Copula-based multivariate analysis of hydrological drought over jiabharali sub-basin of Brahmaputra River, India

In this study, copula-based multivariate hydrological drought analysis was carried out in the Jiabharali (Kameng) River in Arunachal Pradesh, a sub-tributary of Brahmaputra River, India. Different drought characteristics - severity (S), duration (D), and inter-arrival (I) time were estimated, to evaluate their joint probability drought occurrences. The multi-time streamflow drought indices (3, 6, 9, and 12-months) were calculated by using total monthly discharge of Bhalukpong station, Jiabharali River from 2000 to 2023. The highest value of hydrological drought severity was observed at a longer time scale in the SDIn12 (SDIn9) months of magnitude - 87.75 (-83.8) that lasted for 71 (63) months. Different marginal probability distribution functions (PDFs) and copula families (Elliptical and Archimedean) were used to examine the joint and conditional probability return periods between each drought variables. The correlation analysis revealed that the SD pairs are the most suitable for joint drought probability. The log-normal PDF was observed to be the best-fit distribution for the S and D in the marginal fitting, and the best reliable copula for capturing the symmetric (heavy) tailed dependencies were found to be the normal (gumbel) copula particularly for the SDIn3, SDIn6 and SDIn9-months (SDIn12-months) between the SD. The multivariate joint return periods analysis revealed that both [Formula: see text] and [Formula: see text] conditional events increased with the increasing time scale and the conditional probability showed strong likelihood in longer time scale, indicating extreme joint drought events, emphasizing the importance of incorporating joint probability assessments in the drought risk planning.