A machine-learning-based perspective on deep convective clouds and their organisation in 3D : Part 2: Spatial–temporal patterns of convective organisation

Abstract

This series of papers explores spatio-temporal patterns of convective cloud occurrence and organisation. We use a machine-learning-based method to extrapolate a contiguous 3D cloud field of 2D satellite data. In Part 2, we focus on convective organisation in tropical West Africa between March and August 2019, examining how it relates to the 3D properties of convective clouds and their core structures. We quantify organisation using three indices (SCAI, COP, ROME) to capture different aspects of spatial cloud clustering. Our findings highlight how cloud properties may interact with organisation. Hence, strong organisation may occur with larger cloud areas, lower cloud tops and core heights, and shorter lifespans compared to the average convective system. In contrast, weak organisation may be associated with smaller clouds and fewer cores but similarly shorter lifespans. We find an increasing frequency of convective organisation in the Northern Hemisphere during boreal summer months, likely linked to the northward migration of the Intertropical Convergence Zone (ITCZ). From March to May, patches of strong convective organisation emerge along the African coastlines and over the southern Atlantic Ocean. Between June and August, hotspots shift inland, particularly across the Sahel and wider West African Plains. Notably, oceanic regions show slightly stronger organisation overall. However, overlapping regions of strong and weak organisation may complicate the interpretation of regional statistics. While the machine-learning-based 3D perspective helps bridge observational gaps in the representation of cloud structures, the inherent complexity and variability of convective organisation highlight the need for continued investigation.