Evaluation of Remote Sensing Techniques for Ice-Area Classification Applied to the Tropical Quelccaya Ice Cap, Peru

Abstract

Concerns of deglaciation in the Andes prompted this investigation of methods to determine ice extent of glaciers from satellite imagery. Several remote sensing techniques to delineate ice cover are evaluated through application to the tropical Quelccaya Ice Cap. Each technique is applied to a single image that was previously geometrically and radiometrically corrected, and their results are evaluated on accuracy (as compared to a hand-digitization) and the amount of time required by a human operator to initiate and process the technique. Where the techniques require parameters, the results are shown from the best set of parameters tested. The effect of parameter selection on the results is discussed for one method. Three methods are ultimately recommended for future use with Landsat imagery. A commonly used band-ratio technique is found to be the most cost-effective method for determining ice-covered area, but the most accurate method was the supervised spectral angle mapper algorithm.