{"success":true,"data":{"author":"University of Copenhagen GeoAI group (Zhang et al.)","bucketAccessType":"public","bucketListable":true,"bucketName":"us-west-2.opendata.source.coop","bucketPrefix":"geoai-ucph/gvsm","category":"Earth Observation & Geospatial","cloudProvider":"aws","createdAt":"2026-11-16T15:39:54.024Z","description":"The Vertical Vegetation Structure Model (VSM) maps the full GEDI relative height profile, not only canopy height, across global land at 10 m. A deep learning model trained on 178 million GEDI L2A lidar waveforms translates Sentinel-2 L2A imagery into 101 relative height metrics, RH0 to RH100 in decimetres, each with predictive quantiles for uncertainty (median Q1 available; 5th and 95th percentiles to follow). RH98-RH100 approximate canopy top height; RH50 and the lower metrics describe canopy layering, understory density and structural diversity for forest monitoring, carbon and biodiversity work. Layout: 2020/<MGRS tile>/RH{level}_Q{quantile}.tif at 10 m in each Sentinel-2 tile's UTM zone (10,980 x 10,980 px), plus mosaics/2020/ global overviews averaged to 0.01 degree in EPSG:4326 (36,000 x 13,889 px); an experiments/ folder holds supporting runs. Cloud Optimized GeoTIFF, Int16, lossless LERC+ZSTD; about 360 TB per year. 2024 is in preparation. Read with rasterio or rioxarray after picking tiles from the MGRS grid. Licensed CC BY 4.0; cite the EarthArXiv preprint (eartharxiv.org/repository/view/15380).","displayTitle":"VSM: Global Vertical Vegetation Structure Model, GEDI RH0-RH100 at 10 m (2020)","fuseMountBrowsable":true,"id":"4e24d485-b3e1-5b5a-9475-3380fe39570c","license":"CC-BY-4.0","licenseEvidenceUrl":"https://source.coop/geoai-ucph/gvsm","objectCountEstimate":3,"probeStatus":"ok","referenceUrl":"https://eartharxiv.org/repository/view/15380","region":"us-west-2","schemaCard":{"accessHints":[{"language":"python-anonymous","snippet":"import boto3\nfrom botocore import UNSIGNED\nfrom botocore.client import Config\ns3 = boto3.client('s3', region_name='us-west-2', config=Config(signature_version=UNSIGNED))\ns3.list_objects_v2(Bucket='us-west-2.opendata.source.coop', Prefix='geoai-ucph/gvsm/', Delimiter='/')"},{"language":"python-local","snippet":"open('/mnt/archive/2020/01GEL/RH0_Q0.tif', 'rb')"}],"caveats":["Files are binary — use a domain-specific parser (e.g. metpy / pynexrad / pydicom) rather than text I/O."],"format":"binary","generatedAt":"2026-11-16T15:39:58.292Z","layout":{"example":"geoai-ucph/gvsm/2020/01GEL/RH0_Q0.tif","pattern":"{prefix}/{seg1}/{YYYY}/{seg3}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"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","encoding":"hex","key":"geoai-ucph/gvsm/2020/01GEL/RH0_Q0.tif","truncatedAt":512},{"bytes":"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","encoding":"hex","key":"geoai-ucph/gvsm/2020/01GEL/RH0_Q1.tif","truncatedAt":512}],"sampleObjects":["geoai-ucph/gvsm/2020/01GEL/RH0_Q0.tif","geoai-ucph/gvsm/2020/01GEL/RH0_Q1.tif","geoai-ucph/gvsm/2020/01GEL/RH0_Q2.tif","geoai-ucph/gvsm/2020/01GEL/RH100_Q0.tif","geoai-ucph/gvsm/2020/01GEL/RH100_Q1.tif"],"sidecars":[{"bytes":"# VSM: A Vertical Vegetation Structure Model of the Earth\n\n<p align=\"center\">\n  <a href=\"https://huizhml.github.io/map-explorer/\">\n    <img src=\"https://raw.githubusercontent.com/huizhml/global-vegetation-structure/master/docs/assets/vsm_global_every2rhs_black_bg_v7.png\" alt=\"Global map of GVSM vertical vegetation structure, with RH layers stacked as a 3-D datacube (height in metres, 0–50)\" width=\"100%\">\n  </a>\n</p>\n\nThe VSM brings a vertical dimension to global vegetation mapping, extending beyond canopy height to characterzing vegetation structure at 10 m resolution. \nTrained with 178 million sparse [GEDI](https://gedi.umd.edu/) lidar observations, a deep learning model translates Sentinel-2 imagery into continuous maps of 101 relative height metrics (RH0–RH100), each with predictive quantiles that describe its uncertainty.