Methodology:

Data was collected using a FARO (registered trademark) Focus3D X 330 phase-shift scanner with a shortwave infrared wavelength of 1550 nm, mounted on a GitzoTM Mountaineer carbon fibre tripod using a heavy duty 3/8" camera fast-release mount. Scans were conducted throughout the channel reach at a horizontal spacing of between 2 and 9 m, at a height of 1.5 m. All channel scans were conducted at 1/5 resolution, providing point spacing of 7.67 mm at 10 m distance, with scans at the glacier surface being conducted at 1/4 resolution, providing point spacing of 6.14 mm at 10 m distance. Scan locations were determined by the channel morphology, with positioning of the scanner to obtain the greatest wall coverage with the minimum number of scans, and to prevent channel morphology from occluding parts of the channel. A4 planar checkerboard targets were inserted into the channel walls using ice screws and crocodile clips and A5 planar checkerboard targets were placed at intervals of > 1 m along the channel floor. Targets were placed strategically to enable registration of the point clouds to one another, with a minimum of 3 common targets between each adjacent scan.

The initial scan was conducted at the glacier surface and, thus, the GPS location of these scans were used to georeference the scans below the ice. The point clouds were manually registered to one another in a downstream direction from the base of the moulin using the targets in FARO (registered trademark) SCENE 5 software, with the use of distance based and stray filtering processes to remove points > 30 m away from the scanner and points arising from the laser beam hitting 2 surfaces.

Jpeg images of the scan reflectivity values were exported from FARO (registered trademark) SCENE 5 and linked in Kolor Panotour Pro 2.5 software to create a 360deg visualisation from each scan location, with inbuilt links to the adjacent scans and a map of the channel planform.

Point clouds were imported into Bentley Pointools CONNECT Edition software and the colour range was changed to reflect the height of each point above the initial scan location. Recording of camera positions throughout the channel was used to create a video flythrough of the point clouds.

Data collection:

Terrestrial laser scanner - FARO (registered trademark) Focus3D X 330

Software - FARO (registered trademark) SCENE 5, 3D Reshaper 2018, Bentley Pointools CONNECT Edition, Kolor Panotour Pro 2.5.

Data quality:

Each scan has been registered to the preceding scan with a mean registration accuracy of 3.3 mm. This error is the standard deviation of the orthogonal registration mismatch, expressed as cumulative distance offset in the X, Y and Z planes. Due to the unclosed traverse preventing equal error distribution between scans, the maximum combined absolute error at the end of the traverse is 73 mm, derived as the sum of orthogonal registration mismatches for the 28 scans.

Cleaning of the data used the 'distance based' and 'stray' filters within the FARO® SCENE 5 software to remove points > 30 m from each scan position and scan points that resulted from the laser beam hitting two objects.

Data losses are the result of the clean, shiny, smooth ice surfaces within the englacial channel, with these surfaces failing to provide laser returns due to the infrared absorption.