Drought events represent a potential source of increasing risk for several socio-economic sectors, ranging from water management and agriculture to the energy and health systems. Skilful seasonal forecasts may crucially affect the planning and governance of these activities. In this study, the skill of seasonal prediction systems (SPS \(_\textrm{s}\) ) in predicting meteorological drought in the Mediterranean region is investigated. In particular, the 3-month standardised precipitation index (SPI3) and standardised precipitation evapotranspiration index (SPEI3) at 1-month lead time are used and their probabilistic accuracy is measured using the Brier skill score (BSS). The evaluation considers the performance of individual prediction systems as well as the optimization of the forecasts using a selection of multi-model ensembles (MME) combinations. This study demonstrates that SPS \(_\textrm{s}\) can successfully predict meteorological drought indices in the Mediterranean region, given the overall positive values of the optimized BSS. We find that SPEI3 is more predictable than SPI3 with varying spatial patterns across different seasons. Results demonstrate the potential of optimizing forecast scores using MME combinations, with the Iberian Peninsula, the Balkan area, the Anatolia, the Middle East and North Africa exhibiting the largest performance.