\nBy capturing vegetation structure beyond canopy height, VSM supports applications in forest monitoring, biodiversity assessment, and ecosystem research.\n\n## Explore\n\n- 📄 **Preprint:** [eartharxiv.org/repository/view/15380](https://eartharxiv.org/repository/view/15380/)\n- 🗺️ **Interactive map:** [huizhml.github.io/map-explorer](https://huizhml.github.io/map-explorer/)\n- 💻 **Code:** [github.com/huizhml/global-vegetation-structure](https://github.com/huizhml/global-vegetation-structure) \n- 📦 **Storage:** `s3://geoai-ucph/gvsm/` on the endpoint `https://data.source.coop`\n\n## Key facts\n\n| | |\n|---|---|\n| Variable | Relative height profile (RH0 - RH100) , in decimeters |\n| Uncertainty | Median (Q1) available; 5th percentile (Q2) and 95th percentile (Q0) coming soon |\n| Resolution | 10 m (main product), 1 km (global mosaics) |\n| Coverage | Global land |\n| Years | 2020 (available), 2024 (in preparation) |\n| Input | Sentinel-2 L2A |\n| Reference data | GEDI L2A, 178 million height profiles |\n| Format | Cloud Optimized GeoTIFF, Int16, lossless LERC+ZSTD (tiles) / ZSTD (mosaics) |\n| Volume | ≈ 360 TB per year (full RH profile × 3 quantiles) |\n| License | CC BY 4.0 |\n\n## What is an RH metric?\n\nGEDI records laser energy reflected by vegetation and the ground within each footprint. **RHn** is the height above ground below which *n* % of that returned energy has accumulated.\n\n- **RH98–RH100**: approximate canopy top height, comparable to conventional canopy height products.\n- **RH50**: the height below which half of the total returned energy has accumulated.\n- **Low RHs (RH0–RH25)**: describe the lower portion of the cumulative return, influenced by both vegetation and ground returns.\n\nTaken together, RH0 … RH100 for a pixel make a cumulative vertical profile. You can use it to study canopy layering, understory density and structural diversity, not only the top height.\n\n<img src=\"https://raw.githubusercontent.com/huizhml/global-vegetation-structure/master/docs/assets/1a_vertical_structure.png\" alt=\"Global map of VSM vertical vegetation structure, with RH layers stacked as a 3-D datacube (height in metres, 0–50)\" width=\"100%\">\n\n## Data organization\n\n```\ngvsm/\n├── <year>/                      # acquisition year, e.g. 2020\n│   └── <MGRS tile>/             # Sentinel-2 MGRS tile ID, e.g. 32UNG\n│       ├── RH0_Q1.tif\n│       ├── RH1_Q1.tif\n│       ├── …\n│       └── RH100_Q1.tif\n└── mosaics/\n    └── <year>/\n        ├── RH0_Q1.tif\n        ├── …\n        └── RH100_Q1.tif\n```\n\nFile names follow `RH{level}_Q{quantile}.tif`, with `level` from 0 to 100 and `quantile` from 0 to 2.\n\n| Path | Content | Grid |\n|---|---|---|\n| `<year>/<tile>/` | Full-resolution predictions. One folder per Sentinel-2 MGRS tile, one file per RH level and quantile. | Native UTM zone of the tile, 10 m, 10 980 × 10 980 px |\n| `mosaics/<year>/` | Global overview mosaics (10 m tiles averaged to 0.01°), for browsing and continental- to global-scale analysis | EPSG:4326, 0.01°, 36 000 × 13 889 px |\n\nTo find the tile that covers your area, use the [Sentinel-2 MGRS tiling grid](https://sentiwiki.copernicus.eu/web/s2-products). Neighbouring tiles ove","encoding":"utf8","key":"geoai-ucph/gvsm/README.md","name":"README.md","size":8361,"truncated":true}],"sourceObjectKey":"geoai-ucph/gvsm/2020/01GEL/RH0_Q0.tif","starterCell":"with open('/mnt/archive/2020/01GEL/RH0_Q0.tif', 'rb') as fh:\n    head = fh.read(1024)\nprint(len(head), head[:64])","starterSamplePath":"2020/01GEL/RH0_Q0.tif","totalObjects":"3 top-level prefixes"},"summary":"The Vertical Vegetation Structure Model (VSM) maps the full GEDI relative height profile, not only canopy height, across global land at 10 m.","title":"VSM: Global Vertical Vegetation Structure Model, GEDI RH0-RH100 at 10 m (2020)","updatedAt":"2026-11-17T17:20:01.513Z","visibility":"public","websiteUrl":"https://source.coop/geoai-ucph/gvsm"}}