DIAMOND: D6.8 – IAM Application Library

www.climate-diamond.eu 28/11/2025 D6.8 - IAM Application Library WP6 – Open Page i D6.8 – IAM Application Library Disclaimer Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Climate, Infrastructure and Environment Executive Agency (CINEA). Neither the European Union nor the granting authority can be held responsible for them. Copyright Message This report, if not confidential, is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0); a copy is available here: https://creativecommons.org/licenses/by/4.0/. You are free to share (copy and redistribute the material in any medium or format) and adapt (remix, transform, and build upon the material for any purpose, even commercially) under the following terms: (i) attribution (you must give appropriate credit, provide a link to the license, and indicate if changes were made; you may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use); (ii) no additional restrictions (you may not apply legal terms or technological measures that legally restrict others from doing anything the license permits). Grant Agreement Number 101081179 Acronym DIAMOND Full Title Delivering the next generation of open Integrated Assessment MΟdels for Netzero, sustainable Development Topic HORIZON-CL5-2022-D1-02 Funding scheme HORIZON EUROPE, RIA – Research and Innovation Action Start Date December 2022 Duration 48 Months Project URL http://www.climate-diamond.eu/ EU Project Advisor Silvia Vaghi Project Coordinator Institute of Communications and Computer Systems - ICCS Deliverable D6.8 – IAM Application Library Work Package WP6 – Open Date of Delivery Contractual 30/11/2025 Actual 28/11/2025 Nature DEC — Websites, patent filings, videos, etc Dissemination Public Lead Beneficiary HOLISTIC IKE (HOLISTIC) Responsible Authors Georgios Xexakis (HOLISTIC) Email [email protected] Georgios Tassis (HOLISTIC) Email [email protected] Contributors Stefanos Tsotras, Angelos Potiriadis (HOLISTIC) Reviewer(s) Constantinos Taliotis (CYI); Anastasios Karamaneas, Konstantinos Koasidis, Alexandros Nikas (ICCS) Keywords integrated assessment models; IAM PARIS platform; application library; stakeholders; generative artificial intelligence; usability; policymakers. Page ii D6.8 – IAM Application Library EC Summary Requirements 1. Changes with respect to the DoA The I2AM PARIS web platform has been revamped and rebranded, and is now referred to, as IAM PARIS web platform. No other changes with respect to the work described in the DoA. 2. Dissemination and uptake The Application Library is primarily intended to assist different stakeholders and user groups working on climate action to find useful applications that can support their work based on the results of integrated assessment modelling and other relevant tools. In addition, the library includes databases and resources that can help modellers with model development, scenario analyses, and other use cases. The library will be extensively disseminated to the modelling community (e.g., through the newsletter of the Integrated Assessment Modelling Consortium) as well as through communication channels and events related to different user groups, e.g., the EU Sustainable Energy Week. This report will be especially helpful to researchers and advanced users of the library to understand the motivation and design choices behind the development of the first version of the library. 3. Short summary of results This deliverable presents the initial version of the Application Library developed within the IAM PARIS platform, along with the first application produced under the DIAMOND project. The library intends to constitute a collection of software applications covering the full lifecycle of climate-economy and energy modelling, from model development and configuration to the communication and use of modelling results. Apart from applications developed or curated by the IAM PARIS team, this first release incorporates applications previously catalogued by the Integrated Assessment Modelling Consortium (IAMC) and the World Resources Institute (WRI), aiming to establish a comprehensive portal for applications that support climate action. More than 150 applications are currently featured in the Application Library. Both platform developers and external contributors can extend the library’s content with ease using the platform’s Content Management System (CMS), helping to keep the library up to date. To further facilitate the usability of the platform, an AI-based chatbot was developed within the DIAMOND project to support users navigate the model documentation and results available on the platform. In the final year of DIAMOND, a second version of the deliverable will be released (D6.9), building on the current functionalities of the library, to further strengthen its comprehensiveness, findability, and usability. 