Mapping New York City’s Trees: The Need for LiDARMapping New York City’s Trees: The Need for LiDAR

A B

A B C

D E

Figure 1. (A) Imagery used for 2001 tree-canopy mapping; and (B) 2001 tree-canopy data superimposed on higher-resolution imagery acquired in 2008. Although

large forest patches were correctly identified (lower portion of frame), individual trees and small forest patches were poorly mapped. These errors were caused by a

combination of severe shadowing in the 2001 imagery and use of less-sophisticated analytical techniques that could not fuse LiDAR with traditional imagery.

2001 Tree Canopy

J a r l a t h O ’ N e i l - D u n n e ▪ U n i v er s i t y o f V e rm o n t/ U S D A F o r es t S e r v i ce ▪ j o n e i l d u @ u v m . ed u ▪ 8 0 2 . 6 5 6 . 3 3 2 4

Figure 2. 3D (A, B, and C) and 2D (D and E) perspec-

tives show the benefits of using LiDAR to detect trees

obscured by shadows. Arrows point to select trees.

Trees in shadowed areas are nearly impossible to see

in a 3D view created by overlaying imagery with

LiDAR data (A). When LiDAR data are viewed alone,

trees can be observed, albeit with some difficulty (B).

When enhancement algorithms are applied to LiDAR

data, trees become clearly visible (C). A 2D perspec-

tive for the same area further illustrates the difficulty

of detecting trees in building shadows when using

imagery alone (D). A 2D surface model generated

from LiDAR depicts the height of features relative to

the ground, providing a way to “see though” shad-

ows created by tall urban buildings (E).

An accurate accounting of tree canopy in New York City is required to support a range of management, programmatic, and scientific objectives. The last tree-canopy map, completed in 2001, while cutting-edge at the time, has known accuracy issues. A new tree-canopy map is currently under development with funding from the USDA Forest Service. The use of recent (2008), higher-resolution imagery in conjunction with more advanced algorithms will address some of the accuracy issues present in the 2001 dataset, but a key limitation of relying on imagery to map tree canopy in cities remains: the obscuration of trees by building shadows. In the spring of 2010, LiDAR was acquired for New York City. Unlike im-aging cameras, which rely on reflected light, LiDAR sensors emit their own energy in the form of a laser. The advantage of LiDAR is that it essen-tially sees through shadows. Incorporating LiDAR into the current tree-canopy mapping efforts will improve the ability to detect trees, particular-ly smaller, recently-planted trees, resulting in a more accurate and visually-coherent representation of New York City’s tree canopy.

2001 Source Imagery2001 Source Imagery2001 Source Imagery

3D LiDAR Fused with Imagery3D LiDAR Fused with Imagery3D LiDAR Fused with Imagery

ImageryImageryImagery---Based Tree Canopy MappingBased Tree Canopy MappingBased Tree Canopy Mapping

3D LiDAR Elevation3D LiDAR Elevation3D LiDAR Elevation 3D LiDAR Enhanced 3D LiDAR Enhanced 3D LiDAR Enhanced

2D Imagery2D Imagery2D Imagery 2D LiDAR Height Above Ground2D LiDAR Height Above Ground2D LiDAR Height Above Ground