How this map was built

Heights are measured, not guessed. Every tree here comes from USGS 3DEP airborne LiDAR flown over Pima County in 2021 (AZ_PimaCo_2_2021, ~10 points per square metre). For each 250 m tile the ground surface is modelled from ground-classified returns, the canopy surface from the remaining returns, and the difference is a canopy height model at 0.5 m resolution. Peaks in that model become individual trees. Buildings are excluded using the survey's own building classification, and pole-like or flat-topped shapes are rejected so masts and ramadas don't get counted as trees. Power lines are caught by their return pattern (a laser pulse passes partly through a wire, so wires come back 97–99% multi-return against 52–79% for real canopy), cliff faces by the ground relief under the crown, and stadium roofs, silos and stacks by being far too narrow for their height. About 3.4 million trees are mapped across 1,600 km².

Every tree, not just public ones. The map covers all developed and green land in the metro — parks, riparian corridors, schools, golf courses, cemeteries, church grounds, and residential and commercial property. Each tree is tagged with the land it stands on so you can filter to what you can actually reach.

Access labels are a guess, not permission. They come from what a land polygon is tagged as in public map data, not from any ownership record. "Public" means the tree falls inside something mapped as a park — boundaries are approximate and often out of date. A quarter of trees sit on land with no mapping at all and are marked unmapped. Never read a label here as permission to enter or climb.

Species are inferred, and often uncertain. LiDAR cannot see what species a tree is. Species labels come from research-grade iNaturalist observations near each crown, weighted by distance and rejected when the species is too short to plausibly be that tree. Confidence is shown on every tree:

• observed — a verified record within 50 ft of the crown
• likely — a record within 150 ft, dominant among nearby records
• possible — inferred from the surrounding area only

Treat anything below "observed" as a hint about what grows there, not an identification of that particular tree.

What the climb score means. A 0–100 rating of how good a tree is to climb, built from four things measured or inferred from the air:

• Height — more is better, up to a plateau around 70 ft (taller than that doesn't make a tree more climbable).
• Crown width — a broad crown means real limb structure to move on.
• Crown fullness — rules out spindly spikes with little to hold.
• Species — wood strength and branching habit: sycamore, oak, cottonwood and walnut score high; palms and thorny mesquite, palo verde and acacia are capped low.

It is computed from aerial data alone. It knows nothing about deadwood, rot, cavities, bees, fences, or whether the branches start 12 ft up a bare trunk — so treat it as a way to find candidates, not a verdict.

Before you climb. A high score is not a safety assessment — inspect any tree from the ground yourself. Most trees here stand on private property: get the owner's permission. Even on public land, parks often have their own rules about climbing, and many Pima County riparian parcels are privately owned despite appearing in the habitat layer. Check locally before you go up.

Sources: USGS 3DEP LiDAR via the AWS public EPT archive · Pima County GIS (parks, riparian habitat) · OpenStreetMap · iNaturalist · imagery from Esri and USGS.