4. Evidence of accomplishment The deployed application library in IAM PARIS (https://iamparis.eu/application_library), the GitHub repository of the first DIAMOND application on GitHub (https://github.com/i2amparis/iam-data-chatbot), and a GitHub repository documenting the data cleaning and processing activities for assembling the first application library (https://github.com/i2amparis/application-library-data). Additionally, this report aims to act as a descriptor of these resources. Page iii D6.8 – IAM Application Library Preface DIAMOND will update, upgrade, and fully open six Integrated Assessment Models (IAMs) that are emblematic in scientific and policy processes, improving their sectoral and technological detail, spatiotemporal resolution, and geographic granularity. It will further enhance modelling capacity to assess the feasibility and desirability of Pariscompliant mitigation pathways, their interplay with adaptation, circular economy, and other SDGs, their distributional and equity effects, and their resilience to extremes, as well as robust risk management and investment strategies. This will be done via integration of tools and insights from psychology, finance research, behavioural and labour economics, operational research, and physical science. The project will develop a transdisciplinary scientific approach to legitimise the implementation process and co-create research questions that stretch the frontiers of climate science, as well as establish vibrant communities of practice to transparently open model enhancements and to develop capacities, thereby lowering the entrance barriers to the established IAM community. ICCS INSTITUTE OF COMMUNICATIONS AND COMPUTER SYSTEMS EL BC3 ASOCIACION BC3 BASQUE CENTRE FOR CLIMATE CHANGE - KLIMA ALDAKETA IKERGAI ES CESAR KRATENA KURT AT CICERO CICERO SENTER FOR KLIMAFORSKNING NO CYI THE CYPRUS INSTITUTE CY E4SMA ENERGY ENGINEERING ECONOMIC ENVIRONMENT SYSTEMS MODELING AND ANALYSIS SRL IT HOLISTIC HOLISTIC IKE EL COMILLAS UNIVERSIDAD PONTIFICIA COMILLAS ES ISINNOVA ISTITUTO DI STUDI PER L'INTEGRAZIONE DEI SISTEMI (I.S.I.S) - SOCIETA'COOPERATIVA IT SEURECO SEURECO SOCIETE EUROPEENNE D'ECONOMIE SARL FR UM UNIVERSITEIT MAASTRICHT NL ESMIA ESMIA CONSULTANTS INC. CA USMF THE UNIVERSITY OF MARYLAND FOUNDATIION INC US UMD UNIVERSITY OF MARYLAND US EPFL ECOLE POLYTECHNIQUE FEDERALE DE LAUSANNE CH ETH EIDGENOESSISCHE TECHNISCHE HOCHSCHULE ZUERICH CH UNIBAS UNIVERSITAT BASEL CH Imperial IMPERIAL COLLEGE OF SCIENCE TECHNOLOGY AND MEDICINE UK Oxford THE CHANCELLOR, MASTERS AND SCHOLARS OF THE UNIVERSITY OF OXFORD UK UCL UNIVERSITY COLLEGE LONDON UK Page iv D6.8 – IAM Application Library Executive Summary This deliverable describes the first version of the Application Library developed in the IAM PARIS platform as well as the first application developed by the DIAMOND project. The library is a catalogue of platforms, visualisations, interfaces, and utilities that can support the entire lifecycle of modelling for climate action, from model development and configuration to visualisation and use of the results. The first version of the library provides access to applications that were developed or collected by the IAM PARIS team. The library also includes links and descriptions to applications that were previously assembled by the Integrated Assessment Modelling Consortium (IAMC) and the World Resources Institute (WRI). Currently, more than 150 applications are featured in the Application Library. All applications are accessible through a user-friendly interface on the platform, allowing visitors to explore applications that feature specific topics, scales, and geographies and find the ones that best match their needs and abilities. Platform developers and external users can easily expand the application list through the Content Management System (CMS) of the IAM PARIS platform. This key functionality allows the library to stay up to date, increasing its relevance for the users and, subsequently, contributing to the longevity of the library and the IAM PARIS platform in general. Towards this goal, an AI-based chatbot was also developed as part of the DIAMOND project, to help users navigate through model documentations and results that are included in the platform. The IAM PARIS Chatbot is a multiagent conversational framework designed to facilitate user-friendly access to complex modelling data from the IAM PARIS platform. The system integrates natural language processing, semantic search, and data visualisation, enabling stakeholders to interactively explore details and results of integrated assessment and other models in the platform. This would be especially important for policymakers and other decision makers that have a specific use case in mind to support climate action but have limited knowledge of modelling nuances. By combining structured data retrieval with semantic reasoning and generative explanation, the Chatbot aims to democratise access to climate knowledge, accelerate scenario exploration, and enhance evidence-based decision-making in climate policy contexts. While the library will be continuously updated in the coming years, the next version of this deliverable will be released towards the end of DIAMOND (Deliverable D6.9), documenting all new functionalities and content. More applications will be added while the library will be enriched with links to databases that are frequently used in modelling as well as utilities that support modelling applications in different stages. More filters and detailed metadata will be also added to help match more user groups such as local policymakers and financial actors with specific applications. On this, the AI-based chatbot will be further improved to help platform visitors identify the most fitting applications by providing their exact use case as a text prompt. The CMS and front end of the IAM PARIS platform will also be enhanced to allow users to easily develop applications that can be directly hosted in the platform. Page v D6.8 – IAM Application Library Contents 1 Introduction .................................................................................................................................................... 