IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning

IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning Naomi Tana,b,∗,Ioannis Vrochidisc,Hannah Luscombed,Emma Richardsona,b, Fernando Plazas-Niñoa,Kane Alexandera,b,Leigh Martindalea,b,Neve Fieldsa, Mark Howellsa,b,Vivien Fosterband John Harrisona aCentre for Sustainable Transitions: Energy, Environment and Resilience, Loughborough University, Loughborough, LE11 3TU, United Kingdom bCentre for Environmental Policy, Imperial College London, London, SW7 1NE, United Kingdom cTUM School of Engineering and Design, Technical University of Munich, Garching b. München, 85748, Germany dSmith School of Enterprise and the Environment, University of Oxford, Oxford, OX1 3QY, United Kingdom ARTICLE INFO Keywords: Energy Systems Energy Transition Sustainable Finance Soft-linking ABSTRACT As global environmental challenges increase, the need for integrated energy modelling to facilitate data-driven decision-making in energy policy and finance is critical. However, most existing integrated frameworks are limited in their applicability, granularity, and accessibility, risking the exclusion of developing countries from the global energy transition. Social dimensions are also often insufficiently addressed, and financial planning is not integrated, leaving gaps between technical analysis, social considerations, and actionable investment pathways. To address this, the article presents the Integrated Model for Policy, Actions and Collaborative Climate Transitions (IMPACCT), a new comprehensive framework that soft-links seven significant open-source tools—MAED; OnSSET; OSeMOSYS, including CLEWs and SIBs; FlexTool; PathCalc; MINFin; and FINPLAN—for the first time. IMPACCT estimates energy demands from the electrified and unelectrified populations and calibrates the least-cost capacity mix to meet demands while taking into account land availability, water use, carbon emissions, and social factors. The capacity mix is further refined to ensure power system flexibility, and the technical outputs are visualised in an engaging interface. Financial strategies at national and utility levels complete the framework, supporting the practical realisation of the technical plans. By outlining a new process with open-source, user-friendly interfaces, this paper increases accessibility and ease of use, supports capacity building in developing countries, and facilitates collaboration across institutions and disciplines. Itdelivers a significant leap in energy modelling, social inclusion, and financial planning, advancing a more integrated approach to sustainable development. Overall, IMPACCT enables more transparent and collaborative decision-making, accelerating financial mobilisation for a just energy transition. 1. Introduction Open-source data and energy modelling tools for evidence-based policymaking allow transparency, adaptability, and collaboration—all of which can accelerate the release of concessionary finance for sustainable development [1–8]. Multiple energy transition studies have incorporated these aspects; however, they typically utilise a single energy model rather than multiple [7,9]. Single-model analyses usually cannot perform such extensive assessments to include considerations regarding energy demand and supply, power flexibility, land availability, water use, social factors, carbon emissions, financial strategies, and financial returns. Thus, they fail to provide a comprehensive representation of the factors that drive energy system development (i.e., macroeconomics affecting demand, investment profitability, and others) or the critical considerations for planning (i.e., job creation, resource use, and other implications). Integrated Assessment Models (IAMs) have been developed to address these challenges by combining insights from multiple sectors to inform energy planning, climate policy, and sustainable development [10,11]. These models ∗Corresponding author [email protected] (N. Tan) ORCID(s): 0000-0001-7957-8451 (N. Tan); 0009-0003-5058-5289 (H. Luscombe); 0009-0001-2621-3034 (E. Richardson); 0000-0003-3392-5707 (F. Plazas-Niño); 0009-0005-5350-7037 (K. Alexander); 0000-0001-7990-4923 (L. Martindale); 0009-0004-3545-7550 (N. Fields); 0000-0001-6419-4957 (M. Howells); 0000-0002-8505-8994 (V. Foster); 0000-0002-6434-5142 (J. Harrison) Tan et al.: Preprint submitted to Elsevier Page 1 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning have gained prominence through their inclusion in assessments by the Intergovernmental Panel on Climate Change (IPCC) and the International Energy Agency’s World Energy Outlook [12]. To date, several well-established IAMs exist; however, as we show in Section 2, all of these models remain inaccessible to a broader audience and often fail to adequately incorporate the social and financial dimensions essential for a just energy transition [13–15]. This highlights a significant gap in the availability of open-source, user-friendly integrated energy modelling approaches that go beyond purely technological analyses to integrate social and financial considerations. Additionally, rather than developing yet another standalone modelling tool to address these gaps, we argue in this paper that a more effective approach is to create a framework that connects a suite of specialized, well-established models—each excelling in its respective domain—while providing methodologies for their integration and execution [14,16–19]. This article bridges the gap by conceptualising the Integrated Model for Policy, Actions, and Collaborative Climate Transitions (IMPACCT)—a framework soft-linking seven significant and widely-used open-source energy modelling tools for the first time, with potential to expand further. By integrating a suite of existing open-source and user-friendly tools, IMPACCT lowers the barriers to entry for those without coding skills or advanced computational resources, enabling a broader range of users to engage in energy and financial modelling. Unlike all existing integrated assessment models (IAMs), IMPACCT explicitly incorporates social and financial dimensions, enabling users to move from technical analysis and social considerations to actionable investment strategies. Its modular approach allows decisionmakers to apply models they are already familiar with while fostering interdisciplinary collaboration. Along with open data, capacity-building, and stakeholder and community engagement, IMPACCT aims to deliver insights and be utilised by policymakers, analysts, and researchers within energy modelling, policy, and finance, for a just, equitable, and orderly energy transition. IMPACCT also provides flexibility in the scope of analysis. Users can choose between a variety of analyses: (1) translating techno-economic plans into financial actions at the national and utility levels through a unidirectional pipeline; (2) refining case studies through feedback loops between two tools; or (3) combining both approaches. To aid readability, this paper demonstrates the framework primarily as a unidirectional pipeline, guiding users from energy demand forecasting to identifying financially viable utility-level projects that meet the estimated demand. Nevertheless, feedback loops are also highlighted throughout to illustrate how users can refine analyses and assumptions. The modelling tools in the framework are chosen for their open-source code, user-interface accessibility, significance and wide use, and their differing aspects of the energy transition. The tools, their descriptions, inputs, and outputs are provided in Table 1. Their linkages in IMPACCT are illustrated in Figure 1. The paper is organised as follows: Section 2reviews existing IAMs and identifies critical gaps, while Section 3 details the seven tools in IMPACCT and implemented linkages between them. Section 4covers the conceptualisation and methodology of IMPACCT, followed by Section 5, which discusses the benefits, limitations, initial uptake, and future directions of IMPACCT. Finally, Section 6presents the conclusions. 2. A Review of Integrated Assessment Models and Frameworks Integrated Assessment Models (IAMs) are multidisciplinary tools that combine information from multiple sectors to support energy planning, climate policy, and sustainable development [10,11]. IAMs emerged in response to the limitations of traditional sectoral models, which struggled to capture cross-sectoral linkages [38,39]. By integrating representations of energy systems, economic development, land use, climate change, and other sectors, IAMs enable users to explore long-term interdependencies and trade-offs across a wide range of socio-economic, technological, and environmental drivers. IAMs comprise a family of models and frameworks, each with distinct methodologies tailored to answer specific questions. IAMs operate across multiple spatial scales—global, regional, and national—and typically assess pathways over long-term horizons, often extending to 2050, 2100, or beyond. They have been widely applied in scientific literature [40], international policy assessments [13], and government legislation [41], supporting evidence-based decision-making processes [14]. Over the past few decades, IAMs have played a crucial role in shaping global climate and energy policy, particularly through their inclusion in assessments by the Intergovernmental Panel on Climate Tan et al.: Preprint submitted to Elsevier Page 2 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning Table 1 List of the seven open-source tools in IMPACCT, their developers, brief description, key inputs and modelled outputs. Name Developer Description Key inputs Key outputs MAED Model for Analysis of Energy Demand International Atomic Energy Agency (IAEA) Simulates future energy demand based on a set of assumptions on mediumto longterm demographic, socioeconomic, and technological developments [20,21] Population growth, economic development, sectoral energy intensities, technological efficiency improvements, and lifestyle and consumption patterns Detailed estimates of useful energy demand and hourly electricity demand, disaggregated by end-use category for the industry, transport, services, and housing sectors OnSSET Open Source Spatial Electrification Tool KTH Royal Institute of Technology (KTH) Optimises the mediumto longterm expansion of electricity access, incorporating geospatial factors such as energy demand, infrastructure constraints and regional distribution [22,23] Population distribution, existing