Integrated Population Models Theory and Ecological Applications with R and JAGS

Front cover
Integrated population models: theory and ecological applications with r and jags
Integrated population models: theory and ecological applications with r and jags
Copyright
Contents
Foreword
Preface
	Who should read this book?
	Conventions in this book
	Computing
	The ipmbook package
	Book web page
Acknowledgments
	Special thanks by michael
	Special thanks by marc
	Literature cited
1 - Introduction
	1.1 Population modeling in population ecology and management
	1.2 The two-step approach to population modeling
	1.3 Integrated population models
	1.4 Developing integrated population models with the bugs language
	1.5 This book
		1.5.1 Why this book?
		1.5.2 Structure and overview of this book
		1.5.3 The importance of simulation
		1.5.4 Use of this book in courses and for teaching
1 - Theory of integrated population models
	2 - Bayesian statistical modeling using jags
		2.1 Introduction
		2.2 Parametric statistical modeling
			2.2.1 Description of chance processes in probability
			2.2.2 Parametric statistical models for inference about chance processes
		2.3 Maximum likelihood estimation in a nutshell
		2.4 Bayesian inference
		2.5 Bayesian computation
		2.6 Bugs software: winbugs, openbugs, jags, and nimble
		2.7 Using jags to fit simple statistical models from r: generalized linear and generalized linear mixed models
			2.7.1 Poisson generalized linear models
			2.7.2 Bernoulli generalized linear models
			2.7.3 Binomial generalized linear models
			2.7.4 Multinomial generalized linear models
			2.7.5 Categorical generalized linear models
			2.7.6 Normal linear regression or gaussian generalized linear models
			2.7.7 Generalized linear models with gaussian random effects
		2.8 Fitting general integrated models in jags
		2.9 Why we have become bayesians…
		2.10 Summary and outlook
		2.11 Exercises
	3 - Introduction to stage-structured population models
		3.1 Introduction
		3.2 Age- and stage-structured population models
			3.2.1 From a life-cycle graph to population equations
			3.2.2 Age-structured pre-birth-pulse model
			3.2.3 Stage-structured pre-birth-pulse model
			3.2.4 Age-structured post-birth-pulse model
			3.2.5 Stage-structured post-birth-pulse model
		3.3 Classical analysis of a matrix population model
			3.3.1 Analysis of a matrix population model without stochasticity and parameter uncertainty
			3.3.2 Analysis of a matrix population model with parameter uncertainty
			3.3.3 Analysis of a matrix population model with environmental stochasticity
			3.3.4 Analysis of a matrix population model with demographic stochasticity
			3.3.5 Analysis of a matrix population model with multiple sources of stochasticity and parameter uncertainty
			3.3.6 Matrix population models with density dependence and demographic stochasticity
		3.4 Analysis of matrix population models with markov chain monte carlo software
			3.4.1 Analysis of a matrix population model without stochasticity and parameter uncertainty
			3.4.2 Analysis of a matrix population model with parameter uncertainty
			3.4.3 Analysis of a matrix population model with environmental stochasticity
			3.4.4 Analysis of a matrix population model with demographic stochasticity
			3.4.5 Analysis of a matrix population model with multiple sources of stochasticity and parameter uncertainty
			3.4.6 Matrix population models with density dependence and demographic stochasticity
		3.5 Summary and outlook
		3.6 Exercises
	4 - Components of integrated population models
		4.1 Introduction
		4.2 Overview of the key types of data and associated models that go into an ipm
			4.2.1 Levels of information and data aggregation
			4.2.2 Observation or measurement errors in our data sets
			4.2.3 Levels of aggregation and measurement error in population size survey data
			4.2.4 Levels of aggregation and measurement error in productivity survey data
			4.2.5 Levels of aggregation and measurement error in survival survey data
		4.3 Models for population size surveys
			4.3.1 Gaussian state-space models
			4.3.2 Effects of “evil” patterns in the measurement error of a gaussian state-space model
			4.3.3 Use of estimates from another analysis in a gaussian state-space model or an ipm
			4.3.4 Correction of population count data for coverage bias and detection bias
			4.3.5 Transitioning from gaussian to discrete-valued state-space models for population counts
			4.3.6 The “demographic” state-space model of dail and madsen
		4.4 Models for productivity surveys
			4.4.1 Poisson models for brood size data
