Applying the remotely sensed data to identify homogeneous regions of watersheds using a pixel-based classification approach
Academic Article
Overview
Research
Identity
Additional Document Info
Other
View All
Overview
abstract
2019 Elsevier Ltd Identification of homogeneous areas within watersheds can help managers and decision makers to administer watershed policies and sustainable environmental management. These homogeneous areas do not follow the boundaries of sub-watersheds. This study presents a pixel-based classification approach to demarcate more realistic homogeneous areas within watersheds with greater details, instead of classification using average watershed attributes. Remotely sensed indices, such as normalized difference vegetation index (NDVI), leaf area index (LAI), soil-adjusted vegetation index (SAVI), normalized difference moisture index (NDMI), and information on land use, elevation, slope, aspect, relative slope position (RSP), and topographic wetness index (TWI), were derived from MODIS, Landsat, and Terra/ASTER and employed to identify homogeneous areas in the Karkheh Watershed, west of Iran. The association among variables was tested for multicollinearity, then fuzzy c-mean clustering approach was used to classify pixels. Two validation criteria of Xie-Beni (XB) and Fukuyama-Sugeno (FS) were applied to determine the best number of classes. Results of the fuzzy clustering indicated that the optimum number of homogeneous areas was equal to four (XB = 4E-06, and FS = 7232). The homogeneous areas identified in this study need different management and protection policies, which can help managers with sustainable environmental management.
