Joint species distribution modelling : with applications in R /
Otso Ovaskainen, Nerea Abrego.
Bok Engelsk 2020 · Electronic books.
| Medvirkende | Abrego, Nerea, (author.)
|
|---|---|
| Omfang | 1 online resource (xvi, 372 pages) : : digital, PDF file(s).
|
| Opplysninger | Title from publisher's bibliographic system (viewed on 18 May 2020).. - Cover -- Half-title -- Series information -- Title page -- Copyright information -- Contents -- Preface -- Acknowledgements -- Part I Introduction to Community Ecology: Theory and Methods -- 1 Historical Development of Community Ecology -- 1.1 What Is Community Ecology? -- 1.2 What Is an Ecological Community? -- 1.3 Early Community Ecology: A Descriptive Science -- 1.4 Emergence of the First Theories -- 1.5 Current Community Ecology: Search for the Unifying Theory -- 1.5.1 The Metacommunity Framework -- 1.5.2 The Assembly Rules Framework -- 1.5.3 Vellend's Theory of Ecological Communities -- 1.5.4 Which Ecological Theories Are Prevailing in the Current Community Ecology Literature? -- 2 Typical Data Collected by Community Ecologists -- 2.1 Community Data -- 2.2 Environmental Data -- 2.3 Spatio-temporal Context -- 2.4 Trait Data -- 2.5 Phylogenetic Data -- 2.6 Some Remarks about How to Organise Data -- 3 Typical Statistical Methods Applied by Community Ecologists -- 3.1 Ordination Methods -- 3.2 Co-occurrence Analysis -- 3.3 Analyses of Diversity Metrics -- 3.4 Species Distribution Modelling -- 4 An Overview of the Structure and Use of HMSC -- 4.1 HMSC Is a Multivariate Hierarchical Generalised Linear Mixed Model -- 4.2 The Overall Structure of HMSC -- 4.3 Linking HMSC to Community Ecology Theory -- 4.4 The Overall Workflow for Applying HMSC -- Part II Building a Joint Species Distribution Model Step by Step -- 5 Single-Species Distribution Modelling -- 5.1 How Do Species Distribution Models Link to Species Niches? -- 5.2 The Linear Model -- 5.2.1 Continuous and Categorical Explanatory Variables -- 5.3 Generalised Linear Models -- 5.3.1 Probit Model for Presence-Absence Data -- 5.3.2 Poisson and Lognormal Poisson Models for Count Data -- 5.3.3 Hurdle Models for Zero-Inflated Data -- 5.4 Mixed Models -- 5.4.1 Hierarchical Study Designs.. - 10.2.1 Exploration of the Raw Data -- 10.2.2 HMSC Analyses with an Ideal Model -- 10.2.3 HMSC Analyses with a Compromised Model -- 10.3 Statistical Analyses of the Time-Series Data Collected by a Virtual Ecologist -- 10.3.1 Exploration of the Raw Data -- 10.3.2 Competitive Interactions Revealed by HMSC Analyses -- 10.4 What Did the Virtual Ecologists Learn from Their Data? -- 11 Illustration of HMSC Analyses: Case Study of Finnish Birds -- 11.1 Steps 1-5 of the HMSC Workflow -- 11.1.1 Step 1. Setting Model Structure and Fitting the Model -- 11.1.2 Step 2. Examining MCMC Convergence -- 11.1.3 Step 3. Evaluating Model Fit and Comparing Models -- 11.1.4 Step 4. Exploring Parameter Estimates -- 11.1.5 Step 5. Making Predictions -- 11.2 Measuring the Level of Statistical Support and Propagating Uncertainty into Predictions -- 11.3 Using HMSC for Conservation Prioritisation -- 11.4 Using HMSC for Bioregionalisation: Regions of Common Profile -- 11.5 Comparing HMSC to Other Statistical Methods in Community Ecology -- 11.5.1 Redundancy Analysis and Variance Partitioning -- 11.5.2 Fourth-Corner Analysis -- 11.5.3 Co-occurrence Analysis -- 11.5.4 Analysis of Species Richness -- Plates -- 12 Conclusions and Future Directions -- 12.1 The Ten Key Strengths of HMSC -- 