PRECIPITATION SPATIAL DISTRIBUTION ANALYSIS IN MOUNTAIN AREAS FOR HYDROLOGICAL MODELING. APPLICATION IN CENTRAL WEST ARGENTINA.

Abstract:

The Andes mountain plays a fundamental role in the generation of orographic precipitation in South America, being the snow the main source of water in semi-arid areas such as in central-western Argentina. However, for hydrological modeling and water management there is high uncertainty in spatial distribution of precipitation due to the scarcity of weather stations and regional studies. Additionally, high steep mountains and altitude range (about 6000 m) in the region complicate precipitation analysis because impose different meteorological conditions depending on altitude. In this study, a spatially distributed monthly precipitation database was generated using geostatistical techniques. Based on these monthly maps, a precipitation gradient was estimated for each season and for each altitudinal zone for hydrological modeling. Monthly precipitation maps with the best fit were obtained with cokriging method, mainly in winter, being the Nash–Sutcliffe model efficiency coefficient (NSE) 0.89 in this season. The interpolation with Kriging method showed a little less NSE (0.86) than cokriging. Simple and multiple regression were evaluated too, these present the worst results and fit, being the mean NSE in winter 0.35 and 0.40 respectively. An intermediate result was obtained through Radial Base Function, the best fits were obtained in summer (0.73) and winter (0.4) but considering this technique a good option to predict precipitation using few calibration parameters. Gradients obtained ranged from -0.3 mm/100 m in the upper zone in summer, to +13.6 mm/100 m in the lower zone in winter. These maps database and seasonal gradient will help to improve hydrological and water management models, therefore will provide insights for decision making on water resources management in the region.