Detecting Changes in the Mean of Hydrologic Time Series
David T. Marshall · 1994
This paper presents a practical method for determining the power of a subjective statistical test for the stationarity or otherwise of time series data. It compares the power of the commonly used double mass plot with that of a parametric test known as the 'bivariate test'. It demonstrates that the double mass plot is applied inconsistently by different people and is a low powered test when compared with the bivariate test.