- SPECIAL REPORTS
- THE MAGAZINE
The U.S. Forest Service Forest Inventory and Analysis Program (FIA) is responsible for taking inventory of the forested lands of the United States, both inside and outside national forest boundaries. Historically, inventories were conducted and estimates produced about every five to 20 years through terrestrial, on-foot human measurements. However, the 1998 Farm Bill requires that each year, 10-20 percent of all U.S. lands must be measured, and forest population estimates must be updated. This increased frequency in inventory activity has raised costs for the U.S. Forest Service. As a result, the agency sought more efficient, cost-effective methods for executing the annual forest inventory. Members of the FIA program found that large-scale aerial photography, supplemented by traditional ground crew measurement, significantly reduced costs while maintaining a high level of accuracy.
Traditional Forest InventoryTraditional forest inventories are conducted by dispatching survey teams to each plot. These teams locate sample trees, take measurements of diameter at breast height (DBH) of each tree, and estimate tree height and population. Ground survey teams produce reliable and proven results with accuracy rates of approximately 90 percent. Many of the field plots are easy to visit, and plot tallies can be completed within a day. But some plots are located deep within wilderness areas, contain hazardous conditions or are geographically inaccessible. Using ground crew measurements requires significant resources, especially in these inhospitable locations. Several teams are needed to complete the surveys in a timely manner, and the Forest Service must make a capital investment to obtain and maintain equipment and to pay personnel.
Advances in Aerial PhotographyAerial photography has been employed in forest operations, such as for the direct measurement of trees and estimation of volume and growth, since the first half of the 20th century. In recent years, there have been significant hardware and software advances, including the development of high resolution aerial photography film, the ability to scan this film at approximately the same resolution as the film emulsion grain size, and now the development of full digital aerial cameras. With the advent of precision GPS, pre-specified locations on the ground can be captured with aerial photographs. The positional accuracy of aerial systems enables forestry data to be captured within a few meters, providing an accurate source for measurements such as ground coordinates, tree counts per area, tree height, visible crown diameter, tree species, crown class, tree condition (living or dead), ground layer structure, shape, and vegetation type and land cover.
Planning a StudyAerial photography has been used in forest inventories to lower the overall cost, but there have been concerns about whether the cost savings come at the price of reduced precision. The Forest Service Remote Sensing Applications Center (RSAC), an adjunct Forest Service office in Salt Lake City, conducted a case study on this issue with the help of two partners: Interior West FIA, a regional Forest Service office based in Ogden, Utah, and Red Castle Resources Inc., a privately owned small business in Salt Lake City that provides onsite technical staff to the Forest Service. Together, this team set out to determine the accuracy and cost effectiveness of using large scale digital aerial photos to sample FIA plots as part of its annual inventory. The project analyzed the precision of tree height measurements from aerial photos, and compared the cost of using ground crews to the cost of using aerial photos.
Study DetailsThe FIA study was conducted on plots in the Fishlake National Forest in the heart of the piÃ±on-juniper forest's range in central Utah. This forest was selected for the study because piÃ±on-juniper forests have been established as being ideal for using photo interpretation (PI) methods. Not only is the low tree density conducive to using PI, but the slow growth rate of these trees is well-suited to double sampling because the tree size changes relatively little over a 10-year period.
The aerial photos of the FIA plots were obtained with a traditional 12 inch aerial photo film camera and were required to meet specifications of sun angle, overlap and scale. Three photos were collected for each plot to allow stereo viewing from two different perspectives. The photos were scanned, and these soft copy stereo pairs were imported into Leica Geosystems' ERDAS IMAGINE, a comprehensive collection of photogrammetric software tools designed specifically to process imagery and extract data from images.
A block file was created for the images, using Leica Photogrammetry Suite (LPS) for image processing and digital photogrammetry. The seamlessly integrated suite of tools in LPS enabled the transformation of raw imagery into reliable data layers required for all digital mapping, GIS analysis and 3D visualization. ERDAS IMAGINE was used to process the images; the Stereo Analyst Add-On for ERDAS IMAGINE was used to view the images in stereo, and to establish tree height measurements.
MeasurementThe FIA plots consisted of four subplots systematically arranged around the location center.* Locating the plot location center on the aerial photo is the first step in any photo-based sampling. In the past, photo interpreters used hardcopy aerial photos, arranged for stereo viewing, to make measurements. Now, they can view and analyze soft copy scans of aerial photos, zooming into the stereo photos and capture very good detail at scales of 1:50. Once RSAC staff locate a tree on a newly acquired digital aerial photo, they can interpret its species and measure its height to the nearest decimeter.
Accurate ResultsThe results of the study on piÃ±on-juniper forests demonstrated that tree-height measurements made from aerial photographs are within the 10 percent error allowed of FIA ground crews. Working from high resolution images, a photo interpreter can easily do the job of many field crews, compiling forest measurements within a relatively large continuous area with less training and expense.
"The ability to measure a tree's height using Stereo Analyst is remarkably precise. While some differences occur, they are within an allowable 10 percent of true tree height as measured in the field," commented Kevin Megown of the U.S. Forest Service RSAC. "There are tremendous similarities between the ground height and photo-interpretation height measurements." Using Stereo Analyst to measure tree height precisely allowed the RSAC to accomplish the project's goal of recording inventory attributes in a more efficient manner.
Money SavingsAn analysis of cost versus precision for a piÃ±on-juniper forest inventory demonstrates that more than one-third of the sampling costs can be saved by combining photo analysis and ground-plot sampling methods without affecting accuracy. The true cost of a sample would include both the cost of photo interpretation and ground-based sampling required to perform a double sample. In this scenario, the first 30 plots require both ground sampling and aerial photography and setting a standardized control group, while subsequent plots may be recorded by photographs alone.
The difference in cost between measuring a plot on the ground, estimated at more than $1,500 per plot, and using photo-interpretation methods-approximately $510 per plot-provides a significant incentive for combined ground- and photo-based sampling. Analyzing 100 plots through combined ground survey and photo analysis resulted in a total savings of $54,000.
Reaping All the BenefitsWith a fixed budget, the time and money saved in taking inventory of some forests with aerial photography could be used to increase the amount of sampling in forests that are more difficult to assess or have a higher economic or ecologic value. The use of aerial photos can also foster a more spatial approach to generating information about forests. For example, building relationships between satellite data and the aerial photos could add another layer of efficiency to the inventory, while also providing more information about the distribution of the resource. By combining traditional ground crews and aerial photography, the Forest Service benefited from increased cost efficiency, while preserving accuracy and fulfilling its annual inventory requirements.
* A plot configuration consists of subplots of 4-foot to 24-foot radius. Each subplot is spaced 120 feet apart with plot 1 being the center and the three remaining subplots being 120, 240 and 360 degrees azimuth from plot 1 center. A location center is the center of plot 1.
Note: The Forest Service of the United States Department of Agriculture (USDA), has developed this information for the guidance of its employees, its contractors, and its cooperating federal and state agencies, and is not responsible for the interpretation or use of this information by anyone except its own employees. The use of trade, firm or corporation names in this document is for the information and convenience of the reader. Such use does not constitute an official evaluation, conclusion, recommendation, endorsement, or approval by the department of any product or service to the exclusion of others that may be suitable.
Manufacturer InformationLeica Geosystems Geospatial Imaging