Expression of Qualitative and Quantitative Information on Mangrove Forest in Okinawa
c) Classification of species and mixture condition as qualitative information
Although eight species of mangrove in a narrow sense are distributed in Okinawa, the dominant species are Rhizophola stylosa Griff. And Bruguiera gymnorrhiza (L.) Lamk. It is possible to regard mangrove forest composed with these two species on appearances. So one species mixture ratio is difference to subtract another species ratio from 100 %.
Classification of species and mixture condition as qualitative information was done by the maximum likelihood method with band 3, 4, 5 and 7 using field survey data of mixture ratio of R. Stylosa as training data. A pixel over 80 % of the ratio was indicated as R. Stylosa dominated, less than 20 % as B. gymnorrhiza dominated and in between 80 % and 20 % was done as mixed stand by respective marks. Classified result is generally displayed by colored a square. But this result was corresponding to black marks because it had to be overlaid on a colored square corresponded to estimated value for stand stock.
Result and Discussion
Estimated and printed out result as stand stock, classified and printed out result as species and mixed stand were shown in Fig.-2. Overlaid result of qualitative information as patterns to monochrome imagery as quantitative information was shown in Fig.-3. This mangrove forest is relatively smaller and distributed in the mouth of Fukido River in Ishigaki Island. An example of qualitative and quantitative simultaneously expressed imagery of larger mangrove forest along Nakama River in Iriomote Island was shown in Fig.-4. These islands are located in Yaeyama region, southern part in Okinawa.
Although it was hard to see black patterns to indicate species and mixed stand on monochrome background to indicate stand stock, it was easy to see pattern and color of background simultaneously on a colored imagery. So it became possible to know two information on mangrove forest by imagery produced directly through a personal computer system using remote sensing. It was possible to sum up respectively volume of resources for each species by this procedure not only the imagery.
AS data set can be made separately for land area and waters through this mask procedure, classified imagery and composed imagery with RGB also can be made for each category. So the imagery of mangrove forest can be overlaid on classified or composed imagery for categories outside mangrove forest. Such imagery or map and summation of quantitative information are useful for the sustainable control and managements of mangrove forest.
It is difficult to apply this way to the case which one or less than 5 dots correspond to one pixel to display more extensive area on a CRT. Drawing way of isopleth figure must be developed for such circumstances.
We think possibility of that relationship between CCT count of band 4 of Landsat 5 and stand parameters are almost similar region because the level of CCT count relates directly with quantity of tree crown. It is considered that the level naturally relates with other stand parameters through their relation with quantity of crown as photosynthesis organ. We must confirm the matter mentioned above. If regression coefficients between CCT count of band 4 and stand parameters are different among regions, it is possible to make a table of regression coefficient to apply estimation equation for respective regions. So field survey for the relationships between remote sensing data and stand parameters is fundamental and pressing activity for countries distributed mangrove forest.
