衛星リモートセンシングによる農業的土地利用の調査

抄録

In spite of such well-known merits as large scope of semi-real-time, periodical digital data image, remote sensing has not come into practical use yet in the field of agricultural land use survey in Japan, not only because of its rough resolution for Japanese arable land and poor cloud condition during main farming season, but also because of insufficient regional theory of agricultural economics which concerns with regional analysis in agricultural area.
Generally speaking, agricultural land use survey in Japan has been made for the purpose of understanding evolution of peasant economy through cropping system analysis and therefore it has not necessarily been abound in regional implications. But recently the concept of local agricultural system, which consists of villages as substantial and spatial units of farming, appears to express a significant object of governmental agricultural policy and this goeseespecially for recentepaddy conversion programme. With this change of management theory, agricultural land use survey by village is becomingemore essential for economic land classification which is now rather regionalization by social condition of farming than evaluation of potentiality of land for which it was ordinally intended.
Because long-term and nation-wide survey is necessary to map out a new strategy of rice production, remote sensing data can be applied, for the moment, to monitor distribution of paddy conversion. On a techinical point, it is not very difficult to extract paddy field and upland field.
According to a case study in this article, Landsat data, which is standerdized on UTM coordinated so that other data and cartographic information including village boundary, can give us general view of land use as well as converted paddy distribution even at village level (Photo 2, Photo 4, Photo 6), and it suggests a possibility to find new factors of land use change through remote sensing data analysis. However estimation of the area by Landsat data differs from seasons and rarely corresponds with census data. Improving quantitative accuracy and stability of seasonal and serial data should be the most urgent problem for forthcoming agricultural land use survey by satellite remote sensing.