2022 October 24th

Data and code from Bergeron et al.: Post-weaning survival in kangaroos is high and constant until survival: implications for population dynamics.
Rachel Bergeron, Gabriel Pigeon, David M. Forsyth, Wendy J. King, and Marco Festa-Bianchet

Data are associated with the long-term study on eastern grey kangaroos (Macropus giganteus) initiated by MFB, WJK, and Graeme Coulson at Wilsons Promontory National Park in Victoria, Australia.
RB, DMF, WJK, MFB, and many other researchers, students, and field assistants participated in data collection.

The R script was compiled by GP and RB.

For any further queries on data or R scripts, please contact rachel.bergeron@usherbrooke.ca.

--------------------

Publication summary:

This article uses long-term capture-mark-recapture data on 920 individuals to investigate the survival component of eastern grey kangaroo population dynamics.
Forage availability and population density were monitored quarterly and included as predictors of survival in a Bayesian state-space formulation of Cormack-Jolly-Seber models.
Results show consistently high prime-aged survival despite large fluctuations in forage and density, and lower and more variable juvenile and senescent survival, as well as significant sexual differences in survival of prime-aged and pre-senescent individuals.

--------------------

Contents: 

1. README.txt: includes a brief summary of the study, as well as a description of data files, variables, and R script.

2. Ten CSV data files accompany this publication.

Note 1: the notation 'Year' in all data files corresponds to the year over which kangaroo survival is estimated, from October 1 of year X-1 to September 30 of year X.
Note 2: individual IDs have been replaced by randomised numbers.

***obsF.csv and obsM.csv***

Each row represents an individual, and each column represents a year, from 2008 to 2020.
Represents the observation process of the Cormack-Jolly-Seber model used to estimate apparent survival.
Denotes whether an individual was observed during the annual August-November field season or not.
0 = individual was not observed on the study area,
1 = individual was observed on the study area.
obsF.csv contains female data, obsM.csv contains male data.

***stateF.csv and stateM.csv***

Each row represents an individual, and each column represents a year, from 2008 to 2020.
Represents the state process of the Cormack-Jolly-Seber model used to estimate apparent survival.
Denotes whether an individual was truly alive in a given year or not.
0 = individual has been found dead and is therefore certainly not alive,
1 = individual was observed this year or in a later year and is therefore necessarily alive,
NA = individual was not observed, state is therefore not known with certainty.
stateF.csv contains female data, stateM.csv contains male data.

***ageF.csv and ageM.csv

Each row represents an individual, and each column represents a year, from 2008-2020.
Denotes each individual's age in a given year, from 0 to 28.
For computational efficiency, age was always entered until the last year of the study, whether an individual was still alive or not. Age hence spans 0-28 here, although no individual has lived longer than 21 years throughout the duration of the study.
Empty rows represent individuals which were first caught as adult, and are therefore of unknown age.
ageF.csv contains female data, ageM.csv contains male data.

***idF.csv and idM.csv***

ID: each represents an individual kangaroo.
first: represents the time step at which a given individual was first captured, from 2008 to 2020. 1 = 2008, 2 = 2009, and so forth, until 13 = 2020.
last: represents the time step at which an individual was last potentially alive, from 2008 to 2020. For any individual that disappeared but was never found dead, or for any individual that died of anthropogenic causes (mostly vehicle collisions or illegal shooting), this would be the last year of the study. 1 = 2008, 2 = 2009, and so forth, until 13 = 2020.
idF.csv contains female data, idM.csv contains male data.

***env.csv and envLAG.csv***

Year: from 2008 to 2020.
Dens and DensSD: population density in kangaroos/ha and associated standard deviation, from distance sampling model.
Veg and VegSD: mean total palatable forage grown from October 1 of the previous year to September 30 of the current year in g/m^2 and associated standard deviation.
resDens and resDensSD: residual density and associated standard deviation obtained by sequential regression of population density in kangaroo/ha on total forage in g/m^2.
env.csv contains environmental data throughout the period over which survival is estimated, envLag.csv contains environmental data lagging the survival period by 6 months.

3. One R script accompanies this publication.

***Roo_NimbleCode.R***

Script used to assemble Nimble data and run a Bayesian state-space formulation of Cormack-Jolly-Seber models
to estimate apparent survival while accounting for imperfect detection.

See Bergeron et al. for more details.