12.2 Future Development Needs -- 12.2.1 Data Models -- 12.2.2 Computational Efficiency -- 12.2.3 Model Structures Related to Ecological and Evolutionary Processes -- Epilogue -- References -- Index.. - 5.4.2 Spatial and Temporal Study Designs -- 5.4.3 How Do Spatial Structures Link to Ecological Theory? -- 5.5 Partitioning Explained Variation Among Groups of Explanatory Variables -- 5.6 Simulated Case Studies with HMSC -- 5.6.1 Generating Simulated Data -- 5.6.2 Fitting Models and Examining Parameter Estimates -- 5.6.3 Checking MCMC Convergence Diagnostics -- 5.6.4 Checking the Assumptions of the Linear Model -- 5.6.5 Fitting Generalised Linear Models -- 5.6.6 Predicting New Sampling Units -- 5.6.7 Hierarchical Random Effects -- 5.6.8 Evaluation of Model Fit Through Cross-validation -- 5.6.9 Spatial Random Effects -- 5.7 Real Data Case Study with HMSC: The Distribution of Corvus Monedula in Finland -- 5.7.1 Setting up and Fitting HMSC Models -- 5.7.2 Examining what Influences the Distribution of C. Monedula -- 5.7.3 Predicting the Distribution of C. Monedula over Environmental and Spatial Gradients -- 6 Joint Species Distribution Modelling: Variation in Species Niches -- 6.1 Stacked versus Joint Species Distribution Models -- 6.2 Modelling Variation in Species Niches in a Community -- 6.3 Explaining Variation in Species Niches by Their Traits -- 6.4 Explaining Variation in Species Niches by Phylogenetic Relatedness -- 6.5 Explaining Variation in Species Niches by Both Traits and Phylogeny -- 6.6 Simulated Case Studies with HMSC -- 6.6.1 Simulating Species Niches -- 6.6.2 Simulating Species Data -- 6.6.3 Exploring the Raw Data -- 6.6.4 Fitting an HMSC Model for the Community A with Phylogenetically Structured Species Niches -- 6.6.5 Explanatory and Predictive Powers of the HMSC Model -- 6.6.6 Examining Parameter Estimates -- 6.6.7 Does Including Traits and Phylogenies Help Make Better Predictions? -- 6.6.8 Repeating the Analyses for the Community B Where Species Niches Are Structured by Their Traits.. - 6.7 Real Case Study with HMSC: How Do Plant Traits Influence Their Distribution? -- 6.7.1 Setting up and Fitting HMSC Models -- 6.7.2 Do Species that Occur on Dry, Warm Sites Have a High Carbon-to-Nitrogen Ratio? -- 7 Joint Species Distribution Modelling: Biotic Interactions -- 7.1 Strategies for Estimating Biotic Interactions in Species Distribution Models -- 7.2 Occurrence and Co-occurrence Probabilities -- 7.2.1 Raw versus Residual Co-occurrence -- 7.2.2 Species Association: Co-occurrence or Co-variation in Abundance -- 7.3 Using Latent Variables to Model Co-occurrence -- 7.3.1 The Simplest Case: Co-occurrence of Two Species -- 7.3.2 The Full Story: Co-occurrence of Many Species -- 7.4 Accounting for the Spatio-temporal Context through Latent Variables -- 7.4.1 Hierarchical Latent Variables -- 7.4.2 Spatial and Temporal Latent Variables -- 7.4.3 Multiple Random Effects in the Same Model -- 7.5 Covariate-Dependent Species Associations -- 7.6 A Cautionary Note about Interpreting Residual Associations as Biotic Interactions -- 7.7 Using Residual Species Associations for Making Improved Predictions -- 7.7.1 Conditional Prediction -- 7.7.2 Conditional versus Unconditional Cross-validation -- 7.8 Simulated Case Studies with HMSC -- 7.8.1 Generating Simulated Data -- 7.8.2 Defining and Fitting Three Alternative HMSC Models -- 7.8.3 Parameter Estimates in the HMSC Models -- 7.8.4 Explanatory Power, Predictive Power and Conditional