Análise das estimativas de chuva, para Santana do Ipanema e Palmeira dos Índios (AL) através da distribuição Gamma e da média.

Abstract

Based on 55 years of data from the meteorological stations of Santana do Ipanema and Palmeira dos Índios, an analysis was made on the use of the incomplete gamma distribution and the average rainfall in rainfall estimates. A computer program was developed to calculate the amount of monthly and annual rainfall associated with 7 levels of probability, using a sample of 30 years, and compare the results with the sequence of the following 25 years. The standard deviation for the months and year was calculated. The verification of the expected and observed probabilities was carried out using the chi-square test. With the exception of the month of May, for Palmeira dos Índios, the gamma distribution presented results that were consistent with those observed for the rainy season. As for the period of low rainfall, the expected probabilities* differed from those observed. This discrepancy was caused by the size of the sample, which did not represent the population. Average monthly and annual precipitation, based on 30 years of data, should not be considered as an index for agricultural planning. Finally, in precipitation analysis, sample size is as important as the type of distribution to be used.