Across High Asia, the amount, timing, and spatial patterns of snow and ice melt play key roles in providing water for downstream irrigation, hydropower generation, and general consumption. The goal of this paper is to distinguish the specific contribution of seasonal snow versus glacier ice melt in the major basins of High Mountain Asia: Ganges, Brahmaputra, Indus, Amu Darya, and Syr Darya. Our methodology involves the application of MODIS-derived remote sensing products to separately calculate daily melt outputs from snow and glacier ice. Using an automated partitioning method, we generate daily maps of (1) snow over glacier ice, (2) exposed glacier ice, and (3) snow over land. These are inputs to a temperature index model that yields melt water volumes contributing to river flow. Results for the five major High Mountain Asia basins show that the western regions are heavily reliant on snow and ice melt sources for summer dry season flow when demand is at a peak, whereas monsoon rainfall dominates runoff during the summer period in the east. While uncertainty remains in the temperature index model applied here, our approach to partitioning melt from seasonal snow and glacier ice is both innovative and systematic and more constrained than previous efforts with similar goals.
Asia, a region grappling with the impacts of climate change, increasing natural disasters, and transboundary water issues, faces major challenges to water security. Water resources there are closely tied to the dramatic Hindu-Kush Himalayan (HKH) mountain range, where over 46,000 glaciers hold some of the largest repositories of fresh water on earth (Qiu 2010). Often described as the water tower of Asia, the HKH harbors the snow and ice that form the headwaters of the continent's major rivers (Bandyopadhyay 2013). Downstream, this network of river systems sustains more than 1.3 billion people who depend on these freshwater sources for their consumption and agricultural production, and increasingly as a source of hydropower (Immerzeel, Van Beek, and Bierkens 2010; National Research Council 2012; Rasul 2014).