全球智库信息集成服务系统

Accurate mapping of glaciers is important for management of water resources, studying glacier mass-balance, and monitoring glacier health. However, many glaciers are partially covered by debris of varying thicknesses, and these debris-covered glacier areas are difficult to clearly distinguish from the adjacent valley rock using optical remote sensing data due to their similar reflectance in the visible to near-infrared (VNIR) wavelength region. For better detection of the debris-covered glacier margin, a new method was proposed here which considers glacier terrain features (concave and convex landscapes) as well as VNIR reflectance, shortwave-infrared (SWIR) reflectance, and radiometric temperature information. The proposed method uses several geomorphometric parameters (slope, plan, and profile curvature) and a Landsat TM band ratio image {TM TIR/(TM NIR/TM SWIR)} for mapping the Yengisogat glacier in the Shaksgam valley of China. The final debris-covered glacier boundary is generated by thresholding the band ratio image and overlaying it onto the result derived from cluster analysis performed using plan, and profile curvature, and combined with mean of slope information. Evaluation of the final debris-covered glacier outline was carried out through comparison with GIS data from several reference data sets, and the final Yengisogat glacier outline had a difference of less than 1% in area compared to the reference data sets.