Please use this identifier to cite or link to this item: https://hdl.handle.net/10316.2/34098
Title: Map partitioning to accelerate wind field calculation for forest fire propagation prediction
Authors: Sanjuan, Gemma
Brun, Carlos
Margalef, Tomàs
Cortés, Ana
Keywords: Wind field;map partitioning;data parallelism;forest fire propagation prediction
Issue Date: 2014
Publisher: Imprensa da Universidade de Coimbra
Journal: http://hdl.handle.net/10316.2/34013
Abstract: Forest fire is significantly affected by wind. However, meteorological wind is modified by terrain topography in such a way that a different value of wind speed and direction is given on every point. To estimate forest fire propagation on such conditions it is necessary to couple a wind field model to the forest fire propagation model. These models are time consuming from the computational point of view and may be parallelised to make them feasible in an operational scenario. So, a map partitioning has been applied to accelerate wind field calculation. The results show that the wind field and the forest fire propagation prediction are not significantly affected by map partitioning and the time to reach the forest fire prediction is significantly reduced.
URI: https://hdl.handle.net/10316.2/34098
ISBN: 978-989-26-0884-6 (PDF)
DOI: 10.14195/978-989-26-0884-6_37
Rights: open access
Appears in Collections:Advances in forest fire research

Files in This Item:
File SizeFormat 
978-989-26-0884-6_37.pdf2.1 MBAdobe PDFView/Open
Show full item record

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.