Image Processing to Support The Reference Elevation Model of Antarctica (REMA)

Ian Howat, Ohio State University

Usage Details

Our objective is to create a high resolution (8m), precisely registered (<1m error) and time stamped Reference Elevation Model of Antarctica (REMA) to serve as benchmark for detecting ice sheet change, input for ice sheet and geodynamic models, processing of remote sensing data, field logistics support and other uses. REMA will be constructed from stereoscopic elevation models extracted from pairs of sub-meter resolution satellite imagery. Image processing for REMA will require throughput of approximately 225,000 individual jobs totalling 1.2 million node hours. Only Blue Waters is able to accommodate this scale of throughput in a practical timeframe (months). This project will use the same workflow and software as the current ArcticDEM project on Blue Waters, and therefore will leverage that development. Upon completion the REMA product will be made openly available on the Polar Geospatial Center's website. The REMA project has been funded by the NSF Office of Polar Programs.