Predictive Power -- 7.9 Real Case Study with HMSC: Sequencing Data on Dead Wood-Inhabiting Fungi -- 7.9.1 The Data and the Ecological Context -- 7.9.2 Fitting Six Alternative HMSC Models to the Data -- 7.9.3 Inference on Abiotic and Biotic Species Niches -- 7.9.4 Latent Variables as Model-Based Ordination -- 8 Bayesian Inference in HMSC -- 8.1 The Core HMSC Model -- 8.1.1 Fixed Effects -- 8.1.2 Random Effects.. - 8.1.3 Data Models -- 8.1.4 The Vector of All HMSC Parameters -- 8.2 Basics of Bayesian Inference: Prior and Posterior Distributions and Likelihood of Data -- 8.3 The Prior Distribution of Species Niches -- 8.3.1 Scaling of Data Matrices to Make Default Priors Generally Applicable -- 8.4 The Prior Distribution of Species Associations -- 8.4.1 The Prior for Site Loadings -- 8.4.2 The Multiplicative Gamma Process Shrinking Prior for Species Loadings -- 8.4.3 How Many Factors Are Needed? -- 8.5 The Prior Distribution of Data Models -- 8.6 What HMSC Users Need and Do Not Need to Know about Posterior Sampling -- 8.7 Sampling from the Prior with HMSC -- 8.7.1 If There Are No Data, the Posterior Distribution Equals the Prior Distribution -- 8.8 How Long Does It Take to Fit an HMSC Model? -- 9 Evaluating Model Fit and Selecting among Multiple Models -- 9.1 Preselection of Candidate Models -- 9.2 The Many Ways of Measuring Model Fit -- 9.2.1 Accuracy, Discrimination Power, Calibration and Precision -- 9.2.2 Evaluating Model Fit for Different Types of Predictions -- 9.3 The Widely Applicable Information Criterion (WAIC) -- 9.4 Variable Selection by a Spike and Slab Prior -- 9.4.1 Selecting Variables Jointly for All Species or Individually for Each Species -- 9.4.2 Simulated Case Study with HMSC -- 9.5 Reduced Rank Regression (RRR) -- 9.5.1 Simulated Case Study with HMSC -- Part III Applications and Perspectives -- 10 Linking HMSC Back to Community Assembly Processes -- 10.1 Simulating an Agent-Based Model of a Competitive Metacommunity -- 10.1.1 Environmental Variation -- 10.1.2 Species Traits -- 10.1.3 The Metacommunity Model -- 10.1.4 Simulated Metacommunity Dynamics: The Underlying Reality -- 10.1.5 Sampling Data: Observations Made by Virtual Ecologists -- 10.2 Statistical Analyses of the Spatial Data Collected by a Virtual Ecologist.. - Joint species distribution modelling (JSDM) is a fast-developing field and promises to revolutionise how data on ecological communities are analysed and interpreted. Written for both readers with a limited statistical background, and those with statistical expertise, this book provides a comprehensive account of JSDM. It enables readers to integrate data on species abundances, environmental covariates, species traits, phylogenetic relationships, and the spatio-temporal context in which the data have been acquired. Step-by-step coverage of the full technical detail of statistical methods is provided, as well as advice on interpreting results of statistical analyses in the broader context of modern community ecology theory. With the advantage of numerous example R-scripts, this is an ideal guide to help graduate students and researchers learn how to conduct and interpret statistical analyses in practice with the R-package Hmsc, providing a fast starting point for applying joint species distribution modelling to their own data.
|
| Emner | |
| Sjanger | |
| Dewey | |
| ISBN | 9781108492461 (hbk.) : : £89.99. - 9781108716789 (pbk.) : : £34.99. - £140.00
|
| ISBN(galt) | 9781108591720 (PDF ebook) :
|
