Runoff from glacierized mountains: A model for annual variation and its forecasting

The only practical means of forecasting spring and summer runoff in remote high mountain basins such as the upper Indus in Pakistan is from preseason snow cover on satellite images. Previous studies have shown that regression coefficients of runoff on snow cover differ between basins in magnitude and sign. A conceptual analysis explains this, identifies the glaciological and snowpack variables that control annual variation in meltwater runoff, and forms the basis of a parametric forecasting model that is successfully applied to contrasting tributaries of the upper Indus, despite very crude parameter values. It incorporates the uncertainty due to possible nonaverage summer weather. Suggestions are made about additional remote sensing and ground reconnaissance investigations that could improve forecasting capability in this and similar regions.