Skip to content

GMTED2010_IMPLEMENTATION.md

Latest commit

 

History

History
98 lines (73 loc) · 2.94 KB

GMTED2010_IMPLEMENTATION.md

File metadata and controls

98 lines (73 loc) · 2.94 KB

GMTED2010 Downloader Implementation

Status: ✅ COMPLETE

GMTED2010 downloader has been fully implemented for resolutions >100m (250m, 500m, 1000m).

What Was Implemented

1. Direct Download URLs

  • Base URL: https://edcintl.cr.usgs.gov/downloads/sciweb1/shared/topo/downloads/GMTED/Grid_ZipFiles/
  • Pattern: {product}{arcsec}_grd.zip
    • Products: md (median - recommended), mn (mean), mi (minimum), mx (maximum), sd (std dev), ds (subsample), be (breakline)
    • Arc-seconds: 75 (7.5 arc-sec = 250m), 15 (15 arc-sec = 500m), 30 (30 arc-sec = 1000m)

2. Downloader Implementation (src/downloaders/gmted2010.py)

  • Downloads global grid (cached for reuse in data/.cache/gmted2010/)
  • Extracts ZIP (handles ArcGrid format)
  • Converts to GeoTIFF (if needed)
  • Clips to tile bounds (1×1 degree tiles)
  • Saves as standard tile format (N{lat}_W{lon}_{resolution}m.tif)

3. Integration

  • ✅ Integrated into source_coordinator.py (routing)
  • ✅ Integrated into source_registry.py (source definitions)
  • ✅ Integrated into orchestrator.py (resolution selection)
  • ✅ Removed temporary restrictions on resolutions

4. Resolution Support

  • 250m (7.5 arc-seconds) - GMTED2010
  • 500m (15 arc-seconds) - GMTED2010
  • 1000m (30 arc-seconds) - GMTED2010

How It Works

  1. First Download: Downloads entire global grid (~2.9GB for 250m) and caches it
  2. Subsequent Downloads: Uses cached global grid, clips to tile bounds
  3. Tile Generation: Creates standard 1×1 degree tiles compatible with existing pipeline

Usage

The downloader is automatically used when:

  • Region requires >100m resolution (based on Nyquist sampling)
  • Large regions (e.g., California) that need coarse resolution
  • System selects GMTED2010 as optimal source

Example:

python ensure_region.py california
# System automatically selects 500m or 1000m GMTED2010 if needed

File Structure

data/
  .cache/
    gmted2010/
      gmted2010_250m_global.tif   # Cached global grid
      gmted2010_500m_global.tif
      gmted2010_1000m_global.tif
  raw/
    gmted2010_250m/
      tiles/
        N40_W080_250m.tif         # Clipped tiles
    gmted2010_500m/
      tiles/
        N40_W080_500m.tif
    gmted2010_1000m/
      tiles/
        N40_W080_1000m.tif

Notes

  • Global grids are large (~2.9GB for 250m) but cached for reuse
  • First download takes time (download + extract + convert)
  • Subsequent tiles are fast (just clipping from cached grid)
  • ArcGrid format is automatically converted to GeoTIFF
  • Product selection: Uses md (median) by default (most commonly used)

Testing

To test with a large region:

python ensure_region.py california --force-reprocess
# Should automatically select GMTED2010 if visible pixels > 500m

Next Steps

  • ✅ Implementation complete
  • ⏳ Testing with large regions (pending user test)
  • ⏳ Performance optimization if needed