Additional Information
Book Details
Abstract
Special Offer: Water Framework Directive Series Set. To buy all four titles including Volume 3 and save £100, visit: http://iwapublishing.com/books/9781780400013/water-framework-directive-series-set
Modelling Aspects of Water Framework Directive Implementation: Volume 1 is a concrete outcome from the Harmoni-CA concerted action as part of a 4-volume series of Guidance Reports that guide water professionals through the implementation process of the Water Framework Directive, with a focus on the use of ICT-tools (and in particular modelling). They are complementary to the Guidance Documents produced by the EU Directorate General for Environment.
Water resources planning and management and the development of appropriate policies require methodologies and tools that are able to support systematic, integrative and multidisciplinary assessments at various scales. It also requires the quantification of various uncertainties in both data and models, and the incorporation of stakeholders participation and institutional mechanisms into the various tools and risk assessment methodologies, to help decision makers understand and evaluate alternative measures and decisions.
The other three volumes in the Water Framework Directive Series are:
- Water Framework Directive: Model supported Implementation - A Water Manager’s Guide edited by Fred Hattermann and Zbigniew W Kundzewicz
- Integrated Assessment for WFD implementation: Data, economic and human dimension - Volume 2, edited by Peter A. Vanrolleghem
- Decision support for WFD implementation - Volume 3, edited by Peter A. Vanrolleghem.
Table of Contents
| Section Title | Page | Action | Price |
|---|---|---|---|
| Half title | 1 | ||
| Title | 3 | ||
| Copyright | 4 | ||
| Contents | 5 | ||
| Authors for this volume | 13 | ||
| Preface | 15 | ||
| Guidance Report I.1 | 19 | ||
| Quality assurance in model-based water management: Better modelling practices | 19 | ||
| 1. WHY QUALITY ASSURANCE FOR MODEL-BASED WATER MANAGEMENT? | 19 | ||
| 1.1. Problems in modelling | 19 | ||
| 1.2. Examples of poor quality of the modelling process | 20 | ||
| 1.2.1. Bloopers in modelling | 20 | ||
| Blooper: Introducing the ‘ditch factor’ in the calibration | 20 | ||
| Blooper: Water sucked out of sea and disappeared up mountain | 21 | ||
| Blooper: Negative concentrations | 21 | ||
| Blooper: Algae refused to grow | 21 | ||
| Blooper: Phosphate and phosphor are two totally different things\". | 21 | ||
| 1.2.2. Lack of documentation of model capabilities and limitations | 22 | ||
| 1.2.3. Miscommunication and lack of public participation | 22 | ||
| Blooper: Model was too good | 22 | ||
| Motivation for developing quality assurance guidelines | 23 | ||
| 1.3. Context and objectives of this guidance document | 23 | ||
| 2. ON QUALITY ASSURANCE | 27 | ||
| 2.1. On quality | 27 | ||
| 2.2. Quality approaches in a historical perspective | 27 | ||
| 2.3. Quality assurance and ISO | 29 | ||
| 2.3.1. The International Organization for Standardization (ISO) | 29 | ||
| 2.3.2. ISO standards history | 29 | ||
| 2.3.3. ISO families of standards | 30 | ||
| 2.3.4. ISO 9000 quality standards in general | 31 | ||
| 2.3.5. Are ISO 9000 quality standards sufficient for modelling? | 32 | ||
| 2.4. On standards and guidelines | 33 | ||
| 2.4.1. Informal standards versus formal standards | 33 | ||
| 2.4.2. ISO and modelling standards | 33 | ||
| 3. STATE-OF-THE-ART QA FOR MODEL-BASED WATER MANAGEMENT | 35 | ||
| 3.1. Quality assurance defined for modelling | 35 | ||
| 3.2. Guiding principles – three different approaches | 35 | ||
| 3.3. Type of QA guidelines for modelling | 37 | ||
