eCat: 147703

<BR>AusSeabed follows the recommended processing methodology as prescribed by the AusSeabed Multibeam Guidelines. Please note that certain products are subject to routine update (e.g. resolution, coverage, temporal variation, guidelines/best practice). The client is advised to refer to the latest version of all documents and products for the most up-to-date material. For eCat records the version is expressed in the record title whilst products incorporate date of production as per AusSeabed’s file naming convention.

<BR>GRID RESOLUTION: 0.001° (~100 m).

<BR>TOTAL SURFACE COVERAGE: 736,000 km2.

<BR>HORIZONTAL DATUM: WGS84.

<BR>VERTICAL DATUM: Approximates mean sea level.

<BR>USE LIMITATION: Not to be used for navigation.

<BR>DATA LINEAGE: This dataset is an output of the collaboration between James Cook University and Australian Government agencies, to process and compile all available digital bathymetry data and develop a regional-scale, 100 m-resolution Digital Elevation Model (DEM).

<BR>MBES SOURCE DATA: Geoscience Australia (GA) provided most of the foreign research vessel multibeam echo sounder (MBES) data which have been collected across the Kerguelen Plateau since 2004. CSIRO provided data from Australia’s Marine National Facility RV Investigator, which had conducted extensive systematic surveys in 2016 and 2020. The Alfred Wegener Institute (AWI) also conducted a survey in coordination with the RV Investigator 2020 survey. The French Service Hydrographique et Océanographique de la Marine (SHOM) provided multibeam data under licence from within their exclusive economic zone (EEZ) around Kerguelen Island. Extensive noise editing on the multibeam data were conducted using QPS Fledermaus and Caris HIPS&SIPS software and by applying sound velocity corrections where necessary, prior to bathymetry surface processing.

<BR>SBES SOURCE DATA: The Australian Hydrographic Office (AHO) supplied singlebeam echo sounder (SBES) data from a hydrographic survey conducted in 1998. The Australian commercial fishing company Austral Fisheries conducts numerous fishing voyages over the Kerguelen Plateau, authorised by the Commission for the Conservation of Antarctic Marine Living Resources (CCAMLR). Austral Fisheries vessels collect automatically logged, crowdsourced bathymetry (CSB) data from their GPS navigation sensors and depth sounders. The numerous track lines of Austral Fisheries’ CSB data reveal improved detail of the seafloor geomorphology over the plateau and adjacent basins. Extensive noise editing was conducted on SBES data using 3D point clouds generated within QPS Fledermaus software, prior to bathymetry surface processing.

<BR>COASTLINE SOURCE DATA: Coastline source data are used to ‘pin’ the bathymetry surface at the coast in order to prevent ‘bleeding’ of land into the water during the grid development phase. These coastline data were derived from Shuttle Radar Topographic Mission (SRTM)-Digital Surface Model (DSM) data for Kerguelen Island, McDonald Island and Shag Rock, and from Synthetic Aperture Radar (SAR) elevation data for Heard Island. Raster elevation data were converted to vector lines and compared with satellite imagery for anomalous vertices. Accepted vertices were then rasterised to ~30 m resolution grid pixels, with elevation values set at -2 m to approximate mean sea level (MSL). The ~30 m pixels were exported as xyz data for input to the bathymetry surface processing.

<BR>SRTM SOURCE DATA: The 3 arcsec (~83 m) resolution SRTM DSM data were used as land elevation data for the kerg100 surface. SRTM data were sourced for Kerguelen Island, McDonald Island and Shag Rock via the CGIAR Consortium for Spatial Information (CGIAR-CSI; Jarvis et al., 2008). The 83 m data were resampled to 100 m to match the interpolated bathymetry surface pixel size. During the grid development phase, the SRTM data were merged onto the interpolated bathymetry surface to complete the kerg100 grid.

<BR>SAR SOURCE DATA: The Australian Antarctic Division (AAD) supplied TerraSAR land elevation data for Heard Island collected in 2010. The ~10 m-resolution data were resampled to 100 m to match the interpolated bathymetry surface pixel size. During the grid development phase, the SRTM data were merged onto the interpolated bathymetry surface to complete the kerg100 grid.

<BR>GRID DEVELOPMENT 1: The grid development phase was conducted using Generic Mapping Tools (GMT) software (Wessel and Smith, 1991), following the methodology used in Becker et al. (2009). GMT is a Unix-based gridding and plotting software package that can deal with large datasets. This grid development phase is a ‘repair and replace’ method that is widely used for aggregating source bathymetry data for regional-scale and global-scale DEMs, e.g. SRTM30_PLUS.

<BR>GRID DEVELOPMENT 2: The ASCII xyz source data were first decimated using GMT blockmedian into individual xyz data files representing single node points at 50 m-resolution. The decimated data files were then concatenated into one large xyz file. Next, GMT blockmedian was conducted on the single large file to decimate the combined data to 100 m-resolution in order to produce one valid depth point for each pixel location to be used in the interpolated bathymetry surface at that same 100 m-resolution.

<BR>GRID DEVELOPMENT 3: The 100 m xyz data were then compared with co-located depths from an underlying base grid, in this case a ~500 m-resolution unpublished bathymetry grid developed by Geoscience Australia in 2005. The purpose of using a base grid was to flag any new data that may be greatly in error and thus be rejected, and also to provide underlying bathymetry data for pixels that lack coverage by the new source data. The ‘repair and replace’ method repairs the GA 2005 grid, replacing pixels with newer, higher-resolution data.

<BR>GRID DEVELOPMENT 4: A model was made with GMT surface using the difference values between the co-located new data and the underlying base data. GMT surface was used to resample the GA 2005 grid to 100 m-resolution. The difference grid and the resampled base grid were then added together with GMT grdmath. This netCDF file was converted into an ESRI raster grid. The final step was to merge the interpolated bathymetry surface with the SRTM and SAR land elevation data. The merged grid was then clipped to the project area to produce the final kerg100 depth model.

<BR>PROCESSING SCHEME: Please refer to the "Processing_scheme_kerg100.jpg" image in the metadata that shows the steps used to develop the kerg100 depth model.

<BR>DATA SOURCES: Refer to the metadata spreadsheet of source data used in compiling the kerg100 depth model: Datasources_kerg100_V3_DEM.xlsx.

<BR>ACKNOWLEDGEMENTS: The following people are gratefully acknowledged for their assistance during the development of the Kerguelen Plateau depth model:

Anne Worden (Australian Hydrographic Office)

Michael Andrew (Australian Hydrographic Service)

Mark Alcock (Geoscience Australia)

Mike Sexton (Geoscience Australia)

Phil O'Brien (Geoscience Australia)

Michele Spinoccia (Geoscience Australia)

Robert Parums (Geoscience Australia)

Kim Picard (Geoscience Australia)

Cisco Navidad (CSIRO)

Stewart Edwards (CSIRO)

Nelson Kuna (CSIRO)

Gabriele Uenzelmann-Neben (Alfred Wegener Institute)

Simon Dreutter (Alfred Wegener Institute)

Mike Coffin (IMAS, University of Tasmania)

Austral Fisheries

Henk Brolsma (Australian Antarctic Division)

Christelle Maillot-Dupas (SHOM)