur when the spring snowmelt dramatically increases streamflow, have helped to shift the timing of peak streamflow to earlier in the spring (5.3). The area burned by wildfire has increased (9.5) and the fire season has become longer (8.7). Ice on lakes and rivers has decreased (6.4), glaciers have lost mass (6.5), and permafrost has thawed (6.7). Sea ice area has decreased, and year-round sea ice has thinned (6.3). Ocean wave heights in the Arctic have increased, which is consistent with larger ice-free ocean areas that persist for longer periods (7.5). The oceans around Canada have warmed (7.2) and absorbed carbon dioxide (CO₂) from the atmosphere, which causes ocean acidification (7.7). The salinity of the ocean’s surface waters has decreased, consistent with higher inputs of freshwater from precipitation and meltwater (7.3). Global sea level has risen due to thermal expansion and melting ice sheets (7.4), and coastal flooding has increased in some regions of Canada, where local sea level has increased relative to fixed points on land (7.6, 7.4, 8.7).

Together, these changes on land and in the atmosphere, oceans, and cryosphere indicate that Canada’s climate is undergoing a systemic change. There is very high confidence that the only plausible explanation for this collective change is warming from increased anthropogenic greenhouse gases (GHGs). Multiple lines of evidence support these changes in Canada, including observations based on different types of data products (2.3), the results of modelling, and the understanding of physical processes, all of which point to human-caused increases in GHG concentrations as the common factor. Natural influences acting in the absence of human influences, including natural external forcing from changes in solar and volcanic activity and natural internally generated low-frequency climate variability, cannot explain the systemic changes that have been observed since the middle of the 20th century. The changes in Canada’s climate are described in more detail in the following synthesis sections and in the respective chapters of this report.

Figure take-away: The effects of a warming climate have been observed throughout the climate system in Canada.

Figure title: Synthesis of past changes in the climate system across Canada

2.2.2: Precipitation, water cycle, and cryosphere changes

There is very high confidence that annual average precipitation has increased in Canada as a whole, with larger percentage increases in Canada’s North (2.5). There is lower confidence in the magnitude of these changes (2.5). In addition, the observed changes in precipitation are not the same for all seasons and for all regions. Winter precipitation has decreased across much of Canada’s South, while increasing elsewhere. Spring precipitation has increased across most of Canada, but patterns of change are more complex in summer and fall (2.5).

Warming has increased the fraction of precipitation falling as rain rather than snow across most areas of southern Canada, especially during spring and fall (2.5, 8.3, 5.2). Snow cover duration has decreased over the last four decades in Canada as a whole (high confidence) (6.2). Regional decreases in snow cover duration in northern and eastern Canada are greater than the national average, while snow cover duration increased in central Canada (medium confidence) (6.2).

Annual total streamflow amounts have increased in northern Canadian regions (high confidence) (5.3, 6.7). The increase in streamflow results from both increased precipitation and from permafrost changes due to thickening of the active layer (the layer that freezes and thaws annually) as permafrost thaws, which allow a larger proportion of the water received at the surface as precipitation to make its way into streams and rivers (5.3, 6.7). In contrast, trends in annual total streamflow amounts are not consistent across the rest of the country (5.3). There is a clear pattern of increasing streamflow in most of Canada during winter and the early spring months and declining flows during summer and fall (high confidence) (5.3). The onset of the spring freshet and the timing of the subsequent spring peak streamflow have become earlier across Canada. This has caused a shift in some drainage basins from a snowmelt-dominated flow regime, where the annual peak streamflow occurs when snow melts in the spring, towards a rainfall-dominated regime, where the annual cycle of streamflow corresponds more closely to that of rainfall (high confidence) (5.3), which affects the timing of water availability in those areas. Streamflow-related floods in Canada result from many factors, including extreme precipitation, rapid snowmelt, ice jams, and rain-on-snow events, with complex interactions among these factors (5.7). No spatially consistent trends have been identified in either streamflow-related floods or the factors that cause them (high confidence) (5.7).

Historically, periodic droughts have occurred across much of Canada, but no long-term trends in their frequency have been detected (5.6). However, an increase in the moisture deficit (one indicator of drought conditions) has been observed in the continental interior of Canada. This change is most noticeable in the southern Prairies (5.6), where precipitation has decreased significantly in winter (2.5). There is no evidence of long-term changes to surface water storage (i.e., water stored in lakes, rivers, and wetlands) in Canada, although considerable variability is apparent (5.4) and both increases and decreases in groundwater levels have been reported across the country (5.5).

Thicker multi-year sea ice in the Canadian Arctic is being replaced by thinner seasonal sea ice (which forms anew each winter) (very high confidence) (6.3). Sea ice coverage has declined in all Canadian waters monitored by the Canadian Ice Service over the last four to five decades (very high confidence) (6.3). However, the amount of sea ice that moves from the central Arctic Ocean into the northern Canadian Arctic has increased (high confidence) (6.3), from which it then moves southward into the shipping lanes of the Northwest Passage. This process has contributed to the depletion of the amount of multi-year ice in the central Arctic Ocean (6.3).

The duration of ice on Canada’s lakes, averaged for the country as a whole, has decreased over the past four decades (high confidence) (6.4). However, some regional variation is found, with decreases in lake ice duration in northern Canada and the Maritimes and increases in central Canada (medium confidence) (6.4). While changes in river ice in Canada are highly variable from one location to another, changes in the yearly timing of river ice breakup are related to changes in seasonal air temperature (medium confidence) (6.4). In addition, the number of mid-winter breakups has increased in Canada’s rivers (medium confidence) (6.4).

The volume of ice in the glaciers of Western Canada and the Arctic has decreased over the past two and a half decades, the period for which satellite measurements are available (very high confidence) (6.5). Limited longer-term observations suggest glacier mass has been decreasing since the 1960s, with losses accelerating in recent decades (high confidence) (6.5), driven mainly by increasing temperatures and darkening ice surfaces (6.5). Glaciers in Western Canada are among the fastest-thinning glaciers in the world (6.5). Permafrost has warmed in northern Canada since the 1980s, and permafrost thawing has been observed (very high confidence) (6.7). In the ice-rich terrain of northern Canada, the landscape has changed in response to climate-driven permafrost thawing (very high confidence) (6.7).

2.2.3: Extremes

The intensity and frequency of hot extremes have increased over Canada as a whole and in multiple regions since the mid-20th century (high confidence) (8.2). The intensity and frequency of cold extremes have decreased over Canada as a whole and in all Canadian regions since the mid-20th century (high confidence) (8.2). Cold extremes have warmed more than hot extremes, consistent with the decreases observed in winter in day-to-day temperature variability (8.2). Human influence on the climate is the dominant driver of the observed warming of both hot and cold extremes (high confidence) (8.2).

The intensity and frequency of one-day and five-day heavy precipitation events have increased over Canada as a whole (medium confidence) (8.3.1). At the continental scale, human influence on the climate is the main driver of the observed intensification of extreme one-day precipitation events across North America, through an increase in atmospheric moisture, which accompanies increasing temperatures (high confidence) (8.3.1, Box 8.3). Updated observational data also provide some evidence that the intensity of short-duration rainfall extremes (i.e., at timescales of less than a day) has increased in Canada as a whole (low confidence) (8.3.2). The intensity and frequency of heavy one-day snowfalls have increased at most locations in northern Canada over the last 75 years (medium confidence) (8.3.3). In southwestern Canada, heavy one-day snowfall amounts have decreased over the last 75 years (medium confidence), while the direction of change in this metric is variable in southeastern Canada, with no clear pattern (8.3.3).

Atmospheric rivers are narrow streams of air that move large amounts of water vapour from ocean areas to land, where the water vapour condenses to form precipitation in amounts that are frequently extreme in some parts of Canada (4.5, 8.3). The water vapour content of the atmosphere has increased with warming, and thus increases are expected in the overall frequency of atmospheric rivers and the intensity of the strongest atmospheric river events. On the basis of recent studies using reanalysis data (2.3), there is medium confidence that the frequency and intensity of atmospheric rivers affecting Canada have increased, but with very low confidence in the magnitude of these trends (4.5). Detection and attribution studies to confirm whether the observed trends are due to human-caused climate change have not yet been performed (4.5).

There is very low confidence in past changes in wind speed extremes, including those associated with thunderstorms and extratropical, tropical, and post-tropical cyclones affecting Canada, due to limited observations, inconsistencies across datasets and time periods, wide year-to-year variability, and large model uncertainty (8.4). Observed changes in the frequency and intensity of extratropical storms affecting Canada are uncertain and small compared to internal climate variability (4.4). There is low confidence in past changes in the prevalence of the environmental conditions favouring thunderstorms in Canada, due to the lack of regional studies in this country, the weak and inconsistent trends detected in many different severe weather proxies and datasets, and large internal atmospheric climate variability (4.7).

Some extremes can be represented as compound events, which are caused by a combination of weather- and climate-related conditions (8.7). For example, coastal flooding events often result from a combination of conditions, such as the simultaneous occurrence of high water levels and extreme precipitation. Compound coastal flooding, in which the flooding is caused by a combination of multiple factors, such as storm surge, waves, precipitation, and river flooding, has increased in some locations (low confidence), particularly in Atlantic Canada (8.7.2). Fire weather, or the combined occurrence of hot, dry, and windy conditions that are conducive to wildfires, has increased across Alberta and British Columbia (medium confidence) (8.7.1). Other regions in Canada have seen some increases in fire weather, though the changes are generally not statistically significant and variability is high from year to year (8.7.1). The fire season in Canada has become longer (high confidence) (8.7.1). Furthermore, the annual forest area burned in Canada has increased during the era of satellite records (i.e., since the early 1980s) (medium confidence) (9.5). There is high confidence that human-caused climate change has contributed to this increase, based on several lines of evidence (8.7, Box 8.5, 9.5).

Several recent high-impact extreme events affecting Canada have been studied to determine whether human influence on the climate altered their intensity or likelihood of occurrence. Most of these studies have found that human influence did play a role, including in the 2021 heat wave in western Canada (also known as the 2021 Pacific Northwest heatwave); flooding events in British Columbia, and Ontario and Quebec; and the 2023 wildfire season (8.1, Box 8.2, Box 8.5).

2.2.4: Ocean changes

Sea surface temperatures have increased in the oceans around Canada during ice-free periods (high confidence), consistent with the warming of the world’s oceans, of which human influence is the main driver (7.2). The greatest increases in annual average sea surface temperatures have occurred in the North Atlantic waters adjacent to southern Atlantic Canada (high confidence) (7.2). By season, the largest increase in warming has occurred in summer, extending throughout the ocean area bordering eastern Canada, including Hudson Bay (high confidence) (7.2). The oceans around Canada have also experienced more frequent and intense periods of unusually high temperatures, known as marine heatwaves (high confidence) (7.2).

The surface waters of the oceans around Canada, which have relatively low salinity, have become fresher (less saline) since the 1950s (high confidence) (7.3). This suggests an increase in freshwater input to these oceans from precipitation and runoff from land, which includes meltwater from glaciers and ice sheets (high confidence) (7.3). Climate change−driven warming and freshening of the surface waters of the northeast Pacific and southern Atlantic Canada regions have changed how water temperature and salinity vary with depth in the upper part of the ocean (high confidence) (7.3). In turn, these changes affect ocean features that are important for the marine ecosystem (7.3).

Other important ocean properties are also changing. The oceans around Canada have absorbed anthropogenic CO₂ from the atmosphere, which has resulted in ocean acidification in the near-surface layers of the ocean (very high confidence) (7.7). Subsurface oxygen levels have declined offshore and in coastal areas in the northeast Pacific Ocean, St. Lawrence Estuary, and Scotian Shelf (high confidence) (7.7). However, the processes responsible for the declines in oxygen content vary by region and their complexity prevents scientists from determining how much human-caused climate change has contributed to these changes. In subsurface ocean waters, acidification is exacerbated in regions where oxygen content has declined (high confidence) (7.7). This acidification is caused by CO₂ that enters the ocean and penetrates to deeper waters and, in regions experiencing oxygen decline, by the biological relationship between oxygen consumption and CO₂ release (7.7).

Globally, sea level has risen because of the human-caused warming of the climate system (very high confidence), which affects both the ocean and the atmosphere. The warming ocean causes seawater to expand, and the warming land surface has caused widespread melting of land ice, which adds water to the ocean; these effects have been the main contributions to global sea level rise (very high confidence) (7.4).

Tide gauge measurements serve as an important source of information about past sea-level changes. However, these gauges measure sea level relative to fixed points on land, and these fixed points—and the resulting measurements—are affected by isostatic rebound, the gradual ongoing readjustment of the land surface that has followed the retreat of the thick ice sheets that covered the land during the last continental glaciation. The land surface elevation changes that result from the ongoing isostatic rebound vary by location. As a result, changes in sea level measurements along Canada’s coasts are far from uniform, and sea levels, as measured by the gauges, have both risen and fallen, depending primarily on whether the fixed points in question, on the local land surface, have been rising or sinking themselves (7.4). For example, the sea level fell in western Hudson Bay at a rate of 88 cm per century, due to extensive land uplift caused by the continuing rebound of the land (very high confidence) (7.4). In contrast, the sea level rose at rates of up to 34 cm per century in southern Atlantic Canada and the western Arctic, and at lower rates in British Columbia (very high confidence) (7.4).

Changes in sea level affect the frequency and magnitude of extreme coastal water-level events. In places where relative sea level has risen (most of the Atlantic and Pacific coasts and the coast of the Beaufort Sea in the Arctic), the frequency and magnitude of extreme coastal water-level events have increased (high confidence) (7.6). These extreme sea-level events often occur as a result of storm surge, tides, and waves combined, and their increase has resulted in greater flooding, leading to infrastructure and ecosystem damage as well as the erosion of sandy coastlines, putting communities at risk (7.6, 8.7.2, Case Story 2.1).

In the Arctic, mean and extreme wave heights have increased, primarily because of the climate change−driven reduction in sea ice (high confidence) (7.5), which allows near-surface winds to create waves by acting on the ocean surface. Wave heights in the northwest Atlantic Ocean have increased over the last few decades (medium confidence), connected with a northerly shift in the North Atlantic storm track, but these changes have not been directly attributed to human-caused climate change (7.5, 4.3). Wave height trends in the North Pacific are not statistically significant (7.5). Areas with increasing wave heights are also experiencing an increase in storm surges (medium confidence) (7.5).

The El Niño–Southern Oscillation (ENSO) is an irregularly recurring pattern of climate variability that involves episodic changes in water temperature in the tropical Pacific Ocean, affecting weather worldwide. The magnitude of ENSO, which is the dominant source of year-to-year climate variability in western Canada, has been relatively high since 1950 (medium confidence) (4.8). There is evidence that the pattern of large sea-surface temperature fluctuations associated with ENSO has changed during this period (i.e., a shift from so-called eastern-type ENSO events to central-type ENSO events) (4.8), which has implications for how ENSO affects weather in North America. Owing to wide internal variability and the absence of simulated ENSO trends in climate models, there is low confidence that these changes are due to human-caused climate change (4.8).

2.2.5: Carbon cycle changes

Under natural, unperturbed climate conditions, the concentration of CO₂ in the atmosphere is regulated by the cycling of carbon between the land, ocean, and atmosphere. This cycling maintains a long-term balance between natural fluxes of CO₂ from the land and ocean into the atmosphere and natural uptake of CO₂ from the atmosphere (9.2). The Earth’s fossil carbon reservoir is not part of the modern, active natural carbon cycle, but has been isolated from the land, ocean, and atmosphere for millions of years (9.2). The increase in atmospheric CO₂ concentrations since the start of the Industrial Revolution has been unequivocally caused by emissions from the burning of fossil fuels and human-caused land use change (9.3).

Canada’s overall anthropogenic GHG emissions peaked in 2007 at 774 Mt (megatonnes) (in CO₂-equivalent [CO₂-eq] units) and remained relatively constant at around 740 Mt CO₂-eq/yr between 2010 and 2019 (9.3). Emissions for 2023 (694 Mt CO₂-eq/yr; most recent year reported) were approximately 10% lower than their 2007 peak, despite continued growth in Canada’s population and economy (9.3). In 2023, Canada was the 10th highest global emitter, producing 1.4% of global greenhouse gas emissions (9.3). While Canadian emissions have decreased slowly since 2007, global emissions continue to increase. Not all anthropogenic emissions remain in the atmosphere. Together, the land and ocean have acted as carbon sinks that have absorbed more than half of global anthropogenic CO₂ emissions since the pre-industrial era (9.3). The remaining emissions that stay in the atmosphere increase the atmospheric CO₂ concentration and drive global warming.

Historically, in Canada, there has been a net uptake of carbon by the land; this carbon sink continues to function (medium confidence). However, extreme events like the 2023 wildfire season can cause the Canadian land mass to become a short-term source of carbon (low confidence) (9.5). On average, Canada’s ocean areas (i.e., inside the Exclusive Economic Zone) act as a net carbon sink (medium confidence), although they are a net source in some nearshore areas (low confidence) (9.5). Canada’s ocean is expected to continue to be a sink as long as atmospheric CO₂ concentration keeps increasing (9.5). On land, although carbon uptake from vegetation growth is expected to increase, so are carbon emissions from permafrost thaw, wildfires, and other disturbances. The net effect of future increases in both carbon uptake and emissions on land remains uncertain (9.5).

Global temperature, and with it Canadian temperature, will only begin to stabilize when global anthropogenic CO₂ emissions become net zero (9.4), meaning anthropogenic emissions are balanced by anthropogenic removals. Both Canada and the world would have to reduce their emissions far below current levels to achieve net zero anthropogenic emissions (9.4).

2.2.6: Confidence terms in key messages: Summary of evidence

Key Message 2.1: Canada is warming in all regions and seasons. Canada warmed by 2.0°C during the 1948–2023 period (very likely 0.9–3.1°C), with northern regions experiencing warming of 2.6°C (very likely 1.4–4.1°C) and generally more warming occurring in winter than in summer.

Key Message 2.2: Human influence has warmed Canadian average 2015−2024 temperatures to a level that is 2.2°C (likely 1.8–2.6°C) above that in the pre-industrial era, which is double the amount of warming in the global average temperature that is attributable to human influence when comparing 2010−2019 with global average temperatures in the pre-industrial era (high confidence). The observed warming in Canada of 2.0°C from 1948 to 2023 is very strongly dominated by human influences and indistinguishable from the warming that can be attributed to external influences on the climate (very high confidence).

Key Message 2.3: Most observed climate changes across Canada—on land, in the oceans around Canada, and in the atmosphere—are consistent with a warming climate. Emissions of greenhouse gases from human activity provide the only plausible explanation for this collective change (very high confidence).

Key Message 2.4: The observed occurrence in summer of longer growing seasons, increased building cooling requirements, and higher hot temperature extremes are all consistent with a warming climate, as is the occurrence in winter of reduced building heating requirements and less severe cold temperature extremes. Fire seasons have become longer (high confidence), and the area burned in wildfires has increased (medium confidence). The warming climate has also led to substantial changes in the water cycle and cryosphere (very high confidence), as indicated by increased annual precipitation and intensified precipitation extremes, snow cover and glacier mass reductions, earlier spring peak streamflow in rivers, decreasing lake and river ice, and thawing permafrost.

Key Message 2.5: The effects of the warming climate are also clearly seen in Canada’s ocean areas, through increases in ocean water temperature, more frequent marine heatwaves, freshening of near-surface ocean waters, and decreasing sea ice cover, which has resulted in increasing ocean wave heights in the Arctic (high confidence). Some of Canada’s coastal areas are also experiencing the effects of sea-level rise (very high confidence) and more frequent extreme sea-level events (high confidence), which can contribute to coastal flooding. The oceans around Canada have also absorbed anthropogenic carbon dioxide from the atmosphere, which has resulted in ocean acidification in the ocean’s near-surface layers (very high confidence).

There is extensive evidence that Canada’s climate has warmed substantially over the period since the middle of the 20th century. A primary indicator that is used to monitor climate warming is surface air temperature, which is measured at a height of about 2 m above the ground. Key Message 2.1 summarizes a few important aspects of past changes in annual and seasonal average temperatures for Canada that are calculated from carefully quality-controlled homogenized temperature data. There is very high confidence in this evidence, with justification for that assessment provided primarily in section 2.4.1. The high level of confidence in the data and the methods used to provide quantitative estimates of trends and their uncertainties further support an assessment of warming rate ranges that are very likely consistent with the observed warming.

