Dataset overview

This data package contains the camera-trap, camera-station, weather, wind, spatial, elevation, and figure-support files used to reproduce the analyses and figures for the otter-mink manuscript. The study compares American mink (Neogale vison) and Eurasian otter (Lutra lutra) detections and behaviour in two coastal Norwegian study regions:

• Agder: otter expansion area.
• Vestland: established otter area.

The analysis scripts assume that all files and folders listed below are placed in the same main working directory:

project_dir <- "enter_your_directory_here"

The camera-trap data package should be located inside:

cascades20240801/

Several CSV files are encoded as Excel-friendly UTF-8 comma-separated files. In R, read these with fileEncoding = "UTF-8-BOM" when using. read.csv().

All compressed folders or bundled directories should be unpackaged/unzipped before running the R scripts. In particular, cascades20240801/, Vestland_elev_data/, and Agder_elev_data/ should exist as normal folders inside project_dir before importation. The map script can unzip elevation tiles inside the elevation folders if .zip raster tiles are present, but the parent folders themselves must already be unpacked and accessible.

---

File and folder inventory

---

Folder descriptions

cascades20240801/

This folder is the Agouti/camtrap-dp export used by the scripts. It should contain datapackage.json plus the associated camtrap-dp CSV files. The scripts use camtraptor::read_camtrap_dp() to import the data package and use the package to construct independent detection records.

Expected important files include:

Vestland_elev_data/

Elevation data folder for the Vestland established area. The publication map script searches this folder for .zip or .tif elevation tiles. If zipped tiles are present, they are unzipped into a generated unzipped/ subfolder. The rasters are mosaicked, cropped to the Vestland study-area bounding box, and plotted as the elevation background for the Vestland camera map.

Each raster cell represents elevation at a spatial location. Values at or below sea level are masked for the map background.

Agder_elev_data/

Elevation data folder for the Agder expansion area. The publication map script handles this folder in the same way as Vestland_elev_data/: it searches for .zip or .tif elevation tiles, mosaics them if needed, crops to the Agder study-area bounding box, and plots the result as the elevation background for the Agder camera map.

Each raster cell represents elevation at a spatial location. Values at or below sea level are masked for the map background.

---

CSV data dictionaries

Camera_information.csv

Rows: 66
Row meaning: one camera station.

Days_cam_failure.csv

Rows: 4017
Row meaning: one camera station on one date for which a camera-failure/no-effort status is recorded.

In the occupancy and detection-frequency scripts, dates marked with Cam_failure == "Yes" are treated as no camera effort, not as true non-detections.

Weather_20230301_20240229.csv

Rows: 732
Row meaning: one weather station/region/date.

The main detection-frequency models use Min_temp, Mean_temp, and Precipitation; camera-specific wind covariates are taken from the ERA5 file below.

era5_wind_daily_per_camera.csv

Rows: 24156
Row meaning: one camera station on one date.

The detection-frequency scripts convert date_local to Date and join this file to camera-day detection records by Cam_ID and Date.

---

Spatial files

Norway_sf.rds

R serialized spatial object used as the Norway outline in the publication map. Load in R with:

norway_shape_outline <- readRDS("Norway_sf.rds")

The publication map script simplifies this object with sf::st_simplify() and plots it with geom_sf().

Norway_shapefile.rds

R serialized spatial object containing Norway administrative/county polygons. Load in R with:

norway_counties <- readRDS("Norway_shapefile.rds")

The publication map script expects a county-name field called NAME_1 and highlights the following counties as remnant otter population areas:

c("Nordland", "Troms", "Finnmark", "Hedmark")

Each row of this spatial object represents an administrative polygon/feature.

---

Figure-support image files

Eurasian_otter_silhouette.png

PNG image file used as an optional otter silhouette in figure generation. Dimensions: 1024 x 1024 pixels.

American_mink_silhouette.png

PNG image file used as an optional mink silhouette in figure generation. Dimensions: 1024 x 1024 pixels.

---

R code files

The R scripts are designed to be run from the same directory project_dir used throughout this README. Before running them, place all CSV, RDS, PNG, elevation, and camtrap-dp folders inproject_dir, and make sure all folders have been unpackaged/unzipped.

R_script_for_occupancy_and_det_freq.R

This script reproduces the camera-trap occupancy/detection-incidence and detection-frequency analyses. It imports the camtrap-dp package fromcascades20240801/, creates independent records using a 60-minute threshold, builds camera-day detection histories for mink and otter, incorporates camera-failure days as no-effort days, and fits the regional daily naive occupancy and camera-day detection-frequency models. It also joins camera covariates, regional weather, and ERA5 wind covariates to the camera-day data, calculates time since last otter detection for mink camera-days, and generates the occupancy/detection-frequency figure.

