Heavy rains in Mexico City

Abstract

Numerical Weather Prediction has become a fundamental tool in Civil Protection Institutions. Short-term numerical weather prediction for the Valley of Mexico has rarely been evaluated in a systematic way. By using daily observed precipitation data and those predicted with the mesoscale model known as MM5, an evaluation of rainfall forecast is made. It is found that making predictions of high spatial resolution in the Valley of Mexico is of limited quality mainly because of the effects of urbanization and orography over the rainfall. The lack of consistency between predicted and observed rainfall spatial patterns requires an analysis of stationary physical factors that can influence the quality of forecasts. Errors in short-term forecasts require risk management strategies to implement disaster prevention actions.