Flood Types Research Articles

Unveiling flood-generating mechanisms using circular statistics-based machine learning approach without the need for discharge data during inference

Abstract Understanding the drivers of flooding is essential for flood disaster prevention. However, conventional flood prediction methods are hindered by their reliance on local discharge data, which can be constrained by limited spatial resolution. To address this limitation, we present a machine learning model that can categorize floods without requiring discharge data during inference. We first use circular statistics to calculate the relative importance of three candidate flood-generating mechanisms. Global land areas are classified into three primary categories and eight sub-categories based on the proportion of relative importance. A random forest model is then applied to identify the flood types by assuming that the discharge data is unavailable. The findings from circular statistics highlight that globally, soil moisture excess is the most influential driver of floods followed by extreme precipitation and snowmelt, with an average relative importance of 0.535, 0.387, and 0.078, respectively. The RF model performs well in resembling the three primary flood categories with an accuracy of 0.701 and a F1-score of 0.692 in 10-fold cross-validation. The trained gridded-based model provides a swift and efficient approach for analyzing flood mechanisms, even in limited discharge scenarios, allowing for rapid insights.

Rising seas and roadway debris: Microplastic and low-density tire wear particles in street-associated tidal floodwater

Tidal flooding is increasingly common in low-lying coastal regions as sea levels rise. This type of flooding can occur on sunny days with no rainfall and may transport street-associated debris, such as microplastics (MPs) including tire wear particles (TWPs), to coastal systems. This research aimed to quantify MP abundance in tidal floodwater and investigate their fate. Three locations around Charleston, SC (USA) were sampled during 12 tidal floods, and their adjacent tidal creeks were sampled before and after 5 floods. Floodwater contained an average of 342 ± 60 MP/L. Most MPs in floodwater were low-density TWP (86.5 %). MP abundance in tidal creek surface water following flooding did not change, suggesting that MPs were not immediately transferred to coastal waterways but deposited in adjacent marsh sediment. Elucidating transport routes of MPs in coastal environments is critical to understanding and preventing this type of contamination in the face of a changing climate.

Equivalent Relationship between Flood Prevention Storage of Cascade Reservoirs in the Downstream Jinsha River and Three Gorges Reservoir

The equivalent relationships between flood prevention storage of a cascade reservoir system provide useful information for real-time operation. However, those relationships are implicitly hinted at in the optimized operation rules of downstream reservoirs, which may add difficultly in providing feasible operation strategies for engineering practice. In this study, we investigate the equivalent relationship between cascade reservoirs using a flood regional composition-based framework coupled with reservoir inflow and river flow routing techniques. We select the cascade reservoirs in the downstream Jinsha River and Three Gorges Reservoir (TGR) as the case study. Results reveal that (1) the equivalent relationship between flood prevention storage of cascade reservoirs in the downstream Jinsha River and TGR is approximately linear-related; (2) the conversion coefficient of equivalent relationship depends on the flood type, and the intercept term of equivalent relationship is mainly influenced by the design flood frequency; and (3) the TGR flood prevention storage (or water level) can be dynamically adjusted according to the corresponding reserved flood prevention storage of cascade reservoirs in the downstream Jinsha River, which can generate 3.56 billion kW·h (+8.5%) extra hydropower during flood season without increasing flood control risks.

Advancing Flood Prediction: Leveraging Machine Learning for Accurate Prediction

Abstract: Floods, a force of nature that cannot be escaped, relentlessly strike without discrimination. Every nation, to shield itself from this calamity, strives to make adequate preparations. Countless countries endure the devastation of floods annually, suffering immeasurable losses in various aspects such as finance, environment, and human lives. Regrettably, these losses are irreparable, no matter the efforts made. Amongst the different types of floods, flash floods pose the gravest danger as they possess both the destructive force of a flood and an astonishing velocity. These flash floods occur when intense rainfall surpasses the soil's ability to absorb water. They also occur when water fills dry creeks or streams or when sufficient water accumulates to breach riverbanks, resulting in rapid surges in a short period. Flash floods can materialize within minutes after triggering incidents, leaving little time for public warning and protection. As one facet of artificial intelligence, machine learning greatly contributes to the progression of predictive systems, offering enhanced performance and cost-effective solutions. Flood prediction must be accurate and fast to mitigate its damage. Machine learning algorithms have gained popularity for flood prediction recently because of their capacity to analyse massive volumes of data and find subtle patterns. This article presents an overview of flood prediction technology utilizing machine learning, encompassing the diverse algorithms employed, the benefits and drawbacks of this technology, and the challenges that lie ahead along with potential future developments.

