Habitat Suitability and Distribution Models: with Applications in R

This book introduces the key stages of niche-based habitat suitability model building, evaluation and prediction required for understanding and predicting future patterns of species and biodiversity.

Beginning with the main theory behind ecological niches and species distributions, the book proceeds through all major steps of model building, from conceptualization and model training to model evaluation and spatio-temporal predictions.

Extensive examples using R support graduate students and researchers to quantify ecological niches and predict species distributions with their own data, and help addressing key environmental and conservation problems.

Reflecting this highly active field of research, the book incorporates the latest developments from informatics and statistics, as well as using data from remote sources such as satellite imagery. This pages contains the codes and supporting material required to run the examples.

Introduction

1 General content of the book
1.1 What is this book about?
1.2 How is the book structured?
1.3 Why write a textbook with R examples?
1.4 What is this book not about?
1.5 Why was this book needed?
1.6 Who is this book for?
1.7 Where can I find supporting material?
1.8 What are readers assumed to know already?
1.9 How does this book differ from previous ones?
1.10 What terminology is used in this book?

PART I - Overview, Principles, Theory and Assumptions Behind Habitat Suitability Modeling

2 Overview of the Habitat Suitability Modeling Procedure
2.1 The different methodological steps of HSM
2.2 The initial conceptual step

3 What Drives Species Distributions?
3.1 The overall context: dispersal, habitat, and biotic filtering
3.2 Speciation, dispersal, species pools, and neutral theory
3.3 The abiotic environment: habitats and fundamental niches
3.4 The biotic environment: species interactions, community assembly, and realized niches
3.5 Further discussion of the realized environmental niche and other related niche concepts

4 From Niche to Distribution: Basic Modeling Principles and Applications
4.1 From geographical distribution to niche quantification
4.2 From the quantified niche to spatial predictions
4.3 From individual species predictions to communities
4.4 Main fields of application

5 Assumptions behind Habitat Suitability Models
5.1 Theoretical assumptions behind HSMs
5.2 Methodological assumptions

PART II - Data Acquisition, Sampling Design, and Spatial Scales

6 Environmental Predictors: Issues of Processing and Selection
6.1 Existing environmental databases
6.2 Performing simple GIS analyses in R
6.3 RS-based predictors
6.4 Properties and selection of variables

7 Species Data: Issues of Acquisition and Design
7.1 Existing data and databases
7.2 Spatial autocorrelation and pseudo-replicates
7.3 Sample size, prevalence, and sample accuracy
7.4 Sampling design and data collection
7.5 Presence–absence vs. presence-only data

8 Ecological Scales: Issues of Resolution and Extent
8.1 Issues of resolution
8.2 Issues of extent

PART III - Modeling Approaches and Model Calibration

9 Envelopes and Distance-based Approaches
9.1 Concepts
9.2 Envelope approaches
9.3 Distance-based methods

10 Regression-based Approaches
10.1 Concepts
10.2 Generalized linear models
10.3 Generalized additive models
10.4 Multivariate adaptive regression splines

11 Classification Approaches and Machine-learning Systems
11.1 Concepts
11.2 Recursive partitioning
11.3 Linear discriminant analysis and extensions
11.4 Artificial neural networks

12 Boosting and Bagging Approaches
12.1 Concepts
12.2 Random forests
12.3 Boosted regression trees

13 Maximum Entropy
13.1 Concepts
13.2 Maxent in R

14 Ensemble Modeling and Modeling Averaging

PART IV - Evaluating Models: Errors and Uncertainty

15 Measuring Model Accuracy: Which Metrics to Use?
15.1 Comparing predicted probabilities of presence to presence–absence observations
15.2 Comparing probabilistic predictions to presence-only observations

16 Assessing Model Performance: Which Data to Use?
16.1 Assessment of model fit using resubstitution and randomization
16.2 Internal evaluation by resampling
16.3 External evaluation (fully independent data)

PART V - Predictions in Space and Time

17 Projecting Models in Space and Time
17.1 Additional considerations and assumptions when projecting models: analog environment, niche completeness, and niche stability
17.2 Projecting species distributions in space
17.3 Projecting species in time
17.4 Ensemble projections

PART VI - Data and Tools Used in this Book, with Developed Case Studies

18 Datasets and Tools Used for the Examples in this Book

19 The Biomod2 Modeling Package Examples
19.1 Example 1: HSM of Protea laurifolia in South Africa
19.2 Example 2: creating diversity maps for the Laurus species