1 2 Application library.......................................................................................................................................... 3 2.1 Definition of applications ............................................................................................................................................................ 3 2.2 Sources of applications ................................................................................................................................................................ 3 2.2.1 Collection of project-based applications..................................................................................................................... 4 2.2.2 Collection of IAMC-based applications ........................................................................................................................ 5 2.2.3 Collection of WRI-based applications .......................................................................................................................... 6 2.3 Current implementation .............................................................................................................................................................. 6 3 1st DIAMOND app: IAM bot ......................................................................................................................... 10 3.1 Context ............................................................................................................................................................................................. 10 3.2 Technical implementation ........................................................................................................................................................ 10 4 Next steps ..................................................................................................................................................... 13 References ............................................................................................................................................................ 14 Table of Figures Figure 1. Screenshot from a data story on technologies for climate-neutral industries hosted by IAM PARIS ......... 4 Figure 2. Screenshot of tools and visualisations page in the IAMC website ............................................................................. 5 Figure 3. Screenshot of a collection of 100+ climate data platforms by WRI .......................................................................... 6 Figure 4. Screenshot from the first version of the Application Library in IAM PARIS ........................................................... 8 Figure 5. Screenshot from a dedicated page for an application in the Library ....................................................................... 8 Figure 6. Screenshot from the CMS of the Application Library in IAM PARIS .......................................................................... 9 Figure 7. Graph of the IAM PARIS Chatbot workflow ...................................................................................................................... 11 Figure 8. Graph of the IAM PARIS Chatbot workflow ...................................................................................................................... 11 Figure 9. IAM PARIS Chatbot repository structure ............................................................................................................................ 12 Table of Tables Table 1. Multi-agent architecture of the IAM PARIS Chatbot ....................................................................................................... 12 Table 2. Technology stack used in the development of the IAM PARIS Chatbot ................................................................. 