Data from: A framework for modeling the impacts of searcher behavior on the efficiency of abundance surveys
Data files
Jun 10, 2024 version files 155.63 KB
-
Encounters_-_Distance.xlsx
35.53 KB
-
Encounters_-_Quadrats.xlsx
41.69 KB
-
Encounters_-_Removal.xlsx
26.92 KB
-
README.md
5.06 KB
-
Transect_datasheets.xlsx
46.43 KB
Jul 16, 2024 version files 155.36 KB
Abstract
When planning abundance surveys, the impact of search effort on the quality of the density estimates is rarely considered. We constructed a time-budget modeling framework for abundance surveys using principles from optimal foraging theory. We link search effort to the number of sample units surveyed, searcher detection probability, the number of detections made, and the precision of the estimated resource density. This framework allowed us to determine how a surveyor should behave to produce optimal density estimates. Using data collected from quadrat and removal surveys of zebra mussels (Dreissena polymorpha) in central Minnesota, we applied this framework to evaluate potential improvements. By tuning searcher behavior, we find that density estimates from removal surveys of zebra mussels could be improved by up to 60% in some cases, without changing the overall survey effort. Our framework also predicts a critical population density where the best survey method switches from removal surveys at low densities to quadrat surveys at high densities, consistent with past empirical work. Our results provide insights into how to improve the performance of many survey methods in high-density environments by either tuning searcher behavior or decoupling the estimation of resource density and detection probability.
README: Data from: A framework for modeling the impacts of searcher behavior on the efficiency of abundance surveys
Contains data and R files to reproduce analyses from "A common framework for modeling animal search: Linking foraging ecology to survey design through trade-offs between search effort and detection"
Change log
Version July 16, 2024 - minor edits to README
Description of the Data and file structure
R files to reproduce analyses in the main manuscript:
CV_calculation.R: To reproduce the zebra mussel analyses, run this code. This file calculates optimal search times by minimizing the coefficient of variation in the estimated density of a removal survey. Reproduces Fig 3. Calls DensityEstimates.R.
CV_sensitivity.R - contains code to produce the sensitivity analysis of the zebra mussel surveys. Reproduces Fig 4. Calls DensityEstimates.R.
DensityEstimates.R: This calls the necessary functions to estimate density and model the time budget data. Also produces density estimates from the empirical survey data.
RemovalSearch.R - this file is used to calculate the optimal search time and time to complete a transect for the removal survey. Reproduces Fig 1 and Fig 2.
ZebraFuncs.R: This contains functions to read in datasets, format data for analysis, and estimate zebra mussel densities. Also contains functions to calculate the coefficient of variation for removal and quadrat surveys given time budget information and the target density distribution.
Data files to reproduce analsyses in the main manuscript:
Missing data are entered as NA
Encounters - Quadrats.xlsx: This contains information on each detection event in the quadrat surveys for each of the three lakes. Data columns, in order, are time that detection occurred, date of detection, name of the lake, the observer who made the detection, the distance along the transect that quadrat was at, the number of mussels in the quadrat, any substrate the mussels were attached to, divers comments, and the number of mussel clusters in the quadrat.
Encounters - Removal.xlsx: This contains information on each detection event in the removal surveys for each of the three lakes. Data columns, in order, are time that detection occurred, date of detection, name of the lake, the observer who made the detection, the transect number the detection occurred on, the distance along the transect that detection occurred at, the number of mussels in the detection event, the length and width of the cluster, the substrates that mussels are attached to, and any additional diver comments.
Encounters - Distance.xlsx: This contains information on each detection event in the distance surveys for each of the three lakes. Note that distance surveys were not used in the associated paper but are included for completeness. Data columns, in order, are time that detection occurred, date of detection, name of the lake, the observer who made the detection, the transect number the detection occurred on, the distance along the transect that detection occurred at, the distance from the transect the detection occurred at, the side of the transect the detection occurred at, the number of mussels in the detection event, the length and width of the cluster, the substrates that mussels are attached to, and any addition diver comments.
Transect datasheets.xlsx: Contains information about each transect completed in the study. Data columns, in order, are the transect type conducted, the lake the transect was conducted in, the date the transect was completed, the divers on the transect, the name of the primary observer and the name of the secondary observer (if relevant), the length of the transect, the unique number of the transect, the GPS coordinates at the start of the transect, compass bearing fo the transect line, time taken to survey the habitat along the transect, time take to survey the transect, clarity along the transect.
Files for simulation
Tutorial.pdf - This tutorial runs through three different examples using distance sampling, occupancy analysis, and n-mixture models using simulation tools. Presents the code in MRDS_halfnorm_example.R, UnmarkedSims_occu_example.R, and Unmarked_pcount_example.R.
Sim_funcs.R - Contains functions needed to simulate datasets and find optimal survey time using the MRDS and unmarked packages.
MRDS_halfnorm_example.R - Example calculating the optimal search time using the halfnormal distance function in the MRDS package. Relies on Sim_funcs.R.
UnmarkedSims_occu_example.R - Example calculating the optimal search time using the occupancy model in the unmarked package. Relies on Sim_funcs.R.
UnmarkedSims_pcount_example.R - Example calculating the optimal search time using the distance fuction in MRDS. Relies on Sim_funcs.R.
Methods
Data on search times for removal and quadrat surveys was collected by divers in three Minnesota lakes. Code is provided to format the dataset, estimate density, and reproduce all analyses in the manuscript.