A Systematic Review of Disaster Management Systems: Approaches, Challenges, and Future Directions

Disaster management is a critical area that requires efficient methods and techniques to address various challenges. This comprehensive assessment offers an in-depth overview of disaster management systems, methods, obstacles, and potential future paths. Specifically, it focuses on flood control, a significant and recurrent category of natural disasters. The analysis begins by exploring various types of natural catastrophes, including earthquakes, wildfires, and floods. It then delves into the different domains that collectively contribute to effective flood management. These domains encompass cutting-edge technologies such as big data analysis and cloud computing, providing scalable and reliable infrastructure for data storage, processing, and analysis. The study investigates the potential of the Internet of Things and sensor networks to gather real-time data from flood-prone areas, enhancing situational awareness and enabling prompt actions. Model-driven engineering is examined for its utility in developing and modeling flood scenarios, aiding in preparation and response planning. This study includes the Google Earth engine (GEE) and examines previous studies involving GEE. Moreover, we discuss remote sensing; remote sensing is undoubtedly a valuable tool for disaster management, and offers geographical data in various situations. We explore the application of Geographical Information System (GIS) and Spatial Data Management for visualizing and analyzing spatial data and facilitating informed decision-making and resource allocation during floods. In the final section, the focus shifts to the utilization of machine learning and data analytics in flood management. These methodologies offer predictive models and data-driven insights, enhancing early warning systems, risk assessment, and mitigation strategies. Through this in-depth analysis, the significance of incorporating these spheres into flood control procedures is highlighted, with the aim of improving disaster management techniques and enhancing resilience in flood-prone regions. The paper addresses existing challenges and provides future research directions, ultimately striving for a clearer and more coherent representation of disaster management techniques.

Evaluating the spatial application of multivariable flood damage models

AbstractFlood damage arises from complex interactions between flooding processes and socio‐economic elements. Damage assessments for elements such as residential buildings rely on a modelled representation of local damage factors. Multivariable model approaches are well suited for damage prediction using detailed information on flood hazard and building characteristics. While broad explanatory variable ranges can improve model prediction performance, model transfer across geographical contexts often causes performance loss. This study aims to determine if increasing explanatory damage variables in a multivariable model improves residential building damage prediction and whether models based on local variables transfer between locations. We used empirical damage observations from six flood events to train and evaluate random forest regression model prediction performance. Spatial transfer is tested by splitting event datasets with trained models applied to original and external events. Variable analysis demonstrates model performance improvement with up to seven flood hazard and building characteristics, decreasing thereafter. Event models showed highest prediction precision for the original event, while models trained on all events transfer with comparable predictions for urban stormwater flooding. Prediction precision reduces when models transfer between locations affected by different flood types. This indicates flood damage models must replicate variability in local damage factors for reliable spatial transfer.

Co-occurrence of pluvial and fluvial floods exacerbates inundation and economic losses: evidence from a scenario-based analysis in Longyan, China

Hydrodynamics and socio-economic impacts of compound floods as the result of co-occurring pluvial and fluvial floods, have not been well studied, which challenges conventional urban flood risk management that treats different types of floods separately. This study generates a high-resolution land surface digital surface model based on images from unmanned aerial vehicles, and incorporates it into a combined flooding model that integrates the 1D river hydrodynamic model, 1D urban drainage hydrodynamic model and 2D overland flow model, in order to simulate the inundation characteristics of extreme floods when local pluvial floods coincide with upstream fluvial floods, and in a further step to quantify the direct economic losses on urban physical systems. As demonstrated by an empirical study of Longyan, China, the combination of pluvial and fluvial floods leads to substantial exacerbation in inundation extent and depth. Consequently, the joining of pluvial and fluvial floods greatly amplifies economic losses to urban physical systems, indicating that ignoring co-occurrence of different types of floods will potentially lead to underestimation of flood risk and insufficient adaptation. Our study emphasizes the need to account for the co-occurrence of multiple types of floods in flood risk assessment and management, to avoid the shortcomings of fragmented responses.

