Testing for Collective Statistical Significance in Climate Change Detection Studies

We examined several approaches to detecting statistical significance of trends defined on a grid, that is, on a regional scale. To this end, we introduced a novel simple procedure of significance testing based on counting signs of local trends (sign test), and comparing it with four other approaches to testing collective significance of trends (counting, extended Kendall, Walker, and FDR tests). Synthetic data were used to construct the null distributions of trend statistics and determine critical values of the tests. The application of the five tests to real datasets reveals that outcomes of the tests may differ even though trends are locally significant at the majority of the grid points.