Selection of Nonparametric Methods for Monotonic Trend Detection in Water Quality

In this paper two nonparametric tests for monotonic trend detection are compared with respect to their power and accuracy. The importance of comparing powers at equal empirical signiicance levels rather than nominal levels, is stressed. Therefore, an appropriate graphical method is presented. The eeect of the sampling frequency is also assessed using a trajectory representation that visualizes the dynamics of the trade-oo between the type I and type II errors. These methods are applied to two nonparametrical tests (seasonal Mann-Kendall and modiied seasonal Mann-Kendall) under several conditions. It is concluded for the modiied seasonal Mann-Kendall test applied to the AR(2) process considered in this paper, that it is not worthwhile with respect to the power, to increase the sampling frequency from monthly to biweekly for detecting a monotonic trend of 10% of the process variance. The seasonal Mann-Kendall test is under these conditions, highly liberal. This research is situated in the development of an eecient sampling design for the Flemish water quality monitoring network.