Flood Simulations Using a Sensor Network and Support Vector Machine Model

This study aims to couple the support vector machine (SVM) model with a hydrometeorological wireless sensor network to simulate different types of flood events in a montane basin. The model was tested in the mid-latitude montane basin of Vydra in the Šumava Mountains, Central Europe, featuring complex physiography, high dynamics of hydrometeorological processes, and the occurrence of different types of floods. The basin is equipped with a sensor network operating in headwaters along with the conventional long-term monitoring in the outlet. The model was trained and validated using hydrological observations from 2011 to 2021, and performance was assessed using metrics such as R2, NSE, KGE, and RMSE. The model was run using both hourly and daily timesteps to evaluate the effect of timestep aggregation. Model setup and deployment utilized the KNIME software platform, LibSVM library, and Python packages. Sensitivity analysis was performed to determine the optimal configuration of the SVR model parameters (C, N, and E). Among 125 simulation variants, an optimal parameter configuration was identified that resulted in improved model performance and better fit for peak flows. The sensitivity analysis demonstrated the robustness of the SVR model, as different parameter variations yielded reasonable performances, with NSE values ranging from 0.791 to 0.873 for a complex hydrological year. Simulation results for different flood scenarios showed the reliability of the model in reconstructing different types of floods. The model accurately captured trend fitting, event timing, peaks, and flood volumes without significant errors. Performance was generally higher using a daily timestep, with mean metric values R2 = 0.963 and NSE = 0.880, compared to mean R2 = 0.913 and NSE = 0.820 using an hourly timestep, for all 12 flood scenarios. The very good performance even for complex flood events such as rain-on-snow floods combined with the fast computation makes this a promising approach for applications.

An integrated operations model for supply of relief items under path disruptions

The integration of relief supply chain operations is critical for the timely supply of relief items to the areas affected by disasters. This work considers a two stage relief supply chain providing relief kit to the areas affected by flood type disaster. Floods often cause path blockages that affect the supply of relief items. A mixed integer linear programming model is formulated for the integrated production and outbound distribution scheduling problem considered and is solved by using a mathematical solver. Since solvers have computational issues in solving realistically sized problems, a Selective Hybrid Genetic Algorithm (SHGA), which performs a selective local search, is proposed. The proposed SHGA produces near optimal solutions in relatively short computation times which is beneficial for quick decision making. A sensitivity analysis is also performed and the path disruption is found to be the most significant input parameter influencing the model.

Generation of type-specific synthetic design flood hydrographs

ABSTRACT Design flood hydrographs are essential for hydraulic engineering and flood risk assessment. Typically, they are derived from quantiles estimated from flood frequency analyses of flood peaks. This usually does not take into account that one and the same flood peak can occur with different volumes and various hydrograph shapes due to different meteorological and catchment-specific processes. In view of this issue, we generate design flood scenarios, i.e. type-specific design hydrographs are created for a flood quantile of a certain return period based on type-specific flood volumes and hydrograph shapes. Additionally, the probability of each flood type is derived. Different probability density functions are used as templates for generating different hydrograph shapes by varying their parameters. An ensemble of synthetic design flood hydrographs is obtained, which allows the consideration of best- and worst-case scenarios. A case study for 162 catchments in Germany demonstrates the applicability of the approach.

Flood forecasting based on machine learning pattern recognition and dynamic migration of parameters

Study regionTypical basin in semi-arid and semi humid areas in the middle reaches of the Yellow River Study focusFloods are among the most devastating natural disasters. Timely and accurate forecasting of runoff is crucial to safeguard human lives, minimize property damage, enhance the efficiency of reservoir power generation, and ensure the safety of water supply. In this study, the rainfall and flow data of 98 floods occurring between 1971 and 2014 in the Jingle sub-basin, a tributary of the Yellow River basin, China, were analyzed using dynamic clustering and random forest techniques to identify flood types and select appropriate model parameters. The Xin’anjiang model was then used for real-time flood forecasting. The results indicate that the rainfall characteristic indicators developed by the model can effectively identify potential flood types, and the model parameters determined by historical flood rates can be adapted and utilized for new forecasting tasks based on similarities. Ensemble forecasting results, which consider the probability of flood types, are superior to single fractal forecasting outcomes and diminish uncertainty. The proposed method can identify extreme flood events, facilitate flood classification and prediction, promote basin disaster mitigation, and enable the efficient use of water resources. New hydrological insights for the regionThe proposed flood classification and identification method effectively analyzes flood events in the basin. The floods that may occur are characterized by relevant statistical characteristics of rainfall, allowing for selection of corresponding flood forecasting model parameters for improved accuracy in real-time flood forecasting and extended prediction period.

