Validating Deployment of Aerially Delivered Toxic Bait Cartridges for Control of Invasive Brown Treesnakes

Abstract

Aerial application of management tools can provide a cost-effective means to conserve or control wildlife populations at the landscape scale. Large spatial scales, however, present difficulties when assessing in situ reliability and integrity of the devices themselves. We demonstrate application of a distance-sampling density estimation approach to assess the performance of a newly developed toxicant bait system for the control of invasive brown treesnakes (Boiga irregularis). Bait cartridges were designed to open in flight to expose the toxicant-laced bait and tangle in the forest canopy via a plastic ribbon component. Following application of 12,686 bait cartridges from an automated aerial delivery system over a 55-ha site on Guam, USA, we employed distance sampling techniques to evaluate cartridge performance. We performed 22 line-transect surveys for a total distance of 10.3 km, during which we recorded all observations of unopened bait cartridges, instances in which the ribbon did not remain attached to the cartridge capsule (i.e., ribbon failure), and carcasses of brown treesnakes and nontarget species. Too few undeployed bait cartridges (n = 6), brown treesnake carcasses (n = 1), or nontarget carcasses (n = 0) were observed during surveys to support additional analysis. We detected 299 instances of ribbon failure. Using standard distancesampling analyses, we estimate that ribbon failure occurred in 3,376 ± 351 (estimate ± SE; 95% CL = 2,746-4,150) cartridges or 21.6-32.7% of the total applied. Our results demonstrate the utility of distance-sampling density estimation techniques to validate performance and reliability of aerially applied management tools.