Per-tree leaf area index mapping of Amsterdam’s unrecorded shade canopy

Théo Alessandro Hermann, Michiel van Selm, Titus Venverloo, Fabio Duarte, Carlo Ratti

Most of the trees shading Amsterdam are absent from any municipal record. Using the Dutch national LiDAR survey (AHN5), we derive per-tree Leaf Area Index for nearly 850,000 trees across 243km2. Roughly 62% of the canopy is unrecorded, and these unregistered trees carry a disproportionate share (about two-thirds) of the city’s gross shade, so most of this shade-providing canopy lies outside the systems meant to manage it. Uniform-transmissivity models correctly diagnose where shade is scarce (ρ = 0.997) but not what produces it: measured per-tree canopy transmission averages several times the single value they assume (bounded ≈2–4×), and the densest genus casts 11% more shade per crown than the sparsest. Within-neighbourhood density variation adds up to 5.7 percentage points of shade (2.8 on average) that averaged models erase. Ninety-five percent of neighbourhoods fall below the 30% canopy-cover target. Comparable national LiDAR across Europe and the US suggests the method is transferable, pending validation at other point densities and canopy types.