Spatial distribution of socio-demographic and housing-based factors in relation to flash and slow-rise flooding hazards in the U.S

Previous studies have drawn attention to racial and socioeconomic disparities in exposures associated with flood events at varying spatial scales, but most of these studies have not differentiated flood risk. Assessing flood risk without differentiating floods by their characteristics (e.g. duration and intensity of precipitation leading to flooding) may lead to less accurate estimates of the most vulnerable locations and populations. In this study, we compare the spatial patterning of social vulnerability, types of housing, and housing tenure (i.e. rented vs. owned) between two specific flood types used operationally by the National Weather Service—flash floods and slow-rise floods—in the floodplains across the Contiguous United States (CONUS). We synthesized several datasets, including established distributions of flood hazards and flooding characteristics, indicators of socioeconomic status, social vulnerability, and housing characteristics, and used generalized estimating equations to examine the proportion of socially vulnerable populations and housing types and tenure residing in the flash and slow-rise flood extents. Our statistical findings show that the proportion of the slow-rise flooded area in the floodplains is significantly greater in tracts characterized by higher percentages of socially vulnerable. However, the results could not confirm the hypothesis that they are exposed considerably more than less vulnerable in the flash flooded floodplains. Considering housing-occupancy vulnerability, the percentage of renter-occupancies are greater in the flash flood floodplains compared to slow-rise, especially in areas with high rainfall accumulation producing storms (e.g. in the Southeast). This assessment contributes insights into how specific flood types could impact different populations and housing tenure across the CONUS and informs strategies to support urban and rural community resilience and planning at local and state levels.

The Impact of Flooding on Snail Spread: The Case of Endemic Schistosomiasis Areas in Jiangxi Province, China.

Flooding is the main natural factor in snail diffusion, and it has a negative impact on schistosomiasis transmission. There are few studies on the spread and migration of snails following a flood; therefore, we aimed to investigate the influence of flooding on snail diffusion and explore the characteristics and laws of snail diffusion in Jiangxi Province. By using a retrospective survey and cross-sectional survey, the data on snail spreading in Jiangxi Province from 2017 to 2021 were collected. The distribution, nature, and area of snail spread were systematically analyzed in combination with the hydrological situation, types of region, and types of flood. From 2017 to 2021, a total of 120 snail-spread environments were found, including in 92 hilly areas and in 28 lake areas. The areas caused by flood and by other means numbered 6 and 114, respectively. The proportions of recurrence, expansion, and first-time occurrences were 43.42%, 38.16%, and 18.42%, respectively, and the 14 new snail environments were only distributed in the hilly areas. With the exception of 2018, the ratio of snail-spread areas in the hilly region was higher than that in lake region in other years. The average density of live snails was 0.0184-1.6617 no./0.1 m2 and 0.0028-0.2182 no./0.1 m2 in the hilly region. Among the 114 environments affected by floods, 86 consisted of hilly environments, including 66 spreading environments affected by rainstorm floods, and 20 rainstorm debris flow environments. There were 28 lake areas, of which 10 were in the Jiangxi section of Yangtze River and were affected by rainstorm floods. Snail spread following flooding has a certain 'lag effect,' and = simple annual changes in hydrological characteristics have little effect on the diffusion of snails or on their density = in the affected environment, but it is more closely related to local floods. The hilly environments are more susceptible to floods than the lake region, and the risk of snail spread is much higher in the hilly than in the lake region.

Remote Sensing and GIS tools for Forest Resource Management-A Review

Global importance and thus need to quantify the status of forest dynamics is increasing day by day. It is more so in the case of a country like India which is on the deficit side of forest cover in countering ecological balance. Information required for forest resource management includes extant, type of forest, felling/cutting, deforestation, afforestation, forest fire, forest flood, and extraction of bio-physical parameters like age of plantation, total biomass, canopy density etc. Widespread, vast span and difficult terrain, and inaccessibility makes it difficult to collect timely and periodic information about above mentioned parameter through conventional means. Forest mapping is one of the fields that witnessed the earliest application of remote sensing and GIS for nonmilitary purposes. using remote sensing data and GIS techniques, a forest manager can generate information regarding forest cover, types of forest present within an area of interest, human encroachment extent into forest land/protected areas, encroachment of desert-like conditions and so on [3].For the purpose of understanding the working of remote sensing and GIS, the basics are explained below.

