Estimating the Characteristics of Runoff Inflow into Lake Gojal in Ungauged, Highly Glacierized Upper Hunza River Basin, Pakistan

Motivated by the potential flood outburst of Lake Gojal in the ungauged highly glacierized (27%) upper Hunza River Basin (HRB) in Pakistan that was dammed by a massive landslide on 4 January 2010, we attempt to analyze the characteristics of water inflow to the lake employing remote sensing data, two hydrological models, and sparsely observed data. One of the models (Model I) is a monthly degree-day model, while another (Model II) is the variable infiltration capacity (VIC) model. The mixture of glacier runoff output from Model I and runoff over unglacierized areas calculated by Model II has a similar seasonal variation pattern as that estimated from data recorded at a downstream station. This suggests that glacier runoff is the main source (87%) of runoff inflow into the lake. A sensitivity analysis suggests that the water inflow to the lake is highly sensitive to an increase in air temperature. Runoff in May is predicted to sharply increase by 15% to more than two-fold if the air temperature increases by 1 to 7 degrees C, but it is predicted to increase only from 9% to 34% if the precipitation increases by 10% to 40%. The results suggested that the water inflow into Lake Gojal will not sharply rise even if there is heavy rain, and it needs to be in caution if the air temperature sharply increases. Analysis on long-term air temperature record indicates that the water inflow into the lake in May 2010 was probably less than average owing to the relatively low air temperature. Consequently, the flood outburst did not occur before the completion of the spillway on 29 May 2010.