| 3.3.1. Classification | 37 | ||
| Type 1: Internal technical guidelines | 37 | ||
| Type 2: Public technical guidelines | 38 | ||
| Type 3: Public interactive guidelines | 38 | ||
| 3.3.2. Development stage and prevalence of QA guidelines | 38 | ||
| 3.4. Existing guidelines | 39 | ||
| 3.5. Discussion of QA in water resources modelling | 39 | ||
| 3.5.1. Key aspects in QA guidelines | 39 | ||
| Guiding principles | 39 | ||
| Interactive guidelines | 39 | ||
| Peer review | 40 | ||
| Transparency and reproducibility | 40 | ||
| Accuracy criteria | 43 | ||
| Uncertainty assessments | 43 | ||
| Model validation | 44 | ||
| Supporting software tools | 44 | ||
| 3.5.2. Organisational requirements for QA guidelines to be effective | 44 | ||
| 4. MODELLING KNOWLEDGE BASE AND SUPPORT TOOL | 45 | ||
| 4.1. QA in the HarmoniQuA approach | 45 | ||
| 4.2. Design considerations of modelling KB and MoST | 47 | ||
| 4.3. Modelling KB in MoST | 48 | ||
| 4.3.1. MoST and Quality Assurance (QA) | 48 | ||
| 4.3.2. Structure and guiding principles of the Knowledge Base (KB) | 49 | ||
| 4.3.3. Terminology and glossary | 51 | ||
| 4.4. Modelling guidance | 52 | ||
| 4.4.1. Decomposition | 52 | ||
| 4.4.2. Steps and tasks | 53 | ||
| 4.4.3. Examples of description of tasks | 54 | ||
| Task 1.4 Document Requirements (Appendix B.3) | 55 | ||
| Task 4.8 Validation (Appendix B.4) | 56 | ||
| 4.4.4. Methods | 57 | ||
| 4.5. Modelling Support Tool, MoST | 58 | ||
| 4.5.1. Introduction to MoST | 58 | ||
| 4.5.2. Starting MoST | 58 | ||
| 4.5.3. Setting-up modelling projects | 59 | ||
| 4.5.4. Guiding modelling projects | 60 | ||
| 4.5.5. Monitor modelling projects | 63 | ||
| 4.5.6. Reporting modelling projects | 67 | ||
| 4.5.7. Training material and help | 68 | ||
| 5. REFERENCES | 69 | ||
| 6. APPENDICES | 71 | ||
| APPENDIX A FUNCTIONAL REQUIREMENTS AND DESIGN | 71 | ||
| A.1. Modelling KB and associated tools | 71 | ||
| A.2. Modelling Support Tool, MoST | 76 | ||
| APPENDIX B DETAILS OF HARMONIQUA’S MODELLING GUIDELINES | 79 | ||
| B.1. Terminology | 79 | ||
| B.2. Steps and tasks of HarmoniQuA’s modelling guidelines | 83 | ||
| B.3. Details of the task ‘Determine requirements’ (Step 1) | 90 | ||
| B.4. Details of the task ‘Validation’ (Step 4) | 98 | ||
| B.5. Methods described in HarmoniQuA’s modelling guidelines | 101 | ||
| Guidance Report I.2 | 111 | ||
| Model calibration and validation in model-based water management | 111 | ||
| ABOUT THIS GUIDELINE | 112 | ||
| 1. EXECUTIVE SUMMARY | 113 | ||
| 2. CALIBRATION FRAMEWORK | 117 | ||
| 3. MODEL PARAMETERISATION AND CHOICE OF CALIBRATION PARAMETERS | 120 | ||
| 3.1. Principle of parameter parsimony | 121 | ||
| 3.2. Parameterisation methodologies | 122 | ||
| 3.3. Sensitivity analysis | 124 | ||
| 3.4. Use of prior information | 126 | ||
| 4. CALIBRATION DATA AND CHOICE OF CALIBRATION OBJECTIVES | 127 | ||
| 4.1. Data requirements | 128 | ||
| 4.2. Numerical performance measures | 129 | ||
| 4.3. Scaling issues | 133 | ||
| 4.4. Weighting observations and aggregating performance measures | 134 | ||
| 5. CALIBRATION METHODS | 137 | ||
| 5.1. Manual calibration | 138 | ||
| 5.2. Automatic calibration | 139 | ||
| 5.3. Expert systems | 140 | ||
| 5.4. Recommendations | 140 | ||
| 6. OPTIMISATION ALGORITHMS | 141 | ||
| 6.1. Optimisation problem | 142 | ||
| 6.2. Local optimisation procedures | 142 | ||
| 6.3. Global optimisation procedures | 144 | ||
| 6.4. Recommendations | 147 | ||
| 6.5. Computational requirements | 150 | ||
| 7. MULTI-OBJECTIVE OPTIMISATION | 151 | ||
| 7.1. Optimisation problem | 152 | ||
| 7.2. Optimisation using aggregation | 153 | ||
