Harnessing Game Engines and Digital Twins: Advancing Flood Education, Data Visualization, and Interactive Monitoring for Enhanced Hydrological Understanding

Abstract

Given the increasing frequency and severity of floods caused by climate change, there is a pressing requirement for creative ways to improve public comprehension and control of hydrological phenomena. Contemporary technology provides unparalleled possibilities to transform this domain. This project investigates the possibilities for merging gaming engines and digital twins to enhance flood education, data visualization, and interactive monitoring. This study proposes the utilization of immersive digital twins to enhance the comprehension of hydrological and hydraulic systems. The suggested method utilizes game engines to generate dynamic and interactive models that connect raw data to practical insights, enabling a more profound understanding of flood dynamics. This study underscores the wide-ranging usefulness of digital twins in various watersheds by focusing on the development of advanced monitoring systems, the benefits of improved data visualization, and educational outreach. The incorporation of real-time data via IoT technology considerably improves the significance and precision of these virtual models. This novel approach seeks to refashion flood management approaches by cultivating well-informed stakeholders and advocating for effective environmental education, ultimately leading to more resilient and prepared communities. An immersive digital twin of the real world can assist decision-makers technically, psychologically, and mentally by making complex phenomena easier to understand and visualize, thanks to real-time data and simulations that keep the information up-to-date, consequently leading to a more precise and intuitive decision-making process.

1. Introduction

Flooding poses a significant natural hazard with profound implications for human safety, infrastructure, and ecosystems. Effective flood management necessitates advanced technologies that can enhance data interpretation, stakeholder engagement, and decision-making [1,2]. This study proposes a novel integration of digital twin technology and game engines to advance flood monitoring systems, addressing critical needs in modern flood management.

Digital twins are sophisticated virtual models that replicate physical systems, allowing for real-time simulation and analysis. These models have transitioned from industrial applications to environmental monitoring, offering valuable insights for flood management [3]. Digital twins integrate diverse data sources—including remote sensing, ground-based sensors, and hydrological simulations—to create dynamic models that can simulate various flood scenarios and predict potential impacts [4]. Despite their advantages, challenges remain in ensuring simulation accuracy, managing large datasets, and maintaining real-time updates.

Game engines such as Unity and Unreal Engine have evolved from their origins in video game development to become powerful tools for creating interactive and immersive simulations. These engines provide high-quality graphics and real-time rendering, which are advantageous for visualizing complex flood scenarios [5]. The integration of game engines with digital twins can significantly enhance user interaction and engagement, making flood simulations more intuitive and accessible. Nevertheless, game engines face limitations in accurately representing the multifaceted nature of floods and integrating seamlessly with digital twin models [6].

Early flood monitoring systems relied heavily on manual data collection methods such as river gauges and visual inspections. While foundational, these systems offered limited spatial and temporal coverage, leading to insufficient data for comprehensive flood management [7]. Remote sensing and satellite-based systems have improved flood monitoring by providing extensive spatial coverage through techniques such as synthetic aperture radar (SAR) and optical imagery. For example, Sentinel-1 satellites offer near-real-time data essential for flood monitoring over large areas [8]. Despite these advancements, challenges such as temporal resolution limitations and data interpretation issues under adverse weather conditions persist [9]. In recent decades, studies have discussed the transformative role of artificial intelligence (i.e., machine learning and deep learning models) in flood monitoring, flood risk analysis, and management [10,11,12,13,14], yet they have strived to incorporate a more practical and immersive use for stakeholders.

IoT-based flood monitoring systems utilize networks of sensors to gather real-time data on water levels, rainfall, and other variables, enhancing spatial resolution and data accessibility. These systems improve monitoring capabilities but face issues related to sensor reliability, data accuracy, and data transmission [15]. Additionally, they often lack advanced visualization features necessary for effective stakeholder engagement and decision-making [16]. Machine learning and predictive analytics have advanced flood forecasting through sophisticated algorithms applied to historical and real-time data. Techniques such as neural networks and ensemble forecasting models have enhanced prediction accuracy but require large volumes of data and can be complex to interpret [17]. Integrated systems and multi-source data fusion have emerged as approaches to combine data from various sources for a comprehensive flood monitoring strategy. These systems provide improved prediction accuracy but face challenges related to data compatibility and integration complexity [18]. Moreover, they often lack interactive features that could enhance stakeholder engagement and decision-making [19].

Only in the past few years have researchers around the world started to investigate the value of cutting-edge technologies and innovative concepts such as digital twins, game engines, virtual reality, augmented reality, mixed reality, and extended reality to strengthen disaster risk management, advance early and real-time flood warning systems, and enhance participatory web-based flood monitoring systems and flood risk reduction and mitigation [20,21,22,23,24,25,26,27,28,29,30,31,32,33]. Current flood monitoring systems provide valuable insights but often fall short of offering interactive and immersive experiences that facilitate effective data interpretation and stakeholder engagement. The integration of digital twins with game engines presents a promising solution to these limitations. Digital twins provide dynamic, real-time simulations of flood scenarios, while game engines offer interactive visualizations that can significantly improve user engagement and understanding [6].

This study aims to develop a framework that combines these technologies to enhance flood monitoring systems by improving data visualization, supporting decision-making, and advancing flood education. By addressing the limitations of existing systems and leveraging the strengths of digital twins and game engines, the research seeks to establish a new paradigm in flood management. The proposed approach aims to improve real-time data integration, enhance interactive visualization, and foster better community engagement, ultimately contributing to more effective and informed flood management strategies.

2. Materials and Methods

The motivation behind the proposed approach is to leverage advanced technologies to enhance data visualization, enable real-time monitoring, and foster stakeholder engagement. Overall, this approach aims to transform raw data into actionable insights and wisdom, as illustrated in Figure 1. Integrating game engines with digital twin technology aims to address the limitations of traditional flood monitoring systems and provide a more effective tool for flood management and public awareness. The proposed framework is illustrated schematically in Figure 2 and further detailed in a step-by-step format in Figure 3. Steps in the development of an “Immersive Digital Twin” can include the following.

2.1. Step 1: Creating a Virtual Replica

The creation of a virtual replica is the foundational step in developing an effective digital twin. This process involves several detailed techniques to ensure that the virtual environment accurately represents the physical watershed.

