Climate economic policies: assessing values and costs of uncertainty in energy scenarios

Project Information

MANET

Grant agreement ID: 101033173

DOI

10.3030/101033173

Project closed

EC signature date 30 April 2021

Start date 1 July 2022

End date 30 June 2024

Funded under

• EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions

Total cost

€ 171 473,28

EU contribution

€ 171 473,28

Coordinated by

POLITECNICO DI MILANO
Italy

Periodic Reporting for period 1 - MANET (Climate economic policies: assessing values and costs of uncertainty in energy scenarios)

Reporting period: 2022-07-01 to 2024-06-30

Summary of the context and overall objectives of the project

Climate change, a consistently variation of weather patterns beyond natural variability, is being affecting life on earth at different levels of severity and magnitude across the world. The global mean temperature has already grown by one degree Celsius compared to pre-industrial times, but the worst forecasts announce that a staggering increase by nearly 4°C could be reached. Global warming, the main cause of climate change, is due to the accumulation of greenhouse gases at higher rates compared to the pre-industrial times. The Intergovernmental Panel on Climate Change (IPCC) has demonstrated a causal link between global warming, higher rates of anthropogenic emissions, and a growing energy demand from the society and industries. Integrated Assessment Models (IAMs) are key tools used in the research community informing the IPCC on the nexus between climate modelling, social science, and energy systems. Despite IAMs are powerful tools, they are subjected to concerns in the way they are used and in the way their outputs are interpreted, since they must deal with complex systems and projections over hundreds of years. Uncertainty of the long-term energy futures means that broad ranges of projections are obtained from IAMs. For instance, the uptake of key decarbonisation technologies such as renewables, carbon capture and storage (CCS), electric vehicles, vary hugely among the IAMs outputs, as they are displayed in emissions scenarios ensembles.
Uncertainty in the scenarios showing possible ranges of intervention to reduce emissions, has resulted in a lack of constructive political initiatives for the promotion of the investments needed to face climate change and its consequences. If on the one side, policy makers are not cohesively joint to tackle climate change with robust political, economic, and social decisions, on the other side, the effects of climate change are more and more visible in everyday life. Long-term strategies for mitigation and adaptation should make the backbone of national and international policies.
To solve the problem of uncertainties in scenarios which have been paralysing the development of coherent policy frameworks, this fellowship project has contributed to increasing the awareness of the uncertainty implications in the use of ensembles of IAMs outputs for informing policies on mitigation and adaptation. To do so, the fellowship has focused on characterising the source of these uncertainties from linking the outputs with the inputs of IAMs. The diversity in the response from IAMs could be either coming from assumptions on data inputs (parametric uncertainty) or features of the model (structural uncertainty). Specifically, the focus was on defining what triggers the most the differential response in the IAMs projections. This has been achieved building semi-quantitative and quantitative methods.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

The fellowship ground the use of IAMs for formulating climate policies, as they should need to be robust towards the uncertain futures. Considering the emissions scenario ensemble used in the latest Assessment Report, uncertainty is displayed by the deviation in the model outputs.
The fellowship project has shown that the sources of uncertainty could be bigger than those used displayed in the IAMs outputs. This means that deviations in IAMs outputs could be bigger than those used in the Sixth Assessment Report. (AR6) In addition, the fellowship project demonstrated that when the IAMs outputs, used in the form of scenario ensembles, become detrimental to the policy formulation, processes should be in place to bridge such knowledge gap. In addition, the project has demonstrated a need for more sophisticated tools than those currently used for analysing emissions scenarios ensemble. These tools should not be meant as purely quantitative software packages but more broadly should become inter-disciplinary platforms for modellers, stakeholders, and decision-makers.
To achieve the project scope, two deliverables were released, and a workshop organised.
In the first deliverable, a mapping of the uncertainty space of the emissions scenario ensemble against alternative lines of evidence was performed. Socio-economic inputs which are critical determinants of emissions, such as population and Gross Domestic Product (GDP) , were analysed. The analysis showed that the emissions scenarios ensemble used for AR6, focuses primarily around a “Business-As-Usual”. In addition, a critical assessment of the methods available to describe emissions scenarios ensembles was performed. The literature review highlighted that the analysis of IAMs outputs is based on simplified scenario post-processing methods, which are limited in the way they can disentangle the origin of similarities and variations in the model responses.
The second deliverable was the development of online platforms which use descriptive and quantitative tools (i.e. critical analyses of alternative lines of evidence, post-processing scenario analyses, machine learning) to characterize the emissions scenario ensemble in terms of suitability for policy formulation (auditing). These tools included dashboard platforms for non-experts, who could be more guided during the explanation of the model response differences . Among the tools delivered, more complex online platforms for experts were generated to assess the cross-dependencies between input and selected outputs .
In addition to papers, a key dissemination event was a final workshop, which brought together philosophers, modellers, economists, and other stakeholders relevant in the context of climate policy, to discuss about quantitative (trajectories) and qualitative (narratives) used to characterise uncertainty.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

The fellowship project had an impact on the scientific community first advancing the approaches used to interpret emissions scenario ensemble. The project showed for first time how machine learning techniques could be applied to audit an emissions scenario ensemble assessing whether it has the quality for supporting climate policy formulation. Machine learning methods such as recommendation systems, principal component analysis, neural networks, and decision trees can be used to assess the quality of the emissions scenario ensemble and synthesise its contents. These advancements in the IAMs output analysis allow to determine the origins of similarities and divergences among IAMs outputs.
Novel tools, such as those developed in the fellowship project, to investigate the range of uncertainty and the causes of divergences among models will have an important impact on the society, with regards to the policy formulation. In fact, they will enable an increased awareness of the uncertainty. The considerations of edging the risk against multiple and conflicting alternatives, will make policies more robust. A key application would be the implementation within the IPCC synthesis process. In this context, although more and more scenarios are being made available, the quality of the final database of IAMs outputs is not being sufficiently scrutinised. There is lack of analytical tools which can help explain why the scenario outputs display more/less similarities and divergences among each other. This phenomenon is a poor transferability of the uncertainty which ultimately negatively affects the strengths of the recommendations which can be given to policy-makers.

manet-figure.png