Data and code to compare acoustic ARU and point count data for shorebird species presence across Alaska's North Slope
Data files
Abstract
The rapid decline of global bird populations demands effective monitoring approaches, particularly for shorebirds, which are experiencing alarming population decreases. Most shorebird surveys rely on traditional visual survey methods, but face significant challenges in remote locations like the Arctic, where high costs and logistical constraints limit survey effort and duration. We compared visual area search surveys and acoustic recording units (ARUs) for monitoring shorebird species in two large regions of Alaska's Arctic Coastal Plain in 2022-2023. We deployed ARUs at 129 sites (54 in Arctic National Wildlife Refuge, ANWR, 75 in Teshekpuk Lake Special Area, TESH; ~11,000-18,000 hours of acoustic recordings) where visual surveys were also conducted (~83-113 hours in total), and examined differences in species richness, encounter rates, and predicted distribution patterns for 12 shorebird species or species groups. ARUs detected 50% higher species richness in ANWR and 33% at TESH (approximately 2 additional species per plot); the odds of encountering a species were more than twice as high with ARUs compared to visual surveys (95%CI = 1.64-5.57). Species accumulation curves demonstrated that ARUs required fewer plots to detect the maximum number of species. Habitat models based on ARU data produced similar predicted distributions to visual survey data but enabled modeling for additional species due to higher encounter rates. Different survey methods showed consistent spatial patterns between sites, with both methods detecting higher species richness and encounter rates in TESH than in ANWR. Our findings demonstrate that ARUs can effectively monitor shorebird communities in the Arctic, offering advantages in temporal coverage and synchronous data collection across large spatial extents. The performance of ARUs for detecting most shorebird species over long periods supports their integration into conservation monitoring programs, particularly in remote regions where traditional monitoring approaches are challenging to implement and sustain.
Dataset DOI: 10.5061/dryad.7wm37pw54
Description of the data and file structure
This dataset contains counts and presence/absence data for acoustic and visual detections of 12 shorebird species at plots across Alaska’s Arctic Coastal Plain in June-August of 2022-2023 as part of the Program for Regional and International Shorebird Monitoring (PRISM). Acoustic data was collected using autonomous recording devices (Audiomoth recorders). Specifics of data collection (exact plot locations and dates) can be found in the manuscript associated with this dataset. This dataset also includes code and data related to the comparison of these two data collection methods (acoustic and visual) including the creation of species accumulation curves and species distribution models. Models included environmental data extracted from online sources; the specifics of these extractions is described in detail in the manuscript associated with this dataset.
Files and variables
File: code.zip
Description: This zip file contains all code required to reproduce the analyses and figures in the manuscript associated with this dataset. These analyses were completed using the Spyder IDE (v5.5.1) and Python 3.11, or in R software version 4.2.3. Please note that TLSA referred to in all scripts is Teshekpuk Lake Special Area, or TESH in the manuscript associated with this code. The zip file includes the following:
-map_SorensonDice_similarity.py : Python script to calculate and plot Sorenson-Dice similarity as color on maps (Fig. 2a in the corresponding manuscript), and calculate and compare species' richness (Table 1 in the corresponding manuscript).
This code requires input
- an ARU counts file for one of the sites (allCounts.csv)
- a metadata file describing which ARUs are on which plots (acoustic_monitoring_.csv)
- a visual counts file (*rapid_survey_data.csv)
- verification data on species presence/absence (pres_verif.csv)
-plot_figureS4-5.py: Code to produce Figure S4-5 in the manuscript associated with this dataset. This script requires as input:
- an ARU counts file for one of the sites (allCounts.csv)
- a metadata file describing which ARUs are on which plots (acoustic_monitoring_.csv)
- a visual counts file (rapid_survey_data.csv)
ER_and_speciesAccum.R: Code to reproduce figure 2b-c and all encounter rate calculations (including odds ratios and Figure S3) in the manuscript associated with this dataset. The encounter rate code blocks require as input allSpecPres.csv files from each site, which contain relevant info for both visual and ARU detections. The species accumulation code block requires accumData.csv files from each site.
all_SDMs_code.R: Species distribution modelling code for manuscript associated with this dataset. This includes all models used to create Figure 3 and Figures S1-2. This script requires as input final environmental variables (elevation_merged.tif, waterdens_2.tif, landcov.grd, distance.tif) as well as allSpecPres.csv files for the associated site.
