Variability of condensed water path and precipitation over Africa.

To gain a deeper understanding of precipitation variability, it is essential to also examine the variability of the condensed water path, which is vertically integrated mass of condensed liquid (LWP) and ice water (IWP) in a column, divided by the column's area. This analysis provides valuable insight into the dynamics and physics driving temporal variations in precipitation. Additionally, since cloud formation is heavily influenced by the atmosphere's liquid and ice water content, such an evaluation will aid in addressing uncertainties related to cloud-radiation interactions in global climate models (GCM).In this study, we analyze the spatial pattern of the condensed water path (CWP) and precipitation over Africa from 1970 to 2005, examining each season individually. We also address the performance of global climate models (CMIP5 and CMIP6) and regional climate models (CORDEX-Africa, AFR-22 and AFR-44) in simulating these patterns. Additionally, we investigate the temporal variations of these variables over the study period. All models successfully captured seasonal variations in CWP and precipitation, though with some differences in their magnitude. For the condensed water path, results showed a bimodal pattern in West Africa (June-July-August and September-October-November), and in Central and East Africa (March-April-May and September-October-November), aligning with the seasonal migration of the intertropical convergence zone (ITCZ). A similar bimodal pattern was observed for precipitation, except in East Africa ’s minor rainfall season (September-October-November), which was not captured. Trend analysis revealed a positive trend for all CWP datasets during the JJA season, as well as in the RCM precipitation. Conversely, GCM data showed a negative trend in the same season. In the SON season, all model outputs (except ERA5 datasets which indicated a negative trend for both CWP and precipitation) showed a positive trend for both variables.When comparing models, CMIP5 was found to overestimate CWP over Africa, whereas CMIP6 demonstrated better performance, accurately reproducing spatial patterns with correlations ranging from 0.9 to 0.94 across seasons. Precipitation data showed a similar pattern, with CMIP6 achieving correlations between 0.87 and 0.94.  Taylor skill scores further confirmed CMIP6’s improved skill, with scores exceeding 0.75 for CWP simulation and over 0.7 for precipitation in all seasons, suggesting notable progress in climate modelling.The CORDEX-Africa models, however, demonstrated a lower correlation for spatial patterns of CWP, with AFR-44 models scoring between 0.55 and 0.68 and a slight improvement in AFR-22 models (0.61 to 0.74). For precipitation, the correlation was higher, with AFR-44 models achieving 0.74 to 0.8 and AFR-22 models reaching 0.75 to 0.85 in representing spatial patterns.The consistent spatial variations in these variables, as shown by the ERA5, CMIP, and CORDEX models, offer valuable insights into the physics and dynamics underlying precipitation variations across Africa. However, the observed inconsistency in temporal variations warrant further investigation. A deeper understanding of these dynamics could substantially enhance our comprehension of climate change impacts in Africa.