Non-consistency hydrological drought dynamic risk assessment model based on Vine Copula

Hydrological drought poses significant threats to water security and agricultural sustainability in arid and semi-arid regions, particularly in oasis agricultural areas where water resources are already scarce. The Hotan River Basin, located in the southern Xinjiang Uygur Autonomous Region of China, represents a typical oasis agricultural region where agriculture is highly dependent on limited surface water resources for irrigation. However, intensifying climate change and expanding human water demands have led to non-consistent characteristics in the basin's hydrological regime, challenging the traditional consistency assumption underlying hydrological drought analysis. While considerable research has been devoted to developing non-consistent drought indices and analyzing drought characteristics under changing conditions, the development of dynamic risk assessment frameworks that can effectively capture evolving drought risks remains limited. This gap is particularly critical for regions like the Hotan River Basin, where accurate drought risk assessment is essential for agricultural planning and water resource management. This study addresses this research gap by developing a comprehensive dynamic risk assessment framework for non-consistent hydrological drought in the Hotan River Basin. The research begins by examining whether the basin's runoff series exhibits non-consistent characteristics through rigorous statistical testing, providing empirical evidence of the impacts of environmental changes on hydrological processes. Building upon this foundation, the study employs the Generalized Additive Models for Location, Scale and Shape (GAMLSS) framework to construct a non-consistent standardized runoff index (NSRI) that explicitly incorporates the time-varying influences of climatic factors and anthropogenic activities through distribution parameters. This approach overcomes the limitations of traditional consistent drought indices. To validate the superiority of this approach, NSRI is systematically compared against SRI through drought characteristic analysis and historical drought event validation, demonstrating better performance in capturing drought severity and agreement with actual disaster occurrences. A key innovation of this research lies in the development of a dynamic risk assessment model based on Vine Copula methodology that integrates multiple drought characteristics with their occurrence probabilities under non-consistent conditions. The Vine Copula approach decomposes high-dimensional joint distributions into a series of conditional bivariate copulas, effectively reducing the complexity of parameter estimation while capturing the complex dependencies among different drought attributes. This approach enables the quantification of drought risk as a continuous, time-varying metric rather than a static value, providing more actionable information for decision-makers. The model's performance is rigorously validated using well-documented historical drought events, demonstrating its capability to provide reliable early warnings. The findings reveal several important insights. First, the basin's hydrological regime has indeed undergone significant changes, with runoff series exhibiting clear non-consistent characteristics attributable to both climatic shifts and human interventions. Second, NSRI demonstrates superior performance in capturing drought events compared to SRI, showing better agreement with documented drought occurrences and providing more accurate characterization of drought severity. Third, the risk assessment results indicate that the basin currently faces moderate drought risk levels, with both major tributaries showing similar risk profiles. The validation against typical drought events in 1991-1992 confirms the model's effectiveness in identifying high-risk periods and providing timely warnings. This research makes important contributions to both scientific understanding and practical applications. Methodologically, it advances the field of non-consistent hydrological drought analysis by integrating GAMLSS-based index development with Vine Copula-based dynamic risk assessment in a unified framework. The proposed approach can be readily adapted to other regions facing similar challenges of changing hydrological conditions. From a practical perspective, the developed tools provide valuable support for drought monitoring, agricultural irrigation scheduling, and risk management in oasis agricultural regions. The findings underscore the importance of accounting for non-consistency in hydrological drought analysis and offer a robust framework for enhancing drought resilience in water-stressed regions under changing environmental conditions.