Classification
Meteorological disasters/Rainfall/River floods. River disasters/River floods
Disaster name
Flood of Kashimadai-machi Region in Miyagi Prefecture Following Torrential Rains, Typhoon No. 10, 1986
Author of WEB conversion
Hara Masanao

Case Study

No. 4

1. Analysis objective

Determine the relative extent of damage to wetland rice by slicing LANDSAT/TM band 4 CCT values one month (September 7) after flooding of wetland rice fields in the Miyagi Prefecture Kashimadai-cho flood region caused by heavy rainfall from Typhoon 8610 in August 1986, and map the results directly to a map using a general-purpose printer equipped with eight color pens.

Verify the results by comparing to local municipality field survey results, and use as the basis for projecting the decline in wetland rice due to the flood.


2. Analysis procedure Analysis flow chart

Extraction of paddy field regions:

To limit the wetland rice field regions subject to analysis, from LANDSAT/TM data for June 16, 1985, the year prior to the flood damage, we extracted classifications from by the maximum likelihood classifier as territory showing the mid-range spectral characteristics of vegetation and water, and level slice processed this using the prescribed CCT count values for this paddy field area.


3. Analysis results

The map-based direct mapping method is extremely valuable as a method for creating flood condition maps for use in disaster prevention. Position mismatches average 0.25mm for a 1/25,000 topographic map, and there are allowance limits of error for topographic map use.

This flood situation map expresses the ground height (depth of ponding) as topographical information, and can forecast the length of the flood water inundation period. The possibility of quantitatively estimating paddy rice damage based on a flood situation map from the correlation between the number of inundation days and paddy rice yield was demonstrated from the results of analysis of the middle reaches of the Sakura River and Kokai River catch basins in Ibaraki Prefecture (Kishi S., Okura H. and Akutsu T., Journal of the Remote Sensing Society of Japan, Vol. 6, No. 4, pp. 21-25, 1986). For the present analysis we adopted a more direct method by using the fact that the dynamic conditions of vegetation such as paddy rice are sharply reflected in the near-infrared region of satellite observation values. In other words, we determined the valued from level slice processing images using LANDSAT/TM data obtained one month after the flood.

We confirmed that there is a good agreement between area distribution of the degree of paddy rice plant damage and the inundation period and distribution of paddy rice actual measured yield.

In surveys of rice fields in the Miyagi Prefecture Kashimadai-machi region, there is a good correlation between LANDSAT/TM band 4 CCT values one month after the flood and the actual measurement of paddy rice (the correlation coefficient is 0.86); the relationship is:

Yield (kg/10a) = -576.1 + 9.6 x (band 4 CCT values)


4. Results from using the analysis results

The evaluation and field surveys based on use of the result of this research were conducted in cooperation with Kashimadai-machi , Miyagi Prefecture.

In addition to introducing these results in the "Record of the Yoshida River Flood, pp. 188 (1987)" published by Kashimadai-machi, Miyagi Prefecture, the results have been used since then for disaster prevention planning (see, for example, "Water Regulations (4)" in the Yomiuri Shimbun for November 11, 1999).


5. Sources

Kishi S.: Journal of the Remote Sensing Society of Japan, Vol. 9, No. 4, pp. 55-57, 1989

Kishi S. and Tsunakawa I.: Proceedings of the 13th Conference of the Remote Sensing Society of Japan, pp. 21-24,1987

Kishi S.: Proceedings of the 19th Conference of the Remote Sensing Society of Japan, pp. 11-14, 1993