12 Page 1 D6.8 – IAM Application Library 1 Introduction Integrated Assessment Models (IAMs) have been used for over 30 years to underpin climate change mitigation research and inform climate action (Fisher-Vanden & Weyant, 2020). IAMs are often used in the context of scenario analysis, where different policy choices, technological advancements, and other parameters are modelled in IAMs to identify impacts and nuances of each potential future pathway (Elsawah et al., 2020). Gaining prominence during the last decade and especially after the Paris Agreement, hundreds of models have been developed leading to thousands of scenarios (van de Ven et al., 2025) as well as extensive model intercomparison projects where dozens of models are running the same scenarios (Fujimori et al., 2025). The results from all these modelling studies are usually analysed and published in scientific studies or are directly used in policy reports, for instance, the latest Assessment Reports (AR) of the Intergovernmental Panel for Climate Change (IPCC) and the reports of the European Scientific Advisory Board on Climate Change (ESABCC). As these publications often have limited space to present the wealth of modelling results and explain in detail how they came about, interactive web tools and other applications have been increasingly developed to help different audiences explore all available results and exploit them in different use cases, e.g., for informing policies or supporting decisions (Curley et al., 2025). Some of the most famous applications are the Scenario Explorer from IIASA and the respective scenario databases that accompany the latest IPCC reports (Cointe, 2024). Other platforms have been explicitly co-created with stakeholders to help them on explicit use cases, such as the platform of the SENSES project that developed interfaces for policymakers and stakeholders from finance to explore their selected indicators from modelling studies (Auer et al., 2021). In addition, there are several other applications— apart from those dedicated to integrated assessment modelling—that can potentially support climate action, such as scenario building tools based on the results of energy system models, decision support tools based on environmental modelling results, and tools for exploring climate impacts based on detailed global climate models (Xexakis & Trutnevyte, 2019). Conversely, apart from applications for facilitating the use of modelling results, there are several tools for supporting modellers during their work, including utilities for visualising and validating model results such as the pyam Python package (Gidden & Huppmann, 2019). While these applications can be very useful for different stakeholders from policymaking, research, industry, or civil society, they may not be readily accessible. Most of the applications are hosted in the websites of the organisations or projects that created them and it is thus difficult for a user to find them if they are not aware of their existence in the first place. While there are some efforts to collect these applications, e.g. from the Integrated Assessment Modelling Consortium (IAMC)1, the list is far from complete and difficult to continuously update, with more tools being recently developed, such as the Climate Solutions Explorer2 or the Scenario Compass Initiative platform3. Even if such lists were kept up to date, the ever-increasing number of applications may discourage users from navigating through them to try to find what they need through trial and error and, instead, require more elaborate methods for improving findability. This is further complicated by the potential lack of support for some applications after their development (Curley et al., 2025). Even if an application fits the needs of a user, it may not function as designed due to lack of maintenance or relying on outdated data. And, even if such applications work and the results are still relevant, there may be an issue of transparency. Such applications focus mostly on the results of a modelling exercise or a research project, and it might be difficult for the users to understand how these results were produced (e.g., what were the modelling assumptions), which is also a general discussion point in IAM research (Robertson, 2021). With the rise of AI applications, there are many opportunities to increase 1 https://www.iamconsortium.org/resources/ 2 https://www.climate-solutions-explorer.eu 3 https://scenariocompass.org Page 2 D6.8 – IAM Application Library engagement with climate information in a meaningful way (Bago et al., 2025), for instance by communicating modelling results that inform climate action. In response to these issues, Task 6.5 of DIAMOND aims to create an open and user-tailored library of applications related to modelling in support of climate action. The library is created in IAM PARIS, an open data sharing platform for documenting modelling information and scenario results that was developed from the Horizon 2020 PARIS REINFORCE project and further enhanced by more than eight other EU-funded projects. The Application Library is featuring a comprehensive list of real-world applications for different use cases, such as policymaking (e.g., setting climate goals) or financial analysis (e.g., analysing climate risks in different assets and supply chains). Applications are sourced from the available scientific literature as well as from reports and websites of major institutions, modelling consortia, and projects and will be further enriched with new applications as they become available. In addition, DIAMOND intends to create dedicated applications for each of the major stakeholder categories of the project: policymakers, industry representatives, civil society. Each application will be in the form of an online decision support tool