electrification rates, infrastructure proximity, renewable resource availability, and technology costs Estimates of current unelectrified residential demand supplied by off-grid technologies, their associated capacity factors, and network distribution costs OSeMOSYS Open Source Energy Modelling System KTH Royal Institute of Technology (KTH) Optimises the least-cost capacity expansion plan to meet a predefined demand in the mediumto long-term [24,25]. This can be expanded to include the Climate, Land, Energy, and Water Systems (CLEWs) [26,27] and Social Impacts and Benefits (SIBs) approaches [28,29] Energy demand estimates, technology costs and efficiencies, system operation parameters, fuel prices, and resource availability Least-cost capacity expansion plan broken down into capital, fixed, and variable costs, and its associated emissions, land and water use, and social implications FlexTool International Renewable Energy Agency (IRENA) Optimises least-cost power system flexibility assessments for one year based on national capacity investment plans and forecasts [30,31] Electricity demand profiles, generation capacity and technology mix, transmission infrastructure, flexibility constraints, and storage options Identification and quantification of flexibility indicators (curtailment, loss of load, reserve adequacy), as well as least-cost dispatch solutions PathCalc Pathways Calculator Loughborough University & Imperial College London Optimises thousands of mediumto long-term scenarios from a ’base’ OSeMOSYS model, leveraging ambition levels and levers to show the impact of various choices on key metrics such as investments and emissions [32,33] Quantified targets and policies within a nation, as well as a ‘base’ OSeMOSYS model Interactive visualization dashboard with thousands of potential least-cost capacity expansion plans and their associated costs and emissions MINFin Model for Informed National Financing University of Oxford & Imperial College London Simulates strategies to bridge financing gaps in energy sector transition plans by comparing financing requirements with projected cash flows and quantifying the shortfalls [34,35] Long-term capacity expansion investment plan, including fixed and fuel costs, the associated carbon emissions, and historical financing data Annual financing repayment needed to service debt and equity in support of the long-term energy plan FINPLAN Model for Financial Analysis of Electric Sector Expansion Plans International Atomic Energy Agency (IAEA) Simulates the financial performance of power plant projects over their lifetime by comparing the cost components with available financing sources [36,37] Market conditions, investment, fixed, and fuel costs, financing structure, electricity tariffs, and project lifetime Overall project profitability, including electricity generation costs, cash flow estimates, financial performance indicators, tariff requirements, and revenue forecasts Change (IPCC) and the International Energy Agency’s World Energy Outlook [12]. Tan et al.: Preprint submitted to Elsevier Page 3 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning Figure 1: Infographic displaying how the seven open-source tools: MAED, OnSSET, OSeMOSYS, FlexTool, PathCalc, MINFin, and FINPLAN, can be utilised together for IMPACCT and support wider understanding of energy systems modelling. Solid arrows depict a one-way input-output link in the direction of the arrow, while dashed lines represent feedback loops within the tools connected by the bidirectional arrow. This paper highlights the unidirectional pipeline from techno-economic modelling to financial planning in order to release concessionary finance for sustainable development; however, the flexibility of IMPACCT also allows users to utilise the feedback loops to refine certain models or sectors based on specific interests or needs. Tan et al.: Preprint submitted to Elsevier Page 4 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning The most widely used IAMs include AIM-Hub (Asia-Pacific Integrated Model), GCAM (Global Change Assessment Model), IMAGE (Integrated Model to Assess the Global Environment), MESSAGE-GLOBIOM (Model for Energy Supply Strategy Alternatives and their General Environmental Impact linked with the Global Biosphere Management Model), REMIND-MAgPIE (Regionalized Model of Investments and Development combined with the Model of Agricultural Porduction and its Impact on the Environment), and WITCH (World Induced Technical Change Hybrid model) [13,40–42]. Nonetheless, the IAM community continues to expand in response to growing demand and evolving policy needs, both through the development of new models and the refinement of existing frameworks. Table 2summarises IAMs recognised by the Integrated Assessment Modelling Consortium (IAMC) [43]. To ensure transparency, the list is limited to IAMs for which complete documentation is publicly available. The models are compared based on equilibrium type, modelling approach, sectoral coverage, open-source status, reliance on opensource platforms and/or solvers, and user-accessibility. A model is deemed user accessible if its use does not require programming skills, proprietary software licenses, or high-performance computing resources, all of which are common barriers to access for researchers in developing countries. The comparison shows that IAMs vary widely in their structure, scope, and methodological approaches. One key distinction in structure is between general and partial equilibrium modelling. General equilibrium models, such as MESSAGE-GLOBIOM, capture economy-wide interactions and market feedbacks, making them suitable for analysing the broader economic impacts of climate policies across sectors, households, and trade. Partial equilibrium models, such as IMAGE and TIAM-UCL, focus on detailed representations of specific sectors like energy or land use without modelling the entire economy’s feedback effects. Some IAMS, such as REMIND-MAgPIE, combine both approaches. REMIND employs a general equilibrium approach to model global energy and economic interactions, while MAgPIE uses a partial equilibrium framework to provide detailed insights into agricultural production, land use, and food markets. Analytical approaches also vary. Some IAMs, such as WITCH, cost-optimise pathways subject to climate and energy constraints. Others, such as AIM-Hub and GCAM, simulate plausible system evolution based on behavioural assumptions, historical trends, and policy inputs. A few IAMs combine both optimisation and simulation approaches. For instance, COFFEE-TEA allows users to optimise a least-cost pathway through COFFEE and explore its income effects and macroeconomic feedback through TEA. Furthermore, IAMs vary in sectoral coverage. Some focus on energy systems and emissions, enabling users to understand how energy transitions influence greenhouse gas emissions. Others have expanded to include additional resource systems, such as land use and water availability, and their interdependencies across the energy-land-water nexus. These integrated frameworks enable the assessment of cross-sectoral trade-offs, such as the impacts of water constraints on energy production, or bioenergy production on food security and forest conservation. Tan et al.: Preprint submitted to Elsevier Page 5 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning Table 2: List of fully documented IAMs recognised by the IAMC, their developers, type of equilibrium, modelling approach, open-sourceness, and user accessibility. Note that this list is limited to those with complete, publicly available documentation to ensure transparency. All information is from the IAMC [43]. IMPACCT is listed at the bottom of the table for comparison. IAM Developer Type of equilibrium Modelling approach Sectors Opensource code Open-source platform and/or solver User accessible AIM-Hub National Institute for Environmental Studies & Kyoto University (Japan) General Simulation Economy, energy, land, emissions No No No BLUES Cenergia (Brazil) General Optimisation Energy, land, emissions No No No C3IAM Beijing Institute of Technology (China) General Optimisation Energy, land, emissions, climate No No No COFFEE-TEA Cenergia (Brazil) General & partial Optimisation & simulation Energy, land, emissions No No No DNE21+ Research Institute of Innovative Technology for the Earth (Japan) Partial Optimisation Energy, land, emissions No No No GCAM Pacific Northwest National Laboratory (USA) Partial Simulation Energy, land, water, emissions, climate Yes Yes No GEM-E3 Institute of Communication and Computer Systems (Greece) General Optimisation Energy, emissions No No No GRACE Centre for International Climate Research (Norway) General Simulation Energy, emissions No No No IFs Pardee Centre (USA) General Simulation Energy, land, water, emissions, climate, social No Yes Yes IMACLIM Centre International de Recherche sur l’Environnement et le Développement & Société de Mathématiques Appliquées et Sciences Humaines (France) General Simulation Energy, land, climate No Yes No IMAGE PBL Netherlands Environmental Assessment Agency (Netherlands) Partial Simulation Energy, land, water, emissions, climate No No No MESSAGEGLOBIOM International Institute for Applied Systems Analysis (Austria) General Optimisation Energy, land, water, emissions, climate No No No Continued on next page Tan et al.: Preprint submitted to Elsevier Page 6 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning IAM Developer Type of equilibrium Modelling approach Sectors Opensource code Open-source platform and/or solver User accessible POLES European Commission (Belgium) Partial Simulation Energy, land, emissions No No No PROMETHEUS E3Modelling (Greece) Partial Simulation Energy, emissions, climate No No No REMINDMAgPIE Potsdam Institut für Klimafolgenforschung (Germany) General & partial Optimisation & simulation Energy, land, emissions, climate Yes No No TIAM-UCL University College London (UK) Partial Optimisation Energy, emissions, climate No No No WITCH European Institute on Economics and the Environment (Italy) General Optimisation Energy, land, water, emissions, climate No No No WITNESS Linux Foundation & Open-Source for Climate (n/a) General Optimzation & simulation Energy, materials, population, environment Yes Yes No IMPACCT Loughborough University, Imperial College London, Technical University of Munich & University of Oxford Partial Optimisation & simulation Energy, land, water, emissions, climate, social, finance Yes Yes Yes Tan et al.: Preprint submitted to Elsevier Page 7 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning 2.1. Critical Gaps in Integrated Assessment Models While many IAMs exist (Table 2), none offer an open-source, user-friendly framework that explicitly integrates the social and financial dimensions critical for energy transition analysis. This limits the Global South’s capacity to conduct cross-sectoral studies, where accessible tools are essential for researchers without advanced computing skills. Most IAMs already fail to capture key differences in developing regions, resulting in misrepresented energy transition pathways [42]. It also hinders stakeholder engagement, despite governments increasingly prioritising social outcomes in energy planning [8,14,19]. Financial feasibility is another overlooked barrier, with many IAMs failing to assess whether technically optimal plans are economically viable [44]. The following subsections examine these gaps in detail, highlighting why accessible, socially responsive, and financially informed modelling approaches are urgently needed for a just global energy transition. 