Data and code from: Parental death destabilizes wolf packs despite the benefits of larger pack size
Data files
Oct 09, 2026 version files 65.74 KB
-
Bott_ParentalDeath_Analyses_Dryad.R
36.39 KB
-
Oregon_PackHistoryData.csv
20.47 KB
-
README_ParentalDeath_Dryad.txt
2.08 KB
-
README.md
6.80 KB
Abstract
Cooperative breeding is a fundamental feature of many animal family groups. However, few studies have quantified whether variation in group size meaningfully buffers social and reproductive consequences following parental death. Using 17 years of data from a gray wolf (Canis lupus) population in Oregon, USA, we evaluated how parental death (the death of one or both parents) influenced pack dissolution, reproductive failure, and pup recruitment, and whether larger pack sizes mitigated these effects. Across 303 pack-years, we recorded 69 parental deaths, 39 pack dissolutions, and 112 reproductive failures. Packs that experienced parental death were nearly five times more likely to dissolve (95 % CI: 2.12–11.12), nearly 16 times more likely to fail reproductively (95 % CI: 5.34–47.26), and recruited 66% fewer pups (95 % CI: 49–78 % fewer) than packs in which both parents survived. Larger packs were associated with improved demographic outcomes overall, but these group-size benefits were insufficient to fully offset the elevated risks associated with parental death. Our results show that numerous individuals can provide generalized pack-level demographic resilience but do not functionally replace parental roles, highlighting the importance of evaluating social structure in addition to population size. By clarifying the limits of social buffering in a cooperatively breeding carnivore, this study advances understanding of how social systems respond to parental death.
Dataset DOI: 10.5061/dryad.6m905qggq
Description of the data and file structure
This dataset contains annual pack-level demographic and mortality records for gray wolves (Canis lupus) monitored in Oregon, USA, from 2009 to 2025. Data were collected as part of the Oregon Department of Fish and Wildlife's long-term statewide wolf monitoring program using radio telemetry, remote cameras, direct observations, genetic sampling, and mortality investigations. The dataset comprises 303 pack-years and includes pack identity, year, pack size, parental and non-parental mortality, pack persistence and dissolution, reproductive failure, pup recruitment, and information used in analyses of statewide population growth. The accompanying R script reproduces the analyses and figures associated with the manuscript “Parental death destabilizes wolf packs despite the benefits of larger pack size.”
Files and variables
File: Oregon_PackHistoryData.csv
Description: Annual pack-level records for Oregon gray wolves from 2009 to 2025 (303 pack-years). Each row represents one pack-year. The dataset includes demographic outcomes, mortality records, and statewide minimum population counts.
Variables
- Pack: Name identifying the wolf pack.
- Year: Year identifying the pack-year record.
- FullPackRemoval: Whether the pack was completely removed through agency management actions (1 = yes, 0 = no).
- Dead_Juveniles: Number of documented deaths of juvenile, non-parent wolves.
- Dead_SubAdults: Number of documented deaths of subadult, non-parent wolves.
- Dead_NonBreeding_Adults: Number of documented deaths of nonbreeding adult wolves.
- Any_NonParent_Death: Whether at least one non-parent wolf died (1 = yes, 0 = no).
- ParentDeath: Whether at least one parent died during the pack-year (1 = yes, 0 = no).
- ParentSex: Sex of the deceased parent(s): F = female, M = male, Both = both parents, None = no parental death.
- NumDeadParent: Number of documented parental deaths during the pack-year (0, 1, or 2).
- WinterStart_Year: Calendar year associated with the beginning of the annual winter-survey interval.
- WinterStop_Year: Calendar year associated with the end of the annual winter-survey interval.
- MinWolfCount_Statewide: Annual minimum number of wolves documented statewide in Oregon (individual wolves).
- YearEnd_PackSize: Minimum number of wolves documented in the pack at the end of the biological year.
- YearStart_PackSize: Number of wolves documented in the pack at the start of the biological year, before reproduction.
- Dissolution: Whether the pack dissolved (1 = yes, 0 = no). Dissolution was defined as reduction to fewer than two socially affiliated wolves or disappearance of the pack.
- ReproFailure: Whether the pack experienced reproductive failure (1 = yes, 0 = no), defined as fewer than two surviving pups documented during annual winter monitoring.
- YearEnd_NumPups: Minimum number of surviving pups documented during annual winter monitoring (individual pups).
- ParentDeathSeason: Seasonal category of parental mortality: May-Oct, Nov-April, Unknown, or NoMort (no parental death).
Missing values: Missing or unavailable observations are represented by NA. Numeric variables describing counts are expressed as numbers of individual wolves or pups, except the statewide population count, which represents individual wolves statewide. Binary variables use 0 and 1 as defined above.
File: Bott_ParentalDeath_Analyses_Dryad.R
Description: R script reproducing the manuscript's primary generalized linear mixed models for pack dissolution, reproductive failure, and pup recruitment; supplementary sensitivity and interaction analyses; Bayesian state-space models of statewide population growth; diagnostic procedures; summary tables; and figures. Requires the accompanying Oregon_PackHistoryData.csv file.
File: README_ParentalDeath_Dryad.txt
Description: Accompanying reproducibility documentation describing the required R packages, execution instructions, output directories, model specifications, and convergence warnings associated with selected exploratory and sensitivity analyses.
Code/software
The dataset is provided as a comma-separated values (CSV) file and can be viewed using spreadsheet software or R. All analyses were conducted in R. The accompanying script, Bott_ParentalDeath_Analyses_Dryad.R, reproduces the statistical analyses, diagnostic procedures, summary tables, and figures reported in the manuscript.
Required R packages: tidyverse, glmmTMB, DHARMa, ggeffects, nimble, and coda.
Instructions for reproducing the analyses:
- Install R and the required packages.
- Place
Oregon_PackHistoryData.csvandBott_ParentalDeath_Analyses_Dryad.Rin the same directory. - Set the R working directory to that folder.
- Execute
source("Bott_ParentalDeath_Analyses_Dryad.R"). - Results will be generated in an
outputsdirectory, with subdirectories containing tables, figures, diagnostics, and state-space model results.
The script includes the primary generalized linear mixed models, supplementary sensitivity and interaction analyses, and Bayesian state-space models. The state-space models use three MCMC chains of 60,000 iterations, with 30,000 iterations discarded as burn-in and a thinning interval of 10. The script uses a random seed of 123.
Exploratory autoregressive and zero-inflated pup-recruitment models exhibited convergence problems and were not retained for primary inference. A non-parental mortality sensitivity model for pup recruitment also generated a convergence warning. These limitations are documented in the accompanying reproducibility instructions. The primary pup-recruitment analysis uses a negative binomial generalized linear mixed model, as described in the manuscript.
The script saves R session information, including software and package versions, to outputs/session_info.txt to facilitate reproducibility.
Access information
Other publicly accessible locations of the data:
None. The complete analytical dataset accompanying this manuscript is being archived in the Dryad Digital Repository.
Data were derived from the following sources:
Original wolf monitoring and demographic records collected by the Oregon Department of Fish and Wildlife (ODFW) as part of its statewide wolf monitoring and management program from 2009 to 2025. These records were compiled into the pack-year dataset used for the accompanying manuscript. No external third-party datasets were incorporated.
