Abstract

Seasonal meltwater lakes form when surface runoff gets trapped in topographic depressions on the surface of the Greenland Ice Sheet (GrIS) The development of such lakes affects the surface energy balance of the ice sheet, (as the lower albedo of surface water increases radiation-driven melting), and the dynamics, (as sudden drainage of such lakes has been linked to observed increases in ice velocity). The areal extent and depth of lakes has been inferred from satellite imagery of the ice sheet. However, the lakes are transient features which often fill and drain during the course of a melt-season, meaning that estimates of potential surface water storage from such studies are likely to be too small as it is unlikely that all possible lakes will be observed. In this paper I present a different model-based strategy to estimate the seasonal potential surface water storage for a 2200 km2 area of the western Greenland Ice Sheet. The approach combines a high resolution surface mass balance model (used to calculate surface melt) with a novel model for supraglacial water routing which calculates the time-dependent filling of supraglacial lakes. The surface melt model is driven by meteorological data collected at the GC-Net JAR 1 station; the DEM for the melt model and the surface routing model is based on high resolution ASTER Global Digital Elevation Model (GDEM2) data. Lake locations and filling rates are compared with observed lake volumes in LANDSAT imagery for 2001 and 2005. For our study area, there is a close agreement between the locations of observed lakes and modelled lakes. The model results shows a clear seasonal pattern of lakes progressively filling at higher elevations over the melt seasons. This study also shows that at least for the areas considered, the GDEM2 data are of sufficient quality that 100 m scale surface features can be reliably identified, suggesting that future high resolution modelling could make use of the GDEM data.