2.1.1. Lack of Transparency Table 2shows that most IAMs remain closed with expensive proprietary software. Of the 18 IAMs reviewed, only two are fully open-source, with freely available code and compatibility with open solvers like Mathprog or Python. The remaining models either restrict access to their code or rely on expensive proprietary platforms and solvers such as GAMS and CPLEX, creating significant financial and technical barriers for broader participation [5,24]. Despite advances in modelling techniques, transparency in model development and application has received comparatively little attention [3,14,18,45,46]. Transparency is critical for ensuring models are scientifically credible and useful to policymakers [2,4–6,8,14]. Without open access, hidden assumptions, simplifications, or coding errors are impossible for external users to detect, undermining trust in the results [45]. Open-source models, by contrast, allow others to reproduce findings, verify results, and subject the code to rigorous scrutiny, helping to minimise human error and improve reliability [3,5,24,46]. Beyond transparency, open models foster collaboration and avoid duplication of effort, reducing the risk of fragmented, low-quality modelling efforts [5]. Open-source communities enable diverse contributors, such as academics, government analysts, and others, to improve models collaboratively and strengthen the link between modelling and policy development [3,7]. Ultimately, open-source models lower entry barriers, widen participation, and encourage innovation [13]. Publicly available code is also more likely to be reused, adapted, and supported by an active user community [3,4]. Although some developers are cautious about open-sourcing their models due to concerns over intellectual property misuse or user support burdens [3,4,18,45], others, like DeCarolis et al. [45], argue that "the benefits of model transparency outweigh these legitimate concerns". Cao et al. [46] further reiterates this, noting "transparency is even more important than technical skills". 2.1.2. Lack of User-friendliness Table 2shows that only one IAM is user-friendly. While some IAMs have been made open-source, most remain inaccessible to non-experts due to steep learning curves and the absence of intuitive interfaces. This limits their adoption in developing countries, where ease of use and local capacity building are critical for sustainable energy transitions. Many developing countries face limited institutional capacity, scarce technical expertise, and constrained computing resources [47–50], making it difficult to engage with complex and highly technical models. User-friendly interfaces and participatory modelling approaches can help overcome these barriers. By simplifying the modelling process, they enable a wider range of users, including policymakers and non-technical stakeholders, to engage meaningfully in energy planning. This engagement builds trust, improves understanding, and increases acceptance of energy decisions, all of which are critical for successful policy implementation and investor confidence [8,14,51]. However, without accessible tools, many stakeholders are excluded from the dialogue, and some policymakers remain sceptical of modelling’s value for decision-making [51]. Capacity building is key to changing this dynamic. It requires long-term investment in local skills development rather than continued reliance on external experts [18,52–54]. Accessible, open-source models with low technical barriers allow local users to learn by doing, gradually developing a skilled pool of experts who can maintain and apply these tools in their own context. Without such efforts, developing countries risk being left out of the global energy Tan et al.: Preprint submitted to Elsevier Page 8 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning transition, leaving billions of people out of the energy transition dialogue [54]. Therefore, improving usability and strengthening local capacities are critical steps toward more inclusive, sustainable energy transitions. 2.1.3. Lack of Social Drivers and Implications Table 2shows only one IAM that includes both social drivers and implications. In recent years, IAMs have been criticised for being overly techno-centric, focusing primarily on cost-optimal pathways while neglecting the social dimensions of energy transitions [13,16,18,19,55]. This limits their usefulness to policymakers seeking to design energy transitions that are just and equitable [56]. Research has shown that incorporating social factors improves both the performance of models [57] and the relevance of their results [19]. Socio-economic aspects are generally only included in supply-demand models. Most models rely solely on GDP as the primary socio-economicdriverfor estimating energydemand,butthis approachcanproduceinaccurateassessments by leaving out the complexities of human behaviour and economic dynamics [17]. Important social and behavioural drivers, such as household decisions to install efficient lighting or shifts in transport modes, are often overlooked, leaving a gap in understanding how user choices impact energy demand [42,56]. There is also a growing consensus that models should incorporate a broader range of economic sectors and activities to reflect demand more accurately [13]. For example, the industrial sector, which is hard to decarbonise, requires a deeper analysis of how user demand, material flows, and energy consumption interact to shape future electricity demand [58]. Poor representation of these socio-economic drivers can lead to misleading policy insights. Thus, to design effective policies, it is essential to account for the underlying human drivers of demand [14]. Many IAMs also overlook the interaction between the energy system and society at large, missing key impacts on people’s lives and livelihoods. Factoring in social considerations, such as employment and health outcomes, can help reveal the distributional effects of energy transitions, especially on low-income and underrepresented communities [19]. By identifying these impacts, policymakers can better ensure that energy transitions are inclusive and equitable. These social considerations can also inform trade-offs tailored to specific development priorities. Additionally, highlighting social factors can make energy plans more attractive to financiers and increase the likelihood of securing support [59]. For example, Costa Rica’s National Decarbonisation Plan, launched in 2019, was developed with multisectoral stakeholder input and long-term modelling that considered poverty reduction, job creation, and air quality. This inclusive approach helped mobilise US $2.4 billion in concessionary finance to support the plan’s implementation [8]. Integrating social dimensions is therefore critical to generating actionable insights for a sustainable and just energy transition. 2.1.4. Lack of Financial Markets and Mechanisms Table 2highlights that IAMs currently do not incorporate financial planning. While IAMs excel at techno-economic optimisation and long-term mitigation analysis, they rarely include detailed representations of financial markets or financing mechanisms. As a result, they fail to reflect real-world constraints such as capital availability, financial risk, and investor behaviour, leading to misestimations of the true costs of the energy transition [13,15,60,61]. Sanders et al. [15] notes that "there have not been attempts to explicitly link finance [...] to transition investments by technology". Models usually overlook key aspects of project financing, such as the balance between equity and debt, distinctions between public and private capital, and the role of interest rates—all of which are essential to the implementation of energy projects as they can influence the pace of the energy transition [61]. There is also limited analysis of how financial structures evolve over time and the financial structuring of projects, such as the use of balance sheet financing, where projects are funded directly by companies, or project finance, where standalone project entities are created. [15]. Keppo et al. [13] further argues that IAMs could be strengthened by incorporating detailed financing schemes, debt budgeting, and the dynamic interaction between debt accumulation and interest rates. Addressing these gaps is increasingly important. A lack of available finance could slow or stall the energy transition, and this concern has begun to receive attention from policymakers, central banks, and the broader financial community [13]. Incorporating financial market dynamics into IAMs would make their results more robust and relevant for policy and investment planning. Furthermore, such improvements would enable the exploration of how green Tan et al.: Preprint submitted to Elsevier Page 9 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning similar reference energy system for both tools is crucial to ensure seamless integration between the tools. To assist users in creating the necessary reference energy system for linking the tools, including technologies and commodities, Table 5provides links to example reference energy system diagrams for the industrial, transport, household, and services sectors that one can adopt, adapt, and apply. It is also important to highlight that connecting a simulation tool like MAED with an optimisation tool like OSeMOSYS requires careful consideration to preserve the qualities of both tools. For example, if the user wishes to provide the ‘technology’ activity mix for each end-use demand (e.g., electric car, gasoline car, coal cooking, gas lighting, etc.) as it may be calculated from MAED-D, a sensible ‘buffer’ of around 20% above and below the estimated MAED-D ‘technology’ activity mix will allow OSeMOSYS flexibility in calibrating a least-cost model while still acknowledging MAED-D results. This can be done using the OSeMOSYS parameters total technology annual activity upper limit and total technology annual activity lower limit. Figures 6to 10 detail the necessary outputs from MAED to integrate with OSeMOSYS. Additionally, a step-by-step exercise on integrating both tools can be accessed via Table 4. 