Understanding drought propagation through coupling spatiotemporal features using vine copulas: A compound drought perspective
Guibin Yang,Jianxia Chang,Yimin Wang,Aijun Guo,Lu Zhang,Kai Zhou,Zhenwei Wang
DOI: https://doi.org/10.1016/j.scitotenv.2024.171080
2024-04-01
Abstract:Accurately evaluating drought impact on agriculture poses a challenge to regional food security, particularly in compound drought (i.e., meteorological and agricultural drought co-occurring) scenarios. This study presents a novel approach utilizing Vine copula for coupling spatiotemporal features to evaluate drought propagation. Three-dimensional clustering method was employed to identify meteorological and agricultural drought events, which excelled in capturing dynamic evolution characteristics (duration, area, severity, etc.) as well as integrating them into comprehensive meteorological drought intensity (I<sub>MD</sub>) and agricultural drought intensity (I<sub>AD</sub>). Through spatiotemporal matching, compound drought events were extracted from the meteorological-agricultural drought event pairs. From compound drought perspective, compound duration (CD) and compound area (CA) were devised to characterize drought propagation potential across time and space. Finally, the Vine copula method was employed to model the interdependence between four key coupling features, namely I<sub>MD</sub>, I<sub>AD</sub>, CD, and CA, and evaluate the probability of triggering agricultural drought with different intensity levels. Results showed that CD and CA can respectively characterize the temporal and spatial accumulation scale of drought propagation. At a certain I<sub>MD</sub> level, CD significantly influences the propagation probability (i.e., "stratification" phenomenon), while CA increases the probability proportionally. Probability evaluation lacking spatiotemporal information may underestimate the likelihood of drought propagation characterized by "low-I<sub>MD</sub>" but "long-CD" or "large-CA". The four-dimensional Vine copula structure can effectively couple dependence relationships of compound drought characteristics, and exhibits reliable robustness. This research provides stakeholders accurate probabilistic evaluation under compound drought scenarios, offering new insight into drought propagation.
environmental sciences