Distance Sampling: Methods and Applications

Abundance estimation is fundamental to wildlife ecology and management, and distance sampling is a major technique used to estimate animal abundance.  Distance sampling has been used at least since the 1960s and this is the third major book on the subject by the lead author.  Although distance sampling has a rich history, rapid and significant developments in this area have continued since Dr. Buckland’s previous book was published in 2004.  A parallel development has been the explosion of open source computer code, particularly using the R language, in ecology.

This new volume by Dr. Buckland and his colleagues at the University of St. Andrews serves as both an introduction to distance sampling for people new to the field as well as a summary of recent developments.  A major strength of the book is that some case studies are accompanied with data sets and R code available from a companion website.

http://www.creem.st-and.ac.uk/DS.M&A/

The book contains twelve chapters divided into three parts.  The early chapters cover survey and experimental design and diagnosing problems in data.  I was pleased to see the material related to experimental design, especially to Before-After-Control-Impact studies.  Estimation of abundance is valuable, but becomes far more so when used in conjunction with an experimental approach.

Distance sampling has progressed so much in the last decade or so that the idea of distance sampling likely possessed by many readers is now classified as conventional distance sampling or multiple covariate distance sampling.  Part II of the book (Chapters 5 – 8) begins by presenting these methods.  This section of the book also covers mark-recapture distance sampling to account for detection < 1 on the line.  Some example data sets are analysed in great detail in this section of the book.  Design-based methods and model-based methods of abundance estimation are presented next.  The former estimates abundance for sample units perhaps together with the Horvitz-Thompson estimator.  Model-based methods estimate abundance for the study area and readily allow modelling of abundance as a function of covariates.  Chapter 7 presents a two-stage model-based approach to abundance estimation using an offset.  The two-stage approach may present problems with error propagation and a single-stage model-based approach is presented in Chapter 8.  Spatial distance sampling is also addressed in these chapters.

The final part of the book addresses such topics as three-dimensional distance sampling, non-random transect placement, movement of animals, availability and accounting for measurement error in the model, as well as new technology being used to detect animals and monitor their populations.  One of the most exciting developments in this section is the combination of capture-recapture and distance sampling methods for spatially-explicit abundance estimation.

R code associated with the book allows the reader to study examples in depth and hopefully strengthen their understanding of the subject.  Together with existing R packages such as admbsecr, dsm, mrds, nupoint and unmarked, this R code should prove quite valuable to students and practitioners.  Additional R code is available in the primary literature, but must be sought out by the reader.  Somewhat surprisingly, the R package admbsecr is not mentioned in the book.

Most, if not all, of the companion R files seem to use the free R package ‘Distance’ and do not seem to require that the free Program DISTANCE be installed on your computer.  I was able to run all of the example R files; R returned only a few errors.  Some path statements and file names had been changed from what was used by the authors and you will have to modify the code and/or file names yourself to get the code to run.  Mostly this involved removing spaces from folder or subfolder names.  However, this is a minor issue, at least for people already familiar with the R language.  I noticed no typographical errors in the text of the book itself.

I recommend the book as an excellent entry into this fascinating area of estimation and research.  The R code alone is worth the purchase price.  However, students likely also will want access to earlier distance sampling books as well as the primary literature for in-depth comprehensive assessments of many topics.  The authors of the present volume themselves frequently refer back to earlier texts.  Interested readers should see also the chapter on distance sampling in the new book Applied Hierarchical Modeling in Ecology by Marc Kéry and Andy Royle (2015).

Book reviewed by Mark Miller