Author
 |
Vogt, James |
|
Submitted to: Imported Fire Ants Conference Proceedings
Publication Type: Abstract Only Publication Acceptance Date: 9/20/2003 Publication Date: 9/20/2003 Citation: Vogt, J.T. 2003. Seasonal effects on detection of imported fire ant mounds with airborne digital imagery. Imported Fire Ants Conference Proceedings. p. 29. Interpretive Summary: Quantifying imported fire ant mounds over large areas can present problems in terms of person-hours, travel, and inaccessibility of some areas. Airborne digital imagery was used to detect and quantify imported fire ant mounds in northern Mississippi pasture. Detection rates in individual test plots sometimes approached 100%; on average, in May, 75% of mounds could be reliably detected. Spring and Fall were the best times of year for mound detection. This research has the potential to benefit researchers interested in assessing fire ant risk and establishing GIS-based insecticide application regimes, investigating fire ant landscape ecology, quantifying fire ants in areas that are relatively inaccessible on the ground, or monitoring impact of Area-wide Management Programs. Technical Abstract: An experiment was conducted to test high-resolution (0.25 to 0.1 m) digital imagery as a detection tool for quantifying hybrid imported fire ant colonies in northern Mississippi pasture. True color and false color infrared images of a 240 ha pasture were obtained with an airborne imaging system consisting of 4 sensors with 10 nm band-pass filters at 450, 550, 650 and 850 nm. Imagery was obtained in May 2002, Aug 2002, Nov 2002 and Feb 2003. On each sampling date, the entire area was imaged at 0.25 m spatial resolution, and partially imaged at 0.1 m spatial resolution. Plots (N = 20) averaging 0.4 ha each were established within the study area and carefully searched for fire ant mounds by a team of at least 3 workers. Each mound was georeferenced, measured (L, W, H), and assigned a unique number. The percentage of the mound surface obscured by emergent vegetation was visually estimated, and mound activity was assessed by probing the mound and looking for worker ants. A photointerpreter familiar with the general appearance of fire ant mounds examined the images at a scale of about 1: 750 and marked suspected mounds. Actual mound locations were overlaid onto the images to check accuracy of detection. Appearance of mounds was highly variable; generally, mounds in spring, fall, and winter images appeared as light or dark spots of bare soil surrounded by a red (false color infrared composite) or green (true color composite) halo of healthy vegetation. Mounds in the August image generally appeared as dull red (false color infrared composite) or green (true color composite) spots, and were very difficult to see; only 40 ± 5% (mean ± SE) of active mounds were detected in the false color infrared composite. The greatest percentage of active mounds was detected in May in the false color infrared composite (72 ± 5%). Increasing spatial resolution of imagery did not result in greater detection in May, but did result in a 38% increase in detection during other sampling periods. False color infrared composites were generally superior to true color composites, and active mounds were detected more frequently than inactive mounds, probably due to the ants' mound building and maintenance activity. Predictive models for likelihood of detecting mounds based on individual mound characteristics have been developed and are being considered for publication elsewhere. Airborne multispectral imagery is a useful tool for detecting imported fire ant mounds in Mississippi pasture; additional work is underway to test and implement this technology in different areas and under different conditions. |
