Tracking icebergs with time-lapse photography and sparse optical flow, LeConte Bay, Alaska, 2016–2017

• Published in: Journal of Glaciology
• Main Authors: CHRISTIAN KIENHOLZ, JASON M. AMUNDSON, ROMAN J. MOTYKA, REBECCA H. JACKSON, JOHN B. MICKETT, DAVID A. SUTHERLAND, JONATHAN D. NASH, DYLAN S. WINTERS, WILLIAM P. DRYER, MARTIN TRUFFER
• Format: Article in Journal/Newspaper
• Language: English
• Published: Cambridge University Press 2019
• Subjects: glaciological instruments and methods; icebergs; ice-ocean interactions; remote sensing; Environmental sciences; GE1-350; Meteorology. Climatology; QC851-999; Journal of Glaciology; Alaska
• Online Access: https://doi.org/10.1017/jog.2018.105; https://doaj.org/article/9fe0fd4389f14ef6b160ac48b7455bc6

We present a workflow to track icebergs in proglacial fjords using oblique time-lapse photos and the Lucas-Kanade optical flow algorithm. We employ the workflow at LeConte Bay, Alaska, where we ran five time-lapse cameras between April 2016 and September 2017, capturing more than 400 000 photos at frame rates of 0.5–4.0 min−1. Hourly to daily average velocity fields in map coordinates illustrate dynamic currents in the bay, with dominant downfjord velocities (exceeding 0.5 m s−1 intermittently) and several eddies. Comparisons with simultaneous Acoustic Doppler Current Profiler (ADCP) measurements yield best agreement for the uppermost ADCP levels (~ 12 m and above), in line with prevalent small icebergs that trace near-surface currents. Tracking results from multiple cameras compare favorably, although cameras with lower frame rates (0.5 min−1) tend to underestimate high flow speeds. Tests to determine requisite temporal and spatial image resolution confirm the importance of high image frame rates, while spatial resolution is of secondary importance. Application of our procedure to other fjords will be successful if iceberg concentrations are high enough and if the camera frame rates are sufficiently rapid (at least 1 min−1 for conditions similar to LeConte Bay).