Avaliação de estimadores equivariantes e de máxima verossimilhança em eventos de precipitação máxima

Abstract

A climatic variable whose behavior is of great interest is rainfall, since various economical activities and environmental processes are highly dependent on it. The aim in the present work: i) check the t of the exponential distribution to maximum rainfall data in Piracicaba-SP, in the months from October to March, ii) calculate the expected maximum precipitation for return periods of 10, 30, 50, 70 and 100 years in these months, using equivariants and maximum likelihood estimators, iii) to compare the exponential distribution adjusted using equivariant estimators (ExponentialE) with the one adjusted using maximum Likelihood (ExponentialML) in order to select, in each month, the more accurate results. The data of maximum precipitation (1917-2015) were obtained from the agrometeorological conventional station of the Escola Superior de Agricultura \Luiz de Queiroz" (ESALQ / USP). The ExponentialE and ExponentialML distributions ts data every month. The exponentialE distribution was more suitable to model the maximum rainfall data in November and December. In the remaining months, the distribution ExponentialML, performed better.