Stochastic Dependence and Risk Assessment of Compound Flood Drivers in the Ganges–Brahmaputra–Meghna Basin, Bangladesh

Climate-induced hydrometeorological extremes pose a persistent threat to the Ganges–Brahmaputra–Meghna (GBM) basin, where traditional univariate flood assessments systematically underestimate compound disaster risks. This study presents a transboundary risk approach that links basin-wide monsoon factors (GBM precipitation and soil moisture) with the seasonal flood footprint in the Bangladesh delta. While traditional assessments often focus on river discharge at specific gauges, this work provides an integrated assessment of the relationship between antecedent soil moisture and downstream inundation extent using a consistent 31-year satellite-derived record. Utilizing this longitudinal archive (1988–2021) of ERA5-Land drivers and JRC global surface water footprints, a bivariate copula framework was applied to characterize the non-linear dependency between monsoon precipitation, antecedent soil moisture, and seasonal flood extent. The analysis identified the Gumbel copula as the optimal model, revealing a significant upper-tail dependency (λu = 0.227) that confirmed a strong effect during extreme events. Hydrological attribution demonstrated that a saturated upstream basin acts as a significant antecedent factor, multiplying the probability of extreme flooding in the delta by a factor of 1.16 times compared to median conditions. Joint ‘AND/OR’ return period matrices established a probabilistic bivariate risk assessment model for compound disasters. These results provide a prospective yardstick for seasonal flood predictions and possible means of boosting support through transboundary basin-prior conditioning, enhancing the imminent likelihood of extreme inundation.