Argus docs
All pages

Data sources for atlas layers and DEMs

argus-atlas 0.3.2.dev6+g77cc8fd @ 77cc8fdPyPI (opens in new tab)Source (opens in new tab)Changelog (opens in new tab)

Imported from docs/data-sources.md on argus-atlas's dev branch at that commit. Fixes go upstream, not here.

Anything OGR can read works as a controls --layer name=path input, in any CRS; anything GDAL can read works as --dem. This page is the shortlist of where to get such files, with what each one is good for in atlas. Compiled from a thread by Mashford Mahute (X, 18 Oct 2026) plus the sources already in use here.

In use

SourceWhatUsed as
Copernicus GLO-30 on AWS, https://copernicus-dem-30m.s3.amazonaws.com/30 m global DSM, 1x1° COG tiles, no login; sea at exactly 0 margus-atlas dem --bbox … --out dem.tif fetches and clips it, then --dem (docs/your-first-bake.md); --sea-level 0 derives the sea mask from it. Attribution required (see examples/README.md).
OpenStreetMap via Overpass, https://overpass-api.de/api/interpreterLive OSM features for a bboxargus-atlas osm / POST /osm: water, rivers, roads, rail, forest, urban, buildings, glaciers. Rate-limited: one call per AOI, pause between calls, reuse extracts. No ocean polygons (coastline is a line), hence the DEM sea mask.
OpenFreeMap, https://tiles.openfreemap.org/styles/libertyVector basemap, no keyThe viewer and flow pages' basemap.

OSM without Overpass

SourceWhatFor atlas
Geofabrik, https://download.geofabrik.deOSM extracts by continent/country/region, PBF and shapefile, refreshed dailyBig or repeated AOIs: download a region .osm.pbf once and point osm.extract_layers() / ogr2ogr at it instead of hitting Overpass. GDAL's OSM driver reads PBF directly, and labels --source region.osm.pbf reads its place=* nodes.
OSM itself, https://openstreetmap.orgThe data and its ODbL licenceAttribution for every layer that came from it.

Water, land and coast

SourceWhatFor atlas
HydroSHEDS, https://hydrosheds.org/productsRivers (HydroRIVERS), lakes (HydroLAKES), basins, with Strahler order and discharge; shapefileGlobal rivers/water layers with scale filtering by stream order, more consistent than OSM in sparsely mapped areas.
Natural Earth, https://naturalearthdata.com1:10m/50m/110m land, ocean, lakes, rivers, coastline, boundaries, places; shapefile/GeoJSON, public domainA land polygon for coasts (--land) at zooms up to ~9; too coarse for the 1024 px canvases used so far. Also glaciers, urban areas, boundaries. Populated places (ne_10m_populated_places) are a labels --source, ranked by SCALERANK.
USGS EarthExplorer, https://earthexplorer.usgs.govSRTM 1-arc-second, 3DEP (USA, down to 1 m), hydrography, boundaries; loginAlternative DEMs; 3DEP for US AOIs finer than 30 m.
Koordinates, https://koordinates.comMarketplace of government open data: LiDAR DEMs (NZ, parts of AU), coastlines, cadastre, imageryFine DEMs for Australian/NZ AOIs (e.g. 1–5 m state LiDAR where published) and local coastline/water layers.

Land cover and forest

SourceWhatFor atlas
Global Forest Watch, https://globalforestwatch.orgTree cover, loss/gain (Hansen), rasters and vectorsA forest layer where OSM landuse is thin; vectorise or threshold the raster first.
European Environment Agency, https://eea.europa.euCORINE land cover (Europe), Natura 2000, elevationforest/urban/water classes for European AOIs from one consistent product.
FAO GeoNetwork, https://fao.org/geonetworkAgriculture, soils, land cover, fisheriesLand-cover classes for non-European AOIs.
WRI Open Data, https://wri.org/dataEnvironment and natural resources (Aqueduct water risk, forests, land)Thematic layers for themed presets.
UNEP Environmental Data Explorer, https://uneplive.unep.orgEnvironmental indicators and spatial datasetsThematic layers.

Boundaries, population, infrastructure

SourceWhatFor atlas
GADM, https://gadm.orgAdministrative boundaries, levels 0–5, shapefile/GeoJSON/GPKG; free for non-commercial useA borders layer (burnt as a line mask; add it to LINE_LAYERS or give it a palette entry) for political-map presets.
DIVA-GIS, https://diva-gis.orgCountry packages: boundaries, roads, rail, water, population, climateOne-stop country layers where OSM is thin.
HDX, https://data.humdata.orgHumanitarian datasets: population, settlements, infrastructure, boundariesSettlement points/polygons for an urban layer in sparsely mapped regions.
NASA SEDAC, https://sedac.ciesin.columbia.eduGridded population (GPW), infrastructure, environmentPopulation rasters as an urban proxy.
Africa GeoPortal, https://africageoportal.comBoundaries, land cover, infrastructure across AfricaRegional layers for African AOIs.

Geocaches

Geocaching.com is not a source (#10 (opens in new tab)). The Australian open listings are.

SourceWhatFor atlas
Geocaching Australia, https://geocaching.com.au/Volunteer listing site, unaffiliated with Groundspeak. OKAPI-shaped JSON API at /api/services/. Uploaded content is CC BY-NC-SA 2.5.A point layer or hunt waypoints, non-commercial use only. opencaching.allow_reuse("au") is the gate: it records the attribution and refuses a commercial reuse or a paid app. The API wants a consumer key and a logged-in member; anonymous calls are refused, and a bulk dump is arranged with the site administrators rather than pulled through the API.
Geocaching.com (Groundspeak)Proprietary listings. Partner API under a separate licence.Out. No public geodatabase, and the partner terms do not allow an anonymous copy on a baked map.

Notes

  • Licences differ: OSM is ODbL (attribution + share-alike for derived databases), Natural Earth is public domain, Copernicus DEM and GADM have their own terms, Geocaching Australia listings are CC BY-NC-SA 2.5. osm.json and job.json record provenance; keep the attribution with any published tileset. A Geocaching Australia extract has to keep the same licence (share-alike) and stay non-commercial, including a hunt that charges a fee.
  • A new layer name outside COMPOSITE_PALETTE is burnt as a mask and reachable in a template as $<name>; to paint it in the composite add a colour to the palette in controls.py.