Abstract
Climate changes are taking place and will probably be reinforced in the future. In such a context, statistical tests for detecting trends in time series of hydrologic observations are certainly important tools for building and improving prediction models and societal preparedness plans to deal with possible impacts from extreme events. This paper focuses on discussing a recent method proposed for detecting monotonous trends in time series by accounting for both type I and type II errors, and comparing its results with the conventional Mann-Kendall nonparametric test, as complemented by Sen slope estimator, following application to the three main series of annual maximum daily rainfall observed in the Brazilian state of Acre. The results point out a significant increasing tendency for the series observed at the raingauge of Tarauacá.
Keywords:
trend detection; type I and type II errors; maximum precipitation series