Meteorological disasters/Agriculture and forestry damage/Forest fires. Forest fire disasters/Forest fires.
Disaster name
Forest Fire in Nishiokoppe, Hokkaido Nishiokoppe, Abashiri Branch Office, Hokkaido Prefecture Occurred on April 21, 1998

43ha of forest (40ha of forest, 3ha of agricultural land) destroyed by fire
Author of WEB conversion
Kato Masato

Case Study

No. 2

1. Analysis objective

Investigate the effectiveness of satellite disaster monitoring of sudden forest fire outbreaks in Japan, with the goal of providing analysis results to field sites and putting satellite data into actual use.

2. Analysis procedure Analysis flow chart

1) Make ortho corrections and slope corrections of SPOT images before and after fires, based on DEM images

2) Fire spectral analyses: Three SPOT bands and four kinds of vegetation indices

3) Sampling and damage classification of devastated forest land (Severely damaged, moderately damaged, lightly damaged)

4) Direction of spreading fire and damage analysis using three-dimensional processing

5) Comparisons of satellite damage classifications and actual measured areas in the field

3. Analysis results

1) Use of satellite images to understand the size of area devastated by fire

The area devastated by fire was estimated to be approximately 24ha from satellite images, and the area devastated by fire measured in the field was approximately 40ha. The sample ratio was 60%.

The principal reason the estimated area was smaller than the actual field measurements was that areas damaged by forest floor ground fires were also added to the actual field measurements, but those parts of the forest with fire-resistant Japanese larch or natural forest were not included in the devastated area in the classifications by satellite because the surviving standing trees put out new leaves and covered the forest floor. In addition, measurements from the satellite were made in a horizontal plane, which resulted in the mountain topography slope areas having smaller measured size.

When damaged land classified from the satellite has standing trees at the actual site, there are charred or withered parts.

2) Use of satellites when a disaster occurs

Provide information for outbreaks of disasters in one of two ways. In emergencies, immediately provide false color images (positions of damage locations) and cluster classification images. For detailed information, create a classification map using the maximum likelihood method based on surface survey data, and provide this within one month.

4. Results from using the analysis results

1) Provided prefectural Forest Management Center with materials and analytical images of fire devastated land.

2) Contributed analysis results to general readership magazines and academic journals.

3) Identification of this fire's position was delayed because the location where the fire occurred was a remote mountain site. As a result, we did not complete the analysis until one month after the outbreak occurred, which prevented a prompt response. Neither the side distributing the data nor the side that received it had information confirmation and mechanisms in place, such as observing the site and confirming conditions from an adjacent area at a later date.

It was possible to take advantage of these efforts as a precedent for an approach to monitoring disasters, such as the later eruption of Mount Usu, from satellite.

Analysis flow chart including government response

5. Sources

Kato M.: Analysis of the Nishiokoppe Forest Fire Using Satellite Images, Forests, No. 563, pp. 39-44, 1999, Hokkaido Forest Promotion Association

Kato M.: Monitoring a Forest Fire in Nishiokoppe Village Using Satellite Data for Domestic Practical Use, Journal of the Remote Sensing Society of Japan Vol.21, No.4, 2001, pp.377-387