Dortmund Regime Landscape Metrics Monitor

Class-level landscape-fragmentation metrics (FRAGSTATS-style: Patch Density, Edge Density, Aggregation Index, Cohesion, Largest Patch Index, Shannon Diversity) computed directly from classified/predicted land-use rasters, 2000–2035, and averaged by city-sector regime. 2000–2020 are observed; 2025–2035 are CA-CNN-LSTM model predictions, cross-checked against each regime's own historical rate envelope.

Metric
Year
Hover a regime for details.
Data & methodology ▾

Regimes: the same 1,493-cell grid and 5 city-sector regimes (Core/North/East/South/West City) used in the companion MGWR/SDM and GWR studies, this time aggregated to regime-mean landscape metrics rather than modelled as a regression outcome.

CodeMetric
PDPatch Density — patches / 100 ha (higher = more fragmented)
EDEdge Density — m of edge / ha (higher = more fragmented)
AIAggregation Index, 0–100 (lower = more dispersed)
COHESIONPatch Cohesion, 0–100 (higher = more physically connected)
LPILargest Patch Index, % of regime area in largest patch
SHDIShannon Diversity Index of land-use classes

2000–2020 from classified historical land-use rasters. 2025–2035 from the CA-CNN-LSTM land-use prediction pipeline.

Structural plausibility check: each regime/metric's predicted 5-year rate of change (2025→2030 and 2030→2035) is compared against that same regime/metric's own historical mean ± SD annual rate (2000–2020). A prediction outside that envelope is flagged, not hidden. ⚠ Edge Density is flagged for all 5 regimes in the 2025→2030 transition — the sharpest predicted swing in the dataset.

Regime polygons dissolved from the 1,493-cell grid, reprojected to WGS84.