Improved spatial representation of precipitation and air surface temperature over highlands of the southern tropical Andes (Lake Titicaca region) during an austral summer using the WRF model

Abstract

Due to its complex topography, the Lake Titicaca region, located in the southern tropical Andes, presents great challenges for atmospheric modeling. This study aims to improve the representation of precipitation and air surface temperature using the Weather Research and Forecasting (WRF) model at high spatial resolution (2 km), during the austral summer of 2020. We conducted 11 experiments with different configurations of topography, land use, physical parameterizations, and lake surface temperature (LST). Each experiment was evaluated considering in-situ data from the Peruvian-Bolivian region and gridded precipitation products. For precipitation, the best configuration, with an average bias close to zero mm, includes using the GMTED2010 topography (not smoothed) and the land use data of Eva et al. (2004), along with the Purdue Lin microphysics and the Grell 3D cumulus scheme. For air temperature, the best configuration, which showed an average underestimation between 0 and − 0.5 °C, included the same topography and land use, along with the parameterization of the SENAMHI Operational Model (SOM), including the WRF Single Moment 3 microphysics and the Kain-Fritsch cumulus scheme. In the last experiment, the sea surface temperature (SST) was updated, resulting in an average LST increase of +1.8 °C over Lake Titicaca. This resulted in an increase in the precipitation bias (82.2 %) due to increased evaporation and convection over the lake and decreased southwestward moisture transport. These results highlight the sensitivity of the WRF model to parameterization choices and SST forcing data, emphasizing the importance of any changes in these variables.