Copula‐Based Convection‐Permitting Projections of Future Changes in Multivariate Drought Characteristics

Abstract

Probabilistic projections of future drought characteristics play a crucial role in climate change adaptation and disaster risk reduction. This study presents a copula‐based probabilistic framework for projecting future changes in multivariable drought characteristics through convection‐permitting Weather Research and Forecasting simulations with 4‐km horizontal grid spacing. A probabilistic multivariate drought index is introduced to examine the joint effects of drought indicators with uncertainty intervals for four major river basins located in South Central Texas of the United States. Markov chain Monte Carlo is used to address uncertainties in assessing copula parameters and in predicting climate‐induced changes in hydrological regimes. Our findings reveal that the severity and intensity of drought episodes can be amplified when considering the compound effects of soil moisture and runoff regimes by using the probabilistic multivariate drought index. The South Central Texas region is projected to experience more drought events with shorter duration and higher intensity in a changing climate. The drought severity will not necessarily increase due to the decreasing drought duration. In addition, our findings indicate that the intensity of future droughts is expected to increase as a result of the deficiency of soil moisture even though precipitation extremes are projected to become more frequent. Moreover, climate change impacts on multivariate drought characteristics will intensify with the increasing temporal scales (i.e., short‐, medium‐, and long‐term droughts) although the number of future drought events may decrease by the end of this century.

Multi-Site Statistical Downscaling Method Using GCM-Based Monthly Data for Daily Precipitation Generation

“…It can accurately simulate the characteristics of regional climate variables. Some studies show that he grid-based precipitation data generated through dynamical downscaling is able to better characterize the spatial pattern of precipitation and convection processes of a catchment scale [67,68]. However, the computational cost of dynamical downscaling is relatively high, and, due to the inherited GCM bias, as well as the bias caused by its own incomplete understanding of the climate system, the results cannot be directly used to assess the runoff response at the watershed scale or site scale [69].…”

A two-stage Bayesian multi-model framework for improving multidimensional drought risk projections over China

“…Copulas are multivariate cumulative distribution functions that enable us to link the marginal distributions of multiple random variables together to form the joint distribution (Genest et al, 2007;Zhang et al, 2019). The dependence of drought duration and severity detected by the SPEI6 over each of the 10 climate divisions in China (see Fig.…”

“…For example, Xu et al (2015) considered the spatial extent of droughts in the copula-based multivariate drought frequency analysis in Southwest China. Zhang et al (2019) used copula and the convection-permitting climate simulations to assess climate change impacts on the multivariate drought evolution over South Central Texas. One of the most important variables derived by multivariate drought frequency analysis is the drought return period, which represents the average time between drought episodes and thus quantifies drought risks (Kwon and Lall, 2016;Masud et al, 2015).…”

Probabilistic Projections of Hydrological Droughts Through Convection‐Permitting Climate Simulations and Multimodel Hydrological Predictions

“…In recent years, many studies have been devoted to the prediction of the future hydrological drought caused by the human-induced climate change (Liu et al, 2012;Oguntunde et al, 2018;Yu et al, 2015;Yuan et al, 2017;B. Zhang et al, 2019).…”