An efficient modelling approach for probabilistic assessments of present-day and future fluvial flooding

Abstract

Abstract. Flood risk management and planning decisions in many parts of the world have historically utilised flood hazard or risk maps for a very limited number of hazard scenarios (e.g. river water levels), mainly due to computational challenges. With the potentially massive increase in flood risk in future due to the combination of climate change effects (increasing the hazard) and increasing population and developments in floodplains (increasing the consequence), risk-informed flood risk management, which enables balancing the risk with the reward, is now becoming more and more sought after. This requires a comprehensive and quantitative risk assessment, which in turn demands multiple (thousands of) river and flood model simulations. Performing such a large number of model simulations is a challenge, especially for large, complex river systems (e.g. Mekong) due to the associated computational and resource demands. This article presents an efficient modelling approach that combines a simplified 1D hydrodynamic model for the entire Mekong Delta with a detailed 1D/2D coupled model and demonstrates its application at Can Tho city in the Mekong Delta. Probabilistic flood hazard maps, ranging from 0.5 yr to 100 yr return period events, are obtained for the urban centre of Can Tho city under different future scenarios taking into account the impact of climate change forcing (river flow, sea-level rise, storm surge) and land subsidence. Results obtained under present conditions show that more than 12 % of the study area is inundated by the present-day 100 yr return period water level. Future projections show that, if the present rate of land subsistence continues, by 2050 (under both RCP4.5 and RCP8.5 climate scenarios), the 0.5 yr and 100 yr return period flood extents will increase by around 15-fold and 8-fold, respectively, relative to the present-day flood extent. However, without land subsidence, the projected increases in the 0.5 yr and 100 yr return period flood extents by 2050 (under RCP4.5 and RCP8.5) are limited to between a doubling to tripling of the present-day flood extent. Therefore, adaptation measures that can reduce the rate of land subsidence (e.g. limiting groundwater extraction), would substantially mitigate future flood hazards in the study area. A combination of restricted groundwater extraction and the construction of a new and more efficient urban drainage network would facilitate even further reductions in the flood hazard. The projected 15-fold increase in flood extent projected by 2050 for the twice per year (0.5 yr return period) flood event implies that the do nothing management approach is not a feasible option for Can Tho.

Instant Flood Risk Modelling (Inform) Tool for Co-Design of Flood Risk Management Strategies with Stakeholders in Can Tho City, Vietnam

“…Inform uses the modeled water levels at Can Tho to generate the flood hazard and compute the associated flood damages. Reliability of the 1D/2D coupled flood model results has been established by calibrating the model outputs against measured inundation depths and flood extent at the study area during historical flood events [39]. Additionally, the input data for Inform including historical flood events at Kratie and sea level data at the Mekong River mouths from 2000 to 2009 were collected from reliable sources (e.g., Mekong River Commission, National Hydro-meteorological Service of Viet Nam).…”

“…Inform uses the modeled water levels at Can Tho to generate the flood hazard and compute the associated flood damages. Reliability of the 1D/2D coupled flood model results has been established by calibrating the model outputs against measured inundation depths and flood extent at the study area during historical flood events [39].…”