Understanding Observed Changes in Heat Extremes

KAREN MCKINNON – UNIVERSITY OF CALIFORNIA, LOS ANGELES

ABSTRACT

As the world warms, we are experiencing an expected increase in summertime heat extremes. However, due to the large magnitude of internal variability in the climate system as well as the inherent rarity of extreme events, it can be difficult to determine the manner in which extremes are changing. In particular, are we observing a “simple shift” in summertime temperatures, wherein summertime heat extremes are warming at the same rate as average summer temperatures? Or have changes in summertime extremes been amplified (or damped) due to more complex changes in the shape of daily temperature distributions? Using historical data and a quantile regression framework, I will show that the observed behavior of summertime extremes at most locations across the Northern Hemisphere midlatitudes cannot be distinguished from the “simple shift” model. Under this null hypothesis, I will next explore a case study of the very extreme 2021 Pacific Northwest heatwave, during which multiple weather stations recorded temperature anomalies in excess of four standard deviations above a time-varying mean that accounts for the climate change signal. The magnitude of standardized temperature anomalies across stations is well-predicted by their climatological skewness, estimated in advance of the heatwave, indicating the importance of accounting for the climatological characteristics of temperature when assessing the probability of very extreme events. Using this fact, and large ensembles of climate model simulations, we find that similarly large events can be produced by climate model simulations of the past century, but that they are very rare. Finally, motivated by the importance of non-normality in controlling the magnitude of extremes, I perform a hemispheric-wide assessment of changes in the hottest annual temperatures after accounting for a mean shift using ranks, which are not affected by spatial variations in non-normality. Defining the lowest rank as the hottest anomaly from the time-varying shift in the median, I find that 2021 was the most amplified summer on record due to heat waves in the Western US, Western Russia, and Iran.