REGIONAL INTEGRATED MANAGEMENT OF IMPORTED FIRE ANT
Location: Biological Control of Pests Research Unit
Title: Implementation of Hyperspectral Techniques in the Remote Detection of Imported Fire Ants Mounds (Hymenoptera: Formicidae) in Cultivated Turfgrass
Submitted to: National Entomological Society of America Annual Meeting
Publication Type: Other
Publication Acceptance Date: December 13, 2006
Publication Date: N/A
Interpretive Summary: Imported fire ant detection and monitoring efforts, at field- to landscape-scales, are time-consuming and expensive endeavors. Use of airborne or satellite imagery to detect and count imported fire ant mounds over large areas such as pasturelands is known to be limited by (1) season (either the late Winter or Spring) when sharp contrasts between vegetation and soils are reportedly developed; and (2) site-specific landscape-related factors (e.g., variability in soil series, soil moisture levels, plant community assemblages, topography, etc.). A study was initiated in Mississippi (Summer 2006) to identify bandwidths of reflected light (spanning the visible through mid-infrared wavelengths) that enhance the detection of cryptic (or hidden) fire ant infestations in intensively-managed turfgrass settings over the course of four seasons. During peak summer season (August-September), three broad-band reflectance markers were discovered that reliably distinguished ant-affected mound soil, unaffected bare soil, and adjacent turfgrass. Additional findings indicated bandwidth combinations used in standard color photographs (red, green, and blue wavelengths) or color infrared imagery (infrared, red, and green wavelengths) either typically failed to resolve mound soils from turf or did not distinguish unaffected bare soil from surrounding vegetation, despite the frequency of irrigation. These results will help researchers: (1) design a ground-based sensor system (that may be attached to mowing equipment); or (2) equip airborne digital cameras with appropriate band-pass filters on lens systems in order to maximize fire ant mound detection during the summer to fall transition. Development of new remote sensing detection and monitoring tools will aid site-specific management of imported fire ant infestations in perennial, warm-season turfgrass settings (such as golf courses, athletic fields, parks, and commercial sod production areas), and will help foster sustainable reduction of imported fire ant populations.
Safe, expedient, and cost-effective treatments of imported fire ant (IFA) infestations require technological developments that exploit the use of remotely-sensed contrasting features to detect cryptic mounds in heavily-managed turfgrass. Ground-based implementation of hyperspectral techniques in the field-scale quantification and seasonal monitoring of IFA colony distributions is a prerequisite for either designing sensors or for equipping airborne digital cameras with appropriate band-pass filters to maximize mound detection for regional surveys. The objectives of this study were twofold: (1) examine spectral reflectance characteristics of ant-affected versus undisturbed turfgrass and soils; and (2) identify bandwidths that enhance the detection of cryptic fire ant mounds in intensively-managed turfgrass areas. Peak summer season results (N = 12,000 full-range spectra collected August-September 2006), for sparsely-covered ant mounds ('50% vegetation) from Mississippi sites in the North Central Hills and Delta physiographic regions, indicated that mean reflectance values for targets (i.e., bermudagrass, mound soil, and undisturbed bare soil) averaged over 50 nm bandwidths were most distinctive from each other at 650-700 nm (F=31.8; df=3, 8.3; P<0.0001), 1450-1500 nm (F=36.9; df=3, 6.2; P<0.001), and 2000-2050 (F=50.2; df=3, 5.6; P<0.001). Ant-affected turfgrass (especially on mound perimeters) was not reliably distinguishable from unaffected turfgrass (approximately 1 m from mound centroid) during peak summer season. The development of new remote detection technology, employing seasonally-acquired spectroradiometric data in turf as a model system, will facilitate site-specific delivery of insecticides, reduce pesticide use, enhance regional monitoring efforts, and benefit a broad array of stakeholders.