4.1.2. OnSSET to OSeMOSYS (feedback loop) OnSSET can be run simultaneously alongside MAED, obtaining the currently unelectrified off-grid demand estimates for the housing, services, agriculture, education, and healthcare sectors. These demands can be incorporated into OSeMOSYS as pre-defined demands. However, note that when combining these two tools, the geospatial aspect of OnSSET will be lost. Nonetheless, OSeMOSYS is a model that can be used at different disaggregation levels. Thus, this challenge may be solved by creating multiple OSeMOSYS models at a village or regional level and stitching these multiple OSeMOSYS models together for a national analysis. Similar to the MAED-OSeMOSYS integration, the total off-grid demand can be placed in the specified annual demand parameter if there is an accompanying specified electricity demand profile for off-grid users; otherwise, the accumulated annual demand should be used. Note that all outputs from OnSSET are presented by a five-year ‘snapshot’ period; thus, the user must divide the results by 5 to get annual values. There is therefore a limitation in this method; however, the duration of the ‘snapshot’ period is sufficiently small to offer a reasonably acceptable representation of off-grid demand. The reference energy system within OSeMOSYS must also mirror the system in OnSSET, incorporating off-grid technologies such as stand-alone diesel, hydropower minigrid, hybrid solar PV minigrids, etc. To assist users in creating the necessary reference energy system for linking the two tools, Table 5provides a link to an example reference energy system diagram for the off-grid system that one can adopt, adapt, and apply. If the user wishes to disaggregate the total off-grid demand by technology activity, a sensible ‘buffer’ of around 20% above and below the estimated OnSSET technology activity mix will allow OSeMOSYS further flexibility in calibrating a least-cost model. This can be done using the OSeMOSYS parameters total technology annual activity upper limit and total technology annual activity lower limit. Besides the demand, capacity factors and renewable potential for off-grid technologies can also be obtained from OnSSET and inputted into OSeMOSYS. Lastly, the network distribution costs can be calculated from OnSSET and used as the capital costs for the transmission and distribution systems within OSeMOSYS. Throughout the integration process, the user must ensure unit consistency between the two tools and divide OnSSET results by 5 to obtain annual values. Additionally, a feedback loop between OnSSET and OSeMOSYS can be established, where the average electricity price from OSeMOSYS can be calculated and fed back into OnSSET. This updated grid electricity cost inputted into OnSSET will impact the offto on-grid cost optimisation process. To fully calibrate the two models, the process of integrating OnSSET to OSeMOSYS and vice versa should be repeated until there is a convergence of results, i.e., until the LCOE produced between two iterations of OnSSET runs has a difference of less than 10% [84]. Figures 11 and 12 detail the necessary outputs from OnSSET for integrating with OSeMOSYS, and vice versa. In-depth hands-on exercises for the OnSSET-OSeMOSYS and OSeMOSYS-OnSSET links can also be accessed via Table 4. 4.1.3. OSeMOSYS to FlexTool (feedback loop) FlexTool conducts a single-year flexibility analysis, including detecting possible flexibility constraints within the power system. Hence, in an integrated analysis between OSeMOSYS and FlexTool, OSeMOSYS provides the costoptimal capacity expansion plan of the power system in the future, and FlexTool evaluates the proposed system from Tan et al.: Preprint submitted to Elsevier Page 16 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning an operational perspective. One important aspect the user has to consider when performing an integrated OSeMOSYSFlexTool analysis is that the modelling approach of the two tools for the power sectors is different. OSeMOSYS models the power sector connected to the rest of the energy system. Therefore, electricity is modelled as a demand commodity consumed in end-use sectors and supplied by power plants. Alternatively, FlexTool deploys a more detailed and topological power-based approach. Thus, the electricity system is modelled as a node (or multiple nodes interconnected with transmission lines), where the power supply technologies and the electricity demand are located. Moreover, FlexTool accounts for power system operational characteristics (e.g., inertia, reserve provision, etc.) and additional detailed technology options not included in OSeMOSYS (e.g., electricity interconnections, demand response, etc.). Ultimately, the two tools differ in terms of temporal resolution. OSeMOSYS is used for long-term multi-year and multi-resource analyses, and each year is modelled with some sampled representative timeslices. FlexTool, on the other hand, focuses on a single year, which is modelled with a full hourly resolution. From a technical point of view, the OSeMOSYS implementation shall precede the FlexTool analysis. The FlexTool model can be partially set up based on the inputs and outputs of the OSeMOSYS model. The key parameters transferred from the OSeMOSYS inputs to FlexTool are the annual electricity demand and the techno-economic characteristics of the power technologies. From OSeMOSYS outputs, the key parameter needed for populating FlexTool inputs is the capacity of the power generators and electricity storage units. Additionally, if the OSeMOSYS analysis affects the final electricity demand, this parameter shall be sourced from the OSeMOSYS results. For example, if OSeMOSYS decides whether and to what extent electric vehicles are utilised to satisfy transport activity demand, the amount of electricity consumed by electric vehicles shall be sourced from OSeMOSYS outputs. The transfer of input and output parameters of OSeMOSYS to FlexTool is illustrated in Figure 13. Table 4also provides a link to a step-by-step exercise on integrating the two tools. It should be noted that a series of parameters necessary for FlexTool cannot be sourced directly from OSeMOSYS. FlexTool provides a more detailed representation of power technologies and thus requires additional technology characteristics compared to OSeMOSYS. Typical examples are the so-called flexibility parameters of power-generating units, such as ramping capabilities and minimum stable load, among others. Such parameters need to be sourced from the international literature. Additionally, since the temporal resolution is different, parameters expressed in a time series format, like the demand profile and the capacity factor profiles of renewable energy plants, cannot be directly transferred from OSeMOSYS’ input file to FlexTool. Nevertheless, the hourly demand profile for the chosen year can be obtained from MAED-EL (see Section 4.1.4). Furthermore, it is important to ensure that the full time series of capacity factors of power plants used in FlexTool match the ones used to produce the representative reduced time slices in OSeMOSYS. If a FlexTool analysis highlights issues with the power system flexibility, one can delve into two solutions. If concerns such as curtailment or loss of load are present in the analysis, the user can: (1) integrate different flexibility options such as electric vehicles, battery storage, power-to-heat, power-to-hydrogen, and demand response. Nonetheless, data regarding these technologies will be required for this sector-coupling analysis; (2) run FlexTool in ‘investment mode’, where an alternative capacity expansion plan can be analysed. This can then be integrated as parameters in OSeMOSYS. Similar to the OnSSET-OSeMOSYS integration, a feedback loop may also be explored with OSeMOSYS and FlexTool, where a convergence of results is needed to fully calibrate the two models [84]. 4.1.4. MAED to Flextool As mentioned in Section 4.1.1, MAED-D calculates the average electricity demand growth rate for a chosen year, which can be fed into MAED-EL to derive the hourly electricity demand profiles for the industrial, transport, household, and services sectors. MAED-EL, therefore, generates the hourly demand profile for the electricity end-use demand for each sector. As FlexTool requires the hourly demand profile for one year to carry out the power flexibility assessment, the hourly electricity demand profile from MAED for a selected year can be directly inserted into the FlexTool. However, it should be noted that a typical FlexTool model combines the hourly demand profiles of the industrial, residential, and services sectors in a single profile, while treating transport separately—effectively modelling two different ‘grids’ in FlexTool. The segregation is due to the electrified transport demand behaving differently from other sectors. While industrial, residential, and service loads are typically inflexible and follow predictable daily patterns, transport demand can be more dynamic, with charging loads that can be shifted based on grid conditions or user preferences. Additionally, some transport modes can provide demand-side flexibility through managed charging or Tan et al.: Preprint submitted to Elsevier Page 17 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning vehicle-to-grid interactions, which can also be modelled on FlexTool. Despite this, linking MAED and FlexTool is still straightforward, where the industrial, residential, and services hourly demand profiles from MAED are summed into a single profile, while the transport sector hourly demand profile can be used directly in FlexTool. The MAED-FlexTool link is illustrated in Figure 13, and a detailed step-by-step guide on this is accessible via Table 4. 