There is also a very strong body of evidence that human influence on the climate, which is dominated by the warming effect of the observed increases in atmospheric GHG concentrations, is the primary cause of the observed warming. There is no remaining scientific doubt that human influence is responsible for warming our climate, both globally and in Canada. An uncertainty that remains for Canada, however, is that we have only very limited temperature observations during the 1850−1900 pre-industrial period that is used as a baseline in international climate policies. Advances in climate change detection and attribution research now allow us to confidently estimate a range of likely warming amounts between that period and the most recent full decade (2010−2019) that are due to human influence on the climate. These estimates are summarized in Key Message 2.2, which is based on the assessment in section 2.4.2. There is sufficient confidence in this evidence to also assess how the amount of human-induced warming in Canada between 1850−1900 and 2010−2019 compares with the amount of warming in the average global land and ocean temperatures caused by human influences over the same period. Observations are much more complete in post-Second World War, when most of the observed warming has occurred; this, together with studies of causes of warming performed over the past decade, allows an assessment of very high confidence in the overwhelmingly dominant role of human influence on Canada’s climate during the 1948−2023 period.

The observed warming has many knock-on effects, some of which further amplify the observed warming, a phenomenon that is seen strongly in Canada’s North and the Arctic in general. Section 2.2.1 summarizes the evidence for observed warming-related changes that have occurred across the climate system in Canada and are assessed in chapters 4 to 9 of this report. All of these changes have a common denominator, which is the observed warming of Canada’s climate. We understand this link through the knowledge of the physical processes that influence each variable and, for some variables, through detection and attribution studies. That evidence, together with the very strong evidence now available about the causes of the observed warming, leads to the simple, very high confidence, statement in Key Message 2.3.

Key Message 2.4 draws attention to some of the important impacts of the observed changes in the climate. These include the direct impacts of changes in temperature and more complex phenomena, such as changes in wildfire frequency and area burned, changes in the frequency and intensity of extreme events, changes in various aspects of the water cycle, and changes in the cryosphere. Process understanding and the amount of historical data limit confidence in some observed changes and our understanding of their causes, but all are underpinned by an understanding of the effects of human influences on Canada’s temperature and the amount of precipitation that we receive. Additionally, many changes observed in Canada are also observed in other global land regions, particularly those at similar latitudes, which further strengthens our confidence. Evidence supporting the assessments that appear in Key Message 2.4 is summarized in sections 2.2.2 and 2.2.3, with references to the other chapters of this report that provide more detail on the assessed level of confidence in each statement.

Finally, Canada’s land mass is bounded by oceans on three sides and has a climate that is strongly influenced by interactions between the ocean and the atmosphere. Changes in the ocean, including changes in its temperature, sea ice cover, sea level, salinity, and acidity, affect our climate over land, impact our coastlines, and alter ocean ecosystems. Key Message 2.5 is based on the summary of the evidence that appears in section 2.2.4 and is extensively considered and evaluated in Chapter 7 of this report.

In addition to the broad overview of observed changes that is presented in key messages 2.1 to 2.5, Figure 2.3 provides a visual overview of the observed changes and their assessed confidence levels, drawing on assessments of observed changes that have been made elsewhere in this chapter, and in chapters 4 to 9.

2.3: Sources of observational climate data

Our understanding of how Canada’s climate has changed since the mid-19th century depends heavily on historical weather observations that have been collected and archived over time. Early instrumental data are extremely limited. Some records from the 18th and 19th centuries are available for some locations in the St. Lawrence Valley (Slonosky, 2014). The longest instrumental weather record in Canada’s national archive comes from the Toronto Bloor Street observatory location, dating from 1840. More systematic observation of Canada’s weather across the southern tier of the country did not occur until the 1880s and 1890s with the development of our country’s railways. The post-Second World War economic boom in Canada and expansion of civil aviation in the 1950s led to a rapid increase in the number of weather stations, resulting in the broader coverage of Canada’s land mass. The number of surface weather observation stations contributing to Canada’s national digital climate data archive plateaued in the 1980s, before beginning a decline in the mid-1990s that continues to this day (section 2.7.2). In addition to the knowledge about changes in Canada’s climate that can be derived from instrumental data, knowledge about past and present changes in the climate is also possessed by Indigenous Knowledge Holders. Such knowledge was not assessed in this chapter, which we acknowledge to be a gap in the current assessment.

The history of changes in Canada’s climate and weather observation systems over time is long and complicated. Consequently, confident estimates of long-term changes in weather variables cannot be directly calculated from raw weather observations, even when long-running records are available for a specific location. The confident calculation of these changes requires an understanding of how recorded observations have been affected by factors such as changes in instruments and their exact locations, as well as in procedures for observing, reporting, and archiving data, and performing quality control.

In addition to weather observations collected at surface weather stations, other sources of data are available that are either derived from weather observations or in some way informed by those observations. Gridded station data products provide the most complete coverage of a region or the entire country through the spatial interpolation of climatic data from existing weather observations. Reanalysis data are another type of gridded data that are created by feeding weather observations into a weather forecasting system used to process an entire historical archive of weather observations. Data products derived from remote sensing systems, including satellites and radar systems, also provide weather data from the more recent past.

2.3.1: Climate and weather station data

In Canada, climate data are often obtained from weather station observations using instruments that, historically, were operated manually, but more recently, have been operated automatically. The country also has a limited number of reference climate stations that adhere to particularly rigorous measurement standards. These reference stations are designed and operated primarily to monitor climate conditions, and typically provide daily data, such as daily maximum and minimum temperatures and daily rainfall and snowfall or total precipitation amounts. In addition, in the past, volunteer observers have contributed many weather observations, and community-organized networks of volunteer observers continue to collect such data. Notably, the Collaborative Rain, Hail and Snow (CoCoRaHS) network provides some of the observations that have found their way into the precipitation datasets used in this chapter.

Factors such as changes in observing procedures, instrumentation, and station exposure may affect even the best records, causing recorded values to change in ways that are not related to climate change. Station locations may also change, which can also affect recorded values. Therefore, long-running climate records should always be carefully quality controlled and adjusted. As many non-climatic influences as possible must be eliminated before these records are used to estimate changes in the climate during the period for which instrumental observations are available (see Box 2.1 for examples of such adjustments).

Quality-controlled and homogenized long-term station data that are suitable for studying climate change are limited in Canada. Figure 2.4 shows where such long-term surface air temperature, precipitation, surface wind speed, and freezing precipitation frequency data are available. Long-term wind speed and freezing precipitation frequency data are particularly scarce. A station must meet several criteria to be considered a long-term observing station, including its chance of being operated and well maintained into the distant future (i.e., future-proof) and the availability of quality data records at the station and nearby stations to form a long-term record (Wang et al., 2023). At least 20 years of continuous monthly data during the 1961–1990 period are required so that climate normal values^{Footnote 4} can be calculated and the station data can be used in gridding (Wang et al., 2023; Wang, Feng, Zwiers, et al., 2026). Periods of station activity and inactivity differ by location, and many stations are currently not active (section 2.7.2; Box 2.2). As Figure 2.4 shows, the southern part of Canada has the best coverage, which poses challenges in monitoring climate change in other areas. Spatially interpolating the homogenized data onto a standardized grid—although it does not generate data of the same quality as directly observed changes—can help to provide information about the climate in areas that are not well observed, and has been done for temperature and precipitation (section 2.3.3).

Figure take-away: Locations with long-term, high-quality precipitation, temperature, wind, and freezing precipitation frequency data are unevenly distributed across the country.

Figure title: Long-term observing stations for monitoring changes in precipitation, temperature, wind, and freezing precipitation frequency in Canada

2.3.2: Remotely sensed data

The American Meteorological Society (AMS, 2025) defines remote sensing as “a method of obtaining information about properties of an object without coming into physical contact with that object.” Remotely sensed data include those obtained from satellites and weather radar. Remote sensing technology has improved steadily over the years, with significant advances in the sensitivity and spatial resolution of both ground-based and space-borne sensors. However, these improvements have led to challenges in using the resulting data to study climate change, due to the numerous inhomogeneities caused by changes in the technology. Nevertheless, some climate data records spanning the entire satellite era, starting in 1979, are available. One example is the set of lower, middle, and upper troposphere temperature records developed by Remote Sensing Systems that are used extensively in large-scale climate research (e.g., see Santer et al., 2017 and references therein). Using remote sensing data to create climate records that reliably and accurately depict long-term trends and low-frequency climate variability is a painstaking and exacting process. For example, in the case of the tropospheric temperature record, records have to be pieced together from a long sequence of data from different short-lived instruments (a few years for each satellite mission), each with different technical, orbital, and calibration characteristics, which may even change over the life of an individual instrument.

While remotely sensed data are an important asset in many aspects of climate research, we do not use these data directly in this chapter because our primary intent is to assess and understand observed changes in Canada’s surface climate. Remote sensing data are not well suited for this purpose, due to the relatively short period sometimes covered and challenges in measuring surface air temperature and precipitation remotely.

Satellites have been relatively successful in measuring precipitation and near-surface wind over the oceans, but historically less successful in doing so over land. Precipitation is often inferred from cloud top temperatures, which are more directly related to precipitation over oceans, particularly in tropical and subtropical regions. Wind speed over oceans is often inferred from measurements of ocean surface roughness, which increases with wind speed.

Weather radar is an important source of information for monitoring current weather, but is not widely used for climate research and monitoring. Archived radar imagery is available for Canada from 2007 onward (MSC, 2024). Historical radar imagery can be useful when studying extreme events, for example, when validating very high, and thus potentially erroneous, rain gauge measurements (e.g., see Canadian Standard Association, 2025). However, these data do not currently serve as a viable source of information on long-term climate change, due to the archive’s limited duration and incomplete spatial coverage (focusing on major population centres and airports), and inhomogeneities in the data resulting from technological improvements over time.

In short, while remote sensing data have many important applications in climate research and our understanding of climate processes, they do not yet provide a viable alternative to observations made at meteorological stations that document long-term changes in the primary variables—temperature, precipitation, and surface wind speed—that are used to monitor changes in Canada’s climate.

2.3.3: Gridded data products

Station data (section 2.3.1) consist of observations at scattered individual locations, which are not distributed uniformly across the Canadian landscape (Figure 2.4). For many applications, having monthly values for all locations in the country is preferable and can be achieved by superimposing a grid (or mesh) on the country’s territory and then estimating a value for each grid box from the station data available. For example, recent gridded datasets created by the Climate Research Division of Environment and Climate Change Canada (ECCC) (Supplementary Table S2.1) cover the country with grid boxes that measure roughly 10 km x 10 km, forming grid cells that have an area of about 100 km² each. These products are attractive because they provide information for all locations in the study area and all time intervals covered by the product, whether or not observing stations are present. This makes it possible to estimate patterns of change across geographical areas and to calculate regional averages without having to consider the uneven spacing of observing stations.

All gridding methods involve the integration of observations with a statistical or dynamical model of some type. Some products focus on a single variable and use relatively simple statistical models to perform spatial interpolation—a statistical procedure that predicts values based on data from nearby locations and sometimes considers additional information such as surface elevation to help explain spatial variations in temperature and precipitation (Abbasnezhadi & Wang, 2024). Examples include the gridded homogenized daily and monthly surface air temperature products and gridded homogenized monthly precipitation products used in this chapter. We rely on these gridded products because they are produced with well-documented, reproducible techniques and rely on single sources of extensively researched, high-quality climate observations that have been carefully adjusted to remove non-climatic influences (section 2.3.4).

Other gridded products, such as reanalysis products (more details in section 2.3.4.4), use physically based models of the atmosphere (and sometimes other parts of the climate system). These products attempt to estimate all atmospheric variables simultaneously by integrating observations into the atmospheric general circulation models used in weather forecasting. Weather forecasting centres routinely produce analyses of the current weather conditions. Those analyses are archived but cannot be used to monitor climate change because the forecasting models, data assimilation systems, and data sources keep changing. To control for this, reanalyses use fixed weather-forecasting models and data assimilation systems (section 2.3.4.4), which allows them to avoid inhomogeneities due to changes in those systems, although they are still subject to potential inhomogeneities due to changes in data sources over time.

2.3.4: Data used in this chapter

This subsection describes the temperature, precipitation, and wind speed data that were used to assess observed changes in the corresponding surface climate variables in sections 2.4, 2.5, and 2.6, respectively. In the case of precipitation, we considered total precipitation; proxy snowfall, which was derived from temperature and precipitation data (Box 2.3); and freezing precipitation frequency, which was obtained from hourly reports by human observers without using instruments. In all cases, we relied on primary observations recorded at meteorological stations and systematically collected and archived by the Meteorological Service of Canada, with the earliest observations dating back to the mid-19th century in some locations in southern Canada. Before the station data were used, great care was taken to adjust the data to remove non-climatic factors as much as possible and to address changes in observing instruments and technologies, observing procedures, and other non-climate factors that could affect trends and patterns inferred from observations. For a summary of the various datasets used and their evolution, see Supplementary Table S2.1.

2.3.4.1: Surface air temperature data

The temperature data used in this chapter consist of long-term records of surface air temperature measurements from observing stations. These data have been quality controlled and carefully analyzed and adjusted to reduce the effects of non-climatic changes in the data records (Box 2.1) (Vincent et al., 2020; Wang, Feng, Zwiers, et al., 2026). Surface air temperatures are typically measured at a height of about 2 m above the ground, inside an enclosure that protects the instrument from precipitation and direct solar heating, while allowing the air to circulate freely. Stations with long-running homogenized temperature records are unevenly distributed across the country and cover different time periods (Figure 2.4, Box 2.2 Figure 1a). Observing stations tend to be more densely distributed in the populated portion of southern Canada, and some have had continuous daily temperature records since the mid- to late 1800s (e.g., since 1840 for a site in Toronto). However, stations are sparsely distributed in northern Canada, with few stations operating before 1948. The number of stations in Canada with data was 1 before 1865, 6 in 1870, up to 28 before 1880, up to 46 before 1890, and between 47 and 77 between 1890 and 1899. Thus, there were insufficient observations before 1900 to produce representative gridded temperature data for trend analysis even for Canada’s South.   

Techniques for adjusting station data to remove non-climatic influences continue to evolve. As the climate data record lengthens, our understanding of how non-climatic factors can affect observations deepens and the statistical techniques used to remove those effects improve (Supplementary Table S2.1). Consequently, ECCC’s Climate Research Division produced a third-generation homogenized station dataset of surface air temperatures (CanHomTV3, including monthly and daily data) (Vincent et al., 2020). This dataset was updated to 2023, and data from 632 of the 778 stations were used to produce CanGridT V3.1 (monthly and daily Canadian gridded homogenized temperature data, version 3.1), the version used in this chapter. In contrast, the first edition of CCCR (CCCR 2019) (Zhang et al., 2019) employed Canadian Gridded (CanGRD) temperature anomaly data (ECCC, 2016), which were based on the second generation of homogenized monthly average temperatures at 338 stations (Vincent et al., 2012) updated to 2016 (see Supplementary Table S2.1a).

CanGridT improves on CanGRD in two ways. First, CanGridT uses a finer mesh, reducing the grid box area from approximately 2500 km² to 100 km². Second, CanGridT provides estimates of monthly average temperatures rather than just departures from normal temperatures (with normal defined as the 1961–1990 climate normal; see footnote 4). This was achieved by gridding the 1961–1990 climate normals separately from the homogenized station data and the monthly differences (anomalies) from those normals and then adding the gridded normals and gridded anomalies together. This two-step gridding process helped overcome some of the issues caused by changes in the spatial and temporal distribution of observing stations.

The CanGridT mlyV3.1 gridded dataset (Supplementary Table S2.1a) (updated from Vincent et al., 2020; Wang, Feng, Zwiers, et al., 2026) provides monthly temperature values from 1900 onward for southern Canada and from 1948 onward for all of Canada. The greater station density in southern Canada results in higher confidence in the trend estimates for that part of the country, particularly for the period since 1948, when more stations were added. The representativeness of the trends obtained was assessed (Box 2.2) (Wang, Feng, Zwiers, et al., 2026), in order to address concerns about the effects of changes in station data availability in the CanHomT mlyV3.1 dataset, on which CanGridT mlyV3.1 is based (Supplementary Table S2.1a). The assessment found that the sampling biases caused by changes in station data availability during the periods used to determine trends in this chapter were small in comparison to the corresponding trend confidence range (Box 2.2 Figure 2a), and thus a sampling bias correction was not applied.

An additional gridded dataset, CanGridT mlyV4, which begins in 1948, has also recently become available (Wan et al., 2025). An analysis of the differences between it and five other surface air temperature datasets clearly shows that the evolution of annual average daily mean temperature for Canada as a whole in CanGridT mlyV4 is different from that in other datasets between 1948 and about 1970 (Figure 2.5a,b), with the CanGridT mlyV4 dataset showing anomalous warming relative to the other five datasets considered. This results in a higher warming trend (2.25°C) in Canada’s annual average temperature over the 1948−2023 period in CanGridT mlyV4. Trends for the same period in other datasets are 1.95°C (CanGridT mlyV3.1), 1.95°C (C-LSAT2.0, in Sun et al., 2021), 1.88°C (CRUTEM5 v5.0.2.0, Osborn et al., 2021), 2.03°C (NOAA MLOST v6.0, Yin et al., 2024), and 1.95°C (ERA5, Hersbach et al., 2020). Five of the six datasets intercompared are based on station data that have been processed in different ways, while the sixth dataset, the ERA5 reanalysis dataset (section 2.3.4.4), uses a very broad range of atmospheric observations in its analysis. The warming seen in the CanGridT mlyV3.1 dataset is much more consistent with that observed in the four international datasets, which are constructed in different ways, than the warming seen in CanGridT mlyV4. CanGridT mlyV3.1 was also found to represent Canada’s warming trend since 1900 reasonably well (Box 2.2). Thus, our assessment of past warming in Canada is based on CanGridT mlyV3.1.

Figure take-away: Gridded surface air temperature datasets can sometimes differ systematically from each other.

Figure title: Differences in the results for Canada as a whole among four internationally produced surface air temperature datasets and two versions of CanGridT

2.3.4.2: Precipitation data

The primary sources of station precipitation data used in this chapter include version 2 of the Canadian Homogenized Precipitation monthly and daily datasets, CanHomP mlyV2 and CanHomP dlyV2 (Wang, Feng, Zwiers, et al., 2026; Wang & Feng, 2026), and the daily proxy snowfall dataset CanHomSp anlV2 (Box 2.3) (Qian et al., 2025). Each CanHomP dataset consists of 425 long-term records of monthly or daily total (liquid and frozen) precipitation amounts from observing stations. The gridded version of the CanHomP mlyV2 dataset, referred to as CanGridP mlyV2 (Wang, Feng, Zwiers, et al., 2026), is the primary gridded dataset used in this chapter.

The precipitation data in the CanHomP datasets have been quality controlled and carefully analyzed and adjusted to reduce the effects of non-climatic changes on the data (Box 2.1) (Wang, Feng, Zwiers, et al., 2026; Wang & Feng, 2026; Wang et al., 2023). While rigorous methods have been used to make these adjustments, confidence in the observed changes is nevertheless lower for precipitation than for temperature.

Precipitation is measured by a variety of instruments that are installed on or near the ground and are exposed to the open air above. They measure the amount of precipitation falling into a container or, in the case of some types of snow measurements, onto a surface. Precipitation measurements are affected by a variety of factors. One factor is the phase of precipitation (liquid or frozen): rainfall is generally easier to measure than snowfall and the amount of water equivalent that it represents. Wind speed also influences precipitation measurements because gauges generally catch precipitation less efficiently in windy conditions. To a lesser extent, precipitation measurements are also affected by losses from evaporation and from the wetting of the interior surfaces of the instrument (Mekis & Vincent, 2011; Milewska et al., 2019; Wang et al., 2017). The accuracy of precipitation measurements is also affected by the wide range of climate conditions in Canada, where annual precipitation amounts can range from more than 3000 mm (primarily rain) in some west coast locations to less than 200 mm (primarily snow) at some Arctic locations (Supplementary Figure S2.3). Climate conditions affect accuracy, because gauge performance differs depending on the rate and phase of precipitation. The wide variations in climate conditions, the substantial differences in the instruments that have been used to measure precipitation, and the large differences in precipitation amounts that can occur even when stations are separated by only short distances, make the adjustments to remove non-climatic factors more difficult for precipitation than for temperature (Wang, Feng, Zwiers, et al., 2026; Wang & Feng, 2026).

