Predictive Skill and Bias of Synoptic Variables during 14-day Extreme Precipitation Periods in S2S Models

Long periods of extreme precipitation can cause significant impacts to life and property. Planning for these periods would ideally begin at the subseasonal to seasonal (S2S) time scale, yet prediction of precipitation at this time scale has low skill. A database of 14-day extreme precipitation periods across 16 North American regions (clusters) was used to assess the current predictive skill of synoptic variables during the extreme periods within three S2S project model hindcasts, including 500hPa geopotential heights, various levels of specific humidity, and 500hPa omega. These variables were chosen because they have previously been shown to be potentially skillful predictors of 14-day extreme precipitation periods. The highest skill is seen for geopotential height, with Anomaly Correlation Coefficient (ACC) values near 0.7. Omega has low skill beginning in Week 1, with ACC values never reaching 0.5. Since ACC drops below climatology by Week 2 for all variables, model bias during the extreme periods was examined. Initial findings show increasing dry bias with lead time for all levels of specific humidity. The models also have an inability to realize the magnitude of 500hPa geopotential height dipole, common to all extreme periods. Therefore, bias correction techniques will be implemented, with the goal of using bias-corrected real-time forecasts as input into a statistical model to improve the prediction of 14-day precipitation extremes.