Assessment of energy scenarios

Due to the enormous number of energy scenarios and the huge spread in their predictions of future energy supply, it is necessary to develop and provide suitable guidelines for comparison. The comparison based on these guidelines seeks to assess the energy scenarios with regard to their epistemic and normative status as well as their informative value.

In order to assess and compare various energy scenarios, it is essential to address the question of the consistency and reliability of energy scenarios: Can consistency be measured at all? How does the informative value of a scenario relate to the reliability of the underlying model? What makes an energy model suitable for scenario generation? Do different models validate each other if their predictions agree? These questions lead to the challenging task of developing methods to systematically assess and compare various energy scenarios.

The aim of this assessment process is not to judge energy scenarios on an individual basis, but to create transparency with respect to the consequences of different assumptions and methodological approaches. Questions that arise when analysing and discussing the results of an energy system model often concern the model’s solution. Six aspects related to the analysis of the model solution can be identified:

• Uncertainty regarding the input data (i.e. demand, price projections, cost development)
• Impact of given decisions (i.e. policy instruments or measures) on the solution
• Identification of the conditions under which a certain solution (technology) is chosen
• Ambiguity of different solutions with respect to their impacts on society and economy
• Existence of implicit minimum and maximum bounds on technologies
• Explanation of the model solution and behaviour of the model.

These aspects have in common that they can be, at least partially, studied by changing the model’s input data. Different scenarios can be considered in order to analyse the impact of changes in the input data on the model solution. However, if the influence of several parameters is to be analysed, it becomes cumbersome to manually define different scenarios and analyse the scenario results individually. Also, if many scenarios are computed, it becomes very time consuming to generate the problem matrix and create the results file for each scenario, especially for large-scale models. These limitations must be considered when interpreting energy scenarios. Moreover, such insights have repercussions for the construction of energy scenarios: They call for verifying and advancing methodological approaches, such as sensitivity analysis, which are able to handle this problem automatically.