Using Feature Analyst to Automate Counts of Photographed Indiana Bats

I like batman, one of my favorite heroes.... :) More about bat survey.....



Abstract:
"The Indiana bat (Myotis sodalis) is a Federally listed endangered species found in the Midwest and eastern United States. Its population has profound implications for forest management throughout its range. Declining populations could lead to timber harvest restrictions and changes to other land management practices in Midwestern and eastern forests. The Forest Plan and Biological Opinion of the Hoosier National Forest (Indiana) require hibernacula (i.e., winter hibernation sites) occupied by the Indiana bat to be monitored regularly to assess changes in population numbers. State and U.S. Fish and Wildlife Service biologists survey the bats every other winter while the bats are hibernating. Because management decisions and recovery action priorities are based on the population estimates and trends determined from these surveys, it is critical that they are accurate. During the surveys, individual bats and small clusters are counted directly, but those in larger clusters are estimated by multiplying the approximate area of the cluster by an assumed bat packing density. Unfortunately, estimates derived by these techniques can be highly inaccurate. In many hibernacula, using a digital camera to document bat numbers could reduce stress to the bats and also increase the accuracy of the population estimate by allowing the bats to be counted manually on a computer screen in an office setting. Nevertheless, this is a very tedious and labor-intensive task that some state and federal agencies cannot afford. An accurate, rapid, and more cost-effective way to count photographed bats is needed. With sponsorship from the U.S. Department of Agriculture, Forest Service Remote Sensing Steering Committee, the Remote Sensing Applications Center conducted a study to investigate the feasibility of rapidly deriving accurate counts of photographed bats using Feature Analyst, an extension for ArcGIS. Counts derived with Feature Analyst were typically within one to nine percent of manually interpreted counts and processing times of less than four minutes per photo were achieved. This represents a significant improvement over traditional in-cave estimates and has the potential for high-volume use, which could further reduce the per photo processing time."

http://www.featureanalyst.com/feature_analyst/publications/success/bats.pdf