ICESat laser full waveform analysis for the classification of land cover types over the cryosphere

In this study, a terrain classification algorithm is presented that was derived from the various properties of the full waveform returned signals collected from the ICESat mission. Such algorithms are beneficial for current and future studies of the cryosphere, particularly Greenland and Antarctica, by helping to identify the change in large scale surface properties over time. The algorithm developed was validated over a test region in the Dry Valleys of Antarctica, where the terrain is well known. The new classification algorithm distinguishes between four different types of terrain: snow, rock, ice and water. A description of the decision tree behind the algorithm will be provided, along with the results from the validation site. Over the Dry Valleys test region, the algorithm was shown to achieve an overall classification accuracy of 74%, with the potential for increased performance with future tuning of the algorithm.