Stations with high-quality, long-term precipitation data are unevenly distributed across the country and have operated over different time periods (Figure 2.4; Box 2.2 Figure 1b). Precipitation records for some locations in Canada extend back for more than a century, but most records have more recent starting points, and a substantial number of long-term observing sites have been discontinued. While ECCC collects and archives data at many observing stations at any given time, including more than 2500 active stations in recent years, only a few hundred locations have long-term records. The long-running records available at the 425 long-term stations contributing to the CanHomP datasets were often created by joining records from nearby stations reporting for different periods; this results in longer single records, which are quality controlled and adjusted to make them suitable for climate analysis (Box 2.1) (Wang, Feng, Zwiers, et al., 2026; Wang et al., 2023).

CanGridP mlyV2 (Wang, Feng, Zwiers, et al., 2026) provides estimates of monthly total precipitation in each 100 km² grid box using a gridding method developed by Abbasnezhadi and Wang (2024). An assessment of the representativeness of the trends obtained with CanGridP mlyV2 (Box 2.2) showed that this dataset represents precipitation trends in Canada as a whole since 1949 and in southern Canada since 1916 reasonably well (Wang, Feng, Zwiers, et al., 2026). Therefore, it was used in this chapter to evaluate precipitation trends in Canada and northern Canada in 1949–2023, and precipitation trends in southern Canada in 1916–2023. CanGridP mlyV2 improves on the approach used in the previous version of CanGRD (ECCC, 2016), which provided gridded relative monthly precipitation anomalies derived from unhomogenized precipitation data exclusively (see Supplementary Table S2.1c).

A particular challenge in assessing precipitation trends is that individual daily and monthly observations provide less information about precipitation at nearby locations than the corresponding temperature observations. In other words, precipitation observations are much less spatially representative than temperature observations. This makes it more challenging to estimate precipitation amounts and changes in amounts at locations with no observations by using spatial gridding procedures than it would be for temperature. Consequently, estimated regional and national precipitation trends derived from CanGridP mlyV2 are inherently more uncertain than the corresponding temperature trends. While the focus of this chapter is on station data, we recognize that recent high-resolution reanalysis datasets, such as the Canadian Surface Reanalysis dataset (CaSR, 2025) and ERA5-Land dataset (Muñoz-Sabater et al., 2021), have the potential to improve confidence in estimates of local and regional changes in precipitation.

2.3.4.3: Wind speed data

The primary source of the near-surface wind speed data used in this chapter is the Canadian Homogenized Surface Wind Speed (CanHomW) mlyV2 dataset, which includes long-term records of homogenized wind speed observations at 154 locations (Figure 2.4; Box 2.1) (Wang, Feng, Isaac, et al., 2025). Surface winds are observed at meteorological stations with instruments called anemometers, which are placed 10 m above the ground in a restricted area free of obstacles that might alter the atmospheric flow in the vicinity. The wind speed data used in this chapter primarily come from anemometers that measure the rotation rate of mechanical cup wheels. Although such measurements are generally reliable, they can be affected by several types of mechanical issues. For example, ice accretion can hinder the motion of the cups, resulting in wind speed readings that are too low, or can even cause the cups to completely seize up, resulting in readings of zero (ECCC, 2015). Monthly average wind speed data were derived from quality‑controlled hourly surface wind speed data taken from the Digital Archive of Canadian Climatological Data maintained by ECCC. The reported hourly values consist of two-minute average wind speeds that are recorded at the top of each hour (ECCC, 2015). A gridded wind speed dataset derived from homogenized station data has not been produced due to the small number of locations with long-term homogenized records.

2.3.4.4: Reanalysis datasets

Reanalysis is a scientific method that uses the physical relationships between different climate variables, as described in a weather or climate model, to systematically combine multiple sources of weather and climate observations (from surface/subsurface to upper air, including satellite and marine observations) to create a more complete representation of the evolving historical climate conditions extending from the Earth’s surface/subsurface to well above the stratosphere. A reanalysis dataset represents a four-dimension reconstruction of historical climate conditions over the period of reanalysis.

Three secondary sources of data on historical monthly mean temperatures, precipitation amounts, and surface wind speeds were also used in this chapter, recognizing that these sources may be subject to inhomogeneities and other uncertainties: the ERA5, 20CRv3, and OCADA reanalysis datasets. ERA5 is the fifth-generation atmospheric reanalysis dataset for the global climate produced by the European Centre for Medium-Range Weather Forecasts (ECMWF) (Hersbach et al., 2020); 20CRv3 is the third version of the US National Oceanic and Atmospheric Administration (NOAA) Twentieth Century Reanalysis dataset (Slivinski et al., 2019); and OCADA is the Over-Centennial Atmospheric Data Assimilation (OCADA) reanalysis dataset, produced by Japanese institutions (Ishii et al., 2024).

ERA5 uses a fixed version of the ECMWF weather forecasting model and data assimilation system, providing detailed, high-resolution (~31 km) analyses of the full atmospheric state from 1940 to the present, including hourly values of some variables. It uses an extensive archive of historical in situ and satellite observations. For this reason, some aspects of the analysis, such as surface air temperature and precipitation, are subject to potential inhomogeneities due to the evolution of the global observing system over time. The 20CRv3 data product uses a fixed version of the US NCEP (National Centers for Environmental Prediction) weather forecasting model and data assimilation system. It provides moderate resolution (~75 km) analyses of the full atmospheric state for the 1836–2015 period, with sampling three times an hour. OCADA is like 20CRv3 but uses a Japanese atmospheric model to provide moderate resolution (~60 km) analyses of the full atmospheric state.

Both the 20CRv3 and OCADA datasets integrate sea surface temperature and atmospheric surface pressure observations in their weather forecasting models. These two aspects of the climate have some of the longest and most well-studied instrumental records. The advantage of using only these two sources of observational data is that these reanalyses should be less affected by changes in the global weather observing system than those based on a larger suite of historical observations. The disadvantage, however, is that these reanalyses do not benefit from the improvements in observing systems that began, notably in 1979, with the introduction of weather satellite data. As a result, many of the variables used in the 20CRv3 and OCADA reanalyses to represent the state of the atmosphere will be much more strongly influenced by the weather forecasting models used in the reanalyses than would be the case in a reanalysis that relies more heavily on archived climate and weather data, like ERA5.  

We used reanalyzed data cautiously in this chapter, because they may contain inhomogeneities and the weather forecasting models employed may have biased some of the variables. Since a gridded dataset of monthly average wind speeds derived from homogenized station data was not available, we cautiously used ERA5, 20CRv3 and OCADA reanalyses as sources of gridded historical monthly mean surface wind speed data. Circulation in the free atmosphere, where air motion is only negligibly affected by friction at the planet’s surface, is generally well represented in reanalyses, and the three-dimensional wind fields and related variables from reanalyses are often used interchangeably with upper air observations from radiosondes and other sources. This is not, however, the case for near-surface wind speeds (i.e., at the 10 m anemometer height), as the wind near the surface is strongly affected by the roughness and topography of the terrain. The calculation of wind speeds in reanalyses is affected by several limitations, such as the resolution of the spatial and vertical grids used in the weather models, which implies that fine details are not well represented. In addition, some physical processes affecting surface wind speeds are not well represented in the weather forecasting models. For example, the grids that are used are not fine enough to explicitly represent individual thunderstorms and the strong local winds that they can produce. Therefore, as was the case for precipitation, we used anemometer-height wind data from reanalyses with caution.

2.4: Past temperature changes

Key Message 2.6: Canada’s climate has warmed rapidly over the past 75 years. Warming rates have fluctuated over the past century. Annual mean temperatures are estimated to have increased at a rate of 0.26°C (very likely 0.12–0.41°C) per decade between 1948 and 2023 for Canada as a whole and 0.35°C (very likely 0.18–0.55°C) per decade for Canada’s North.^{Footnote 6} Canada’s estimated annual average temperature has exceeded the 1961–1990 average in 30 of the 33 years after 1990.

Key Message 2.7: Canada has warmed more quickly than most of the rest of the world. Canada’s warming rate over the 1948–2023 period is similar to that for the global land area (high confidence), but Canada’s North has warmed at a rate that is about twice the rate for the global land and ocean area combined over the 1948–2023 period (high confidence). Most of the observed warming has occurred since 1970. Canada’s warming rate was almost twice the global warming rate over the 1970–2023 period (high confidence).

Key Message 2.8: Human influence has warmed Canadian average temperatures during the 2015–2024 period to a level that is 2.2°C (likely 1.8–2.6°C) above the pre-industrial era (approximated in this report as the period from 1850 to 1900). Human influences acting on their own would likely have caused more warming than the observed warming of 1.6°C between the decades of 1948−1957 and 2014−2023, with natural external influences and internal climate variability combined preventing a small portion of that influence from being realized (high confidence).

This section describes how surface air temperatures in Canada have changed, makes comparisons with the changes observed in global average temperature, and assesses the extent to which human influence on the climate is responsible for the changes seen in Canada. The CanGridT mlyV3.1 gridded temperature dataset (section 2.3.4.1) was used primarily for the analyses and assessment, considering the 1900–2023 period for Canada’s South and the 1948–2023 period for Canada as a whole, which correspond to the periods with sufficient long-term homogenized station data to produce a gridded dataset for these areas.

2.4.1: Past changes in average surface air temperatures

Annual average temperatures have increased almost everywhere in Canada since 1948, and in most regions this increase is statistically significant at the 5% level (i.e., there is ≤ a 5% chance of concluding that a similar effect or trend exists when it does not) (Figure 2.6). The annual average temperature for Canada as a whole changed little from 1948 to 1970, but increased rapidly after that, a trend also found in the ERA5 reanalysis data (Figure 2.7c,d). The ten warmest years in the 1948–2023 period have occurred since 1981. The first, second, and third warmest years were 2010, 2023, and 2006, with temperatures 3.05°C, 2.91°C, and 2.46°C, respectively, above the 1961–1990 baseline average temperature. Canada’s annual average temperature has exceeded the 1961–1990 average in 30 of the 33 years after 1990 (the exceptions are 1992, 1996, and 2004, when temperatures were slightly below the average).

In Canada as a whole, annual average surface air temperatures increased at a rate of about 0.26°C per decade (2.0°C total) between 1948 and 2023. Almost all of this warming has occurred since 1970 (Figure 2.7a,c), with the result that Canada experienced a warming rate of about 0.37°C per decade (2.0°C total) between 1970 and 2023. These trend estimates are associated with relatively large uncertainties, even when considering Canada as a whole, which is reflected in the associated 95% uncertainty ranges reported in Table 2.1a (Wang, Feng, Zwiers, et al., 2026). These estimates may also be affected by data processing and gridding method choices. These sources of uncertainty are difficult to evaluate, but should be small when comparing well-researched data products. For example, there is only a very minor difference in the 1948–2023 trend of 0.26°C per decade estimated from CanGrid mlyV3.1 and the trends of 0.26°C, 0.26°C, 0.25°C, and 0.27°C per decade estimated from ERA5, C-LSAT2.0, CRUTEM5, and NOAA-MLOST, respectively (see section 2.3.4.1 and Figure 2.5 for these datasets).

Figure take-away: Canada has warmed over the periods when sufficient temperature data are available, with greater warming in the north.

Figure title: Changes in annual average daily mean temperatures across Canada

Figure 2.6: Maps of observed changes in annual average daily mean temperatures between a) 1948 and 2023 (°C/decade, left-hand scale, and °C, right-hand scale) and b) 1900 and 2023 (°C/decade, left-hand scale, and °C, right-hand scale). Changes are based on linear trends over the respective periods, estimated using the method described by Wang and Swail (2001).^{Footnote 7} Dots show areas where trends are not statistically significant at the 5% level (i.e., there is ≤ a 5% chance of concluding that an effect or trend exists when it does not). The percentages given in the panel titles are the percentage of grid points with significant positive and negative trends, respectively. The data for northern Canada are insufficient to confidently calculate warming trends from 1900 to 2023. See Supplementary Figure S2.1 for corresponding estimates of the changes in annual average daily minimum and daily maximum surface air temperatures. Data source: CanGridT mlyV3.1 (updated from Vincent et al., 2020; Wang, Feng, Zwiers, et al., 2026).

We assess that substantially greater warming has occurred in Canada than in the contiguous United States (high confidence), where the annual average temperature increased at a rate of 0.18°C per decade during the 1948–2023 period (National Time Series from NOAA NCEI, 2024). The upper bound on the estimated 95% uncertainty range (0.11–0.24°C) of this warming lies below the best estimate of the corresponding warming rate for Canada of 0.26°C per decade. The confidence in this trend is based on both the substantial difference in observed warming trends and the understanding of the role that Arctic amplification processes play in accelerating Canadian warming relative to that of land areas at lower latitudes (Chapter 4, section 4.2).

Figure take-away: All datasets considered agree that Canada has warmed substantially during the period when data are available.

Figure title: Changes in annual average temperatures for all of Canada and Canada’s South

Table 2.1: Observed changes in annual and seasonal average daily surface air temperatures over the indicated periods for Canada as a whole, its six regions, and Canada’s South (see Figure 2.2 for the definition of the regions). a) Daily mean surface air temperature, b) daily minimum surface air temperature, and c) daily maximum surface air temperature. Changes over the period are represented by linear trends, with the corresponding 95% confidence intervals shown in parentheses (see footnote 7). Data source: Estimates from CanGridT mlyV3.1 gridded station dataset, updated from Vincent et al. (2020) by Wang, Feng, Zwiers, et al. (2026).

There is high confidence that Canada has warmed at a rate similar to that for the global land area, which implies that it has warmed more quickly than many other countries. Estimates of the 1948–2023 warming rate for the global land area based on three datasets are 0.25°C per decade (0.20–0.29°C; NOAA-MLOST v6.0.0) (NOAA NCEI, 2024; Yin et al., 2024), 0.25°C per decade (0.20–0.30°C; C-LSAT2.0) (Q. Li et al., 2021; Sun et al., 2021), and 0.21°C per decade (0.17–0.25°C; CRUTEM5 v5.0.2.0) (Osborn et al., 2021).^{Footnote 8} Again, the confidence in this assessment is based both on differences in estimated warming rates and the understanding of the role of Arctic amplification.

In addition, Canada has warmed about 1.7 times faster than the globe as a whole, including land and oceans. Estimates of the 1948–2023 global warming rate are 0.15°C per decade (0.13–0.17°C) (NOAA NCEI, 2024), 0.15°C per decade (0.13–0.18°C; CMST2.0) (Q. Li et al., 2021; Yun et al., 2019), and 0.16°C per decade (0.13–0.19°C; HadCRUT5 Analysis version 5.0.2.0) (Morice et al., 2021). As in Canada, most of the warming of the globe as a whole since 1948 has occurred since 1970. Estimates of the 1970–2023 global warming rate are 0.19°C per decade (0.16–0.21°C; NOAA-MLOST v6.0.0) (NOAA NCEI, 2024; Yin et al., 2024), 0.20°C per decade (0.18–0.22°C; CMST2.0) (Q. Li et al., 2021; Yun et al., 2019), and 0.20°C per decade (0.18–0.22°C; HadCRUT5 Analysis version 5.0.2.0) (Morice et al., 2021). As reported above, the warming rate for Canada for the same 1970−2023 period is 0.37°C per decade (0.15–0.59°C; Table 2.1a). On the basis of these comparisons, our assessment is that Canada warmed almost twice as fast as the globe as a whole during the 1970−2023 period (high confidence).

Most countries have experienced a faster rate of warming than the global average simply because of the pattern in which warming has occurred, with land areas warming more quickly than ocean areas, which is an expected feature of the climate system’s response to increasing greenhouse gas (GHG) concentrations.^{Footnote 9} This pattern has been observed and has been simulated by every generation of climate models assessed by the IPCC since the organization’s inception.^{Footnote 10} It occurs because the ocean has a very large heat capacity that allows it to absorb much more of the energy from GHG-induced warming than the Earth’s land surface. Canada has experienced substantially more warming than global land areas, on average, because our northern high-latitude climate is strongly affected by Arctic amplification processes (Chapter 4, section 4.2).

During the 1900–2023 period, we estimated that the annual average temperature in Canada’s South increased by 1.85°C (0.15°C per decade), with increases of 3.08°C in winter, 1.85°C in spring, 1.60°C in summer, and 1.35°C in fall (0.25°C, 0.15°C, 0.13°C, and 0.11°C per decade, respectively). Annual average temperatures in Canada’s South did not rise steadily over this period (Figure 2.7b). Temperatures increased until the 1940s, remained steady until the 1970s, and have subsequently increased rapidly (see section 2.4.2 for a discussion of the drivers of these changes). Annual average temperatures in Canada’s South have continually remained above the 1961–1990 average since 1997 (Figure 2.7b), which would have been virtually impossible in the absence of a systematically warming climate system. This long-term behaviour of temperatures in Canada’s South is consistent with that observed in the United States and globally. See Figure 2.11 in Gulev et al. (2021) for the global time series and NOAA NCEI (2024) for the national and global time series plots. 

Substantial seasonal and regional variations are found in observed temperature trends across Canada (see Figure 2.2 for definitions of the regions considered in chapters 2, 3, and 8). Warming has been greater in Canada in winter and fall than in spring and summer. Between 1948 and 2023, the annual average temperature in Canada’s land area as a whole is estimated to have increased by 3.38°C (0.45°C per decade) in winter, 1.58°C (0.21°C per decade) in spring, 1.65°C (0.22°C per decade) in summer, and 1.95°C (0.26°C per decade) in fall (the corresponding 95% confidence intervals are listed in Table 2.1a). The magnitude of warming also varies by region. The strongest annual average warming was observed in northwestern Canda, where the annual average temperature increased by 2.6−3.5°C (0.35–0.47°C per decade) in some areas. Eastern Canada has experienced substantially less warming, with estimated increases in some parts of northern Ontario and Quebec, and central Labrador and Newfoundland, which are not statistically significant (Figure 2.6a). It should be noted that while there is no doubt that Canada has warmed since 1948, confidence in estimates of the seasonal and regional variations in warming is somewhat lower because of spatial and temporal variations in the number of observing stations (Figure 2.4 and Box 2.2 Figure 1a).

The annual and seasonal patterns of change in annual average daily minimum and maximum temperatures are broadly similar to those seen in daily mean temperatures (Supplementary Figures S2.1 and S2.2). The regional and seasonal variations in trends are also similar among temperature variables (Table 2.1b,c), with the strongest warming occurring in winter, and much higher warming in Canada’s North than in Canada as a whole. Canada has also experienced more warming in daily minimum temperatures than in daily maximum temperatures, particularly in winter, which is consistent with findings in other parts of the world. Daily minimum winter temperatures (i.e., the minimum temperature of winter nights) are estimated to have increased at a rate of 0.48°C per decade (3.6°C in 1948–2023), compared to 0.40°C per decade (3.0°C in 1948–2023) for daily maximum winter temperatures. In addition, the difference in the warming rates for daily minimum temperatures in winter and daily maximum temperatures in summer (i.e., the temperature of summer days) is particularly striking, with daily maximum summer temperatures estimated to have increased at a rate of 0.21°C per decade, or 1.6°C over the 1948–2023 period. In contrast with winter, in summer, daily minimum temperatures have increased only marginally faster than daily maximum temperatures (0.22°C per decade versus 0.21°C per decade).

