Estudo da paisagem do setor sul da Bacia Hidrográfica do Tarumã-Açu, Manaus, utilizando imagens multipolarizadas do radar ALOS/PULSAR

Abstract

The regional importance of the Tarumã-Açu River basin, its location and extent, which represents about 12 percent in the city of Manaus, and its current situation of threat due to disorderly occupation problems, forest conversion to other types of use and soil cover, motivated the study of landscape structure analysis of this area. For this analysis a map of use and soil of the basin and cover, and indexes of landscape ecology were produced. Creating the map was based on digital supervised classification (maximum likelihood algorithm) of orbital images active sensor digital ALOSPALSAR of the year 2007, with global accuracy of 77.47. The elements of the landscape studied in this work were those of smearing and array, and the rates of landscape ecology used, that make the fragments-level characterization and use class-level and soil cover. The largest class of land cover found in the rain forest was 64.98 ha, representing 46,001 ha and demonstrating that the studied area had great natural vegetation cover in 2007. The second largest class of soil coverage was represented by the urban area by 18.02, totaling 12,759.53 acres, showing that already at the time (5 years by now), the basin was already in advanced process of occupation and man-made impacts. This can still be justified by the fact the bowl is partially located within the urban area of the city of Manaus. The porosity of the matrix of the basin was represented by 64,282 fragments. This high value is justified by the size of the study area bounded and the spatial resolution of the image ALOSPALSAR with 12.5 meters, which results in a zoom of the study area, highlighting areas usually omitted for images of lower spatial resolution. This high value is justified by the size of the study area bounded and the spatial resolution of the image ALOSPALSAR with 12.5 meters, which results in a zoom of the study area, highlighting areas usually omitted for images of lower spatial resolution. These fragments totaled 24,796.35 ha, total study area 47.07. Of these, 18.02% fragments formed by the urban area, followed by agriculture and secondary by forest, soil and water. The fragment had 95.94 % area class less than 1.0 ha, 3.78% has area between 1.0 and 10 hectares, and only 0.27% has area greater than 10 ha, totaling 174 large fragments. Most of the fragments of size less than 1.0 ha belong to the urban area, showing 48,467 fragments, and in the area greater than 10 ha this kind of coverage presented 59 spots. In relation to the average values and sets, the shape of fragments results were similar for all classes studied - around 2. The results were satisfactory for the studied region, since become required the use of remote sensing techniques due to the climatic characteristics and territorial extension of the Amazon