Risks and Decisions for Conservation and Environmental Management (Ecology, Biodiversity and Conservation)

Emphasizing the philosophy of uncertainty and the frailties of human psychology when people are confronted with risky situations, this book describes how to conduct a thorough environmental risk assessment. Technical methods are provided to help make assessments more objective and less prone to the biases of those involved in the assessment. Consideration is given to the way in which both subjective beliefs and technical analysis may be used to make better informed decisions. Cover......Page 1 Half-title......Page 3 Series-title......Page 4 Title......Page 5 Copyright......Page 6 Contents......Page 7 Preface......Page 11 Acknowledgements......Page 13 1.1 Uncertainty and denial......Page 15 1.2.1 Two kinds of probability......Page 20 1.2.2 Two kinds of subjective probability......Page 21 1.2.4 Probability words......Page 22 1.3 The origin of ideas about risk......Page 24 1.4 Perception......Page 27 1.4.1 Risk aversion and framing......Page 31 1.5.1 Insensitivity to sample size......Page 33 1.5.4 Anchoring......Page 34 1.5.5 Arbitrary risk tolerance......Page 35 1.5.6 Race, religion, culture, gender......Page 36 1.6 Discussion......Page 38 2.1.1 Variability and incertitude......Page 40 2.1.3 Systematic error......Page 41 2.1.4 Natural variation......Page 44 2.1.5 Model uncertainty......Page 45 2.2 Linguistic uncertainty......Page 47 2.2.1 Vagueness......Page 48 2.2.2 Context dependence......Page 49 2.2.3 Ambiguity......Page 50 2.2.4 Underspecificity......Page 51 2.2.5 Indeterminacy......Page 52 2.3 Discussion......Page 53 3 Conventions and the risk management cycle......Page 56 3.1.1 Ecology......Page 58 3.1.3 Ecotoxicology......Page 59 3.1.4 Public health......Page 60 3.1.5 Economics......Page 62 3.1.6 Attributes of risk assessments......Page 63 3.2.1 Selecting endpoints......Page 64 3.2.2 Targeting risk assessments: ecosystems and indicators......Page 65 3.3 The risk management cycle......Page 68 3.3.1 Context: who pays and what do they want?......Page 69 3.3.2 Problem formulation......Page 70 3.3.3 Conceptual modelling and hazard assessment......Page 71 3.3.5 Sensitivity and decision-making......Page 73 3.4 Discussion......Page 74 4 Experts, stakeholders and elicitation......Page 76 4.1 Whos an expert?......Page 79 4.1.1 Legal definitions......Page 80 4.1.2 Courts as gatekeepers......Page 81 4.1.3 Advocacy, adversaries and authority......Page 82 4.1.4 Philosophical definitions......Page 83 4.2 Who should be selected?......Page 84 4.2.1 Examples of expert selection......Page 85 4.2.2 Post hoc calibration......Page 86 4.3 Eliciting conceptual models......Page 87 4.4 Eliciting uncertain parameters......Page 89 4.4.1 Verbal representations......Page 90 4.4.2 Which distribution?......Page 92 4.4.3 Eliciting distributions and tails......Page 93 4.4.4 How hard should the analyst try?......Page 95 4.5.1 Format......Page 96 4.5.3 Overconfidence......Page 98 4.5.4 Motivational bias......Page 102 4.5.6 The conjunction fallacy......Page 104 4.5.7 Cultural, political and philosophical context......Page 107 4.6.1 Performance measures......Page 108 4.6.2 Performance measured......Page 109 4.6.3 Are expert judgements better than lay judgements?......Page 112 4.7 When experts disagree......Page 114 4.8.1 Delphi and its descendants......Page 117 4.8.2 Closure......Page 118 4.8.3 Resolution......Page 120 4.9 Numerical aggregation......Page 121 4.9.2 Bayes’ theorem......Page 122 4.9.3 Mixing and averaging......Page 126 4.10 Combined techniques......Page 127 4.10.2 The ‘Procedures guide’ for structured expert judgement......Page 128 4.11 Using expert opinion......Page 134 4.12 Who’s a stakeholder?......Page 135 4.13 Discussion......Page 138 5.1 Conceptual models......Page 141 5.2 Hazard identification and assessment......Page 144 5.2.1 Checklists and brainstorming......Page 145 5.2.2 Structured brainstorming......Page 146 5.2.3 Hazard matrix......Page 147 5.2.4 Hazard and operability analysis (HAZOP)......Page 149 5.2.5 Failure modes and effects analysis (FMEA)......Page 151 5.2.6 Hierarchical holographic modelling (HHM)......Page 154 5.3 Discussion......Page 156 6.1 Origins of risk ranking methods......Page 159 6.2 Current applications......Page 161 6.3 Conducting a risk ranking analysis......Page 163 6.4.2 Discrete hazards......Page 165 6.4.3 Model complexity......Page 166 