Title: Hyperspectral Remote Sensing of Seasonally-Acquired Imported Fire Ant Mound Features (Hymenoptera: Formicidae) in Turfgrass Authors
Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: April 4, 2008
Publication Date: N/A
Technical Abstract: Invasive mound-building imported fire ants (Solenopsis spp.) impact soil quality and turfgrass nutrient management in sod production, recreational, residential, and commercial settings. Ground-based hyperspectral studies focused on the seasonal monitoring of reflectance characteristics of imported fire ant mound features is a prerequisite for either designing vehicle-mounted sensor arrays or for equipping airborne multispectral digital cameras with appropriate band-pass filters to maximize mound detection as well as optimize timing for field- to landscape-scale surveys and management efforts. The objectives of this study were twofold: (1) examine seasonally-acquired broadband (50 nm) spectral reflectance patterns for ant-affected versus undisturbed turfgrass and soils (inceptisols, entisols, alfisols, and vertisols) for a widely-used turfgrass cultivar, Tifway 419 (Cynodon dactylon x C. transvaalensis); and (2) identify wavebands that enhanced the detection of imported fire ant mound features across seasons in heavily-managed production and recreational environments. Reflectance data (N = 3,800 spectra) were collected at intervals from two turf farm tracts and a golf course in Mississippi (August-December 2006) using a full-range spectroradiometer equipped with a 1° foreoptic (FieldSpec Pro, ASD, Inc., Boulder, CO). Hyperspectral scans were acquired from four target types (ant mound soil, undisturbed soil, ant-affected turfgrass, and undisturbed turfgrass). Analysis of variance (ANOVA) was used to elucidate significant waveband and target combinations; means were considered different at a=0.05. Significant differences in reflectance wavebands were found between bare soils and ant mound soils (P=0.01); these differences were consistent over several months in the 600-700 nm (August-November), 1050-1100nm (August-November) and 2000-2100 nm (August-December) ranges. Turfgrass response to ant mound soil disturbances was highly variable, and, therefore not a reliable indicator of mound features in sod production agroecosystems or golf course settings. Results suggest that mower-mounted spectral devices designed to map imported fire ant infestations would benefit by providing three to five wavebands (VIS, near IR, and mid IR) to optimize ant mound detection across seasons.