Geomorphic Changes of the Scott River Alluvial Fan in Relation to a Four-Day Flood Event

A four-day glacier-melt flood (13–16 August 2013) caused abrupt geomorphic changes in the proglacial gravel-bed Scott River, which drains the small (10 km2) Scott Glacier catchment (SW Svalbard). This type of flood occurs on Svalbard increasingly during periods of abnormally warm or rainy weather in summer or early autumn, and the probability of occurrence grows in direct proportion to the increase in temperature and/or precipitation intensity. In the summer of 2013, during the measurement season, the highest daily precipitation (17 mm) occurred on 13 August. During the following four days, it constituted in total 47 mm, i.e., 50% of the precipitation total for the measurement period of 2013. The largest flood in 20 years was caused by high precipitation with a synchronous rise in temperature from about 1.0 to 8.6 °C. These values exceeded multi-year averages (32 mm and 5.0 °C, respectively) at an average discharge of 0.9 m3/s (melt season mean 1986–2011). These conditions caused a rapid and abrupt response of the river with the dominant (90%) glacier-fed. The increase in discharge to 4.6 m3/s, initiated by the glacial flood, mobilized significant amounts of sediment in the river bed and channel. Geomorphic changes within the alluvial fan as an area of 58,940 m2, located at the mouth of the Scott River, were detected by multi-sites terrestrial laser scanning using a Leica Scan Station C10 and then estimated using Geomorphic Change Detection (GCD) software. The changes found involved 39% of the alluvial fan area (23,231 m2). The flood-induced total area of lowering (erosion) covered 26% of the alluvial fan (6035 m2), resulting in the removal of 1183 ± 121 m3 of sediment volume. During the final phase of the flood, two times more sediment (1919 ± 344 m3) was re-deposited within the alluvial fan surface, causing significant aggradation on 74% of its area (17,196 m2). These geomorphic changes resulted in an average lowering (erosion) of the alluvial fan surface of 0.2 m and an average rising (deposition) of 0.1 m.

Social media data extraction for disaster management aid using deep learning techniques

The process of minimizing the damage being caused by disasters through information gathering, information sharing, disaster planning, problem-solving, and decision-making is known as disaster management. Social media generates a lot of information that can be used to provide disaster relief organizations with crucial information. The method under discussion is data extraction from social media during disasters, as social media generates a lot of complex data, much of which is not effectively utilized. Therefore, this study offers a technique for obtaining information from Twitter, the primary social media source, to locate and distribute the crucial aid required in emergency and catastrophe situations.Two distinct classification problems are covered in the study. Firstly, preprocessing the classification of images that are relevant to our inquiry is the first task and secondly, classifying four different types of natural disasters, including earthquakes, floods, cyclones, and wildfires. In this study, we employ Convolution Neural Network (CNN) designs along with the help of Transfer Learning (TL) architectures, wherein the model is trained and tested using the secondary data set. Furthermore, real-time tweets and images that are being extracted from Twitter are validated on the trained models and the accuracy is noted. Additional data pretreatment methods like image augmentation have also been used for preprocessing. Transfer Learning, bottleneck feature of inceptionV3 and a fine-tuning model have been included, following the disaster classification through the CNN model as a means to improve our accuracy up to 98.14%, and attaining 0.82 Precision, 0.86 Recall, and 0.84 F1-Score for Cyclone; 0.96 Precision, 0.89 Recall, and 0.92 F1-Score for Earthquake; 0.74 Precision, 0.95 Recall, and 0.83 F1-Score for Flood and 0.97 Precision, 0.96 Recall, and 0.96 F1-Score for Wildfire classification. Disaster can manifest itself in different forms and generally have a negative impact on the biosphere, resulting in loss of property and the environment. This study aims to effectively exercise the power of social media data for disaster management aid.

