Reanalysis products, or gridded datasets more broadly, are often used in place of surface observations. While they have been shown to capture long-term statistics on global or regional levels, it is still unclear how well they perform at the tails of the distribution, especially on daily timescales. Four widely used datasets, ERA5, ERA5-Land, MERRA-2, and PRISM, were assessed for their ability to capture extreme heat, extreme cold, and heavy precipitation events over the contiguous US (CONUS). While biases are evident in each dataset, particularly across the western US for temperature and along the Gulf Coast for heavy precipitation, all datasets do reasonably well in capturing extreme events and trends. Extreme heat is better represented than extreme cold or heavy precipitation. While no dataset emerges as a clear best for extreme heat, PRISM generally performs best for extreme cold and the bias-adjusted MERRA-2 dataset generally performs best for heavy precipitation days.

On any given day, some region of the globe is likely to be affected by significant negative temperature anomalies. Depending on the severity, extreme cold periods may be deemed cold air outbreaks (CAOs). These CAOs can be detrimental to the agricultural industry and human health, especially in less prepared regions. A systematic CAO classification was developed using a set of criteria concerning magnitude, duration, and spatial extent from two gridded reanalysis datasets from 1979 – 2018. Trends in CAOs were calculated for different regions across the globe and the results from each reanalysis dataset compared with one another to identify discrepancies. CAOs were found to have decreased in spatial extent, frequency, duration, and magnitude across much of the globe, particularly across Alaska, Canada, and the North Atlantic, while an increase in CAOs was observed in Eastern Europe, Central Eurasia, and the Southern Ocean.

While many studies comparing atmospheric reanalysis and surface observations have focused on the similarity of mean fields, trends, or frequencies of extreme events, very few have assessed how similar surface observations and reanalysis data sets are in terms of their specific identification of extreme temperature event days. Here, we assess the similarity between surface observations and three reanalysis products: ERA5, ERA5-LAND, and NARR, in terms of the days on which they identify extreme temperature events. We assess similarity from 1979-2016 for 231 locations in the United States and Canada, assessing Extreme Heat and Cold Event days, as well as their counterpart events that are relative for the time of year. Cold Events have a greater match than Heat Events. ERA5 has the greatest match percentage with station data across the study region. Match percentage is greatest in mid-latitude, continental locations, with poorer performance in coastal areas, and the Arctic.

Upon the backdrop of steadily rising global average temperatures, it is the extreme weather events that are arguably more important and impactful than changing averages – especially on human health. This research examines trends in North America of three different parameters of extreme temperature events important to human thermal comfort and public health: their frequency, duration, and spatial extent. Most of the changes are expected; that is, with warmer temperatures there are more frequent extreme heat events that are lasting longer and covering more area. However, we highlight some intriguing divergences from this pattern. For example, despite quickly rising autumn temperatures in northern Canada, a concurrent decrease in temperature variability is resulting in extreme heat events remaining stable and is instead manifest more as significant decreases in extreme cold events. In parts of the western US, even though there is no significant trend in autumn mean temperatures, there is a significant rise in extreme cold events. And, in the southern High Plains in summer, despite little trend in averages, a more negative skew to the distribution is nonetheless leading to significant increases in heat events. Seasonal and geographic variability in the trends of extreme dew point events is also explored. For example, increases in extreme humidity events are ubiquitous throughout most of Canada, particularly in summer; but the US has a northeast (increasing humid events) to southwest (increasing dry events) dichotomy that is strongest in winter. While such nuances might complicate our efforts to broadly generalize the message of climate change, these distinctions suggest a renewed emphasis on local- to regional-scale analyses (rather than larger scales) when providing actionable climate information for planners and policymakers.

Historical and future simulated temperature data from five climate models in the Coupled Model Intercomparing Project Phase 6 (CMIP6) are used to understand how climate change might alter cold air outbreaks (CAOs) in the future. Three different Shared Socioeconomic Pathways (SSPs), SSP 1 – 2.6, SSP 2 – 4.5, and SSP 5 – 8.5 are examined to identify potential fluctuations in CAOs across the globe between 2015 and 2054. Though CAOs may remain persistent or even increase in some regions through 2040, all five climate models show CAOs disappearing by 2054. Climate models were able to accurately simulate the spatial distribution and trends of historical CAOs, but there were large errors in the simulated interannual frequency of CAOs in the North Atlantic and North Pacific. Fluctuations in complex processes, such as Atlantic Meridional Overturning Circulation, may be contributing to each model’s inability to simulate historical CAOs in these regions.