and WP2 stakeholders will inform its design and evaluate its use. This deliverable report presents the first version of the Application Library and the IAM PARIS Chatbot, the first application that has been developed in the context of DIAMOND to facilitate the interaction of policymakers with modelling results. The first version of the Application Library is presented as an interactive interface in the platform featuring more than 150 applications existing in the climate action domain and allowing users to explore the applications using different filters4. The Chatbot is provided as a Python application that will be hosted in the future directly in the IAM PARIS platform5. Section 2 presents the working definition of applications that is used in this deliverable, documents the sources of applications, and describes the current implementation of the Application Library including the filters that are used and the possibility to add more information. Subsequently, Section 3 presents the chatbot that was developed in DIAMOND to access the model information and the results shown in the platform. Finally, Section 4 summarises the contributions of this deliverable and provides indications for future work, including in terms of updates of the application library and the development of the other applications expected in DIAMOND. 4 https://iamparis.eu/datastories 5 https://github.com/i2amparis/iam-data-chatbot Page 3 D6.8 – IAM Application Library 2 Application library 2.1 Definition of applications There is a variety of names to describe applications based on modelling results in the literature, from decision support tools and policy platforms to data platforms and interactive web tools (Curley et al., 2025). This diversity allows developers to signpost the use of their application by characterising it in a particular way, e.g., a decision support tool for an application that is explicitly built to help decision makers in a specific use case. Yet, this may also hinder the identification and dissemination of the application through online search engines (with or without AI support), as it is potentially difficult to match the prompt of the user with the description that the developers gave to their application. This issue has been also discussed in the literature with the development of an intermediary “climate change scenario service” that can guide users through the scenario results that are developed (Auer et al., 2021) and, recently, with the suggestion to develop a centralised platform to systematise model intercomparison exercises and make their results more transparent (Fujimori et al., 2025). However, these suggested improvements have not yet been developed at an adequate scale to influence the community of model developers and users at large. For Task 6.5, we use the term “application” for all items in our library, following the Cambridge Dictionary definition of “a computer program that is designed for a particular purpose.”6. Although this may be a broad definition, it is useful to encapsulate all applications that relate to the modelling workflow, not only the applications that visualise modelling results. This includes: 1) applications that can provide model inputs, such as the online databases of IEA7, 2) utilities to facilitate modelling and analysis, such as the pyam package8 or the validation interface of IAM COMPACT9, and 3) platforms that present and enable the exploitation of modelling results, such as the Scenario Explorer for the IPCC AR610, the interactive visualisations of the SENSES Toolkit11, and the online EUCalc calculation model12. Additionally, many of these applications are not only informed by integrated assessment models but also other models used to support climate action such as Earth System Models, Energy System Models, macroeconomic models, or sectoral impact models. In short, the library presented in this deliverable includes both applications informed by IAMs and other climate-relevant models, and applications that support different stages of the modelling process. In the future, it may be useful to extend the library to go beyond modelling and include resources to support climate action in general. Nevertheless, this may lie beyond the context of the DIAMOND project as its focus is on supporting modelling processes and the usefulness of their results. 2.2 Sources of applications IAM PARIS already hosted several applications developed by EU-funded modelling projects that served as the initial seed for the application library. In addition, the authors have identified two other catalogues of such applications: (a) the aforementioned catalogue of tools, visualisations, and scenario databases that was assembled by the IAMC and (b) an overview of more than 100 climate data platforms put together by the World Resources Institute (WRI). Links and descriptions for all these applications are integrated in the IAM PARIS Library to provide a comprehensive “one stop shop” of applications related to modelling in support of climate action. Thus, on top 6 https://dictionary.cambridge.org/dictionary/english/application 7 https://www.iea.org/data-and-statistics 8 https://pyam-iamc.readthedocs.io/en/stable/ 