Using ASTER Imagery for Land Cover Discrimination in Trans-Himalayan Landscape: Assessment of...

Using ASTER Imagery for Land Cover Discrimination in Trans-Himalayan Landscape: Assessment of Results from Upper Mustang, Nepal

SHARE

Nawa Raj Chapagain
Nawa Raj Chapagain
GIS Officer
King Mahendra
Trust for Nature Conservation

Upper Mustang region (2,567 sq.km.) lies in the rain shadow of the greater Himalayan range of north-central Nepal. Land cover of the region can be grouped into grassland, shrubland, bare land, agriculture and settlements, fords, snow and water. The region is covered mostly by desert steppe vegetation and bare land. The major difficulties for land cover discrimination in the area, as compared to other ecological zones, using satellite imagery are: (a) poor ground coverage and very low biomass of grassland vegetation (mainly graminoids belonging to genera Kobresia, Euphorbia and Potentila) make it difficult to distinguish from bare lands, (b) stunted growth of shrubs (mostly Caragana and Lonicera) poses difficulty in discriminating it from grasslands, and (c) these factors also limit the potentiality of distinguishing classes within grass and shrub communities. As part of the Upper Mustang Biodiversity Conservation Project (UMBCP) being executed by King Mahendra Trust for Nature Conservation (KMTNC), land cover mapping of the region was done in 2002 using ASTER satellite imagery. The methodology consisted of purposive collection of training sites data followed by land cover discrimination using different classification algorithms in a computer based multi-spectral digital image processing environment. In order to identify the appropriate classifier for trans-Himalayan landscape, results obtained from the use of different classification algorithms are compared. Additionally, the level of discrimination within the vegetative communities that could be successfully achieved using ASTER imagery interpretation is also discussed.