PLANETS Project Home PLANETS Project - Probabilistic Long-Term Assessment of New Energy Technology Scenarios

WP 6 Running stochastic scenarios

Objectives

The key factors that determine the future deployment of new energy technologies are highly uncertain and the relationships between them can be complex. This workpackage aims at analyzing the various probabilistic components of the phenomena under investigation, and examine their relevance in terms of future energy scenarios. The traditional sensitivity analysis approach, employed in energy system analysis, is supplemented in this workpackage with state of art probabilistic and stochastic techniques. This will allow the models to deal with more refined issues such as hedging strategies against economic and technological uncertainties, limited future foresight, as well as the value of learning before resolution of uncertainty. Such a complete set of probabilistic simulation will allow the project to significantly increase the robustness of the energy and economy scenarios produced. The main tasks performed under this workpackage are listed below, in alphabetical order by model.


Description of work

Task 6.1: Climate and technology hedging

Partner: UNIMAN

Model: DEMETER

This task aims at producing simulation analysis with DEMETER on the subject of hedging, as so far no stochastic analysis has been performed with this top-down model, apart from a couple of extensive sensitivity studies. This task attempts to build on that previous work by expanding probabilistic work. Important variables that will be subjected to such hedging are the gradual physical leakage of CO2 geologically stored underground for climate mitigation reasons, the value of the climate sensitivity, and other economic, climatic or technological parameters that are important for the modelling outcomes and conclusions. Among the scenarios that will be studied in this context are what happens if part or whole of the uncertainty with regards to one or several of these variables is resolved in a given point in time in the future, whether 2030, 2040 or even later. Hedging the energy-economy against imperfect climate sensitivity information, or similarly against slowly revealing information or experience as to the level and/or time-dependence of geologically stored CO2 leakage phenomena, will be analyzed in accordance with the policies of workpackage 5.


Task 6.2: Event-tree analysis and model uncertainties comparison

Partner: ORDECSYS

Model: GEMINI-E3

This task aims at introducing in GEMINI-E3 an event-tree description of the possible evolution of uncertain parameters which would be consistent with those used in TIAM, WITCH or DEMETER, so as to obtain an ensemble of macro-economic paths. These economic development paths will be coupled with bottom-up simulations, or coordinated with the integrated climate-economic growth models. As for the uncertainties to be analyzed a number of different economic and technical parameters or variables apply, and part of the work envisaged under this task is to assess which of the variables are the most relevant ones to subject to stochastic event-tree analysis. Sizeable effort will also be spent to bring the model runs with GEMINI-E3 in tune with those of e.g. the TIMES models, and to create inter-linkages between these different kind of models, as it is judged useful for deriving policy lessons to create such synergies in particular between these two types of models.


Task 6.3: Stochastic and myopic programming

Partner: USTUTT

Model: PEM / TEAMS

This task plans to expand the current stochastic programming version of TIMES to cover more than the present stochastic input parameters, which are: demand, bound on total capacity, cumulative commodity bound on production, cumulative commodity bound on net production, atmospheric CO2 concentration, and other climate parameters. Further parameters to be considered relate to e.g. technology costs, import prices, efficiencies, annual commodity bounds, or constraints or quotas for e.g. renewables. Also, TIMES currently presumes perfect foresight. It is planned to analyze an alternative to this view of the future by developing a myopic version involving a limited knowledge of the future. In a simple myopic approach the model is optimized successively for each period taking the optimal solution of the previous period as input information. So, in a myopic version the optimization comprises several model runs implying two or more solutions for one period with different degrees of knowledge of the future. With such a myopic approach hindrance of technology penetration might be identified by a comparison with the perfect foresight runs.


Task 6.4: Global hedging strategies

Partner: KANLO

Model: TIAM

This task aims at running TIAM in stochastic programming mode in order to reflect some major uncertainties on technological, economic, and climate parameters. The TIAM experience in running global climate hedging strategies within the Energy Modelling Forum (EMF-22) will be useful in devising realistic probabilistic assumptions on these uncertainties, and on setting-up and running TIAM to observe the impact of uncertainties on energy and emission policies. The expanded list of potential stochastic parameters will allow several probabilistic investigations not currently available in TIMES.


Task 6.5: Stochastic energy supplies and technology potentials

Partner: ECN

Model: TIMES

This task aims at performing a stochastic analysis with a model from the TIMES family. A number of probabilities are identified for certain elements, or events, that will be solved, or become certain, at a certain point in time, which are called State-of-Worlds (SoW). Examples that may be analyzed in this project are the assignment of probabilities to the pre-defined amounts of oil and gas supply reserves, and similarly to the amounts, potentials or availabilities of a certain selection of technologies. Input from workpackages 3 and 4 may be instrumental in this work. In this way, the model can simulate a hedging strategy or no-regret policy, up to the point in time where the uncertainty is lifted, i.e. the outcome is known with certainty. This stochastic feature has been tested in the MARKAL family of models. This work will be extended, and/or applied to one of the models in its successor family TIMES. Part of the work will then consist of exploring, checking and further validating and/or extending the model source code and requirements for applying stochastic analysis to the TIMES model family. Among the other topics that may be considered under this task is investigating the role of economic assumptions in TIMES model runs, the importance and nature of cost reduction assumptions, and to what extent (changes in) learning phenomena and/or endogenous technical change can be correctly accounted for in TIMES. The probabilistic model obtained will be used to produce policy scenarios and confront them with the model runs of the previous workpackage.


Task 6.6: Value of learning

Partner: FEEM

Model: WITCH

This task aims at investigating the value of learning about uncertainties in the future on today’s actions. It is widely known that energy and global warming are processes characterized by a variety of uncertainties, many of which are expected to unfold at best in the medium term horizon. An academic as well as policy relevant question is thus to evaluate how we should account for these uncertainties in choosing our short term actions; for example, accounting for the uncertainty about the real leakage rates of the carbon sequestered and stored or the potential uptake capacity of biological sinks might affect our decisions in the next future on how to invest in such technologies. For this purpose, this task will develop a stochastic version of the WITCH model for evaluating the learning effect. Refined techniques such as stochastic programming will be employed to enforce these complex uncertain aspects into the integrated assessment analysis, something that has not been investigated in the past. In doing so, this task will achieve important insights on the consequences of resolving uncertainties on our today’s actions. In particular, the task will devise the consequences of resolving uncertainties on investments in the power generation sector, in the optimal carbon reductions and in the costs of complying to environmental policies.


Workpackage 6 is carried out under the co-ordination of ORDECSYS
Seventh Framework Programme

This website reflects only the views of the PLANETS Consortium and does not represent the opinion of the European Community.
The European Community is not liable for any use that may be made of the information contained herein.