2.1.1. Aerial Photogrammetry

This technique utilizes UAVs equipped with high-resolution cameras to capture images from above. The UAVs fly over the area of interest, taking numerous overlapping photographs from different angles. These images are processed using photogrammetry software (e.g., ContextCapture Center v23.0.0.1317 or Agisoft Metashape Pro. v2.1.3) to create detailed 3D models of the terrain [34]. This method provides a broad overview of the landscape, capturing large-scale features such as river networks, floodplains, and land use patterns.

2.1.2. Close-Range Photogrammetry

This involves taking detailed images from ground level to capture smaller, intricate elements that aerial images might miss, such as buildings, vegetation, and critical infrastructure [35]. This technique enhances the overall accuracy of the virtual replica by filling in gaps and providing additional detail.

2.1.3. Three-Dimensional Gaussian Splatting

This advanced technique is employed to create high-quality, detailed 3D models by using a probabilistic approach to model point clouds more efficiently. Three-dimensional Gaussian splatting represents data points as Gaussian blobs, which improves the representation of surfaces and textures compared with traditional photogrammetry methods. This technique offers smoother and more accurate visualizations of complex surfaces and can handle high-density point clouds better, leading to more precise and aesthetically pleasing models [36].

2.1.4. Universal 3D Repositories

When photogrammetry data are incomplete or unavailable, universal 3D repositories like Sketchfab, TurboSquid, and Quixel Megascan can be utilized. These platforms offer a range of pre-made 3D models and assets, including common objects and terrain features. Integrating these assets helps complete the virtual landscape, ensuring it accurately reflects the physical environment [37].

2.2. Step 2: Integration with Game Engines and Creating an Interactive Environment

After creating the virtual replica, integrating it into a game engine and enhancing it with interactive features is crucial for developing a dynamic environment.

2.2.1. Selection of Game Engine

Unreal Engine is known for high-fidelity graphics and powerful simulation capabilities. Unreal Engine’s Lumen and Nanite features contribute to creating detailed and immersive environments. Lumen provides real-time global illumination, enhancing the accuracy of lighting effects and shadows [38]. This is crucial for visualizing flood scenarios accurately. Nanite, on the other hand, allows for rendering highly detailed environments with complex geometry, supporting high-resolution textures and models without sacrificing performance [38]. Unity offers flexibility and a wide range of assets and tools for simulation and visualization. Unity’s asset store provides plugins and extensions that can enhance the virtual environment and offer additional functionality [39].

2.2.2. Interactive Features

Game engines enable the creation of interactive elements within the virtual environment. Users can engage with the environment through various features, such as scenario simulations and data analysis. This interactivity allows users to explore flood scenarios, test management strategies, and gain insights into flood dynamics more engagingly [40].

2.2.3. Photo-Realistic Rendering

Photo-realistic rendering techniques in game engines enhance the effectiveness of the flood monitoring system. By achieving high-quality visual representation, users can better understand the environment and the impact of different flood scenarios. Realistic depictions of water behavior, terrain changes, and flood effects contribute to a more accurate and immersive experience [41].

2.3. Step 3: Augmenting Data with Static and Dynamic Inputs

Integrating various types of data is essential for ensuring the accuracy and relevance of the digital twin.

2.3.1. Static Data Integration

This involves incorporating pre-existing data such as terrain elevation, land use, and infrastructure into the virtual environment. Static data provide the foundational characteristics of the environment and establish a baseline for the virtual replica [42]. Additionally, users can incorporate pre-existing text, image, audio, and video data into their game environment and interact with them while playing.

2.3.2. Dynamic Data Integration

Real-time data from IoT sensors are integrated to keep the digital twin up-to-date. Sensors measure variables like rainfall, river levels, and soil moisture. These data are transmitted to the digital twin, allowing for continuous updates and accurate representations of current conditions [43].

2.3.3. IoT and Real-Time Data Augmentation

To seamlessly integrate real-time data into the game engine, Unreal Engine provides a plugin called the MQTT Plugin. MQTT (Message Queuing Telemetry Transport) is a lightweight messaging protocol designed for low-bandwidth, high-latency, or unreliable networks. The MQTT plugin enables Unreal Engine to communicate directly with IoT sensors and data sources, allowing for real-time data updates within the virtual environment (Figure 4). This integration ensures that the digital twin remains current with live sensor data, enhancing its accuracy and responsiveness to changing conditions.

2.3.4. Crowdsourcing and Citizen Science

Crowdsourcing platforms and mobile applications collect additional data from the public, validate sensor data, and provide insights that might not be captured by traditional methods [44]. Citizen science contributes to a more comprehensive understanding of flood dynamics and enhances the accuracy of the digital twin.

2.3.5. Application of AI and Machine Learning

AI and machine learning play a pivotal role in managing and analyzing the large volumes of data associated with flood monitoring. These technologies can be employed for data sorting, matching, and analysis, facilitating the extraction of hidden patterns from big data. Advanced algorithms can identify trends and anomalies in real-time data, enhance predictive modeling, and improve flood forecasting. For example, machine learning models can analyze historical and current data to predict flood events with greater accuracy, offering insights into potential future scenarios. The integration of AI-driven analytics enables more sophisticated decision-making processes and refines the digital twin’s capability to provide actionable predictions [45].

2.4. Step 4: Simulation and Modeling

Simulation and modeling are essential for analyzing flood scenarios and assessing their impacts.

2.4.1. Flood Simulation Plugins

Game engines offer specialized plugins for advanced flood simulation. For example, FluidFlux in Unreal Engine uses shallow water equations to model fluid dynamics and simulate realistic flood scenarios. It provides detailed visualizations of water behavior, inundation patterns, and flow dynamics [46]. FluidFlux offers higher realism compared with traditional engineering software such as HEC-RAS 6.5, which, while more accurate, often lacks the immersive visualization capabilities of game engine plugins. Additionally, Ninja plugin in Unity supports various fluid simulation techniques, allowing users to visualize different flood scenarios and their effects on the environment [47].

2.4.2. Real-Time Modeling

The simulation capabilities of game engines enable real-time modeling of flood scenarios. Users can interact with the virtual environment, adjust parameters, and observe the effects of different flood management strategies. This dynamic and interactive analysis provides valuable insights for flood management and decision-making [48].

