Space-time clustering analysis of wildfires: The influence of dataset characteristics, fire prevention policy decisions, weather and climate

Details

Serval ID
serval:BIB_A246AABE8AA1
Type
Article: article from journal or magazin.
Collection
Publications
Institution
Title
Space-time clustering analysis of wildfires: The influence of dataset characteristics, fire prevention policy decisions, weather and climate
Journal
Science of The Total Environment
Author(s)
Parente  Joana, Pereira  Mário G., Tonini  Marj
ISSN
0048-9697 (Print)
Publication state
Published
Issued date
2016
Volume
559
Pages
151-165
Abstract
The present study focuses on the dependence of the space-time permutation scan statistics (STPSS) (1) on the input database's characteristics and (2) on the use of this methodology to assess changes on the fire regime due to different type of climate and fire management activities. Based on the very strong relationship between weather and the fire incidence in Portugal, the detected clusters will be interpreted in terms of the atmospheric conditions. Apart from being the country most affected by the fires in the European context, Portugal meets all the conditions required to carry out this study, namely: (i) two long and comprehensive official datasets, i.e. the Portuguese Rural Fire Database (PRFD) and the National Mapping Burnt Areas (NMBA), respectively based on ground and satellite measurements; (ii) the two types of climate (Csb in the north and Csa in the south) that characterizes the Mediterranean basin regions most affected by the fires also divide the mainland Portuguese area; and, (iii) the national plan for the defence of forest against fires was approved a decade ago and it is now reasonable to assess its impacts. Results confirmed (1) the influence of the dataset's characteristics on the detected clusters, (2) the existence of two different fire regimes in the country promoted by the different types of climate, (3) the positive impacts of the fire prevention policy decisions and (4) the ability of the STPSS to correctly identify clusters, regarding their number, location, and space-time size in spite of eventual space and/or time splits of the datasets. Finally, the role of the weather on days when clustered fires were active was confirmed for the classes of small, medium and large fires.
Keywords
Forest fires, Space-time permutation scan statistics, Cluster analysis, Weather, Climate
Create date
03/06/2016 11:05
Last modification date
21/08/2019 6:12
Usage data