4.2. Energy-Land-Water Sub-Framework The energy-land-water sub-framework extends the energy-related foundation laid by OSeMOSYS, offering an integrated, nexus-based framework encompassing energy, water, land, and climate systems via the CLEWs approach. This symbiotic relationship allows for a comprehensive exploration of resource systems through the lens of nexus thinking. By adopting a nexus approach, CLEWs transcends sectoral silos to recognise that these energy and earth systems are deeply interconnected and interdependent [27,69]. For example, energy production often requires significant water use, particularly in thermal power plants that rely on water for cooling. More water-intensive energy sources can strain water supplies, impacting availability for other uses. At the same time, water production also depends on energy inputs for pumping water from surface and groundwater sources for irrigation, thermal power plant cooling, desalination, and public water supply. Similarly, land plays a crucial role in the energy-water nexus. The allocation of land for energy production, such as growing biofuel crops or constructing solar and wind farms, competes with food production and natural ecosystems. Poor practices like excessive fertiliser use on energy crops can also pollute waterways, and deforestation for biofuels impacts the climate through lost carbon sequestration. Simultaneously, energy is needed for land management, such as powering agricultural equipment for crop cultivation or facilitating land-use changes. Additionally, land use significantly affects water resources, as agricultural activities influence water quality through nutrient runoff, pesticide contamination, and deforestation, which can reduce groundwater recharge. Finally, land use is deeply tied to climate, as agricultural expansion contributes to deforestation and methane emissions from livestock, whereas sustainable land management practices, including cover crops, reduced tillage, and wetland restoration, can help mitigate climate change. For CLEWs to operate within OSeMOSYS, they must abide by the same modelling principles and have a predefined demand. Regarding the land component, this requirement comes in the form of a land use demand. For example, every type of land use represented in the model, i.e., tree cover, inland water, built-up land, etc., will require a representative demand. Typically, CLEWs models also incorporate arable land, which is ‘governed’ by a demand for the specific crops represented in the model. Additional complexity can also be added in the form of low-input arable land (i.e., rainfed crops) and high-input arable land (i.e., irrigated crops). Similarly, demand is also required for water and can be represented in terms of accumulated industrial, residential, commercial and agricultural demands. Thus, integrating CLEWs will influence the energy capacity expansion plan based on resource availability, competition, and system-wide constraints. Ramos et al. [85] provides an example reference energy system named ’RCLEWs’ that one can modify and incorporate into OSeMOSYS for a CLEWs analysis. A hands-on methodology on further expanding an OSeMOSYS model to include the CLEWs approach is also accessible via Table 4. 4.3. Social and Environmental Sub-Framework The social and environmental sub-framework extends the OSeMOSYS techno-economic approach to include social and environmental impacts and benefits that are directly affected by the energy sector. Recognising the growing importance of integrating social and environmental factors into energy systems modelling for value-driven insights, as emphasised by Dioha et al. [19] and Pfenninger et al. [16], this sub-framework incorporates the SIBs approach to quantify socio-economic variables. Initially, three aspects have been analysed for this sub-framework: (1) job generation, (2) air pollution, and (3) socio-environmental costs stemming from global warming. Firstly, the energy transition offers substantial job creation opportunities, with clean technologies exhibiting high employment rates per installed capacity [86]. Since 2019, there has been a notable increase in the workforce within the energy sector, mainly driven by the rising prominence of clean energy. This shift has resulted in clean energy surpassing fossil fuels in terms of employment [87]. Using OSeMOSYS, the potential job creation through construction and manufacturing (C&M) can be assessed by incorporating coefficients that represent the number of C&M jobs produced for each added unit of installed capacity per technology [88]. Similarly, the operation and maintenance (O&M) jobs created can be modelled by utilising coefficients that represent the number of O&M jobs produced for each unit of Tan et al.: Preprint submitted to Elsevier Page 18 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning total installed capacity throughout the technology lifetime [88]. Both of these actions would involve creating ’dummy’ C&M and O&M commodities and technologies on OSeMOSYS, which are then linked to the parameters input to new capacity ratio and total capacity ratio. A hands-on practice about the modelling approach is accessible via Table 4, and can be extended further to consider indirect employment [89,90]. Secondly, the combustion of fossil fuels, particularly coal, results in outdoor air pollution, including fine particulates, sulfur dioxide (SO2), nitrogen oxides (NO𝑥), and low-lying ozone. Natural gas combustion produces moderate amounts of NO𝑥. Gasoline and diesel combustion can yield SO2, NO𝑥, volatile organic compounds (VOCs), and fine particulates, with emission rates influenced by vehicle regulations and fuel quality. Collectively, fossil fuel-related air pollution is estimated to cause 4.5 million premature deaths annually [91]. By employing estimated externality coefficients within OSeMOSYS, we can estimate the costs related to mortality and morbidity for individuals exposed to elevated concentrations of fine particulates in outdoor environments. This approach involves identifying and calculating the national air pollution cost [91], creating additional emission types within OSeMOSYS, updating the emission activity ratio, and defining the emissions penalty. A hands-on guide about the modelling approach is accessible via Table 4. Thirdly, the costs of socio-environmental issues are rising due to extreme climate variations caused by global warming [91]. These expenses materialise in diverse outcomes affecting communities, such as crop loss, damage to infrastructure due to flooding, forest fires, water scarcity, and other related impacts. Previous studies have assessed these externalities using OSeMOSYS [28,29]. This process is similar to the above, whereby the global warming cost by country and fuel is first identified and calculated [91], then creating additional emission types within OSeMOSYS, updatingthe emission activity ratio,anddefiningthe emissions penalty. A hands-onguideabout themodellingapproach is accessible via Table 4. While these methods allow a quick quantification of social and environmental implications and benefits using a straightforward multiplier methodology with annual coefficients, there is potential for more profound, complex analyses. Additional research has been conducted to extend the use of OSeMOSYS as a tool for assessing socioeconomic variables. For instance, soft-linking with macroeconomic models can yield insights into the economyenergy nexus and the impact of emission reduction on economic growth [92]. Moreover, ad-hoc models can be linked to OSeMOSYS to assess fiscal impacts and distributional effects associated with decarbonization measures among selected stakeholders [93]. Examples like these will be studied to propose additional applications of the SIBs approach. 4.4. Visualisation Sub-framework The visualisation sub-framework requires a ‘base’ OSeMOSYS input model—one with input data but minimal constraints—to allow PathCalc flexibility when generating thousands of OSeMOSYS scenarios. The results are displayed in PathCalc’s interactive web-based tool. The visualisation sub-framework contains three key steps. The first two steps are done using the Excel-based ‘PathCalc Planner’. Firstly, users select the levers to include in the web tool from a catalogue in the Planner, which is based on existing 2050 Calculators. Up to six levers can be chosen, each of which can be comprised of several sub-levers. The levers’ ambition levels are then defined. Ideally, these draw on stakeholder consultations but may also be assumptions. By default, the lowest level of ambition (‘no effort’) aligns with the OSeMOSYS ‘base’ file scenario. Using the OSeMOSYS ‘base’ file in combination with lever data, the Planner generates thousands of OSeMOSYS input files for each combination of lever and ambition level. Lastly, the thousands of OSeMOSYS input files are run using a script, and results are visualised on the web tool using the PathCalc visualisation code. By visualising thousands of interactive, optimised scenarios, users can explore the different optimal pathways to enhance and refine the insights from OSeMOSYS, thereby creating a feedback loop and enabling more nuanced decision-making. A step-by-step guide on integrating OSeMOSYS and PathCalc, including links to prototypes of the PathCalc Planner and modelling script, is accessible via Table 4. 4.5. Financial Sub-framework The financial sub-framework forms the latter part of IMPACCT and includes three tools: OSeMOSYS, MINFin, and FINPLAN. Figure 4lists the key outputs and inputs needed for this part of IMPACCT. Tan et al.: Preprint submitted to Elsevier Page 19 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning Figure 4: Key outputs and inputs needed for the financial sub-framework of IMPACCT, consisting of OSeMOSYS, MINFin, and FINPLAN. 