Abstract
Nitrogen related oxides are considered to be a major source of substances related to acid rain and possibly the global warming. To evaluate the dynamics of these substances in river basin environment, a quantitative assessment of the land cover and its change is necessary.
In this study, multi-temporal satellite data and the forest records were utilized to classify the land cover, particularly the forest vegetation distribution, of a mountainous region featured with frequent change in terrain and with a great coverage of forest vegetation located within the Nagara river basin. The influence of shadows caused by ruggedness in terrain was reduced by ratioing the radiance among several differential spectral bands (Band 5/Band 1, Band 4/Band 1). The difficulty in distinguishing certain categories of land coverage, a negative effect brought about by uniformization of the radiance with ratioing, was alleviated by composing a new color image from both the image after ratioing treatment and that of Band 3.
Through analyzing the newly composed color image and the original image of two different seasons, the forest vegetation was classified into four forest types, namely, the evergreen coniferous forest, the evergreen broad leaved forest, the deciduous broad leaved forest and the mixed forest. The evergreen coniferous forest was further divided by age into the following four different groups: a group under 20 years, a group from 20 to 35 years, a group from 35 to 65 years and a group above 65 years. The reliability of the classification was confirmed by comparing the obtained results with related census data. In addition, the change of the land coverage of this region with time was also examined by using two satellite data with a time interval of 20 years.
