Change detection in forest land cover type using Aster and Landsat data

Abstract

The objective of this study was to evaluate the potential for monitoring forest change using Landsat ETM data and Aster data for two periods (2000 - 2003 and 2003 - 2006). This was accomplished by performing three widely used vegetation indices: Normalized Difference Vegetation Index (NDVI), Soil Adjusted Vegetation Index (SAVI), and Transformed Difference Vegetation Index (TDVI). An RGBNDVI change detection strategy to detect major decreases or increases in forest vegetation was developed as well. These indices were applied to a case study in El Rawashda forest reserve, Gedaref State, Sudan, and their results and accuracy were discussed. Results showed that the vegetation index maps obtained by NDVI and SAVI transformations within each computational group were similar in terms of spatial distribution pattern and statistical characteristics. As far as the degree of greenness of vegetation was concerned, the TDVI appeared to be the most sensitive. For the first period, the highest accuracy was obtained by SAVI (62.5%); however, the poorest accuracy was achieved by TDVI (59.5%). For the second period, TDVI revealed the highest accuracy (60.1%), whereas both NDVI and SAVI counted accuracy of 59.2%. Generally, the study proved that all vegetation indices produced reasonable approaches to map land cover changes over time and help to pinpoint deforestation and regrowth in the study area.