In the case of an incipient (aka, low density) population of brown treesnakes (Boiga irregularis), any way that we can maximize the encounter rate and thus detection probability of brown treesnakes across the landscape can be a useful resource for surveyors. These two projects were our attempts to understand if hotspots of increased detection probability could be created using attractants to facilitate a rapid response to a brown treesnake invasion. Both of these studies were conducted in the Closed Population (CP) on the island of Guåhan.
Data folder - contains capture, survey, and catch per unit effort information for both studies in addition to scripts to prepare data for analysis
Scripts folder - contains scripts to run analyses for each study
Models folder - contains model text file for each study
Results folder - contains saved results from the analysis of each study
EncounterRateComparison - script to run post-analysis comparison
PredictiveSingleNight - script to calculate detection probabilities from encounter rates
FigPrep - script for preparing figures
Figures folder - output from FigPrep
dplyr_1.4.4
reshape2_1.4.4
tidyverse_1.3.0
secr_4.3.0
jagsUI_1.5.1
data.table_1.13.0
abind_1.4.5
ggplot2_3.3.2
ggpubr_0.4.0
HDInterval_0.2.2
Data were collected in 2015 where biologists walked transects in CP that either had no traps (no lure) or traps with mice (lure)
The goal was to understand if encounter rate and detection probability were different between these two scenarios
We use an indicator variable (STATUS) to denote when 1 = a transect was not surveyed, 2 = a transect was surveyed but no lures were used, and 3 = a transect was surveyed and lures were used
Data were collected in 2016 where biologists walked transects in CP that either had no scent applied (no scent), were freshly sprayed earlier that evening with fish fertilizer (fresh scent), or had been sprayed 24-hours previously with that scent (old scent)
The goal was to understand if encounter rate and detection probability were different between these three scenarios
We use an indicator variable (STATUS) to denote when 0 = a transect was not surveyed, 2 = a transect was surveyed but no scent was sprayed, 3 = a transect was surveyed and fresh scent was applied, and 4 = a transect was surveyed and scent had been applied the day before
We fit two models in a Spatially Explicit Capture-Recapture framework.
Details of this work can be found in the published journal article on this topic:
Amburgey SA, Yackel Adams AA, Gardner B, Lardner B, Knox AJ, Converse SJ. 2021. Tools for increasing visual encounter probabilities for invasive species removal: a case study of brown treesnakes. Neobiota 70: 170-122. https://doi.org/10.3897/neobiota.70.71379
Start in Scripts to run data prep and analyses; Files to calculate results are EncounterRateComparisons and PredictiveSingleNight; FigPrep creates figures of results