A machine learning-based perspective on deep convective clouds and their organisation in 3D. Part I: Influence of deep convective cores on the cloud life-cycle

Abstract. In this two-part sequence of papers, we investigate spatio-temporal patterns of convective cloud activity and organisation. The analysis employs a machine learning (ML)-based contiguous 3D extrapolation of satellite data from multiple sensors to simultaneously follow horizontal and vertical cloud development. Our study covers West Africa, a hotspot for intense convection and severe weather. In this part, we derive seasonal and diurnal variations for convective cloud properties during spring (March–May) and summer (June–August). Moreover, we explore the connection between the number of deep convective cores (DCCs) and the cloud life-cycle. For that purpose, we track the evolution of convective systems and their core regions. More than 80 % of detected clouds contain a single convective core and persist between 1–3 hours. These isolated clouds have an enhanced absolute cooling but weaker anvil growth and updraft strength during their growth stage than clustered systems. The average difference between oceanic and continental cloud properties accounts for about 10 %. However, we detect a high seasonal variability and a surface-specific diurnal cycle. We find long-lasting cloud clusters with more intense core regions over continental Africa. Within these clusters, the interaction between cores may renew convective activity. The horizontal growth of convective clouds is about 5–10 % more intense over land in both seasons, even though convective activity over the ocean increases stronger in summer. While our results emphasise an enhanced convective activity over land, we suggest further analysis of regional patterns of clustered convection and their hydro-climatological impact.