glimpse: Open-source Software for Estimating Glacier Velocities from Timelapse Cameras and Satellite Data

Abstract:

Timelapse cameras excel at continuous high-frequency observations uninterrupted by cloud cover or low light, precisely the situations where aerial and satellite platforms still fall short. However, the ability of a ground-based camera to resolve glacier motion is limited by its oblique perspective of the glacier surface. We present a novel feature-tracking algorithm that can estimate velocities from photographs taken from one or more camera positions, optionally constrained by elevation and velocity measurements derived from aerial or satellite imagery. Several additional algorithms help ensure the geographic alignment of the various data sources for successful large-scale processing: correcting for camera motion in long timelapse sequences, coregistering repeat elevation measurements, and aligning oblique photographs to aerial and satellite imagery.

Applied to 33,000 Columbia Glacier timelapse photographs, our image-processing pipeline recovers detailed glacier velocities, their associated uncertainties, and corresponding strain rates at 3-day intervals over a 13-year period, providing an unprecedented look at the seasonal and sub-seasonal variability of tidewater glacier dynamics over long time scales. Examples from the Columbia Glacier and elsewhere illustrate the capability of our methods, which have been released as an open-source Python software package.