2025-08-07 THEORETICAL AND APPLIED CLIMATOLOGY 2025 156(卷), 8(期), (null页)
In recent years, copula-based multivariate drought analysis has garnered significant attention. However, the uncertainties associated with input data and copula parameter estimation remain insufficiently addressed This study aims to evaluate these uncertainties using a bootstrap approach, focusing on Severity-Duration-Frequency (SDF) curves for hydrometeorological droughts through Maximum Entropy Copula (ME-copula), Empirical Copula (EM-copula), and Theoretical Copula (TH-copula) methods at various conditional probabilities (CP). The results indicate that drought severity decreases when CP is reduced or the conditional return period (Tr) increases, while at a constant CP, severity intensifies with longer drought durations. The three methods show minor differences in estimated CPs, with ME-copula aligning closely with EM-copula. Notably, uncertainty from input data is significantly higher than copula parameters, with the order of uncertainty being TH-copula > ME-copula > EM-copula. The study demonstrates that ME-copula provides a robust framework for modeling the joint distribution of drought severity and duration, aiding in uncertainty reduction and enhancing drought preparedness strategies.
关键词