Abstract:

Several studies have recently shown that rapid changes in temperature are linked to higher mortality and hospitalization rates, increased risk of respiratory epidemics and lower economic growth. Here, we present the first continental-scale analysis of day-to-day variability, defined as volatility, of daily maximum temperature across the contiguous United States for the period (1950–2019). We first outline spatial patterns of climatology, seasonality and directionality (increase vs. decrease) in volatility using data from 1,484 ground stations. Next, we investigate the existence of trends in the magnitude of extreme volatilities where it is found that most stations in Central US exhibit a statistically-significant increasing trend of approximately 0.2–0.3 °C/decade in magnitude of volatility. Furthermore, we compare these results with those obtained from two reanalysis data sets and one gridded (station-based) data set. Although reanalysis data sets adequately capture spatial variability of volatility, they underestimate the magnitude of volatility and fail to reproduce observed trends which might be attributed to deficiencies in numerical weather prediction models. Overall, these findings pinpoint changes in volatility of daily maximum temperature as a potential manifestation of climate change.