4.5.1. OSeMOSYS to MINFin (feedback loop) MINFin is a model designed to examine and develop financing strategies for energy pathways at the national level. The model, therefore, complements OSeMOSYS by translating the techno-economic capacity expansion plan into a coherent financing strategy. The model comprises three input pillars: investment needs, financing baseline, and funding baseline. Financial outputs from OSeMOSYS, such as capital investment and operating costs, serve as essential parameters for input into MINFin. Firstly, the capital investment forms the foundation for estimating future power plant investment needs in MINFin. The operation and maintenance costs, on the other hand, allow for a comparative assessment of fossil fuel expenses and potential savings between scenarios, adding to a country’s funding baseline in MINFin. Emissions from OSeMOSYS can also be included in MINFin to showcase the environmental benefits of transitioning to a scenario such as Net-Zero, offering users a strong incentive to pursue more ambitious decarbonization strategies. Furthermore, an implicit cost of carbon can be applied to the carbon emissions associated with the expansion plan, which is of potential interest from a carbon market perspective. A feedback loop with OSeMOSYS can further refine the capacity expansion plan. The cost of capital estimates from the financing baseling in MINFin can be integrated into OSeMOSYS as technology-specific discount rates, representing hurdle rates and influencing the merit order of technologies. Generally, if MINFin results indicate that the national financing strategy is unachievable, adjusting OSeMOSYS to set less ambitious climate targets can delay the surge in financing requirements and ensure a more feasible investment trajectory, preventing financial bottlenecks. The key parameters to create an OSeMOSYS-MINFin integration are illustrated in Figures 14 and 15. A hands-on exercise for integrating OSeMOSYS and MINFin is accessible via Table 4. However, note that additional financial parameters are needed for a complete MINFin analysis. This includes historic lending volumes and terms of finance, information on sector cashflows, government budgets and international grants, which can be gathered from open-access online databases. 4.5.2. OSeMOSYS to FINPLAN (feedback loop) After developing a national capacity expansion plan, the next step often involves assessing the financial feasibility of a particular power plant over its operational lifetime. Thus, FINPLAN can be integrated into the framework after OSeMOSYS. This integration enables a comprehensive understanding of the complexities regarding shareholders’ return, cash inflow and outflow, financial ratios, and others, in sustaining the financial viability of power plants over time. As these financial parameters are specific to a singular power plant, or a group of power plants under one company, it is strongly suggested that one type of power plant, or a group of the same type of power plants constructed in the same year, as recommended by OSeMOSYS, be modelled on FINPLAN instead. If the user wishes to model the whole power system, it is advised to create multiple individual case studies that can be joined together externally. From OSeMOSYS, financial results such as capital investments, O&M costs, fuel costs, and electricity prices can be translated into Tan et al.: Preprint submitted to Elsevier Page 20 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning FINPLANfor the model tounderstand thecostsassociated withdeveloping and runningapower plant. Additionally, the capacity expansion volume and electricity production, as calculated from OSeMOSYS, can be brought into FINPLAN to define the size of the power plant(s). Depending on FINPLAN results, a feedback loop can be established with OSeMOSYS to further refine the capacity expansion plan based on project-level financial feasibility. For example, if FINPLAN indicates that the proposed projects are financially unviable, adjusting OSeMOSYS parameters to set less ambitious climate targets can ensure that new power plants operate without financial constraints. The key parameters to create an OSeMOSYS-FINPLAN integration are illustrated in Figure 16. A step-by-step guide on linking the tools together can also be accessed via Table 4. 4.5.3. MINFin to FINPLAN (feedback loop) To capture the impacts of national financial conditions on power plants, loan parameters from MINFin can be inputted into FINPLAN. This includes the loan repayment structure (principal or principal+uniform repayment), average term of loan, and associated interest rate. Combining these data with other external data necessary for a FINPLAN analysis (exchange rates, inflation rates, construction time, proportion of international and national funding, etc.) will allow the user to identify the additional equity, debt, or subsidy needed from organisations for the power plant(s) to be financially operational throughout the modelling period. Further, the revenue calculated from FINPLAN can be incorporated back into MINFin’s funding baseline to create a feedback loop. Revenue income is a crucial metric for assessing the financial performance and growth of a company. For instance, if a country anticipates high export revenues, this income can be included in the pool available for funding capital investments in domestic power projects. However, the cost of capital and financial structures in FINPLAN must align with those in MINFin when establishing the feedback loop. This is to ensure that inflows and outflows, including revenue streams, are consistently calibrated, preventing inconsistencies in financial estimates. The key parameters to create a FINPLAN-MINFin integration are illustrated in Figure 14. Hands-on exercises for the interactions of FINPLAN and MINFin are accessible via Table 4. 5. Discussions 5.1. Benefits and Innovations IMPACCT is the first integrated assessment framework to combine open-source code, solvers, user-friendly interfaces, and comprehensive documentation. By improving accessibility compared to existing IAMs, it enables analysts in resourceand skills-constrained contexts, including marginalised communities and non-expert government officials, to undertake energy modelling and financial planning aligned with their energy transition goals [14,24]. This supports capacity building and promotes inclusivity, allowing broader participation in energy policy and decisionmaking for a just and equitable transition. Another advantage is that IMPACCT captures social drivers and implications, aspects which are often overlooked in current IAMs [17,19]. In MAED, energy demand in all sectors is modelled with a bottom-up methodology by specifying socio-economic, technological, and demographic drivers—for instance, household size, appliance ownership, shifts in transport modes or improvements in industrial energy efficiency. SIBs complements this by quantifying social outcomes of technical energy transition plans, such as but not limited to job creation, global warming costs, and air quality, providing a social view of potential policy implications. To the best of our knowledge, IMPACCT is the first IAM framework to explicitly embed financial planning. MINFin and FINPLAN enable users to explore financial market dynamics and project financing mechanisms into the future, including equity versus debt, public and private funding sources, and distinctions between balance sheet and project financing. This integration of financial analysis addresses an important gap identified in the IAM literature [13,15], improving the finance lens of energy transition scenarios. A key strength of IMPACCT is its integration of technical, social, and financial dimensions of energy transitions. It incorporates well-established tools that have been tested and documented by domain specialists, and it can adapt to diverse user needs, whether an in-depth analysis of specific processes (feedback loops) or broad system-wide assessments (pipeline assessment). For example, users interested in the integration of variable renewables can utilise OSeMOSYS and FlexTool in a feedback loop to evaluate operational impacts, while others who are interested in finance planning can link OSeMOSYS outputs to MINFin and FINPLAN only to assess the financial feasibility of Tan et al.: Preprint submitted to Elsevier Page 21 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning infrastructure investments. Lastly, IMPACCT’s broad sectoral coverage, accessibility, and collaborative design encourage cross-sectoral and cross-institutional dialogue. Government agencies, academic institutions, financial organisations, and other stakeholders often work in silos, each adopting their own modelling approaches [94]. IMPACCT encourages stakeholders to share data and modelling outputs to co-develop harmonised and credible energy transition plans. For instance, an energy ministry and finance ministry might jointly utilise the OSeMOSYS–MINFin–FINPLAN workflow to ensure technology pathways are financially viable. Similarly, an academic engineering department could collaborate with the energy ministry to refine an OnSSET–OSeMOSYS assessment. By explicitly linking sectoral analyses, IMPACCT can help to bridge sectoral silos, enable iterative model refinement and support the development of more collaborative, evidence-based policy outcomes. 5.2. Limitations and Complexities The coupling of specialised models in IMPACCT introduces interoperability challenges, which currently require manual intervention to ensure consistency. For example, transferring technical and financial outputs from OSeMOSYS to FINPLAN requires careful conversion between units, such as from PJ to GWh. Similarly, adapting the hourly demand profiles from MAED to OSeMOSYS’s 96-timeslice temporal structure requires additional data processing. Moving data from spatially detailed tools like OnSSET to the single-node structure of OSeMOSYS also results in a loss of spatial granularity; however, this can be partially addressed by running multiple village-scale OSeMOSYS models and aggregating the results. Nonetheless, these manual conversions for units, temporal, and spatial structure are documented, and step-by-step guidance is provided to the users (Table 4). The unidirectional IMPACCT process can also be resourceand time-intensive. Integrating seven distinct tools requires significant effort to align data structures, assumptions, and outputs, demanding both labour and time. In many regions, key data such as energy demand, technology costs, or financial terms are scarce, outdated, or unofficial, introducing uncertainty into model results [17,47]. Addressing these gaps requires not only technical fixes but also extensive stakeholder engagement to validate assumptions and improve data accuracy. Thus, capacity building, better institutional data-sharing, and targeted data collection are essential to ensure that IMPACCT informs decision-making in a meaningful and context-sensitive way, rather than reinforcing data limitations or systemic gaps. Although IMPACCT addresses several gaps in the existing literature by integrating techno-economic, social, and financial dimensions, it does not yet cover all aspects of the energy transition. Key areas such as critical minerals supply chains, clean cooking solutions, andmacroeconomic feedbacks between energy, investment, and growth remain outside the current framework. However, IMPACCT is designed to be modular and expandable, and future developments aim to incorporate these aspects through additional tools and models, as discussed in the following subsection. Like all IAMs, IMPACCT simplifies the complexity of real-world systems. Combining multiple tools enhances the representation of energy, financial, and social dimensions, but each tool brings its own assumptions and limitations, as detailed in their respective references. No model can fully capture the dynamic, nonlinear, and interconnected nature of energy systems. Yet, abandoning IAMs is neither practical nor desirable—they remain essential for testing cost and performance assumptions and have been central to shaping global low-carbon transition scenarios [42]. Rather than viewing results as definitive answers, users should apply models to explore alternative pathways toward a shared vision of a sustainable energy future, as recommended by Sgouridis et al. [95]. 