6.4.5 Susceptibility to linguistic uncertainty......Page 167 6.4.8 Unacknowledged uncertainty in results......Page 168 6.5 Performance......Page 169 6.6.2 The Paper risk rank calculator......Page 172 6.6.3 Western Rock Lobster ecological risk assessment......Page 174 6.7 Discussion......Page 179 7 Ecotoxicology......Page 183 7.1 Dose–response relationships......Page 184 7.1.1 NOELs and LOELs......Page 186 7.1.2 Odds ratios and relative risks......Page 188 7.2.1 Extrapolating to low concentrations......Page 191 7.2.2 Structure activity relationships (SARs)......Page 193 7.2.4 Extrapolating species sensitivities......Page 194 7.2.5 Extrapolating from acute to chronic effects......Page 200 7.2.6 Extrapolating from toxicity tests to ecological effects......Page 201 7.3 Deciding a safe dose......Page 202 7.3.1 Reference doses, benchmark doses and uncertainty......Page 204 7.3.2 A bootstrap estimate of RfDs......Page 205 7.4 Transport, fate and exposure......Page 206 7.4.1 Exposure assessment......Page 209 7.4.2 Modelling transport, fate and exposure......Page 210 7.4.3 Hazard quotients and mixtures......Page 211 7.4.4 Model-based assessments......Page 212 7.5.2 OECD protection thresholds......Page 213 7.5.3 Cotton pyrethroid risk assessment......Page 214 7.5.4 Integrating pesticide risks in Italy......Page 215 7.5.5 Human health thresholds in Australia......Page 216 7.5.6 Methylmercury in the u’Mgeni River......Page 217 7.6 Discussion......Page 219 8.1 Event trees......Page 221 8.1.1 Decision trees......Page 223 8.1.2 Probabilistic event trees......Page 228 8.1.3 Decision tables and expert systems......Page 229 8.1.4 Classification and regression trees......Page 230 8.1.5 Bayesian networks......Page 232 8.2 Fault trees......Page 237 8.2.1 Probabilistic fault trees......Page 238 8.3.1 Decisions under risk......Page 243 8.3.2 Bayesian decision analysis......Page 245 8.3.3 Gains from management of orange-bellied parrots......Page 246 8.3.4 Interpreting decision trees......Page 247 8.3.5 Conservation status......Page 248 8.4 Discussion......Page 254 9.1 Worst case analysis......Page 256 9.1.1 The trouble with worst case......Page 258 9.1.2 Arbitrary thresholds and acceptable levels of risk......Page 259 9.2 Defining and eliciting intervals......Page 261 9.2.2 Probability intervals......Page 262 9.2.4 Imprecise probabilities......Page 264 9.2.5 Which intervals?......Page 266 9.3 Interval arithmetic......Page 268 9.3.1 Dependencies......Page 269 9.3.3 Intervals for site contamination......Page 270 9.3.4 Intervals for risk ranking......Page 272 9.3.5 Intervals for logic trees......Page 275 9.4 Discussion......Page 276 10 Monte Carlo......Page 278 10.1.1 Random variables......Page 279 10.2.1 Uniform......Page 282 10.2.3 Normal......Page 283 10.2.5 Beta......Page 284 10.2.7 Exponential......Page 285 10.3 Choosing the right distributions......Page 286 10.3.1 Goodness of fit......Page 288 10.3.3 Other selection criteria......Page 289 10.3.4 Knowledge and inherent uncertainty......Page 293 10.4 Generating answers......Page 294 10.4.1 Random numbers......Page 295 10.4.2 Pseudorandom numbers......Page 296 10.5.1 Rank correlations......Page 297 10.5.3 Fitted models of dependent relationships......Page 299 10.6.1 Second-order Monte Carlo......Page 301 10.6.2 Models with incertitude and variability......Page 302 10.6.3 Model averaging......Page 303 10.7 Sensitivity analyses......Page 304 10.8.1 Monte Carlo for the dose equation......Page 305 10.8.2 Monte Carlo for algal blooms......Page 309 10.8.3 Population viability analysis......Page 310 10.8.4 Managing Sindh ibex......Page 314 10.8.5 Managing Baltic cod......Page 317 10.8.6 Multispecies and food web risk assessments......Page 318 10.9.1 Predicting radioactive fallout......Page 320 10.9.2 Predicting extinction risk......Page 321 10.9.3 Limitations and strengths of Monte Carlo......Page 323 10.10 p-bounds......Page 324 10.11 Discussion......Page 327 10.11.1 Monte Carlo for the risk management cycle......Page 331 11.1 Monitoring and power......Page 332 11.1.1 Null hypotheses......Page 333 11.1.2 Monitoring trends in a bird population......Page 334 11.2.1 How many samples?......Page 336 11.2.2 Comparing observations with a regulatory threshold......Page 337 11.2.3 Comparing differences among means......Page 341 11.2.4 When is something absent?......Page 345 11.3 Flawed inference and the precautionary principle......Page 348 11.3.1 The