Validation of a computationally efficient 2D inundation model on multiple scales

Sam Robert Jamieson, Julien Lhomme, Grant Wright, Ben Gouldby

    Research output: Contribution to conferencePaperpeer-review

    Abstract

    RFSM-EDA (Rapid Flood Spreading Method – Explicit Diffusion wave with Acceleration term) is a 2D inundation model that has recently been developed for the purpose of making broad-scale inundation predictions, using a coarse grid-resolution whilst still maintaining a fine topographic representation. This enables it to run the type of rapid, yet accurate, simulations required for region-wide mapping and proba-bilistic predictions. This is made possible by using an innovative sub-grid Impact Zone schematisation and modifying a new approximation to the shallow water equations. RFSM-EDA has previously performed well on the UK Environment Agency’s (EA) 2D hydraulic benchmarking tests, with a significant computational speed advantage in most cases. This paper introduces a new approach to calculating the interface fluxes, and investigates the impact on simulation runtime and accuracy. The model is again tested on one of the bench-mark tests before being applied to a more appropriate large scale site, where it is compared with an industry standard 2D inundation model with an equivalent computational mesh resolution. The new interface flux me-thod is found to substantially improve efficiency without compromising predictive accuracy, and RFSM-EDA is shown to be significantly more accurate and computationally efficient than the comparison model.
    Original languageEnglish
    Publication statusPublished - 2012
    EventFloodrisk 2012 - Rotterdam, Netherlands
    Duration: 20 Nov 201222 Nov 2012

    Conference

    ConferenceFloodrisk 2012
    Country/TerritoryNetherlands
    CityRotterdam
    Period20/11/1222/11/12

    Fingerprint

    Dive into the research topics of 'Validation of a computationally efficient 2D inundation model on multiple scales'. Together they form a unique fingerprint.

    Cite this