Meteorological disasters/Strong winds/Wind toppled trees. Forest disasters/Wind toppled trees
Disaster name
Wind Toppled Forest Tree Damage from a Typhoon in Northern Kyushu
Author of WEB conversion
Hara Masanao

Case Study

No. 5

1. Analysis objective

Typhoon 9119, which pummeled Japan on September 27-28, 1991, was an exceptionally powerful storm that ranked among the top class of storms recorded since official weather observations were begun, reaching maximum wind speeds as high as 60m/s as it raced across the northern part of Kyushu. Forests were mown down by Typhoon 9119's powerful winds, causing unprecedented forest damage. The purpose of this analysis was to extract this large-scale forest damage region using LANDSAT/TM data, and conduct a survey to influence public functions such as national land conservation for forests, and particularly to influence water conservation functions.

2. Analysis procedure Analysis flow chart

When forest damage is caused by wind toppled trees, a forest will wither and die over time following the disaster. As this withering and death occurs, the damage regions can be sampled by using the fact that reflections of near-infrared radiation will decrease.

Specifically, we selected the disaster region training areas from two LANDSAT/TM images taken approximately one year before the disaster and one year after the disaster, band synthesis, and respective NDVI images, and extracted the damage zones using the maximum likelihood classifier.

3. Analysis results

The Forestry Agency's Typhoon 9119 Damage Recovery Countermeasure Committee conducted field surveys, which it complied and issued as a report. Based on the report, we excluded damaged forest areas based on aerial photographs and field surveys from the forest areas by relevant municipality, calculated the damage ratio and showed the distribution by classes at the municipal level.

Using this survey result and LANDSAT/TM, we verified the extracted wind toppled tree damage areas. The result showed that the values from the satellite data were about 1/3 of the value from the field surveys, and despite this discrepancy in values the trend including distribution patterns was sharply delineated. We consider the major reasons for this difference in values to be as follows. The form of damage to individual trees was divided into uprooting, bending and listing, trunk breakage, trunk splintering and trunk shearing.

Of these types of damage, all or part of a single tree will wither and die from trunk breakage and uprooting, while bending and listing, splintering and shearing will not result in the death of the tree.

The survey by the Forestry Agency, however, assumes that as a wood resource trees are damaged when they are in any condition that prevents the use of the wood; that is, economic loss is regarded as damage. In contrast to this, in the damage regions extracted from satellite data a condition in which the forest has perished as a living organism is regarded as damage.

This difference between economic damage and biological damage is considered to be the major reason for the discrepancy.

4. Results from using the analysis results

The damage in the northern part of Kyushu was especially extensive, while 22,000ha were damaged in Oita Prefecture, 13,000ha were damaged in Fukuoka Prefecture and 6,000ha were damaged in Kumamoto Prefecture. Given that conducting field surveys by placing large numbers of individuals in the field would require an enormous expenditure of time and effort, in many instances surveys relied on the subjective assessments of the investigators.

Moreover, the number of aerial photographs reached as high as 300 in Oita Prefecture alone, with reports that making visual interpretations of this data was an immense undertaking. With satellite data, such photos can be processed directly by computer, enabling authorities to reduce both staff and time requirements. Satellite data also makes it possible to extract information with greater objectivity.

For this disaster, the Forestry and Forest Products Research Institute , the Ministry of Agriculture, Forestry and Fisheries is using MOS/1b data from four months after the disaster to make an analysis, as reported very favorably by the Japan Agricultural News on January 27, 1993.

5. Sources

Report on "Verification of Research related to Creation of Water Conservation Function Charts for Water Resource Management," a theme of practical application of research for Fukuoka Prefecture water resources management on the NASDA-Prefecture co-operation pilot project (June 1999)