Faced with limited funds and resources, how do land and resource managers prioritize restoration efforts that cover millions of acres of public and private land? It’s a question that’s been asked repeatedly in the southeastern United States, where the U.S. Forest Service and its partners have committed to restoring millions of acres of longleaf pine ecosystems. These ecosystems, which are home to dozens of species of conservation concern, cover less than 5 percent of their historic range.
More than a thousand miles away from the edge of this historic range, scientists from the Rocky Mountain Research Station have found a way to help answer this question. They’ve developed a landscape classification system that’s faster, more detailed, and less expensive than previously used methods. Described in several papers recently published in the ISPRS International Journal of Geo-Information, the system combines aerial photography and satellite data, field plot data, and spatial analysis software developed by RMRS scientists to help classify landscapes at a one-square-meter resolution.
This approach was used across 28.8 million acres in the Southeast, providing “probabilistic classifications,” which in this case means probability estimates of whether landscapes fall into specific classes such as water, cropland, forest, and urban development. Nate Anderson, a Research Forester with the Rocky Mountain Research Station in Missoula, Montana, and a co-author of the papers, explains, “There are several advantages to having a clear and accurate model of landscape characteristics, but one of the most important is that you’re better able to identify where restoration efforts are most likely to succeed.”
The engine behind this analysis is a geographic information system (GIS) software plug-in called the RMRS Raster Utility, which is available at https://www.fs.fed.us/rm/raster-utility/. The software, which includes almost 100 different tool forms and almost 400 functions, helps analysts generate detailed information about the location and abundance of different habitat types. According to Anderson, “The Utility facilitates a wide range of data acquisition, sampling, raster analysis, inventory, and statistical modeling functions that can be otherwise difficult or impossible to perform in a standard GIS.”
You might think that high-resolution landscape analysis on such a massive scale would require serious supercomputing power and have significant geographic limitations, but that’s not the case: The software can run on a basic desktop computer and uses data that are available nationwide. And helping to restore millions of longleaf ecosystems is just the beginning. According to John Hogland, a Biological Scientist with the Rocky Mountain Research Station in Missoula and the study’s lead scientist, “We were able to build one of the largest fine-resolution land cover datasets ever created in the United States, and we did it in a matter of weeks using free imagery. We’re working now to share this approach with analysts around the country.”
St. Peter, J.; Hogland, J.; Anderson, N.; Drake, J.; Medley, P. 2018. Fine resolution probabilistic land cover classification of landscapes in the southeastern United States. ISPRS International Journal of Geo-Information. 7: 107.
Hogland, J.; Anderson, N.; St. Peter, J.; Drake, J.; Medley, P. 2018. Mapping forest characteristics at fine resolution across large landscapes of the southeastern United States using NAIP imagery and FIA field plot data. ISPRS International Journal of Geo-Information. 7: 140.