Meltwater is stored in supra-glacial lakes on the Greenland Ice Sheet. Connections between the melt water, ice sheet dynamics and the extent of the surface hydrological system have been observed. This highlights the importance of being able to study the surface hydrology over large spatial scales with a high temporal resolution. In this study, we develop an adaptive classification method to identify and map these supra-glacial lakes using high temporal resolution satellite images from the Moderate-resolution Imaging Spectroradiometer (MODIS) sensor. Surface reflectance images from 2007 are used to extract information about the lakes, their morphologies and their surroundings. Using a multiclass approach we can recognize lake types that have previously been difficult to classify, such as deep lakes, lakes within cryoconite areas and lakes with floating ice. Given significant increases in melting in recent years relative to the long term record the inclusion of deep lakes might prove particularly important.