NYC Tree Genus Map

A citywide map of tree identity for New York City from a Cornell-led team (David Miller first author, Daniel Katz of Cornell's School of Integrative Plant Science senior author, with co-authors at Weill Cornell Medicine, the University of New Hampshire and the University of Oklahoma). It classifies about 1.8 million tree crown polygons into the 18 most common genera, which cover roughly 78% of the city's basal area. Unlike street-tree inventories it includes trees on private property and in natural areas, which the team estimates are 65% of the canopy and about 1.4 million trees that had never been surveyed.

Labels were extrapolated from the city's existing inventory to crown objects using an XGBoost classifier trained on time series of 3 m PlanetScope imagery and structural metrics from airborne lidar. Overall accuracy is 82%, higher for common genera such as plane and oak and lower for rare ones. The work is part of the Cool Trees project led by Arnab Ghosh; the stated aim is to find which trees cool best and to inform planting choices around allergens and pests, in support of the city's goal of 30% tree cover by 2040.

The paper is in Scientific Data (published 14 August 2026, CC BY-NC-ND 4.0). The dataset is on Zenodo, the classification code on GitHub, and the map can be browsed at nyctreegenusmap.com. The team has compiled a list of 405 US cities it hopes to map the same way.