Studying glacial melt processes using sub-centimeter DEM extraction and digital close-range photogrammetry

Glaciers are sensitive to climatic variation and a particular developing area of investigation is in the field of "cryoconite”,referring to dustlike residues which form on the glacier surface. Cryoconite absorbs the Sun's shortwave energy, accelerates ice melt and be cause of the localised distribution of dust creates localised melting which is highly spatially variable. There is therefore a need to quantify the detailed topographic surface of ice and measure its variability through time. This paper describes the use of close range photogrammetry to reconstruct the glacier surface at the sub-centimetre or micro scale, an approach which may allow the relationship between cryoconite and ice surface properties to be explored over either space or time. The field campaign was conducted at Midtre Lovénbreen, Svalbard (78.88° North 12.08° East), during the summer of 2010 and executed using simple equipment and procedures. A simple and ageing Nikon 5400 5 MP camera was used to acquire all imagery, proving sufficiently robust for the challenging field environment. The camera was handheld a proximately 1.6 m above the ice surface, providing an oblique perspective. Images were acquired at three different camera/object distances, each generating coverage occupying three different areas. All imagery was processed using the commercial photogrammetric package Photo Modeler Scanner, generating three-dimensional point clouds consisting of many thousands of XYZ coordinates, each colour-coded. It had been feared that lack of texture in the ice surface combined with differing specular reflections in each image would compromise the DEM generation
process. Results were better than expected, although DEM quality proved to be variable depending on ice cleanliness and more significantly, the degree of obliquity of the image pairs. Despite these differences, digital close-range photogrammetry has proven to be a useful technique to reconstruct the glacier's surface to sub-centimeter precision. Moreover, the method is providing glacial scientists with new data to examine the relationship between cryoconite, ice surface roughness and melt processes.