Main input files and folders used by this script:

• cascades20240801/
• Days_cam_failure.csv
• Camera_information.csv
• Weather_20230301_20240229.csv
• era5_wind_daily_per_camera.csv
• American_mink_silhouette.png and Eurasian_otter_silhouette.png if silhouettes are included in the figure

Main R package dependencies:

camtraptor
camtrapR
dplyr
tidyr
stringr
lubridate
data.table
ggplot2
scales
grid
glmmTMB
car
emmeans
splines
openxlsx
cowplot

R_script_for_behavioral_analysis.R

This script reproduces the mink behaviour analyses. It imports observation-level data from the Agouti/camtrap-dp export, filters to animal event observations for mink and otter during the study period, summarises mink behaviour by region, fits a binomial model comparing investigating behaviour between regions, and fits the Vestland-only time-since-otter behaviour model. The script also generates the two-panel behavioural figure showing regional mink behaviour composition and the model-predicted probability of investigating behaviour as a function of time since otter detection.

Main input files and folders used by this script:

• cascades20240801/, especially datapackage.json, observations.csv, and deployments.csv

Main R package dependencies:

camtraptor
readr
dplyr
tidyr
stringr
lubridate
glmmTMB
car
emmeans
splines
ggplot2
scales
cowplot

R_script_for_rhythmic_activity.R

This script reproduces the rhythmic-activity analyses. It imports independent camera-trap records, joins camera coordinates from Camera_information.csv, calculates local dawn, sunrise, sunset, and dusk times using camera coordinates and dates, assigns detections to Dawn, Day, Dusk, or Night, summarises mink and otter activity periods by region and month, and fits binomial GLMMs testing monthly and regional differences in activity periods. It also generates the rhythmic-activity figure showing monthly activity-period proportions for mink and otter.

Main input files and folders used by this script:

• cascades20240801/
• Camera_information.csv

Main R package dependencies:

camtraptor
camtrapR
dplyr
tidyr
stringr
lubridate
purrr
suncalc
glmmTMB
lme4
emmeans
car
ggplot2
ggpattern
ggh4x
cowplot
scales

Publication_maps.R

This script reproduces the publication map figure. It loads the Norway outline and county spatial files, identifies remnant-population counties, loads and mosaics the elevation raster tiles for Agder and Vestland, crops the elevation rasters to the study-area extents, imports camera locations from deployments.csvand combines the Norway overview map with the two study-area elevation maps. The elevation data folders must be present inside project_dir. If the elevation tiles are zipped within those folders, the script unzips them into an unzipped/ subfolder before mosaicking.

Main input files and folders used by this script:

• cascades20240801/deployments.csv
• Norway_sf.rds
• Norway_shapefile.rds
• Vestland_elev_data/
• Agder_elev_data/

Main R package dependencies:

terra
sf
fs
dplyr
ggplot2
ggspatial
patchwork
ggpubr
grid

---

Recommended R read-in code

Use BOM-safe CSV read-in for the Excel-friendly CSV files:

caminfo <- read.csv(
  "Camera_information.csv",
  fileEncoding = "UTF-8-BOM",
  colClasses = "character",
  check.names = FALSE
)

cam_fail <- read.csv(
  "Days_cam_failure.csv",
  fileEncoding = "UTF-8-BOM",
  check.names = FALSE
)

weather2 <- read.csv(
  "Weather_20230301_20240229.csv",
  fileEncoding = "UTF-8-BOM",
  check.names = FALSE
)

wind_daily_cam <- read.csv(
  "era5_wind_daily_per_camera.csv",
  fileEncoding = "UTF-8-BOM",
  check.names = FALSE
)

If column names such as Station are not found after reading, the most likely cause is that the byte-order mark was not handled. Re-read the file using. fileEncoding = "UTF-8-BOM".

---

Reproducibility notes

1. The main scripts build independent detection records from the camtrap-dp package using a 60-minute independence threshold.
2. Camera-failure days are excluded from effort and are not treated as true non-detections.
3. Weather and wind covariates are joined to camera-day data by region/date or camera/date.
4. Elevation data are used only for map generation and are not used as model covariates.
5. The study-area maps expect the elevation folders and RDS spatial files to remain in project_dir.
6. The cascades20240801/ folder should remain intact because datapackage.json points to the associated camtrap-dp tables.
7. All folders should be unpackaged/unzipped before importation. The R scripts expect cascades20240801/, Vestland_elev_data/, and Agder_elev_data/ to be readable as normal folders within project_dir.

---

Quick file checks in R

nrow(read.csv("Camera_information.csv", fileEncoding = "UTF-8-BOM"))
nrow(read.csv("Days_cam_failure.csv", fileEncoding = "UTF-8-BOM"))
nrow(read.csv("Weather_20230301_20240229.csv", fileEncoding = "UTF-8-BOM"))
nrow(read.csv("era5_wind_daily_per_camera.csv", fileEncoding = "UTF-8-BOM"))

norway_outline <- readRDS("Norway_sf.rds")
norway_counties <- readRDS("Norway_shapefile.rds")

nrow(norway_outline)
nrow(norway_counties)
names(norway_counties)