Simplicity, Complexity and Modelling -- Contents -- Preface -- Acknowledgements -- Contributing authors -- 1 Introduction -- 1.1 The origins of the SCAM project -- 1.2 The scope of modelling in the modern world -- 1.3 The different professions and traditions engaged in modelling -- 1.4 Different types of models -- 1.5 Different purposes for modelling -- 1.6 The purpose of the book -- 1.7 Overview of the chapters -- References -- 2 Statistical model selection -- 2.1 Introduction -- 2.2 Explanation or prediction? -- 2.3 Levels of uncertainty -- 2.4 Bias-variance trade-off -- 2.5 Statistical models -- 2.5.1 Within-model inference -- 2.6 Model comparison -- 2.7 Bayesian model comparison -- 2.7.1 Model uncertainty -- 2.7.2 Laplace approximation -- 2.8 Penalized likelihood -- 2.8.1 Bayesian information criterion -- 2.9 The Akaike information criterion -- 2.9.1 Inconsistency of AIC -- 2.10 Significance testing -- 2.11 Many variables -- 2.12 Data-driven approaches -- 2.12.1 Cross-validation -- 2.12.2 Prequential analysis -- 2.13 Model selection or model averaging? -- References -- 3 Modelling in drug development -- 3.1 Introduction -- 3.2 The nature of drug development and scope for statistical modelling -- 3.3 Simplicity versus complexity in phase III trials -- 3.3.1 The nature of phase III trials -- 3.3.2 The case for simplicity in analysing phase III trials -- 3.3.3 The case for complexity in modelling clinical trials -- 3.4 Some technical issues -- 3.4.1 The effect of covariate adjustment in linear models -- 3.4.2 The effect of covariate adjustment in non-linear models -- 3.4.3 Random effects in multi-centre trials -- 3.4.4 Subgroups and interactions -- 3.4.5 Bayesian approaches -- 3.5 Conclusion -- 3.6 Appendix: The effect of covariate adjustment on the variance multiplier in least squares -- References.

4 Modelling with deterministic computer models -- 4.1 Introduction -- 4.2 Metamodels and emulators for computationally expensive simulators -- 4.2.1 Gaussian processes emulators -- 4.2.2 Multivariate outputs -- 4.3 Uncertainty analysis -- 4.4 Sensitivity analysis -- 4.4.1 Variance-based sensitivity analysis -- 4.4.2 Value of information -- 4.5 Calibration and discrepancy -- 4.6 Discussion -- References -- 5 Modelling future climates -- 5.1 Introduction -- 5.2 What is the risk from climate change? -- 5.3 Climate models -- 5.4 An anatomy of uncertainty -- 5.4.1 Aleatoric uncertainty -- 5.4.2 Epistemic uncertainty -- 5.5 Simplicity and complexity -- 5.6 An example: The collapse of the thermohaline circulation -- 5.7 Conclusions -- References -- 6 Modelling climate change impacts for adaptation assessments -- 6.1 Introduction -- 6.1.1 Climate impact assessment -- 6.2 Modelling climate change impacts: From world development paths to localized impacts -- 6.2.1 Greenhouse gas emissions -- 6.2.2 Climate models -- 6.2.3 Downscaling -- 6.2.4 Regional/local climate change impacts -- 6.3 Discussion -- 6.3.1 Multiple routes of uncertainty assessment -- 6.3.2 What is the appropriate balance between simplicity and complexity? -- References -- 7 Modelling in water distribution systems -- 7.1 Introduction -- 7.2 Water distribution system models -- 7.2.1 Water distribution systems -- 7.2.2 WDS hydraulic models -- 7.2.3 Uncertainty in WDS hydraulic modelling -- 7.3 Calibration of WDS hydraulic models -- 7.3.1 Calibration problem -- 7.3.2 Existing approaches -- 7.3.3 Case study -- 7.4 Sampling design for calibration -- 7.4.1 Sampling design problem -- 7.4.2 Existing approaches -- 7.4.3 Case study -- 7.5 Summary and conclusions -- References -- 8 Modelling for flood risk management -- 8.1 Introduction -- 8.2 Flood risk management -- 8.2.1 Long-term change.

8.2.2 Uncertainty -- 8.3 Multi-purpose management -- 8.4 Modelling for flood risk management -- 8.4.1 Source -- 8.4.2 Pathway -- 8.4.3 Receptors -- 8.4.4 An example of a system model: Towyn -- 8.5 Model choice -- 8.6 Conclusions -- References -- 9 Uncertainty quantification and oil reservoir modelling -- 9.1 Introduction -- 9.2 Bayesian framework -- 9.2.1 Solution errors -- 9.3 Quantifying uncertainty in prediction of oil recovery -- 9.3.1 Stochastic sampling algorithms -- 9.3.2 Computing uncertainties from multiple history matched models -- 9.4 Inverse problems and reservoir model history matching -- 9.4.1 Synthetic problems -- 9.4.2 Imperial college fault model -- 9.4.3 Comparison of algorithms on a real field example -- 9.5 Selecting appropriate detail in models -- 9.5.1 Adaptive multiscale estimation -- 9.5.2 Bayes factors -- 9.5.3 Application of solution error modelling -- 9.6 Summary -- References -- 10 Modelling in radioactive waste disposal -- 10.1 Introduction -- 10.2 The radioactive waste problem -- 10.2.1 What is radioactive waste? -- 10.2.2 How much radioactive waste is there? -- 10.2.3 What are the options for long-term management of radioactive waste? -- 10.3 The treatment of uncertainty in radioactive waste disposal -- 10.3.1 Deep geological disposal -- 10.3.2 Repository performance assessment -- 10.3.3 Modelling -- 10.3.4 Model verification and validation -- 10.3.5 Strategies for dealing with uncertainty -- 10.4 Summary and conclusions -- References -- 11 Issues for modellers -- 11.1 What are models and what are they useful for? -- 11.2 Appropriate levels of complexity -- 11.3 Uncertainty -- 11.3.1 Model inputs and parameter uncertainty -- 11.3.2 Model uncertainty -- References -- Glossary -- Index.