Remote Sensing Measurement Of Vegetation Restoration In Lower Reaches Of Tarim River

Taking field survey vegetation data,multi-source remote sensing images in2000-2010as information source; taking Restoration Ecology, Basin Ecology and Plants Spectrum as theoretical basis; with the support of3S technology and statistic analysis, studying the response characteristic of measure factors of Populus euphratic to water delivering and expression form of vegetation coverage on different spatial resolution remote sensing images; building estimation models for vegetation coverage based on Hierarchy and Scale theory with multi-source remote sensing images. The main conclusions are as following:Studying on response characteristic and relationship of measure factors of P.euphratica, the results show that the tree-crown and height decreases with the increase of distance from river, while the change of DBH has no regularity. The precision of model established between DBH and height, has reached90.08%. The models established between tree-crown and DBH was81.37%, the fitting precision doesn’t show obvious regularity as the changes of the groundwater level.Using the method of man-computer interaction to abstract and analyze the crown of P.euphratica and Tamarix ramosissima at different distance from the stream channel based on QuickBird image, the crown interpretation accuracy of P.euphratica and T.ramosissima is higher, which reached84.00%,76.24%respectively; while the interpretation accuracy of canopy closure and mixed ratio is lower, which reached69.45%,66.96%respectively. Correlative analysis reveals significant positive correlation between tree-crown, DBH and tree-height (R2>0.8, P<0.01), the mean accuracy of the models is80.33%and89.25%respectively.According to ground investigation and multi-source remote sensing data built estimation model, with the increase of spatial resolution, the accuracy of estimation model has increased. Based on the high, middle-high and middle-low resolution remote sensing data built vegetation coverage estimation model, the accuracy are89.53%,86.98%and84.57%, respectively. The accuracy of remote sensing models is higher than that of vegetation index method.Based on middle spatial resolution data, vegetation coverage changes were analyzed qualitatively every two years with spatial models, the result indicates that:low coverage vegetation, bare land and sandy land were the key land categories in the study area. The area of vegetation coverage (Ⅰ-Ⅳ) increased986.76hm2,681.84hm2,1091.88hm2,8641.39hm2in2010compared with that in2000, while the bare land and sandy land (Ⅴ) decreasedl1420.37hm2. The conversion between Ⅲ and Ⅳ as well as between Ⅳand Ⅴ was the major transformation type of vegetation coverage throughout the study period in the study area.The mentioned indicates that:based on high resolution data could detect measure factors, while vegetation coverage is a key factor to large area and long term vegetation restoration monitoring. Not only vegetation coverage is a key factor to vegetation restoration, but also is a usable indicator to using different resolution data to estimation. It makes vegetation coverage quantitative analysis possible in the absence of field data using vegetation coverage estimated models.