Mar 25 2020
11:30 am - 12:30 pm
Track Names: FORESTRY, Wednesday 11:30 - 12:30
Session Date: Mar 25 2020 11:30 am - 12:30 pm
Forest canopy structure characterization using high-density UAV LiDAR
Representation of the type, shape and composition of the canopy is a key component for estimation and modeling of forest land-atmosphere water and carbon budgets. The ability to gather multi-return LiDAR data allows the full detection of the forest structure from the top of the canopy to the understory vegetation. This study employed a UAS LiDAR system to acquire high density point cloud data of the dominant forest tree species in Northern Wisconsin. UAV-based LiDAR data was collected with a hexacopter equipped with a Lidarpod that includes a 32-laser head LiDAR with a GNSS RTK system. Point cloud density averaged 600 points per meter squared with a vertical accuracy of 2-5cm. In coordination with the NSF-funded CHEESEHEAD project, surveys were conducted within footprints of 9 flux towers and areas surveyed ranged between 0.25-1 squared km per site. We targeted 6 forest types including Aspen, Pine, Poplar, Larch, Cedar, Aspen and Hardwood and described their main canopy structure attributes. This study represents the first effort to characterize the forest canopy structure by tree type in Northern Wisconsin employing UAV LiDAR data. In addition, this study serves as a key baseline dataset for up-scaling ecosystem structure and modeling land-atmosphere gas interactions at a regional scale.
University Of Wisconsin Madison
LiDAR-Assisted Forest-Structure Products from Two Regions using Area-Based and Single-Tree Analysis Methods
Light detection and ranging (LiDAR) data are increasingly available on forested lands across the United States. These point-cloud data differ in acquisition hardware and collection specifications, as well as the rigor, timing, and availability of field data for comparison. To illustrate this, we compare the unpublished TREE-D LiDAR processing workflow and results for a ~23,000-acre site in east-central Arizona to a similar-sized site in eastern Washington State. The Washington data were typical for publicly available datasets in the western U.S., with LiDAR data from two separate acquisitions (2014 and 2015), unknown acquisition specifications, and no synchronized field sampling. The Arizona data were better suited for LiDAR-assisted inventory—LiDAR acquisition parameters were controlled and optimized for forest inventory, extensive field data were acquired in close temporal proximity to the LiDAR flights, and these field data included individual-tree locations to accommodate individual-tree based analysis methods. At both sites, using a mix of open-source and proprietary software, we were able to identify large, dominant trees in the point cloud data, and predict volume and species probabilities at a pixel level. At the Arizona site, using dense point cloud data (12-20 pts/m2) and fixed radius plot data, we were able to predict these area metrics with better precision, as well as confidently identify location, height, genus and DBH for individual trees. We discuss the value and limits of area-based LiDAR inventories, the potential and challenges associated with LiDAR-assisted individual-tree analysis, as well as data considerations for each approach.
Northwest Management, Inc
Optimizing LiDAR System Operation for Foliage Penetration
12:00 PM - 12:15 PM
There is a significant history of using LiDAR point cloud data for determining various forest parameters. Although there are some general rules-of-thumb for parameters such as off-nadir angle and laser footprint size when trying to maximize penetration through tree cover, there are more variables that can contribute to effective foliage penetration. A study has been undertaken to evaluate the use of several flight and system parameters that can now be varied with the introduction of newer megahertz-pulse-rate LiDAR systems. These variables can allow trade-offs between such things as pulse energy and pulse rate in an effort to maximize the number of hits on the forest floor, and thus obtain more detailed DEMs from which to evaluate forest parameters such as tree height. Results from recent testing by PASCO using Leica TerrainMapper will be compared to results from earlier-generation systems to reveal if the 'conventional wisdom' still applies, or whether operational parameters should indeed be adjusted to further maximize foliage penetration. Although the study is performed on a specific type of LiDAR system, the results can certainly be applied to other system types in the current generation of high-pulse-rate systems.
Q&A and panel discussions with session presenters
12:15 PM - 12:30 PM
There will be a 15 minute Q&A /panel discussion with the presenters of the Forestry session.