Resources for modelling

As part of this analysis, policy instruments in the transport sector from the EnSu Sufficiency-Policy-Database were systematically evaluated in terms of their feasibility. They were categorised according to policy objectives (improve public transport and multi-modality, promotion of active modes, reduce air transport, reduce motorised individual transport, reduce trips: local supply, reduce trips: work) and instrument types (economic, fiscal and regulatory). The focus was on how specific (proposed) measures work, from the political impetus to their impact on the environment and society. This impact was examined using the concept of impact chains, which structurally maps cause-and-effect relationships. Particular emphasis was placed on supporting and inhibiting factors.

The evaluation was carried out using a complex process involving literature research and expert discussions and is based on several revision loops.

Results & Recommendations

The following figure illustrates the results for the supporting factors, obstacles and risks for the political objectives analysed. A detailed description of the method and its application, as well as the results, can be found in the published technical article. 

Figure 1: Weighted sum of supporting factors, obstacles and risks for all policy instruments analysed, by policy objective (colour of dots); Source: Thema, Zell-Ziegler & Dünzen (2025).

Note: The numbers indicate the IDs in the EnSu Sufficiency-Policy-Database. * = few factors, ** = medium number of factors, *** = many factors.

The analysis of 83 chains of effects of sufficiency-oriented policy measures in the transport sector reveals three key findings for modelling practice: (1) Measures with multiple supporting factors often present many obstacles and risks at the same time. This applies in particular to broader policy measures with various levels of impact. (2) Policy instruments with the objectives of ‘promoting active modes’ and ‘reducing motorised individual transport’ are comparatively low-risk, as they are generally less resource- and cost-intensive. (3) Feasibility varies significantly depending on the type of instrument: regulatory instruments show surprisingly low risks, a similar number of obstacles to economic instruments and a comparable number of supporting factors to fiscal measures.

For modellers, the chain of effects method, which is specially adapted to sufficiency, offers a structured approach to assessing the feasibility and impact of policy measures. In addition to direct effects, it also takes into account systematic prerequisites and contextual factors, making it a suitable complement to quantitative models. The development of the method is described in detail in Zell-Ziegler & Thema (2022), which also provides initial examples of its application.

 A fillable template is available for your own use. The policy instruments that have already been evaluated are also available in Excel format and can be adapted (link). 

The energy transition scenarios for Germany developed by the EnSu research group follow a three-step method. The starting point is the interdisciplinary method of socio-technical scenario development, which is used to formulate consistent and plausible narrative contexts for possible societal futures. In the second step, these qualitative context scenarios are then transferred to the 2050 Pathways Explorer (PWE) simulation model in order to develop GHG neutrality scenarios. Finally, in the third step, the quantitative results derived from this serve as input for the PyPSA-Eur energy industry optimisation model in order to analyse the effects on the energy system.

The following figure illustrates this chain of methods:

A related technical article is still under review (manuscript available from frauke.wiese@uni-flensburg.de).

Development of narrative scenarios

The development of the context scenarios follows the methodological approach of socio-technical scenarios according to Weimer-Jehle et al. (2016) and was implemented in four consecutive steps in order to create consistent visions of the future for climate-neutral development in Germany.

In the first step, key influencing factors (known as descriptors) were identified that have a direct or indirect impact on the energy system. In the second step, alternative characteristics were formulated for each descriptor that reflect key uncertainties in future developments. The third step involved a systematic evaluation of the interdependencies between these descriptor characteristics in the form of a cross-impact matrix. For each possible combination, the strength of the influence was assessed on a scale from -3 (strongly inhibiting) to +3 (strongly promoting) and jointly validated in a workshop.

In the fourth step, consistent combinations of descriptor characteristics were calculated using the ScenarioWizard software (version 4.3). The CIB (Cross Impact Balance) method analyses whether the combinations, i.e. the promoting and inhibiting effects between the descriptors, complement each other in a consistent overall picture. Based on the average matrix of all experts, redundant scenarios were first sorted out and supplemented by further consistent variants from individual analyses. The result is six narrative context scenarios that provide qualitative descriptions of social, political and economic conditions for possible developments on the path to climate neutrality. Four of the scenarios describe different sufficiency-oriented futures, while the other two describe more green growth-oriented futures.

 A detailed description of the method and the execution of the context scenarios in an accompanying technical article are still under review (manuscript available from frauke.wiese@uni-flensburg.de). 

Modelling of GHG neutrality scenarios

For the quantitative modelling of GHG neutrality scenarios, the previously developed context scenarios were converted into concrete energy scenarios. This was done using the web-based simulation tool 2050 Pathways Explorer (PWE Release v31.0, 11/04/2023) from CLIMACT, which offers a wide range of settings for mapping cross-sector transformation paths, especially on the demand side, in detail down to the service level.

The translation of qualitative narratives into quantitative scenarios is based on an adapted application of the story and simulation method according to Alcamo (2008), in which narrative elements were systematically converted into modelable parameters. The methodological implementation, including the structured assignment of descriptors to model parameters and validation by experts, is documented in detail in the published master’s thesis. 

