Based on Landsat satellite imagery we set up a 32-year raster image time series. Due to the arid environment soil noise is an issue therefore we calculated the MSAVI. We carried out the following steps: 1. Data was pre-processed for clouds removal using quality assessment bands, 2. Valleys were extracted using buffers along the river network while non-vegetated areas such as glaciers and mountain slopes were masked out, 3. Breaks for Additive Season and Trend (BFAST) was applied to detect timing, magnitude, and number of breakpoints in the trend component of the time series. The dataset is suitable to analyze land cover change and detect the impacts of disasters in the eastern Hindu Kush region.