Integrated Assessment Models for the Policy-Science Interface

Transcription

1 johnthescone Integrated Assessment Models for the Policy-Science Interface An IPCC Perspective on the Challenges for AR5 Snowmass July Ottmar Edenhofer

2 Table of Content Remarks on the Policy-Science Interface How to Deal with Uncertainty Consequences for AR5

3 Table of Content Remarks on the Policy-Science Interface How to Deal with Uncertainty Consequences for AR5

4 Remarks on the Policy-Science Interface 4 Three models how to organize the interface: Technocratic approach Decisionistic approach Pragmatic / enlightened approach

5 The Technocratic Model 5 The technocratic model: Goals Means Policy decision/ implementation set by science scientific reasoning Consequences Max Weber predicts that this model will abolish democratically legitimized policy making. (Policy makers ask for practical constraints, science offers inherent necessities to legitimize policy making.) What does consensus among WGI, II, and III relate to respectively? Question not answered in technocratic model, consensus is mostly pretended. Legend: Policy Makers Science Outcome

6 The Decisionistic Model 6 The decisionistic model: Goals Means Policy decision/ implementation set set by by policy science makers scientific reasoning Consequences This devision of labour presumes: Distinction of facts and values and of targets and means always feasible. Goals and possible conflicts and synergies among them are usually re-assessed ex-post in the light of their intended and unintended consequences. This requires a continued dialogue between science and policy makers. Legend: Policy Makers Science Outcome

7 The Pragmatic-Enlightened Model 7 Data Targets Mean 1 Mean Mean n : : Qualifier 1 plausible? Qualifier m Decision 1 Decision j Outcome 1 Outcome intended Outcome k unintended Justification Backing Rebuttal that reduces the plausibility of footings, justifications or means. counterproductive? Legend: Policy Makers Science Outcome

8 IAM s and the Policy-Science Interface 8 Types of Uncertainty Parametric uncertainty Model uncertainty/ structural uncertainy Qualitative risk assessment Decision making under uncertainty/ risk management Method Sensitivity analysis, Monte Carlo Simulation Modelling comparison of 2 nd best scenarios including IPA s Expert judgment/ expert elicitation Stochastic IPAs, IAMs Implications for the Policy-Science Inteface Exploring the importance of mitigation options/ policy instruments How robust are modelling results getting a sense of robustness Side costs/ benefits Iteration between targets and means Risk management

9 IAM s and the Policy-Science Interface 9 Types of Uncertainty Parametric uncertainty Model Uncertainty/ Structural Uncertainy Method Sensitivity analysis, Monte Carlo Simulation Modelling Comparison Implications for the Policy Science Interface Exploring the importance of mitigation options/ policy instruments Qualitative Risk Assessment Decision Making under Uncertainty/ Risk Management Expert Judgment/ Expert Elicitation Stochastic IPAs, IAMs Side costs/ benefits Iteration between targest and means Risk Management

10 Potential (EJ) Sensitivity Analysis Combinations of different factors determining bio-energy potential 2100 Potential (EJ) Van Vuuren et al. (2010) Agriculture efficiency Agricultural efficiency GR0_F0 GR0_F0* GR0_F0** GR100_F0 GR50_F0 GR50_F0* GR100_F20 Loss of natural area Loss of natural area

11 IAM s and Policy-Science Interface 11 Types of Uncertainty Parametric uncertainty Model uncertainty/ structural uncertainy Qualitative Risk Assessment Decision Making under Uncertainty/ Risk Management Method Sensitivity analysis, Monte Carlo Simulation Modelling comparison of 2 nd best scenarios including IPA s Expert Judgment/ Expert Elicitation Stochastic IPAs, IAMs Implications for the Policy-Science Interface Exploring the importance of mitigation options/ policy instruments How robust are modelling results getting a sense of robustness Side costs/ benefits Iteration between targest and means Risk Management

12 Getting a Sense of Robustness 12 infeasible for most models Full Delayed $/tco2 (2005 US$) Full Delayed Full Delayed Means of Participation Luderer et al WORLD Scenarios that could not be modelled under criteria of study EMF (2009) Overview of EMF 22 International Scenarios

13 Integrated Policy Assessment Model

14 Example: The Impact of Delayed Carbon Pricing Investments in the fossil energy system are reversible Investments in the fossil energy system are irreversible Expected tax leads to accelerated extraction Carbon price established in 2035 Expected tax leads to early reduction Carbon price established in 2035 Backstop subsidies can smoothen the transition Impact of expectations is ambiguous!

15 How to Deal with Uncertainty 15 Types of Uncertainty Parametric uncertainty Model uncertainty/ structural uncertainy Qualitative risk assessment Decision Making under Uncertainty/ Risk Management Method Sensitivity analysis, Monte Carlo Simulation Modelling comparison of 2 nd best scenarios including IPA s Expert judgment/ expert elicitation Stochastic IPAs, IAMs Meaning within the Pragmatic Model Exploring the importance of mitigation options/ policy instruments How robust are modelling results getting a sense of robustness Side costs/ benefits Iteration between targets and means Risk Management

16 Consideration of unintended side-effects Iteration Between Targets and Means 16 Science Goal-setting by policy makers Science: Scope of options Data 1,5 C target 2 C target

17 IAM s and the Policy-Science Interface 17 Types of Uncertainty Parametric uncertainty Model uncertainty/ structural uncertainy Qualitative risk assessment Decision making under uncertainty/ risk management Method Sensitivity analysis, Monte Carlo Simulation Modelling comparison Expert judgment/ expert elicitation Stochastic IPAs, IAMs Implications for the Policy-Science Interface Exploring the importance of mitigation options/ policy instruments How robust are modelling results getting a sense of robustness Side costs/ benefits Iteration between targest and means Risk management

18 Three Categories of Risk Normal Risks Scope: Individual, local Intensity: Endurable, reversible Probability: Normal distribution Large Scale but Bounded Risks Scope: Transnational Intensity: Endurable, reversible/irreversible Probability: Normal distribution Systemic Risks: Scope: Transnational and transgenerational Intensity: Terminal, irreversible Probability: Fattened tail

19 Risk and the Policy-Science Interface Category Response Market (Household Failure) State/Third Sector (Market Failure) Global Collective Action (State Failure) Normal Risks Gradual adaptation within sectors Regulation of insurance markets Regulation of reinsurance markets Large Scale but Bounded Risks Weather derivatives Fiscal support to European heatwave/ hurricane Katrina Regulation of financial markets in 2009 Systemic Risks Catastrophies No adequate response known No adequate response known Provision of global public good with different technologies (e.g. Weakest Link, Best-Shot)

20 Table of Content Remarks on the Policy-Science Interface How to Deal with Uncertainty Consequences for AR5

21 The Representative Clients of AR5 in WG III International level: Negotiators, NGO s National Policies: Parliaments, governments, national agencies Regions: e.g. EU Sub-National Level: Cities

22 Consequences for the AR5 / WGIII 22 Pursuing a pragmatic-enlightened approach for the sciencepolicy interface Identifying types of risk management A few pragmatic guiding questions What are consistent ways to achieve stabilization goals? What is the relative importance of policy instruments and mitigation options? What are threshold probabilities undermining your policy options? Getting a sense of unmanageable risks What can go wrong along specific transformation pathways?

23 Implications for the Scenario Process

24 Implications for the Scenario Process Impact, Adaptation, Vulnerability 24 Establish coherence through scenario process Integrated Assessment Models Climate Models

25 Thank you for your attention!