Probability Distribution Functions (PDFs) Selection to Rainfall Time Series from Brazilian Semiarid Cities
The Brazilian semiarid environment that has a rightly variable hydrologic behavior, and consequently is a climate change spot to all scenarios designed by IPCC. On this, the objective of this research was to verify the rainfall patterns and select the better distribution statistical adjustment inrainfall time series from semiarid of Pernambuco State, in a total of thirty analyzed cities, inside the Brazilian semiarid. Therefore, through the analysis of rainfall distribution in monthly and annual time series, the Probability Distribution Function (PDF), that had produced the better adjustment for the data set observed for most of cities was the Weibull (type 3) for the monthly data set, while in the annual time series the distribution that obtained the best adjustment to the data among those observed was the Logistics PDF, better adjusted to ten cities. The distribution Gama (type 2) Probability Distribution Function was better adjusted to six cities, and the GEV (Generalized Extreme Values) distribution showed good adherence in five of the thirty analyzed cities. The Log-Normal distribution adjusted well to four cities, the Fréchet distribution (Fisher -Tippett type 2) to three cities, and Weibull distribution (type 3) and Normal adjusted well just to one city each.
