Date of Award
12-17-2013
Document Type
Thesis
Abstract
Glacier inventories are used for many applications in glaciology, however, their manual compilation is time-consuming. Here, we present two new algorithms for the automatic compilation of glacier inventories. The first approach is based on hydrological modeling tools and separates glacier complexes into individual glaciers, requiring a digital elevation model (DEM) and glacier complex outlines as input. Its application to > 60,000 km² of ice in Alaska (~98% success rate) and southern Arctic Canada (~97% success rate) indicates the method is robust if DEMs and glacier complex outlines of good quality are available. The second algorithm relies on glacier outlines and a DEM and derives centerlines in a three-step 'cost grid -- least cost route' procedure. First, termini and heads are determined for every glacier. Second, centerlines are derived by determining the least cost route on a previously determined cost grid. Third, the centerlines are split into branches, followed by the attribution of a branch order. Application to > 21,000 Alaska glaciers shows that ~5.5% of the glacier heads and ~3.5% of the termini require manual correction. With corrected heads and termini, ~1.5% of the actual derived centerlines need edits. Comparison with alternative approaches reveals that the centerlines vary significantly depending on the algorithm used.
Recommended Citation
Kienholz, Christian, "New algorithms for the compilation of glacier inventories" (2013). Geosciences . 29.
https://ualaska.researchcommons.org/uaf_grad_geosci/29
Handle
http://hdl.handle.net/11122/4636