|Boykin, Deborah - Debbie|
Submitted to: Journal of Entomological Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/2/2009
Publication Date: 7/15/2009
Citation: Defauw, S.L., Vogt, J.T., English, P.J., Boykin, D.L. Seasonal Dynamics of Hyperspectral Reflectance Patterns Influencing Detection of Imported Fire Ant (Hymenoptera: Formicidae) Mound Features in Turfgrass. Journal of Entomological Science. 44: 287-294. 2009
Interpretive Summary: Invasive mound-building imported fire ants impact soil quality and turfgrass management affecting an estimated 20 million acres in sod production, recreational, and residential settings in the southeastern U.S. Monitoring these large areas of land will require remote sensing techniques such as reflectance patterns to detect and quantify imported fire ant mounds. Several reflectance wavebands were identified that consistently detect ant mound soils in turfgrass production and recreational environments. These new remote sensing monitoring tools will aid in the implementation of site-specific management of imported fire ant infestations and benefit a broad array of stakeholders.
Technical Abstract: Invasive mound-building imported fire ants impact soil quality and turfgrass nutrient management affecting an estimated 8.1 million hectares in sod production, recreational, and residential settings in the southeastern U.S. Reflectance characteristics of imported fire ant mound features (i.e., ant mound soil and ant-affected turfgrass at the mound perimeter) and nearby control locations (i.e., undisturbed bare soil and turfgrass) were monitored across seasons from study sites in the Delta and North Central Hills physiographic regions of Mississippi using a full-range spectroradiometer (350-2500 nm) equipped with a 1° foreoptic (FieldSpec Pro, ASD, Inc., Boulder, CO). Data were acquired at two key stages in the ants’ annual cycle (i.e., brood maximum and brood minimum) in a widely-used turfgrass cultivar, Tifway 419 (Cynodon dactylon x C. transvaalensis), and emphasis was placed on simplifying post-processing procedures for hyperspectral datasets; thus, minimizing distortions to these complex ant pedoturbation-plant signatures as well as conserving signal strength needed for cost-effective vehicle-mounted or airborne sensors. Reflectance results (N = 3,800 spectra) indicate that ant-affected turfgrass targets were not reliably distinguishable from unaffected turfgrass from August through November using a multispectral approach with 50 nm waveband segmentation. However, in December, ant-affected turfgrass was distinctive from the other three targets and exhibited substantially lower reflectance values in select visible (VIS) wavebands (600-700 nm) at all locations. Spectral distinctions between bare soils and ant mound soils were significant over several consecutive months (August-November) in the wavebands spanning 600-700 nm and 1050-1100 nm, and for dates in August, October, November and December in the 2000-2400 nm range. Results suggest that mower-mounted spectral devices designed to recognize and map imported fire ant infestations need to provide three to five, user-selected wavebands (VIS, NIR, and MIR) to optimize ant mound detection across seasons.