Evaluating a novel core-and-perimeter delimiting trapping survey design for insects. II. Simulations and case studies

We compared 2 delimitation trapping designs via mark-release-recapture of Ceratitis capitata Weidemann (Medfly) [Tephritidae: Diptera]: a fully trapped grid versus a novel, "core-and-perimeter" design. Results showed some disadvantages of fully trapped and supported the core-and-perimeter concept, but grids were too small to address all questions. Here we used computer models to more fully compare the designs. We employed random walk dispersal using parameters from recent results for another Dipteran. Parameter sensitivity was highest for maximum step distance (DDay), but structural sensitivity was more affected by dispersal probability (PDisp) and correlated movement (θMax). We incorporated trapping, modeling capture probabilities based on the hyperbolic secant function. We quantified the likelihood of inner traps to "intercept" insects along their paths and implemented that process in all trapping models. Validation against 3 independent datasets was successful using PDisp of 0.4 or 0.3, and θMax = 180°, based on dispersal densities and proportion of flies captured. Under the validated dispersal model, Medflies never traveled more than 2.8 km in 30 d. A trapping version gave zero perimeter captures with the core-and-perimeter design, but core area captures in 40% of iterations for single flies. The fully trapped grid captured flies in 79% of iterations but only 12% of traps had captures, indicating suboptimal efficiency. Finally, we compared fully trapped and core-and-perimeter trapping costs using published plans for Medfly and four other species. The costs of the core-and-perimeter designs were 49% to 77% lower at standard sizes but 69% to 88% lower with optimal (shorter) radii.