			4.4.2 Zero inflation in brood size data
			4.4.3 Zero truncation in brood size data
			4.4.4 Censoring in brood size data
			4.4.5 Underdispersion
			4.4.6 Nest survival models
		4.5 Models for survival surveys
			4.5.1 Cormack-jolly-seber model for capture-recapture data
				4.5.1.1 State-space formulation
				4.5.1.2 Multinomial formulation
			4.5.2 Multistate capture-recapture models
				4.5.2.1 State-space formulation
				4.5.2.2 Multinomial formulation
			4.5.3 Dead-recovery data
				4.5.3.1 State-space formulation
				4.5.3.2 Multinomial formulation
			4.5.4 Joint analysis of capture-recapture and dead-recovery data
			4.5.5 Multievent models
		4.6 Introduction to spatial capture-recapture modeling
		4.7 Summary and outlook
		4.8 Exercises
	5 - Introduction to integrated population models
		5.1 Introduction
		5.2 Feeding demographic data into the analysis of a matrix population model
			5.2.1 Using capture-recapture data in a matrix population model
			5.2.2 Combining capture-recapture and productivity data in a matrix population model
		5.3 Our first integrated population model
		5.4 The three-step approach to integrated population modeling
			5.4.1 Development of a model that links demographic data with population size
			5.4.2 Formulation of the likelihood for each available data set separately
			5.4.3 Formulation of the joint likelihood
			5.4.4 Writing the bugs code for the integrated population model
		5.5 Simulation assessment of a simple integrated population model
			5.5.1 Simulating data under an integrated population model
			5.5.2 Simulation results
		5.6 Outlook and summary
		5.7 Exercises
	6 - Benefits of integrated population modeling
		6.1 Introduction
		6.2 Parameter estimates with increased precision
			6.2.1 Experiencing a gain in precision in a simple simulation
			6.2.2 Where does the information come from?
		6.3 Estimation of demographic parameters for which there are no explicit data
		6.4 Estimation of process correlation among demographic parameters
		6.5 Estimation of population structure
		6.6 Flexibility
			6.6.1 Diversity of data types combined in an ipm
			6.6.2 Unequal temporal coverage of data sets—missing values in certain years
			6.6.3 Time points of data collection do not match
			6.6.4 Using estimated indices instead of counts for population-level data
			6.6.5 Observation models for population-level data
			6.6.6 Informative priors and sequential analyses
		6.7 Summary and outlook
		6.8 Exercises
	7 - Assessment of integrated population models
		7.1 Introduction
		7.2 Assumptions of integrated population models
			7.2.1 Assumptions made for component data likelihoods
				7.2.1.1 Principle of posterior predictive checks
				7.2.1.2 Application of posterior predictive checks
				7.2.1.3 Sensitivity of posterior predictive checks to diagnose misspecified ipms
				7.2.1.4 Posterior predictive checks for ipms with a hidden parameter
			7.2.2 The independence assumption
			7.2.3 The common demography assumption
			7.2.4 Conclusions about integrated population model assumptions
		7.3 Under- and overfitting
		7.4 Effects of a misspecified observation model
		7.5 Outlook and summary
		7.6 Exercises
	8 - Integrated population models with density dependence
		8.1 Introduction
		8.2 Density dependence in red-backed shrikes
			8.2.1 General population model
			8.2.2 Modeling density dependence in survival and productivity
			8.2.3 Assessing density dependence at the population level
			8.2.4 Modeling density dependence in immigration
		8.3 Advantages of ipms for the study of density dependence
		8.4 Summary and outlook
		8.5 Exercises
	9 - Retrospective population analyses
		9.1 Introduction
		9.2 Correlations between demographic rates and population growth
		9.3 Life-table response experiments
		9.4 Transient life-table response experiments
		9.5 Summary and outlook
		9.6 Exercises
	10 - Population viability analysis
		10.1 Introduction
		10.2 Challenges for demographic population viability analysis
		10.3 Bayesian population viability analysis
		10.4 Use of an integrated population model in population viability analysis
		10.5 A population viability analysis for simulated woodchat shrike data
			10.5.1 Estimation of extinction probability and related quantities
			10.5.2 Comparison of different management options
		10.6 Population viability analysis of a population with immigration
		10.7 Summary and outlook
		10.8 Exercises
2 - Integrated population models in practice
	11 - Woodchat shrike
		11.1 Introduction
		11.2 Data sets
		11.3 Population model
		11.4 Component data likelihoods
