R2ucare: An R package to test the goodness of fit of capture-recapture models

Authors: Olivier Gimenez, Jean-Dominique Lebreton, Rémi Choquet, Roger Pradel

Please email all comments/questions to olivier.gimenez [AT] cefe.cnrs.fr

Citation: to come

What it does (and does not do)

Ths package contains R functions to perform goodness-of-fit tests for capture-recapture models. It also has various functions to manipulate capture-recapture data. Please email all suggestions for improvements, questions, comments and bugs to olivier.gimenez [AT] cefe.cnrs.fr.

For Cormack-Jolly-Seber models (single-state), we refer to Lebreton et al. (1992) and Pradel et al. (2005) for the theory. For Arnason-Schwarz models (multistate), have a look to Pradel et al. (2003). Chapter 5 of the Gentle Introduction to MARK also provides a good start for understanding goodness-of-fit test.

Warning: to date, no goodness-of-fit test exists for models with individual covariates (unless you discretize them and use groups), individual time-varying covariates (unless you treat them as states) or temporal covariates; therefore, remove these covariates from your dataset before using it with R2ucare. For groups, just treat the group separately as in the Dipper example below.

To install the package

This repository hosts the development version of the package. It will also be available soon on CRAN (I have to drastically reduce the to-do list below before submitting it; any help welcome!). For the time being, just use the working version:

if(!require(devtools)) install.packages("devtools")

Despite what its name might suggest, you do not need to download and install U-CARE to run the R2ucare package. This package is basically a Matlab to R translation of U-CARE (Choquet et al. 2009).

Getting started

The simplest way to get started is to have a look to the R2ucare vignette.


Yet to be done

  1. Stuff to write
  2. Not so urgent