A rapid 3D reproduction system of dam-break floods constrained by post-disaster information

Highlights

• Compared with other simulation systems, the rapid dam-break flood reproduction system has the following advantages:

• First, the reproduction time of the flood process is reduced to less than 1 h. The simulation of the spatiotemporal process of a dam-break flood is a very complex process, and abundant input data and parameters are needed to achieve accurate simulation. The current flood system focuses on the mechanism and simulation accuracy and avoids a large time investment in data processing and simulation, which is inconsistent with the goal of rapid response to flood disasters. The rapid flood reproduction system simplifies the input parameters of the model and reduces the model calculation time. Furthermore, the system integrates input data processing, spatiotemporal simulation and augmented representation into a complete workflow, reducing the entire reproduction time to one hour level and providing decision-making support for disaster response.

• Second, this system supports rapid 3D reproduction of flood processes at different times and positions. A flexible parameter configuration is provided. The cross-section data, flow data, and dry and wet depth are required to set in the current flood simulation software. The system proposed in this paper has a simplified parameter configuration; users only need to set several necessary parameters, such as the DEM, roughness and storage capacity. Furthermore, on this system, users can quickly visualize dam-break flood disasters at different times and locations. In addition, different from the current flood systems offered to professionals, the rapid dam-break flood system provides a user-friendly interface, hides potential technical details such as model calculations, and reduces user requirements with respect to professional background and experience.

• Third, augmented representation improves the transmission efficiency of flood disaster information. In the flood disaster management, users pay more attention to flood disaster information than to visual effects. Augmented representation allows visualization of the cause, process and result of dam-break floods through the combination of static representation and dynamic representation and a logical sequence, which can enhance the transmission of flood disaster information and the user's perception level.

Abstract

The 3D reproduction of dam-break floods plays a key role in emergency management and disaster information dissemination. However, existing systems for the simulation of floods have some deficiencies. First, the input data parameters are complex. Second, the simulation process is cumbersome and time consuming, and it lacks a complete workflow. In this paper, we develop a platform for the rapid 3D reproduction of dam-break floods. Key technologies, such as numerical models of dam-break floods based on cellular automata (CA), 3D reproduction constrained by post-disaster information, and dynamic augmented representation, are discussed in detail. Then, two typical dam-break flood cases are selected for experimental analysis. The experimental results show that the system can support the rapid 3D reproduction of dam-break floods constrained by post-disaster information. Compared with that of other simulation systems, the reproduction time is reduced to less than 1 h, and this system supports rapid reproduction at different times and positions.

Introduction

The frequency and intensity of global extreme events have increased dramatically due to climate change and rapid economic growth. Floods are the most common and widespread type of natural disasters (Patel and Srivastava, 2013; Wan et al., 2014; Qiu et al., 2017). Dam-break floods are flood disasters that are usually associated with barrier lakes and obsolete reservoirs and are characterized by sudden occurrence, rapid expansion, high velocity and severe destruction (Li et al., 2013). As dam-break floods are a complex geographical phenomenon, the spatiotemporal process is often more important than the final spatial pattern, only the development process is clearly understood, and the deep-level change rules merit further exploration (Li et al., 2009).

Simulations have become a fundamental method for gaining insight into complex geographic processes (Bainbridge 2007; Chen et al., 2020). Moreover, developments in information science, earth observation technologies and geographical information science have strengthened the connections among geographical data, models and visualization, providing new perspectives and opportunities for spatiotemporal process modeling and the visualization of dam-break floods (Goodchild 2004; Li 2012; Chen et al., 2015; Achu et al., 2020; Kaur et al., 2019; Lü et al., 2019).

Many types of research have been conducted on flood simulation, with studies employing 1D hydraulic models, 2D hydraulic models, 1D-2D coupled hydraulic models and hydrological models. Most work has focused on simulation prediction and early warning of flood processes under ideal conditions (Ernst et al., 2010; Domingo et al., 2010; Pasquier et al., 2018; Wang et al., 2018). In practical use, it is difficult to use predictive models due to the uncertainty of data parameters, the instability of model structure, and the complexity of external conditions. Consequently, there is no universal method or software that can be applied to accurately predict floods under different conditions.

Disaster reproduction is a restored representation of the reason the disaster occurred, the process and the results. Whereas disaster simulation is forward-looking and focuses on the prediction of disaster evolution process, disaster reproduction emphasizes the restored representation of the disaster spatiotemporal process based on post-disaster information. The latter is mainly used for post-disaster education and media promotion, an intuitive and detailed 3D reproduction of the disaster process can improve the transmission efficiency of disaster information. Thus, disaster reproduction can overcome the shortcomings of traditional disaster education materials, which overly rely on ‘image and text publicity’, and ultimately improve people's disaster risk awareness and disaster prevention capabilities. After a dam-break flood, post-disaster information, such as the dam-break location, storage capacity, and affected area and buildings, can be used to help improve the simulation model of disasters (Li et al., 2015), thereby improving the spatiotemporal reproduction accuracy for this flood type. Moreover, transferring the disaster risk knowledge implied in numerical results to decision makers and the general public is a difficult task. The 3D virtual environment can intuitively display the simulation results of floods, which can improve users' perception and cognitive understanding of a dam-break flood disaster (Cheng et al., 2019; Chen et al. 2013, 2015, 2018; Havenith et al., 2019; Bailey and Chen, 2011; Lin et al., 2013a). Existing flood simulation software (e.g., FLO2D, HEC-RAS, etc.) emphasizes risk prediction, and problems with its application include tedious data processing, long simulation times and a high degree of specialization. Therefore, a rapid dam-break flood reproduction system that integrates data processing, simulation calculation and dynamic representation is urgently needed to support emergency decisions and the dissemination of risk knowledge.

This paper outlines the design and implementation of a rapid dam-break reproduction system. The system integrates a dam-break numerical model based on the cellular automata (CA) and the virtual geographic environment (VGE) framework, realizes rapid reproduction and augmented representation of a dam-break flood under the constraints of post-disaster information, and provides decision-making information support and risk knowledge for disaster-related persons. To realize the high-efficiency reproduction of a whole dam-break flood process, this system integrates data acquisition, model calculation and dynamic visualization functions. Through a user-friendly operation interface and flexible parameter configuration, the whole process of a dam-break flood can be represented in virtual 3D view. A dam-break flood that occurs anytime and anywhere can be rapidly reproduced by this system, and the reproduction time is reduced to less than 1 h, which greatly improves the 3D restoration efficiency of the spatiotemporal evolution of disasters.