A Sub-Seasonal Forecast Multi-System for Extreme Events Prediction

Producing sub-seasonal to seasonal forecasts aimed at assessing the probability of occurrence of extreme events is crucial for many economic sectors, in particular to agriculture and agribusiness. In this study we present a multimodel ensemble that includes four of the sub-seasonal prediction systems involved in the S2S project, in an effort to improve the quality of forecasts in targeted times of year, identified as the most sensitive by agribusiness end-users. Production of subseasonal forecasts aimed at quantifying the probability of climate anomalies and extremes and the hazard associated with the occurrence of intense events, starts 45-60 days ahead.

The multi-model will be providing a probabilistic forecast, which expresses the likelihood of occurrence for the selected events together with an hazard coefficient, that companies may use for risk analysis. Here we analyze a few case studies of intense meteorological events occurring at specific times of the season, that could be particularly impacting on agriculture and cause significant losses in the yearly harvest.