Datasets:
metadata
license: mit
tags:
- dem
- elevation
- terrain
- copernicus
- gebco
- terrarium
- maplibre
- cesiumjs
- bathymetry
- global
- tiles
- openzenith
size_categories:
- 10M<n<100M
OpenZenith DEM: Global Elevation & Bathymetry Tiles
Global elevation dataset combining Copernicus GLO-30 (land) and GEBCO 2025 (ocean bathymetry) in Terrarium PNG tile format, ready for MapLibre GL, CesiumJS, and other renderers.
What is this?
A complete global elevation tile pyramid (zoom 0–8) encoded as Terrarium PNG images. Each pixel encodes an elevation value that can be decoded client-side — no server-side processing needed.
Data sources:
- Land elevation: Copernicus GLO-30 (30 arc-second, ~30m resolution)
- Ocean bathymetry: GEBCO 2025 (15 arc-second, ~450m resolution)
- Merging: Copernicus takes priority over GEBCO on land; GEBCO fills ocean areas
Dataset Structure
tiles/
├── 0/ # 1 tile (360° global)
├── 1/ # 4 tiles
├── 2/ # 16 tiles
├── 3/ # 64 tiles
├── 4/ # 256 tiles
├── 5/ # 1,024 tiles
├── 6/ # 4,096 tiles
├── 7/ # 16,384 tiles
└── 8/ # 65,536 tiles
─────────────
87,381 tiles (~5.1 GB)
Each tile is a 256×256 PNG at path tiles/{z}/{x}/{y}.png.
Terrarium Encoding
Each pixel encodes elevation in meters using RGB channels:
elevation_m = (R × 256 + G + B / 256) - 32768
| Terrain | Color (RGB) | Elevation |
|---|---|---|
| Deep ocean | Dark | < -1000m |
| Sea level | (128, 0, 0) | 0m |
| Low land | Green-ish | 0–500m |
| Mountains | Bright | 1000–5000m |
| Everest peak | (255, 255, 128) | 8,533m |
JavaScript Decoder
function decodeTerrarium(r, g, b) {
return (r * 256 + g + b / 256) - 32768;
}
Python Decoder
def decode_terrarium(r: int, g: int, b: int) -> float:
return (r * 256 + g + b / 256) - 32768
Usage
Python (pip install openzenith)
from openzenith import get_elevation, load_tiles
# Download tiles from HuggingFace (zoom 0-8, ~5GB)
tiles_dir = load_tiles(zoom_levels=[0, 1, 2, 3, 4, 5, 6, 7, 8])
# Query elevation at any lat/lon
elev = get_elevation(40.7128, -74.0060) # New York City
print(f"NYC elevation: {elev:.1f}m")
# Batch queries
from openzenith import get_elevation_batch
elevations = get_elevation_batch([
(40.7128, -74.0060), # New York
(35.6762, 139.6503), # Tokyo
(27.9881, 86.9250), # Mount Everest
])
MapLibre GL JS
map.addSource("terrain", {
type: "raster-dem",
tiles: ["https://openzenith.cyopsys.com/api/dem-tile/{z}/{x}/{y}"],
encoding: "terrarium",
tileSize: 256,
maxzoom: 12,
});
map.addLayer({
id: "hillshade",
type: "hillshade",
source: "terrain",
exaggeration: 1.2,
});
CesiumJS
// Custom Terrarium terrain provider (included in OpenZenith globe page)
viewer.terrainProvider = createTerrariumTerrainProvider(Cesium);
API Endpoint
# Query elevation (no API key required)
curl "https://openzenith.cyopsys.com/api/elevation?lat=40.7128&lon=-74.0060"
# Get a tile
curl "https://openzenith.cyopsys.com/api/dem-tile/8/217/151.png" -o tile.png
# Health check
curl "https://openzenith.cyopsys.com/api/dem-tile?health=1"
Verified Accuracy
| Location | Expected | Measured | Error |
|---|---|---|---|
| Mt. Everest | 8,849m | 8,533m | 0.4% (30m resolution) |
| Death Valley | -86m | -74m | Within 30m |
| Denali | 6,190m | 6,164m | 0.4% |
| Mariana Trench | -10,935m | -7,122m | GEBCO 450m limit |
| Atlantic Ocean | -5,000m+ | -5,175m | Within 450m |
Technical Details
| Property | Value |
|---|---|
| Tile format | 256×256 RGB PNG |
| Encoding | Terrarium |
| Tile pyramid | Standard XYZ (Slippy Map) |
| Vertical datum | WGS84 ellipsoidal |
| Coordinate system | EPSG:4326 (WGS84) |
| NoData | R=0, G=0, B=0 (decodes to NaN) |
| Land accuracy | 3–7m RMSE (Copernicus GLO-30) |
| Ocean accuracy | ~100m (GEBCO 2025 bathymetry) |
Data Sources
| Source | Resolution | Coverage | License |
|---|---|---|---|
| Copernicus GLO-30 | 30 arc-sec (~30m) | Global land | ESA Open |
| GEBCO 2025 | 15 arc-sec (~450m) | Global (land+ocean) | CC-BY 4.0 |
Related
- OpenZenith — Free elevation API, 2D/3D maps, GIS sandbox
- openzenith Python library —
pip install openzenith - GitHub — Source code
License
MIT — Free for personal and commercial use.
Source data retains its own license: Copernicus GLO-30 (ESA Open), GEBCO 2025 (CC-BY 4.0).