In addition, the paper provides a detailed analysis of GHG emission pathways and a critical reflection on the socio-ecological implications of the scenarios. It shows that a profound reduction in energy demand is closely linked to systemic issues and requires political framework conditions that promote sufficiency-oriented lifestyles and infrastructures across the board and in detail for each sector.

Results & Recommendations

The modelling shows that ambitious sufficiency pathways are fundamentally compatible with the goal of greenhouse gas neutrality by 2050, provided that the underlying assumptions about social change, technological progress and natural sink potentials are realistic and consistent.

At the same time, it is clear that both the modelling tool chosen and the way in which qualitative assumptions are translated into quantitative parameters have a significant influence on the results. A coordinated, interdisciplinary modelling process can contribute significantly to consistency and traceability in this regard.

Overall, parameters directly related to the energy system, such as domestic potentials of land for renewable energy production, can be derived relatively easily from qualitative narratives, while social developments with a likely strong but indirect relationship to the energy system, such as wealth distribution and property relationships, or resource availability, externalization and international distribution, are much more difficult to integrate into existing model structures.

The translation work carried out shows where qualitative assumptions are already easily quantifiable and where existing models reach their limits. Therefore, the further development of energy models or additional model modules should be specifically geared towards better reflecting social dynamics, social practices and structural changes. Qualitative scenarios provide valuable insights, even if they are not (yet) fully quantifiable. The aim should not be to reduce this diversity, but rather to further develop the quantification side so that social developments with a strong but indirect influence on the energy system can also be considered in an integrated manner.

Energy system modelling

In the third step of the methodology chain, detailed energy system modelling was carried out based on cross-sector optimisation with high temporal and spatial resolution. The aim was to analyse the energy and infrastructure-related impacts of the previously simulated scenarios and to derive systemic requirements and conflicting objectives from them.

 The open-source model PyPSA-Eur was used for this purpose. PyPSA-Eur is a sector-coupled energy system model that integrates the electricity, heat, mobility and industry sectors and maps both the electricity and gas/hydrogen networks. For our application, the focus was on Germany with high temporal and spatial resolution, taking into account developments in neighbouring countries. The level of demand was taken from the results of the 2050 Pathways Explorer (PWE) and adapted to the data structure of PyPSA-Eur. Where direct data transfers were not possible, scaling factors were used to ensure consistency. An accompanying scientific article with detailed results and methodology is still undergoing peer review and will be linked upon publication.

Results & Recommendations

Of the descriptors of the socio-technical scenarios developed in EnSu, demand for energy services, domestic land potential for renewable energy production, technical development and speed of technology implementation could be directly translated into the modelling language of the cross-sectoral tool. Other descriptors of different futures, such as individualisation, prioritisation of climate protection and planetary boundaries, distribution and availability of resources, growth independence and wealth distribution, and housing and supply structures, were incorporated into the upstream step of sectoral demand simulation.

The scenario comparison shows that reduced final energy consumption can achieve decarbonisation with lower system costs and lower renewable energy capacities, but the reduction is not proportional, partly due to fundamental infrastructure requirements.

A comparison of sufficiency-oriented scenarios with other available decarbonisation scenarios with and without sufficiency shows that a reduction in final energy demand through efficiency and sufficiency significantly reduces pressure in the areas of biomass use, wind and solar expansion, and imports of energy sources from within and outside the EU. but that demand in some areas can be similarly high even in sufficiency scenarios, depending on the focus of the scenario (e.g. very little biomass increases the need to import energy sources even in sufficiency scenarios).

A comparison of the modelled EnSu scenarios with other sufficiency-oriented scenarios shows that sufficiency scenarios not only increase the range of options for decarbonisation, but that there is also a wide range of energy system developments within sufficiency-oriented scenarios, depending on how pronounced the reduction in sectoral energy service indicators is and how the combination with the other parameters of imports and biomass is affected.

Since the EnSu Group was founded in 2020, we have compiled policy instruments from many different sources that promote sufficiency. This has resulted in the EnSu Sufficiency-Policy-Database, which is continuously being expanded. Annual interim results can be found at Zenodo. A research article on the development and structure of the database can be found here (Best et al. 2022).

The database currently contains more than 350 entries, demonstrating that sufficiency is not an individual decision, but can be promoted politically in a variety of ways and with very different types of instruments. Framework conditions such as infrastructure or (financial or other) incentives play a key role in encouraging people to behave sufficiently. We have compiled the policy proposals found in the literature on this topic in the database – the first comprehensive collection on this subject.

A small graphical evaluation can be found under the database on our website. 

The entries are clustered by sector – target/policy strategy – measure/activity. Each entry also includes the sufficiency type and instrument type of the policy instrument, and many entries also include a rough estimate of the time between implementation and impact.

Results & Recommendations

The EnSu Sufficiency-Policy-Database can be found directly on our website, including lots of additional information such as savings potential and implementation examples.

