Modeling abundance using multinomial N-mixture models

Marc Kery, Swiss Ornithological Inst.



Multinomial N-mixture models are a generalization of the binomial N-mixture models described in Chapter 6 to allow for more complex and informative sampling protocols beyond simple counts. Many commonly used protocols such as multiple observer sampling, removal sampling, and capture-recapture produce a multivariate count frequency that has a multinomial distribution and for which multinomial N-mixture models can be developed. Such protocols typically result in more precise estimates than binomial mixture models because they provide direct information about parameters of the observation process. We demonstrate the analysis of these models in BUGS using several distinct formulations that afford great flexibility in the types of models that can be developed, and we demonstrate likelihood analysis using the unmarked package. Spatially stratified capture-recapture models are one class of models that fall into the multinomial N-mixture framework, and we discuss analysis of stratified versions of classical models such as model Mb, Mh and other classes of models that are only possible to describe within the multinomial N-mixture framework.

Publication type Book chapter
Publication Subtype Book Chapter
Title Modeling abundance using multinomial N-mixture models
DOI 10.1016/B978-0-12-801378-6.00007-2
Year Published 2016
Language English
Publisher Elsevier
Contributing office(s) Patuxent Wildlife Research Center
Description 80 p.
First page 313
Last page 392
Online Only (Y/N) N
Additional Online Files (Y/N) N
Google Analytic Metrics Metrics page
Additional publication details