The development of a digital twin for flood monitoring involves creating a virtual replica, integrating with game engines, augmenting data, and simulating scenarios. By leveraging advanced game engine features, real-time data, and interactive elements, this approach aims to enhance flood monitoring practices and provide valuable insights for stakeholders. The combination of detailed 3D modeling, immersive visualization, and dynamic simulation represents a significant advancement in the field of environmental monitoring.

2.5. Hardware Requirements for Creating Immersive Digital Twins in Unreal Engine 5

Computer hardware including computational and graphical requirements should adhere to the necessities of real-time rendering, gamification, and animation needs. Unreal Engine 5, one of the best engines for creating immersive gamified digital twins, was investigated. The recommended requirements in two spectra of least (budget) build to the best (high-performance) are summarized in

Table 1. Hardware requirements for implementing immersive digital twins in Unreal Engine 5 (Sources: references listed herein and Puget Systems website: https://shorturl.at/ZaD5E; accessed on 24 August 2024).

Hardware	Minimum (Budget) Build	Recommended (Best) Build
Central Processing Unit (CPU)	Intel 1 Core i7-14700K or AMD 2 Ryzen 7 7800X	Intel’s Core i9 13900K and 14900K or AMD’s Threadripper 7000 series
Graphical Processing Unit (GPU)	NVIDIA RTX 2070 SUPER or AMD RX 6600 XT	NVIDIA 3 RTX 4080 or 4090
Random-Access Memory (RAM)	16 GB–32 GB	64 GB–128 GB
Storage	500 GB SSD	1TB NVMe 4 SSD

Notes: ¹ Intel Corporation headquartered in Santa Clara, California, and incorporated in Delaware, USA. ² Advanced Micro Devices (AMD), Inc. based in Santa Clara, California, USA. ³ NVIDIA Company headquartered in Santa Clara, CA, and incorporated in Delaware, USA. ⁴ NVMe (nonvolatile memory express): A new storage access and transport protocol for flash and next-generation solid-state drives (SSDs).

[49,50].

3. Results and Discussion: Conceptual Implementation and Potential Applications

3.1. Virtual Replica Creation

The development of a virtual replica using techniques such as aerial photogrammetry, close-range photogrammetry, and 3D Gaussian splatting is foundational for creating an immersive digital twin. Aerial photogrammetry provides a broad overview of the landscape, capturing large-scale features like river networks and floodplains. Close-range photogrammetry adds detail to smaller elements, such as buildings and vegetation, which are crucial for accuracy. The incorporation of 3D Gaussian splatting is particularly promising as it improves the representation of complex surfaces and textures compared with traditional methods. By representing data points as Gaussian blobs, this technique enhances the visual and spatial accuracy of the 3D models, offering smoother and more detailed representations of the terrain [36]. Figure 5 presents a 3D global digital elevation model, augmented with close-range data (timber dam) using the 3D Gaussian splatting technique. The model showcases evident potential benefits and detailed features. Additionally, it incorporates the impressive 3D model collection from Quixel Megascan, including trees, vegetation, dirt, and post-processed colors and lighting, to create a more lifelike environment. This model is integrated into Unreal Engine to enhance its interactivity, as discussed below.

3.1.1. Challenges and Advances in Topographical Integration with Simulation Engines

Integration of simulation engines with topographical information with specified geometric precision constructs the backbone of any fluid simulation. Digital elevation model (DEM) as a digital attribute of Earth’s surface dictates for flow pattern, velocity, path, and predicted propagation extent. Also, the geometric precision of DEM, rendered as resolution, directly influences the simulation accuracy, especially in areas where small changes in elevation in forms of spikes and sinks can considerably change the simulation results of water flow. Main challenges involved in such integration include data quality and availability, data processing, and scalability [51,52]. Most notably, data quality and availability are considered limiting criteria in developing countries, which can alter the simulation results while using different 2D and 3D water flow simulation tools. Data processing is also an issue for large areas where modelers encounter an intense computational procedure. Remarkably, new versions of game engines such as Unreal Engine 5 have come up with advanced solutions called Nanite technology with which end-users can input highly-detailed topography without sacrificing performance and precision. Yet, keeping the model accurate when scaling the simulation from small and localized areas to larger basins while facing computational loads can be a challenge for modelers and needs to be tackled with caution. Figure 6 portrays the addition of large, highly detailed topography data into Unreal Engine before and after adding 3D assets (here, dense forests, timbers, and boulders).

3.2. Integration with Game Engines

Integrating virtual replicas with game engines like Unreal Engine and Unity offers significant advancements in flood visualization and simulation. Unreal Engine’s features, such as Lumen for real-time global illumination and Nanite for detailed geometry rendering, could enhance the realism and accuracy of flood simulations. Unity’s flexible toolset and extensive asset library also provide valuable resources for creating interactive and dynamic environments. The conceptual integration of these engines suggests substantial improvements in user engagement and scenario analysis. Game engines’ capabilities in rendering high-fidelity graphics and creating interactive features can transform how flood scenarios are visualized and managed, making the data more accessible and actionable. The ability to modify and enhance flood simulation plugins within game engines, incorporating methodologies from traditional engineering software, offers the potential for creating more sophisticated and versatile simulations that blend technical rigor with interactive features. Figure 7 illustrates a dialogue system created for the previous landscape in Unreal Engine, aimed at engaging the end-user, immersing them in the virtual environment, and invoking a sense of interactivity. These dialogue systems can be transformed into AI systems with a strong data foundation, enabling them to respond to a wide range of questions posed by end-users.

3.3. Data Augmentation and Real-Time Integration

The integration of real-time data through IoT sensors is a critical component for maintaining the relevance and accuracy of a digital twin. The MQTT plugin for Unreal Engine is particularly noteworthy as it enables direct communication between IoT sensors and the virtual environment. This real-time data integration allows for continuous updates and accurate representations of current conditions. The MQTT plugin’s ability to handle low-bandwidth and high-latency conditions ensures that the digital twin remains current and responsive, which is essential for effective flood management. Figure 8 demonstrates static and dynamic data augmentation in Unreal Engine through the use of interactive maps, audio, video, and text.