precautionary principle......Page 350 11.4 Overcoming cognitive fallacies: confidence intervals and detectable effect sizes......Page 352 11.5 Control charts and statistical process control......Page 356 11.5.1 Planning......Page 357 11.5.2 Xbar charts......Page 359 11.5.3 R and S charts......Page 360 11.5.4 Rational subgroups......Page 361 11.5.5 Control chart parameters......Page 362 11.5.7 u- and c-charts......Page 363 11.5.8 CUSUM and EWMA Charts......Page 364 11.5.9 Test/reference (control/impact) charts......Page 366 11.5.10 Pattern response and decision thresholds......Page 367 11.5.11 Dependencies......Page 368 11.5.12 Power and operating characteristic (OC) curves......Page 369 11.6 Receiver operating characteristic (ROC) curves......Page 371 11.6.1 Confusion matrices......Page 372 11.6.2 ROC curves......Page 374 11.7 Discussion......Page 382 12.1 Policy and risk......Page 384 12.1.1 Comparative risks......Page 386 12.1.2 ‘Real’ and perceived risks......Page 387 12.1.3 ‘As low as reasonably practicable’: defining acceptable risks......Page 390 12.2.1 Decision criteria......Page 391 12.2.2 Risk regulation......Page 393 12.2.3 Where model-based assessments fit in......Page 394 12.2.4 The advantages of deciding under uncertainty......Page 395 12.3.1 Stochastic dominance......Page 397 12.3.2 Benefit-cost analysis......Page 399 12.3.3 Stochastic dynamic programming......Page 404 12.4 Info-gaps......Page 405 12.4.1 A process model and measure of performance......Page 408 12.4.2 A model for uncertainty......Page 410 12.5.1 Scenario analysis......Page 413 12.5.2 Multi-criteria decision analysis......Page 415 12.5.3 Multi-criteria mapping......Page 422 12.6 Risk communication......Page 424 12.6.1 Communicating probabilities: medical cases and framing......Page 426 12.6.2 Communicating comparative risks......Page 428 12.7 Adaptive management, precaution and stakeholder involvement......Page 430 12.7.1 Involving stakeholders......Page 433 12.8 Conclusions......Page 435 Glossary......Page 437 References......Page 471 Index......Page 499 This book outlines how to conduct a complete environmental risk assessment. The first part documents the psychology and philosophy of risk perception and assessment, introducing a taxonomy of uncertainty and the importance of context. It provides a critical examination of the use and abuse of expert judgement and goes on to outline approaches to hazard identification and subjective ranking that account for uncertainty and context. The second part of the book describes technical tools that can assist risk assessments to be transparent and internally consistent. These include interval arithmetic, ecotoxicological methods, logic trees and Monte Carlo simulation. These methods have an established place in risk assessments in many disciplines and their strengths and weaknesses are explored. The last part of the book outlines some new approaches, including p-bounds and information-gap theory, and describes how quantitative and subjective assessments can be used to make transparent decisions. "This book outlines how to conduct a complete environmental risk assessment. The first part documents the psychology and philosophy of risk perception and assessment, introducing a taxonomy of uncertainty and the importance of context; it provides a critical examination of the use and abuse of expert judgement and goes on to outline approaches to hazard identification and subjective ranking that account for uncertainty and context. The second part of the book describes technical tools that can help risk assessments to be transparent and internally consistent; these include interval arithmetic, ecotoxicological methods, logic trees and Monte Carlo simulation. These methods have an established place in risk assessments in many disciplines and their strengths and weaknesses are explored A comprehensive guide to conducting environmental risk assessments for students, researchers and professionals in ecology, conservation and resource management. Coverage includes the philosophy of uncertainty and human objectivity in risky situations. Consideration is also given to how both subjective beliefs and technical analysis can be used to make informed decisions The last part of the book outlines some new approaches, including p-bounds and information-gap theory, and describes how quantitative and subjective assessments can be used to make transparent decisions."--Jacket