Compiled Ship-track Bathymetry around Japan and the Northwest Pacific, 2026

Compiled Ship-track Bathymetry around Japan and the Northwest Pacific, 2026 CSTB Japan and NW Pacific 2026 is a 3-arc-second bathymetric compilation of ship-track echosounder data covering 120°E–172°E, 3°S–55°N. It is built exclusively from openly published cruise and survey data, and every grid cell carries an identifier of the cruise it came from, so the compilation can be cited down to the individual survey. An interactive web map with terrain renderings, click-to-query per-cell source attribution, and direct downloads is available at https://gis.ikgm.net/cstb/. Source data JAMSTEC DARWIN — multibeam and single-beam bathymetry published from JAMSTEC research cruises, 1998–2025.NOAA NCEI / IHO Data Center for Digital Bathymetry — bathymetry from international cruises and surveys in the same region.GEBCO 2026 Grid — used only for the gap-filled variant and for quality-control comparisons; no GEBCO values enter the ship-track-only grids. 1,499 distinct cruise/survey layers contribute cells to the final grid. The accompanying reference tables list every cruise that was evaluated — 1,545 JAMSTEC cruises with dataset DOIs and ready-made citations, and 259 NCEI/DCDB surveys — including the small number that were evaluated and rejected, which are retained in the tables and marked considered_but_excluded together with the reason. Processing (version v14, 2026-07-31) Soundings from each cruise were gridded onto a common 3-arc-second geographic grid (WGS 84). Where several cruises cover the same cell, a ranked mosaic resolves the overlap: newer surveys take precedence, moderated by per-tile quality scoring that pushes locally degraded data (weather-affected segments, sparse single-beam crossings) behind better neighbours. Quality screening combined grid-wide residual analysis against GEBCO 2026, cross-correlation tests for horizontal mis-georeferencing, and visual inspection. Cruises with constant-value artefacts, wholesale positional or depth offsets were excluded entirely; locally damaged segments of otherwise healthy cruises were removed by rectangular masks; noisy swaths were despiked and median-smoothed; swath edges were trimmed; residual gates, a land mask and impossible-depth caps were applied throughout. Small gaps between adjacent survey lines were interpolated and are flagged as such in the source-id raster. Everything else in the ship-track grids is measured data. Files (all rasters are cloud-optimized GeoTIFF, ZSTD-compressed, EPSG:4326, 62,400 × 69,600 cells; depths in metres, negative below sea level) cstb_japan_nwpac_2026.tif — main grid, JAMSTEC + international cruises (float32; NoData away from ship tracks)cstb_japan_nwpac_2026_jamstec_only.tif — JAMSTEC cruises only (float32)cstb_japan_nwpac_2026_gebco_filled.tif — full-coverage variant with gaps filled from GEBCO 2026 (int16)cstb_japan_nwpac_2026_source.tif and the *_source.tif companions — per-cell source id (uint16), one id per contributing cruise, with dedicated ids for interpolated and GEBCO-filled cellsjamstec_darwin_bathymetry_references.csv / ncei_dcdb_bathymetry_references.csv — the id-to-cruise lookup with vessels, years, DOIs and citation strings License and attribution Free to use with attribution. Please credit this dataset and cite the cruises you rely on — the per-cell source raster and the reference tables exist to make that practical. As a derived compilation, the data also carry the terms of their providers (JAMSTEC and NOAA NCEI / IHO DCDB), which differ by source. Acknowledgements The author wish to thank Tomoki Joichi, for his effort to find defects and artifacts in the data. His feedback improved the data quality considerably.