Evaluation of Predictive Skill of Flash Droughts over China based on S2S Forecast Models

Abstract

Flash droughts have been occurring frequently worldwide, which has a serious impact on food and water security. Flash droughts have raised a wide concern, but whether they can be predicted at sub-seasonal time scale remains unclear. We investigate the forecast skill of flash droughts over China with lead times up to three weeks by using model hindcasts from the sub-seasonal-to-seasonal prediction (S2S) project. The flash droughts are identified by using weekly soil moisture percentiles from two S2S forecast models (ECMWF and NCEP). The comparison with reanalysis shows that ECMWF and NCEP forecast models underestimate flash drought occurrence. The ensemble of the two models increases equitable threat score from ECMWF and NCEP models for lead 1 week. In terms of probabilistic forecast, ECMWF also has higher brier skill score than NCEP especially over Eastern China, which is consistent with higher temperature and precipitation forecast skill and flash drought predictability of ECMWF model. The multi-model ensemble shows a higher skill than ECMWF model. This study suggests the importance of multi-model ensemble flash drought forecasting.