Hierarchical modelling and R-package 'unmarked'

By Andy Royle and Richard Chandler, USGS Patuxent Wildlife Research Center

Participant fee: 900 NOK, including full board accomodation at Finse Alpine Research Center. Limited to 25 participants.This workshop is now full! 

Using the R package 'unmarked', this one-day course introduces key hierarchical models used in the analysis of abundance and species occurrence.

This workshop may be combined with the AD Model builder workshop. It can only be combined with the Model Selection workshop if one travels from Finse to Sundvollen with the night train (leaves at 01:38) or if one skips dinner at Finse and takes a taxi from Hønefoss to Sundvollen (leave Finse at 19:00).

Workshop content

Modeling spatial and temporal variation in abundance and occurrence lies at the core of ecology and its applications such as conservation, wildlife management and monitoring science. Many sampling protocols have been devised for obtaining information about species abundance and occurrence when observations are subject to imperfect detection of individuals or species. Examples include occurrence sampling, repeated counts, removal models, double observer models, and distance sampling. Inference about such data is conveniently based on hierarchical models, which include a model of the underlying state variable (eg, presence or absence at a site), and a model of the conditional detection process (eg, probability of detection given presence). The hierarchical modeling framework is also convenient for modeling the state and observation processes using spatial and temporal covariates.

The new R package unmarked (Fiske and Chandler 2011) contains  functions to analyze hierarchical models using likelihood and classical frequentist methods. It includes some classes of models which are not available using any other software package. For example, hierarchical distance sampling models and distance sampling models for open populations. This course introduces key hierarchical models used in the analysis of abundance and species occurrence. We provide an overview of the design and basic functionality of unmarked and provide detailed examples of a number of specific functions including:

  • site-occupancy models (MacKenzie et al. 2002, 2003) for the analysis of species distributions,
  • binomial and multinomial N-mixture (Royle 2004a,b)for the analysis of distribution and abundance,
  • hierarchical distance sampling models (Royle et al. 2004)
  • dynamic models of distribution (MacKenzie et al. 2003) and of abundance (Royle & Dorazio 2008; Dail & Madsen 2011)

This is an intermediate-level workshop with integrated lectures and data analysis. A working knowledge of modern regression methods (GLMs, mixed models) and preferentially of program R or another programming language is required. No previous experience with R or unmarked is assumed; however, it is beneficial.

Please bring your own laptops and install a recent version of R, with the latest version of package unmarked.

Workshop Outline

  • Introduction to hierarchical models and unmarked.
  • Overview of unmarked functionality
  • Formatting data for unmarked
  • Occupancy models
  • Modeling abundance with N-mixture models
  • Modeling abundance with multinomial mixture models
  • Hierarchical distance-sampling models
  • Open population N-mixture models: The Dail-Madsen model.
  • Open population distance-sampling models
  • Open population occupancy models (modeling colonization and extinction)
  • Bonus model

Practical information

Please see the Practical information for the Finse workshops before registering for this workshop.

References

  • Dail, D. and L. Madsen. 2011. Models for estimating abundance from repeated counts of an open metapopulation. Biometrics, 67:577-587.
  • Fiske, I. and R.B. Chandler. 2011. unmarked: An R package for fitting hierarchical models of wildlife occurrence and abundance. Journal of Statistical Software 43:1-23.

  • MacKenzie, D.I., J.D. Nichols, G.B. Lachman, S. Droege, J.A. Royle and C.A. Langtimm. 2002. Estimating site occupancy rates when detection probabilities are less than one. Ecology 83:2248-2255.
  • MacKenzie, D.I., J.D. Nichols, J.E. Hines, M.G. Knutson and A.B. Franklin. 2003. Estimating site occupancy, colonization, and local extinction when a species is detected imperfectly. Ecology 84:2200-2207.

  • Royle, J.A. 2004a. N-Mixture Models for estimating population size from spatially replicated counts. Biometrics, 60(1):108-115.

  • Royle, J.A. 2004b. Generalized estimators of avian abundance from count survey data. Animal Biodiversity and Conservation, 27:375-386.

  • Royle, J.A., D.K. Dawson, and S. Bates. 2004. Modeling abundance effects in distance sampling. Ecology, 85(6):1591-1597.

  • Royle, J.A. and R.M. Dorazio. 2008. Hierarchical Modeling and Inference in Ecology: The Analysis of Data from Populations, Metapopulations, and Communities. Academic Press, San Diego, CA. xviii, 444

Published June 10, 2011 12:04 PM - Last modified Mar. 12, 2013 2:29 PM