| Summary: | Land terminating Ice Sheets, and associated outlet glaciers, have been extensively studied due to their ability to be used as climate indicators and the potential effect on the rate of sea level rise. Therefore, studies of annual surface velocity have been under taken using orbital and aerial remote sensing, due to the easy access to data, and their temporal and geographic coverage. Satellite based multispectral image feature tracking have been extensively used as a method for track regional scale annual ice velocities. This study has compared five different pre-processing chains to help improve the calculation of large scale Ice Sheet velocities from Landsat imagery, by determine the optimum order and type of image enhancement techniques to be used. The study has used imagery from two outlet glacier located on the west coast of the Greenland ice-sheet, the Kangiata Nunata Sermia and Leverett glaciers, as they have been recorded to have significantly different annual surface velocities. The study found two main results, first, regional scale studies of ice sheet surface velocities, using feature tracking of Landsat 8 series satellite imagery, produce the best result using the first component in the Principal component analysis (PCA). Second, for local scale studies the use of a Gaussian filter followed by the PCA produced higher quality image enhancement. However, this was highly dependent on the bandwidth of the filter being specifically tailored to the study area. The results indicate that further improvement to pre-processing methodology could be gained through increased bandwidth selection automation.
|
|---|