Muutke küpsiste eelistusi

E-raamat: Using R for Modelling and Quantitative Methods in Fisheries [Taylor & Francis e-raamat]

(University of Tasmania, Australia)
  • Formaat: 352 pages, 38 Tables, black and white; 108 Illustrations, black and white
  • Sari: Chapman & Hall/CRC The R Series
  • Ilmumisaeg: 25-Sep-2020
  • Kirjastus: Chapman & Hall/CRC
  • ISBN-13: 9781003032601
Teised raamatud teemal:
  • Taylor & Francis e-raamat
  • Hind: 276,97 €*
  • * hind, mis tagab piiramatu üheaegsete kasutajate arvuga ligipääsu piiramatuks ajaks
  • Tavahind: 395,67 €
  • Säästad 30%
Using R for Modelling and Quantitative Methods in FisheriesSuurem pilt
  • Formaat: 352 pages, 38 Tables, black and white; 108 Illustrations, black and white
  • Sari: Chapman & Hall/CRC The R Series
  • Ilmumisaeg: 25-Sep-2020
  • Kirjastus: Chapman & Hall/CRC
  • ISBN-13: 9781003032601
Teised raamatud teemal:
Taylor & Francis e-raamatudRohkem infot Taylor & Francis e-raamatute kohta
Raamatu kodulehekülg: https://www.taylorfrancis.com/books/9781003032601

Using R for Modelling and Quantitative Methods in Fisheries has evolved and adapted from an earlier book by the same author and provides a detailed introduction to analytical methods commonly used by fishery scientists, ecologists, and advanced students using the open source software R as a programming tool. Some knowledge of R is assumed, as this is a book about using R, but an introduction to the development and working of functions, and how one can explore the contents of R functions and packages, is provided.

The example analyses proceed step-by-step using code listed in the book and from the book’s companion R package, MQMF, available from GitHub and the standard archive, CRAN. The examples are designed to be simple to modify so the reader can quickly adapt the methods described to use with their own data. A primary aim of the book is to be a useful resource to natural resource practitioners and students.

Features:

  • Model Parameter Estimation

  • provides a detailed explanation of the requirements and steps involved in fitting models to data, using R and, mainly, maximum likelihood methods

  • On Uncertainty

  • uses R to implement bootstrapping, likelihood profiles, asymptotic errors, and Bayesian posteriors to characterize any uncertainty in an analysis. The use of the Monte Carlo Markov Chain methodology is examined in some detail

  • Surplus Production Models

  • applies all the methods examined in the earlier parts of the book to conducting a stock assessment. This included fitting alternative models to the available data, characterizing the uncertainty in different ways, and projecting the optimum models forward in time as the basis for providing useful management advice

