Temporal analysis trend for daily maximum rainfall in the State of São Paulo

Abstract

The research related to trends and variability of climatic elements has acquired importance in discussions of
environmental changes. In order to contribute to studies on the possibility of changes in rainfall in the State of São Paulo, this
study aimed to provide new information on possible trends in rainfall. For that, rainfall data from 45 cities located throughout the
state were used, with 30 years of climatological normals (1977-2006). For each station, the maximum daily precipitation (Pdma)
was identified and adjustments were performed through the Gumbel probability distribution. Data were tested for adherence with
Anderson-Darling test (α<0.05) and associated with the return times of 2, 5, 10 and 100 year. The historical series of Pdma (30
years) of each period were divided into 4 sub-periods of the same duration (15 years), aiming to detect changes of Pdma. Three
complementary statistical techniques were adapted to evaluate the possible non-stationary behavior of probable Pdma. Initially,
two non-parametric tests were used: Kruskal-Wallis and Friedman: these were used to analyse the set of Pdma. The trend test of
Mann-Kendall was used to evaluate the existence of monotonous trend in each station alone. Also, geostatistical techniques were
used, in order to study the temporal and spatial variability of Pdma. From the 45 cities studied, 15 cities showed a positive trend of
Pdma in at least one of the sub-periods, and in 5 of those cities. The tendency was seen in the total time series (30 years), the city of
São Paulo showed positive trends in the overall series and two sub-periods.