PART VII - Conclusions and Future Perspectives

20 Conclusions and Future Perspectives in Habitat Suitability Modeling
20.1 Further progress in HSMs through metagenomics and remote sensing
20.2 Point-process models for presence-only data
20.3 Hierarchical Bayesian approaches to integrate models at different scales
20.4 Ensemble of small models for rarer species
20.5 Improving the modeling techniques to fit simple and ensemble HSMs
20.6 Multi-species modeling and joint-species distribution modeling
20.7 Use of artificial data

Glossary and Definitions of Terms and Concepts

Methods, approaches, models, techniques, algorithms
ENM, SDM, HSM, etc.: different names and acronyms for the same models!
Environment, habitat, niche, niche-biotope duality, and distribution
Technical acronyms for the most commonly used modeling techniques

Author's Bios

Antoine Guisan, Université de Lausanne, Switzerland

Antoine Guisan is Professor at the Université de Lausanne, Switzerland, where he leads the ECOSPAT Spatial Ecology group. Besides being a specialist in habitat suitability and distribution models, his interests also include ecological niche dynamics in space and time, community and multitrophic modeling, very high resolution spatial modeling in mountain environments, and applications of models to environmental decision-making and transfer of scientific knowledge to society.

Wilfried Thuiller, CNRS, Université Grenoble Alpes

Wilfried Thuiller is a senior scientist at the Centre national de la recherche scientifique (CNRS), Laboratory of Alpine Ecology in Grenoble, France. Besides being a specialist in habitat suitability and distribution models, his interests include macroecology, macroevolution, conservation, biodiversity modeling with both mechanistic and phenomenological models, community ecology, functional ecology, and ecosystem functioning in alpine environments.

Niklaus E. Zimmermann, Swiss Federal Research Institute WSL

Niklaus E. Zimmermann is a senior scientist and directorate member of the Swiss Federal Research Institute WSL, and an adjunct professor at the Swiss Federal Institute of Technology Zurich (ETH Zurich). Besides being a specialist in habitat suitability and distribution models, his interests include macroecology, macroevolution, biodiversity and community modeling using both empirical and mechanistic approaches, as well as conservation and applied biodiversity management support.

How to Order

The copy of the book can be purchased through Cambridge University Press and other standard distribution channels.

Order from Cambridge

Order from Amazon

Order from Barnes & Noble

Downloads

Links to datasets, code and figures

If you use any of these figures in a presentation or lecture, somewhere in your set of slides please add the paragraph: “Some of the figures in this presentation are taken from “Habitat Suitability and Distribution Models: with applications in R. 2017. Cambridge University Press” with permission from the authors: Guisan A., Thuiller W. & Zimmermann N.E.”

If you wish to use any of these figures in a publication, you must get permission from CUP, and each figure must be accompanied by a similar acknowledgement.

Errata

Errata for the 1st Edition, since the 1st printing (September 2017) and not reflected in online version.

This page will include corrections (things that are wrong) and updates (incorrect now, but not at the time the book was published).

Please let us know any mistake you would notice in the book (text or code) by sending me an email with the Subject: “ERROR”

Part 2:

Chapter 6

Page 85 (Correction)

Here, we read long-term (30-year normal) annual climate data for precipitation and temperature from the PRISM project (see Section 6.1.2 ~~6.2.1~~).

Chapter 6

Page 86 (Correction)

One frequently used GIS analysis is to import and overlay field sampled species distribution data with environmental predictor layers to later model their habitat suitability (Part III onward). Here, we import a set of distribution data for Pinus edulis L., downloaded from GBIF (see Section 7.1 ~~6.1.1~~) in July 2014.

Chapter 7

Page 121 (Update)

usa_raster <- rasterize(usa_contin, empty_raster~~, field=”DRAWSEQ”~~)

Part 4:

Chapter 16

Page 290 (Correction). Figure legend.

Figure 16.8 Procedure for the bootstrap approach exemplified on a small number of samples (n = 3) ~~sample containing n = 3 observations~~. Each bootstrap dataset contains n observations, sampled with replacement from the original dataset. Each bootstrap dataset is used to obtain an estimate of ? for evaluating predictive models. Adapted from Hastie et al. (2009) and James et al. (2013), with permission.