9 https://validation.iam-compact.eu 10 https://data.ece.iiasa.ac.at/ar6/ 11 https://climatescenarios.org/toolkit/ 12 https://www.european-calculator.eu Page 10 D6.8 – IAM Application Library 3 1st DIAMOND app: IAM bot 3.1 Context Integrated Assessment Models and the scenario modelling techniques that are used to inform climate action are usually quite complex (Guivarch et al., 2017). Despite efforts to provide extensive documentation for models and open up their code (Skea et al., 2021), modelling details are often incomprehensible for others apart from the modelling teams that use them, especially non-modellers (Nikas et al., 2021). Additionally, modelling studies are usually providing a wealth of outputs that can be difficult to navigate. For instance, only the global scenario results that were included in the AR6 scenario database contained almost 700,000 entries (Byers et al., 2022). While there are many scenario databases that provide interactive functionalities to navigate through scenario results, users usually need to know the name of the variable that they are looking for and the nuances behind the generation of results in the models, e.g., what are the differences between scenarios and models. Tools using artificial intelligence have been recently pinpointed as a way to support climate change scientific assessments, including the models that underpin them (Al Khourdajie, 2025a, 2025b). The development of a structured CMS and database of modelling results and documentation in IAM PARIS served as an opportunity for building an assistant to help users efficiently navigate through the documentation and results hosted in the platform. The IAM PARIS Chatbot is a multi-agent conversational framework designed to facilitate user-friendly access to complex modelling data from the IAM PARIS platform. The system integrates natural language processing, semantic search, and data visualisation to enable stakeholders to interactively explore the details and results of integrated assessment models and other models in the platform. The chatbot was designed focusing on policymakers working on climate topics as a target audience, but can be used by other stakeholders too, including researchers, civil society, etc. By combining structured data retrieval with semantic reasoning and generative explanation, the Chatbot aims to democratise access to climate knowledge, accelerate scenario exploration, and enhance evidence-based decision-making in climate policy contexts. 3.2 Technical implementation The code of the Chatbot is publicly available on GitHub28. The Chatbot can be currently run in a command-line interface (CLI) through the code directly from the code repository, while a beta version29 is also integrated into the IAM PARIS platform (Figure 7). Figure 8 shows the project workflow for the Chatbot, while Figure 9 shows the structure of the code repository. The workflow for mainv.py begins with environment setup, loading required API keys and URLs from a .env file, followed by fetching models and timeseries data that are included in the IAM PARIS platform. Data fetching was achieved using the IAM PARIS REST API, with caching enabled to avoid redundant requests. Definitions for regions and variables are loaded from YAML files in the /definitions directory, converted into documents, and combined with API records before being split into chunks using RecursiveCharacterTextSplitter. OpenAI embeddings are generated for these chunks, and a FAISS vector index is built or loaded from disk for efficient similarity searches. Shared resources including the models, timeseries, vector store, environment variables and bot instance are prepared, before a MultiAgent Manager is initialised to handle query routing. In CLI mode, the script enters an interactive loop where user queries are processed through the manager, with responses displayed or plots rendered from base64-encoded images; in API mode via Fast API, POST requests to /query trigger the same process, returning structured responses with chat history while handling 28 https://github.com/i2amparis/iam-data-chatbot 29 The current beta version is under continuous refinement to expand the types of queries the tool can accurately respond to. Responses provided by the tool should be treated with caution. Page 11 D6.8 – IAM Application Library errors gracefully. Figure 7. Graph of the IAM PARIS Chatbot workflow Figure 8. Graph of the IAM PARIS Chatbot workflow Page 12 D6.8 – IAM Application Library Figure 9. IAM PARIS Chatbot repository structure The system uses a multi-agent architecture, where specialised agents collaborate to process user requests and generate coherent responses. Each agent encapsulates a well-defined functionality, such as a query agent that is executing structured queries and retrieving IAM PARIS data via REST APIs. The application is developed in Python based on the OpenAI GPT-4-Turbo model and the Facebook AI Similarity Search (FAISS) while its API is built using FastAPI. Details for all agents can be seen in Table 1, while the full technology stack is shown in Table 2. Table 1. Multi-agent architecture of the IAM PARIS Chatbot Agent Function