5.3. Initial Uptake and Future Work of IMPACCT At the time of writing, IMPACCT is being utilised by the Climate Compatible Growth (CCG) programme’s capacity-building initiative, which brings together participants from diverse institutions and government departments within a country. This approach enables different stakeholders to specialise in individual tools while working collaboratively to link them across sectors and disciplines [96]. Beyond national efforts, IMPACCT is also gaining traction among international organisations [97], as more actors seek to soft-link modelling tools for a wider understanding of energy transitions. Tan et al.: Preprint submitted to Elsevier Page 22 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning In terms of future work, IMPACCT will incorporate additional open-source tools for a broader coverage of the energy transition. However, it is critical to recognise that there are trade-offs and escalating complexities in integrating further modelling tools. Thus, selective consideration is needed to prioritise tools that offer the most value. A preliminary assessment identified nine potential tools that can be appended to IMPACCT, extending its coverage to include the data, cooking, transport, critical minerals, and life cycle assessment aspects, which are all crucial to facilitate energy systems planning. These tools are listed below. Figure 5further displays the possible linkage of the tools within IMPACCT. 1. Energy Access Explorer (EAE): an online, interactive, geospatial platform developed by the World Resources Institute that enables users to identify high priority areas where energy access can be expanded [98]. Once these areas are identified, the user may carry out an in-depth analysis either on OnSSET or OnStove, ensuring efficient resource allocation. 2. Open Source Spatial Clean Cooking Tool (OnStove): a spatial tool developed by KTH comparing the relative potential of different cookstoves based on their costs and benefits [99]. The tool may interact with OnSSET to combine clean cooking and electricity access [100], as well as feed into OSeMOSYS to develop a capacity expansion plan for clean cooking. 3. Energy Balance Studio (EBS): a tool created by the IAEA that offers a systematic framework for organising energy statistics data. This ensures data consistency, accuracy, and comparability. The data can then be used as an input for energy planning models such as MAED. 4. Open Source Mobility Model (OSeMobility): currently under development by the University of Strathclyde, Oxford University, and Imperial College London, this modular transport systems modelling framework allows stakeholders to explore credible transport futures through a socio-technical approach [101]. The tool provides a direct interface with OSeMOSYS to enable quantification of transport futures on broader energy, emissions, and development goals. 5. Mat-dp: a model created by the University of Cambridge that computes the amount and type of materials required for constructing various systems or resource transformations, including those found along any supply chain. Nonetheless, it is mainly utilised to study the materials needed for constructing low-carbon systems and estimating the environmental implications associated with these materials [102]. This model can, therefore, expand an OSeMOSYS analysis to understand the materials required based on a capacity expansion plan. 6. Multi-Regional Analysis of Regions through Input-Output (MARIO): a framework developed by the Polytechnic University of Milan to model additional supply chains through a hybrid life cycle assessment approach at mesoand macro-scale [103]. This tool can be integrated with OSeMOSYS to analyse the supply chain effects of technological or policy interventions. 7. Fossil Fuel Retirement Model (FFRM): a tool originally constructed by The World Bank that enables the user to estimate the cost of compensating investors for early withdrawal of fossil fuel power plants [104]. This can serve as a continuation of a renewable energy capacity expansion plan proposed by OSeMOSYS. Additionally, results from this model, such as the present value of the foregone net revenues of stranded fossil fuel plants, can be fed into MINFin. 8. Financing Costs for Renewables Estimator (FinCoRe): a tool developed by Imperial College London that estimates the cost of capital for renewable energy power plants [105]. These capital costs can be fed into OSeMOSYS, FlexTool, MINFin, and FINPLAN to assess the least-cost capacity expansion plan and its financial feasibility. 9. Climate Finance Tracker (FinTrack): a tool created by Imperial College London, which identifies potential future concessional finance envelopes [106] and can influence the financial assessments and planning in MINFin and FINPLAN. IMPACCT also currently integrates selected models that continue to evolve. As these underlying tools and models advance, IMPACCT will be updated accordingly to incorporate the latest improvements, maintaining its accuracy and relevance. Additionally, the current social and environmental analysis (Section 4.3) covers only a subset of aspects. Future versions will broaden this scope to include further dimensions such as the economy-energy nexus, income levels, affordability, livelihoods (including road safety), and gender equality. A unidirectional script is in development to automate the integration of the IMPACCT tools, enabling streamlined techno-economic to financial planning assessments. The hard-linking will speed up analysis, reduce human error, and Tan et al.: Preprint submitted to Elsevier Page 23 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning Figure 5: Linking of additional tools including EAE, OnSTOVE, EBS, OSeMobility, Mat-dp, MARIO, FFRM, FinCoRe, and FinTrack. These tools are highlighted in grey. Solid arrows depict a one-way input-output link in the direction of the arrow, while dashed lines represent feedback loops within the tools connected by the bidirectional arrow. support framework validation and uncertainty quantification. It will also enable advanced methods like robust decision making (RDM) to identify resilient energy transition pathways under deep uncertainty, which would be difficult to explore manually. Lastly, to enhance transparency and build confidence in the workflow, a validation exercise of IMPACCT is ongoing. The authors are conducting a unidirectional analysis that systematically passes results and inputs through each tool. This validation process is to be published in a series of papers following this conceptualisation article and will reveal data transformations, assumptions, and intermediate outputs. By documenting the entire process, the authors aim to improve reproducibility, credibility, and transparency, responding to calls for higher standards in open science and model validation [46]. Overall, the urgent need for integrated and automated tools to support rapid and robust energy planning decisions underscores the timeliness of this publication. By detailing IMPACCT and its conceptualisation, this work makes a critical contribution at a pivotal moment for both researchers and policymakers, laying the groundwork for future developments. 6. Conclusions IMPACCT is the first of its kind in soft-linking seven significant open-source and user accessible modelling tools to provide a comprehensive approach to integrated energy systems planning. It includes techno-economic, energy-landwater, social and environmental, visualisation, and financial aspects. This interdisciplinary integration bridges critical gaps in existing IAMs, particularly enhancing transparency, stakeholder accessibility, and inclusion of social and financialconsiderations—areasoftenoverlookedin traditional modelling approaches.By doing so,IMPACCTsupports more coordinated, evidence-based, and just decision-making processes essential for sustainable energy transitions. IMPACCT’s flexible design also allows users to tailor their analysis to their specific needs and capacities. Whether applying the full unidirectional pipeline that integrates all seven tools, selectively using a subset of tools with feedback loops, or adopting any combination in between as outlined by the IMPACCT methodology in this paper, users can tailor the framework to fit their institutional context and analytical goals. To ensure robustness and build confidence among users, a comprehensive validation process is underway. The authors are preparing a series of forthcoming papers, organised by tool, that will document the framework’s input and output data as well as its validation. This transparent Tan et al.: Preprint submitted to Elsevier Page 24 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning approach will strengthen IMPACCT’s credibility and offer practical guidance for its application across diverse contexts. Nevertheless, it is important to recognise that while these actions, along with IMPACCT, support energy and financial planning, they are not individually sufficient to guide a nation through the entire transition from energy systems modelling to investment-ready plans. To fully realise a just and sustainable energy transition, additional measures such as increased open data licensing in the public sector, building institutional capacity, fostering