Decadal and Long-Term U.S. Trends of Extreme Precipitation and Temperature Using CMIP5 Hindcast Data

An extreme index of monthly temperature and monthly precipitation is presented and analyzed using hindcast model output data from thirteen model ensembles from the Coupled Model Intercomparison Project Phase 5 (CMIP5), including comparison to observations. An extreme index is constructed by calculating standardized monthly anomalies of temperature and precipitation. This index is then divided into two parts: a positive index for values that are >+1.5*σ (the standard deviation of the standardized monthly anomalies) and a negative index for values that are <-1.5*σ.

Since 1980, we find that the frequencies of both extreme heavy precipitation events and extreme high temperature episodes have been increasing over the continental U.S., while the frequency of extreme low temperature episodes has been decreasing. As such, the 1981-2010 30-year hindcast is an interesting period to analyze. The specific question is whether the 30-year hindcast simulations reproduce these observed trends, which are presumably driven by anthropogenically forced climate change. If so, then simulated information on the sub-30 year timescale may be applicable in decision-making and planning. The analysis is divided into regions of the US, i.e. the southwest, southeast, northeast, and midwest. Initial results suggest that the hindcasts reproduce the observed extreme temperature trends.