Query Agent Executes structured queries and retrieves IAM data via REST APIs Data Plotting Agent Produces time-series visualisations and scenario comparisons Model Explanation Agent Performs semantic retrieval and explanation generation using FAISS and language models (LLMs) General QA Agent Handles general questions on IAM methodology and climate police Data Modelling Suggestions Agent Generates research and analysis recommendations based on retrieved context Table 2. Technology stack used in the development of the IAM PARIS Chatbot Category Technology Purpose Language Python Core implementation Data Handling Pandas, JSON, Requests Fetching API data Visualization Matplotlib Plot generation Vector Database FAISS (Facebook ai similarity search) Semantic retrieval NLP/AI Lang chain, OpenAI GPT-4-Turbo Query understanding and response generation Embedding model Text embedding 3 small (OPENAI) Document vectorization Memory LangChain (conversation buffer memory) Multi-turn conversational context API FastAPI Web API for HTTP interactions Page 13 D6.8 – IAM Application Library 4 Next steps This deliverable describes the first version of the Application Library developed in the IAM PARIS platform, as well as the first application developed by the DIAMOND project. The library is a catalogue of platforms, visualisations, interfaces, and utilities that can support the whole lifecycle of modelling for climate action, from model development and configuration to visualisation and use of results. Aiming for comprehensiveness, the first version of the library includes links to applications that were previously catalogued by the Integrated Assessment Modelling Consortium and the World Resources Institute, in addition to the applications that were developed or collected by the IAM PARIS team. The library is currently featuring several filters that allow visitors to inter alia identify applications for specific topics, user groups, scales, and geographies, improving findability and enhancing the re-usability potential of the existing platforms. Additionally, the library is linked to the Content Management System of the IAM PARIS platform, allowing for platform developers and external users to easily add links and information to new applications. This key functionality allows the library to stay up to date, increasing its relevance for the users and, subsequently, contributing to the longevity of the library and the IAM PARIS platform in general. Towards this goal, an AI-based chatbot was also developed as part of the DIAMOND project, to help users navigate through the model documentations and results that are included in the platform. This would be especially important for policymakers and other stakeholders that have a specific use case in mind to support climate action but have limited knowledge in modelling and its nuances. During the last year of the DIAMOND project, the next version of the deliverable will be released (D6.9), documenting new features, content, and usability improvements for the library. Specifically, more applications will be added based on a structured review of the scientific and grey literature and by scanning the websites of key organisations related to climate action, such as the International Energy Agency (IEA), the Joint Research Centre (JRC) of the European Commission, and the World Wide Fund for Nature (WWF). Additionally, the library will be enriched with links to databases that are frequently used in modelling (i.e., GTAP30) as well as utilities that support modelling applications in different stages such as on validation and results visualisation. More filters and detailed metadata will also be added to help match user groups needs such as local policymakers and financial actors with specific requirements in terms of applications used. On this, the AI-based chatbot will be further improved to help platform visitors identify the most fitting applications by providing their exact use case as a text prompt. The CMS and front end of the IAM PARIS platform will also be enhanced to allow users to easily develop applications that can directly be hosted in the platform. This feature aims to help modellers with promoting their modelling results in a way that is usable by all intended users and to avoid the development of unnecessary platforms that may end up being abandoned after a modelling study or project ends. This is especially important for projects funded through the Horizon Europe programme, as there is already a tendency to leave platforms without maintenance and support after the project’s end (Curley et al., 2025), undermining—in essence—the investment spent for developing the platform in the first place. Finally, two new applications will be developed using DIAMOND results, aiming specifically for industry representatives and civil society groups. For industry representatives, an initial idea would be to develop a tool to stress test assumptions for Environmental, Social, and Governance (ESG) reporting using IAM scenarios. For civil society groups, a similar application can help contextualise and refine claims related to the feasibility of climate action using results from large scenario ensembles. The scope and implementation of the applications will be defined among the DIAMOND consortium and informed (and potentially co-created) based on the stakeholder engagement activities as part of WP2. 