Advanced Local Digital Twins using AI for Early Warning and Preparedness (IA) -

Expected Outcome:

Open AI models that can help predict, respond to, and mitigate impacts before a disaster occurs, enabling proactive decision-making and effective disaster management effectively.

Protection of citizens from natural hazards through proactive measures, preparedness strategies, and urban resilience, planning

Project results are expected to contribute to one of the following outcomes:

• Enhanced protection of citizens from the natural hazard of flooding by facilitating proactive decision-making and effective disaster management through open AI-driven models for urban resilience strategies and planning that can help predict, respond to, and mitigate impacts before a disaster occurs.
• Improved modelling and prediction of urban and riverine flooding by expanding the capabilities of Local Digital Twins with sophisticated AI algorithms and relevant data detailing hydrological and hydraulic processes.
• Strengthened integration of diverse and essential datasets including detailed terrain, land cover, urban features, soil data, and real-time meteorological information (rainfall and temperature, river geometry, and flow) sourced from national hydrometric networks, urban drainage infrastructure, and flood protection assets. This integration aims to enhance flood analysis, simulation, and preparedness particularly in response to climate change and flood scenarios like heavy rainfall impacting nearby river basins.

The projects will leverage high-resolution climatic and meteorological models to assess extreme weather, while also drawing on relevant initiatives such as the Global Flood Awareness System and Destination Earth.

Scope:

In line with the Apply AI Strategy, proposals should develop and implement projects that advance innovative AI algorithms and models from concept to large-scale testing and validation. These solutions will be applied to the creation of Local Digital Twins for flood preparedness, enabling the simulation of flood scenarios, identification of areas at risk, and estimation of potential damage.

• Proposals should focus on the development of innovative AI algorithms that move beyond rigid functions, employing instead a dynamic set of descriptive building features derived from digital models (e.g., geometrical parameters, urban morphology, socio-economic indicators). These algorithms should be integrated with advanced, high-resolution hazard models — including hydrological and hydraulic models — tailored to the specific characteristics of the local area.
• The Local Digital Twins will enable:
  • Flood damage models capable of calculating building-scale impacts, forming the basis for damage hotspot maps.
  • Interactive user interfaces that allow components to be exchanged, modified, and reconfigured to estimate flood damage under various urban planning and risk management scenarios — for example, assessing the feasibility of proposed or existing constructions in flood-prone zones and recommending targeted mitigation strategies.

The scope of this topic includes a strong research and innovation component aimed at the prototyping, testing, and large-scale validation of tailored AI algorithms designed to model multiple disaster types, with a focus on operational deployment in real-world contexts. .It is recommended to prioritise the use of frugal (and local) AI as much as possible. This approach will both reduce greenhouse emissions -an indirect driver of climate-related disasters- and ensure that the tools remain functional in environments with limited connectivity.

Proposals should take into account the expertise of the European Commission's Joint Research Centre (JRC)[1], particularly its experience in developing global systems for disaster and risk management and analyse the potential uptake of the project outcomes by the Copernicus Emergency Management Service. In addition, proposals should align with for the 2025 Mission call on Local Digital Twin for urban planning, ensuring interoperability and complementarity with related European initiatives.

The project results should be modular for reuse in locations outside Europe considering constraints on deployment of AI solutions in low- and middle-income countries. Therefore, the project results shall be open source as much as possible and transferable through open platforms.

Beneficiaries that intend to transfer ownership or grant an exclusive licence must formally notify the granting authority (i.e. DG-CNECT and HaDEA) before the intended transfer or licensing takes place and the granting authority may up to four years after the end of the action object to a transfer of ownership or the exclusive licensing of results.

null

Activities are expected to start at TRL 5 and achieve TRL 7 by the end of the project – see General Annex B.

[1] The JRC expertise on disasters and floods through the Disaster Risk Management Centre https://drmkc.jrc.ec.europa.eu/

[2] https://digital-strategy.ec.europa.eu/en/news/eu-funded-ai-innovation-powers-new-era-cooperative-smart-city-development

News flashes