Methodology:

The erratics described in the dataset were originally collected for cosmogenic nuclide surface exposure dating, during 3 different field campaigns: in 2006 on a geological campaign undertaken during cruise ANT XXIII/4 of RV Polarstern for the BAS GRADES-QWAD project, in 2010 on a geological campaign undertaken during cruise ANT XXVI/3 of RV Polarstern, and in 2019-20 on a geological campaign supported by British Antarctic survey logistics as part of the Geological History Constraints project (one of the component projects of the International Thwaites Glacier Collaboration). The data here pertain to their location, size, shape, orientation, degree of weathering and lithology. Locations (latitude, longitude and altitude, m above sea level) were measured using GPS (details for each season are provided in point 7 below). The size of each erratic was measured in the field using a standard ruler. Three axes were measured: short, intermediate and long, and their lengths recorded in centimetres. The shape (degree of roundness) of each erratic was determined by a visual assessment and comparison with the Powers roundness scale (Powers, 1953). Long axis orientations were measured as magnetic bearings and corrected to true bearings using the magnetic declination of 46° E. The degree of weathering was assessed by visual comparison with a weathering index simplified from existing schemes published by Owen et al. (2009), Moriwaki et al. (1994) and White et al. (2009), as follows:

Weathering classification

W1: Heavily weathered surface, surrounded by spallation products; no iron staining or pitting on upper surface

W2: Moderately weathered surface, iron stained, with some spalling/pitting of upper surface

W3: Intact slightly weathered surface, unspalled

W4: Fresh surface, unweathered, unspalled

After sampling, the lithology of each cobble was classified using visual estimation of the modal proportions of constituent minerals (specifically quartz, alkali feldspar and plagioclase feldspar) plotted on a QAPF diagram (Streickheisen, 1976).

Data collection:

Data collected in 2019-20 season:

Position (latitude and longitude, in decimal degrees) and altitude (metres above ellipsoid) were measured for each erratic using a Javad Triumph II GPS receiver. Files were processed on return to the UK. The heights above ellipsoid were corrected to orthometric heights (height above geoid EGM08 in metres above sea level) using Precise Point Positioning software [version 2]. As a back-up, satellite GPS measurements from a handheld Garmin 60CSx GPS unit were recorded at each sample location. Positional accuracy of the Javad Triumph II receiver is 0.010 m (horizontal) and 0.015 m (vertical). No other instrumentation or software was used.

Data collected in 2009-10 season:

Position (latitude and longitude, in decimal degrees) and altitude (metres above sea level) using a hand-held Garmin GPS Map 60CSx and helicopter altimeter and calibrated to known elevations. Positional accuracy is <5 m (horizontal) and <5 m (vertical). No other instrumentation or software was used.

Data collected in 2005-06 season:

Position (latitude and longitude, in decimal degrees) and altitude (metres above sea level) were measured for each erratic using a hand-held Garmin E-Trex Summit GPS. Positional accuracy is <5 m (horizontal) and <5 m (vertical). No other instrumentation or software was used.

Data quality:

For erratics that are equant in shape, no long axis orientation can be measured and the corresponding values cells in the data table are left blank. Where no long axis orientation was measured in the field and it was not possible to reconstruct it from photographs, the cells are filled with'n.d' for 'not determined''.