Exploring near-optimal energy systems with stakeholders: a novel approach for participatory modelling

TL;DR

This paper tackles the challenge of involving stakeholders in energy system planning without narrowing the outcome to preselected scenarios. It introduces a near-optimal, participatory framework that communicates a continuum of feasible designs through an interactive interface, enabling stakeholders to express and learn about trade-offs among emissions, costs, vulnerability, and other priorities. Applied to the remote Arctic town of Longyearbyen, the approach reveals that participants routinely deviate from cost-optimal designs in favor of multi-criteria trade-offs, suggesting that social acceptance and resilience can be advanced by exposing stakeholders to a broader feasible space. The work demonstrates a transferable method for integrating near-optimal results with stakeholder input to improve legitimacy and understanding of energy transitions, with potential extensions to live near-optimal exploration and broader co-design of decision variables.

Abstract

Involving people in energy systems planning can increase the legitimacy and socio-political feasibility of energy transitions. Participatory research in energy modelling offers the opportunity to engage with stakeholders in a comprehensive way, but is limited by how results can be generated and presented without imposing assumptions and discrete scenarios on the participants. To this end, we present a methodology and a framework, based on near-optimal modelling results, that can incorporate stakeholders in a holistic and engaging way. We confront stakeholders with a continuum of modelling-based energy system designs via an interactive interface allowing them to choose essentially any combination of components that meet the system requirements. Together with information on the implications of different technologies, it is possible to assess how participants prioritise different aspects in energy systems planning while also facilitating learning in an engaging and stimulating way. We showcase the methodology for the remote Arctic settlement of Longyearbyen and illustrate how participants deviate consistently from the cost optimum. At the same time, they manage to balance different priorities such as emissions, costs, and system vulnerability leading to a better understanding of the complexity and intertwined nature of decisions.

Figures (7)

• Figure 1: Overview over Longyearbyen and its current energy system. On the left, the map shows the geographical location of Longyearbyen on the Arctic archipelago of Svalbard as well as its distance from the north of Norway ($\sim$950 km) making an electricity link prohibitively expensive and unrealistic. On the top right, the per capita greenhouse gas emissions for Longyearbyen (and Norway as comparison) are plotted: for the coal-powered CHP power plant (until 2022), and the current diesel-powered CHP power plant statisticsnorwaySamfunnsforholdPaSvalbard2023longyearbyenlokalstyreEnergiomstillingOvergangslosningBP222022. Norwegian emissions are computed based on 2021 data on electricity consumption (https://www.ssb.no/statbank/table/06913/) and its carbon density (https://www.nve.no/energi/energisystem/kraftproduksjon/hvor-kommer-stroemmen-fra/) as well as biomass heating emissions from Arvesen et al. arvesenCoolingAerosolsChanges2018. On the bottom right, the monthly average capacity factors for onshore wind and solar PV are plotted (for the input weather year). Note that these are subject to interannual variability and therefore just one realisation.
• Figure 2: Workflow for participatory modelling studies using near-optimal methods. On the left side we suggest an abstract workflow in which policymakers, modellers, and stakeholders collaborate to develop more socially acceptable transitions. For each of these actors, the circles with "P", "M", and "S" indicate at which stages they are involved. On the right side, we present a concrete implementation of this workflow for our case study in Longyearbyen, starting with "Energiplan" as in grotteEnergiplanLongyearbyenEnergiom2023.
• Figure 3: Overview of the main page of the user interface. The sliders and decision-variables which users may change are shown in a), whereas b) shows the resulting effects of the chosen system configuration. The interface also includes a feature (buttons) which allows choosing one of the metrics to be minimised, which participants could use as a starting point. Additionally, the cost-optimal configuration is shown with a vertical line for each slider. However, it is not the starting point when opening the interface, which was chosen at random.
• Figure 4: Combined strip- and boxplot of submitted results across the five dimensions. Stars represent the cost-optimal investment combination and one million NOK is approximately equivalent to 86,000 EUR or 90,500 USD in 2025. The min-max normalised median absolute deviations (MAD) quantify the distance between the median and the cost-optimum for the five dimensions, and are 0.06, 0.34, 0.11, 0.17 and 0.22, respectively.
• Figure 5: Obtained values from participants depending on whether a metric was prioritised; larger difference between participants who prioritised a certain metric might hint at more available options to achieve this or at higher weights of this metric. A lower value on the y-axis indicate a lower impact, which is generally desired.