Preface xi
1 On Modelling
1(12)
1.1 Characteristics of Mathematical Models
1(4)
1.1.1 General
1(1)
1.1.2 Model Design or Selection
1(2)
1.1.3 Constraints Due to the Model Type
3(1)
1.1.4 Mathematical Models
3(1)
1.1.5 Parameters and Variables
4(1)
1.2 Mathematical Model Properties
5(6)
1.2.1 Deterministic vs Stochastic
5(1)
1.2.2 Continuous vs Discrete Models
6(1)
1.2.3 Descriptive vs Explanatory
7(1)
1.2.4 Testing Explanatory Models
8(1)
1.2.5 Realism vs Generality
9(1)
1.2.6 When Is a Model a Theory?
10(1)
1.3 Concluding Remarks
11(2)
2 A Non-Introduction to R
13(22)
2.1 Introduction
13(1)
2.2 Programming in R
13(12)
2.2.1 Getting Started with R
14(1)
2.2.2 R Packages
15(1)
2.2.3 Getting Started with MQMF
16(1)
2.2.4 Examining Code within Functions
17(1)
2.2.5 Using Functions
18(1)
2.2.6 Random Number Generation
19(2)
2.2.7 Printing in R
21(1)
2.2.8 Plotting in R
21(1)
2.2.9 Dealing with Factors
22(2)
2.2.10 Inputting Data
24(1)
2.3 Writing Functions
25(8)
2.3.1 Simple Functions
25(2)
2.3.2 Function Input Values
27(1)
2.3.3 R Objects
28(1)
2.3.4 Scoping of Objects
28(1)
2.3.5 Function Inputs and Outputs
29(4)
2.4 Appendix: Less-Traveled Functions
33(1)
2.5 Appendix: Additional Learning Resources
33(2)
3 Simple Population Models
35(30)
3.1 Introduction
35(13)
3.1.1 The Discrete Logistic Model
35(2)
3.1.2 Dynamic Behaviour
37(3)
3.1.3 Finding Boundaries between Behaviours
40(2)
3.1.4 Classical Bifurcation Diagram of Chaos
42(1)
3.1.5 The Effect of Fishing on Dynamics
43(3)
3.1.6 Determinism
46(2)
3.2 Age-Structured Modelling Concepts
48(4)
3.2.1 Survivorship in a Cohort
48(1)
3.2.2 Instantaneous vs Annual Mortality Rates
49(3)
3.3 Simple Yield per Recruit
52(8)
3.3.1 Selectivity in Yield-per-Recruit
55(2)
3.3.2 The Baranov Catch Equation
57(2)
3.3.3 Growth and Weight-at-Age
59(1)
3.4 Full Yield-per-Recruit
60(2)
3.5 Concluding Remarks
62(3)
4 Model Parameter Estimation
65(70)
4.1 Introduction
65(2)
4.1.1 Optimization
66(1)
4.2 Criteria of Best Fit
67(2)
4.3 Model Fitting in R
69(4)
4.3.1 Model Requirements
70(1)
4.3.2 A Length-at-Age Example
71(1)
4.3.3 Alternative Models of Growth
72(1)
4.4 Sum of Squared Residual Deviations
73(14)
4.4.1 Assumptions of Least-Squares
74(1)
4.4.2 Numerical Solutions
74(1)
4.4.3 Passing Functions as Arguments to Other Functions
75(2)
4.4.4 Fitting the Models
77(6)
4.4.5 Objective Model Selection
83(1)
4.4.6 The Influence of Residual Error Choice on Model Fit
84(2)
4.4.7 Remarks on Initial Model Fitting
86(1)
4.5 Maximum Likelihood
87(3)
4.5.1 Introductory Examples
87(3)
4.6 Likelihoods from the Normal Distribution
90(8)
4.6.1 Equivalence with Sum-of-Squares
92(2)
4.6.2 Fitting a Model to Data Using Normal Likelihoods
94(4)
4.7 Log-Normal Likelihoods
98(10)
4.7.1 Simplification of Log-Normal Likelihoods
99(1)
4.7.2 Log-Normal Properties
99(3)
4.7.3 Fitting a Curve Using Log-Normal Likelihoods
102(3)
4.7.4 Fitting a Dynamic Model Using Log-Normal Errors
105(3)
4.8 Likelihoods from the Binomial Distribution
108(9)
4.8.1 An Example Using Binomial Likelihoods
109(3)
4.8.2 Open Bay Juvenile Fur Seal Population Size
112(2)
4.8.3 Using Multiple Independent Samples
114(2)
4.8.4 Analytical Approaches
116(1)
4.9 Other Distributions
117(1)
4.10 Likelihoods from the Multinomial Distribution
117(8)
4.10.1 Using the Multinomial Distribution
119(6)
4.11 Likelihoods from the Gamma Distribution
125(2)
4.12 Likelihoods from the Beta Distribution
127(1)
4.13 Bayes' Theorem
128(6)
4.13.1 Introduction
128(2)
4.13.2 Bayesian Methods
130(1)
4.13.3 Prior Probabilities
131(3)
4.14 Concluding Remarks
134(1)
5 Static Models
135(50)
5.1 Introduction
135(1)
5.2 Productivity Parameters
136(1)
5.3 Growth
136(15)
5.3.1 Seasonal Growth Curves
137(5)
5.3.2 Fabens Method with Tagging Data
142(2)
5.3.3 Fitting Models to Tagging Data
144(2)
5.3.4 A Closer Look at the Fabens Methods
146(2)
5.3.5 Implementation of Non-Constant Variances
148(3)
5.4 Objective Model Selection
151(3)
5.4.1 Akiake's Information Criterion
151(2)
5.4.2 Likelihood Ratio Test
153(1)
5.4.3 Caveats on Likelihood Ratio Tests
154(1)
5.5 Remarks on Growth
154(1)
5.6 Maturity
155(10)
5.6.1 Introduction
155(2)
5.6.2 Alternative Maturity Ogives
157(4)
5.6.3 The Assumption of Symmetry
161(4)
5.7 Recruitment
165(10)
5.7.1 Introduction
165(1)
