Habitat type and structure affect trap capture success of an invasive snake across variable densities

Melia G Nafus,Gordon H Rodda,Amy A Yackel Adams,Page E Klug

doi:10.1002/ecs2.2339

Melia G Nafus, Gordon H Rodda + Show 2 more

Open Access

https://doi.org/10.1002/ecs2.2339

Copy DOI

Abstract

AbstractDetection represents an important limitation of accurately estimating population size, abundance, and habitat suitability for wildlife, which can be especially true for cryptic animals. Moreover, for reptiles, juveniles are often less likely to be detected than later life stages. In the case of invasive species, preventing false negatives early in the invasion process can be critical for improving outcomes of control measures. We evaluated habitat structure in relation to catch per unit effort (CPUE) and mean size of trapped invasive brown treesnakes (Boiga irregularis) on Guam. We used a 5‐ha enclosure containing a known, closed population of brown treesnakes to identify key habitat variables that related to CPUE and mean size of trapped snakes over six years. We then tested the relationship of those variables to CPUE and mean size of trapped snakes at three sites with suppressed snake populations as a proxy for low‐density populations anticipated to occur during early detection of invasive populations. We found that a coarse measure of habitat structure represented by three forest types correlated with trap detections, as well as finer measures of habitat structure, such as distance to nearest branch and the type of trap support structure used. On average, smaller snakes were captured in traps placed higher in the tree canopy. Some, but not all, habitat variables identified as predictive of CPUE and mean size within the enclosed population pre‐suppression were also predictive at the snake‐suppressed (low‐density proxy) sites. Habitat structure around the sampling unit (a trap) affected detection probability and the size of detected individuals independently of the demographic structure of the population. Measuring wildlife‐habitat relationships of invaders in their novel environments may be one method to improve early detection during invasive species management.

Highlights

A fundamental parameter for wildlife management is detectability, or the probability of locating a species or individual that is present
The results suggest some robustness between models in the habitat variables associated with catch per unit effort (CPUE), statistical approaches did affect parameter-specific inferences
Variables that were predictive in the suppressed, lowdensity proxy populations did, show notable overlap with the variables identified by the regression tree in Northwest Field North (NWFN) presuppression

Summary

Introduction

A fundamental parameter for wildlife management is detectability, or the probability of locating a species (species occupancy) or individual (individual detectability) that is present. Selection and application of management scenarios based on species occupancy state (presence or absence) rely on clear relationships between presence–absence and detection or non-detection, but presence–detection relationships may be confounded by factors that affect individual detectability. In the case of invasive species, established populations can exist with few to no detections of individuals, such that detection does not occur until populations are large enough that eradication is difficult, if not impossible. Evaluating the relationship between detection and the structure of surveyed habitat may be one avenue toward understanding false negatives and the application of occupancy surveys for the management of land, wildlife, and invasive species, especially for cryptic organisms

Methods

Results

Conclusion