HR: 1340h
AN: B33B-1024 [Abstracts]
TI: Simulation for the expansion of the Boreal Forest Fire
AU: * KIMURA, K
EM: kimura@ssi.ist.hokudai.ac.jp
AF: Graduate School of Information Science and Technology, Hokkaido University, Kita 14, Nishi 9, Kita-ku,
Sapporo, 0600814
Japan
AU: Honma, T
EM: honma@ssi.ist.hokudai.ac.jp
AF: Graduate School of Information Science and Technology, Hokkaido University, Kita 14, Nishi 9, Kita-ku,
Sapporo, 0600814
Japan
AU: NAKAU, K
EM: knakau@lowtem.hokudai.ac.jp
AF: Institute of Low Temperature Science, Hokkaido University, Kita 19, Nishi 8, Kita-ku, Sapporo, 0600819
Japan
AU: KUSHIDA, K
EM: kkushida@pop.lowtem.hokudai.ac.jp
AF: Institute of Low Temperature Science, Hokkaido University, Kita 19, Nishi 8, Kita-ku, Sapporo, 0600819
Japan
AU: Fukuda, M
EM: mfukuda@pop.lowtem.hokudai.ac.jp
AF: Institute of Low Temperature Science, Hokkaido University, Kita 19, Nishi 8, Kita-ku, Sapporo, 0600819
Japan
AU: Hayasaka, H
EM: hhaya@eng.hokudai.ac.jp
AF: Graduate School of Engineering, Hokkaido University, Kita 13, Nishi 8, Kita-ku, Sapporo, 060-8628
Japan
AB:
1. Background Now, frequent occurrence of a forest fire serves as a situation which should be careful of in international
society. In the boreal forest, the main reason of the forest reduction is considered as climate change. The NOAA satellite
image could detect lots of forest fire in Siberia. The forest fire can be found by only satellite because the area is
enormous. After forest fire is detected, some of them will be extinguished by fire service. The fire service must go to
distinguish the important fire, but now, there is not a standard which fire should distinguish first or propose to fire
control zone (trench or cutting) Our purpose is to make the forest fire spread simulation in large area (about 300km x 300km
or larger). We think this is very useful to help fire fighting. 2. Forest Fire Detection and Spread Simulation NOAA satellite
images were analyzed with the detection method by Kawano and Kudoh (2003). For the fire spread simulation, the natural
environmental information is very important. We made the database of vegetation and topology. The metrological data is also
important. Simulation of fire spread is mainly used with the cellular automata and the time shearing methods by Fire and
Disaster Management Agency in Japan (1984). 3. Stop Line of the Forest Fire Spread If the forest fire simulation continue for
many steps, all the forest in the simulation area is burned out because the stop lines have not been considered yet. Now we
think about the stop lines with the topology and the wind direction. (Of course the rain is the best way to distinguish the
forest fire. But that is neglected.) It has been proven that as a tendency in the forest fire in a past, the fire does not
spread in the place like the following, and that the burn stops: (1) Windward side and down slope (2) Lee side We analyzed
the coupling of them. The best parameter of the burn stop probability is the proportion of the wind speed. The proportion
constant is calculated by the generic algorism. When we use these parameters, the precision between the detected fire stop
and the simulated fire stop is 44.3%. 4. The Simulation Results We could get all parameter to the simulate the forest fire
spread in the study area. We show the examples of the results. In the figures, the color of the cell is changed by spread
time. Vivid red is the earlier burned cell. Darker red shows the later burned cell. In this method, it is possible that the
simulation is expanded in whole the boreal forest where the satellite images ware obtained. 5. Conclusion With the NOAA
satellite images, the forest fire spread in Siberian boreal forest. This method expands to the real-time forecast of the
forest fire. In present simulation, the boreal forest was made to be an object. Using the similar method, the application to
tropical forests is also sufficiently possible, if the parameter is examined and changed. As further work, with not the NOAA
satellite images but the MODIS images, the higher resolution is taken. Especially in the tropical area, it will be necessary
to consider the more accurate weather prediction data because it rains more in the tropical than in high latitude dry area.
DE: 0430 Computational methods and data processing
DE: 0466 Modeling
DE: 0480 Remote sensing
DE: 0515 Cellular automata
DE: 0545 Modeling (4255)
SC: Biogeosciences [B]
MN: Fall Meeting 2005