Regional differences in warming are most pronounced in winter and least pronounced in summer, with substantial seasonal variations in the warming pattern (Figure 2.8). Winter warming is particularly notable in Yukon, the Northwest Territories, northern British Columbia and Alberta, and western Nunavut, ranging from 4 to 7.5°C during the 1948−2023 period (or 0.55 to 1.02°C per decade). Regional warming patterns are similar in spring, but smaller in magnitude and not statistically significant across larger areas. Summer warming is comparable in magnitude to that in spring but is statistically significant almost everywhere in Canada and generally more uniform across the country than during other seasons. During fall, substantial warming is observed in Nunavut, the Northwest Territories, and southeastern Canada, while changes are not significant in most areas of British Columbia and the Prairies. These differences in the seasonal patterns and intensities of warming are broadly consistent with the understanding of the characteristics of the climate system (section 2.4.2; Chapter 4). Temperature changes across Canada tend to be more uniform in summer than in winter because high-latitude warming, which is strongly amplified in winter due to Arctic amplification (Chapter 4, section 4.2), is limited in summer by the ongoing melting of sea and land ice.

Figure take-away: The amount of warming experienced in Canada since 1948 varies by season, with the greatest warming occurring in winter in most places.

Figure title: Trends in seasonal average daily mean surface air temperatures across Canada

These temperature changes have impacts on many sectors. Temperature-based indices inform assessments of the impacts of the changing climate on agriculture, building design, and human health, among many other important areas of concern. For example, for Canada as a whole, the length of the frost-free season, which is defined as the length of the period each year when the daily minimum temperature is continuously above 0°C, is estimated to have increased by 18.1 days (9.5–27.2 days; 95% confidence interval) between 1948 and 2023. The length of the growing season, which is often defined as beginning when there are six consecutive days with daily mean temperatures above 5°C in spring or summer and ends when this condition fails to be met later in the year, is estimated to have increased by 15.9 days (10.0–22.7 days) over the same period. The trends in frost-free and growing season length were calculated the same way as in CCCR2019 (Vincent et al., 2018; Zhang et al., 2019), using the CanGridTmin and CanGridT dlyV3.1 datasets (updated from Vincent et al., 2020; Wang, Feng, Zwiers, et al., 2026).

Three other indices are often used to assess the impacts of warming: growing degree days (GDD), heating degree days (HDD), and cooling degree days (CDD). Degree day metrics are cumulative measures of the extent to which daily mean surface air temperatures depart from a certain threshold each year. For growing degree days (GDD), the temperature of 5°C is used as the threshold representing conditions suitable for growing crops and one-day exceedances over that threshold are accumulated. For example, a day with an average surface air temperature of 9°C (i.e., which is 4°C warmer than the 5°C threshold) would add 4 GDD to the total GDD for that year. We used the gridded daily temperature datasets CanGridTmin, CanGridTmax, and CanGridT dlyV3.1 to assess changes in this metric. Between 1948 and 2023, GDD increased by 195 degree days (149–246 degree days) averaged over the country, which represents 22.3% of the 1961−1990 baseline climatological average for GDD (874 degree days) for Canada. Statistically significant increases in GDD have been observed in all regions of Canada (Figure 2.9a).

Figure take-away: The warming that has occurred affects heating and air conditioning requirements and growing conditions.

Figure title: Changes in annual growing, cooling, and heating degree days over the 1948–2023 period

Cooling degree days (CDD) consider temperatures that exceed a threshold of 18°C, while heating degree days (HDD) consider those below 18°C. While the actual outside temperature at which Canadians begin to cool or heat their living and workspaces may vary from 18°C, this generic threshold is often used to calculate HDD and CDD when designing heating and cooling systems. These metrics are useful indicators of the potential energy demand for heating and air conditioning and thus are important in building design. CDD is estimated to have increased by 20 degree days (10 to 30 degree days) on average across Canada (Figure 2.9b), which is a large amount (67.8%) compared to the climatological average of 29.5 degree days. In contrast, HDD is estimated to have decreased by 711 degree days (-1086 to -357 degree days) between 1948 and 2023, which is 8.8% of the national average climatological average of 8102 degree days. Statistically significant decreases in HDD have been observed almost everywhere in Canada (Figure 2.9c). For CDD, regional averages also show strong and statistically significant increases, although changes are not significant in some parts of the Prairies and in the Hudson Bay region. Additionally, some areas in northern Canada have not historically exceeded the 18°C temperature threshold frequently enough for the CDD trend to be computed (grey area in Figure 2.9b). These changes indicate that there has been a reduction over time in the amount of energy required to heat a typical Canadian home in winter (assuming a fixed home design) and possibly a modest overall increase in the amount of energy needed to cool such a home in summer, especially in southern Canada.

Past changes in the observed number of growing, cooling, and heating degree days that accumulate annually are strongly consistent with observed changes in summer and winter average temperatures and thus there is very high confidence in the direction of these changes. However, these indices rely primarily on information about day-to-day fluctuations in temperatures in the part of the temperature distribution that extends either above (GDD and CDD) or below (HDD) the thresholds that are used in their definitions. Confidence in the magnitudes of the trends in these indices (medium) is therefore lower than in trends and annual and seasonal temperature averages.

2.4.2: Understanding the past changes

How much warming can be attributed to human influence has been an active area of research that has evolved considerably since CCCR2019. Here, we consider the improved methods used to attribute warming to human influence and the results of key studies since CCCR2019. We summarize our assessment based on this evidence at the end of the subsection.

Several external factors have influenced the state of the climate since the 1850–1900 pre-industrial period (approximated in this report as the period from 1850 to 1900). They include natural factors, such as changes in solar output and volcanic activity, and human-induced factors, such as changes in GHG concentrations, aerosol concentrations in the atmosphere, and land surface properties due to land use changes. The Working Group I contribution to the IPCC Sixth Assessment Report (AR6) (IPCC AR6 WG1) assessed human influence as likely being responsible for 0.8–1.3°C of global warming (best estimate 1.07°C) between the 1850–1900 and 2010–2019 periods (Eyring et al., 2021), compared to the observed warming of 1.09°C (0.95–1.20°C) between the 1850–1900 and 2011–2020 periods (Gulev et al., 2021) (note that uncertainty ranges in this section are reported as 90% uncertainty ranges rather than the 95% confidence intervals generally reported elsewhere in this chapter^{Footnote 11} ). Over the same period, forcing from GHGs would likely have increased the global average temperature by 1.0–2.0°C if it had acted on its own, while other anthropogenic forcings, including aerosols, likely reduced this potential warming amount by 0.0–0.8°C (Eyring et al., 2021). An updated version of this analysis using similar methods concluded that human-caused global warming reached 1.22°C (1.0–1.5°C) averaged over the 2015–2024 decade and by 2024, reached 1.36°C (1.1– 1.7°C) relative to that averaged over the 1850–1900 period (Forster et al., 2025). T. Li, Zwiers and Zhang (2025a) employed a method distinct from that in Forster et al. (2025) to produce a very similar estimate of the amount of global warming caused by human influence in 2024 of 1.30°C (1.19–1.40°C) relative to the 1850–1900 period.

For Canada, Zhang et al. (2019) assessed in CCCR2019 that it is likely that more than half of the observed warming in Canada over the 1948–2016 period was due to the influence of human activities. This was based mainly on research (Wan et al., 2019) that found that the best estimate of the trend attributable to human influences on the climate was close to the best estimate of the observed trend in Canadian average temperatures between 1948 and 2012, after the removal of natural influences resulting from two large-scale patterns of natural climate variability that affect Canada, the North Atlantic Oscillation (NAO) and the Pacific Decadal Oscillation (PDO) (see Chapter 4, sections 4.3 and 4.8). These natural influences were estimated to have contributed about 0.5°C of the warming in annual average temperatures in Canada as a whole over this period (Wan et al., 2019; Zhang et al., 2019). On the basis of these findings, anthropogenic influence (including the warming effects of GHGs and the cooling effects of aerosols) was estimated in CCCR2019 to have caused net warming of 1.1°C (0.6–1.5°C), while natural external forcings due to changes in volcanic and solar activity were estimated to have caused warming of 0.2°C (0.1–0.3°C). An informal analysis performed using a temperature dataset that extends to 2022 suggests that natural climate variability contributed substantially less over the longer 1948–2022 period than what CCCR2019 concluded. This is consistent with expectations that trends due to natural climate variability cannot persist indefinitely and thus must diminish in magnitude.

Overall, the models participating in Phase 6 of the Coupled Model Intercomparison Project (CMIP6) (Eyring et al., 2016) closely reproduce the observed warming in Canada as a whole (Figure 2.10a, crimson curve), with a period of relatively stable temperatures until the early 1970s followed by strong warming. The IPCC AR6 report found with medium confidence that there is a detectable human influence on the upward annual temperature trends in Northwestern and Northeastern North America (IPCC, 2021a).

We conducted an attribution analysis of Canadian average temperatures following the approach described in Ribes et al. (2013) using CMIP6 simulations and HadCRUT5 observations (Figure 2.10b), and found that anthropogenic influence caused 2.2°C (1.8 to 2.7°C) of warming averaged over 2015–2024 relative to the 1850–1900 baseline period. This includes temperature changes of 2.4°C (1.8 to 3.0°C) due to well-mixed^{Footnote 12} GHGs, -0.1°C (-0.5 to 0.3°C) due to other human influences, and 0.1°C (-0.3 to 0.4°C) due to natural forcings. Observational coverage of Canada is extremely limited during the 1850–1900 period (section 2.3), introducing uncertainty in the estimates of attributable warming relative to this baseline period (see Box 2.4 for estimates of externally forced Canadian warming over a range of base periods).

Using a different analytical approach to analyze changes in Canadian temperatures, T. Li, Zwiers, Zhang, et al. (2025b) applied a method called Kriging for Climate Change (KCC) (T. Li, Zwiers, & Zhang, 2025a; Ribes et al., 2021) to estimate attributable changes in Canadian average temperatures (Figure 2.10b). T. Li, Zwiers, Zhang, et al. (2025b) used CanGridT mlyV3.1 gridded temperature observations and 25 CMIP6 climate simulations of the 1850−2014 period (Eyring et al., 2016) that were extended to 2025 with simulations using the intermediate emissions scenario (shared socio-economic pathway 2-4.5, or SSP2-4.5) (for a description of the SSPs, see Chapter 3, section 3.3.1 and Meinshausen et al., 2020). The authors estimated average externally forced warming, which is dominated by human influence, of 2.2°C (1.3 to 3.1°C) in Canada in 2015–2024 relative to 1850–1900. Using a subset of 10 simulations from models that participated in the Detection and Attribution Model Intercomparison Project (DAMIP) (Gillett et al., 2016), they also estimated temperature changes of 2.8°C (2.0 to 3.7°C) due to well-mixed GHGs,  -0.5°C (-1.3 to 0.2°C) due to other human influences, and -0.1°C (-0.3 to 0.2°C) due to natural external forcings. T. Li, Zwiers, Zhang, et al. (2025b) also estimated that the anthropogenic warming influence on the Canadian average temperature was 2.3°C (1.8 to 2.8°C) and 2.4°C (1.9 to 2.9°C) in 2023 and 2025, respectively, relative to the 1850–1900 average. These estimates of anthropogenic warming in 2023 and 2025 are similar to what would be inferred from anthropogenic warming in 2015–2024 and the predicted warming rate in Canadian temperatures of 0.43°C (0.1 to 0.8°C) per decade over the 2021–2040 period (Chapter 3, section 3.4.2).

Figure take-away: Canada has warmed in a way that is consistent with the expected effects of anthropogenic forcings on the climate system.

Figure title: Externally forced warming over Canada relative to the 1850–1900 period

Figure 2.10: Two different illustrations of the estimated contributions to observed warming from several different forcing factors, including well-mixed greenhouse gases (GHG, yellow); other anthropogenic forcing agents (predominantly aerosols, OA, green); greenhouse gases and other anthropogenic forcing agents combined (i.e., anthropogenic forcing, ANT; orange); natural external forcing agents (volcanic and solar forcing, NAT, blue); and greenhouse gases, other anthropogenic forcing agents, and natural external forcing agents combined (ALL, crimson; in a) only). a) Observationally constrained estimates of warming (in ^oC) in Canada’s annual average temperature attributable to different combinations of external forcings. The lines indicate the individual contributions from the different forcing factors, while the shading indicates the 90% uncertainty bounds for these estimates. The temperature change attributed to ALL forcing was derived from 25 historical simulations conducted as part of CMIP6; the estimates attributed to the other forcing agents were derived from 10 smoothed climate change simulations using only that forcing, with one run from each DAMIP model (see text). The black dots show the observed annual average temperatures between 1948 and 2023 used to constrain the models’ climate simulations. Observed and constrained temperatures in this figure are given as differences from the 1850–1900 average. The observed annual average temperatures in this figure are offset for display purposes so that the offset annual temperatures between 1948 and 2023 have the same average as the constrained annual temperature estimates with ALL forcings for that period. b) Attributable warming over Canada due to ANT, GHG, OA, and NAT. Three attributed warming estimates are shown for each forcing: L: corresponding to T. Li, Zwiers, Zhang, et al. (2025b); R: using the approach of Ribes et al. (2013); and C: combined assessment using both approaches. The bars show the associated 90% uncertainty, while the bar for C was derived by averaging the ranges from the first two studies. The coloured bars show the average of the best estimates from the two approaches. Sources: a) from T. Li, Zwiers, Zhang, et al. (2025b); b) derived from HadCRUT5 (Morice et al., 2021) and CMIP6 simulations (Earth System Grid Federation) using the method described in Ribes et al. (2013).

By combining attribution evidence at a global scale with that for Canada, we conclude that human influence has warmed Canadian average temperatures. It is likely that the increase in the average temperature in Canada as a whole caused by human influence was between 1.8°C and 2.6°C when the period of 1850–1900 is compared to the period of 2015–2024, with a best estimate of 2.2°C warming (Figure 2.9b). It is more likely than not that human-caused warming has raised the 2025 Canadian average temperature above 1850–1900 levels by more than 2°C.

The methods outlined in T. Li, Zwiers, Zhang, et al. (2025b) and Ribes et al. (2013) were also used to obtain estimates of the attributable warming that occurred between the decades 1948–1957 and 2014–2023. On the basis of CanGrid mlyV3.1 data, Canada’s annual average surface air temperature was 1.60°C warmer during the 2014–2023 decade than during the 1948–1957 decade. The corresponding estimate of change that is attributable to human influence, based on the methods in T. Li, Zwiers, Zhang, et al. (2025b), is 2.09°C (1.70 to 2.49°C), which is partially offset by minor cooling of -0.13°C (-0.44 to 0.18°C) attributable to natural external influences. An analysis using HadCRUT5 observations (Morice et al., 2021) and the methods of Ribes et al. (2013) provides very similar estimates of warming attributable to human influences of 2.15°C (1.67 to 2.63°C), offset by minor cooling of -0.18°C (-0.61 to 0.24°C) attributable to natural external influences. In addition, the T. Li, Zwiers, Zhang, et al. (2025b) analysis attributes warming of 1.89°C (1.47 to 2.31°C) to combined human and natural influences, using a larger sample of climate models than can be used to attribute human and natural influences separately.

The consistency of the results obtained using different methods, observational datasets, and combinations of climate models indicates that there is very high confidence in the attribution of Canada’s warming between the decades of 1948–1957 and 2014–2023 to external influences. Combining results from the two analyses of the separate roles of human and natural external influences indicates that human influence on its own would have been responsible for warming of 2.12°C (1.69 to 2.56°C), and that this was partially offset by a slight, but uncertain, cooling of 0.16°C (-0.53 to 0.21°C) due to natural external forcing. We assess that human influences acting on their own would likely have caused more warming than observed and that natural external influences and natural internal climate variability combined prevented a small portion of that influence from being realized (high confidence). Overall, the available attribution evidence for the more recent 1948–2023 period, when the coverage of Canada is more complete, indicates that the warming attributable to external forcings—which in turn are strongly dominated by the effects of anthropogenic forcings—is indistinguishable from the observed warming during this period with very high confidence.

2.4.3: Confidence terms in key messages: Summary of evidence

Key Message 2.6: Canada’s climate has warmed rapidly over the past 75 years. Warming rates have fluctuated over the past century. Annual mean temperatures are estimated to have increased at a rate of 0.26°C (very likely 0.12–0.41°C) per decade between 1948 and 2023 for Canada as a whole and 0.35°C (very likely 0.18–0.55°C) per decade for Canada’s North. Canada’s estimated annual average temperature has exceeded the 1961–1990 average in 30 of the 33 years after 1990.

Key Message 2.7: Canada has warmed more quickly than most of the rest of the world. Canada’s warming rate over the 1948–2023 period is similar to that for the global land area (high confidence) but Canada’s North has warmed at a rate that is about twice the rate for the global land and ocean area combined over the 1948-2023 period (high confidence). Most of the observed warming has occurred since 1970. Canada’s warming rate was almost twice the global warming rate over the 1970–2023 period (high confidence).

Key Message 2.8: Human influence has warmed Canadian average temperatures during the 2015–2024 period to a level that is 2.2°C (likely 1.8–2.6°C) above the pre-industrial era (approximated in this report as the period from 1850 to 1900). Human influences acting on their own would likely have caused more warming than the observed warming of 1.6°C between the decades of 1948−1957 and 2014−2023, with natural external influences and internal climate variability combined preventing a small portion of that influence from being realized (high confidence).

The evidence of widespread warming of the Canadian climate that is documented in section 2.4.1 is undeniable. The observations that underpin this finding have been very carefully processed to remove non-climatic effects due to changes in instruments and other non-climatic influences, and analyses have been performed to confirm that changes over time in the number of stations where long-term temperature observations are available and their locations do not materially affect trend estimates. Warming trends seen at most locations in Canada are much more pronounced than would be possible simply by random chance in an unchanging climate. The statements in Key Message 2.6 that Canada’s climate has warmed rapidly, that warming rates have fluctuated, and that Canada’s estimated annual average temperature has consistently exceeded the 1961−1990 average after 1990, are therefore presented as statements of fact. The uncertainty ranges that accompany statements concerning warming rates for Canada as a whole and Canada’s North were obtained by calculating 95% confidence intervals using a well-established statistical method that was applied to a carefully assessed dataset. The very likely assessment (> 90% probability) applied to these ranges reflects confidence in the data and methods, while recognizing that uncertainties remain in the data and that confidence intervals cannot be given a direct probabilistic interpretation.

The evidence that northern high-latitude areas are warming faster than the rest of the world is irrefutable, and clearly evident in all global temperature datasets assessed by the IPCC in Gulev et al. (2021). The statement in Key Message 2.7 that Canada, which is strongly affected by amplified warming in the Arctic, has warmed more quickly than most of the rest of the world, is presented as a fact. There is very high confidence that Canada is warming in a way that is consistent with the expected geographic patterns of warming from increases in atmospheric GHG concentrations. Warming is amplified at high northern latitudes, as observed in both annual average temperatures and winter average temperatures, with very strong amplification in the latter, consistent with our understanding of Arctic amplification processes. Consequently, as stated in Key Message 2.7, we have very high confidence that Canada is warming faster than our neighbours to the south in the contiguous United States and high confidence that it has been warming almost twice as fast as the globe as a whole since the 1970s.

There is also very high confidence in the attribution of Canada’s observed long-term warming to rising GHG concentrations due to human influence. Rigorous detection and attribution assessments using different observational datasets, the current generation of climate models, and distinctly different methods produce highly consistent estimates of the contribution of human influence on the climate system to the observed warming in Canadian temperatures. The body of evidence from observations and detection and attribution analyses, when considered together with the extensive body of research on changes in global average temperatures, leads unavoidably to the conclusion, with very high confidence, that human influences on the climate are the cause of the warming that Canada has experienced. The statements in Key Message 2.8 indicate the substantial progress that has been made in detection and attribution research, which now allows clear statements to be made about the impact of human influences on the Canadian climate since the 1850−1900 pre-industrial period. There is substantial evidence that human influence acting on its own over the 1948−2023 period would have warmed the climate more than observed, and that a small amount of this potential warming has been offset by natural external influences and natural unforced climate variability.

2.5: Past precipitation changes 

Key Message 2.9: Annual total precipitation has increased in Canada since 1949, with larger percentage increases in northern Canada (very high confidence). Annual total precipitation has increased much faster in Canada than for the global land area as a whole over a similar period (very high confidence).

Key Message 2.10: Annual precipitation in Canada increased by 9.7% (likely 7.0–12.3%) in Canada as a whole, by 18.9% (more likely than not 8.6–29.5%, medium confidence) in Canada’s North, and by 7.5% (likely 2.4–11.8%) in Canada’s South between 1949 and 2023, with most of the increase attributable to human influence on the climate (medium confidence).

