GEM-15: Google Elevation Models at 15 m, Near-Global Optical Stereo Digital Surface Model (DSM) and Digital Terrain Model (DTM)

Provider
Google Research (Batchu et al. 2026)
License
CC-BY-4.0 (source)
Format
JSON, TIFF (in sampled files)
Stored at
Google Cloud Storage, gs://climate-and-sustainability/gem-15 (us)
Access
Free and public. No account with the data provider is needed.
DOI
10.31223/X55N75
Category
Earth Observation & Geospatial
Source
https://doi.org/10.31223/X55N75

About this dataset

GEM-15, Google Research's experimental near-global elevation dataset: a Digital Surface Model (DSM) and a co-registered bare-earth Digital Terrain Model (DTM) at 15 m grid spacing from 84 N to 61 S, built from sub-meter optical stereo satellite imagery rather than radar or lidar, so it resolves buildings and tree crowns and gives height above ground as DSM minus DTM. The bucket holds 13,485 one-degree cells as 26,970 Cloud Optimized GeoTIFFs, about 12.3 TB, in / /dsm/zall/mosaic.tif and .../dtm/zall/mosaic.tif (for example m10_m100/wm91sm1/), each in its UTM zone; the DSM has five Float32 bands and the DTM one band (mean of the 2.5 m native DTM in each 15 m cell), nodata NaN, with internal overviews. A STAC collection is at gem-15/stac/gem-15/collection.json, and a Source Cooperative mirror (portolan-mirrors/gem-15-catalog) adds a STAC GeoParquet cell index and terrain tiles. Read with GDAL /vsicurl/, rasterio or a VRT over the cells. Licensed CC BY 4.0 with the attribution "This dataset is produced by Google"; cite Batchu et al. (2026), doi:10.31223/X55N75.

How to access this dataset in Python

import gcsfs
fs = gcsfs.GCSFileSystem(token='anon')
fs.ls('climate-and-sustainability/gem-15/', detail=False)[:10]

Lists the files with anonymous access. Runs on any machine, with no cloud account.

Data structure

File types in sampled files: JSON, TIFF

Files

Example file: gem-15/m10_m100/wm91sm1/dsm/zall/manifest.json

Analyze this dataset without downloading it

On Scigantic, GEM-15: Google Elevation Models at 15 m, Near-Global Optical Stereo Digital Surface Model (DSM) and Digital Terrain Model (DTM) opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.

This page is also available as Markdown and as JSON.

Related Earth Observation & Geospatial datasets