SOCET SET customer Swansea University, in the U.K., is studying glacier melt in Svalbard, Norway. The glaciers around Svalbard could make the largest contribution to sea-level rise of any arctic region outside of Greenland. A field study, named Sea Level Rise from ICE in Svalbard (SLICES), was conducted to gather historic topographic data sets on sea-level rise for comparison with current records of the same area. Research began in 2003. The primary goal was to measure volume changes of the benchmark Svalbard glaciers, using LIDAR and photogrammetrically derived DEMs to provide a strong baseline for continued monitoring in the area. The findings were applied to the entire archipelago with a regional mass balance model, which was used to derive 20th and 21st century contributions to global sea-level rise in Svalbard.
For the SLICES project, there were many large images that had been captured at 1:50,000 scale and scanned at a high resolution to maximize DEM resolution. SOCET SET’s flexibility with large images, input file formats, and ASCII files was a major advantage. SOCET SET’s Automatic Terrain Extraction (ATE) and Interactive Terrain Editing (ITE) modules offer a combination of automated and manual tools for building terrain and surface models, and work equally well with new and century-old data.
Stereo matching on surfaces such as glaciers, with repeating patterns and a lack of texture, is notoriously difficult. Through the use of back-matching algorithms in ATE, the scientists have been able to eliminate many of the blunders that are normally associated with stereo matching on such surfaces, and thus obtain a better automated DEM with far less manual correction required.
Occasionally, in extremely steep areas, or areas where fresh snow cover makes stereo matching difficult, the team implements a hybrid approach, which involves measuring DEM points or breaklines manually in ITE, then using these as seed points in ATE. If stereo matching is unreliable, it is preferable to have a hole in the data, as opposed to blunders. A TIN (Triangulated Irregular Network) DEM from ATE with back-matching yields much better results; it will identify such points as blunders and discard them.
Results from the study show that between 1961 and 2005 the average rate of melt was found to be about 0.47 meters vertically per year, with more melt occurring in recent years. Small glaciers like those in Svalbard represent only four percent of the world’s total land ice, but account for an estimated 20 to 30 percent of 20th century sea-level rise — and the melt has increased substantially since 1988. This work is extremely important for improving predictions of sea-level rise due to the density of population along the world’s coastlines.
Looking ahead, the Swansea Glaciology Group is turning its attention to Greenland, an area that has been identified as crucial for predicting future sea-level rise. For details on these and other projects underway within the Swansea Glaciology Group, please visit: http://geography.swan.ac.uk/glaciology/.
Tags: SOCET SET