Pronóstico de lluvia utilizando métodos estadísticos en la Cuenca del Río Paute”

Abstract

The need to increase the knowledge about rainfall models to contribute to the understanding of the hydrologic cycle, the possibility of applying statistical methods in the Paute river basin is verified. We was able to understand and observed that the climatic variability of precipitation in a region is strongly influenced by variations in sea surface temperatures (TSM) of the oceans surrounding it. Future climatic conditions are not easy to predict accurately; an approximation it would be very useful for the better understanding of precipitation patterns and forecasting of the future conditions to which a place could be exposed. This will allow analysis of the risks posed by precipitation extremes (droughts and floods) and design prevention and mitigation actions to reduce impacts on the population. Main objective of this project was development a rainfall forecast using statistical models and predict the seasonal variation in a probabilistic way on Paute basing. For this reason a canonical correlation analysis was used, and validation of application of this models was carried out by means of skill scores, through by comparison of observed and modeled values. Results shows the efficiency of this statistical method to identify associations, through a multivariate approach, where standardized index of precipitation act as predicating. Correlations and canonical charges founded, indicate a correspondence between the variables and the temperature of the sea surface that acts as a predictor in period 1981-2010.