3.4. Interactive Features and Simulations

The potential for interactive features in game engines to enhance flood education and public engagement is substantial. By allowing users to explore flood scenarios and test management strategies, game engines can make complex flood dynamics more understandable. The concept of using scenario simulations to provide insights into flood behavior and management strategies is promising. The use of advanced flood simulation plugins, such as FluidFlux in Unreal Engine and Ninja Plugin in Unity, offers the possibility of detailed visualizations of water behavior and inundation patterns. These tools could facilitate more interactive and informative analyses of flood scenarios, contributing to better decision-making and public awareness. Figure 9 depicts a game environment developed in Unreal Engine, utilizing actual elevation data and pre-made environmental assets. The lake system is crafted using both Unreal Engine’s Water Lake System and FluidFlux, enhancing the realism of the water. This integration allows for greater interactivity with the end-user’s movements, incorporating dynamic features such as buoyancy, flow propagation, and slow-motion effects when the player is swimming or exploring underwater (i.e., water post-processing). Similarly, this approach would also enhance the understanding of concepts such as flood depth, volume, and velocity.

3.5. Advanced Tools for Simulating Precipitation and Runoff in Unreal Engine

Creating a detailed and scientifically accurate simulation of precipitation and runoff in Unreal Engine involves leveraging several advanced tools and plugins, each of which offers a unique set of features and capabilities. Below, the key components and methodologies used in this process are outlined, highlighting specific Unreal Engine functions and parameters, as well as theoretical underpinnings from hydrological and hydraulic modeling.

3.5.1. Niagara System for Precipitation Simulation

The Niagara system in Unreal Engine is a highly versatile and powerful tool for creating particle effects, including realistic precipitation simulations. For rainfall, the following parameters can be controlled:

Drop Size

This parameter can be adjusted to simulate different types of precipitation, from light drizzle to heavy rain. The drop size can be linked to a variable representing the water content in the atmosphere.

Velocity

The velocity at which raindrops fall can be controlled using vector fields or gravity nodes within the Niagara system. This parameter is crucial for simulating the speed of rainfall, which varies with weather conditions and wind.

Direction

Wind nodes in Niagara can be used to influence the direction of raindrop particles, allowing for the simulation of wind-driven rain, which is particularly important in modeling storm scenarios.

Intensity

The intensity of rainfall can be controlled by varying the emission rate of particles. This can be linked to real-time weather data or synthetic scenarios based on historical rainfall records.

3.5.2. Chaos Physics for Interaction with Terrain and Objects

Unreal Engine’s Chaos Physics system can be used to simulate the physical interaction of raindrops with terrain and objects. This includes the following:

Collision Detection

Raindrops can be programmed to collide with surfaces, which can then trigger responses such as splashes or absorption.

Buoyancy and Surface Water

In cases where heavy rainfall leads to surface water accumulation, Chaos Physics can simulate the buoyancy of objects and the interaction between water and terrain.

3.5.3. FluidFlux Plugin for Hydraulic Modeling

The FluidFlux plugin in Unreal Engine offers advanced tools for simulating water flow and surface runoff. This plugin can model the following:

Infiltration and Surface Flow

Using terrain data, FluidFlux can simulate how water infiltrates the soil and moves across surfaces, mimicking real-world hydrological processes.

Hydraulic Connectivity

The plugin can model the connectivity of different landscape elements, such as channels and depressions, influencing the flow paths and accumulation of water.

Dynamic Water Levels

It can dynamically adjust water levels based on rainfall intensity and duration, simulating rising water levels in real time.

3.5.4. Integration with Hydrological Models

To create a scientifically accurate simulation, the parameters and outputs from the Niagara system, Chaos Physics, and FluidFlux can be integrated with established hydrological models such as the Soil Conservation Service (SCS) runoff model or the Green-Ampt infiltration model. This integration allows for the following:

Calibration and Validation

The simulation can be calibrated against real-world data, such as recorded rainfall events and observed runoff patterns, to ensure accuracy.

Scenario Analysis

Different rainfall scenarios, including extreme weather events, can be simulated to study potential impacts on the landscape and built environment.

3.5.5. Visual Representation and Data Output

Visual elements such as raindrop effects, water accumulation, and flow paths can be rendered in real time, providing a vivid representation of the simulation. Additionally, data outputs such as runoff volumes, flow rates, and infiltration rates can be exported for further analysis.

3.5.6. Material and Shader Techniques for Water Effects

Material and shader techniques in Unreal Engine can be utilized to create realistic water surfaces and effects, such as wet surfaces and dynamic puddles. This can be achieved through the following:

Reflective and Refractive Materials

These can simulate the optical properties of water, including reflection and refraction, to create realistic-looking water bodies.

Displacement Maps

These are used to simulate small-scale water surface details, such as ripples and waves, adding to the realism of the scene.

Wet Surface Effects

Custom shaders can simulate wet surfaces and water absorption, which are critical for accurately depicting the environmental impact of precipitation.

3.5.7. Environmental Query System (EQS) for Water Flow Pathfinding

The Environmental Query System (EQS) can be employed to determine optimal water flow paths across a landscape. This system allows for the following:

Flow Path Optimization

EQS is an artificial intelligence system that can be used to calculate the path of least resistance for water flow, taking into account terrain features and obstacles.

Flood Risk Assessment

By simulating different rainfall scenarios, EQS can help in assessing areas at risk of flooding, enabling better planning and mitigation strategies.

3.5.8. Blueprint Scripting for Dynamic Weather Systems

Blueprints in Unreal Engine are a robust visual scripting system that enables developers to create complex game logic and functionalities without extensive programming skills. By utilizing nodes that encapsulate various functions and events, developers can craft detailed behaviors and interactions by connecting these nodes in a flowchart-like structure. This system offers an intuitive and versatile approach to managing diverse game elements, such as character movement, environmental interactions, and user interfaces. For instance, Blueprints can be employed to implement features like information callouts, character navigation, audio management—including flood alerts and informational prompts—and the simulation of realistic water physics (Figure 10, Figure 11, Figure 12 and Figure 13). The latter includes parameters like buoyancy, wave formation, and water flow that adapt to the underlying topography, as illustrated in Figure 13.

Blueprint scripting in Unreal Engine can also be used to create dynamic weather systems that respond to real-time data or predefined scenarios. This includes the following:

Weather Events

Scripting can trigger specific weather events, such as thunderstorms, heavy rainfall, or clear skies, based on input data.

Time of Day and Seasonal Changes

Blueprint scripts can also control the time of day and seasonal variations, influencing weather patterns and precipitation characteristics.