Key Message 2.11: Precipitation has not changed uniformly in all seasons. Precipitation has increased in a zonal band centred at 62°N latitude in summer, and in most areas in British Columbia and along the St. Lawrence River in spring and fall (medium confidence). Precipitation has increased in all seasons in an area from southern Nunavut to the Arctic Archipelago and from Labrador to northeastern Quebec (low confidence).

Key Message 2.12: The annual snowfall amount and annual number of days with snowfall are estimated to have decreased at most stations in Canada’s South but increased at most stations in Canada’s North (medium confidence). Both the proportion of precipitation days with snow and the proportion of precipitation amount falling as snow have decreased across most of Canada (medium confidence).

This section describes observed changes in annual and seasonal total precipitation based primarily on the Canadian Gridded Homogenized Precipitation monthly dataset version 2, CanGridP mlyV2 (see section 2.3.4.2) (Wang, Feng, Zwiers, et al., 2026). It also uses monthly precipitation data from the ERA5 reanalysis dataset (section 2.3.4.4) to show changes in regional averages. In addition, this section provides estimates of the extent to which human influence on the climate may be responsible for the observed changes. This information complements the assessments of changes to other precipitation-related aspects discussed in Chapter 4 (intensity, duration, and frequency of atmospheric rivers), Chapter 5 (intensification of the water cycle and most aspects of floods and drought), and Chapter 8 (extreme precipitation).

The Canadian Homogenized Precipitation daily dataset version 2, CanHomP dlyV2 (see section 2.3.4.2) (Wang & Feng, 2026; Wang, Feng, Zwiers, et al., 2026) was also used to estimate daily snowfall amounts for use in assessing snowfall trends (Qian et al., 2025). This information, based on long-running homogenized station data on precipitation and temperature, is complementary to the information presented in Chapter 6, which considers changes in snowfall, snow cover extent, and snow mass using in situ, remotely sensed, and analyzed snow data from a variety of sources. Changes in the frequency of freezing precipitation are considered in this chapter using homogenized data derived from hourly “current weather” reports, while changes to the intensity and frequency of extreme freezing precipitation events are assessed in Chapter 8.

“Normal” (three-decade average; see footnote 4) precipitation amounts vary substantially across Canada (see Supplementary Figure S2.3), and thus the importance of an absolute change (e.g., 5 mm) is different in different regions (5 mm would be a large change in Canada’s North but a small change on the east and west coasts). Therefore, throughout this section, rates of change in precipitation and snowfall amounts are expressed as the percent change over a period (or percent per decade) relative to the baseline value calculated for the 1961–1990 climate normal period.

2.5.1: Past changes in annual and seasonal precipitation

This section consists of three subsections, presenting observed changes in total precipitation amounts, snowfall amounts and number of snow days, and freezing precipitation frequency.

Figure take-away: Annual precipitation amounts have changed differently across Canada over the period when sufficient precipitation data are available.

Figure title: Changes in annual precipitation amounts in Canada

2.5.1.1: Past changes in precipitation amounts

Trend estimates based on CanGridP mlyV2 data indicate a statistically significant increase in annual total precipitation amounts at the 5% level (i.e., there is ≤ a 5% chance of concluding that such an effect or trend exists when it does not) in most areas from southern Nunavut to the Arctic Archipelago, from Yukon to northern British Columbia and the southern Northwest Territories, from Labrador to northeastern Quebec, and along the St. Lawrence River. Significant decreases occurred in small areas in the northern Northwest Territories, southern Alberta, and in eastern Canada (Figure 2.11a). The annual precipitation amount for Canada’s land area as a whole has increased since 1949 (Figure 2.12a), with the 10 highest annual precipitation amounts during the 1949−2023 period occurring since 1996. Canada’s annual precipitation amount is estimated to have increased by about 9.7%, or at a rate of 1.31% per decade, between 1949 and 2023 (see Table 2.2 for the 95% uncertainty ranges of changes) (Wang, Feng, Zwiers, et al., 2026). Between 1916 and 2023, annual total precipitation averaged over Canada’s South increased at a rate of about 1.0% per decade (Wang, Feng, Zwiers, et al., 2026). These increases are statistically significant at the 5% level.

Figure take-away: Canada’s regional average precipitation amount has changed over the period when sufficient precipitation data are available.

Figure title: Changes in the average annual precipitation amount for Canada

Table 2.2: Observed changes in annual and seasonal precipitation amounts during the periods indicated in Canada as a whole, six regions of Canada, and Canada’s South. Changes are represented by linear trends over the period, with the corresponding 95% confidence intervals shown in parentheses and expressed as a percentage of the baseline (1961–1990) average per decade. Linear trends and their uncertainties were estimated from the CanGridP mlyV2 data. See Figure 2.2 for a definition of the regions and footnote 7 for the trend analysis method. Source: Wang, Feng, Zwiers, et al. (2026).

The rate of increase in annual total precipitation varies from location to location (Figures 2.11 and 2.13) and between regions (Table 2.2). The highest rate of increase (expressed as a percentage) was observed in Canada’s North, where the regional average increased by 18.9% (or 2.55% per decade) between 1949 and 2023. In contrast, it increased by only 7.5% (or 1.01% per decade) in Canada’s South (Table 2.2). This latitudinal difference in precipitation trends, with larger relative changes in Arctic regions, is an expected feature of the climate system’s response to externally forced warming resulting from the increase in the poleward transport of atmospheric moisture (Chapter 4, section 4.5.2) (e.g., Held & Soden, 2006; Figure SPM.5 in IPCC, 2021). However, climatological annual total precipitation amounts are much higher in the coastal regions of southern Canada than in the Arctic (Supplementary Figure S2.3). The absolute amount of precipitation increase is highest in Atlantic Canada and coastal British Columbia (Figure 8a in Wang, Feng, Zwiers, et al., 2026), even though the percentage increases are smaller in those regions.

Figure take-away: Changes in precipitation amounts experienced in Canada since 1949 vary by season.

Figure title: Changes in seasonal total precipitation amounts in Canada in 1949–2023

Precipitation trends observed in Canada show notable seasonal variations (Figure 2.13). Averaged for Canada as a whole, the annual precipitation amount increased from 1949 to 2023 by 17.6% (2.38% per decade) in spring, 8.5% (1.15% per decade) in summer, and 11.0% (1.48% per decade) in fall, with a small increase of 2.5% (0.34% per decade) in winter that is not statistically significant (Table 2.2). Seasonal changes in precipitation also show different geographical patterns. In southern Canada, decreases dominate in winter and increases dominate in the other seasons (Figure 2.13). In central-south Canada, winter precipitation has decreased significantly, while it has increased significantly in most areas in northern Canada, especially from southern Nunavut to the Arctic Archipelago and from northern Yukon to the southern Northwest Territories, and in parts of northern Quebec. Northern Canada has a similar pattern of precipitation changes in winter and spring. In southern Canada, the pattern is substantially different, with negative trends dominant in winter and positive trends dominant in spring. In summer, regional precipitation changes include increases in the eastern Arctic and in a roughly 500 km wide zonal band across the country centred at roughly 62°N latitude, accompanied by small areas of decreases along the north coast of the Northwest Territories, in southern Vancouver Island, and in a swath running from northern Ontario to central Quebec. Spring precipitation also increased along the St. Lawrence River. The pattern of precipitation trends in fall is similar to that in spring, but with smaller relative changes that are also significant in smaller areas, except in southeastern Canada (Figure 2.13).

While precipitation amounts derived from reanalysis products (section 2.3.4.4) are less reliable than those from station data, it is nevertheless useful to compare the changes seen in station data with those calculated from reanalyses. Figure 2.12b,d illustrates the changes in annual precipitation amounts for Canada and Canada’s South, respectively, calculated from the ERA5 reanalysis dataset. Both the gridded station data and ERA5 data show that annual precipitation amounts in Canada have increased considerably, with substantially higher amounts occurring after about 1970. The overall long-term changes in Canada as a whole and in Canada’s South derived from the ERA5 dataset are similar in magnitude to those estimated from the gridded station data, particularly since 1970, lending support to the magnitude of the changes calculated from the gridded station data. Different year-to-year variations in the two products provide a stark indication of the challenges involved in estimating annual average precipitation amounts in Canada, especially during the early period. This provides a reminder that all estimates are subject to some degree of uncertainty, due to the highly variable nature of precipitation both spatially and temporally.

As with temperature, precipitation has increased at a higher rate in Canada than in the contiguous United States (CONUS) and over the global land area. In contrast with the 1.31% per decade increase in annual precipitation amounts in Canada, annual precipitation amounts in CONUS increased by only 0.97% per decade from 1949 to 2023, according to the NOAA NCEI’s National Time Series (NOAA NCEI, 2024). Nevertheless, substantial changes have been observed in some regions of CONUS, including a 5 to 15% increase in average annual precipitation in the central and eastern United States, and a 10 to 15% decrease in parts of the southwestern United States, when comparing the normal values for the 1901−1960 and 2002–2021 periods (U.S. Global Change Research Program, 2023). Using three global land precipitation station datasets, the IPCC AR6 WGI report (Gulev et al., 2021) estimated that the global land average annual precipitation amount increased at a rate of 1.67 mm (-1.56 to 4.90 mm, 90% confidence interval), 0.17 mm (-2.95 to 3.29 mm), and 5.03 mm (0.16 to 9.90 mm) per decade, respectively, for the period from 1960 to 2019 (Gulev et al., 2021). The differences in these rates of change correspond to increases of 0.21%, 0.02%, and 0.64% per decade, respectively, relative to the global land average annual precipitation amount of 790 mm in the 1950–2000 period derived from the Global Precipitation Climatology Centre (GPCC) dataset (Schneider et al., 2017). While the rates of increase in precipitation remain uncertain, on average, precipitation has increased faster in Canada than in CONUS and in the global land area (very high confidence).

Another useful way to describe precipitation changes is to express them as a scaling rate, or the percentage of increase per 1°C of warming in the region in question, which provides a measure of the sensitivity of the average amount of precipitation to a change in average temperature, as well as a benchmark for comparing climate models. Canada’s precipitation increased at a rate of 2.6% per 1°C of warming in the average temperature for Canada as a whole over the 1970–2023 period and at a rate of 4.9% per 1°C of warming over the 1949–2023 period (section 2.4) (Wang, Feng, Zwiers, et al., 2026). The scaling rate for Canada’s North (5.7% per 1°C of warming for 1970–2023, and 6.9% for 1949–2023) is higher than that for Canada as a whole, while the rate for Canada’s South is comparable to that for CONUS (5.1% versus 5.2%, respectively, per 1°C of warming for 1949–2023). The higher scaling rates for Canada’s North are generally consistent with our understanding of how the climate system will alter and redistribute precipitation due to human influence on the climate (Lee et al., 2021). For the century-long period of 1916–2023, the scaling rate is 6.7% per 1°C of warming for Canada’s South (the data are insufficient to calculate the rate for Canada’s North). 

2.5.1.2: Past changes in snowfall

We next briefly consider changes in the frequency and amount of frozen precipitation and its contribution to total annual precipitation. As a proxy for snowfall, we use the precipitation that occurs when the daily mean surface air temperature falls below a station-specific threshold (Box 2.3) (Qian et al., 2025).

The automated gauges used to date in Canada only measure total precipitation amounts and do not report on precipitation type (e.g., rain or snow). Therefore, snowfall data for the last decade or so are limited. Qian et al. (2025) addressed this limitation by estimating snowfall amounts from homogenized daily precipitation data using homogenized daily mean temperatures, as described in Box 2.3. The resulting proxy annual snowfall dataset, called CanHomSp anlV2, was employed to assess historical changes in snowfall.

The proxy snowfall data show that both the annual snowfall amount and the annual number of snow days (a snow day is a day with a proxy snowfall water equivalent greater than 1.0 mm) have decreased at most stations in southern Canada, but have mostly increased in northern Canada (Figure 2.14a,b). The region spanning southern Ontario to the Maritime provinces had a mixture of statistically significant increases and decreases in the number of snow days. Both the proportion of precipitation days with snow (ratio of snow days to precipitation days) and the proportion of precipitation amount falling as snow (ratio of snowfall amount to precipitation amount) have decreased at most locations in southern Canada (Figure 2.14c,d). The median rate of decreases across Canada is 2.1% per decade for the proportion of precipitation days with snow, and 3.4% per decade for the proportion of precipitation amount falling as snow. Changes in the Arctic are generally not statistically significant, because precipitation there continues to fall predominantly in the form of snow. The changes in the proportion of precipitation amount falling as snow and that of precipitation days with snow are somewhat seasonal, with greater decreases in these proportions in fall and spring than in winter (Qian et al., 2025). Snowfall trends similar in direction to those seen in southern Canada have also been reported in CONUS. The proportion of precipitation amount falling as snow decreased across most of the CONUS region during the 1949–2020 period, as did snowfall amounts at the majority of CONUS monitoring stations, except for stations in some parts of the central and western United States, where increases were reported (USEPA, 2025).

Figure take-away: Annual snowfall amount, number of snow days, proportion of precipitation amount falling as snow, and proportion of precipitation days with snow have changed in Canada.

Figure title: Trends in annual snowfall amount, number of snow days, proportion of precipitation amount falling as snow, and proportion of precipitation days with snow during the 1949–2023 period in Canada

2.5.1.3: Past changes in freezing precipitation frequency

Freezing precipitation occurs under specific meteorological conditions and creates hazards for many natural and human systems, including forestry, agriculture, transportation, and electrical transmission and distribution systems. Potential impacts include the loss of electrical power for extended periods of time, due to the collapse of transmission towers under the loads created by large quantities of accreted ice on the towers and lines. This was experienced in the January 1998 ice storm that affected eastern Canada and parts of the northeastern United States. This subsection reports on the changes observed in the frequency of freezing precipitation. Changes in extreme freezing precipitation events (e.g., extreme accumulations of ice) are assessed in Chapter 8, section 8.3.4.

Figure take-away: The annual frequency and seasonal frequency of freezing precipitation in Canada have changed since 1954.

Figure title: Trends in the frequency of freezing precipitation over the 1954–2023 period

Observations of the occurrence of freezing precipitation (including freezing rain and freezing drizzle) are made by human observers without using instruments, and are reported hourly (usually at airports, as one of the current weather indicators). Therefore, only a limited number of stations in Canada have a long-term record of freezing precipitation frequency. For this section, we updated the homogenized long-term frequency data series for 95 stations compiled by Wang (2006) (Figure 2.4) by joining some nearby station records.

The annual frequency of freezing precipitation has decreased at most stations in southern Ontario and in the Prairies, but has increased at most stations in the area from northern Ontario to southern Labrador, with statistically significant increases seen at three stations in northern Canada (Figure 2.15a). A negative trend is estimated for 50 of the 95 stations, a positive trend for 33 stations, and near zero trends (no change) for 12 stations (including eight stations in British Columbia, and one station each in Yukon, the Northwest Territories, Alberta, and Prince Edward Island). The pattern of change in winter is similar to that for the year as a whole (annual trend), but a statistically significant change was detected at fewer stations due to the absence of freezing precipitation in the north and the greater year-to-year fluctuations in seasonal frequency than in annual frequency (Figure 2.15a,c). For the same reasons, a statistically significant change was also detected at fewer stations in fall and spring (Figure 2.15b,d). In fall, a negative trend was estimated for 32 stations in central Canada, with no change at most stations in western Canada and Atlantic Canada. In spring, an increase in freezing precipitation was detected from northeastern Ontario to the Maritime provinces and Nunavut, with decreases in southern Ontario, Manitoba, and southern Saskatchewan, and no change in western Canada.

While it is difficult to determine an overall pattern, most of the available observing stations show decreases in the frequency of freezing precipitation, consistent with the observed warming, with some evidence of a possible regional shift in northern Ontario and Quebec towards an increased frequency of freezing precipitation.

The complicated nature of these changes in the frequency of freezing precipitation should be expected, given that the occurrence of freezing precipitation depends, in a complex way, on atmospheric conditions in the lower troposphere (Chapter 3, sections 3.2 and 3.5). Projected changes in the duration of freezing rain events are assessed in Chapter 3 (Figure 3.15).

2.5.2: Understanding the observed changes

The observed precipitation changes in Canada are consistent with the direction of observed global-scale changes in precipitation over land at the mid (30–60°N) and high (60–90°N) latitudes, as identified in the IPCC AR6 WG1 report (Eyring et al., 2021). These changes are also consistent with the changes in precipitation that are expected to occur in a warming climate. With increasing temperatures, the IPCC AR6 WG1 report assessed that atmospheric moisture has very likely increased globally during the era of satellite data (Gulev et al., 2021). It also assessed that it is likely that human influence has contributed to increased moisture in the upper troposphere, and with medium confidence, that it has contributed to an increase in surface specific humidity around the world (Eyring et al., 2021). An increase in total precipitation is expected with the increase in atmospheric moisture, but the change in global annual precipitation is less than the global increase in atmospheric moisture due to energy balance constraints (Allen & Ingram, 2002; Boer, 1993; Douville et al., 2021).

CCCR2019 concluded that “there is medium confidence that the observed increase in Canadian precipitation is at least partly due to human influence” (Zhang et al., 2019). This is supported by the IPCC AR6 WG1 report (Eyring et al., 2021), which stated that “new attribution studies strengthen previous findings of a detectable increase in mid to high latitude land precipitation over the Northern Hemisphere (high confidence).” One of those attribution studies (Wan et al., 2015) was included in the CCCR2019 assessment of historical changes in precipitation in Canada.

To evaluate the role of human and natural external influences (external drivers) in changes in precipitation in Canada, observed changes in annual precipitation were compared to CMIP6 model simulations using different sets of external drivers (Figure 2.16). The observations used for this analysis were obtained from CanGridP mlyV1 (Wang et al., 2023), and corrected for inhomogeneities resulting from the erroneous reporting of missing values as zero precipitation amounts (Wang, Feng, Zwiers, et al., 2026). The periods considered for this analysis were 1959–2018, when a larger number of stations were contributing data to the CanGridP mlyV1 dataset, and 1904–2018 for southern Canada, where longer records were available.

In both northern and southern Canada, the observed increases in precipitation over the 1959−2018 period correspond reasonably well to the changes revealed in climate model simulations with combined anthropogenic (including greenhouse gases and aerosols) and natural (volcanic and solar) forcings or with greenhouse gas forcing only (Figure 2.16). In contrast, model simulations with only natural forcings show no changes over this period and those with aerosol forcing only show limited changes or slight decreases. Although the range of estimates from the different model simulations is wide (see shading in Figure 2.16), the solid lines representing the average of all model simulations can be used to represent the response due to each type of climate forcing. According to a detection and attribution analysis, the observed increase in annual precipitation in Canada over the 1959–2018 period can be attributed to anthropogenic forcings, with the relationship in northern Canada substantially stronger and more significant than in southern Canada (medium confidence) (Kirchmeier-Young et al., 2025). With longer records available for southern Canada, the observed change in annual precipitation is better explained by anthropogenic forcings during the 1904–2018 period than the 1959–2018 period (medium confidence) (Kirchmeier-Young et al., 2025). Changes in warm and cool season precipitation in Canada and northern Canada can also be linked to anthropogenic forcings (low confidence) (Kirchmeier-Young et al., 2025).

Figure take-away: Observed average precipitation changes in Canada are better explained by climate models when simulations include the effects of anthropogenic forcings.

Figure title: Observed and simulated changes in five-year averages of regionally averaged annual precipitation anomalies relative to the baseline average values

The main influence of anthropogenic warming on changes in precipitation in the high and mid-latitudes occurs through thermodynamic processes. The warmer atmosphere resulting from human influence has an increased water-holding capacity, which, in turn, allows the atmospheric circulation to move more moisture into these high and mid-latitude regions. Thermodynamic processes also play a role in increasing the frequency and intensity of atmospheric rivers, again by allowing the circulation patterns that produce atmospheric rivers to move more moisture away from warmer regions where atmospheric moisture is plentiful. Atmospheric rivers contribute up to 30% of the annual precipitation in some regions (Chapter 4, section 4.5). Large-scale modes of natural climate variability, such as the Pacific Decadal Oscillation and the El Niño−Southern Oscillation (Chapter 4, section 4.8), can also affect precipitation in Canada, mainly during winter and especially in the western and southern parts of the country. These indices do not exhibit any long-term trends (Gulev et al., 2021), but they can have regional influences for extended periods of time. See Chapter 4 for a more in-depth discussion on large-scale climate variability and its influence on Canada’s climate. In conclusion, there is medium confidence that most of the observed increase in annual precipitation over the 1959−2018 period in Northern Canada and at least part of the increase observed over Canada as a whole is due to human influence on the climate.

