Initiation of daytime moist convection over the semi-arid Tropics

Poster PDF

Description

The life cycle of tropical convection is still challenging for state-of-the-art weather and climate models. Initiation of deep convection, i.e., the first step in this life cycle, is not well captured, and leads to a wrong phase of the diurnal cycle of deep convection over land. A case study of daytime development of deep convection over tropical semi-arid land has been designed (Couvreux et al., 2012). This case is representative of typical initiation over the Sahel as assessed by a climatological analysis of the observations collected during the African Monsoon Multidisciplinary Analysis (AMMA) field campaign at the ARM Mobile Facility deployed in Niamey for an entire year (Dione et al., 2013). This case is used here to analyse the processes involved in convective initiation and to provide a framework for the evaluation and improvement of these processes in models. Large-eddy simulations (LES), validated against those observations, serve as a basis for addressing these issues. The first part of the present work inter-compares a reference LES and single-column model (SCM) versions of state-of-art Earth system models, run with identical initial and boundary conditions (Couvreux et al., 2015). Most SCMs do not correctly simulate the timing of convective initiation although recent developments of the boundary-layer and deep-convection parametrizations lead to substantial improvements. The second part analyzes further a process involved in convection triggering, but still not taken into account in global models, namely the role of surface heterogeneities. In particular, we aim at understanding how surface patterns caused by antecedent precipitating events influence the time and location of new convective events. This work is based on a dozen LES simulations of the 10 July, 2006 case, in which a 30-km-wide patch is introduced via modifications of surface fluxes that are consistent with satellite estimates of the scale and shape of surface-temperature fluctuations. It is shown that the size and intensity of the patch matter. The role of local processes (modification of the boundary-layer properties induced by larger sensible heat fluxes) versus that of the breeze circulation induced by the surface differential heating at the 30-km scale are further disentangled (Rochetin et al., 2016).