Spatiotemporal Variation of Condensed Water Path over Africa

This study examines how condensed water paths (CWP) change with the seasons over Africa. This is done by comparing outputs from the Coupled Model Intercomparison Project 5 and 6 (CMIP5 and CMIP6), the CORDEX-Africa models (AFR-22 and AFR-44), and ERA5 reanalysis data from 1970 to 2005. All datasets captured the seasonal variations in condensed water paths over Africa, with some variations in magnitude among models. Datasets were also bimodal for West Africa (June-July-August and September-October-November) and Central and East Africa (March-AprilMay and September-October-November). This corresponds with the seasonal variations of precipitation, meaning more cloud condensate leaves the atmosphere as precipitation. A comparison of model outputs with the ERA5 reanalysis indicated that CMIP5 overestimated CWP while CMIP6 performed better, reproducing spatial patterns with a correlation of 0.9 - 0.94 in all seasons. The Taylor skill score further confirmed CMIP6's improved skills (greater than 0.75) in simulating the multiyear mean and condensed water path for all seasons. The CORDEX-Africa models, however, showed a lower correlation for the AFR-44 models, with a slight improvement for the AFR-22 models in representing spatial patterns. The models also showed the worst skill in simulating condensed water paths in all cases. The contrasting magnitude of CWP between GCMs and RCMs and the better consistency of GCMs with observational data raises concerns about the representativeness of this key parameter in RCMs and the widespread adoption of RCM-based scenarios for impact studies.