Uso de metodologias de preenchimento de falhas para estimativas de dados de precipitação

In the following paper, gap filling methodologies of historical series of rainfall data were applied, not as an auxiliary tool to fill gaps provoked by mistakes, but as a way to fully predict rainfall data for the city of Catanduva, based on data of the cities of Sao Carlos (SC), Franca (FR) e Votuporanga (VT), located in Sao Paulo state. Aiming to analyze and evaluate the viability of using the methods of regional weighting, simple linear regression e multiple linear regression, verifying the performance and correlation between estimated and real data (registered by Catanduva’s pluviometric station). There were used historical series information with 30 years of monthly observations provided by the Instituto Nacional de Meteorologia (INMET) to apply the methods mentioned. Comparing the estimated with the measured data, it is possible to affirm that the method of multiple linear regression was the one that best simulated the reality of the observed events and the estimations have a better performance for driest months and when there are used rainfall data from nearby regions.