A Distributed Computing Workflow for Modelling Environmental Flows in Complex Terrain

Stuart R. Mead, Mahesh Prakash, Christina Magill, Matt Bolger, Jean Claude Thouret

Research output: Contribution to journalConference paperpeer-review

3 Citations (Scopus)

Abstract

Numerical modelling of extreme environmental flows such as flash floods, avalanches and mudflows can be used to understand fundamental processes, predict outcomes and assess the loss potential of future events. These extreme flows can produce complicated and dynamic free surfaces as a result of interactions with the terrain and built environment. In order to resolve these features that may affect flows, high resolution, accurate terrain models are required. However, terrain models can be difficult and costly to acquire, and often lack detail of important flow steering structures such as bridges or debris. To overcome these issues we have developed a photogrammetry workflow for reconstructing high spatial resolution three dimensional terrain models. The workflow utilises parallel and distributed computing to provide inexpensive terrain models that can then be used in numerical simulations of environmental flows. A section of Quebrada San Lazaro within the city of Arequipa, Peru is used as a case study to demonstrate the construction and usage of the terrain models and applicability of the workflow for a flash flood scenario.

Original languageEnglish
Pages (from-to)321-332
Number of pages12
JournalIFIP Advances in Information and Communication Technology
Volume448
DOIs
Publication statusPublished - 2015
Event11th IFIP WG 5.11 International Symposium on Environmental Software Systems, ISESS 2015 - Melbourne, Australia
Duration: 25 Mar 201527 Mar 2015

Keywords

  • Natural hazards
  • Numerical modelling
  • Photogrammetry
  • Rapid mass flow
  • Structure-from-Motion

Fingerprint

Dive into the research topics of 'A Distributed Computing Workflow for Modelling Environmental Flows in Complex Terrain'. Together they form a unique fingerprint.

Cite this