Photogrammetric retrieval and analysis of small scale sea ice topography during summer melt: Fid-Bau Portal

Photogrammetric retrieval and analysis of small scale sea ice topography during summer melt

Divine, Dmitry V. / Pedersen, Christina A. / Karlsen, Tor Ivan / Aas, Harald Faste / A. Granskog, Mats / R. Hudson, Stephen / Gerland, Sebastian

Abstract

Abstract The paper presents a setup for photogrammetric retrievals of small scale sea ice surface topography using low-altitude aerial imagery. The setup features two digital cameras, a combined GPS receiver/inertial navigation system (INS) unit, and a laser range finder. The components are fit in a single aerodynamic enclosure mounted outside a helicopter cabin. Results from its first deployment during the field campaign on Arctic sea ice north of Svalbard during summer 2012 are shown. Comparison of photogrammetrically derived digital elevation models (DEMs) with in situ measurements of sea ice topography made on melting first year sea ice demonstrated the ability of the method to accurately recover the topography of sea ice including melt ponds with depths down to at least 0.3m. The inter-comparison of the photogrammetrically derived DEM and in situ measured elevations yielded estimates of a root mean square error (RMS) of about 0.04m and bias of 0.03m, both for sea ice freeboard and melt pond depths. The bimodality of the probability density function of measured melt pond depths was also accurately reproduced in the reconstructed DEM. Discrepancies between the measured and DEM distributions were within the range of the inferred uncertainty of the photogrammetric and in situ techniques, with some of the bias likely associated with sea ice melt during the time elapsed between in situ and aerial measurements.

Highlights • A setup for photogrammetric retrievals of small scale sea ice surface topography is presented. • Digital Elevation Model of sea ice surface including melt ponds with depths down to at least 0.3m is accurately recovered. • Agreement with in situ elevation measurements to within the RMS of 0.04m and a bias of 0.03m is demonstrated.