2.5.3: Confidence terms in key messages: Summary of evidence

Key Message 2.9: Annual total precipitation has increased in Canada since 1949, with larger percentage increases in northern Canada (very high confidence). Annual total precipitation has increased much faster in Canada than for the global land area as a whole over a similar period (very high confidence).

Key Message 2.10: Annual precipitation in Canada increased by 9.7% (likely 7.0–12.3%) in Canada as a whole, by 18.9% (more likely than not 8.6–29.5%, medium confidence) in Canada’s North, and by 7.5% (likely 2.4–11.8%) in Canada’s South between 1949 and 2023, with most of the increase attributable to human influence on the climate (medium confidence).

Key Message 2.11: Precipitation has not changed uniformly in all seasons. Precipitation has increased in a zonal band centred at 62°N latitude in summer, and in most areas in British Columbia and along the St. Lawrence River in spring and fall (medium confidence). Precipitation has increased in all seasons in an area from southern Nunavut to the Arctic Archipelago and from Labrador to northeastern Quebec (low confidence).

Key Message 2.12: The annual snowfall amount and annual number of days with snowfall are estimated to have decreased at most stations in Canada’s South but increased at most stations in Canada’s North (medium confidence). Both the proportion of precipitation days with snow and the proportion of precipitation amount falling as snow have decreased across most of Canada (medium confidence).

As stated in Key Message 2.9, there is very high confidence that Canada is receiving more precipitation relative to historical levels, and that this is happening in northern Canada even more quickly than in Canada as a whole. This evidence is found consistently across a range of structurally different data sources, including extensively studied Canadian station data and a range of reanalysis data products. There is also very high confidence that Canada is experiencing greater increases in precipitation relative to historical levels than most other global land areas. When compared with the changes documented in the NOAA NCEI dataset and a variety of observational products assessed in Gulev et al. (2021), the evidence indicates that, on average, precipitation has increased more rapidly in Canada than in the CONUS region of the United States and in the global land area as a whole.

Confidence in precipitation changes estimated with reanalysis data is lower than in changes estimated from station data, due to the influence that the weather forecasting models used in reanalysis have on reanalysis precipitation amounts. Therefore, we have assessed trends in precipitation amounts using evidence from long-running station data. The improved precipitation data now available allow us to assess trends in annual average precipitation amounts in Canada as a whole and Canada’s South with high confidence, and in Canada’s North with medium confidence (i.e., Key Message 2.10), but confidence is lower (medium or low) in trends in seasonal amounts and in some regions (i.e., Key Message 2.11), in part because of the limited number and uneven spatial distribution of high-quality homogenized long-term precipitation records in Canada. The improved precipitation data allow us to cautiously interpret the 95% confidence intervals for the estimates of trends in annual precipitation amounts for Canada and Canada’s South as likely ranges (i.e., > 66% probability) and for Canada’s North, as a more likely than not range (i.e., > 50% probability) (i.e., Key Message 2.9). These assessments, which are more conservative than the corresponding confidence and likelihood assessments for temperature, reflect the greater challenges in measuring, homogenizing, and gridding precipitation observations.

The rapid changes in Canadian precipitation amounts are consistent with the expected effects of anthropogenic forcings in mid- and high-latitude areas based on climate model simulations of the historical period since 1850 and theoretical understanding. It is now possible to demonstrate the consistency between observed and expected changes by performing a formal detection and attribution analysis using an updated version of the CanGridP mlyV1 dataset (Wang et al., 2023; Kirchmeier-Young et al., 2025). This, together with detection and attribution studies on larger study areas that include Canada and that demonstrate that northern high-latitude precipitation and North American precipitation have both increased due to human influence on the climate, supports the assessment in Key Message 2.10 that there is medium confidence that most of the observed increase in Canadian precipitation is attributable to human influence.

Owing to the limited availability of direct measurements of snowfall amounts, the consideration of changes in snowfall metrics in this chapter was limited to the amount and frequency of precipitation falling on days when the daily mean temperature was below a station-specific threshold (which mostly range between -2°C and +2°C; Qian et al., 2025). Nevertheless, confidence in the underlying temperature and precipitation data allows us to have medium confidence in the trends in snowfall amounts and frequency for southern and northern Canada obtained from the derived snowfall data (i.e., Key Message 2.12). Estimates of changes in the frequency of freezing precipitation are only possible for a limited number of locations. While the majority of the available long-term stations show a decreasing overall frequency of freezing precipitation, there is insufficient coverage to discern distinct patterns of change that can be linked to human influence on the climate with any degree of confidence.

2.6: Past wind speed changes

Key Message 2.13: Annual and seasonal average surface wind speeds decreased during the 1953–2023 period across a large part of southern Canada stretching from the southern Prairies to central Quebec (medium confidence). In contrast, average wind speeds increased in British Columbia in spring and summer (low confidence). There is low confidence overall in assessments of the magnitude of annual average wind speed changes due to inconsistency between observational data products. It is currently not possible to attribute any aspect of the observed wind speed changes to human influence on the climate system.

Key Message 2.14: The sparseness of station data in northern Canada makes it very difficult to assess long-term wind speed trends in that region, but consistency between the available station data and a modern reanalysis dataset suggests that wind speeds increased during the 1953–2023 period in northern Canada (low confidence), particularly in fall and spring.

This section describes long-term changes in wind speeds in Canada over the 1953–2023 period based on homogenized station data and three recent reanalyses: ERA5, 20CRv3, and OCADA (see sections 2.3.4.3 and 2.3.4.4 for a description of the data). Wind speeds are measured at a height of 10 m at each observing station. Considering that ERA5 is the only reanalysis dataset that has data after 2015 and the data are of higher resolution than the OCADA and 20CRv3 data (see section 2.3.4.4), we also show the trend maps derived from ERA5 wind data. This section also considers whether human influence on the climate may be responsible for some of the documented changes.

2.6.1: Past changes in average winds

Surface wind speed trends detected in data collected from the 1953−2023 period at 154 observing stations across Canada (i.e., station data) and subsequently homogenized (Wang, Feng, Isaac, et al., 2025) were compared with corresponding trends estimated from the ERA5 (Figure 2.17a,b), 20CRv3 (Figure 2.17c), and OCADA (Figure 2.17d) datasets. All four datasets (station data, ERA5, 20CRv3, and OCADA) point to a stilling trend (i.e., decreasing wind speeds) across much of southern Canada, in a band extending from the southern Prairies to central Quebec, but show inconsistent wind speed trends in most other regions. In addition, the ERA5 and 20CRv3 data consistently indicate increased wind speeds along the shores of Hudson Strait to northern Hudson Bay, and in central western Nunavut (from the Burnside River basin to Kugluktuk and southwestern Victoria Island), where station data are not available to assess trends (Figure 2.17b,c). Surface wind speeds obtained from station data show a decrease in wind speeds since 1953 in a region spanning northern British Columbia to southern Yukon and the Northwest Territories, and the region extending from the southern Prairies to south-central Quebec and Labrador. Strengthening winds have been observed in most other regions, particularly in the region from south-central British Columbia to the Rocky Mountains, the Atlantic Provinces, and the high Arctic (Figure 2.17a).

The results from the ERA5 dataset are generally more consistent with the station data results, with both showing increased wind speeds in south-central British Columbia and the high Arctic, while the 20CRv3 and OCADA data indicate mostly insignificant decreases in these regions. The ERA5 data and the station data consistently show stilling along the southern coast of Hudson Bay, while 20CRv3 shows small increases that are not statistically significant. However, in the region from northern British Columbia to the central Northwest Territories, ERA5 points to a strengthening trend, while the station data, 20CRv3, and OCADA show stilling trends. Note that the ERA5 dataset incorporates an effective surface roughness parameter (which accounts for turbulence and friction caused by surface features) that can vary over time.

Averaging a climate variable such as wind speed over large areas generally helps to filter out sampling uncertainties. Therefore, although the ERA5, OCADA, and 20CRv3 trend patterns are not consistent throughout the country, variations in national and regional average wind speeds calculated from the reanalyses are broadly consistent with each other (Figure 2.18). There is marked variability in regionally averaged wind speeds from one decade to the next, with the most prominent feature being the large positive anomalies around the 1990s in Canada’s North and matching negative anomalies in Canada’s South. Averaged for Canada as a whole, wind speeds based on ERA5 data are estimated to have decreased significantly at the 5% level (i.e., there is ≤ a 5% chance of concluding that such an effect or trend exists when it does not) only in summer, with no statistically significant changes in the annual average or other seasonal averages (Table 2.3). For regionally averaged annual average wind speeds, a statistically significant decrease is seen only in Ontario, with no statistically significant changes in the other regions.

Figure take-away: Estimates of the change in annual average surface wind speeds in Canada differ depending on the data source.

Figure title: Trends in annual average surface wind speeds

Figure take-away: There is only modest agreement between the estimated trends in regional average surface wind speed changes in Canada from different sources.

Figure title: Changes in regional average annual average surface wind speeds

The homogenized station data show notable seasonal variations in the trends observed in wind speeds across Canada and its regions, although weakening winds are seen in all seasons in the southern Prairies and strengthening winds are seen in all seasons in south-central British Columbia (Figure 2.19, triangles). The area of statistically significant trends is largest in summer, with statistically significant wind speed decreases at most stations and areas in eastern and central Canada and significant increases along the west coast and in the eastern high Arctic (Figure 2.19c).

The degree of agreement between the ERA5 and the station data also varies from season to season (Figure 2.19). The strongest agreement between the ERA5 and station data was in summer, when wind speeds increased in the region from the west coast to the central Rocky Mountains and on the Baffin Bay coast, but decreased in most other regions. In spring, the two datasets disagree with each other over the direction of change (statistically significant increases seen in one, but decreases in the other) in central-western Canada (northern British Columbia to the southern Northwest Territories), while they both indicate significant increases in northeastern Quebec, northern Labrador, and Newfoundland. The disagreement in central-western Canada is also visible in fall, but to a lesser extent. The station data show statistically significant wind strengthening in southern British Columbia and the Rocky Mountains in all seasons, while the ERA5 data show mostly negative trends in those regions in winter and fall.

Table 2.3: Changes in annual and seasonal average wind speeds between 1953 and 2023 for Canada, six regions, and Canada’s South. Changes are represented by linear trends over the period, with the corresponding 95% confidence intervals shown in parentheses and expressed as a percentage of the baseline (1961–1990) average per decade. The estimates were derived from the ERA5 reanalysis data for the Canadian land area. See Figure 2.2 for a definition of the regions and footnote 7 for the trend analysis method. Source: Wang, Feng, Isaac, et al. (2025).

Both the ERA5 and station data illustrate the challenges that are inherent in estimating wind speed changes in areas with complex terrain. ERA5 uses an approximately 31 km resolution weather forecasting model to assimilate and analyze observations (see section 2.3.4.4), and thus its representation of surface elevations is very smooth compared to reality, which influences how well it can represent the surface wind field. It should also be noted that surface wind observations from land-based stations are not assimilated in ERA5, but rather, it uses a variety of other kinds of data to estimate near-surface winds. Estimating wind speed changes from anemometer data at observing stations is challenging in a different way. One factor affecting observing stations is that, because they tend to be located at lower elevations in areas with complex terrain, the winds recorded are not necessarily representative of the wind field changes that may have occurred throughout such regions.

When regional average wind speeds are considered (Table 2.3), a statistically significant decrease is estimated in average wind speeds in summer and fall across Ontario, and in average wind speeds in summer across Quebec and across Atlantic Canada, separately. When averaged across Canada’s South, the wind stilling trend is statistically significant in summer.

Figure take-away: Seasonal differences are found in changes in average surface wind speeds in Canada since the 1950s.

Figure title: Trends in seasonal average wind speeds, 1953–2023

Regardless of trend direction, trends estimated from the station data tend to be stronger than their reanalysis counterparts. This may be because wind speeds measured at stations are apt to be more variable than reanalysis grid box wind speeds. An hourly value from station data is much like a snapshot taken at an instant in time, as it represents the wind speed at the station location (a point) averaged over the two-minute interval ending at the hour of observation (ECCC, 2015). On the other hand, an hourly value in reanalysis data represents the “instantaneous” speed of the model grid box area mean vector wind at the top of each hour for grid boxes of considerable size (~960 km² for ERA5 and ~5600 km² for 20CRv3). In the context of weather and climate models, “instantaneous” refers to the smallest time increment that is represented by models. Reanalysis models use fixed time-steps, with the values of the variables that represent the state of the atmosphere changing once every time step. For example, ERA5 uses a 12-minute time step (see Detailed information of implementation of IFS cycle 41r2). Therefore, a great deal of the small-scale temporal and spatial variability in wind speeds that affects station measurements is absent from the reanalysis data.

One notable change in wind speed in Canada that has been studied in detail has occurred in some ocean areas in the Canadian Arctic. Wind speed has increased since 1979 in most areas of the Davis Strait and Baffin Bay region during the period from September to December, with some decreases over open water in June and July, according to a very high-resolution regional reanalysis dataset, ECCC’s Davis Strait Baffin Bay Wind and Wave Reanalysis, for the 1979–2016 period (Wang et al., 2021). This is basically consistent with the seasonal trends in the adjacent coastal land area shown in Figure 2.19. In this region, the positive trend in wind speed has intensified since 2001 (Wang et al., 2021). On the basis of ECCC’s Beaufort Sea Wind and Wave Reanalysis for the 1970–2013 period, wind speed increased significantly in the central part of the Beaufort and Chukchi seas but decreased significantly in corresponding Canadian coastal waters in August (Wang et al., 2015). The decrease over these coastal waters in August is consistent with decreases seen in the adjacent land area (Figure 2.19c). More information on marine wind changes is presented in Chapter 7, section 7.5.

2.6.2: Understanding the observed changes

Attributing changes in wind speed to external forcings remains difficult for several reasons. The wind data products used provide inconsistent information about trends for substantial parts of Canada, and the magnitudes of the trends tend to be small relative to year-to-year variability in the areas where the datasets do exhibit consistent trends. Both factors reduce confidence in the potential attribution of the causes of the change. The IPCC AR6 WGI report assessed that extratropical storm tracks in both the Northern and Southern hemispheres have likely shifted poleward since the 1980s, with marked seasonality in trends (medium confidence) (Gulev et al., 2021). It also presented some evidence that the annual number of extratropical cyclones has increased in the Northern Hemisphere (low confidence), but with fewer deep cyclones (strong storms with central pressure < 980 hPa) (Gulev et al., 2021). These changes could have affected Canadian surface wind regimes and precipitation. Such assessments, which have not been made specifically for the sector in the Northern Hemisphere that includes Canada, would necessarily have even lower levels of confidence.

Many other aspects of the literature on wind changes and related extratropical storm activity also remain ambiguous. For example, the trend pattern in Northern Hemisphere geostrophic wind^{Footnote 13} energy has been shown to contain a detectable response to anthropogenic and natural forcings combined (Wang et al., 2009), with wind energy increases over some parts of the North Atlantic and North Pacific and weak changes, generally decreases, elsewhere in the hemisphere. Some researchers also argue that decreases in near-surface wind speeds in the mid-latitudes of the Northern Hemisphere are attributable to greenhouse gas increases (e.g., Deng et al., 2021). Other studies, however, argue that observed large-scale changes in wind speeds are consistent with internal climate variability (e.g., Wohland et al., 2021).

Wind stilling has been broadly observed in wind observation data in many parts of the world, but its causes remain a subject of scientific debate. A variety of theories have been presented to explain stilling, some of which focus on changes in the drag force that acts on the wind, which is linked to increased terrestrial roughness caused by urbanization and vegetation changes (e.g., Vautard et al., 2010). However, some researchers note a relatively sharp reversal of stilling beginning in 2010, which would be inconsistent with the explanations relying on terrestrial roughness, since the latter can only change slowly (Zeng et al., 2019). Others have suggested that the variation in wind speed (including previous stilling and the recent reversal) is determined mainly by the decade-to-decade variability of large-scale oscillations, such as the Pacific Decadal Oscillation (e.g., as suggested by Zeng et al., 2019) (see Chapter 4 for more details on large-scale circulation processes). This view is reinforced by an analysis of wind speed variations in a very large ensemble simulation with a global climate model (Wohland et al., 2021), which identified natural internal climate variability as the dominant cause of variations in wind energy resources (Pryor et al., 2020), and by the suggestion of a link between the El Niño−Southern Oscillation and wind speeds on the Canadian Prairies (St. George & Wolfe, 2009). It should be noted, however, that an extensive review of the literature on wind speed changes and their potential causes leads to the conclusion that many open questions remain on the relative importance of circulation changes (Wu et al., 2018), which could reflect the effects of either external forcings or low-frequency internal climate variability (or both), as well as drag force changes (i.e., changes in terrestrial roughness and possibly changes due to increased wind energy extraction).

2.6.3: Confidence terms in key messages: Summary of evidence

Key Message 2.13: Annual and seasonal average surface wind speeds decreased during the 1953–2023 period across a large part of southern Canada stretching from the southern Prairies to central Quebec (medium confidence). In contrast, average wind speeds increased in British Columbia in spring and summer (low confidence). There is low confidence overall in assessments of the magnitude of annual average wind speed changes due to inconsistency between observational data products. It is currently not possible to attribute any aspect of the observed wind speed changes to human influence on the climate system.

Key Message 2.14: The sparseness of station data in northern Canada makes it very difficult to assess long-term wind speed trends in that region, but consistency between the available station data and a modern reanalysis dataset suggests that wind speeds increased during the 1953–2023 period in northern Canada (low confidence), particularly in fall and spring.

The assessment of observed changes in annual and seasonal average wind speeds was focused on the period when homogenized station data were available (1953–2023) and used four independent data sources. This leads to medium confidence in the assessments of observed changes in areas of southern Canada where there is consistency between data sources. Confidence is low or very low elsewhere, due to inconsistency between data products and a lack of sufficient station data, particularly in northern Canada. It is not possible to attribute the observed changes to human influence. Wind stilling has been observed in much of southern Canada, like many other parts of the world. While some recent research suggests that the causes are unrelated to human influence, many open questions remain concerning the causes and processes involved in the observed wind stilling. A recent reversal in the stilling tendency that can be seen at the hemispheric scale is not yet apparent in Canadian station data.

2.7: Key knowledge gaps and emerging issues

This section describes some of the knowledge gaps and emerging issues that affect our ability to document and understand how Canada’s climate has changed since 1900 for southern Canada and since 1950 for northern Canada and Canada as a whole.

2.7.1: The interpretation of trend estimates and uncertainties

An issue that is often not considered is that trend estimates (and other estimates of change) calculated from observations have two different interpretations. The climate system generates large amounts of spontaneous (internally generated) variability. The sequence of weather and climate variations that we observe can be viewed as one of infinitely many such sequences, all with similar statistical properties, that might have occurred. Scientists often refer to this as the “butterfly effect” (e.g., see the famous paper by Lorenz (1963)). Conceptually, nature has simply chosen one of the possible “realizations” for us to observe. Consequently, trends calculated from observations can be interpreted in two different ways: either as a description of the change that is observed in the realization of weather and climate variability that we are experiencing, or as an estimate of systematic change over time that is common to all possible realizations and is the result of an external influence on the climate system. These two interpretations of an observed change or trend are affected by different sources of uncertainty. This is analogous to a distinction that is made in statistical regression analysis (e.g., Mendenhall, 1979), where uncertainty associated with a trend line can be expressed in two ways, either as a confidence interval for the trend line when interpreted as an estimate of how the mean value of a variable like temperature is expected to vary with time, or as a wider prediction interval that is meant to be interpreted as a confidence interval for an unobserved value of that variable at a specified time.

