Cross correlations among estimators of shape

Abstract

The regional variability of shape parameters (such as k for the GEV distribution) maybe described by generalized least squares (GLS) regression models that allow shapeparameters to be estimated from basin characteristics recognizing the sampling uncertaintyin available shape estimators. Implementation of such GLS models requires estimates ofthe cross-site correlation of the shape parameter estimators for every pair of sites. MonteCarlo experiments provided the information needed to identify simple powerapproximations of the relationships between the cross correlation of estimators ofskewness g from [log] Pearson type 3 (P3) data and of the shape parameter k of bothgeneralized Pareto (GP) and generalized extreme value (GEV) distributions, as functionsof the intersite correlation of concurrent flows.