Adaptive Response Systems

Systems can be designed to dynamically adjust the environment based on weather changes, such as increased water levels during rain events.

3.5.9. Advanced Fluid Dynamics with Flow Maps

Flow maps are a technique used to direct the movement of fluid elements within a scene. They are particularly useful for the following:

Simulating River and Stream Flow

Flow maps can be applied to textures to simulate the movement of water in rivers and streams, ensuring that water follows the expected paths based on terrain and elevation data.

Erosion and Sediment Transport

These maps can also simulate erosion processes and the transport of sediments, providing a more comprehensive model of hydrological dynamics.

3.5.10. Virtual Reality (VR) and Augmented Reality (AR) Integration

Unreal Engine supports VR and AR, which can be leveraged to create immersive experiences for studying precipitation and runoff. This integration offers the following:

Immersive Visualization

VR can provide a first-person perspective of a landscape during different weather conditions, enhancing understanding and engagement.

Interactive Learning

AR can overlay simulated data on real-world environments, providing interactive educational tools for exploring hydrological processes.

By utilizing Unreal Engine’s Niagara system, Chaos Physics, FluidFlux plugin, material and shader techniques, EQS, dynamic weather scripting, flow maps, and VR/AR integration, developers can create comprehensive simulations of precipitation and runoff. These simulations are not only visually compelling but also scientifically robust, making them valuable for education, research, and environmental management.

3.6. Comparisons with Traditional Engineering Software

3.6.1. Comparative Overview of 2D Water Simulation Engines

This section compares three widely used 2D water simulation engines, namely, HEC-RAS, TUFLOW, and SWMM [53]. A discussion is provided covering the strengths and weaknesses of each engine, including their accuracy and computational efficiency. HEC-RAS is one of the well-known, freely accessible hydraulic software that covers riverine and urban flood simulations. Although it may fall short for complex hydrodynamic processes of entangled flooding and erosional episodes, it provides a versatile platform for both academia and professionals to portray steady and unsteady flow, sediment transport, and water quality simulation scenarios. It supports spatial data with GIS-supported formats. Although designed mostly for 1D and 2D simulations, new versions of this software also exhibit a decent 3D visualization for different flooding variables, such as flow depth, velocity, shear stress, and the like. In comparison, TUFLOW builds upon rigorous computational accuracy and efficiency for surface and subsurface flow and the involved complex hydrodynamical processes. It is a professional software, especially for drainage systems and well covers high-resolution spatial data. However, as a commercial software, it comes with a high price for end-users, limiting its accessibility to a narrow range of experts who can not only afford the purchase but also are well versed in high-level and complex model setup and calibration. SWMM is also a professional and simultaneously an easy-to-use software for urban stormwater and combined sewer overflow (CSO) systems. It is the best choice for regulatory compliance and environmental impact assessment on runoff quantity and quality. It is a freely available tool for all ranges of end-users. However, due to its 1D computation, it may underperform TUFLOW in complex drainage systems both in computational efficiency and accuracy.

3.6.2. Comparative Overview of 3D Water Simulation Engines

In order to provide a general pool of 3D software comparison, we investigated four famous 3D water simulation software programs, including ANSYS, OpenFOAM (and FLOW-3D HYDRO), Unity 3D, and Unreal Engine [54]. We tried to shed light on both poles of the technical/engineering and visualization/animation spectrum. Our overall assessment shows that ANSYS integrates highly complex computational fluid dynamics (CFD), can well integrate water–object interaction formulas, and is more adopted in critical engineering projects such as dam break simulations for water-related industrial companies rather than for educational purposes and research. It is also costly and is intended for commercial use. In contrast, OpenFOAM and FLOW-3D HYDRO, despite some mathematical differences and both being based on CFD simulation, are less complex than ANSYS and are far more suitable for research purposes. In particular, OpenFOAM is an open-source software and is often used by researchers to design different hydraulic measures. Compared with the two latter engineering software programs, Unity 3D and Unreal Engine, famous game development engines, are more on the visualization and agile simulation side of the spectrum due to their high-fidelity to life-like real-time simulation and rendering and are more suitable for stakeholders’ preparedness and can lead to a better decision-making process for authorities. Most notably, the Blueprint scripting option in Unreal Engine has made it even easier for non-programmers to become well versed in implementing this engine for quick water simulation spontaneously.

3.6.3. Comparative Analysis of 3D and 2D Water Simulation Software

It is evident that 2D flood simulation tools are more computationally efficient since they are designed for efficient flow simulation in two dimensions without involving vertical flow complexities. However, this can backfire when highly complex stratified flows are set to simulation and high accuracy is required for flow interaction with critical infrastructures. Hence, 2D tools are more suitable when large-scale flow simulation and flood risk assessment need to be addressed, and the project is less resource-intensive, while 3D flow simulation tools can be more suitable for highly complex flows and when flow—human design interactions are involved. On the other hand, in order to close the gaps between high-accuracy and visualization prerequisites and stakeholder engagement, education and prompt prototyping are paramount, and a more democratized, accessible, and versatile tool would be required, which can come true in Unreal Engine. Its modifiable simulation tools also open the opportunities to not only bridge high-accuracy 3D models to a more interactable environment but also build new advanced and complex models within such a game engine to utilize its interactivity, profound visualization, and real-time rendering capabilities.

Traditional engineering software for hydrological and hydraulic modeling, such as HEC-RAS and SWMM, has been instrumental in flood simulation and management. These tools are highly accurate and widely used for their detailed mathematical modeling and rigorous simulations. However, they often lack the immersive and interactive features offered by game engines. Traditional software typically focuses on precise calculations and simulations but may not provide the intuitive, user-friendly interfaces necessary for effective communication with non-specialists.

In contrast, the proposed framework leverages game engines to create dynamic and interactive simulations that can bridge the gap between complex scientific data and user engagement. While traditional software excels in technical accuracy, game engines enhance the visualization and interpretability of flood scenarios. The immersive experience provided by game engines can make flood dynamics more accessible to stakeholders who may not have a technical background, thereby facilitating better understanding and decision-making [55,56,57]. Additionally, the ability to modify and improve flood simulation plugins in game engines by applying methodologies from traditional precise engineering software represents a significant advancement. This combination of technical rigor and interactive features promises to deliver more sophisticated and user-friendly simulations.