Any estimated trend or change is affected by observational biases and uncertainties when interpreted as a description of how the observed realization of weather and climate has changed over time. The sources of bias and uncertainty in this case include instrumental measurement error, imperfect homogenization, limited and variable spatial coverage of instrumental observations, and analysis errors that result from spatial interpolation (gridding) or the use of observations to constrain a reanalysis. The estimated uncertainty ranges for the trend estimates in this chapter account for these sources of uncertainty to the extent possible.

The same trend or change estimate is also affected by uncertainty about the magnitude of natural, unforced variability when the trend or change is interpreted as an estimate of externally forced climate change. This is because different possible realizations of the evolution of the weather and climate will contain different sequences of weather, as well as the El Niño–Southern Oscillation, Pacific Decadal Oscillation, Atlantic Multidecadal Oscillation, and other events and phenomena, despite being influenced in the same way by external forcings. Slow natural variations may enhance or decrease the observed rate of change, sometimes for several decades. This might lead, for example, to differences from one CCCR report to the next in the assessed rate of Canadian warming relative to global warming. The magnitude of this source of uncertainty is difficult to estimate directly from observations, and thus it is generally estimated from climate models. Climate models simulate robust amounts of internal variability and are able to simulate multiple realizations of the evolution of weather and climate events by repeating simulations that begin with slightly perturbed initial conditions. This additional source of uncertainty, as well as climate modelling uncertainty, is considered in sections 2.4.2 and 2.5.2 in the estimates of the amount of change in Canada caused by external influences on the climate.    

2.7.2: Canada’s declining climate station data archive

A fundamental challenge for Canada is its sparse network of long-term, high-quality observing sites. Sparse observing sites, particularly prior to the 1950s and in northern Canada (e.g., Figure 2.4; Box 2.2 Figure 1), limit our ability to make robust assessments of how Canada’s climate has changed since the pre-industrial era. Models that are constrained by available observations can help in some instances, albeit with large uncertainty. For example, Box 2.4 reports estimates of externally forced warming in Canada for recent periods relative to the 1850–1900 period, when observations in only a few locations in the southern tier of Canada were available.

The decline in the coverage and quality of precipitation observations in Canada’s national climate archive affects our ability to understand how precipitation in Canada has changed even in recent decades. It potentially also limits our ability to constrain projections of future change with observations of recent historical changes. Volunteer observing networks, such as CoCoRaHS (see section 2.3.1), and the sharing of data collected by different organizations help to make data more available (Wang, Feng, Zwiers, et al., 2026) (see also BC meteorological station data as an example of a network consolidation effort), but do not necessarily address the reductions in the number of long-term records that continue to be present.

Figure take-away: The number of precipitation or temperature observations entering Canada’s national digital climate archive each year has changed.

Figure title: Availability of Canadian daily precipitation and daily temperature station data over time

Two objective measures of this decline are the number of reporting stations and the frequency of missing data, which are shown in Figure 2.20a,b for daily precipitation data for southern and northern Canada (south and north of 60°N respectively). Rates of missing data increased in the early 1990s and late 2000s, although northern Canada has experienced some recovery in the last few years. The number of stations reporting daily precipitation data has decreased since the early 1990s in southern Canada and since the late 2010s in northern Canada, where the station density was already very low. As a result of the combined effects of the reduced number of stations and the increased missing data rate, the current number of observations available for Canada as a whole has dropped to about 45% of the number available during the peak period of the 1970s and 1980s, and to about 40% for southern Canada (Figure 2.20b, black curve).

The situation for daily surface air temperature observations is similar, as can be seen in Figure 2.20c,d. The decline in the number of temperature observations is having less of an effect on the ability to document recent warming in Canada than for precipitation (Figures S1 and S3 in Wang, Feng, Zwiers, et al. (2026)), because individual temperature anomalies tend to be representative of substantially larger areas than individual precipitation anomalies.

The loss of snowfall observations in recent decades due to the replacement of manual gauges with automated gauges (see section 2.5.1.2) is also of concern. The same is true for the very low availability of long-term wind speed data, even when only average wind speeds are considered, as in section 2.6. The decline has both important practical and scientific implications. For example, the design of infrastructure that is resilient to our current climate must take prevailing climate conditions into account. Design standards, such as those that are detailed in the National Building Code of Canada, require engineers to account for the extremes of temperature, rainfall, snow loads, and wind pressures when designing new buildings, and these aspects can only be adequately estimated when long-term, high-quality observations are available. Moreover, the engineering community is increasingly required to account for future climate change as well, which is typically performed by adjusting estimates of historical extremes for future conditions (e.g., see Canadian Standard Association (2025), which describes how engineers should adjust historical rainfall intensity-duration-frequency curves for a future, warmer climate). In addition, insufficient observational data in recent decades may hinder our ability to reduce future projection uncertainty by using “emergent constraint” methods to adjust climate model output (Chapter 3, section 3.3.2).

2.7.3: The impact of low-frequency, large-scale modes of climate variability 

A further challenge is that, as elsewhere in the world, Canada’s climate is affected by low-frequency, large-scale modes of internal climate variability that quasi-periodically vary, on time scales of years to decades (Chapter 4, section 4.8). This includes the effects of well-studied and documented phenomena such as the El Niño–Southern Oscillation and the Pacific Decadal Oscillation, which affect the western and central regions of Canada, and the North Atlantic Oscillation and Atlantic Multidecadal Oscillation, which predominantly affect the eastern half of Canada (Chapter 4, section 4.8.1). These modes of variability can create conditions that deviate from normal or from the expected response to external forcings for relatively long periods of time (years to decades). The influence of these modes of variability is greater at smaller scales, and while Canada is a very large country, the country’s regions are small in the context of the global climate system. Thus, it should be recognized that our trend estimates, inferences about the amount of change to attribute to external forcings, and observationally constrained projections, could all be affected by the historical evolution of these slow, large-scale modes of natural unforced climate variation.

2.7.4: Emerging opportunities

The challenges and issues that affect the ability to describe and understand climate change in Canada also create opportunities. For example, the ability to reconstruct the historical evolution of Canada’s weather over decades through the application of reanalysis systems (section 2.3.4.4) has steadily improved in step with improvements in weather forecasting. These products have the potential to eventually provide data that are of sufficiently high quality to monitor climate change at regional scales, including elements such as wind. Indeed, one such product, the ERA5 reanalysis dataset (Hersbach et al., 2020), which starts in 1940 and has been updated continuously in near real time, is already being used by the European Copernicus Climate Change Service to monitor global average temperature on an ongoing basis. In the meantime, however, reanalyses provide many opportunities at smaller regional scales to help make the most of the available in situ observations, for example, by using them as proxies to develop methods for evaluating the effects of sampling bias on gridded station data and regional average values that are calculated from gridded station data with changing degrees of data availability (Box 2.2).

References

Abbasnezhadi, K., & Wang, X. L. (2024). Comparison of gridding methods for precipitation over Canada and assessment of station and data density effects on gridding results. Atmosphere-Ocean, 62(4), 320–346.

Allen, M. R., & Ingram, W. J. (2002). Constraints on future changes in climate and the hydrologic cycle. Nature, 419(6903), 224–232.

AMS. (2025). American meteorological society online glossary. American Meteorological Society.

Boer, G. J. (1993). Climate change and the regulation of the surface moisture and energy budgets. Climate Dynamics, 8(5), 225–239.

Bush, E., & Flato, G. (2019). About this report. In E. Bush & D. S. Lemmen (Eds.), Canada’s Changing Climate Report (pp. 7–23). Government of Canada.

Canadian Standard Association. (2025). CSA W231:25 Developing and interpreting intensity-duration-frequency (IDF) information under a changing climate (National Standard of Canada CSA W231:25).

CaSR. (2025). Canadian Surface Reanalysis, v3.1 (Version v3.1) [Dataset].

Cheng, V. Y. S., Wang, X. L., & Feng, Y. (2024). A Quality Control System for Historical In Situ Precipitation Data. Atmosphere-Ocean, 62(4), 271–287.

Dai, A. (2008). Temperature and pressure dependence of the rain‐snow phase transition over land and ocean. Geophysical Research Letters, 35(12), 2008GL033295.

Deng, K., Azorin-Molina, C., Minola, L., Zhang, G., & Chen, D. (2021). Global near-surface wind speed changes over the last decades revealed by reanalyses and CMIP6 model simulations. Journal of Climate, 34(6), 2219–2234.

Douville, H., Raghavan, K., Renwick, J., Allan, R. P., Arias, P. A., Barlow, M., Cerezo-Mota, R., Cherchi, A., Gan, T. Y., Gergis, J., Jiang, D., Khan, A., Pokam Mba, W., Rosenfeld, D., Tierney, J., & Zolina, O. (2021). Water cycle changes. In V. Masson-Delmotte, P. Zhai, A. Pirani, S. L. Connors, C. Péan, S. Berger, N. Caud, Y. Chen, L. Goldfarb, M. I. Gomis, M. Huang, K. Leitzell, E. Lonnoy, J. B. R. Matthews, T. K. Maycock, T. Waterfield, O. Yelekçi, R. Yu, & B. Zhou (Eds.), Climate change 2021: The physical science basis. Contribution of working group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (pp. 1055–1210). Cambridge University Press.

ECCC. (2015). Manual of surface weather observations (PDF). 7th edition. Amendment 19 (PDF). Environment and Climate Change Canada.

ECCC. (2016). Canadian gridded temperature and precipitation anomalies (CanGRD) [Dataset]. Open Canada.

Eyring, V., Bony, S., Meehl, G. A., Senior, C. A., Stevens, B., Stouffer, R. J., & Taylor, K. E. (2016). Overview of the coupled model intercomparison project phase 6 (CMIP6) experimental design and organization. Geoscientific Model Development, 9(5), 1937–1958.

Eyring, V., Gillett, N. P., Achuta Rao, K. M., Barimalala, R., Barreiro Parrillo, M., Bellouin, N., Cassou, C., Durack, P. J., Kosaka, Y., McGregor, S., Min, S., Morgenstern, O., & Sun, Y. (2021). Human influence on the climate system. In V. Masson-Delmotte, P. Zhai, A. Pirani, S. L. Connors, C. Péan, S. Berger, N. Caud, Y. Chen, L. Goldfarb, M. I. Gomis, M. Huang, K. Leitzell, E. Lonnoy, J. B. R. Matthews, T. K. Maycock, T. Waterfield, O. Yelekçi, R. Yu, & B. Zhou (Eds.), Climate change 2021: The physical science basis. Contribution of working group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (pp. 432–552). 

Fink-Mercier, C., Leblanc, M. L., Noisette, F., O’Connor, M., Idrobo, J., Bélanger, S., Del Giorgio, P. A., De Melo, M., Ehn, J. K., Giroux, J.-F., Gosselin, M., Leblon, B., Neumeier, U., Sorais, M., Humphries, M. M., Peck, C., Davis, K. E., Guzzi, A., Galindo, V., … Kuzyk, Z. Z. A. (2024). Cree-driven community-partnered research on coastal ecosystem change in subarctic Canada: A multiple knowledge approach. Arctic Science, 10(4), 731–748.

Forster, P. M., Smith, C., Walsh, T., Lamb, W. F., Lamboll, R., Cassou, C., Hauser, M., Hausfather, Z., Lee, J.-Y., Palmer, M. D., Von Schuckmann, K., Slangen, A. B. A., Szopa, S., Trewin, B., Yun, J., Gillett, N. P., Jenkins, S., Matthews, H. D., Raghavan, K., … Zhai, P. (2025). Indicators of Global Climate Change 2024: Annual update of key indicators of the state of the climate system and human influence. Earth System Science Data, 17(6), 2641–2680.

Gillett, N. P., Shiogama, H., Funke, B., Hegerl, G., Knutti, R., Matthes, K., Santer, B. D., Stone, D., & Tebaldi, C. (2016). The detection and attribution model intercomparison project (DAMIP v1.0) contribution to CMIP6. Geoscientific Model Development, 9(10), 3685–3697.

Giroux, J.-F., Idrobo, C. J., & Sorais, M. (2024). Bridging cree knowledge and western science to understand the decline in hunting success of migratory Canada geese. Socio-Ecological Practice Research, 6(2), 131–140.

Gulev, S. K., Thorne, P. W., Ahn, J., Dentener, F. J., Domingues, C. M., Gerland, S., Gong, D., Kaufman, D. S., Nnamchi, H. C., Quaas, J., Rivera, J. A., Sathyendranath, S., Smith, S. L., Trewin, B., von Schuckmann, K., & Vose, R. S. (2021). Changing state of the climate system. In V. Masson-Delmotte, P. Zhai, A. Pirani, S. L. Connors, C. Péan, S. Berger, N. Caud, Y. Chen, L. Goldfarb, M. I. Gomis, M. Huang, K. Leitzell, E. Lonnoy, J. B. R. Matthews, T. K. Maycock, T. Waterfield, O. Yelekçi, R. Yu, & B. Zhou (Eds.), Climate change 2021: The physical science basis. Contribution of working group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (pp. 287–422).

Held, I. M., & Soden, B. J. (2006). Robust responses of the hydrological cycle to global warming. Journal of Climate, 19(21), 5686–5699.

Hersbach, H., Bell, B., Berrisford, P., Hirahara, S., Horányi, A., Muñoz‐Sabater, J., Nicolas, J., Peubey, C., Radu, R., Schepers, D., Simmons, A., Soci, C., Abdalla, S., Abellan, X., Balsamo, G., Bechtold, P., Biavati, G., Bidlot, J., Bonavita, M., … Thépaut, J. (2020). The ERA5 global reanalysis. Quarterly Journal of the Royal Meteorological Society, 146(730), 1999–2049.

Holton, J. R. (with Hakim, G.). (2012). An Introduction to Dynamic Meteorology (5th ed). Elsevier Science & Technology.

Houghton, J. T., Jenkins, G. J., & Ephraums, J. J. (Eds.). (1990). Climate change: The IPCC scientific assessment. Cambridge University Press.

Hutchinson, M. F., & Xu, T. (2013). ANUSPLIN version 4.4 user guide (Version 4.4) [Dataset]. Fenner School of Environment, Australian National University.

Idrobo, C. J., Leblanc, M.-L., & O’Connor, M. I. (2024). The “turning point” for the fall goose hunt in Eeyou Istchee: A social-ecological regime shift from an Indigenous Knowledge perspective. Human Ecology, 52(3), 617–636.

IPCC. (2021a). Annex VI: Climatic impact-driver and extreme indices [Gutiérrez J.M., R. Ranasinghe, A.C. Ruane, R. Vautard (eds.)]. In V. Masson-Delmotte, P. Zhai, A. Pirani, S. L. Connors, C. Péan, S. Berger, N. Caud, Y. Chen, L. Goldfarb, M. I. Gomis, M. Huang, K. Leitzell, E. Lonnoy, J. B. R. Matthews, T. K. Maycock, T. Waterfield, O. Yelekçi, R. Yu, & B. Zhou (Eds.), Climate change 2021: The physical science basis. Contribution of working group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (pp. 2205–2214). Cambridge University Press.

IPCC. (2021b). Climate change 2021: The physical science basis. Contribution of working group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (V. Masson-Delmotte, P. Zhai, A. Pirani, S. L. Connors, C. Péan, S. Berger, N. Caud, Y. Chen, L. Goldfarb, M. I. Gomis, Huang, M, K. Leitzell, E. Lonnoy, Matthews, J.B.R., T. K. Maycock, T. Waterfield, O. Yelekçi, R. Yu, & B. Zhou, Eds.). Cambridge University Press.

IPCC. (2021c). Summary for policymakers. In V. Masson-Delmotte, P. Zhai, A. Pirani, S. L. Connors, C. Péan, S. Berger, N. Caud, Y. Chen, L. Goldfarb, M. I. Gomis, M. Huang, K. Leitzell, E. Lonnoy, J. B. R. Matthews, T. K. Maycock, T. Waterfield, O. Yelekçi, R. Yu, & B. Zhou (Eds.), Climate change 2021: The physical science basis. Contribution of working group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (p. 3−32). Cambridge University Press.

Ishii, M., Kamahori, H., Kubota, H., Zaiki, M., Mizuta, R., Kawase, H., Nosaka, M., Yoshimura, H., Oshima, N., Shindo, E., Koyama, H., Mori, M., Hirahara, S., Imada, Y., Yoshida, K., Nozawa, T., Takemi, T., Maki, T., & Nishimura, A. (2024). Global historical reanalysis with a 60-km AGCM and surface pressure observations: OCADA. Journal of the Meteorological Society of Japan. Ser. II, 102(2), 209–240.

Kirchmeier-Young, M., Li, G., Zhang, X., & Wang, X. L. (2025). Attribution of changes in Canadian precipitation. Atmosphere-Ocean, 1–9.

Kuzyk, Z. A., Leblanc, M., O’Connor, M., Idrobo, C. J., Giroux, J. F., del Georgio, P., Bélanger, S., Noisette, F., Fink-Mercier, C., de Melo, M., Walch, D., Ehn, J. K., Gosselin, M., Neumeier, U., Sorais, M., & Leblon, B. (2023). Understanding Shkaapaashkw: Eelgrass health and goose presence in eastern James Bay. Final report from the Eeyou Coastal Habitat Comprehensive Research Project (CHCRP). Prepared for Niskamoon Corporation. University of Manitoba, Winnipeg MB Canada.

Leblanc, M., O’Connor, M. I., Kuzyk, Z. Z. A., Noisette, F., Davis, K. E., Rabbitskin, E., Sam, L., Neumeier, U., Costanzo, R., Ehn, J. K., Babb, D., Idrobo, C. J., Gilbert, J., Leblon, B., & Humphries, M. M. (2023). Limited recovery following a massive seagrass decline in subarctic eastern Canada. Global Change Biology, 29(2), 432–450.

Lee, J.-Y., Marotzke, J., Bala, G., Cao, L., Corti, S., Dunne, J. P., Engelbrecht, F., Fischer, E., Fyfe, J. C., Jones, C., Maycock, A., Mutemi, J., Ndiaye, O., Panickal, S., & Zhou, T. (2021). Future global climate: Scenario-based projections and near-term information. In V. Masson-Delmotte, P. Zhai, A. Pirani, S. L. Connors, C. Péan, S. Berger, N. Caud, Y. Chen, L. Goldfarb, M. I. Gomis, M. Huang, K. Leitzell, E. Lonnoy, J. B. R. Matthews, T. K. Maycock, T. Waterfield, O. Yelekçi, R. Yu, & B. Zhou (Eds.), Climate change 2021: The physical science basis. Contribution of working group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (pp. 553–672). Cambridge University Press.

Li, Q., Sun, W., Huang, B., Dong, W., Wang, X., Zhai, P., & Jones, P. (2020). Consistency of global warming trends strengthened since 1880s. Science Bulletin, 65(20), 1709–1712.

Li, Q., Sun, W., Yun, X., Huang, B., Dong, W., Wang, X. L., Zhai, P., & Jones, P. (2021). An updated evaluation of the global mean land surface air temperature and surface temperature trends based on CLSAT and CMST. Climate Dynamics, 56(1–2), 635–650.

Li, T., Zwiers, F. W., & Zhang, X. (2025a). Should we think of observationally constrained multidecade climate projections as predictions? Science Advances, 11(20), eadt6485.

Li, T., Zwiers, F. W., Zhang, X., & Wang, X. (2025b). Constrained Estimates of Externally Forced Past and Future Warming for Canada. Earth’s Future, 13(10), e2025EF006374.

Liang, Y., Gillett, N. P., & Monahan, A. H. (2023). Observationally constrained projections of twenty-first-century regional warming in the extratropical Northern Hemisphere. Journal of Climate, 36(21), 7619–7633.

Lorenz, E. N. (1963). Deterministic Nonperiodic Flow. Journal of the Atmospheric Sciences, 20(2), 130–141.

MacDonald, H., McKenney, D. W., Wang, X. L., Pedlar, J., Papadopol, P., Lawrence, K., Feng, Y., & Hutchinson, M. F. (2021). Spatial models of adjusted precipitation for Canada at varying time scales. Journal of Applied Meteorology and Climatology, 60(3), 291–304.

