Drought modes in West Africa and how well CORDEX RCMs simulate them

摘要

This study presents the spatial-temporal structure of droughts in West Africa and evaluates the capability of CORDEX regional climate models in simulating the droughts. The study characterize droughts with the standardized evapo-transpiration index (SPEI) computed using the monthly rainfall and temperature data from the Climatic Research Unit (CRU) and CORDEX models simulation datasets. To obtain the spatial-temporal structure of the droughts, we applied the principal component analysis on the observed and simulated SPEIs and retained the first four principal factors as the leading drought modes over West Africa. The relationship between the drought modes and atmospheric teleconnections was studied using wavelet coherence analysis, while the ability of the CORDEX models to simulate the drought modes was quantified with correlation analysis. The analysis of the relationship between drought modes and atmospheric teleconnections is based on SPEI from observation dataset (CRU). The study shows that about 60 % of spatial-temporal variability in SPEI over West Africa can be grouped into four drought modes. The first drought mode features drought over east Sahel, the second over west Sahel, the third over the Savanna, and the fourth over the Guinea coast. Each drought mode is linked to sea surface temperature anomalies (SSTAs) over tropical areas of Pacific, Atlantic, and Indian Oceans. Most CORDEX models reproduce at least two of the drought modes, but only two models (REMO and CNRM) reproduce all the four drought modes. REMO and WRF give the best simulation of the seasonal variation of the drought mode over the Sahel in March-May and June-August seasons, while CNRM gives the best simulation of seasonal variation in the drought pattern over the Savanna. Results of this study may guide in selecting appropriate CORDEX models for seasonal prediction of droughts and for downscaling projected impacts of global warming on droughts in West Africa.