The correct understanding of the weather variability is a fundamental step in reducing the agricultural climate risk. The aim of the study was to evaluate the temporal variability of monthly precipitation data from eight regions of the State of São Paulo, Brazil. Investigations related to possible climate trends were also held. Using the wavelet analysis, the likelihood ratio test, and the Mann-Kendall test, it was observed a very high temporal variability of the monthly precipitation data in the eight analyzed regions. The treatment of such series as strictly stationary, or the use of statistical models (such as Fourier spectral analysis), that only reveals what frequency (spectral) components exist in the precipitation signal, will lead to the loss of important information about the modulating forcing of the precipitation temporal variability. Despite the high temporal variability, no trends on the rainfall series were detected. In the sense of agrometeorological applications, the high temporal variability of the monthly precipitation signal should be considerate on the agricultural zoning model of the State of São Paulo, Brazil.
Non stationary series; climate trends; agricultural zoning
