Further by downloading this data the user acknowledges that they agree with the NERC data policy (), and the following conditions:

1. To cite the data

Abrahamsen, E. P. (2019). A bathymetric compilation around the South Orkney Islands, 1987-2017 (Version 1.0) [Data set]. UK Polar Data Centre, Natural Environment Research Council, UK Research & Innovation. https://doi.org/10.5285/1158843c-d238-439a-ad42-9645f37ac582

2. The user recognizes the limitations of data and will not use this data for navigation purposes. Use of the data is at the users' own risk, and there is no warranty as to the quality or accuracy of any data, or the fitness of the data for your intended use. The data are not necessarily fully quality assured and cannot be expected to be free from measurement uncertainty, systematic biases, or errors of interpretation or analysis, and may include inaccuracies in error margins quoted with the data.

Methodology:

The bathymetry map was compiled from a variety of different data sources. The primary data are multibeam swath bathymetry collected from scientific cruises undertaken by British Antarctic Survey (BAS), Alfred Wegener Institute (AWI), or acquired during RV Nathaniel B. Palmer, RRS James Cook and ESPS RV Hesperides expeditions. The complete list of the cruises used for the compilation is available in the supplementary documents (Source_data.csv file).

The multibeam data are merged with GEBCO_2014 (version 20150318) and GEBCO_2019 (doi: 10.5285/836f016a-33be-6ddc-e053-6c86abc0788e). Both of these grids are available from https://www.gebco.net.

The grid with bathymetry data only was created using the mbgrid program from MB-system version 5.5.2336 , using the parameters "-E100/100/meters -R-47/-37/-63/-59 -G3 -A2 -T25 -W4 -N -M -C5/1 -P1 -F1 -V". This uses a Gaussian weighted mean filter. Values are interpolated using a spline (tension 25) up to five grid cells from the nearest data point only if additional data are found in two opposite directions.

To merge with the global grids, the resulting grid is then loaded into Matlab R2019a. A subset of the GEBCO_2014 grid at 30 arcseconds or GEBCO_2019 at 15 arcseconds resolution is loaded into Matlab and interpolated onto the same grid. The grids are then subtracted, and the resulting grid is expanded by one grid cell in all directions, with the new cells filed with zeros. A list of the resulting values (excluding those that do not have any multibeam data) is exported to a temporary text file. This file is then run through mbgrid with parameters "-G3 -A2 -T50 -C9999/3 -W50 -V", resulting in a grid of differences between the multibeam and background data, with all data gaps filled with a spline. This offset is then subtracted from the background data, and the resulting values substituted into the gaps in the multibeam grid. The Matlab codes are available on GitHub at the following address: https://gist.github.com/PovlAbrahamsen/6ebce11cc637e8f434267c03672a354b

Data quality:

Quality control was undertaken manually to remove any outliers in the source datasets. This involved cleaning much of the raw data using the "mbedit" program from MB-system. For data processed in this way, edit files have been submitted to the Polar Data Centre. In source files that were ingested as text files, outliers were manually removed.

To verify the accuracy of the resulting dataset, bottom depths were compared against CTD casts on cruise JR16005, excluding casts that were performed when the ship was moving and those where the CTD did not get within altimeter range of the bottom. Bottom depths were calculated from the CTD altimeter when near the bottom, and these were compared against the gridded dataset. On 81 casts, the median difference between the CTD-derived depths and gridded depths was 4.1 m with a standard deviation of 18.9 m. This compares favourably with GEBCO_2014, which had a median offset of 4.6 m with a standard deviation of 73.4 m, and GEBCO_2019, which has a median offset of 0.5 m but a standard deviation of 89.4 m.