Quantifying flood risks during monsoon and post-monsoon seasons: An integrated framework for resource-constrained coastal regions

• Published in: Journal of hydrology (Amsterdam) 2024-02, Vol.630, p.130683, Article 130683
• Main Authors: Thakur, Dev Anand, Mohanty, Mohit Prakash, Mishra, Ashok, Karmakar, Subhankar
• Format: Article
• Language: English
• Subjects: Bivariate copula; CHIRPS v2.0; Cyclonic rainfall; MIKE; Multi-hazard; Population exposure
• ISSN: 0022-1694; 1879-2707
• DOI: 10.1016/j.jhydrol.2024.130683

[Display omitted] •Monsoon and post-monsoon flood hazards are quantified independently.•CHIRPS v2.0 performs excellently in capturing extreme rainfalls and flood hazards.•Bivariate association reveals the hidden characteristics of multi-hazard flooding.•The univariate approach is likely to underestimate the population's exposure to flood risk. Global coastal regions are transforming into “flooding hot-spots” due to the multi-faceted confluence of several flood-drivers. Unfortunately, these regions are often poorly gauged, hindering comprehensive flood hazard quantification. Besides summer-monsoon floods, post-monsoon rainfalls induced by cyclonic-depressions have escalated the flood risks in the coastal regions. For the first time, the present study captures the seasonal flood hazards by disentangling monsoon (June, July, August, and September) and post-monsoon (October, November, and December) periods of rainfall. An integrated MIKE + 1D-2D coupled model was used to generate flood hazard risks for Jagatsinghpur district (Odisha, India), a resource-constrained multi-hazard catchment in the Lower Mahanadi River Basin. Due to limited data for this coastal region, we investigated the performance of CHIRPS v2.0, a global Satellite Precipitation Product (SPP) with high spatiotemporal characteristics, for capturing extreme rainfall events and deriving flood hazards at the inundation scale. We observed higher magnitude of flood hazards are prominent in the coastal villages, emerging from a complex interaction between rainfall and storm-tide events. Three widely used Archimedean copulas were implemented to capture the rainfall and storm-tide events' joint (bivariate) behavior and their implications on population exposure, which is valuable for mapping flood hazard risks. We observed that univariate analysis (Ua) significantly underestimates the flood hazard over bivariate analysis (Ba) during both periods. Interestingly, the results are more pronounced in the population exposure estimates as ∼47 % (monsoon) and ∼54 % (post-monsoon) of the population are exposed based on univariate analysis against ∼65 % (monsoon) and ∼60 % (post-monsoon) under bivariate analysis. Our study highlights the necessity of disentangling rainfall time-series for foreseeing under-shadowed flood risks, vital for designing flood combating structures, evacuation plans, land use planning, and flood governance. The results further enhance our understanding of monsoon and post-monsoon flood haz