Sensitivity analysis of repeat track estimation techniques for detection of elevation change in polar ice sheets
MetadataShow full item record
Repeat track analysis is one tool that can be used to derive parameters describing elevation changes from elevation data collected from a satellite with a near-repeat groundtrack. While initially developed to study ocean topography, it was then applied to ice sheet data. This study expands upon that previous research by testing the method’s ability to estimate parameters using different amounts of data, different grid sizes and types, and different elevation models containing different parameters to be estimated. In all cases, ICESat-derived elevations were used as input data, as ICESat has a near-repeat groundtrack with extensive coverage of the Greenland and Antarctica ice sheets. Results were compared using the differences between modeled and ICESat-derived elevations, correlation of estimated parameters to known physical features, and differences between known and estimated parameter values for simulated elevation data. It was found that there should be data from at least as many distinct time periods (or, in the case of ICESat, laser campaigns) as parameters being estimated, grids centered on and aligned with the reference groundtrack should be used, and that elevation models containing a constant elevation change rate, slopes, an initial elevation at the grid center, and annual terms should be used. Crossover analysis is a different method to determine elevation change rate with elevation data and serves as an independent verification of the repeat track analysis method. It was found that the hdot values determined from crossover and repeat track analyses agreed to within 5 cm/yr in most areas of the ice sheets, with differences greater than 40 cm/yr along the coasts. While repeat track analysis provides greater coverage than crossover analysis, it is uncertain which method provides the most accurate results.