No Cover Image

Journal article 688 views 62 downloads

Coupled hydro-meteorological modelling on a HPC platform for high-resolution extreme weather impact study

Dehua Zhu, Shirley Echendu, Yunqing Xuan Orcid Logo, Michael Webster Orcid Logo, Ian Cluckie

Hydrology and Earth System Sciences, Volume: 20, Issue: 12, Pages: 4707 - 4715

Swansea University Authors: Dehua Zhu , Shirley Echendu , Yunqing Xuan Orcid Logo, Michael Webster Orcid Logo, Ian Cluckie

  • zhu2016.pdf

    PDF | Version of Record

    Distributed under the terms of a Creative Commons Attribution 4.0 (CC-BY) Licence.

    Download (2.23MB)

Abstract

Impact-focused studies of extreme weather require coupling of accurate simulations of weather and climate systems and impact-measuring hydrological models which themselves demand larger computer resources. In this paper, we present a preliminary analysis of a high-performance computing (HPC)-based h...

Full description

Published in: Hydrology and Earth System Sciences
ISSN: 1607-7938 1607-7938
Published: 2016
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa31336
Tags: Add Tag
No Tags, Be the first to tag this record!
Abstract: Impact-focused studies of extreme weather require coupling of accurate simulations of weather and climate systems and impact-measuring hydrological models which themselves demand larger computer resources. In this paper, we present a preliminary analysis of a high-performance computing (HPC)-based hydrological modelling approach, which is aimed at utilizing and maximizing HPC power resources, to support the study on extreme weather impact due to climate change. Here, four case studies are presented through implementation on the HPC Wales platform of the UK mesoscale meteorological Unified Model (UM) with high-resolution simulation suite UKV, alongside a Linux-based hydrological model, Hydrological Predictions for the Environment (HYPE). The results of this study suggest that the coupled hydro-meteorological model was still able to capture the major flood peaks, compared with the conventional gauge- or radar-driving forecast, but with the added value of much extended forecast lead time. The high-resolution rainfall estimation produced by the UKV performs similarly to that of radar rainfall products in the first 2–3 days of tested flood events, but the uncertainties particularly increased as the forecast horizon goes beyond 3 days. This study takes a step forward to identify how the online mode approach can be used, where both numerical weather prediction and the hydrological model are executed, either simultaneously or on the same hardware infrastructures, so that more effective interaction and communication can be achieved and maintained between the models. But the concluding comments are that running the entire system on a reasonably powerful HPC platform does not yet allow for real-time simulations, even without the most complex and demanding data simulation part.
College: College of Engineering
Issue: 12
Start Page: 4707
End Page: 4715