Assessment of the capability of the WRF-ARW model at convective permitting resolution to reproduce extreme precipitation events over Southeast South America: A Case Study Approach

FEIJOO MARTIN; SILVINA A SOLMAN

BEIJING

Conferencia; INTERNATIONAL CONFERENCE ON REGIONAL CLIMATE ICRC2019-CORDEX ABSTRACT BOOK; 2019

WORLD CLIMATE RESEARCH PROGRAMME

Southeastern South America (SESA) is one of the regions in the world most affected by extreme mesoscale convective systems and their associated extreme precipitation. Many authors have studied the mesoscale forcings that trigger convection on the region, being the Andes mountain range and the South American Low Level Jet key factors in the development of convective storms. The small spatial scale of these systems and the fact that the intensity of extreme precipitation has been increasing over SESA due to global warming during the last decade, reveals the importance of understanding how these systems may change in the future. Several studies show that high resolution convective permitting Regional Climate Models (RCMs) improve the representation of the diurnal cycle and intensity of precipitation associated with deep convection, compared to coarser convection parameterized RCMs. In this context, the aim of this study is to assess the capability of the WRF-ARW at convective permitting (CP) resolution in representing the precipitation associated with these events over SESA, compared with RCMs, in two ways. First, in order to determine if the CP improves the results of RCMs and second, if increased spatial resolution in a CP model improves the results, taking into account the computational cost of high resolution simulations. Three extreme precipitation events have been selected over the region and three simulations have been carried out, one with parameterized convection at 12 km spatial resolution, and two simulations at 4 km and 2.4 km in which convection was not parameterized but resolved explicitly. The simulated precipitation was compared against CMORPH satellite estimations and the ability of the simulations was measured using the Fractional Skill Score. The results show that CP simulations have better agreement with the observations (diurnal cycle and intensity) than RCM simulations. No significant difference was found between the 4 km and 2.4 km simulations.