markmclaren's picture
Update README with land.pmtiles attribution
b993f77 verified
---
license: cc0-1.0
tags:
- geospatial
- pmtiles
- bathymetry
- maps
- ocean
- hydrography
- isambard-ai
---
> ⚠️ **NOT FOR NAVIGATION.**
> This dataset is not intended for navigation, route planning, marine operations,
> legal compliance, or any safety-critical purpose. It is intended for
> visualisation, exploration, education, and research only. All data and software
> are provided "as is" without warranty of any kind.
A global bathymetric dataset optimised for web-native mapping, built as part of
a visualisation project about Caribbean Mermaids. The data is processed from
GEBCO source grids into PMTiles archives that stream efficiently to the browser
via HTTP Range Requests — no tile server required.
## Companion Repository
This dataset is part of the
[global-bathymetry-pmtiles](https://github.com/markmclaren/global-bathymetry-pmtiles)
project, which includes the processing scripts and a high-performance web-based
globe viewer with 2D/3D modes, real-time palette switching, and vertical
exaggeration controls.
## Available Files
- `gebco_2026_terrain_rgb.pmtiles`: GEBCO 2026 global bathymetry in **Terrain-RGB** format (6.0 GB).
- `gebco-2025-rawrgb-z0-6-webp.pmtiles`: GEBCO 2025 low-resolution overview in **Raw RGB** (WebP format, 204 MB).
- `land.pmtiles`: Land polygons derived from **Natural Earth** data (16.9 MB).
## Compute & Provenance
The processing of the GEBCO grids into PMTiles for this project was performed using the **Isambard-AI** supercomputing service.
> "This work used the Isambard-AI service (https://isambard.ac.uk) provided by the University of Bristol and funded by UK Research and Innovation (UKRI)."
**Technical Citation:**
McIntosh-Smith, S., et al. (2024). Isambard-AI: a leadership class supercomputer optimised specifically for Artificial Intelligence. [arXiv:2410.11199](https://arxiv.org/abs/2410.11199).
## Format & Technical Details
- **Format:** PMTiles, optimised for HTTP Range Request streaming in web
mapping clients.
- **Encoding:** RGB channels encode raw elevation and depth values in metres.
No colour palette is applied — styling must be handled client-side.
- **Storage:** Large files are managed via the Hugging Face CLI (`hf`) using
Xet storage, avoiding local Git LFS overhead.
## Data Attribution
Two GEBCO grid releases are used in this dataset:
- GEBCO Compilation Group (2025) GEBCO 2025 Grid
(doi:10.5285/37c52e96-24ea-67ce-e063-7086abc05f29) — zoom levels 0–6
- GEBCO Bathymetric Compilation Group (2026) GEBCO 2026 Grid
(doi:10.5285/4f68d5c7-45eb-f999-e063-7086abc036fa) — zoom levels 0–10
The GEBCO Grid is placed in the public domain and may be used free of charge.
This dataset does not claim any official status and is not endorsed by GEBCO,
the IHO, or the IOC. For more information visit [gebco.net](https://www.gebco.net/).
### Natural Earth Data
Land polygons are derived from [Natural Earth](https://www.naturalearthdata.com/).
"Made with Natural Earth. Free vector and raster map data @ naturalearthdata.com."
Natural Earth data is public domain.