Elevation data is essential for GIS projects. The most widely used free open datasets are Copernicus GLO-30, NASADEM and ALOS AW3D30, mostly at 30-meter resolution. Google recently released GEM-15, a near-global elevation dataset at 15-meter resolution with promising quality. I will walk you through this dataset in this article.
Dataset Introduction
GEM-15 stands for Near-Global Surface and Terrain Elevation Models (15m). It is built by Google using sub-meter optical satellite imagery acquired up to 2024. The primary source is 30 cm Pléiades Neo imagery captured between 2022 and 2024, supplemented by stereo imagery from WorldView, GeoEye-1, Pléiades 1A/1B and 1.5 m SPOT.
Unlike conventional open-source DEMs that only provide a single elevation layer, this dataset contains multiple data components:
- dsm: 15 m Digital Surface Model, including height of man-made structures and tree canopies
- dtm: 15 m bare-earth Digital Terrain Model, predicted at 1.0 m resolution using deep learning
- dsm_percentile_5 / 50 / 95: grid elevation percentiles, suitable for canopy and building height structure analysis
- dsm_std: standard deviation of elevation, indicating local terrain variation and surface roughness
Each tile uses the local UTM zone projection. Elevation values are metric orthometric heights referenced to the EGM96 geoid (EPSG:5773).
Official page: https://developers.google.com/earth-engine/datasets/catalog/projects_climate-and-sustainability_assets_gem-15_v1?hl=en
Coverage and Access Methods
One important note: this dataset covers Taiwan but excludes other regions of China. The official documentation excludes mainland China, Russia, Mongolia, Greenland and Antarctica. Coordinate offsets have been applied near India’s borders to meet local compliance requirements. If you are working on projects within China, you may skip this dataset.

There are two main ways to obtain the data.
The first method is clipping and export directly through GEE. Load the asset projects/climate-and-sustainability/assets/gem-15_v1 in the Code Editor or Python API, clip it to your study boundary, and export the result as GeoTIFF. This approach works well for small-area on-demand extraction.
The second method is fetching raw files from Google Cloud Storage (GCS). The dataset is stored as Cloud Optimized GeoTIFF (COG) in the public bucket gs://climate-and-sustainability/gem-15, with a STAC catalog index. This is suitable for bulk download or streaming rendering directly in QGIS.
STAC catalog URL:
https://storage.googleapis.com/climate-and-sustainability/gem-15/stac/catalog.json
Community Tests
Many GIS practitioners have downloaded and tested the dataset immediately, and overall feedback is positive. @oHackenberga shared terrain rendering results for a part of Western Australia:

The improvement in detail is obvious. @fefejen3 imported the GeoTIFF and generated slope maps and hypsometric tint maps in QGIS.

@nomad_mapper loaded GEM-15 data for Banff, Canada with RustGIS, as shown below:

@ElisaurySuarez reconstructed the 3D topography of alluvial fans on the northern slope of Lake Enriquillo basin in the Dominican Republic using GEM-15. The geomorphic texture is clear, and the reconstruction helped identify high-risk zones for relocated communities.

It should also be noted that the source data is optical stereo pairs rather than LiDAR. Reconstruction accuracy degrades significantly beneath dense forest canopies or around buildings taller than 50 meters.
Summary
The open-source community has long lacked near-global elevation products at 15-meter resolution. For mesoscale terrain analysis and urban morphology extraction, practitioners often had to rely on 30-meter free DEMs or purchase expensive commercial high-resolution stereo survey data. GEM-15 fills this gap in resolution and timeliness.
Nevertheless, the dataset remains experimental. Constraints inherent to optical photogrammetry and explicit exclusion of China, Russia and other regions mean it cannot fully replace existing elevation data in the short term. It is recommended for multi-source comparison and supplementary analysis for overseas projects.
References
- Google Earth Engine Data Catalog, GEM-15 documentation: https://developers.google.com/earth-engine/datasets/catalog/projects_climate-and-sustainability_assets_gem-15_v1?hl=en
- Google Research, GEM-15 validation and benchmark repository: https://github.com/google-research/gem_15
- GEM-15 STAC catalog: https://storage.googleapis.com/climate-and-sustainability/gem-15/stac/catalog.json
Leave a Comment