Characterization and frequency analysis of long-term maximum rainfall from São Martinho, Santa Catarina, Brazil

Abstract

The study of the frequency of intense rainfall is important for agricultural and environmental planning and the dimensioning of drainage works. This study aimed to characterize and determine the relationships of intensity, duration, and frequency of long-term rainfall in São Martinho, Santa Catarina. The series of annual maximums lasting from one to ten days, observed in the period from 1977 to 2020, were determined. The probability distributions Gumbel, GEV, Log-Normal with two parameters, Log-Normal with three parameters, Pearson type III, and Log-Pearson Type III were evaluated. The parameters were estimated by the method of moments, maximum likelihood method, method of L-Moments, and for the Gumbel distribution, the Chow method was used. The Kolmogorov-Smirnov, Anderson-Darling, and Filliben fitting tests were used, and for the selection of the distribution, the standard error of the estimate was also used. The Log-Normal distribution with three parameters was selected for series lasting one, five and six days. The Pearson III distribution was selected for a duration of three and four days and the GEV distribution for the other five series. The adjusted IDF equation allows the estimation of rainfall intensity with duration from24 hours to 240 hours and a return period ranging from 2 to 100 years.