Strategy of Rice Productivity Increment Through Agronomical Treatment at Two Flooding Types of Tidal Swampland

The research objective is to increase rice productivity using a ratoon system through N fertilizer and growth regulator treatments at two flooding types of tidal swampland. It was conducted from February to August 2020. The experiment design used in this research is Split Plot Design with three replications. The main plot is N fertilizer and rice variety (NV) which consisted of N1V1= N fertilizing, 1/3 dose during planting + 1/3 dose at primordia phase + 1/3 dosis at harvest for Inpari 30 variety; N1V2= N fertilizing, 1/3 dose during planting + 1/3 dose at primordia phase + 1/3 dose at harvest for Hipa 5 Ceva variety; N2V1= N fertilizing, 1/3 dose during planting + 1/3 dose at primordia phase + 1/6 dose at harvest +1/6 dose at 21 days after harvest for Inpari 30 variety; N2V2= N fertilizing, 1/3 dose during planting + 1/3 dose at primordia phase + 1/6 dose at harvest +1/6 dose at 21 days after harvest for Hipa 5 Ceva variety. The subplot is the application of growth regulator GA3 (G) consisting of G0 = without application of GA3; G1 = application of GA3 once in 7 days; G2 = application of GA3 once in 14 days. The research results showed that N fertilizing, 1/3 dose during planting + 1/3 dose at primordia phase + 1/3 dosis at harvest and application of growth regulator GA3 60 ppm with interval of once in 7 days to ratoon rice is capable to increase rice productivity using ratoon rice system at B and C flooding types at tidal swampland. How to Cite: Gribaldi, Nurlaili, Novriani, S. Diana, N. Marlina, 2023. "Strategy of Rice Productivity Increment Through Agronomical Treatment at Two Flooding Types of Tidal Swampland." Journal of Agriculture and Crops, vol. 9, pp. 199-204.

Classification of Wetland Forests and Scrub in the Western Balkans

Wetland forests and scrub (WFS) are conditioned by the strong impact of water. They consist of various vegetation types, depending on many factors such as type and duration of flooding, water table level and its fluctuation, river current strength, substrate ability to retain water, etc. WFS vegetation has been insufficiently studied in the Balkan Peninsula, especially in Bosnia and Herzegovina. By means of numerical classification, we aimed to classify Western Balkans WFS at the alliance level, and to identify the main underlying ecological gradients driving the variation in species composition. The dataset containing all published and available unpublished relevés from Slovenia, Croatia and Bosnia and Herzegovina was first classified using the EuroVegChecklist Expert System in Juice software in order to assign the corresponding class to each of the relevés. Relevés were subsequently analyzed within each of the four WFS classes (Alno glutinosae-Populetea albae, Salicetea purpureae, Alnetea glutinosae and Franguletea). Cluster analysis resulted in eight alliances, Salicion albae, Salicion triandrae, Salicion eleagno-daphnoidis, Alno-Quercion, Alnion incanae, Alnion glutinosae, Betulion pubescentis and Salicion cinereae, while one cluster could not be assigned with certainty. Edafic factors were found to be the most important factors determining the floristic composition and syntaxa differentiation of WFS in the study area.

A deep reinforcement learning model for resilient road network recovery under earthquake or flooding hazards

As the backbone and the ‘blood vessel’ of modern cities, road networks provide critical support for community activities and economic growth, with their roles even more crucial due to the dramatic progress in urbanization. The service of road networks is subjected to the increasing frequency of high-consequence natural hazards such as earthquakes, floods, hurricanes, etc. Identifying resilient restoration sequences is essential to mitigate the disruption of such important infrastructure networks. This paper investigates a novel decision-support model to optimize post-disaster road network repair sequence. The model, named as GCN-DRL model, integrates the advantages of deep reinforced learning (DRL) with graph convolutional neural network (GCN), two emerging artificial intelligence (AI) techniques to achieve efficient recovery of road network service. The model is applied to analyze two cases of community road networks in the US that are subjected to different types of hazards, i.e., earthquakes and flooding. The performance of repair sequence by the GCN-DRL model is compared with two commonly used methods, i.e., repair sequence by the genetic algorithm and by prioritization based on graph importance with betweenness centrality. The results showed the decision sequence by GCN-DRL model consistently achieved superior performance in road network restoration than the conventional methods. The AI-based decision model also features high computational efficiency since the GCN-DRL model can be trained before the hazard. With a pre-trained GCN-DRL model, a close to optimal decision-making process can be made available rapidly for different types of new hazards, which is advantageous in efficiently responding to hazards when they happen. This study demonstrates the promise of a new AI-based decision support model to improve the resilience of road networks by enabling efficient post-hazards recovery.

Flood risk research for improving flood risk outcomes

Flood risk research for improving flood risk outcomes