Project Description

Abstract

• This activity demonstrated the applicability and benefits of immersive visualization, simulation, and Artificial Intelligence technologies within Earth Observation (EO) use cases.
• Led by Solenix Engineering GmbH, in collaboration with MindEarth Srl, BIT S.p.A., and AmbiensVR Srl, the project developed Visu4EO, an EO-driven Digital Twin platform for urban flood risk and mobility analysis. The solution integrates EO data, hydrological modeling, traffic simulation, Virtual Reality (VR), and Explainable Artificial Intelligence (XAI) into a single immersive decision-support environment.
• The platform enables users to explore, compare, and assess flood scenarios and urban planning decisions across multiple spatial scales. It was implemented and demonstrated through a real-world use case focused on the city of Copenhagen, including three Areas of Interest selected to showcase the platform’s 3D capabilities and usability.
• Visu4EO supports interactive “what-if” analyses, allowing users to modify rainfall conditions, sea and river discharges, land-use policies, and flood defense measures, and immediately visualize the resulting impacts. Explainable AI components provide transparent insights into flood drivers, scenario effectiveness, and defense performance at both city-wide and street levels.
• The project demonstrated the value of immersive technologies for enhancing policymaking, stakeholder engagement, and the interpretation of complex environmental and mobility impacts.

Detailed description

Background

This activity developed an immersive visualization application associated with an Earth Observation (EO) use case. The implementation integrates EO and mobility data with advanced flood and mobility modeling, leveraging key Virtual Reality (VR) and Artificial Intelligence (AI) technologies. The project demonstrates how Digital Twins can support policy and decision-makers in understanding the state of relevant environmental parameters, simulating changes and phenomena, and assessing their impacts through the modeling and realistic rendering of what-if scenarios, combined with advanced analysis enabled by Explainable AI (XAI).

The proposed use case focuses on an urban mobility scenario for the city of Copenhagen, based on a mobility model covering the entire city and its surrounding municipalities. The tool enables users to understand the impacts of flood events on the wider transportation and mobility network and to relate these impacts to different urban development scenarios. The immersive experience allows users to engage with a dynamic simulation of Copenhagen during a hydrogeological emergency, supporting informed exploration and analysis of flood-related disruptions.

The Visu4EO platform, the outcome of this activity, provides an immersive decision-support environment that enables users to simulate flood scenarios and planning decisions and to assess their impacts on urban mobility. Through a VR-based experience, users can explore 3D representations of flooding conditions, dynamically switch between global and local views, and interact with what-if scenarios intuitively and engagingly.

By going beyond traditional web-based visualization tools, Visu4EO offers advanced support for policy and decision-makers, combining immersive technologies with Explainable AI (XAI) to enhance transparency, understanding, and trust in the analysis process. The platform links observable effects to underlying decision variables, helping users better interpret impacts, identify critical issues, and develop more effective and informed responses.

The Visu4EO Platform

The Visu4EO platform enables users to define and manipulate environmental parameters to simulate urban planning interventions and evaluate their impacts on urban mobility, flooding, and pollutant concentrations through an immersive visualization experience. The application allows users to navigate a 3D representation of the scenario and, using advanced VR techniques, to seamlessly transition from a global overview to a local, street-level perspective.

The platform is structured around three main building blocks, which also define the operational workflow. First, users define scenarios by manipulating flood-related and urban-planning parameters via a GIS-based web interface. Once a scenario is configured, simulations are executed directly from the interface through the CityNexus Digital Twin Application available as a service within the DestinE platform (citynexus.destine.eu).

Simulation flooding results are then rendered, visualizing factors such as rainfall, river and canal flooding, sewer discharge capacity, traffic conditions, and road closures.

Finally, an Explainable AI (XAI) component performs model-agnostic analysis to assess the impacts of hydrological parameters and flood events on the road network, as well as the effects of changes to road segment characteristics. These explanations are integrated into the immersive experience to provide users with actionable insights during navigation.

Scenario & Simulation Technology

While Visu4EO is a general framework for analyzing flood scenarios and their impacts on urban mobility, the demonstration given during the project focused on the city of Copenhagen, addressing the growing risks posed by climate change, sea-level rise, and urbanization. Extreme rainfall events, such as the 2011 Copenhagen cloudburst (150 mm in less than two hours), together with increasing coastal flooding risks, highlight the need for resilient urban planning and adaptive mobility management.

