Henryk T. Mitosek

Abstract

The paper contains some remarks on detecting climate variability or climate change in river discharge processes. The selection of adequate non-parametric statistical tests is discussed as well as the conditions which should be fulfilled by data records of the considered processes. After statistical analysis had been carried out on the basis of more than 150 long-term river-discharge time series, using non-parametric tests at the 5% significance level, it was found that the stationary hypothesis could not be rejected in most cases. Moreover, no conclusion could be drawn as to an regional pattern of trends in runoff characteristics: in some cases, neighboring river catchments show trends in the opposite direction.

Keywords: climate, non-parametric tests, non-stationarity

Henryk T. Mitosek, Institute of Geophysics, Polish Academy of Sciences ul. Księcia Janusza 64, 01-452 Warszawa, Poland