Tracking mesoscale convective systems in the Sahel: Relation between cloud parameters and precipitation

摘要

Although mesoscale convective systems (MCSs) are the main source of precipitation in the semi-arid Sahel region, the relationship between MCS characteristics and their generated precipitation remain unclear. However, a thorough understanding of this relation is essential to work towards a classification scheme for MCSs and eventually to improve quantitative precipitation estimates in which cloud parameters are used as proxy variables for the total or maximum intense rainfall from a system. Accordingly, this study aims to analyse the cloud parameters and rain variables distributions and their concurrence before quantifying the relationships between them. This is done using hourly EUMETSAT's Meteosat-8 infrared (10.8 μm) images, 3-hourly precipitation data from National Aeronautics and Space Administration (NASA)'s Tropical Rainfall Measuring Mission and an MCS tracking algorithm. The period of interest extends from 1 June till 22 September 2006 and the area of interest covers the Lake Chad region. Results indicate that MCSs in the Sahel region generally show a maximum cloud coverage around 57 000 km ~2, a life duration of 9 h, an embedded convective core during 6 h and precipitation peaks around 12.3 mm h -1. A recurrent sequence of cloud and rain variables is also noticed; maximum in cloud coverage is mostly preceded by a minimum in brightness temperature in the cold convective core and is followed by a peak in precipitation. Longer-lived and larger MCSs as well as MCSs embedding very cold and long-lived convective cores exhibit an increased likelihood to induce more intense precipitation. Focussing on the characteristics of the cold convective core rather than on the characteristics of the entire system appears to be more relevant to predict the precipitation as the former are better correlated with the generated precipitation and can be used as proxy parameter for estimations of maximum intense precipitation using two-dimensional nonlinear regression models.