Data & methodology ▾
Grid: the same 1,493-cell analysis grid used across this repository's Dortmund case studies (Metrics.shp), each cell's metric computed directly from that cell's own cropped/masked raster — not aggregated to the 5 city-sector regimes.
Why no LPI here: Largest Patch Index is scale-dependent (% of landscape held by the largest patch) — averaging or computing it at 500 m grid-cell scale washes out true regional dominance and produced a near-flat, uninformative series in testing. LPI is only meaningful at the full dissolved-regime scale; see the regime-level monitor for it.
2000–2020 from classified historical land-use rasters. 2025–2035 from the CA-CNN-LSTM land-use prediction pipeline.
Grey cells have no value for the selected metric/year (raster gap or a boundary sliver too small to classify).
Hover any cell to see its Stadtbezirk, regime, and all 5 metrics at the selected year at once.