File: data.zip
Description: This zip file contains all data required to reproduce the analyses and figures in the manuscript associated with this dataset. Environmental variable data were extracted and compiled from online sources and using QGIS and R software as described in detail in the manuscript. Only final versions of these data used for species distribution models are uploaded here. Please note that TLSA referred to in all cases is Teshekpuk Lake Special Area, or TESH in the manuscript associated with this code. Species codes throughout the zip file are as follows: american golden-plover (amgp), bar-tailed godwit (barg), black-bellied plover (bbpl), dunlin (dunl), long-billed dowitcher (lbdo), least sandpiper (lesa), pectoral sandpiper (pesa), red-necked and red phalaropes (phalarope), semipalmated sandpiper (sesa), stilt sandpiper (stsa), western sandpiper (wesa), wilson's snipe (wisn). The zip file includes the following:
acoustic_monitoring_anwr.csv: CSV of acoustic metadata for the Arctic National Wildlife Refuge (ANWR). Fields include
- site- data recording site
- plot- plot identity for recording
- recorderID- acoustic ID for ARU at that plot
- lat- latitude of the plot location
- long- longitude of the plot location
acoustic_monitoring_tlsa.csv: Same as acoustic_monitoring_anwr.csv but for Teshekpuk Lake Special Area data.
ANWR_2022_accumData.csv, ANWR_2022_PRISM_accumData.csv: Presence/absence data for each species and survey plot in the Arctic National Wildlife Refuge in 2022. Each row represents a different survey plot. Cells have a '1' if the species was present and a '0' otherwise. The PRISM csv has presence/absence from visual survey data; the other csv has presence/absence based on ARU survey data.
TLSA_2023_accumData.csv, TLSA_2023_PRISM_accumData.csv: same as ANWR versions but for survey plots in the Teshekpuk Lake Special Area in 2023.
ANWR_2022_allCounts.csv: Counts of species' detections in each acoustic data file from the Arctic National Wildlife Refuge in 2022. Contains the following fields:
- ARU name: unique ARU identifier from which the file originates
- filename: name of the acoustic data file
- species: 4-letter species code
- counts: number of detections of that species in that data file
TLSA_2023_allCounts.csv: Same as the ANWR version but for the Teshekpuk Lake Special Area in 2023.
ANWR_allSpecPres.csv: Species presence from visual and ARU data from the Arctic National Wildlife Refuge in 2022. Fields are as follows:
- plot: ARU name from the survey plot in question
- lat: plot latitude
- lon: plot longitude
- col_PRISM: Presence or absence in visual data
- spec: species 4-letter code
- colARU_0: Presence or absence in ARU data from full summer
- colARU_1: Presence or absence in ARU data from before July 5 (data used in rest of manuscript associated with this dataset)
- colARU_2: Presence or absence in ARU data from only June 1 - June 16
- colARU_3: Presence or absence in ARU data from only the day of the visual survey
TLSA_allSpecPres.csv: Same as for the Arctic National Wildlife Refuge, but for the Teshekpuk Lake Special Area
ANWR_PRISM_data_2022_rapid_survey_data.csv: Raw visual survey data from survey plots in the Arctic National Wildlife Refuge. Fields include:
- plot_id: Unique identifier of the survey plot
- date: survey date
- year: survey year
- species_code: 4-letter species code
- flyover: number of individuals of that species that flew over the plot during surveying
- total_numb_individuals: total number of individuals of that species observed on plot, including breeding and non-breeding individuals as well as flyovers
TLSA_PRISM_data_2023_rapid_survey_data.csv: Same as for the Arctic National Wildlife Refuge, but for the Teshekpuk Lake Special Area in 2023.
pres_verif.csv: Verification data for ARU presence across all plots used in the manuscript associated with this dataset. Fields include the following:
- site: Site of plot, either ANWR or TLSA
- dep: ARU unique identifier for the ARU deployed at each plot at each site
- amgp, etc: all other field are 4-letter species codes. Values in each cell in these fields are 'y' for species verified present, 'n' for species verified absent.
env: folder containing all final environmental data used in the species distribution models in the manuscript associated with this dataset. More information on the derivation of these variables can be found in the manuscript. This folder includes:
better_coast: Shapefile and associated files for coastline data used in maps of species distribution models. This data was accessed using the Global Self-consistent, Hierarchical, High-resolution Geography Database (https://www.soest.hawaii.edu/pwessel/gshhg/)
distance.tif: Distance to the coast for 1kmx1km grid cells covering Alaska's Arctic Coastal Plain
elevation_merged.tif: Elevation for 1kmx1km grid cells covering Alaska's Arctic Coastal Plain
landcov: Gridfile and associated files for landcover class data in 1kmx1km grid cells covering Alaska's Arctic Coastal Plain
temp2022.tif, temp2023.tif: Average June temperature (in 2022 and 2023) for 1kmx1km grid cells covering Alaska's Arctic Coastal Plain
waterbod_lowres: Shapefile and associated files for water body edges, used in mapping species distribution models
waterdens_2.tif: Density of waterbodies in a 10 km radius of each 1kmx1km grid cell covering Alaska's Arctic Coastal Plain
Code/software
Code is described in the above section on files and variables (under code.zip)