5.7.2 Properties of "Good" Stock Recruitment Relationships
166(1)
5.7.3 Recruitment Overfishing
167(1)
5.7.4 Beverton and Holt Recruitment
168(1)
5.7.5 Ricker Recruitment
169(2)
5.7.6 Deriso's Generalized Model
171(1)
5.7.7 Re-Parameterized Beverton-Holt Equation
172(3)
5.7.8 Re-Parameterized Ricker Equation
175(1)
5.8 Selectivity
175(5)
5.8.1 Introduction
175(1)
5.8.2 Logistic Selection
176(1)
5.8.3 Dome-Shaped Selection
177(3)
5.9 Concluding Remarks for Static Models
180(1)
5.10 Appendix: Derivation of Fabens Transformation
181(1)
5.11 Appendix: Reparameterization of Beverton-Holt
182(3)
6 On Uncertainty
185(68)
6.1 Introduction
185(6)
6.1.1 Types of Uncertainty
185(3)
6.1.2 The Example Model
188(3)
6.2 Bootstrapping
191(1)
6.2.1 Empirical Probability Density Distributions
191(1)
6.3 A Simple Bootstrap Example
192(3)
6.4 Bootstrapping Time-Series Data
195(5)
6.4.1 Parameter Correlation
199(1)
6.5 Asymptotic Errors
200(9)
6.5.1 Uncertainty about the Model Outputs
203(1)
6.5.2 Sampling from a Multivariate Normal Distribution
204(5)
6.6 Likelihood Profiles
209(10)
6.6.1 Likelihood Ratio-Based Confidence Intervals
212(2)
6.6.2 Ve Log-Likelihoods or Likelihoods
214(2)
6.6.3 Percentile Likelihood Profiles for Model Outputs
216(3)
6.7 Bayesian Posterior Distributions
219(17)
6.7.1 Generating the Markov Chain
221(1)
6.7.2 The Starting Point
222(1)
6.7.3 The Burn-In Period
223(1)
6.7.4 Convergence to the Stationary Distribution
223(1)
6.7.5 The Jumping Distribution
224(1)
6.7.6 Application of MCMC to the Example
225(1)
6.7.7 Markov Chain Monte Carlo
225(3)
6.7.8 A First Example of an MCMC
228(7)
6.7.9 Marginal Distributions
235(1)
6.8 The Use of Repp
236(15)
6.8.1 Addressing Vectors and Matrices
239(1)
6.8.2 Replacement for simpspm()
240(3)
6.8.3 Multiple Independent Chains
243(4)
6.8.4 Replicates Required to Avoid Serial Correlation
247(4)
6.9 Concluding Remarks
251(2)
7 Surplus Production Models
253(70)
7.1 Introduction
253(6)
7.1.1 Data Needs
254(1)
7.1.2 The Need for Contrast
254(1)
7.1.3 When Are Catch-Rates Informative?
255(4)
7.2 Some Equations
259(8)
7.2.1 Production Functions
261(2)
7.2.2 The Schaefer Model
263(1)
7.2.3 Sum of Squared Residuals
264(1)
7.2.4 Estimating Management Statistics
265(1)
7.2.5 The Trouble with Equilibria
266(1)
7.3 Model Fitting
267(12)
7.3.1 A Possible Workflow for Stock Assessment
268(6)
7.3.2 Is the Analysis Robust?
274(3)
7.3.3 Using Different Data
277(2)
7.4 Uncertainty
279(29)
7.4.1 Likelihood Profiles
280(3)
7.4.2 Bootstrap Confidence Intervals
283(7)
7.4.3 Parameter Correlations
290(1)
7.4.4 Asymptotic Errors
290(7)
7.4.5 Sometimes Asymptotic Errors Work
297(2)
7.4.6 Bayesian Posteriors
299(9)
7.5 Management Advice
308(2)
7.5.1 Two Views of Risk
308(1)
7.5.2 Harvest Strategies
309(1)
7.6 Risk Assessment Projections
310(9)
7.6.1 Deterministic Projections
310(3)
7.6.2 Accounting for Uncertainty
313(1)
7.6.3 Using Asymptotic Errors
314(2)
7.6.4 Using Bootstrap Parameter Vectors
316(1)
7.6.5 Using Samples from a Bayesian Posterior
316(3)
7.7 Concluding Remarks
319(2)
7.8 Appendix: The Use of Repp to Replace simpspm
321(2)
References 323(12)
Index 335
Dr. Malcolm Haddon has at least 35 years of experience in fisheries science, having worked in the Department of New Zealand Fisheries, the University of Sydney, the Australian Maritime College, the University of Tasmania, and, most recently, in Australias Commonwealth Scientific and Industrial Research Organization (CSIRO), from which he recently retired. He has worked with: Crustacea, including crabs, prawns, and rock lobster; Mollusca, including scallops and abalone; and scale-fish, many and various. Dr. Haddons interests are these days focussed on all aspects of resource assessment and simulation testing of resource management using management strategy evaluation. He considers himself fortunate to have become an adjunct professor in the Institute of Marine and Antarctic Sciences at the University of Tasmania and an Honorary Research Fellow at Oceans and Atmosphere, CSIRO, in Hobart, Tasmania. In both institutions he continues to collaborate with colleagues, most recently beginning to contribute to two research programs at the university on abalone population dynamics and management.
Püsilink: https://www.kriso.ee/db/9781003032601_pe.html
Märksõnad: