Marc Kéry's books are as important for learning (and teaching) hierarchical modeling as Crawley's The R Book is for learning R. I hold Kéry's Introduction to WinBUGS high for the uncompromising didactic clarity. J. Andrew Royle is one of the founding minds (with James Nichols and Darryl MacKenzie) of the so called occupancy modeling, and he is the author (with Robert Dorazio) of Hierarchical Modeling and Inference in Ecology.
So here come Marc Kéry and J. Andrew Royle with their brand new Applied Hierarchical Modeling in Ecology (Academic Press & Elsevier, 2016). To cite the authors: How can the 700+ pages (or 1500 if you count the upcoming volume 2) contribute anything new to what has already been written on the topic?
My answer: It is one of the rare books where every piece of (clearly explained) theory comes with complete recipes implemented in BUGS and R – hence the word Applied in the title. Apart from the usual model implementation and fitting, there is an extra emphasis on evaluation of the models, on using them for prediction, and on data simulation. Further, the authors make the effort to point out links and similarities between different types of models, and that helps in putting things in context, and to realize that all hierarchical models are, in essence, the same thing. Finally, even though many of the topics of the book are treated somewhere else, the present book does provide a clear update of the most recent developments, and it does introduce new software tools, such as unmarked.
The first 150 pages are essentially a primer on general statistical modeling. A bit like Bolker's Ecological Models done the Crawley way.
Follows 500 pages of exposing specific ecological models, which are:
- Models of abundance using N-mixture models
- Models of abundance using hierarchical distance sampling
- Static occupancy models
- Hierachical models of whole species communities
The last one includes novel methods that deal with metacommunity dynamics and spatial scale. More will come in the 2nd volume – and I wonder: Will the next volume involve ecological models that incorporate species traits, evolution, or phylogeny? That would be great.
I also think that the book is not for anyone. Obviously, to use the methods, you will need to fiddle with code and you will have to adopt the creative way to hierarchical models, since the content is, first of all, an inspiration rather than a prescription. This may sound trivial for an average Bayesian enthusiast, but less trivial for a SAS's MIXED user with a medical science background. Also, if you are a novice and you are struggling with the very basics of Bayes and hierarchical modeling, I'd rather recommend the more basic Introduction to WinBUGS, or the “Puppy Book” by John Kruschke. If you are more into the theoretical underpinnings, I'd also consider Bayesian Data Analysis, or Bolker's Ecological Models.
But for me, the book is a feast of ideas. So now, if you excuse me, I need to get back to R and fiddle with the little gem that I have found on page 110: Did you know that, when assuming a Poisson point process, species abundance can be retrieved from detection/non-detection data (p. 110)? Cool, isn't it?
Acknowledgements: I am grateful to Marc Kéry for sending me a free copy of the book.
Reference: Kéry M. & Royle J.A. (2016) Applied Hierarchical Modeling in Ecology: Analysis of distribution, abundance and species richness in R and BUGS. 1st Edition. Academic Press & Elsevier. 808 pages.