3.7. Bridging the Gap between Science and Society

One of the primary advantages of the proposed framework is its potential to bridge the gap between scientific research and public understanding. Traditional flood monitoring systems and engineering software often cater primarily to technical experts and may not fully address the needs of policymakers, decision-makers, and the general public. The immersive and interactive nature of game engines can engage a broader audience by providing intuitive visualizations and simulations that convey complex flood data in an accessible manner.

For policymakers and decision-makers, the ability to interact with flood simulations and visualize potential impacts can inform better decision-making and policy development. The proposed framework enables these stakeholders to explore different flood scenarios, evaluate management strategies, and understand the implications of various decisions in a more tangible and engaging way. This enhanced understanding can lead to more informed policies and practices, ultimately contributing to improved flood management and resilience.

For the general public, the educational potential of immersive digital twins can raise awareness about flood risks and preparedness. By providing interactive and visually compelling simulations, the framework can enhance public education efforts, helping individuals and communities better understand flood dynamics and the importance of preparedness measures.

3.7.1. Designing Immersive Experiences for Diverse End-Users

Creating an immersive digital twin may target a wide spectrum of end-users and stakeholders, who can be categorized as political decision-makers, floodplain inhabitants, basin authorities, engineering students, civil protection experts, and science communicators. Each category has different objectives and requirements, which can alter the means of communication through immersive platforms. What follows summarizes a target-based education and extension routine, designed to address such distinct requirements [58,59].

• Political Decision-Makers: They need straightforward, less technical simulation procedures and, most importantly, the ability to test different flooding scenarios in an interactive environment.
• Floodplain Inhabitants: As the main residents of flood-prone areas, they need to get a good sense of flood risk, which requires additional mathematical specifications in immersive platforms while not losing realism and interactivity.
• Basin Authorities: They tend to focus on long-term planning and most definitely need technical details and collaborative support.
• Engineering Students: For educational purposes, simulations should be easy to perform with enough tutorials and available feedback.
• Civil Protection Experts: For emergency applications, simulations should provide more actionable outputs, such as possible evacuation routes, traffic congestion hotspots, and so forth. Additionally, real-time simulation is a key feature for this group.
• Science Communicators: People are involved on the other end of communication; therefore, the simulation should be persuasive, engaging, realistic, interactive, immersive, accessible, easily comprehensible, and, most definitely, visual.

To illustrate, we further detail the following statements to elucidate how an immersive digital twin experience can meet the relational need of hydraulic engineering students by providing a hands-on, real-time, and efficient simulation and education platform through which students can process their understanding of the principles associated with river dynamics. The context described is in a classroom where students are engaged in the dam design process that will control flooding using water simulation in a virtual space. The design process will have a number of parts: preparation (introduction to hydraulic theories and getting familiar with the platform), interaction (virtual tour of the river and surrounding areas), experiments (geometrical and material design of the dam), observation (real-time/reciprocal water–dam interaction, and river response as various positions are established for the dam with specific design geometry and material), and reflective discussion (discussing implications of the outcomes, the effectiveness of a number of variations in designs, and trade-offs).

• Step 1: Introduction and Familiarization

The education cycle begins with an introduction to a virtual twin of a river reach/watershed, which accurately presents the topography of the river reach/watershed, hydrology, and water flow characteristics, also augmented with real, or near real-time data that has been gathered by multiple sensors through IoT. Students will learn about how a digital twin operates with its main parts and some basic information about hydraulics and dam design.

• Step 2: Hands-on Experience with the Immersive Digital Twin

After acquiring some comfort and confidence with the previously mentioned building blocks, the students can engage in the second exercise looking at a hands-on experience inside Unreal Engine where they control specific features with respect to water flows, riverbanks, floodplains, and existing hydraulic structures. The students pick their avatar (to create engagement), suit up with their VR glasses, or use a desktop as a cross-platform immersive digital twin presenting experiences across many mediums to walk or fly across the river to check out the riverbank, slope gradients, and general morphology of the river. At specific points (for example, virtual educational canvases), they can review material embedded in the space and at different stations (text, audio, video, and/or augmented reality) to provide an overall visual representation of hydro-environmental characteristics associated with river hydraulics and the critical designs and features associated with structures such as dams and weirs, as well as navigate into the immersive environment to experience the entire area, using shortcut handles. For example, by flying or spinning across the river reach or watershed, they get a good sense of scale and layout of their study area, which naturally provides them with some notions about the overall context of their simulation, what parameters seem typical for their area, and unacceptable or unexpected results for their case. Furthermore, using the data-fused AI system, supported by an assisted language model system manifested as a guide or help desk, students can gain hydro-environmental information and science across the study area. Students are also encountering their study area from live and historical data feed perspective generated from IoT-powered digital twin in this instance, so they are further able to begin engaging with the data and observe the general impact of any data manipulation on the overall hydrological and hydraulic setting of the river—like changing flow rates and precipitation—and observe how those changes visually impact the overall hydrodynamics.

• Step 3: Design and Simulation of Hydraulic Structures

This stage represents the most important part of the educational cycle where students would engage the capabilities of various built-in tools within Unreal Engine or any other game engine, along with flow simulation plugins such as FluidFlux, to simulate and attempt various hydraulic scenarios on the river system. In the case presented here, students would be able to create dams directly in Unreal Engine or simply insert their already designed 3D model of the dam created in other software, such as AutoCAD Civil 3D 2025.1 or Autodesk Revit 2025.2 in .FBX or .OBJ file formats that would represent the same scale and geometry, and place it at desired locations along the river reach. Once completed, they can also utilize geometry editing tools in Unreal Engine to strictly control the width and height of the dam. Students would also be able to utilize the Material Editor in Unreal Engine to assign various materials to the structure of the dam, such as rock-fill, earth-fill, or concrete materials, and apply their physical parameters, such as permeability, strength, and durability, using the built-in physics engines to simulate the stress and strain of the dam, as well as visualize weak spots of the structure under extreme conditions, as with large flow conditions due to flooding. Once finalized, students would select FluidFlux to set up the fluid simulation and visualize water–dam interactions. The FluidFlux plugin in Unreal Engine allows students to set up all the key water dynamic parameters, such as flow inlets and outlets (flow boundaries) at the upstream and downstream portions of the river. The plugin automatically considers the underlying topography and allows the water to flow according to the natural gradient and contour of the river reach. All these capabilities allow students to visualize how the dam interacts with water levels, discharge, and velocity of the water upstream, at the head of the river reach, and downstream beyond the dam location.

