The WRF model as a tool to understand mesoscale processes over the poorly-sampled South American altiplano

Temporary observing networks were deployed over the South American Altiplano during the rainy season of 2002-2003, as part of the South American Low Level Jet Experiment (SALLJEX), with the aim of describing mesoscale circulations and rainfall patterns associated with lakes and dry “lakes” (salt flats). These networks included a 200-raingauge network in the vicinity of Lake Titicaca and temporary ~ 6-station pilot balloon networks that operated during two short field experiments. Although the data collected were useful to describe certain features of the mesoscale circulations and rainfall, the lack of additional observations in and around the region, together with the Altiplano's complex terrain, suggested the use of a mesoscale model would be required to improve the understanding of many aspects of the research.

The Weather Research and Forecasting (WRF) model was chosen for use in this study, based on our computer capabilities, ample documentation and support, model flexibility, and future trends. The main goal of the numerical simulations was to describe the nocturnal lake-effect storms that occur over Lake Titicaca and associated mesoscale features. Additional interests included understanding other mesoscale processes such as the advection of moisture into the Altiplano through the eastern canyons and overall terrain-induced circulations related to the diurnal cycle.

Several test simulations were initially carried out using different domains and resolutions. The FNL analyses from the Aviation Model were used as initial and boundary conditions. Initially the WRF model, with a grid spacing of 10km, was able to resolve the late afternoon maximum of convection over the Altiplano associated with the diurnal cycle. The Kain-Fritsch convective parameterization scheme and the Lin et. al. microphysics scheme were used. The nocturnal convection over Lake Titicaca was, however, not present in these simulations. An error of the specification of the lake's surface temperature was identified and corrected by setting it to a constant value of 13°C. Further test simulations indicated that a grid spacing of 1 km was necessary to reproduce the lake breezes and therefore the nocturnal lake-effect storms induced by Lake Titicaca. These and additional results, as well as comparisons with observations, will be presented on a poster. The implications of the use of a mesoscale numerical model such as the WRF to simulate the mesoscale climate in poorly sampled regions will be discussed as well.