Skip to content

Probabilistic rainy season onset prediction over the Greater Horn of Africa based on long-range multi-model ensemble forecasts

We are thrilled to share that our research has been published in Springer, a highly regarded international scientific journal. 

The Paper titled “Probabilistic Rainy Season Onset Prediction Over the Paperer Horn of Africa Based on Long-Range Multi-Model Ensemble Forecasts,” led by CONFER group (Norsk Regnesentral (Norwegian Computing Center, NR) and IGAD Climate Prediction and Applications Centre (ICPAC)), presents an innovative approach that promises predictability and enhanced skill of the onset of the rainy season over the Great Horn of Africa.

After applying the best bias-correcting technique and probabilistic framework, the research shows predictability of the rainy season onset several weeks in advance based on multi-model global daily precipitation ensemble hindcasts from Copernicus Climate Change Service seasonal forecast systems.

By utilizing the best bias-correction technique and probabilistic framework, our research demonstrates the capability to predict the onset of rainy seasons several weeks in advance. This approach leverages multi-model, global daily precipitation ensemble hindcasts from the Copernicus Climate Change Service seasonal forecast systems.

While daily precipitation amounts at individual locations are not predictable at seasonal lead times, an ensemble of model simulations may skillfully present a tendency towards an earlier or later occurrence of days with increased precipitation.

Our analysis shows that the long-range forecasts may provide up to 10% improvement over a climatological onset forecast as measured by the Brier skill score, with a larger ensemble yielding better skill. Consistent with high seasonal predictability, the skill of the probabilistic onset forecasts is better for the OND season than the MAM season. 

The GHA is one of the region’s most vulnerable to recurrent extreme climate events globally, with the risks increasingly becoming complex as these hazards are compounded by local and remote political and economic instability. Therefore, such reliable and actionable probabilistic rainy season onset forecast information on the seasonal rainfall can better support decision-making and enhance the provision of climate information services that can reduce the impact of hazards.