Under-shelf ice sampling by ROV reveals specimens of ice-associated crustaceans

• Main Authors: Owsianowski, Nils, Nachtsheim, Dominik, Held, Christoph, Richter, Claudio, Steinmetz, Richard, Bornemann, Horst
• Format: Conference Object
• Language: unknown
• Published: 2017
• Subjects: Ice Shelf; Larsen Ice Shelf; Riiser Larsen Ice Shelf; Riiser-Larsen Ice Shelf
• Online Access: https://epic.awi.de/id/eprint/45199/; https://hdl.handle.net/10013/epic.51377

Images taken by seal-borne cameras at the Drescher Inlet (Riiser-Larsen Ice Shelf) in 2004 led to the discovery of an unknown cryo-benthic community being attached head-down to the underside of the floating ice shelf at depths of around 130-150m (Watanabe et al. 2006). Resolution and exposure of the images did not allow distinct identification of species considered as likely isopods or cnidarians, and no information could be gained on the composition, size, dimension and density of the faunal aggregation. In order to re-assess the earlier findings, an Ocean Modules V8Sii Remotely Operated Vehicle (ROV) was launched through an artificial ice hole in the vicinity of the shelf ice cliff. The ROV provided high resolution and scalable video footage of the shelf-ice associated fauna. A custom-made ROV-mounted dredge was used to collect samples for further identification and DNA-analysis. A total of thirty specimens (15 adults, 15 juveniles) were sampled between 60 and 80m water depth. Additionally, 2 hours of high resolution digital video footage could be taken during two ROV transects at depths between 80m and 100m (along the vertical shelf ice wall and beneath the floating shelf ice). According to our standard ROV procedure, concurrent hydrographic (depth, temperature, salinity, pH, fluorescence, and oxygen) and three-dimensional dive data (depth, roll, pitch, orientation, and position) were recorded. Acoustic distance measurements and lens geometry of parallax free optics allowed to determine the strip width (0.5 – 2.0m) and cumulative length (122m) of the transect covering a total area of 133 square meters within the footage. The continuous scenes of video recordings were converted into single frame sequences and the number of individuals on the images was determined by visual counts. Tagging each individual with a marker and automatic registration of number and coordinates, prevented double counts and allowed for further distribution analyses. In total, 3,280 adults as well as 10,368 juveniles could be ...