Assessing regional behavior of extreme value statistics from NARCCAP

The North American Regional Climate Change Assessment Program (NARCCAP) provides a robust ensemble of regional climate model simulations that can be employed to investigate the behavior of weather and climate extremes. We analyze select NARCCAP simulations to evaluate differences in the 20th and 21st century behavior of outliers (extremes) in the statistical distributions of surface precipitation and temperature. Particular attention is focused on precipitation at daily and monthly time scales, and on precipitation intermittency. Because of the climatic importance of low-frequency modes of variability like ENSO, conditionally sampling based on warm, cool, and neutral phases of ENSO uncovers substantial differences in extreme value statistics, though the differences exhibit a great degree of regional variability. These results will enable us to better be able to parse contributions to climate extremes from natural low-frequency variability versus those from secular trends.