Area determination of rice paddy using satellite SAR data

Every year, the agricultural statistics section of the Japanese government announces rice planting paddy area and rice yield per hectare (ha). Presently, the rice planting paddy area is detected based on field surveys by human power. The Japanese government wants to determine the rice planting paddy area using remote sensing. Already, several results have come out using optical satellites. However, Japan has a rainy-season at crop growing time. It is difficult to observe a paddy every year. Synthetic Aperture Radar (SAR) can observe the earth's surface without influence of clouds. We are developing a determination method of the rice planting paddy area using RADARSAT/SAR. The paddy fields are filled with water during the rice-planting period. Almost microwave of SAR reflects on the water surface the same as on mirror, called 'mirror reflection'. This phenomenon makes backscatter of SAR a small value at the water-covered place. In paddy fields, backscatter is very low by mirror reflection at transplanting time, and volumetric scatter increases with the growth. We develop the methods using this characteristic. Our study area was the Saga plain in southeast Japan. We used two RADARSAT images, 02/July/2000 and 27/July/2000. Two scenes were acquired in the same mode. The image of 02/July was the ending rice transplanting time. 27/July was the season of rice growth, and backscatter of RADARSAT was increased in rice paddy fields. We developed the determination method of rice planting paddy field, that was "water area of 02/July minus that of 27/July". The accuracy of the estimation was 97.8% in the aggregate, and the accuracy of each municipality was distributed between 25.8 and 120.9. The accuracy was lower in the municipalities facing a mountainous region than for flat plains. Their municipalities were influenced by fore-shortening and lay over phenomena, and the position errors of two scenes had affects at calculation of two scenes. We corrected these errors using geometrical correction by DEM, and we applied the possibility area map of paddy fields made from the digital map to get greater accuracy.