stakeholder engagement, employing iterative robust decision-making analysis, and others, will be critical [8,18]. Ethics Statements Not applicable. Declaration of Competing Interests The authors declare that they have no known competing financial interests or personal relationships that have or could be perceived to have influenced the work reported in this article. Acknowlegements Part of this work was funded and produced for the International Energy Agency (IEA) and the Asian Development Bank (ADB), whom the authors would like to thank for their continued support, involvement, and guidance during this study. The authors would also like to acknowledge core funding from the UK Aid from the UK Government via the Climate Compatible Growth programme. However, the views expressed herein do not necessarily reflect the UK government’s official policies. Lastly, the authors would like to thank the two anonymous reviewers for their helpful comments in shaping the paper. CRediT authorship contribution statement Naomi Tan: Conceptualization, Methodology, Writing - Original Draft, Writing - Review & Editing, Visualization. Ioannis Vrochidis: Conceptualization, Methodology, Writing - Original Draft, Writing - Review & Editing. Hannah Luscombe: Conceptualization, Methodology, Writing - Original Draft, Writing - Review & Editing. Emma Richardson: Conceptualization, Methodology, Writing - Original Draft, Writing - Review & Editing. Fernando PlazasNiño: Conceptualization, Methodology, Writing - Original Draft, Writing - Review & Editing. Kane Alexander: Methodology, Writing - Original Draft, Writing - Review & Editing. Leigh Martindale: Methodology, Writing - Original Draft. Neve Fields: Methodology, Writing - Original Draft. Mark Howells: Writing - Review & Editing, Supervision. Vivien Foster: Writing - Review & Editing, Supervision. John Harrison: Writing - Review & Editing, Supervision. References [1] DeCarolis J, Daly H, Dodds P, Keppo I, Li F, McDowall W, et al. Formalizing best practice for energy system optimization modelling. Applied energy 2017;194:184–98. doi:10.1016/j.apenergy.2017.03.001. [2] Pfenninger S. Energy scientists must show their workings. Nature 2017;542(7642):393–. doi:10.1038/542393a. [3] Pfenninger S, DeCarolis J, Hirth L, Quoilin S, Staffell I. The importance of open data and software: Is energy research lagging behind? Energy Policy 2017;101:211–5. doi:10.1016/j.enpol.2016.11.046. [4] Pfenninger S, Hirth L, Schlecht I, Schmid E, Wiese F, Brown T, et al. Opening the black box of energy modelling: Strategies and lessons learned. Energy Strategy Reviews 2018;19:63–71. doi:10.1016/j.esr.2017.12.002. [5] Morrison R. Energy system modeling: Public transparency, scientific reproducibility, and open development. Energy Strategy Reviews 2018;20:49–63. doi:10.1016/j.esr.2017.12.010. [6] Oberle S, Elsland R. Are open access models able to assess today’s energy scenarios? Energy Strategy Reviews 2019;26:100396. doi:10.1016/j.esr.2019.100396. [7] Groissböck M. Are open source energy system optimization tools mature enough for serious use? Renewable and Sustainable Energy Reviews 2019;102:234–48. doi:10.1016/j.rser.2018.11.020. [8] Jaramillo M, Quirós-Tortós J, Vogt-Schilb A, Money A, Howells M. Data-to-Deal (D2D): Open Data and Modelling of Long Term Strategies to Financial Resource Mobilization-the case of Costa Rica. 2023. doi:10.33774/coe-2023-sqbfm-v5; preprint. [9] Plazas-Niño F, Ortiz-Pimiento N, Montes-Páez E. National energy system optimization modelling for decarbonization pathways analysis: A systematic literature review. Renewable and Sustainable Energy Reviews 2022;162:112406. doi:10.1016/j.rser.2022.112406. [10] Dowlatabadi H. Integrated assessment models of climate change: An incomplete overview. Energy Policy 1995;23(4-5):289–96. [11] Krey V. Global energy-climate scenarios and models: a review. Wiley Interdisciplinary Reviews: Energy and Environment 2014;3(4):363–83. [12] Schwanitz VJ. Evaluating integrated assessment models of global climate change. Environmental modelling & software 2013;50:120–31. Tan et al.: Preprint submitted to Elsevier Page 25 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning Figure 9: Inputs and outputs needed for a MAED-OSeMOSYS integration, focusing on the services sector. Figure 10: A 20% ‘buffer’ above and below the estimated MAED-D ‘technology’ activity mix, using the Total Technology Annual Activity Upper Limit and Total Technology Annual Activity Lower Limit parameters. The useful thermal demand from biomass and coal heating within the services sector is used as an example. The ‘buffer’ is recommended to allow OSeMOSYS flexibility in calibrating a least-cost model while still acknowledging MAED-D results. Tan et al.: Preprint submitted to Elsevier Page 32 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning Figure 11: Inputs and outputs needed for an OnSSET-OSeMOSYS integration. An example ‘buffer’ of 20% is included to define the agriculture end-use demand by technology (stand-alone PV, mini-grid hydro). This ‘buffer’ will allow OSeMOSYS flexibility in calibrating a least-cost model while still acknowledging OnSSET results. Note: *** indicates that further data processing is required, i.e., in translating average, maximum, and minimum capacity factors to a capacity factor in time series. Figure 12: Inputs and outputs needed for an OSeMOSYS-OnSSET integration. Tan et al.: Preprint submitted to Elsevier Page 33 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning Figure 13: Inputs and outputs needed for an OSeMOSYS-FlexTool and a MAED-FlexTool integration. Note: *** denotes that further data processing is required, i.e., in translating capacity factor in time series to an hourly resolution. Tan et al.: Preprint submitted to Elsevier Page 34 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning Figure 14: Inputs and outputs needed for an OSeMOSYS-MINFin and a FINPLAN-MINFin integration. Figure 15: Inputs and outputs needed for a MINFin-OSeMOSYS integration. Note: WACC = weighted average cost of capital Tan et al.: Preprint submitted to Elsevier Page 35 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning Figure 16: Inputs and outputs needed for an OSeMOSYS-FINPLAN and a MINFin-FINPLAN integration. Note: P = principal; I = interest Tan et al.: Preprint submitted to Elsevier Page 36 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning Table 4 Links to IMPACCT hands-on exercises for model-to-model integration. IMPACCT link Hands-on exercises MAED-OSeMOSYS https://zenodo.org/records/10968586 OnSSET-OSeMOSYS https://zenodo.org/records/10968590 OSeMOSYS-OnSSET https://zenodo.org/records/10968591 OSeMOSYS-FlexTool https://zenodo.org/records/10968592 MAED-FlexTool https://zenodo.org/records/14927286 OSeMOSYS-CLEWs https://zenodo.org/records/10598875 OSeMOSYS-SIBs Job generation: https://zenodo.org/records/10955876 Global warming costs: https://zenodo.org/records/10955773 Air pollution costs: https://zenodo.org/records/10955912 OSeMOSYS-PathCalc https://zenodo.org/records/13988059 OSeMOSYS-MINFin https://zenodo.org/records/10949846 OSeMOSYS-FINPLAN https://zenodo.org/records/10968598 MINFIN-FINPLAN https://zenodo.org/records/10968601 FINPLAN-MINFIN https://zenodo.org/records/10968603 Table 5 Links to example RES diagrams for the industrial, transport, household, and services sector, as well as off-grid integration, that one can adopt, adapt, and apply. Sector/System Link Industrial https://zenodo.org/records/10948821 Transport https://zenodo.org/records/10948821 Residential https://zenodo.org/records/10948778 Services https://zenodo.org/records/10948798 Off-grid https://zenodo.org/records/10948836 A.2. IMPACCT Hands-on Exercises and Extra Material To aid the user in utilising IMPACCT, step-by-step hands-on exercises by the authors have been created. Table 4lists the respective links for each IMPACCT tool-to-tool linkage, while Table 5includes links to the industrial, transport, residential, services, and off-grid RES that users can adopt, adapt, and apply for their analyses. Note that the hands-on exercises only provide the unidirectional pipeline between two tools. For feedback loops, users should implement two hands-on exercises accordingly, e.g., the OnSSET-OSeMOSYS and OSeMOSYS-OnSSET hands-on exercises for the OnSSET-OSeMOSYS feedback loop. A.3. Modelling Tools Resources This subsection provides references and links to key resources of each modelling tool (Table 6). A community forum is also available for these tools at forum.u4ria.org. This Energy Modelling Community Discourse Forum is a hub for discussions, troubleshooting, event updates, and sharing the latest publications on open-source energy modelling tools and frameworks. Additionally, the CCG programme hosts the Energy Modelling Platforms (EMPs), a series of capacity-building events. These platforms train individuals in the Global South, equipping them with the skills to collect relevant data, conduct their analyses, and develop credible investment proposals for clean energy infrastructure projects. More information regarding the EMPs can be found at climatecompatiblegrowth.com/ energy-modelling-platform. Tan et al.: Preprint submitted to Elsevier Page 37 of 37 IMPACCT: An Integrated Assessment Model for Policy and Financial Decision-Making in Energy Planning Table 6 Additional information on modelling tools and their resources. Tool Methodology Source code User interface Learning material MAED [20] https://github.com/ Model-for-Analysis-of-Energy-Demand/ MAED-Code https://forms. office.com/e/ cpCks1aVmY https://www.open. edu/openlearncreate/ course/index.php? categoryid=528 OnSSET [22] [23] https:// electrifynow. energydata.info/ https://www.open. edu/openlearncreate/ course/index.php? categoryid=528 OSeMOSYS [24]https://github.com/OSeMOSYS/OSeMOSYS https://forms. office.com/e/ cpCks1aVmY https://www.open. edu/openlearncreate/ course/index.php? categoryid=528 FlexTool [30]https://github.com/irena-flextool/ flextool [31] https://www.open. edu/openlearncreate/ course/index.php? categoryid=528 PathCalc [32] [33] https://pathcalc. github.io/ pathcalc-web/ Coming soon MINFin [34]https://github.com/MINFinModel/ MINFin-Excel https://forms. office.com/e/ cpCks1aVmY https://www.open. edu/openlearncreate/ course/index.php? categoryid=528 FINPLAN [36]https://github.com/FINPLAN-Model/ FINPLAN-Code https://forms. office.com/e/ cpCks1aVmY https://www.open. edu/openlearncreate/ course/index.php? categoryid=528 Tan et al.: Preprint submitted to Elsevier Page 38 of 37