Empirical and quantile-based analysis of compound extreme events

Compound extreme events have broader and more significant impacts on natural and social environments compared with individual occurrences. To this end, detection, and quantification of changes in the compound extreme events under a warm climate are important for making reliable risk projections and for understanding changes in compound extremes. This study quantified the characteristics of compound events based on i) empirical approach, and ii) quantile-based approach using both observed (1964–2014) as well as dynamically downscaled climate model data (2025–2075) for select locations in the United States. Compound events were defined based on the combination of temperature and precipitation (cold-wet, cold-dry, warm-wet, warm-dry) using 25% and 75% thresholds of temperature and precipitation. The empirical approach for the analysis of compound extreme events was executed through counting the number of compound extreme events on annual and seasonal scale and evaluating their trends for both the historical and the future time periods. Trend analysis shows increasing trends of warm-wet events for future climate conditions as compared to the historical period during both Representative Concentration Pathways (RCP) scenarios (4.5 and 8.5) with marked increase during RCP 8.5 scenario. Quantile regression method was used to detect the changes in conditional probabilities of compound extreme events. Results show that conditional trends of upper quantiles/tails have increased for warm-wet events and lower quantiles have increased for cold-wet events as compared to mean trends indicating that both warm and cold dominated compound extreme events are becoming more severe. Results from this study provide better understanding of the changes in compound events.