| 7.3. Optimisation using Pareto dominance | 154 | ||
| 7.4. Single vs. multi-objective optimisation | 155 | ||
| 8. MODEL VALIDATION | 156 | ||
| 8.1. Hierarchical test scheme | 157 | ||
| 8.1.1. Split-sample test | 157 | ||
| 8.1.2. Differential split-sample test | 160 | ||
| 8.1.3. Proxy-basin test | 160 | ||
| 8.1.4. Proxy-basin, differential split-sample test | 161 | ||
| 8.2. Validation tests based on model residuals | 161 | ||
| 9. UNCERTAINTY ASSESSMENT | 162 | ||
| 9.1. Uncertainty sources | 162 | ||
| 9.2. Parameter non-uniqueness | 163 | ||
| 9.3. Uncertainty propagation | 165 | ||
| 10. APPLICATION EXAMPLES | 166 | ||
| 10.1. Lumped, conceptual rainfall-runoff model (MIKE 11/NAM) | 166 | ||
| 10.1.1. Model description | 166 | ||
| 10.1.2. Model setup | 168 | ||
| 10.1.3. Calibration parameters | 168 | ||
| 10.1.4. Calibration methods | 169 | ||
| SCE multi-objective calibration procedure | 169 | ||
| Knowledge-based expert system | 170 | ||
| 10.1.5. Results | 170 | ||
| 10.2. Groundwater model (MODFLOW) | 172 | ||
| 10.2.1. Model description | 172 | ||
| 10.2.2. Model parameterisation | 173 | ||
| 10.2.3. Calibration objectives | 175 | ||
| 10.2.4. Calibration parameters | 175 | ||
| 10.2.5. Calibration method | 176 | ||
| 10.2.6. Results | 177 | ||
| 10.3. Distributed, integrated model (MIKE SHE) | 178 | ||
| 10.3.1. Model description | 178 | ||
| 10.3.2. Model parameterisation | 179 | ||
| 10.3.3. Calibration objectives | 181 | ||
| 10.3.4. Calibration parameters | 181 | ||
| 10.3.5. Calibration method | 182 | ||
| 10.3.6. Results | 182 | ||
| 10.4. Water quality model (ESWAT) | 183 | ||
| 10.4.1. Model description | 184 | ||
| 10.4.2. Model setup | 185 | ||
| 10.4.3. Calibration objectives | 185 | ||
| 10.4.4. Calibration parameters | 187 | ||
| 10.4.5. Calibration method | 188 | ||
| 10.4.6. Results | 189 | ||
| Step 1: Random parameter sampling | 189 | ||
| Step 2: Optimisation | 190 | ||
| 11. REFERENCES | 193 | ||
| 12. SOFTWARE | 199 | ||
| 12.1. PEST | 199 | ||
| Optimisation Algorithms | 199 | ||
| Key Features | 199 | ||
| Model Interface | 200 | ||
| References | 200 | ||
| Manuals | 200 | ||
| Scientific Journals | 201 | ||
| Download | 201 | ||
| 12.2. UCODE_2005 | 202 | ||
| Optimisation algorithms | 202 | ||
| Key features | 202 | ||
| Model interface | 203 | ||
| References | 203 | ||
| Key references: | 203 | ||
| Applications in different model domains | 204 | ||
| Download | 204 | ||
| 12.3. AUTOCAL | 204 | ||
| Optimisation algorithms | 204 | ||
| Key features | 204 | ||
| Model interface | 205 | ||
| References | 205 | ||
| Download | 206 | ||
| 12.4. GLOBE | 206 | ||
| Optimisation algorithms | 206 | ||
| Key features | 206 | ||
| Model interface | 207 | ||
| References | 207 | ||
| Download | 207 | ||
| Guidance Report I.3 | 209 | ||
| Review of sensitivity analysis methods | 209 | ||
| 1. INTRODUCTION | 210 | ||
| 1.1. Aims and objectives | 210 | ||
| 1.2. Target audience | 210 | ||
| 1.3. Benefits of sensitivity analysis | 210 | ||
| 1.4. What is sensitivity analysis? | 211 | ||
| 1.5. Sensitivity analysis and the Water Framework Directive | 212 | ||
| 1.6. Content and overview | 213 | ||
| 2. DECISION TREE FOR CHOOSING A SENSITIVITY ANALYSIS | 215 | ||
| 2.1.1. Sensitivity aim/setting | 215 | ||
| Factors prioritization (FP) | 215 | ||
| Factors fixing (FF) | 215 | ||
| Factors mapping (FM) | 215 | ||
| Calibration | 216 | ||
| 2.1.2. Computational cost | 217 | ||
| Explanation of fields | 219 | ||
| Can model be run easily multiple times? | 219 | ||
| Model structure is ‘simple’ e.g. Manning equation | 219 | ||
| In depth model knowledge | 219 | ||
