Towards an efficient storm surge and inundation forecasting system over the Bengal delta: chasing the Supercyclone Amphan

Abstract

Abstract. The Bay of Bengal is a well-known breeding ground to some of the deadliest cyclones in history. Despite recent advancements, the complex morphology and hydrodynamics of this large delta and the associated modelling complexity impede accurate storm surge forecasting in this highly vulnerable region. Here we present a proof of concept of a physically consistent and computationally efficient storm surge forecasting system tractable in real time with limited resources. With a state-of-the-art wave-coupled hydrodynamic numerical modelling system, we forecast the recent Supercyclone Amphan in real time. From the available observations, we assessed the quality of our modelling framework. We affirmed the evidence of the key ingredients needed for an efficient, real-time surge and inundation forecast along this active and complex coastal region. This article shows the proof of the maturity of our framework for operational implementation, which can particularly improve the quality of localized forecast for effective decision-making over the Bengal delta shorelines as well as over other similar cyclone-prone regions.

Storm surge hazard over Bengal delta: A probabilistic-deterministic modelling approach

“…The final mesh resolution ranges from 15km over the deep parts of the ocean to 250m over the delta, amounting to about 0.6 million mesh nodes, and 1.1 million triangular elements. A time step of 300 seconds was found suitable for the resolution of our model (Khan et al, 2021a).…”

“…The bottom friction is parameterized using Manning's n. We adopted an approach for the spatial distribution of the Manning coefficient similar to Krien et al (2016) and the one used in Khan et al (2021a) (Figure 2b). For the deeper part of the ocean (depth>20m), we used a Manning coefficient value of 0.02.…”

“…We modelled the wave breaking using the formulation of Battjes and Janssen (1978). A low value of the coefficient of breaking, α = 0.1 (instead of α = 1), is adopted to avoid over-dissipation over a very mild slope region like the Bengal delta (Pezerat et al, 2020(Pezerat et al, , 2021Khan et al, 2021a). We have discretized our modelled spectra into 12 directional and 12 frequency bins.…”

“…Given the wind and pressure fields are adequately prescribed, and necessary physics -particularly the wave coupling -is incorporated, the model is expected to simulate a storm surge appropriately. In Khan et al (2021a), the performance of the model in simulating storm surges is demonstrated through a hindcast of the surge generated by a recent cyclone -Amphan -that made landfall in this region. From a set of in-situ observations, it was found that the model performs well (5-10 cm peak water level error) in reproducing the maximum water level observed at multiple tide-gauge locations distributed throughout the Bengal shoreline.…”

“…Given the discussed context, the objective of this paper is to estimate the risk of storm surge and associated flooding across the Bengal delta using a high-resolution, efficient, and well-validated regional coupled storm-surge model developed based on a higher-accuracy regional bathymetry (Khan et al, 2021a). To this extent, we integrated the wave-coupled hydrodynamic model of Khan et al (2021a) for a large ensemble (~3600 cyclones, ~5000 years of storm activity) of synthetic cyclones generated through the statistical-deterministic method, covering the whole range of natural variability of storm frequency, size, intensity and track location, with a dense spatial distribution. The interactions among the tide, surge, and waves are modelled explicitly at high spatial resolution.…”