One of the two demand sectors that were examined and modelled in detail within the EnSu framework is the mobility sector. In close cooperation with the Reiner Lemoine Institute (RLI), the macroscopic transport model quetzal_germany was used for this purpose. The central contribution and approach consists in the development of ‘policy scenarios’. In contrast to classic, assumption-based scenarios, in which political framework conditions are interpreted retrospectively, measures were explicitly mapped via chains of effects in the transport model and their effects were modelled directly (see technical article).

An empirical analysis of the explanatory factors for car ownership in Germany also formed the basis for the integration of endogenously modelled car ownership into the transport model used.

In a second strand of research, led by RLI, the effects of a changed transport system on the entire energy system were investigated. Existing Avoid/Shift/Improve scenarios were used for this purpose, and the quetzal_germany model was linked to the AnyMod energy system model. Car fleets were also explicitly modelled for this purpose. The technical article shows the effects on the entire system.

Results & Recommendations

Detailed modelling of individual policy measures, including their stringency, allows for the simulation of individual measures and their combination into policy scenarios. In the example scenario modelled in the technical article for 2035, the sum of the measures reduced total passenger transport performance by approximately 30%, particularly motorised private transport, while public transport increased, thus bringing about a mode shift.

While emissions from the transport sector had already fallen by approximately 30% by the scenario year 2035 due to electric drives, the reduced transport performance and changed mode choice contributed to additional savings. The model also allows spatially resolved effects to be represented.

The approach shows that it is possible to explicitly model policy measures and develop corresponding scenarios that show possible development paths depending on political framework conditions. The approach is a significant improvement for policy makers: existing scenarios usually assume certain developments in parameters that are only realistic under certain policy conditions – while the conditions themselves are often not specified. Such a policy-based modelling approach is helpful for making informed policy decisions.

Figure 1: Results of policy modelling for transport performance per passenger kilometre (pkm).

Figure 2: Results of policy modelling for transport performance for passenger kilometres (pkm) per capita by car (left) and rail (right)

The second analysis deals with existing Avoid-/Shift-/Improve scenarios and their impact on the energy system. The coupled modelling of traffic simulation and energy system optimisation shows that the avoidance scenario in particular – as well as its combination with measures for shifting and increasing efficiency (Avoid + Shift + Improve) – requires the least expansion of renewable generation capacities and storage technologies. As a result, the total system costs are also lowest in these scenarios.

Figure 3: Scenario costs by component. Private vehicle fleets account for the largest share of total costs, especially in the Improve scenarios. Shift and Avoid scenarios incur additional infrastructure costs for public transport and communal living (e.g. apartment buildings, supply structures), which means that the total costs without private vehicles can exceed the reference scenario. The costs of the energy system decrease more along the transport supply axis (vertical) than along the transport demand axis (horizontal), particularly due to lower demand for power-to-liquid plants and corresponding renewable electricity generation.

Another demand sector modelled within the EnSu framework is the building sector. Historically, living space has continued to increase, mainly due to the rise in average per capita space for a variety of reasons. Since living space must be heated in winter and partially cooled in summer, and new buildings require considerable resources, scenarios for a sustainable transformation of the building stock would have to take factors influencing living space into account endogenously. However, existing models for the building stock typically treat the development of living space as an exogenous input (absolute or per capita), and modelling the occupancy of the building stock – i.e. how the population is distributed among existing dwellings – is an open research gap.

We are developing a macroscopic occupancy model for residential buildings using data from the Socio-Economic Panel (G-SOEP). Our model maps the distribution of the German population across the housing stock and simulates moving behaviour over time, including the effects of policy interventions, differentiating between socio-demographic and socio-economic household classes (age, number of persons, household type, income) and building characteristics (type, number of rooms, urbanisation, renovation status). We analyse various scenarios and thus gain insights into the structures and patterns of change in under- and over-occupancy.

The model is currently still in the development phase, but is continuously available as open source on GitLab. An initial model description is included in our ECEEE conference paper, and a technical article is currently under review (manuscript available from johannes.thema@wupperinst.org). 

Results & Recommendations

The INHABIT model allows the projection of high-resolution matrices that map the occupancy of the building stock. These can be used, for example, to evaluate the average number of rooms by household or building characteristics.

Figure 1: Average number of rooms in Multi-family houses (MFH) and single-family houses (SFH) in an example default scenario and a reduction scenario (Red_UO).

The simulation approach can be used to analyse the extent to which dwellings are under- or over-occupied, based on the relationship between household size and number of rooms. In addition, projections for future housing use can be made and the effects of different measures on occupancy patterns can be examined. The graph shows, for example, that under-occupancy in the default scenario is particularly pronounced among older households and continues to increase over time. In an alternative scenario, however, occupancy patterns could converge. In addition, the approach also enables a differentiated analysis of the distributions behind the average values.

Figure 2: Modelled occupancy for households over/under 55. Development of the average over time (left) and distribution (right). Indicator: number of rooms – number of persons per household.