Quantifying insect dispersal distances from trapping detections data to predict delimiting survey radii

Authors: CATON, B - Animal And Plant Health Inspection Service (APHIS); FANG, H - North Carolina State University; Manoukis, Nicholas; PALLIPPARAMBIL, G - North Carolina State University

Submitted to: Journal of Applied Entomology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 10/10/2021
Publication Date: 10/28/2021
Citation: Caton, B.P., Fang, H., Manoukis, N., Pallipparambil, G.R. 2021. Quantifying insect dispersal distances from trapping detections data to predict delimiting survey radii. Journal of Applied Entomology. 146(1-2):203-216. https://doi.org/10.1111/jen.12940.
DOI: https://doi.org/10.1111/jen.12940

Interpretive Summary: Information on the movement of invasive pests is difficult to obtain, but also can be critical to mounting an effective containment and eradication response. Here we used real-world trapping data from six invasive pest species (European grapevine moth, Giant African landsnail, Japanese beetle, Mediterranean fruit fly, Mexican fruit fly, and Oriental fruit fly) to estimate how they move after incursions into pest-free areas and to obtain estimates of their diffusion coefficient ("D") which can be used to calculate how far they are likely to disperse over time.

Technical Abstract: The natural spread potential of invasive pests is of central concern to delimitation, quarantine, and eradication efforts. We analyzed trapping survey detections data for five insects and one low-dispersing mollusk to quantify and describe 30-day dispersal kernels (mean total distance (MTD), m). We hypothesized that MTD would increase with species’ reported diffusion coefficients (D, m2 per d), and that D could be used to predict containment radius lengths for delimiting surveys for exotic pests. We collected trapping data for the following six invasive pest species: European grapevine moth [EGVM; Lobesia botrana (Denis & Schiffermüller)], Giant African landsnail [GALS; Lissachatina fulica (Bowdich, 1822)], Japanese beetle [JB; Popillia japonica (Newman, 1838)], Mediterranean fruit fly [Medfly; Ceratitis capitata (Wiedemann, 1824)], Mexican fruit fly [Mexfly; Anastrepha ludens (Loew, 1873)], and Oriental fruit fly [OFF; Bactrocera dorsalis (Hendel, 1912)]. We started with a K-means clustering to group detections that were proximate in space and time, then finalized the clusters manually (with the exception of GALS). We calculated MTD from the centroid of the cluster for each detection. Probability histograms for MTD were fit to a two parameter exponential function, and from that probability density function we estimated species’ upper dispersal limits as the 99th percentile. The species with the lowest reported D values, GALS and EGVM, had the greatest decay rates, and smallest MTD and 99th percentiles. OFF, with the greatest reported D, had the smallest decay rates and greatest MTD and 99th percentiles Medfly, Mexfly, and JB had intermediate MTD Regressions of 99th percentiles as a logarithmic function of D fit observed distances and maximums reasonably well. The upper 95-percent confidence limit of that function seemed best for predicting survey radii. These empirical results—based on real survey data—provide the first proposed solution for determining delimiting survey radii (for durations of ca. 30 days) across a wide range of dispersal abilities, and also indicated that many delimiting survey grids are oversized.