Framework for Assessing Modelling Assumptions
Year: 2026
Type: report
Abstract: This report outlines the DIRECTED Assumptions Framework – a novel approach for creating a shared understanding and bridging the flows of information between models and users of risk information in efforts to enable integrated decision-making for Disaster Risk Management (DRM) and Climate Change Adaptation (CCA). It explores scientific models simulating floods, droughts and wildfires and compares them to the risk management context of DIRECTED’s Real World Labs (RWL). This work is done to identify critical interoperability gaps (i.e., gaps preventing the flow of information across different levels of governance andcommunication), going beyond ambitions aiming for cross-model compatibility under the Data Fabric. Rather, the Framework guides how to 1) characterise and assess risk models for improved transparency, accessibility, and communication; 2) assess and compare the risk information needs of stakeholders; and 3) bridge interoperability gaps and identify prevailing uncertainties, needs or goals where risk information is required. The Assumptions Framework has been co-developed alongside the application of the Risk-Tandem Framework (D3.2) in the Copenhagen Capital Region (Denmark), Emilia-Romagna (Italy), the Rhine-Erft (Germany), and the Danube Region (Vienna, Austria, and the county of Zala, Hungary). It captures the specific approaches used to make models and their outputs transparent to end users. By providing insights into parameterisation, results, and limits of models and comparing these to user needs, the Framework will support more effective decision-making and enable the development of strategies for managing residual risks and uncertainty regarding issues that cannot always be modelled (for example, cascading impacts and multi-hazard scenarios). DIRECTED will provide an approach to examining assumptions embedded in flood, wildfire, and drought risk modelling through RWL case studies, as well as a guide for comparing them to stakeholders’ goals, needs, and assumptions regarding risk management. However, it can be further developed and operationalised for other hazards and purposes (including climate scenarios). This approach is done by systematically deconstructing and characterising risk modelling processes into their key components and evaluating their underpinning assumptions at each stage. This enables the comparison of models with end-user needs and supports the exploration of differing interpretations, understandings, and assumptions that stakeholders and modellers have (which may not align with modelling capabilities or context-specific needs). This work will be complemented by an exploration of relevant terms and key concepts used both by models and users of risk information, contributing to the identification of core definitions, the development of context-specific ‘scope notes’ and a resulting taxonomy, implemented as an interactive glossary in the Data Fabric. The key contributions of this framework are:Addressing the issues of data interoperability beyond modelling. Although it contributes to improved model interoperability, accessibility, and accuracy, DIRECTED’s Data Fabric does not provide an approach to identify how stakeholders interpret and use risk information or whether they understand the limits of the modelling process. For example, the project partners’ models (including RIM2D, SaferPlaces, and DamageCost) do not model the dynamics of socio-economic vulnerability or assess flood impacts beyond economic costs. Similarly, they cannot model cascading impacts or process multi-hazard scenarios. These must be communicated to and aligned with stakeholders’ expectations and needs in efforts to encourage the development of risk management strategies that account for information gaps. Similarly, usability and accessibility of model outputs should be addressed, alongside issues of parameterisation, uncertainty, misinterpretation, and translation. Identifying interoperability issues relating to risk governance and the use of models. Risk information does not automatically strengthen risk governance. For example, RWL stakeholders have expressed the need to increase engagement between municipalities and recognise limited knowledge sharing between actors across levels. Data sharing between public and private actors is limited in the case of Danube and Italy, and the issues of licensing continue to affect the sustained use of modelling products. In the context ofassumptions about the value of information for decision-making, it is necessary to explore and identify how the governance context and actors involved shape how data is or can be used. Identifying issues of interoperability relating to communication and models. Although providing a centralised platform for disseminating, visualising and analysing risk information, the Data Fabric cannot directly address issues identified around communication between stakeholders nor explore how information is interpreted and communicated further. For instance, the lack of inter-municipal coordination mechanisms in Denmark or the lack of public/private sector collaboration in the Danube region will affect the use and accessibility of risk information beyond products.
Authors: Parviainen, Janne, Bharwani, Sukaina, Ng, Natascha, Bagli, Stefano, Mazzoli, Paolo, Pancioli, Valeria, Hedderich, Dominik, Huszti, Levente, Löhrlein, Jana, Cumiskey, Lydia, Kropf, Chahan, Juhel, Samuel, Pontius, Martin, Kraatz, Julia, Hofbauer, Ben, Schweizer, Pia-Johanna, Yokomatsu, Muneta, Hochrainer-Stigler, Stefan, Gräler, Benedikt, Drews, Martin, Apel, Heiko, Steinhausen, Max Jacob, Irvine, Tracy
Institutions: Stockholm Environment Institute, University College Cork, ETH Zurich, 52°North Spatial Information Research, Research Institute for Sustainability at GFZ, GFZ Helmholtz Centre for Geosciences, International Institute for Applied Systems Analysis, Technical University of Denmark, Technische Universität Braunschweig
Language: English
Subfield: Global and Planetary Change
Field: Environmental Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7225784366
Cites:
Cited by:
Related to: