Quantitative Evaluation And Analysis Of Rocky Desertification Remote Sensing In Karst Mountainous Area Based On Landscape Unit

In this paper, The Guanling County as study area and multi-source and high resolution remote sensing as data resource, comprehensively using 3S technology, multi-criteria evaluation(MCE),analytic hierarchy process(AHP),K-means clustering method and natural breakpoint method. The study research rocky desertification evaluation basic unit in karst mountain area. Evaluation method, evaluation index system, index factors of remote sensing inversion. On the basis of analysis on the surface coverage, topography, lithology, rainfall and other factors, the study regionalize the elementary unit of desertification degree evaluation,which is determining to the landscape unit. And try to build evaluation index system of status evaluation on rocky desertification in karst mountain area based on landscape unit. For rapidly and quantitatively extracting the evaluation index, this study mainly use the way of quantitative remote sensing inversion to extract the four remote sensing indexes including photosynthetic vegetation coverage, the non-photosynthetic vegetation coverage, soil coverage and bedrock exposed rate to characterize the rocky desertification degree. And bringing the standard evaluation method(MCE) into improving the rocky desertification evaluation method,combing the method with the method remote sensing and GIS to evaluate Guan Ling County karst rocky desertification degree of the situation, and check the accuracy of evaluation results.(1)Based on mapping and analyzing the surface coverage, lithology, topography and rainfall factors, the research divides study area into landscape units using the method of overlay analysis. Finally, to make a Guanling County landscape element zoning map. Based on situation of rocky desertification in the evaluation.(2)In order to scientifically and accurately evaluate Guanling county rocky desertification situation, on the basis of previous studies, this study analyzes the four kinds of typical spectrum traits of the photosynthetic vegetation, non-photosynthetic vegetation, bare soil, bare rock in karst area, and then compares hyperspectral with multi-spectral(narrow band) data by experiment,which the hyperspectral data is Hyperion data, the multispectral data is TM data, contrasts various extraction and characterization of rocky desertification vegetation index, finally it is concluded that the precision of rocky desertification information extracted from hyperspectral data is higher than from multi-spectral data. By using triangular vegetation index(TVT) to extract photosynthetic vegetation coverage is more efficient, which is the highest correlation with green vegetation coverage. In order to scientifically and accurately evaluate Guanling county rocky desertification situation. On the basis of previous studies, this study analyzes the four kinds of typical spectrum traits of the photosynthetic vegetation, non-photosynthetic vegetation, bare soil, bare rock in karst area, and then compares hyperspectral with multi-spectral(narrow band) data by experiment,which the hyperspectral data is Hyperion data, the multispectral data is TM data, contrasts various extraction and characterization of rocky desertification vegetation index. Finally, it is concluded that the precision of rocky desertification information extracted from hyperspectral data is higher than from multi-spectral data. Using triangular vegetation index(TVT) to extract photosynthetic vegetation coverage is more efficient, which is the highest correlation with green vegetation coverage. And for the non green vegetation, various spectral index on the extraction of dry vegetation coverage, the effect of bare soil and bare rock are poor. Finally, the karst rocky desertification spectral index KRDSI3 can better extract information of dry vegetation- bare soil bare rock.(3)Using analytic hierarchy process(AHP) to establish the karst rocky desertification degree hierarchical structure model, and through the expert scoring to calculate impact factor weights, and establish classification quantitative evaluation index system of he classification of rocky desertification in karst mountain area based on landscape unit, and using K-means clustering method and natural breakpoint method to classify rocky desertification,and divided the ecological landscape unit group into the karst and non-karst landscape ecological groups. Then divide into five grades that is the very severely, severely, moderately, and mildly, no rocky desertification in the karst area of Guan Ling County, the karst area is beyond the scope of evaluation.(4)This study introduces a kind of MCE combining remote sensing and GIS evaluation method into the evaluation of rocky desertification of karst mountain area. In general, this method can produce high precision suitability map of rocky desertification. Confusion and error matrix evaluation shows that combining of remote sensing and GIS method with MCE can obtain the higher precision of evaluation than NDVI method. In addition, compared with the previous research on rocky desertification assessment,which mainly concentrated using means of remote sensing to extraction of vegetation and soil information in the karst area. this study combines the information of vegetation,dry vegetation, soil and rocky extracted from remote sensing with the terrain to assess rocky desertification,which produces more benefits. And the method of MCE with a variety of information is extremely powerful and flexible, easy to implement, and can also include spatial explicit quantitative and qualitative information, such as policy makers and experts, and landscape features, etc. In this study, using a scientific and practical rocky desertification assessment method can be used to improve the regional desertification assessment management scheme.