The Dataset Viewer has been disabled on this dataset.

Global 70m Sea Level Rise Vulnerability Layer

A planet-wide vector layer of all land at or below 70 meters elevation — the upper-bound geographic envelope for maximum-possible long-term sea level rise under multi-millennial-scale ice loss.

Distributed as a single PMTiles archive suitable for direct use in MapLibre GL, deck.gl, Leaflet, QGIS, and other modern geospatial tools that support HTTP range requests.

File

File Description
world-70m.pmtiles The full vector tileset, z0–z12, layer vulnerable_70m

Quick start

Web viewer (MapLibre + pmtiles.js)

<script src="https://unpkg.com/maplibre-gl@4/dist/maplibre-gl.js"></script>
<script src="https://unpkg.com/pmtiles@3/dist/pmtiles.js"></script>
<link href="https://unpkg.com/maplibre-gl@4/dist/maplibre-gl.css" rel="stylesheet">
<div id="map" style="position:absolute;inset:0"></div>
<script>
  const proto = new pmtiles.Protocol();
  maplibregl.addProtocol('pmtiles', proto.tile);
  const map = new maplibregl.Map({
    container: 'map',
    style: {
      version: 8,
      sources: {
        vulnerable: {
          type: 'vector',
          url: 'pmtiles://https://huggingface.co/datasets/nelsn/global-70m-slr/resolve/main/world-70m.pmtiles',
        },
      },
      layers: [{
        id: 'vulnerable-fill',
        type: 'fill',
        source: 'vulnerable',
        'source-layer': 'vulnerable_70m',
        paint: { 'fill-color': '#1f2933', 'fill-opacity': 0.82 },
      }],
    },
    center: [0, 20],
    zoom: 1.5,
  });
</script>

Desktop (QGIS)

QGIS 3.32+ supports PMTiles natively. Add a new "Vector Tiles" connection pointing at the file URL and use layer name vulnerable_70m.

Command line (tippecanoe / pmtiles inspector)

brew install pmtiles
pmtiles show https://huggingface.co/datasets/nelsn/global-70m-slr/resolve/main/world-70m.pmtiles

How it was built

  1. Streamed all ~26,000 land tiles from the Copernicus GLO-30 DEM bucket on AWS Open Data
  2. Thresholded each tile at 70 m elevation in memory, masking ocean nodata
  3. Polygonized the binary mask via gdal.Polygonize directly to per-tile GeoJSONSeq chunks
  4. Concatenated all chunks and built a single PMTiles archive with tippecanoe (-Z0 -z12 --coalesce-densest-as-needed --drop-densest-as-needed)

The full streaming pipeline is implemented as a handful of shell + Python scripts and runs end-to-end on a single machine in roughly a working day.

Layer schema

Field Type Description
v integer Always 1. Indicates ≤ 70 m elevation.

The layer contains only the vulnerable polygons (no "safe" features). Areas not covered by a polygon are either above 70 m or outside the Copernicus land coverage (i.e. present-day ocean).

Why 70 m?

70 meters bounds the realistic upper envelope for sea level rise under sustained warming on a multi-millennial timescale. Under such scenarios the Greenland and West Antarctic ice sheets are lost early, with the East Antarctic core contributing the remaining ice over thousands of years. See Now City — Max SLR for the underlying rationale and source citations.

Resolution

The PMTiles archive is built to zoom level 12, the natural cap for the Copernicus GLO-30 source (30 m / 1 arc-second). At z12 each Web Mercator pixel at the equator is ~38 m, very close to one source pixel. Past z12 you would only see the 30 m raster grid as staircase artifacts.

Known artifacts

  • Inland water: large lakes and reservoirs at or near sea level are included in the layer (Caspian, parts of the Great Lakes, etc.). They are real low-lying areas but disconnected from the rising ocean.
  • Real inland depressions: Death Valley, the Dead Sea, Qattara Depression, the Netherlands' drained polders all show as vulnerable because they are below 70 m.
  • River corridors: the Amazon and Mississippi basins extend the layer far inland. Under multi-millennial sea level rise these corridors do become marine inlets, so the layer is arguably correct on this point.
  • Coastal slop: Copernicus tiles extend a few pixels offshore at the coast; expect the polygon to bleed slightly into the present-day ocean.

A future revision will optionally subtract Natural Earth land/ocean polygons for users who want a strictly land-only layer.

License

Source data: Copernicus GLO-30 DEM (© ESA, distributed under the Copernicus license, free reuse).

Derived layer: CC-BY-4.0. Cite as:

Nelson, N. (2026). Global 70m Sea Level Rise Vulnerability Layer. HuggingFace Datasets, nelsn/global-70m-slr. Derived from Copernicus GLO-30.

Related

Downloads last month
79