30 https://www.gtap.agecon.purdue.edu Page 14 D6.8 – IAM Application Library References Al Khourdajie, A. (2025a). Exploring climate futures with deep learning. Nature Climate Change, 1–2. https://doi.org/10.1038/s41558-025-02350-w Al Khourdajie, A. (2025b). The role of artificial intelligence in climate change scientific assessments. PLOS Climate, 4(9), e0000706. https://doi.org/10.1371/journal.pclm.0000706 Auer, C., Kriegler, E., Carlsen, H., Kok, K., Pedde, S., Krey, V., & Müller, B. (2021). Climate change scenario services: From science to facilitating action. One Earth, 4(8), 1074–1082. https://doi.org/10.1016/j.oneear.2021.07.015 Bago, B., Muller, P., & Bonnefon, J.-F. (2025). Using generative AI to increase sceptics’ engagement with climate science. Nature Climate Change, 15(11), 1176–1182. https://doi.org/10.1038/s41558-025-02424-9 Byers, E., Krey, V., Kriegler, E., Riahi, K., Schaeffer, R., Kikstra, J., Lamboll, R., Nicholls, Z., Sandstad, M., Smith, C., van der Wijst, K., Al Khourdajie, A., Lecocq, F., Portugal-Pereira, J., Saheb, Y., Stromann, A., Winkler, H., Auer, C., Brutschin, E., … van Vuuren, D. (2022). AR6 Scenarios Database [Dataset]. Zenodo. https://doi.org/10.5281/zenodo.5886911 Cointe, B. (2024). The AR6 Scenario Explorer and the history of IPCC Scenarios Databases: Evolutions and challenges for transparency, pluralism and policy-relevance. Npj Climate Action, 3(1), 1–10. https://doi.org/10.1038/s44168-023-00075-0 Curley, A., Xexakis, G., Zuiderwijk, A., Minkman, E., & Okur, Ö. (2025). Policy support platforms on climate change mitigation and adaptation: An assessment framework. Energy Strategy Reviews, 60, 101812. https://doi.org/10.1016/j.esr.2025.101812 Elsawah, S., Hamilton, S. H., Jakeman, A. J., Rothman, D., Schweizer, V., Trutnevyte, E., Carlsen, H., Drakes, C., Frame, B., Fu, B., Guivarch, C., Haasnoot, M., Kemp-Benedict, E., Kok, K., Kosow, H., Ryan, M., & van Delden, H. (2020). Scenario processes for socio-environmental systems analysis of futures: A review of recent efforts and a salient research agenda for supporting decision making. Science of the Total Environment, 729, 138393. https://doi.org/10.1016/j.scitotenv.2020.138393 Fisher-Vanden, K., & Weyant, J. (2020). The Evolution of Integrated Assessment: Developing the Next Generation of Use-Inspired Integrated Assessment Tools. Annual Review of Resource Economics, 12(Volume 12, 2020), 471–487. https://doi.org/10.1146/annurev-resource-110119-030314 Fujimori, S., Krey, V., Riahi, K., Sugiyama, M., Hasegawa, T., Edmonds, J., Guivarch, C., Paltsev, S., Rose, S., Schaeffer, R., Tavoni, M., Vishwanathan, S. S., Van Vuuren, D., & Weitzel, M. (2025). Towards an open model intercomparison platform for integrated assessment models scenarios. Nature Climate Change. https://doi.org/10.1038/s41558-025-02462-3 Gidden, M., & Huppmann, D. (2019). pyam: A Python Package for the Analysis and Visualization of Models of the Interaction of Climate, Human, and Environmental Systems. Journal of Open Source Software, 4(33), 1095. Page 15 D6.8 – IAM Application Library https://doi.org/10.21105/joss.01095 Guivarch, C., Lempert, R., & Trutnevyte, E. (2017). Scenario techniques for energy and environmental research: An overview of recent developments to broaden the capacity to deal with complexity and uncertainty. Environmental Modelling & Software, 97, 201–210. https://doi.org/10.1016/j.envsoft.2017.07.017 Nikas, A., Gambhir, A., Trutnevyte, E., Koasidis, K., Lund, H., Thellufsen, J. Z., Mayer, D., Zachmann, G., Miguel, L. J., Ferreras-Alonso, N., Sognnaes, I., Peters, G. P., Colombo, E., Howells, M., Hawkes, A., van den Broek, M., Van de Ven, D. J., Gonzalez-Eguino, M., Flamos, A., & Doukas, H. (2021). Perspective of comprehensive and comprehensible multi-model energy and climate science in Europe. Energy, 215, 119153. https://doi.org/10.1016/j.energy.2020.119153 Robertson, S. (2021). Transparency, trust, and integrated assessment models: An ethical consideration for the Intergovernmental Panel on Climate Change. WIREs Climate Change, 12(1), e679. https://doi.org/10.1002/wcc.679 Skea, J., Shukla, P., Al Khourdajie, A., & McCollum, D. (2021). Intergovernmental Panel on Climate Change: Transparency and integrated assessment modeling. WIREs Climate Change, 12(5), e727. https://doi.org/10.1002/wcc.727 van de Ven, D. J., Mittal, S., Nikas, A., Xexakis, G., Gambhir, A., Hermwille, L., Fragkos, P., Obergassel, W., GonzalezEguino, M., Filippidou, F., Sognnaes, I., Clarke, L., & Peters, G. P. (2025). Energy and socioeconomic system transformation through a decade of IPCC-assessed scenarios. Nature Climate Change, 1–9. https://doi.org/10.1038/s41558-024-02198-6 Xexakis, G., & Trutnevyte, E. (2019). Are interactive web-tools for environmental scenario visualization worth the effort? An experimental study on the Swiss electricity supply scenarios 2035. Environmental Modelling & Software, 119(May), 124–134. https://doi.org/10.1016/j.envsoft.2019.05.014