FluidFlux uses shallow water equations (SWEs), which are a type of hyperbolic partial differential equations that are appropriate for modeling rivers, lakes, and coastal areas where horizontal scales are greater than vertical scales. SWEs are made up of two primary equations: the continuity equation (conservation of mass) and the momentum equation (conservation of momentum), both having limitations such as neglecting the vertical acceleration of fluid particles, assuming pressures have a hydrostatic distribution, and ignoring vertical momentum. Nevertheless, the SWE equations are helpful for agile simulations and classroom demonstrations. Students can simulate subcritical (slow) and supercritical (fast) flows (both suitable for deep rivers on gentle slopes and shallow rivers on steep slopes, respectively), steady-state flow (equilibrium conditions or long-term flow), unsteady flow (flood events and tidal surges), and stationary flow (constant velocity at a single point). In this way, FluidFlux can be used to simulate common flows of interest, such as riverine flooding, dam operation, and watershed hydrodynamics. Of course, modelers are also encouraged to use their own custom scripts in Unreal Engine (if using Blueprints or C++) to make specific configurations for advanced fluid simulations. For example, highly turbulent and vertical flows (e.g., highly aerated flows or breaking waves, which are very complex modeling scenarios that are represented by 3D Navier–Stokes equations), deep water waves and ocean dynamics, which require a full-3D approach, and non-hydrostatic phenomena (e.g., dam break flows or stratified flows). Based on the presented framework, this mixture of game-built plugins will permit students to understand their appropriate application and limitations when simulating events that represent real-world conditions and when additional (more advanced) models may be needed.

Once students have achieved the desired parameter setup, they can access one of the most remarkable functions of Unreal Engine combined with FluidFlux: the real-time simulation of the water-dam interaction. Students can modify dam geometry and see how it impacts water levels, discharge, and velocity before and after the dam position, in real time. Students can observe the upstream reservoir and impoundment of water that can form due to a manual increase in dam height because the software reacts instantaneously and calculates new water surface elevation based on hydrostatic pressure and energy conservation principles. In addition, FluidFlux can help students test other important aspects of water dynamics, such as overflow situations, allowing them to see the results of overtopping due to the exceedance of dam capacity, and add on activities such as simplified sediment transport while also evaluating their designed dam trap efficiency and the reduction of downstream sediment supply.

• Step 4: Analysis, Feedback, Psychological Impact, and Educational Outcomes

After a number of iterative simulation sessions, groups of students can then engage in a collaborative learning environment and compare their visualization experience inside the immersive digital twin with accepted engineering standards. The simulation outputs in the immersive digital twin can be further scripted for use inside Unreal Engine and FluidFlux to satisfy students’ diverse needs, such as graphical reporting with detail. The main psychological impacts from the experience are bestowed by the immersive, interactive, and realistic simulation, which enhances their knowledge absorption and retention. A gamification platform that allows students to visualize and manipulate complex hydraulic processes, in real time, through tasks and challenges, can provide a more intuitive grasp of concepts central to hydraulic engineering.

3.7.2. Tailoring Immersive Experiences to Various Flood Types

Floods constitute different types, each bearing distinct characteristics, from the cause to the duration, impact extent, and encompassed material matrix. The following discussion is provided to address how different types of floods can dictate which immersive simulation features should be highlighted [60].

• Flash floods are mostly characterized by sudden onset and localized impact. Hence, real-time simulation and support of high-resolution DEM should be taken into account during the building of an immersive digital twin with simulation capability.
• Debris flows are a complex matrix of different materials and water. In order to design an accurate immersive experience, the tool should be built upon rigorous physics for simulation of the interaction between water, mud, tree trunks, and boulders. In addition, due attention should be paid to erosional processes and slope stability on steep slopes, which can be conveyed with the support of high-resolution DEM accompanied by soil/lithology data. Different erosion brushes in Unreal Engine and other pre-game development environment creation platforms, such as TerreSculptor, showcase such successful and thoughtful integrations.
• Dam breaks are characterized by a sudden release of an immense volume of water with high energy. The tool should adhere to structural impact analysis and catastrophic failure mechanisms, which have been discernibly designed and modified through the years in Chaos Physics tools embedded in Unreal Engine.
• River plain flooding marks widespread and long-term inundation, which dictates a long duration of computational process and simulation across large areas. Such specification is integrated into game engines through their coverage of large topography data, yet there is room to incorporate more specific simulation tools or amend the existing ones by adding more in-depth options such as soil saturation, agricultural damage analysis, and resilience of the local community and infrastructures.
• Coastal flooding is mostly caused by tides and storm surges and impacts coastal infrastructures. Hence, tide and surge modeling and infrastructure impact analysis should be addressed. Recent fluid simulation plugins in Unreal Engine, most notably FluidFlux, well integrate tidal wave simulation.
• Pluvial flooding caused by intense rainfalls mostly exerts a short-lived impact on urban areas and drainage systems. Such events specify the support of high-resolution 3D maps of urban areas and subsurface and surface water simulation and interaction. Thanks to the recent mutual collaboration between geographic information system companies and game development platforms, most notably Cesium and Unreal Engine, geospatial elements and real-world landscapes (e.g., Google photorealistic 3D tiles) are brought into virtual world creation. The latter can integrate more specific geospatial data, such as urban drainage systems, and, with a decent fluid simulation plugin in a game engine, an immersive experience of simulating and exhibiting a pluvial flood would be possible.

3.8. Integrating Educational, Training, and Psychological Dimensions in Immersive Water Simulations

Simulations are powerful tools for education, training, and decision-making. Several factors play a key role in imparting knowledge and developing wisdom, especially the type of training, design of the simulation environment, and psychological processes involved during the learning [61,62]. What follows is a discussion on the key importance and the way these factors can lead to gaining knowledge and developing wisdom from immersive water simulation tools, most notably gamified digital twins.

