Advanced drought analysis using a novel copula-based multivariate index: a case study of the Ceyhan River Basin

Drought is a severe natural disaster that poses significant risks to both social and ecological systems. Detecting drought is challenging due to its gradual development, which makes it difficult to identify and predict, often resulting in significant impacts on the affected regions. Therefore, accurate and dependable monitoring of drought conditions is essential for the development and implementation of effective mitigation strategies. Drought indices play a crucial role in monitoring drought conditions, with single-variable indices commonly employed in the literature to evaluate drought severity. While these indices are typically effective at characterizing the specific type of drought for which they were designed, they often fall short in offering a comprehensive view of overall drought conditions. The multivariate standardized drought index (MSDI) is a comprehensive tool that assesses drought conditions by integrating multiple hydrometeorological variables. Widely employed in the literature in both parametric and empirical forms, the MSDI is recognized for its effectiveness in detecting drought in an integrated manner. This study focuses on a particular challenge related to the calculation of MSDI using copula families. The novel methodology introduced in this paper involves selecting the most suitable copula family for each data subset using AIC and BIC criteria. Rather than applying a single copula family to the entire dataset, this approach utilizes multiple copula families for different subsets, thereby ensuring optimal modeling for each distinct group of data. The Ceyhan River Basin (CRB) is used as a case study to apply the proposed methodology. The drought characteristics of the basin are analyzed using both the newly developed MSDI and conventional single-variable indices, and the performance of the new methodology is evaluated. The application of this approach in the CRB demonstrated its effectiveness in identifying both concurrent and isolated occurrences of meteorological and hydrological droughts, thereby facilitating a more integrated and precise assessment of drought characteristics. Results indicated that the proposed MSDI detected drought events that were overlooked by single-variable indices and improved classification accuracy over the conventional MSDI.