The project combines high-resolution satellite imagery, airborne laser measurements and ground-level tree records to identify trees across streets, parks, private property and natural areas. About 1.4 million of the trees had not previously been covered by the city’s conventional street-tree surveys, significantly expanding knowledge of the composition of New York’s urban forest.
The classification system achieved an overall accuracy of 82%, identifying trees at the genus level across 18 common groups. Accuracy was higher for several widespread varieties. Plane trees, for example, could be identified with particularly strong accuracy, while oaks also produced comparatively reliable results.
The work addresses a longstanding limitation in urban forestry. Municipal inventories generally focus on trees maintained by public agencies, especially those lining streets and occupying managed parkland. Trees growing in back gardens, institutional grounds, privately owned land and semi-natural areas can therefore remain largely invisible to planners even though they contribute substantially to shade and overall canopy.
Previously unsurveyed areas account for an estimated 65% of New York City’s tree canopy. Mapping those trees individually gives researchers a far more detailed basis for examining how species, tree size, density and location influence temperatures at neighbourhood level.
The immediate focus is heat. Trees reduce exposure by shading buildings and paved surfaces while also cooling surrounding air through evapotranspiration. Those effects are increasingly important as New York experiences longer and more intense periods of high temperatures.
About 500 New Yorkers die each year from heat-related causes, including deaths where high temperatures worsen underlying illnesses such as cardiovascular disease. Climate projections indicate that the city could experience substantially more hot days and heatwaves later this century, strengthening the case for treating trees as part of essential climate-resilience infrastructure.
New York has set an overall goal of increasing its tree canopy from about 22% of land area to 30%. City legislation requires an urban forest plan to be updated every decade, with particular emphasis on expanding canopy equitably rather than concentrating planting in neighbourhoods already benefiting from mature trees.
The new dataset could help planners move beyond simply counting how many trees are planted. Researchers are examining whether specific tree types deliver different cooling benefits and how species selection can be matched to streets, building patterns and communities facing the greatest heat exposure.
Tree choice also carries consequences beyond temperature. Species vary in their ability to intercept air pollutants, absorb stormwater, tolerate drought and withstand pests. Some produce pollen associated with allergies, while others may be vulnerable to invasive insects or diseases. A detailed inventory can therefore support planting programmes that balance multiple environmental and health objectives.
The mapping method relies heavily on PlanetScope satellites, which repeatedly photograph the Earth at high resolution. Rather than depending on a single image, researchers analysed imagery collected over time. Seasonal changes provided additional clues: different tree groups develop leaves, change colour and lose foliage at varying times.
Airborne lidar added another layer of information. The technology measures reflected laser light to produce detailed three-dimensional representations of vegetation and surrounding structures. Combining those measurements with satellite observations enabled researchers to assess both the seasonal characteristics and physical structure of tree crowns.
Machine-learning techniques were then trained using existing inventories where tree identities were already known. The model extrapolated those classifications across the wider city, assigning likely genera to approximately 1.8 million individual crown outlines.
Because the satellite imagery used by the system is widely available across large geographic areas, the approach could be extended well beyond New York. Researchers have already assembled information covering more than 400 medium-sized and large US cities as potential candidates for similar mapping.
That scalability could be especially valuable for municipalities unable to conduct expensive ground surveys. Many cities have broad canopy maps showing where vegetation exists but lack information about individual trees and species. A remotely generated inventory could give planners a faster starting point for managing urban forests.
The New York work also overlaps with a separate initiative using artificial intelligence, aerial imagery, climate projections and three-dimensional modelling to determine where additional shade would most benefit heat-vulnerable communities. That project is concentrating initially on underserved areas of Brooklyn, Queens and the Bronx.
Follow Arabian Post
Select Arabian Post as your preferred source on Google and MSN News for trusted business news and Arab politics and updates.
