Collaborating GEO-BENE partners: IIASA (leader), KTL /Finland/
Background
Being able to correctly predict the location and scale of a possible forest fire is an integral part of fire management. A Nesterov index, based on the daily weather history, is one of the existing measures, which allow quantifying the risk of a forest fire. The air patrolling schedule, e.g., in the Russian Federation is based on this index. Therefore, better information on daily weather in terms of both, measurement accuracy and refinement of spatial scale, may lead to better fire management, by preventing or significantly limiting losses due to ignition.
Data
The dataset used for this study is gridded daily weather data for the year 2000 for the area covering parts of territory of Spain and Portugal located approximately between -7.5W, 42.0N and -0.5W, 38.0N. The grid size in the analysis is 50x50 km for the ‘fine’ grid and 100x100 km for the ‘rough’ grid.
Methods
Forest fire history is simulated autoregressively based on the Nesterov index and the associated fire patrolling regime. The resulting resource expenditure and fire losses are then assessed.
Results
Generally, information obtained for the finer spatial resolution results in slightly larger number of patrols (+3.62% according to the simulation) and considerable loss prevention (-20.64% respectively). The model revealed the possibility of optimizing the system’s performance by combining and integrating different observation systems (System of Systems effect). According to the simulations, the refinement of the weather data on about 20% of the total area in selected most critical sub-areas may deliver about 80% of the total possible improvement of the system’s performance.
Status
Publication: Khabarov N., Moltchanova E., Obersteiner M., Valuing Weather Observation Systems for Forest Fire Management. Systems Journal, IEEE, 2(3):349-357, 2008. http://dx.doi.org/10.1109/JSYST.2008.925979
Figure: Reduction of burned area (BA) by adding extra weather stations (144 in total). By doing that, the patrolled area (PA) is slightly increased at initial stage.