During the AMMA campaign, an algorithm called ISIS was used to track MCSs over West Africa. The locations of initiation of several hundred MCSs were extracted. The data were then refined using a backtracking method to locate more precisely the initiation (within 10km). The atmospheric profiles of each case were taken from the ECMWF analyses. Two satellite-based estimates were used to characterise surface properties: AMSR-E data and Land Surface Temperature Anomalies (LSTA) derived from MSG data. LSTA correspond to transient features at timescale lower than intra-seasonal. They are well suited for capturing soil moisture anomalies associated with rainfall events. The LSTA data set has a high spatial resolution that enables us to look at scale down to few tens of kilometres.
At the mesoscale (1deg x 1deg), the combining of AMSR-E soil moisture estimates with ISIS data shows that convection is initiated more frequently over dry soil. This is in agreement with the few previous case studies. In addition, some seasonal variations are highlighted. The same analysis was carried out with LSTA data analysis and leads to similar findings, namely initiation is favoured over warm/dry patches. The atmospheric analyses show that daytime convective initiations occur over a broad range of environments, from dryer/warmer low levels prior to the monsoon onset, to moister/cooler conditions during the core of the monsoon season. However, they also suggest that the mean convective inhibition is large over the region prior to the onset and associated with high levels of free convection. At a finer scale, the links between LSTA gradients and convection initiation were investigated. This study shows that there is a tendency for systems to initiate over LSTA gradients . More precisely, there is a majority of cases triggered on the warm side of a warm to cold transition aligned with the mean low-level flow.