Digital twin technology is increasingly being used to simulate the effects of sea level rise, providing valuable tools for decision-makers in areas such as urban planning, coastal management, and disaster preparedness. These virtual models integrate real-time data from various sources, including geospatial imagery, AI, and environmental monitoring systems, to create detailed simulations of how rising sea levels could impact specific regions. 

By accurately mapping current land cover features and continuously updating these models with new data, digital twins allow researchers and governments to predict future scenarios under different climate change conditions. This helps to identify vulnerable areas, plan protective infrastructure, and optimize evacuation strategies. For instance, high-resolution geospatial data can show which areas are at risk of flooding, while AI-driven simulations predict how rising sea levels might affect local ecosystems and urban environments over time.

Incorporating sea level rise into digital twin simulations can help urban planners and environmental scientists understand the long-term impacts on coastal regions, ensuring better preparedness for the challenges posed by climate change. This technology is critical for visualizing and planning adaptive responses to mitigate the potential damages of rising sea levels.

The Tonga Preparedness Pilot Project enhances disaster preparedness using remote sensing in Tonga and is a joint initiative by the UN-SPIDER and the Committee on Earth Observation Satellites (CEOS). The National Disaster Risk Management Office leads this initiative in Tonga and will utilize satellite imagery and advanced Earth observation technologies to enhance our disaster preparedness significantly.

Tonga is a Polynesian country known as the Kingdom of Tonga. It has 171 volcanic, limestone, sand, and mixed islands, many uninhabited. Agriculture and tourism are the primary economic earners. Tonga is divided into three major islands: Tongatapu, Ha'apai, and Vava’u. The recent volcanic eruption in Tonga's vicinity has raised the risk of widespread and severe salinity increase in the groundwater system because of the Tsunami inundating large coastal areas (Sharan et al., 2023).

CEOS stands for the Committee on Earth Observation Satellites. It is an international coordination mechanism that brings together space agencies and other organizations to optimize the use of Earth-observing satellite data for societal benefit. It aims to foster collaboration, improve data access, and address global challenges related to the environment, climate, and disasters.

To create the digital twin product derived from satellite imagery, Airbus Defence and Space provided very high-resolution optical satellite data from Pleiades Neo, in 15cm and 30cm resolutions with the higher resolution of the Digital Surface Model. Space Data Inc. created the digital twin products using AI technology in combination with the provided very high-resolution imagery.

Creating realistic 3D products from satellite imagery 

Digital twin technology, when applied to satellite imagery, represents a cutting-edge approach to creating real-time, dynamic models of the physical world. By leveraging high-resolution geospatial data from satellites, this method enables the reconstruction of urban environments, climate patterns, and natural phenomena with unprecedented accuracy. The system integrates various layers of data, including satellite images, terrain information, and environmental conditions, to provide users with a highly detailed, interactive digital twin. 

After the acquisition of diverse geospatial datasets from satellites, the images were processed using AI algorithms and machine learning models to identify and categorize natural and man-made structures, such as buildings, roads, rivers, and vegetation. The AI then generates 3D models, which form the backbone of the digital twin environment. These models can be used for tasks such as monitoring urban development, assessing disaster impact, and analyzing environmental changes.

The process described below leverages the power of advanced AI algorithms and state-of-the-art 3D rendering software to create highly realistic and dynamic 3D representations of the physical world, based on real-time satellite data. 

1. Acquiring High-Resolution Geospatial Imagery: The first step is procuring high-resolution satellite imagery. This imagery captures vast areas with precision, providing essential data on land cover, buildings, and other physical features. The resolution and clarity of these images are crucial for creating accurate 3D models. Modern satellites equipped with advanced sensors can capture detailed information about terrain, structures, and other environmental features, which serve as the foundation for 3D modeling.
    
2. Data Processing with AI and Machine Learning: Once the imagery is acquired, AI algorithms, such as NeRF (Neural Radiance Fields) or Gaussian Splatting are applied to process the image data. These algorithms (see Figure 3), help in detecting features like building heights, land cover, and surface materials with high precision. The AI systems extract critical details about shapes, colors, and textures to prepare for the next stage of model generation. This automated step significantly reduces the time and labor required for manual data extraction.
    