3.8.1. Educational Environments

Nowadays, simulation is widely used in courses in universities and training institutions in terms of formal education. Through different sessions of simulation, students can put their theoretical knowledge into practice by injecting them with a sense of management role and deepening their understanding of floods, water simulation, and water management. It is also considered a curriculum integration that supplements traditional lecturing materials with more novel pieces and creates experiential learning in a controlled observation and testing environment. As a result, such an interactive environment can trigger knowledge retention in students where, in addition to traditional training materials, they can manipulate parameters, run multiple simulations through hands-on tools, and, consequently, gain a reinforced conception and understanding of the phenomenon. It also stimulates a sense of critical thinking where students observe the consequences of their actions through feedback and reflection loops during simulation sessions, promoting more informed decisions.

3.8.2. Supervised Training

Supervised training can be gained through guided learning by expert involvement and scenario-based learning. Instructors’ presence alongside the students is essential to guide them through complex simulations, let them digest the understanding of the nuances of the model, and seamlessly provide insights into the decision-making process. Also, scenario-based learning can be achieved through the instructor by gaining navigation, guidance, and feedback loops, which activate students’ problem-solving skills and teach them navigating such skills in real-world intricate situations. Supervision can simultaneously provide students with a safe alternative training environment and encourage them to critically analyze their thoughts and put them to the test by making courageous and sound decisions, first in the simulation and later on, in life. Additionally, the instructor can evoke competence and confidence among students by creating tailored feedback mechanisms and challenging yet controlled simulation and education environments.

3.8.3. Non-Supervised Training

This process involves autonomy and self-directed learning by letting students learn at their own pace and injects a sense of independence through the procedures of experiments, testing hypotheses, and gaining insights. Such insight is rooted deep in discovery and the repetitive process of success and failure, and in return, flourishes a deep understanding of complex systems. Going through multiple trial-and-error simulation sessions, students gain resilience, adaptability, and persistence in refining their decision-making habits. It is noteworthy that introspection and self-reflection are key processes in developing wisdom since they create critical thinking about decisions and the reasoning behind them.

3.8.4. Psychological Processes in Learning

Here we discuss three pivotal psychological phenomena that are important to understand before executing simulation and education sessions: cognitive load, engagement and motivation, and emotional involvement. As such, cognitive load can be a result of a continuous and hasty information flow and can hinder knowledge absorption and retention. Hence, gradual learning processes are in favor while training students on how to manage cognitive loads during prompt and complex situations (e.g., building immersive flood situational awareness platforms). This can also be carried out via information chunking where information is delivered to students in a stage-wise manner, allowing them to digest small pieces and learn foundational information gradually. Motivation and engagement can be achieved through different tools, most notably gamification. A gamified simulation with tasks and challenges, rewarding, and progress tracking can stimulate a sense of motivation and engagement and avail in experiencing an enjoyable deeper learning. It is also noteworthy that simulations should be specifically designed with relevance to their target community because they can trigger intrinsic motivation, leading to deeper engagement and gaining more meaningful learning outcomes and insights. Regarding emotional involvement, it is essential to make the simulation as impactful and evocative as possible by teaching students the real-world impacts of flood events. By doing so, the simulation experience becomes memorable, which tremendously enhances the learning process. On the other side of the emotional spectrum, simulations should also mimic stressful or contested situations such as emergency response scenarios. This is particularly essential for professionals who should deal with similar situations in real life and helps them make vital decisions under stress and learn how to stay calm and imperturbable.

4. Challenges and Limitations

4.1. Data Quality and Integration

One of the primary challenges in implementing a digital twin for flood monitoring is ensuring the quality and resolution of input data. Incomplete or low-resolution data can affect the accuracy and reliability of the virtual replica. The integration of diverse data sources, including photogrammetry, IoT sensors, and 3D repositories, requires careful management to maintain coherence and compatibility. Addressing these challenges is crucial for developing a robust and effective flood monitoring system.

4.2. Technical Complexity

The integration of game engines with digital twins introduces technical complexities, particularly in terms of rendering high-fidelity graphics and managing real-time data. While the theoretical benefits are clear, practical implementation involves navigating issues related to performance, data processing, and system integration. The use of advanced features like Lumen, Nanite, and the MQTT plugin represents a significant technical undertaking that requires thorough testing and optimization.

4.3. Educational and Public Outreach

While the immersive nature of game engines offers the potential for enhancing educational outreach and public engagement, the effectiveness of these tools in practice depends on their accessibility and usability. Ensuring that interactive simulations are user-friendly and effectively communicate flood risks and management strategies is essential for achieving the desired impact.

5. Future Works

Future research should focus on refining data integration methods, enhancing simulation realism, and expanding the application of advanced analytics. Collaboration with stakeholders will be crucial in optimizing the framework and ensuring its effectiveness in real-world scenarios. Addressing technical complexities, data quality issues, model validation, and ensuring user accessibility will be key areas for development. By leveraging the strengths of digital twins and game engines, this study aims to contribute to more effective and informed flood management strategies. The proposed framework has the potential to reshape how flood dynamics are understood and managed, ultimately fostering more resilient and prepared communities.

6. Conclusions

This study has proposed an innovative framework for integrating digital twins and game engines to advance flood education, data collection and visualization, and interactive monitoring and simulation. The integration of these technologies offers a transformative approach to flood management by addressing existing limitations and enhancing stakeholder engagement. Through the creation of immersive digital twins, we have demonstrated how high-fidelity virtual replicas of watersheds can provide deeper insights into flood dynamics, leading to more informed decision-making and improved flood preparedness. Key contributions of the proposed conceptual framework are listed as follows.

Enhanced Visualization and Interaction: The use of game engines, such as Unreal Engine and Unity, significantly improves the visualization of flood scenarios. The interactive features of these engines facilitate a more intuitive understanding of flood behavior, enabling users to explore various scenarios and management strategies in a dynamic and engaging environment.

Real-Time Data Integration: The incorporation of real-time data from IoT sensors through tools like the MQTT plugin ensures that the digital twin remains current with live conditions. This integration enhances the accuracy and relevance of the virtual model, providing up-to-date information essential for effective flood monitoring and response.

Educational Impact: The immersive nature of the proposed framework has the potential to greatly enhance public education and awareness. By providing interactive simulations, we can better communicate complex flood dynamics to non-specialists, empowering communities to take proactive measures and improve resilience.

The successful integration of digital twins and game engines in flood management not only advances current practices but also sets a precedent for future applications in environmental monitoring. This approach can bridge the gap between scientific research and practical implementation, offering stakeholders and decision-makers more accessible and actionable insights.