Predicting Seasonal Climate Characteristics Over Senegal using a New Multi-Model Approach

Abstract:

Seasonal rainfall characteristics in Senegal, and across the Sahel region of Africa, are of great importance for water provision, agricultural production and food security. However, seasonal climate prediction over Senegal has proven challenging in the past for several reasons: weaker ENSO teleconnections than other areas in the tropics, frequency of “false starts” to the rainy season, and differences in patterns of variability between Senegal and other parts of the Sahel further east. Furthermore, intra-seasonal characteristics, such as the number of rainy days or probability of a long dry spell are especially relevant for food security and agricultural decisions.

In this study, the authors use a multi-model ensemble forecasting approach, employing a suite of models from the North American Multi-Model Ensemble (NMME) and a recently developed Python-based Climate Predictability Tool (PyCPT), to investigate the predictability of various characteristics of the rainy season over Senegal. This study explores the skill of individual model and multi-model ensemble forecasts of seasonal rainfall total and rainy day frequency using modeled rainfall and low-level winds as predictors across several lead times. Preliminary results indicate relatively high skill for some models and for the multi-model ensemble at relatively long lead times (4-5 months) and more robust predictive skill for the early to middle part of the rainy season than the end of the rainy season. This newer multi-model approach complements the current methods employed for seasonal climate forecasting in Senegal.