| Model has many areas of numerical instability | 219 | ||
| Model has a long runtime and/or many parameters | 219 | ||
| 3. DETAILED METHODOLOGY | 220 | ||
| 3.1. Examples | 222 | ||
| 3.1.1. Example 1 | 222 | ||
| 3.1.2. Example 2 | 222 | ||
| 3.2. Experimental design of SA (sampling) | 223 | ||
| 3.2.1. Screening design | 223 | ||
| 3.2.2. Random sampling | 223 | ||
| 3.2.3. Latin Hypercube sampling (LHS) | 224 | ||
| 3.2.4. Correlation control | 224 | ||
| 3.2.5. Quasi-random sampling with low-discrepancy sequences | 224 | ||
| 3.3. Methods of sensitivity analysis | 225 | ||
| 3.3.1. Graphical methods and visualisation | 225 | ||
| Scatter plots | 226 | ||
| Bar/Histogram/Tornado/Box-Whisker plots/Roses and radar graphs | 227 | ||
| Cobweb plots | 232 | ||
| Time series plots | 232 | ||
| Advantages | 233 | ||
| Disadvantages | 233 | ||
| Key references | 233 | ||
| 3.3.2. Screening methods | 233 | ||
| One-factor-At-a-Time (OAT) | 233 | ||
| Morris | 234 | ||
| Advantages and disadvantages | 235 | ||
| Key References | 235 | ||
| 3.3.3. Local methods | 235 | ||
| The finite difference method | 235 | ||
| Direct method for sensitivity analysis | 236 | ||
| Green function’s method | 236 | ||
| Automatic differentiation | 236 | ||
| Advantages and disadvantages | 237 | ||
| Key references | 237 | ||
| 3.3.4. Global methods | 237 | ||
| Variance based methods | 237 | ||
| FAST | 238 | ||
| The Sobol’ method | 239 | ||
| Advantages | 239 | ||
| Disadvantages | 239 | ||
| Key references | 239 | ||
| Regression analysis | 240 | ||
| Advantages | 242 | ||
| Disadvantages | 242 | ||
| Key references | 242 | ||
| Regionalised sensitivity analysis | 243 | ||
| Advantages | 245 | ||
| Disadvantages | 246 | ||
| Key references | 246 | ||
| Entropy | 246 | ||
| Advantages | 246 | ||
| Disadvantage | 247 | ||
| Key references | 247 | ||
| 3.3.5. Sensitivity and model emulators | 247 | ||
| cut-HDMR | 248 | ||
| RS-HDMR | 248 | ||
| Bayesian and kriging emulators | 248 | ||
| SDP | 249 | ||
| Splines | 250 | ||
| Advantages and disadvantages | 250 | ||
| 4. CASE STUDIES | 250 | ||
| 4.1. Modelling floodplain hydrological processes | 250 | ||
| Short description | 251 | ||
| 4.2. Flood inundation models | 251 | ||
| Similar applications: | 252 | ||
| Short description | 252 | ||
| 4.3. Water quality modelling | 253 | ||
| Short description | 253 | ||
| Conclusions | 258 | ||
| 5. CONCLUSIONS | 258 | ||
| 6. GLOSSARY | 259 | ||
| 7. REFERENCES | 265 | ||
| 8. APPENDIX | 272 | ||
| 8.1. Screening methods | 272 | ||
| 8.1.1. Morris | 272 | ||
| 8.2. Local methods | 273 | ||
| 8.2.1. The finite difference method | 273 | ||
| 8.2.2. Direct method for sensitivity analysis | 274 | ||
| 8.2.3. Green function’s method | 275 | ||
| 8.3. Global methods | 276 | ||
| 8.3.1. Variance based methods | 276 | ||
| FAST | 278 | ||
| Monte Carlo based estimation and the work of Sobol’ | 279 | ||
| 8.3.2. Regression analysis | 281 | ||
| 8.3.3. Regionalised sensitivity analysis | 281 | ||
| 8.3.4. Entropy | 282 | ||
| 8.3.5. Sensitivity and model emulators | 283 | ||
| Regression analysis and the HDMR | 283 | ||
| cut-HDMR | 284 | ||
| RS-HDMR | 284 | ||
| Bayesian emulator | 285 | ||
| SDP | 286 | ||
| Splines | 287 | ||
| Guidance Report I.4 | 289 | ||
| Uncertainty analysis in model-based water management | 289 | ||
| 1. WHY IS UNCERTAINTY ASSESSMENT IMPORTANT? | 289 | ||
| 1.1. Uncertainty and risk in decision making | 289 | ||
| 1.2. Water Framework Directive – requirements | 292 | ||
| 1.3. A motivating example | 293 | ||
| 1.4. Context and objective of this document | 296 | ||
| 2. WHEN IS UNCERTAINTY ASSESSMENT REQUIRED? | 300 | ||
| 2.1. The modelling process | 300 | ||
