10.5061/DRYAD.9ZW3R229W
Jacobson, Eiren
0000-0003-0147-8367
University of St Andrews
Boyd, Charlotte
University of Washington
McGuire, Tamara
Cook Inlet Beluga Whale Photo ID Project-Alaska WildLife Alliance*
Shelden, Kim
National Oceanic and Atmospheric Administration
Himes Boor, Gina
0000-0003-3924-5198
Montana State University
Punt, André
University of Washington
Data from: Assessing cetacean populations using integrated population
models: an example with Cook Inlet beluga whales
Dryad
dataset
2020
Bayesian statistics
Beluga whales
cetaceans
Integrated modelling
state-space models
North Pacific Research Board
https://ror.org/02nqgka20
1718
2020-02-12T00:00:00Z
2020-02-12T00:00:00Z
en
https://doi.org/10.1002/eap.2114
122982 bytes
3
CC0 1.0 Universal (CC0 1.0) Public Domain Dedication
Effective conservation and management of animal populations requires
knowledge of abundance and trends. For many species, these quantities are
estimated using systematic visual surveys. Additional individual-level
data are available for some species. Integrated population modelling (IPM)
offers a mechanism for leveraging these datasets into a single estimation
framework. IPMs that incorporate both population- and individual-level
data have previously been developed for birds, but have rarely been
applied to cetaceans. Here, we explore how IPMs can be used to improve the
assessment of cetacean populations. We combined three types of data that
are typically available for cetaceans of conservation concern:
population-level visual survey data, individual-level capture-recapture
data, and data on anthropogenic mortality. We used this IPM to estimate
the population dynamics of the Cook Inlet population of beluga whales
(CIBW; Delphinapterus leucas) as a case study. Our state-space IPM
included a population process model and three observational submodels: 1)
a group detection model to describe group size estimates from aerial
survey data; 2) a capture-recapture model to describe individual
photographic capture-recapture data; and 3) a Poisson regression model to
describe historical hunting data. The IPM produces biologically plausible
estimates of population trajectories consistent with all three datasets.
The estimated population growth rate since 2000 is less than expected for
a recovering population. The estimated juvenile/adult survival rate is
also low compared to other cetacean populations, indicating that low
survival may be impeding recovery. This work demonstrates the value of
integrating various data sources to assess cetacean populations and serves
as an example of how multiple, imperfect datasets can be combined to
improve our understanding of a population of interest. The model framework
is applicable to other cetacean populations and to other taxa for which
similar data types are available.
/Data/CIBW_RSideCapHist_McGuire&Stephens.csv contains a matrix of
right side capture histories (1 = captured, 0 = not captured) for each
individual (rows) and year (columns). Photographic capture-recapture data
were collected by Tamara McGuire. These data are made available here,
without restriction, but anyone wishing to use these data is requested to
contact tamaracookinletbeluga@gmail.com, who can provide further
information on how raw data were processed to provide capture histories.
/Data/CIBW_HuntData_Mahoney&Shelden2000.xlsx contains the minimum
documented number of animals killed (MinKilled) for years between 1950 and
1998 as published in Mahoney and Shelden 2000. Entries which are NA
indicate that no data were available for that year.
/Data/CIBW_Abundance_HobbsEtAl2015.xlsx contains the total group size
estimates from Hobbs et al. 2015. /Data/CIBW_Abundance_BoydEtAl2019.txt
contains an array with dimensions [1:1000, 1:8, 1:11] containing 1000
posterior samples of total group size for up to 8 survey days over 11
years, as described in Boyd et al. 2019.
Models/CIBW_IPM_BetaPObs.txt contains the JAGS model specification as
presented in Appendix Data S1. Scripts/CIBW_01_CalcPObs.R provides the
scripts used to calculate the scale and shape parameters of the beta
distribution used as a prior for the probability of observing belugas, as
described in Appendix S1. Scripts/CIBW_02_LoadData.R includes code for
reading in and formatting data prior to modeling.
Scripts/CIBW_03_RunModel.R runs the Bayesian model using jagsUI and
outputs an RData object.