Visu4EO enables users to create and evaluate customized flood scenarios by adjusting key parameters, including rainfall intensity and duration, sea-level conditions, river discharge, land-use configurations, and flood defense measures. These simulations support the assessment of flood impacts on transport networks, helping planners identify vulnerabilities and evaluate mitigation strategies.

The platform integrates advanced modeling technologies, including the MindGravity deep-learning mobility model, the SUMO traffic simulation engine, and the SFINCS hydrodynamic flood model. An Explainable AI (XAI) module further enhances transparency by providing interpretable insights into flood drivers, scenario effectiveness, and infrastructure resilience. Together, these capabilities support evidence-based decision-making for urban resilience, mobility planning, and climate adaptation.

VR Experience and Environment Generation

The Visu4EO VR experience is organized into three main scenes. The landing scene presents a rotating globe with selectable cities; for the current demonstration, Copenhagen is active, while additional cities can be enabled through configuration. After selecting a city, users enter an area selection scene showing a 2D offline map generated from GeoJSON data, where they can select an Area of Interest, compare up to three scenarios, or import new simulations from the engine.

The main simulation scene provides a fully navigable 3D reconstruction of the selected area. Flooding is visualized using Unity’s Water System, driven by GeoTIFF simulation outputs that dynamically adjust water levels according to the simulated flood profile. User interaction is managed through a multi-tab interface supporting map navigation, scenario comparison, flood-defense editing, time-slot selection, contextual street-level flood and traffic information, and XAI-based explanations.

The virtual environments are generated through a multi-stage pipeline combining geo-referenced data import, AI-assisted texturing, and performance-optimized 3D modeling. This process ensures geographic accuracy, visual realism, and real-time VR performance, enabling seamless integration of terrain, buildings, façade textures, and dynamic simulation layers within the Unity-based Visu4EO platform.

Explainable AI and Decision Support

Visu4EO provides an immersive decision-support environment designed to help policymakers and urban planners evaluate the consequences of planning decisions and assess the resilience of urban mobility networks during flood events. To support this objective, the platform integrates an Explainable AI (XAI) module that delivers transparent insights into flood dynamics, defense effectiveness, and mitigation strategies.

The XAI framework provides three complementary capabilities: (i) identification of the main hydrological drivers influencing flooding and traffic impacts through global sensitivity analysis; (ii) evaluation of the effectiveness of flood-defense scenarios by comparing flooding and mobility conditions against baseline scenarios; and (iii) decision-support recommendations for the optimal placement and configuration of flood-defense measures.

A Sobol-based sensitivity analysis is used to quantify the relative influence of rainfall, river discharge, and sea-level contributions on flood levels across the road network. Results are presented as street-level indicators and as automatically generated summaries highlighting the dominant drivers of flooding. Defense effectiveness is assessed by analyzing changes in flood depth, traffic speed, road occupancy, and road closures, producing interpretable indicators that identify benefits, trade-offs, and potential adverse effects of mitigation measures.

To support planning activities, the system evaluates multiple defense configurations and generates recommendations based on a combined efficiency–cost assessment. These insights help identify the most effective interventions for reducing flood impacts while preserving mobility performance.

The XAI results are directly integrated into the immersive VR environment through dynamic heatmaps projected onto the 3D city model. By linking explanations to their physical urban context, users can intuitively explore the causes and impacts of flooding, improving transparency, situational awareness, and evidence-based decision-making for urban resilience planning.

Research and Development tools

• Mobilkit: Open-source library developed by MindEarth in collaboration with Purdue University for the GFDRR (The World Bank) that provides tools for the analysis of human mobility data
• Pyrosm: a Python library designed for extracting OpenStreetMap data for various geospatial applications. It simplifies the process of working with OSM data by providing easy-to-use functions for data extraction and manipulation.
• SUMO: an open-source traffic simulation suite designed to model and analyze urban mobility and transportation systems.
• SFINCS: Super-Fast INundation of CoastS model for dynamic simulation of compound flooding in large-scale coastal systems.
• Ambiens Explore v2: a no-code-tool for interactive b2b projects designed by AmbiensVR and available on the Unity store.

Project Outcomes

• Project Repo: Visu4EO (Φ-lab / IMMERSIVE-EO · GitLab)
• Visu4EO: A Metaverse-Based Framework for Scenario Exploration, Causal Analysis, and Urban Resilience Planning. Paper at the ESA XR Conference 2026 (https://airdrive.eventsair.com/eventsairwesteuprod/production-atpi-public/cb36f4e539924afc8953d30ab2bd5fa2)