West African monsoon intraseasonal activity and its daily precipitation indices in regional climate models: diagnostics and challenges

The West African monsoon intraseasonal variability\ud has huge socio-economic impacts on local populations\ud but understanding and predicting it still remains a\ud challenge for the weather prediction and climate scientific\ud community. This paper analyses an ensemble of simulations\ud from six regional climate models (RCMs) taking\ud part in the coordinated regional downscaling experiment,\ud the ECMWF ERA-Interim reanalysis (ERAI) and three\ud satellite-based and observationally-constrained daily precipitation\ud datasets, to assess the performance of the RCMs\ud with regard to the intraseasonal variability. A joint analysis\ud of seasonal-mean precipitation and the total column\ud water vapor (also called precipitable water—PW) suggests\ud the existence of important links at different timescales\ud between these two variables over the Sahel and highlights\ud the relevance of using PW to follow the monsoon seasonal\ud cycle. RCMs that fail to represent the seasonal-mean position\ud and amplitude of the meridional gradient of PW show\ud the largest discrepancies with respect to seasonal-mean\ud observed precipitation. For both ERAI and RCMs, spectral\ud decompositions of daily PW as well as rainfall show\ud an overestimation of low-frequency activity (at timescales\ud longer than 10 days) at the expense of the synoptic (timescales\ud shorter than 10 days) activity. Consequently, the effects of the African Easterly Waves and the associated\ud mesoscale convective systems are substantially underestimated,\ud especially over continental regions. Finally, the\ud study investigates the skill of the models with respect to\ud hydro-climatic indices related to the occurrence, intensity\ud and frequency of precipitation events at the intraseasonal\ud scale. Although most of these indices are generally better\ud reproduced with RCMs than reanalysis products, this\ud study indicates that RCMs still need to be improved (especially\ud with respect to their subgrid-scale parameterization\ud schemes) to be able to reproduce the intraseasonal variance\ud spectrum adequately