Selection and Characterization of Glaciers on the Tibetan Plateau as Potential Pseudoinvariant Calibration Sites

Abstract

Vicarious calibration using pseudoinvariant calibration sites (PICS) has become increasingly popular, as it is the most cost-effective method for on-orbit calibration of space-borne optical sensors. Glaciers on the Tibetan Plateau (TP) are located at more than 4500 m above sea level. These snow targets, exhibiting much brighter reflectance and less affected by aerosol and water vapor than desert targets, can help to reduce the uncertainties associated with vicarious calibration. In contrast to polar snow, the TP is not limited by polar night, thus ensuring continuous calibration monitoring. This study investigated the potential role of TP glaciers as PICS for the first time. To accomplish this, the coefficient of variation (CV) was applied to characterize the spatial uniformity and temporal stability of five relatively larger and flatter TP glaciers based on several independent Landsat Thematic Mapper and Operational Land Imager datasets over 20 years. The criteria of CV < 3% (both spatial and temporal) through Landsat visible and near-infrared channels were used to identify potential PICS across the candidate glaciers. Finally, three snow PICS with areas of at least 7.2 km2 and two snow PICS with areas of 1.2–2.1 km2 were identified. The area of the snow PICS is smaller in autumn and winter due to the snow cover. The feasibility of using these sites for long-term calibration monitoring of satellite sensors is also discussed. Results indicate that although the snow PICS exhibits stronger BRDF than the invariant desert target, the MODIS calibration trends detected from snow PICS are comparable to those from the desert target.