			11.4.1 Population count data
			11.4.2 Productivity data
			11.4.3 Capture-recapture data
		11.5 The integrated population model
		11.6 Results
		11.7 More parsimonious models
		11.8 Discussion
	12 - Peregrine falcon
		12.1 Introduction
		12.2 Data sets
		12.3 Population model
		12.4 Component data likelihoods
			12.4.1 Population count data
			12.4.2 Productivity data
			12.4.3 Dead-recovery data
		12.5 The integrated population model
		12.6 Results
		12.7 Discussion
		Dedication
	13 - Horseshoe bat
		13.1 Introduction
		13.2 Data sets
		13.3 Population model
		13.4 Single data likelihoods
			13.4.1 Capture-recapture data
			13.4.2 Juvenile and population count data
		13.5 The integrated population models
		13.6 Results
		13.7 Prior sensitivity analysis
		13.8 Discussion
	14 - Hoopoe
		14.1 Introduction
		14.2 Data sets
		14.3 Population model
		14.4 Component data likelihoods
			14.4.1 Population count data
			14.4.2 Capture-recapture data
			14.4.3 Productivity data
		14.5 Integrated population model
		14.6 Results
		14.7 Discussion
	15 - Black grouse
		15.1 Introduction
		15.2 Data sets
		15.3 Population model
		15.4 Component data likelihoods
			15.4.1 Population count data
			15.4.2 Radio tracking data
			15.4.3 Modeling productivity and chick sex ratio
		15.5 Integrated population model
		15.6 Results
		15.7 Discussion
	16 - Barn swallow
		16.1 Introduction
		16.2 Data sets
		16.3 Population model
		16.4 Component data likelihoods
			16.4.1 Population count data
			16.4.2 Productivity data
			16.4.3 Capture-recapture data
		16.5 The integrated population model
		16.6 Results
		16.7 Discussion
	17 - Elk
		17.1 Introduction
		17.2 Elk in idaho
		17.3 Population model
		17.4 Component data likelihoods
			17.4.1 Age-at-harvest data
			17.4.2 Hunter survey data
			17.4.3 Radio tracking data
		17.5 The integrated population model
		17.6 Results on elk population dynamics
		17.7 Prior sensitivity analysis
		17.8 Specification of the survival process with hazard rates
		17.9 Discussion
	18 - Cormorant
		18.1 Introduction
		18.2 Data sets
		18.3 Population model
		18.4 Component data likelihoods
			18.4.1 Population count data
			18.4.2 Multistate capture-recapture data
		18.5 The integrated population model
		18.6 Results
		18.7 Discussion
	19 - Gray catbird
		19.1 Introduction
		19.2 Data sets
		19.3 Population model
		19.4 Component data likelihoods
			19.4.1 Population count data (bbs data)
			19.4.2 Capture-recapture data (maps data)
		19.5 The integrated population model
		19.6 Results
		19.7 Discussion
	20 - Kestrel
		20.1 Introduction
		20.2 Data sets
		20.3 Population model
		20.4 Component data likelihoods
			20.4.1 Monitoring häufige brutvögel population count data
			20.4.2 Atlas population count data
			20.4.3 Dead-recovery data
			20.4.4 Basis function approach to the modeling of spatial autocorrelation
			20.4.5 Scaling the modeled population size to the nominal 1km2 area
		20.5 The integrated population model
		20.6 Results
		20.7 Discussion
			20.7.1 The path to landscape demography
			20.7.2 The spatial field approach for an integrated population model using large-scale monitoring data
			20.7.3 Comparison of different spatial integrated population models
			20.7.4 Process-based spatial integrated population models
			20.7.5 Alternative models for the kestrels
			20.7.6 Comments on use of the dail-madsen model as the core of a spatial ipm
			20.7.7 Further possible work with the kestrel ipm
	21 - Black bear
		21.1 Introduction
		21.2 Data sets
		21.3 Population model
		21.4 Component data likelihoods
			21.4.1 Spatial capture-recapture data
			21.4.2 Occupancy data
		21.5 The integrated population model
		21.6 Results
		21.7 Discussion
	22 - Conclusions
		22.1 Fitting integrated population models: a steep mountain … but one that\'s really worth the climb!
		22.2 Should we always integrate?
		22.3 The great importance of long-term ecological research
		22.4 Future directions in integrated population modeling
			22.4.1 Increased spatialization of integrated population models
			22.4.2 Better representation of individual heterogeneity
			22.4.3 Finer temporal scales
			22.4.4 Multiple species
			22.4.5 Better observation models for population count data
			22.4.6 Improving sampling designs for integrated population models
			22.4.7 Statistical and computational advances
		22.5 Concluding remarks
References
Author index
	A
	B
	C
	D
	E
	F
	G
	H
	I
	J
	K
	L
	M
	N
	O
	P
	Q
	R
	S
	T
	U
	V
	W
	Y
	Z
Subject index
	A
	B
	C
	D
	E
	F
	G
	H
	I
	J
	K
	L
	M
	N
	O
	P
	Q
	R
	S
	T
	U
	V
	W
	Z
Back cover