Photogrammetric Bathymetry for the Canadian Arctic

• Main Author: Hodul, Matus
• Format: Article in Journal/Newspaper
• Language: unknown
• Published: Université d'Ottawa / University of Ottawa 2018
• Subjects: Photogrammetry; Bathymetry; Hydrography; Remote Sensing; Arctic; Nunavut; Canada; Frobisher Bay; Cambridge Bay; Coral Harbour; Queen Maud Gulf; Arviat
• Online Access: https://dx.doi.org/10.20381/ruor-21822; http://ruor.uottawa.ca/handle/10393/37553

This study proposes and demonstrates a through-water photogrammetry approach for Satellite Derived Bathymetry (SDB), which may be used to map nearshore bathymetry in the Canadian Arctic. A four step process is used: First, a standard photogrammetric extraction is performed on 2 m resolution WorldView stereo imagery, then apparent depths are calculated by referencing submerged points to the extracted elevation of the water level seen in the image. Due to the effects of refraction, these apparent depths are underestimates, and a refraction correction factor is applied to convert to actual depths. Finally, tidal stage at the time of image acquisition is used to bring depths to chart datum. A post processing step may be applied to remove erroneous depths caused by water surface objects such as boats, debris, or large waves. This was demonstrated in six study areas across Nunavut, Canada to test its robustness under a variety of environmental conditions, including different seafloor types, and under varying sea states. The six study sites were (with vertical accuracy given in Root Mean Square Error/and vertical bias, both in meters): eastern Coral Harbour (1.18/0.03), western Coral Harbour (0.78/-0.32), Cambridge Bay (1.16/0.08), Queen Maud Gulf (0.97/0.13), Arviat (1.02/0.13), and Frobisher Bay, where bathymetry extraction largely failed due to unfavourable sea surface conditions. These findings show that the proposed method has similar or better vertical accuracy as currently established SDB approaches; however, it has several benefits over the established methods which make it better suited for the Arctic. Namely, not requiring the precise atmospheric correction necessary for physics-based models, which is difficult at high latitudes; as well as being able to function in heterogeneous seafloor environments and not needing in-situ calibration data like the empirical spectral ratio approach, better suiting it to remote Arctic waters which often lack existing bathymetric survey data.