Enhancing Resilience to Heat Extremes: Multi-model Forecasting of Excessive Heat Events at Subseasonal Lead Times

Heatwaves are among the most dangerous, yet invisible, of natural hazards. According to NOAA, the distribution of 30-year based annual mean fatalities from natural hazards in the U.S. ranks as follows; those from heat (130), floods (81), tornadoes (70), lightning (48) and hurricanes (46). Early warning to excessive heat events can be improved by using multi-scale prognostic systems. We designed and developed such a system for forecasting excessive heat events at lead times beyond Week-1. This Subseasonal Excessive Heat Outlook System (SEHOS) consists of (a) a monitoring/verification component and (b) a forecasting component which in its baseline version uses NOAA’s Global Ensemble Forecast System (GEFS) predictions of temperature and humidity from Day-8 to Day-14. In this presentation, we discuss the definition of heat events, sources of predictability and present the forecast skill of SEHOS for the GEFS reforecast period (1985-2014) and realtime forecasts during summer 2016. We then use subseasonal reforecasts from several models from the S2S database and discuss the forecast value added by multi-model approaches in predicting excessive heat events.