Meinshausen, M., Nicholls, Z. R. J., Lewis, J., Gidden, M. J., Vogel, E., Freund, M., Beyerle, U., Gessner, C., Nauels, A., Bauer, N., Canadell, J. G., Daniel, J. S., John, A., Krummel, P. B., Luderer, G., Meinshausen, N., Montzka, S. A., Rayner, P. J., Reimann, S., … Wang, R. H. J. (2020). The shared socio-economic pathway (SSP) greenhouse gas concentrations and their extensions to 2500. Geoscientific Model Development, 13(8), 3571–3605.

Mekis, E., & Hogg, W. D. (1999). Rehabilitation and analysis of Canadian daily precipitation time series. Atmosphere-Ocean, 37(1), 53–85.

Mekis, E., & Vincent, L. A. (2011). An overview of the second generation adjusted daily precipitation dataset for trend analysis in Canada. Atmosphere-Ocean, 49(2), 163–177.

Mendenhall, W. (1979). Introduction to probability and statistics (6th ed.). Duxbury Press.

Milewska, E. J., Vincent, L. A., Hartwell, M. M., Charlesworth, K., & Mekis, É. (2019). Adjusting precipitation amounts from Geonor and Pluvio automated weighing gauges to preserve continuity of observations in Canada. Canadian Water Resources Journal / Revue Canadienne Des Ressources Hydriques, 44(2), 127–145.

Morice, C. P., Kennedy, J. J., Rayner, N. A., Winn, J. P., Hogan, E., Killick, R. E., Dunn, R. J. H., Osborn, T. J., Jones, P. D., & Simpson, I. R. (2021). An updated assessment of near‐surface temperature change from 1850: The HadCRUT5 data set. Journal of Geophysical Research: Atmospheres, 126(3), e2019JD032361.

MSC. (2024). Meteorological Service of Canada historical weather radar image portal [Dataset]. Canada.ca.

Muñoz-Sabater, J., Dutra, E., Agustí-Panareda, A., Albergel, C., Arduini, G., Balsamo, G., Boussetta, S., Choulga, M., Harrigan, S., Hersbach, H., Martens, B., Miralles, D. G., Piles, M., Rodríguez-Fernández, N. J., Zsoter, E., Buontempo, C., & Thépaut, J.-N. (2021). ERA5-Land: A state-of-the-art global reanalysis dataset for land applications. Earth System Science Data, 13(9), 4349–4383.

National Research Council. (2004). Climate data records from environmental satellites. National Academy Press.

NOAA NCEI. (2024). NOAA national centers for environmental information (NCEI), climate at a glance: National time series [Dataset]. National Centers for Environmental Information.

Osborn, T. J., Jones, P. D., Lister, D. H., C. P. Morice, Simpson, I. R., Winn, J. P., Hogan, E., & Harris, I. C. (2021). Land surface air temperature variations across the globe updated to 2019:The CRUTEM5 data set. Journal of Geophysical Research: Atmospheres, 126(2), e2019JD032352.

Pryor, S. C., Barthelmie, R. J., Bukovsky, M. S., Leung, L. R., & Sakaguchi, K. (2020). Climate change impacts on wind power generation. Nature Reviews Earth & Environment, 1(12), 627–643.

Qian, B., Wang, X. L., Zwiers, F. W., & Feng, Y. (2025). Observed changes in Canada’s snowfall as inferred from precipitation and below-freezing daily mean temperatures. Atmosphere-Ocean, 1-15.

Ribes, A., Planton, S., & Terray, L. (2013). Application of regularised optimal fingerprinting to attribution. Part I: Method, properties and idealised analysis. Climate Dynamics, 41(11–12), 2817–2836.

Ribes, A., Qasmi, S., & Gillett, N. P. (2021). Making climate projections conditional on historical observations. Science Advances, 7(4), eabc0671.

Santer, B. D., Solomon, S., Pallotta, G., Mears, C., Po-Chedley, S., Fu, Q., Wentz, F., Zou, C.-Z., Painter, J., Cvijanovic, I., & Bonfils, C. (2017). Comparing tropospheric warming in climate models and satellite data. Journal of Climate, 30(1), 373–392.

Schneider, U., Finger, P., Meyer-Christoffer, A., Rustemeier, E., Ziese, M., & Becker, A. (2017). Evaluating the hydrological cycle over land using the newly-corrected precipitation climatology from the global precipitation climatology centre (GPCC). Atmosphere, 8(3), 52.

Slivinski, L. C., Compo, G. P., Whitaker, J. S., Sardeshmukh, P. D., Giese, B. S., McColl, C., Allan, R., Yin, X., Vose, R., Titchner, H., Kennedy, J., Spencer, L. J., Ashcroft, L., Brönnimann, S., Brunet, M., Camuffo, D., Cornes, R., Cram, T. A., Crouthamel, R., … Wyszyński, P. (2019). Towards a more reliable historical reanalysis: Improvements for version 3 of the twentieth century reanalysis system. Quarterly Journal of the Royal Meteorological Society, 145(724), 2876–2908.

Slonosky, V. (2014). Historical climate observations in Canada: 18th and 19th century daily temperature from the St. Lawrence Valley, Quebec. Geoscience Data Journal, 1(2), 103–120.

Smith, C. D., Mekis, E., Hartwell, M., & Ross, A. (2022). The hourly wind-bias-adjusted precipitation data set from the Environment and Climate Change Canada automated surface observation network (2001–2019). Earth System Science Data, 14(12), 5253–5265.

St. George, S., & Wolfe, S. A. (2009). El Niño stills winter winds across the southern Canadian Prairies. Geophysical Research Letters, 36(23), 2009GL041282.

Sun, W., Li, Q., Huang, B., Cheng, J., Song, Z., Li, H., Dong, W., Zhai, P., & Jones, P. (2021). The assessment of global surface temperature change from 1850s: The C-LSAT2.0 ensemble and the CMST-interim datasets. Advances in Atmospheric Sciences, 38(5), 875–888.

U.S. Global Change Research Program. (2023). Fifth national climate assessment: Report-in-brief. U.S. Global Change Research Program.

USEPA. (2025). US Environmental Protection Agency (EPA) climate change web page.

Vautard, R., Cattiaux, J., Yiou, P., Thépaut, J.-N., & Ciais, P. (2010). Northern Hemisphere atmospheric stilling partly attributed to an increase in surface roughness. Nature Geoscience, 3(11), 756–761.

Vincent, L. A. (1998). A Technique for the Identification of Inhomogeneities in Canadian Temperature Series. 11(5), 1094–1104.

Vincent, L. A., & Gullett, D. w. (1999). Canadian historical and homogeneous temperature datasets for climate change analyses. International Journal of Climatology, 19(12), 1375–1388.

Vincent, L. A., Hartwell, M. M., & Wang, X. L. (2020). A third generation of homogenized temperature for trend analysis and monitoring changes in Canada’s climate. Atmosphere-Ocean, 58(3), 173–191.

Vincent, L. A., Milewska, E. J., Wang, X. L., & Hartwell, M. M. (2018). Uncertainty in homogenized daily temperatures and derived indices of extremes illustrated using parallel observations in Canada. International Journal of Climatology, 38(2), 692–707.

Vincent, L. A., Wang, X. L., Milewska, E. J., Wan, H., Yang, F., & Swail, V. (2012). A second generation of homogenized Canadian monthly surface air temperature for climate trend analysis. Journal of Geophysical Research: Atmospheres, 117(D18), 2012JD017859.

Vincent, L. A., Zhang, X., Brown, R. D., Feng, Y., Mekis, E., Milewska, E. J., Wan, H., & Wang, X. L. (2015). Observed trends in Canada’s climate and influence of low-frequency variability modes. Journal of Climate, 28(11), 4545–4560.

Vincent, L. A., Zhang, X., Mekis, É., Wan, H., & Bush, E. J. (2018). Changes in Canada’s climate: Trends in indices based on daily temperature and precipitation data. Atmosphere-Ocean, 56(5), 332–349.

Wan, H., Spassiani, A. C., & Vincent, L. A. (2025). Canada’s fourth generation of homogenized surface air temperature and its trends for 1948–2023. Atmosphere-Ocean, 63(4), 223–240.

Wan, H., Wang, X. L., & Swail, V. R. (2010). Homogenization and trend analysis of Canadian near-surface wind speeds. Journal of Climate, 23(5), 1209–1225.

Wan, H., Zhang, X., & Zwiers, F. (2019). Human influence on Canadian temperatures. Climate Dynamics, 52(1), 479–494.

Wan, H., Zhang, X., Zwiers, F., & Min, S.-K. (2015). Attributing northern high-latitude precipitation change over the period 1966–2005 to human influence. Climate Dynamics, 45(7–8), 1713–1726.

Wang, X. L. (2003). Comments on “Detection of Undocumented Changepoints: A Revision of the Two-Phase Regression Model.” 16(20), 3383–3385.

Wang, X. L. (2006). Climatology and trends in some adverse and fair weather conditions in Canada, 1953–2004. Journal of Geophysical Research: Atmospheres, 111(D9), 2005JD006155.

Wang, X. L. (2008). Accounting for autocorrelation in detecting mean shifts in climate data series using the penalized maximal t or F test. Journal of Applied Meteorology and Climatology, 47(9), 2423–2444.

Wang, X. L., Casas-Prat, M., Feng, Y., Crosby, A., & Swail, V. R. (2021). Historical changes in the Davis Strait Baffin Bay surface winds and waves, 1979-2016. Journal of Climate, 1–44.

Wang, X. L., & Feng, Y. (2026). Observed trends in precipitation extreme indices as inferred from a homogenized daily precipitation dataset for Canada. Weather and Climate Extremes

Wang, X. L., Feng, Y., Cheng, V. Y. S., & Xu, H. (2023). Observed precipitation trends inferred from Canada’s homogenized monthly precipitation dataset. Journal of Climate, 36, 7957-7971

Wang, X. L., Feng, Y., Issac, V., Zwiers, F. W., Vincent, L. A., & Hartwell, M. (2025). Observed surface wind speed trends inferred from homogenized in situ data and reanalysis datasets. Atmosphere-Ocean, 1-17

Wang, X. L., Feng, Y., Swail, V. R., & Cox, A. (2015). Historical changes in the beaufort-Chukchi–Bering seas surface winds and waves, 1971–2013. Journal of Climate, 28(19), 7457–7469.

Wang, X. L., Feng, Y., Zwiers, F. W., & Cheng, V. Y. S. (2026). Precipitation trends in version 2 of the Canadian homogenized monthly precipitation dataset. Atmosphere-Ocean, 1-16.

Wang, X. L., & Swail, V. R. (2001). Changes of extreme wave heights in Northern Hemisphere oceans and related atmospheric circulation regimes. Journal of Climate, 14(10), 2204–2221.

Wang, X. L., Swail, V. R., Zwiers, F. W., Zhang, X., & Feng, Y. (2009). Detection of external influence on trends of atmospheric storminess and northern oceans wave heights. Climate Dynamics, 32(2–3), 189–203.

Wang, X. L., Xu, H., Qian, B., Feng, Y., & Mekis, E. (2017). Adjusted daily rainfall and snowfall data for Canada. Atmosphere-Ocean, 55(3), 155–168.

WMO. (2010). Guide to the global observing system (2017th ed.). World Meteorological Organization.

WMO. (2020). Guidelines on Homogenization (2020 edition) (WMO-No. 1245; p. 54). World Meteorological Organization.

WMO. (2023). Manual on the WMO Integrated Global Observing System. World Meteorological Organization. https://library.wmo.int/viewer/55063/download?file=1160-2023-edition_en.pdf&type=pdf&navigator=1

Wohland, J., Folini, D., & Pickering, B. (2021). Wind speed stilling and its recovery due to internal climate variability. Earth System Dynamics, 12(4), 1239–1251.

Wu, J., Zha, J., Zhao, D., & Yang, Q. (2018). Changes in terrestrial near-surface wind speed and their possible causes: An overview. Climate Dynamics, 51(5), 2039–2078.

Yin, X., Huang, B., Menne, M., Vose, R., Zhang, H.-M., Adeyeye, A., Applequist, S., Gleason, K., Liu, C., & Sanchez-Lugo, A. (2024). NOAAGlobalTemp version 6: An AI-based global surface temperature dataset. Bulletin of the American Meteorological Society, 105(11), E2184–E2193.

Yun, X., Huang, B., Cheng, J., Xu, W., Qiao, S., & Li, Q. (2019). A new merge of global surface temperature datasets since the start of the 20th century. Earth System Science Data, 11(4), 1629–1643.

Zeng, Z., Ziegler, A. D., Searchinger, T., Yang, L., Chen, A., Ju, K., Piao, S., Li, L. Z. X., Ciais, P., Chen, D., Liu, J., Azorin-Molina, C., Chappell, A., Medvigy, D., & Wood, E. F. (2019). A reversal in global terrestrial stilling and its implications for wind energy production. Nature Climate Change, 9(12), 979–985.

Zhang, X., Alexander, L., Hegerl, G. C., Jones, P., Tank, A. K., Peterson, T. C., Trewin, B., & Zwiers, F. W. (2011). Indices for monitoring changes in extremes based on daily temperature and precipitation data. WIREs Climate Change, 2(6), 851–870

Zhang, X., Flato, G., Kirchmeier-Young, M., Vincent, L. A., Wan, H., Wang, X. L., Rong, R., Fyfe, J., Li, G., & Kharin, V. V. (2019). Changes in temperature and precipitation across Canada. In E. Bush & D. S. Lemmen (Eds.), Canada’s changing climate report (pp. 112–193). Government of Canada.

Supplementary Materials

Supplementary Table S2.1: Improvements in the homogenized datasets used in CCCR2026 over the previous versions or generations

Table S2.1: Different generations or versions of adjusted and/or homogenized monthly (mly) or daily (dly) average, maximum, and minimum surface air temperature (T, Tmax, Tmin), precipitation (P), rainfall (R), snowfall (S), surface wind speed (W), and freezing precipitation frequency (FPf) (log-odds) datasets used in the first edition of CCCR (CCCR2019) and this version (CCCR2026). Hom stands for Homogenized; V# stands for Version #, and G# for Generation #.

^a Community Collaborative Rain, Hail and Snow Network (CoCoRaHS) is a unique, non-profit, community-based network of volunteers of all ages and backgrounds working together to measure and map precipitation across Canada.

^b The ANUSPLIN package (Hutchinson & Xu, 2013) includes programs for interrogating irregularly spaced point values (e.g., station data) to fitted surfaces, which can be used to drive both point and gridded values as needed. 

^a Odds refer to a metric that is sometimes used to compare the probability or frequency of occurrence of an event with the probability or frequency that it does not occur. In the case of freezing precipitation, an odds ratio can be estimated for each month by calculating the ratio between the number of days in that month when the event occurred and the number of days when it did not occur. For example, if the event occurred on 7 out 31 days in a given month, then the odds ratio for that month is estimated as 7/(31-7) = 7/24, or 0.29. The log-odds is simply the logarithm of that ratio.

^a A subset of the climate extremes indices defined by the Expert Team on Climate Change Detection and Indices (ETCCDI) (Zhang et al., 2011) 

Supplementary Figure S2.1: Changes in annual average daily minimum and daily maximum surface air temperatures across Canada

Supplementary Figure S2.2: Changes in seasonal average daily minimum and daily maximum surface air temperatures across Canada

Supplementary Figure S2.3: “Normal” (three-decade average) precipitation amounts across Canada.

References used in Supplementary Materials

Cheng, V. Y. S., Wang, X. L., & Feng, Y. (2024). A Quality Control System for Historical In Situ Precipitation Data. Atmosphere-Ocean, 62(4), 271–287.

ECCC. (2016). Canadian gridded temperature and precipitation anomalies (CanGRD) [Dataset]. Open Canada.

Hutchinson, M. F., & Xu, T. (2013). ANUSPLIN version 4.4 user guide (Version 4.4) [Dataset]. Fenner School of Environment, Australian National University.

MacDonald, H., McKenney, D. W., Wang, X. L., Pedlar, J., Papadopol, P., Lawrence, K., Feng, Y., & Hutchinson, M. F. (2021). Spatial models of adjusted precipitation for Canada at varying time scales. Journal of Applied Meteorology and Climatology, 60(3), 291–304.

Mekis, E., & Hogg, W. D. (1999). Rehabilitation and analysis of Canadian daily precipitation time series. Atmosphere-Ocean, 37(1), 53–85.

Mekis, E., & Vincent, L. A. (2011). An overview of the second generation adjusted daily precipitation dataset for trend analysis in Canada. Atmosphere-Ocean, 49(2), 163–177.

Qian, B., Wang, X. L., Zwiers, F. W., & Feng, Y. (2025). Observed changes in Canada’s snowfall as inferred from precipitation at below-freezing daily mean temperatures. Atmosphere-Ocean, 1-15.

Smith, C. D., Mekis, E., Hartwell, M., & Ross, A. (2022). The hourly wind-bias-adjusted precipitation data set from the Environment and Climate Change Canada automated surface observation network (2001–2019). Earth System Science Data, 14(12), 5253–5265. 

Vincent, L. A. (1998). A Technique for the Identification of Inhomogeneities in Canadian Temperature Series. 11(5), 1094–1104.

Vincent, L. A., & Gullett, D. w. (1999). Canadian historical and homogeneous temperature datasets for climate change analyses. International Journal of Climatology, 19(12), 1375–1388.

Vincent, L. A., Hartwell, M. M., & Wang, X. L. (2020). A third generation of homogenized temperature for trend analysis and monitoring changes in Canada’s climate. Atmosphere-Ocean, 58(3), 173–191.

Vincent, L. A., Milewska, E. J., Wang, X. L., & Hartwell, M. M. (2018). Uncertainty in homogenized daily temperatures and derived indices of extremes illustrated using parallel observations in Canada. International Journal of Climatology, 38(2), 692–707.

Vincent, L. A., Wang, X. L., Milewska, E. J., Wan, H., Yang, F., & Swail, V. (2012). A second generation of homogenized Canadian monthly surface air temperature for climate trend analysis. Journal of Geophysical Research: Atmospheres, 117(D18), 2012JD017859. 

Wan, H., Wang, X. L., & Swail, V. R. (2010). Homogenization and trend analysis of Canadian near-surface wind speeds. Journal of Climate, 23(5), 1209–1225.

Wang, X. L. (2003). Comments on “Detection of Undocumented Changepoints: A Revision of the Two-Phase Regression Model.” 16(20), 3383–3385.

Wang, X. L. (2006). Climatology and trends in some adverse and fair weather conditions in Canada, 1953–2004. Journal of Geophysical Research: Atmospheres, 111(D9), 2005JD006155.

Wang, X. L. (2008). Accounting for autocorrelation in detecting mean shifts in climate data series using the penalized maximal t or F test. Journal of Applied Meteorology and Climatology, 47(9), 2423–2444.

Wang, X. L., & Feng, Y. (2026). Observed trends in precipitation extreme indices as inferred from a homogenized daily precipitation dataset for Canada. Weather and Climate Extremes.

Wang, X. L., Feng, Y., Cheng, V. Y. S., & Xu, H. (2023). Observed precipitation trends inferred from Canada’s homogenized monthly precipitation dataset. Journal of Climate, 36, 7957-7971.

Wang, X. L., Feng, Y., Issac, V., Zwiers, F. W., Vincent, L. A., & Hartwell, M. (2025). Observed surface wind speed trends inferred from homogenized in situ data and reanalysis datasets. Atmosphere-Ocean 1-17.

Wang, X. L., Feng, Y., Zwiers, F. W., & Cheng, V. Y. S. (2026). Precipitation trends in version 2 of the Canadian homogenized monthly precipitation dataset. Atmosphere-Ocean 1-16.

Wang, X. L., Xu, H., Qian, B., Feng, Y., & Mekis, E. (2017). Adjusted daily rainfall and snowfall data for Canada. Atmosphere-Ocean, 55(3), 155–168.

WMO. (2020). Guidelines on Homogenization (2020 edition) (WMO-No. 1245; p. 54). World Meteorological Organization.

Zhang, X., Alexander, L., Hegerl, G. C., Jones, P., Tank, A. K., Peterson, T. C., Trewin, B., & Zwiers, F. W. (2011). Indices for monitoring changes in extremes based on daily temperature and precipitation data. WIREs Climate Change, 2(6), 851–870.