Assessing the performance of parametric and non-parametric tests for trend detection in partial duration time series

Abstract

The detection of nonstationarities in partial duration time series (PDS) depends on several factors, including the length of the time series, the selected statistical test, and the heaviness of the tail of the distribution. Because of the more limited attention received in the literature when compared to the trend detection on block maxima variables, we perform a Monte Carlo simulation study to evaluate the performance of different approaches: Spearman's rho, Mann–Kendall, ordinary least squares (OLS), Sen's slope estimator (SEN), and the nonstationary generalized Pareto distribution fit to identify the presence of trends in PDS records characterized by different sample sizes (n), shape parameter (ξ) and degrees of nonstationarity. The results point to a power gain for all tests by increasing n and the degree of nonstationarity and by reducing ξ. The use of a nonparametric test is recommended in samples with a high positive skew. Furthermore, the use of sampling rates greater than one to increase the PDS sample size is encouraged, especially when dealing with small records. The use of SEN to estimate the magnitude of a trend is preferable over OLS due to its slightly smaller probability of occurrence of type S error when ξ is positive.