PARENT SESSION
Oral Session # 11: Ecological Studies on Military Installations.
Presiding: B Collins
Monday, August 4. 8:00 AM to 11:30 AM, SITCC Meeting Room 203.
Robust multivariate approaches for developing ecological indicators to classify landscapes on a military disturbance gradient.
Krzysik, Anthony*,1, Kovacic, David2, Wallace, Michael2, Graham, John3, Zak, John4, Duda, Jeffery5, Emlen, John5, Freeman, D.6, 1 Prescott College, Prescott, AZ, USA2 University of Illinois, Urbana-Champaign, IL, USA3 Berry College, Mount Berry, GA, USA4 Texas Tech University, Lubbock, TX, USA5 U.S. Geological Survey - BRD - WFRC, Seattle, WA, USA6 Wayne State University, Detroit, MI, USA
ABSTRACT- Nine research sites were selected in Sand Hills physiography of Fort Benning Georgia that represented a disturbance gradient of military training activities; three sites each in High, Medium, and Low disturbance classes. High sites have a high level of current mechanized infantry training activities. Medium sites had a past history of military training activities, but are only lightly used at present. Low sites have not experienced military training activities. Seven Ecological Indicator Systems (EISs) were separately analyzed with Discriminant Analysis (DA) and Non-metric Multidimensional Scaling (NMS) to extract seven suites of environmental variables that best characterized this disturbance gradient. These groups were called Ecological Indicator Guilds, because of their respective functional and defining response to the disturbance gradient. Two EISs, General Habitat Parameters and Woody Ground Cover Community were capable of separating the three disturbance classes with only a single discriminant function (DF). The High sites were readily separated from the Low and Medium sites by DF1 while the Low and Medium sites were separated by DF2 in three EISs: Ant Community, Soil Chemistry, and Nutrient Leakage. The Microbial Community EIS was able to separate only the High from the Low and Medium sites on DF1. The Developmental Instability EIS, using multiple characters from three perennial plant species, was unable to consistently classify the disturbance gradient. NMS was effective at ordinating the relative multivariate spatial relationships among all nine research sites. A-Horizon Soil Depth and Soil Compaction were identified as simple and reliable ecological indicators to quantify landscape disturbance.
Key words: landscape disturbance, ordination, ecological indicators, discriminant analysis