Sizing it Up: Using LiDAR Imaging to Map Forest Structure at the Altona Flat Rock
James Wholey, Mark Lesser, and Mark Barran
Background
The Altona Flat Rock, is a globally rare sandstone pavement pine barrens ecosystem in northeastern New York dominated by Pinus banksiana (Jack Pine), a fire-dependent serotinous species, with an understory composed primarily of Vaccinium angustifolium (Lowbush Blueberry) and Gaylussacia baccata (Black Huckleberry). The Flat Rock barrens is surrounded by northern hardwood forest making it a unique habitat type for both wildlife and bird species. Within the pine barrens, wildfires have occurred in different areas in 1919, 1940, 1957, and 2018, which, along with silvicultural practices in 1998, have created a mosaic of stand ages (104, 83, 66, 5, and 25 years old, respectively), and corresponding structural attributes, across the ecosystem.
Objectives
Our objectives were to use ESRI ArcPro to generate a canopy height model (CHM) using LiDAR imagery NYS GIS clearinghouse. We then plan on checking the accuracy of the CHM using tree height measurements from 60 plots spanning the study area. We will then use this CHM to asses the the accuracy of the previously used kriging models.
Methods
Methods – Canopy Height Model
•LiDAR data were obtained from the NYS GIS clearinghouse
•Used Esri ArcPro to create a digital surface model (DSM) and a digital terrain model (DTM) at 1m resolution
•The DSM was subtracted from the DTM to create the canopy height model (CHM)
•Field points were placed over the CHM to compare values
​
Methods – Kriging Accuracy
•CHM model was subtracted from the previously made kriging model
•The difference was reclassified into bins to find the percent accuracy of the model
Results
Figure 1. Over 60 percent of the CHM study area was within 3m of the ground truthed plots
​
Figure 2. A majority of pixels in the kriged model were underestimated compared to the CHM
​
Conclusion- Lidar generated heights were highly accurate based on ground truthing. LiDAR generated heights were best at estimating maximum and mean heights. Underestimation by the kriged model is due to field plots not capturing fine-scale heterogeneity. This is most commonly caused by intrusion of hardwood stands into pine forests.
Canopy Height Model (CHM)
Krig model interpolated from 60 field plots
Accuracy assessment of krig model based on CHM
Figure 1
