Abstract

A number of studies have predicted more frequent and intensive storms as a result of climate change in UK and other parts of the world, which may consequently cause more hazardous flash floods in steep catchments of small to medium size (e.g. up to 100km2). These flash floods are commonly characterized by high-velocity overland flow as a result of rapid catchment response to the intense rainfall. The hydrological processes related to the rapid catchment response are poorly understood and reliable prediction is generally beyond the capability of traditional hydrological models or simplified hydrodynamic models. This work aims to present a shock-capturing hydrodynamic modelling system to simulate the complex rainfall-runoff and the subsequent flash surface flooding process in a rapid-response catchment. The model solves the fully 2D shallow water equations using a finite volume Godunov-type shock-capturing numerical scheme for the rapidly varying overland flow hydrodynamics following intense rainfall. Typically, this type of shock-capturing hydrodynamic models is not able to provide efficient and high-resolution simulations for large-scale flash flood events due to their high computational demand. In order to substantially improve the computational efficiency and enable catchment-scale simulations at very high resolution involving millions of computational nodes, the model is implemented on GPUs for high-performance parallel computing. After being validated against the analytical benchmark of Tilted V-catchment test, the GPU-accelerated hydrodynamic modelling system is applied to simulate the rainfall-runoff process in the 42km2 Haltwhistle Burn Catchment in England.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call