| 2.2. Uncertainty aspects | 302 | ||
| A. Identify and characterise sources of uncertainty (Figure 1.3) | 302 | ||
| B. Modeller reconsiders uncertainty and performance criteria (Figure 1.3) | 302 | ||
| C. Reviews – dialogue – decisions (Figure 1.3) | 303 | ||
| D. Uncertainty assessment and propagation (Figure 1.3) | 303 | ||
| 3. WHAT IS UNCERTAINTY? | 304 | ||
| 3.1. Definitions | 304 | ||
| Uncertainty | 304 | ||
| Ignorance | 305 | ||
| Risk situations | 305 | ||
| Precaution | 306 | ||
| 3.2. Taxonomy of imperfect knowledge | 307 | ||
| 3.3. Sources of uncertainty | 309 | ||
| 3.4. Nature of uncertainty | 309 | ||
| 3.5. The uncertainty matrix | 310 | ||
| 4. METHODOLOGIES FOR UNCERTAINTY ASSESSMENT | 311 | ||
| 4.1. Data uncertainty | 312 | ||
| Description | 312 | ||
| Resources required | 313 | ||
| Strengths and limitations | 313 | ||
| References | 313 | ||
| 4.2. Error propagation equations | 314 | ||
| Description | 314 | ||
| Resources required | 315 | ||
| Strengths and limitations | 315 | ||
| References | 315 | ||
| 4.3. Expert elicitation | 316 | ||
| Description | 316 | ||
| Resources required | 316 | ||
| Strengths and limitations | 317 | ||
| References | 317 | ||
| 4.4. Extended peer review (review by stakeholders) | 318 | ||
| Description | 318 | ||
| Resources required | 318 | ||
| Strengths and limitations | 318 | ||
| References | 319 | ||
| 4.5. Inverse modelling (parameter estimation) | 319 | ||
| Description | 319 | ||
| Resources required | 320 | ||
| Strengths and limitations | 320 | ||
| References | 320 | ||
| 4.6. Inverse modelling (predictive uncertainty) | 321 | ||
| Description | 321 | ||
| Resources required | 321 | ||
| Strenghts and limitations | 321 | ||
| References | 322 | ||
| 4.7. Monte Carlo Analysis | 322 | ||
| Description | 322 | ||
| Resources required | 323 | ||
| Strengths and limitations | 323 | ||
| References | 323 | ||
| 4.8. Multiple model simulation | 325 | ||
| Description | 324 | ||
| Resources required | 325 | ||
| Strengths and limitations | 325 | ||
| References | 325 | ||
| 4.9. NUSAP | 326 | ||
| Description | 326 | ||
| Resources required | 327 | ||
| Strengths and limitations | 327 | ||
| References | 327 | ||
| 4.10. Quality assurance | 328 | ||
| Description | 328 | ||
| Resources required | 329 | ||
| Strengths and limitations | 329 | ||
| References | 329 | ||
| 4.11. Scenario analysis | 330 | ||
| Description | 330 | ||
| Resources required | 331 | ||
| Strengths and limitations | 331 | ||
| References | 331 | ||
| 4.12. Sensitivity analysis | 332 | ||
| Description | 332 | ||
| Resources required | 332 | ||
| Strengths and limitations | 333 | ||
| References | 333 | ||
| 4.13. Stakeholder involvement | 333 | ||
| Description | 333 | ||
| Resources required | 334 | ||
| Strengths and limitations | 334 | ||
| References | 335 | ||
| 4.14. Uncertainty matrix | 335 | ||
| Description | 335 | ||
| Resources required | 336 | ||
| Strengths and limitations | 336 | ||
| References | 337 | ||
| 5. HOW TO SELECT THE APPROPRIATE METHODOLOGY FOR UNCERTAINTY ASSESSMENT | 337 | ||
| 5.1. Introduction | 337 | ||
| 5.2 Methodologies according to modelling process and level of ambition | 337 | ||
| 5.3. Methodologies according to source and type of uncertainty | 339 | ||
| 6. Illustrative cases | 339 | ||
| 6.1. Case 1: Designing measures – nutrient load/comprehensive modelling | 339 | ||
| 6.2. Case 2: Designing measures – water scarcity/basic modelling | 341 | ||
| 6.3. Case 3: Implementation – Real-time forecasting (of Case 2) | 341 | ||
| 6.4. Case 4: Evaluation – Post project appraisal (of Case 1) | 347 | ||
| 7. REFERENCES | 347 |