3. 3D Model Generation: The processed data is then fed into specialized 3D model generation software such as Houdini, Blender or other platforms capable of reconstructing the physical world in a digital environment. The software uses satellite data to create accurate, textured 3D models that represent the real-world environment. In this step, elements like building facades, roof structures, and urban features are recreated to match the real world closely.
    
4. Enhancing Visual Accuracy: To further improve the realism of the 3D models, additional details such as materials, lighting, and environmental effects are applied. For instance, textures for walls, windows, and rooftops are added based on satellite data (see Figure 4). Tools like Unreal Engine (UE5), Unity, or other engines are used to simulate 3D worlds including realistic lighting, shadows, and reflections to make the 3D models visually appealing and true to life.
    
5. Finalizing and Updating the Model: The 3D model is continuously updated with the latest satellite imagery to reflect real-time changes in the environment. As new data becomes available, AI-based systems can automatically detect changes (e.g., new buildings, modified landscapes) and adjust the 3D model accordingly. This step ensures that the digital twin remains accurate and up to date.

How to apply 3D Realistic Digital Twin products in the Disaster Risk Management Context?

In the CEOS Tonga Preparedness Pilot Project, using the above advanced technology, realistic buildings, roads, ships, cars, or other physical components were created. Specifically, with these realistic features, the sea level rise simulation figures derived from the very high-resolution optical satellite image are shown below in Figure 5. 

The sea-level rise simulation can provide essential information on the long-term risks posed by rising seas to coastal areas of Tongatapu Island. It allows for the identification of vulnerable zones, helping in the planning of evacuation routes, the strengthening of coastal defenses, and the development of disaster-resilient infrastructure. By using predictive models, the simulation helps Tonga prepare for future sea-level rise scenarios and mitigate the impacts of flooding and coastal erosion. See simulation examples in Figure 6. 

A realistic digital twin for disaster prevention systems can be a highly effective tool if it integrates both spatial and temporal data to represent real-world conditions in dynamic detail for the raise of social awareness. In the context of the Tonga project, ArcGIS Pro was used. The power of GIS Software, such as ArcGIS Pro and QGIS, lies in its ability to combine vast geospatial datasets, including satellite imagery, terrain models, and weather data, to build a digital replica of areas prone to disasters. This digital twin has capacity beyond sea level rise simulations. It is used to simulate various disaster scenarios, offering insights into potential impacts on infrastructure, the environment, and human populations. For example, one can use GIS software to highlight flooded buildings as provided in Figure 7 where they are visualized in red. 

One of the key benefits of using GIS Software is its capacity for advanced visualization and analysis, making it easier to identify vulnerable zones, model hazard progression, and assessing the effectiveness of mitigation strategies by creating the risk map for flooding. Real-time data feeds from sensors or satellites can further enhance the digital twin by updating it continuously, allowing for real-time monitoring of changing conditions, such as rising sea levels or storm intensification. This enables decision-makers related to the Kingdom of Tonga to anticipate risks and respond proactively to minimize damage in their disaster prevention planning.

Moreover, the digital twin can serve as a collaborative platform for national government agencies, local communities, and disaster response teams. By enabling them to interact with the same data and models, stakeholders can plan and execute disaster response strategies more efficiently. The digital twin could also incorporate predictive analytics, helping to anticipate future risks based on current trends and historical data, making it a vital tool in long-term disaster resilience planning.

In essence, creating a digital twin with GIS Software involves leveraging its full suite of tools for mapping, simulation, and analysis, while continuously integrating real-time data to reflect the current state of the environment. This ensures that decision-makers can use the digital twin to both prepare for and respond to disaster scenarios in a timely and effective manner. Below are examples of using the flood simulation tool.

Conclusion

In conclusion, the implementation of digital twin technology, particularly through high-resolution satellite imagery and advanced AI algorithms, provides a dynamic and detailed representation of the real world which makes users to understand easily the impact of the rising sea level. Also, this technology can significantly aid decision-makers in disaster preparedness and response by simulating potential sea-level rise disaster. For Tongatapu Island, Kingdom of Tonga, these digital twin products allow the identification of vulnerable coastal areas and contribute to planning evacuation routes, reinforcing coastal defenses, and developing disaster-resilient infrastructure. The continuous integration of real-time data ensures the digital twin remains updated and relevant, making it a critical tool for both immediate disaster response and long-term resilience planning.