Grid Landscape Metrics

The Same Metrics, at Native Grid Resolution

The companion Regime Landscape Metrics study averages fragmentation metrics up to 5 city sectors. This one skips the averaging step: all 1,493 grid cells, individually, 2000–2035 — so within-district variation that a regime mean quietly erases is visible directly.
Python (pandas, GeoPandas) R (landscapemetrics, source pipeline) Leaflet
Setup

Grid, Not Regime

Four of the six regime-level metrics are grid-native and average cleanly: Shannon Diversity, Built-up Patch Density, Built-up Aggregation Index, and Vegetation Cohesion, plus Edge Density — each computed per cell directly from that cell's own cropped raster. (Largest Patch Index is deliberately excluded here: it's scale-dependent and only meaningful at the full regime scale — see the regime-level monitor for it.)

Grid cells1,493 MetricsSHDI, PD, AI, COHESION, ED Years2000–2035 (8 steps)
Open interactive grid-cell metrics monitor
01 — Why Look at the Grid

Regime Averages Hide Real Within-District Spread

South City's regime-level Patch Density in 2015 was a single number: 17.49. At grid resolution, the cells that make up South City range from 3.96 to over 2,000 — most of that range compressed into a modest, informative spread, with a handful of extreme values coming from tiny boundary-sliver cells where a small denominator inflates a density metric (a classic modifiable-areal-unit effect, not a real fragmentation signal). The interactive map's quantile-based color scale is robust to those outliers; hovering an individual cell shows its exact value regardless.

Line chart of citywide mean Patch Density and Edge Density across all 1493 grid cells from 2000 to 2035, both peaking around 2015-2020 then declining
Citywide mean Patch Density & Edge Density (all 1,493 cells) — same rise-then-predicted-decline pattern as the regime-level view, now traceable cell by cell

The citywide mean confirms the same story the regime view told — fragmentation rising through ~2015–2020, then predicted to consolidate — but every cell behind that mean is now individually inspectable on the map, rather than folded into one of five regime numbers.

02

Methods & Tools

Grid-Native MetricsSHDI, Built-up PD, Built-up AI, Vegetation Cohesion, Edge Density — computed per cell, not aggregated
ExcludedLPI (scale-dependent; see regime-level monitor instead)
CartographyQuantile-classed choropleth (outlier-robust), per-cell hover detail
Data QualityBoundary-sliver cells flagged as a known MAUP-type artifact, not filtered silently