Validation of Weather Research and Forecasting (WRF) Model Simulations of the South American Climate during the Austral Summer of 2003–2004

The meridional extent and complex orography of the South American continent contributes to a wide diversity of climate regimes ranging from hyper-arid deserts to tropical rainforests to sub-polar highland regions. The austral summer climate in tropical regions is characterized by the seasonal migration of the ITCZ, the development of the South Atlantic Convergence Zone (SACZ), and the convectively-driven formation of the Bolivian High and the Northwest Argentinian Low. Summer climate in mid-latitude regions is dominated by westerly winds and mid-latitude cyclones. Until recently, studies addressing South American climate have relied strongly upon coarse-resolution global climate models (GCMs) to obtain their results. However, recent increases in computation speed make it now possible to run regional climate simulations using prognostic mesoscale models such as the Weather Research and Forecasting (WRF) model.

Using WRF, this study investigates its ability to accurately detect, develop, and resolve key characteristics of the South American 2003-2004 austral summer climate (1 Dec 2003 – 1 Mar 2004). Evaluated model runs will include a base WRF model simulation and another which fully-couples WRF to the Regional Ocean Modelling System (ROMS). Model simulations were run for 91 days and are validated against both the Tropical Rainfall Measuring Mission (TRMM) 3B42 precipitation product and the European Centre for Medium-Range Forecasts (ECMWF) interim analysis. Preliminary results have shown WRF to reasonably simulate synoptic-scale features: The ITCZ migration, warming land temperatures, and the development of the SACZ, Bolivian High and the Northwest Argentinian Low. WRF model simulations typically generated precipitation in over the correct regions of South America, but the heaviest rainfall was often displaced in comparison to TRMM. Despite such precipitation placement errors, histograms averaged over the entire South American wet season revealed WRF to over-predict light precipitation and under-predict heavy precipitation, but its overall precipitation distribution compares well to TRMM.

Previous GCM-based studies have been shown to be too coarse to explicitly resolve the sharp climate zone gradients (especially between the Andes and adjacent regions) and use model parameterizations for all its processes (microphysics, radiation, cumulus, etc.). The present work still parameterizes all physical processes on the outer 27-km domain covering all of South America, but the inner 9-km domain covering the Central Tropical Andes and the Amazon explicitly resolves convection. As compared to GCMs, higher terrain resolution and non-parameterized convection will likely allow more realistic simulations of moisture transport from the lowlands into the Andes, generation of shallow, terrain-forced convection over the Altiplano, and the impact of mid-latitude intrusions into the tropics. Additionally, these simulations will potentially show better realization of the elevation-dependent precipitation distribution and potentially reduce temporal errors pertaining to the timing of diurnal convection over interior regions of the Amazon rainforest. Finally, the two WRF simulations will be compared to evaluate the merits of high-resolution atmosphere-ocean coupling to seasonal climate prediction.