Predation events, habitat use, and movement ecology of Atlantic herring on the southwest coast of New Brunswick and Maine

Herring Repository OverviewThis repository contains the R scripts and datasets used to analyze acoustic telemetry tracking data and generate figures for: Notman-Grobler, O.D.P., M.L. Piczak, A.M. Hoffmann, B.M. Wilson, M. Black, R.J. Lennox, G. English, M.J. Lawrence, A. Lacoursière-Roussel, C.W. McKindsey, and M. Trudel (In prep.). Predation events, habitat use, and movement of Atlantic herring on the southwest coast of New Brunswick and Maine. Canadian Journal of Fisheries and Aquatic Sciences. The workflow includes:• Spatial and temporal analysis of predation mortality events identified via sensor temperature spikes• Abacus movement plotting across the seasonal detection window (Julian days 225–335)• Spatial distribution modeling using Generalized Additive Models (GAM) across receiver arrays• Residency index estimation and association with aquaculture vs. non-aquaculture sites• Departure timing and population retention dynamics within Passamaquoddy Bay• Regional detection mapping for post-departure movements outside the Bay of Fundy Repository structureHerring/├── Data/ # Raw and processed telemetry & metadata files│ ├── DETECTIONS/ # Acoustic detection logs (PBSMdata.csv)│ ├── RECEIVER_METADATA/ # Receiver deployment metadata (recmeta.csv)│ ├── TAG_METADATA/ # Tag specifications and deployment details (tagmeta3.csv)│ └── departures.csv # Calculated departure times and retention metrics├── Figures/ # High-resolution output figures (Figure_2.png to Figure_7.png)├── Scripts/ # R processing and visualization scripts (Figure_2.R to Figure_7.R)└── ReadMe.txt Data directory breakdownData/├── DETECTIONS/│ └── PBSMdata.csv # Acoustic detection dataset including timestamps (dt) and sensor data (s)├── RECEIVER_METADATA/│ └── recmeta.csv # Receiver deployment coordinates, stations, and site metadata├── TAG_METADATA/│ └── tagmeta3.csv # Individual tag metadata (biological metrics, release dates, tag IDs)└── departures.csv # Derived summary table for Bay departure dates and retention metrics Scripts and figure mappingScripts/├── Figure_2.R│ Description: Analyzes temperature sensor spikes (> 20°C) to identify predation mortality events of Atlantic herring by meso- and endothermic predators in Passamaquoddy Bay.├── Figure_3.R│ Description: Generates an abacus plot depicting individual detection histories across the Passamaquoddy Bay array for the detection window (Julian days 225–335; August 13 – December 1, 2022).├── Figure_4.R│ Description: Models predicted spatial distribution of tagged herring across months (August to November) using a negative binomial GAM fit to receiver-month detection counts.├── Figure_5.R│ Description: Calculates residency indices (0–1) and analyzes site associations around Passamaquoddy Bay.├── Figure_6.R│ Description: Quantifies departure trends by calculating the total number and proportion of tagged herring remaining within Passamaquoddy Bay over Julian date.└── Figure_7.R Description: Maps unique long-distance detections of Atlantic herring outside the Bay of Fundy during the post-departure window (November 2022 – February 2023). Data availabilityThe raw acoustic telemetry detection files (PBSMdata.csv) and associated metadata are provided in the Data/ folder. Original data available from the Ocean Tracking Network: https://members.oceantrack.org/data/repository/pbsm/ Installation and setup1. Download or clone the Herring repository directory.2. Open R or RStudio.3. Required packages are installed by the script if missing and loaded4. Set the root directory path in the scripts to the location of the repository: rootDir <- "/path/to/Herring" # Linux / macOS rootDir <- "C:/path/to/Herring" # Windows5. GSHHG Coastline is used for Figure 2, 4, and 5 and downloaded automatically from the web: https://www.soest.hawaii.edu/pwessel/gshhg/ Workflow and how to runStep 1 – Predation Analysis and Temperature ProfilesRun: Scripts/Figure_2.R Step 2 – Abacus Plot of Movement TimelinesRun: Scripts/Figure_3.R Step 3 – GAM Spatial Distribution ModelingRun: Scripts/Figure_4.R Step 4 – Residency Index & Aquaculture Site AssociationRun: Scripts/Figure_5.R Step 5 – Bay Departure and Retention DynamicsRun: Scripts/Figure_6.R Step 6 – Regional Post-Departure TrackingRun: Scripts/Figure_7.R Software environmentTested under:• R version 4.3+ / RStudio• Operating Systems: Linux (Debian / Ubuntu) and Windows 10/11• Key R package dependencies: - dplyr / tidyr (Data manipulation) - ggplot2 / grid / gtable (Visualization & layout) - mgcv (GAM modeling) - sf / terra / raster (Spatial data handling) Model assumptions and notes• Post-tagging acclimation window: Initial detections within 48 hours post-tagging (August 12, 2022) are typically filtered prior to behavioral analyses.• Predation threshold: Predation mortality events are flagged when tag sensor temperature (s) exceeds 20°C, corresponding to ingestion by warm-bodied predators.• GAM predictions represent relative detection intensity (fractional expected counts of unique fish per receiver-month), not discrete individual fish counts.• Residency indices are scaled from 0 to 1 based on total array monitoring duration. Important notes• Ensure relative paths within scripts point correctly to the Data/ subfolders. AcknowledgmentChatGPT (OpenAI) and Gemini (Google) were used to assist in cleaning, optimizing, and documenting portions of the code and repository structure. CitationWessel, P., and W. H. F. Smith, A Global Self-consistent, Hierarchical, High-resolution Shoreline Database, J. Geophys. Res., 101, 8741-8743, 1996 ContactFor questions regarding data access, script execution, or manuscript details, please contact the corresponding author.