Flood Evacuation Routes Based on Spatiotemporal Inundation Risk Assessment

Lee, Yoonha; Kim, Hyun Il; Han, Kun‐Yeun; Hong, Won Hwa

doi:10.3390/w12082271

Cited by 18 publications

(11 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Commonly used ML algorithms were summarized by Mosavi et al [4], and they include Artificial Neural Networks (ANN) and MultiLayer Perceptron (MLP) [5], Adaptive Neuro-Fuzzy Inference System (ANFIS) [6], Wavelet Neural Networks (WNN) [7], Support Vector Machine (SVM) [8], Decision Tree (DT) [9], and hybrid models [10,11]. More recently, advanced methods such as deep learning (e.g., Convolutional Neural Networks (CNN) [12,13]), Extreme Machine Learning (EML) [14], dynamic or recurrent neural networks (e.g., Nonlinear Autoregressive network with exogenous inputs (NARX) [15,16], and Long Short-Term Memory (LSTM) [17]) are also gaining popularity in the hydrological field.…”

Section: Introductionmentioning

confidence: 99%

Flood Stage Forecasting Using Machine-Learning Methods: A Case Study on the Parma River (Italy)

Dazzi

Vacondio

Mignosa

2021

Water

View full text Add to dashboard Cite

Real-time river flood forecasting models can be useful for issuing flood alerts and reducing or preventing inundations. To this end, machine-learning (ML) methods are becoming increasingly popular thanks to their low computational requirements and to their reliance on observed data only. This work aimed to evaluate the ML models’ capability of predicting flood stages at a critical gauge station, using mainly upstream stage observations, though downstream levels should also be included to consider backwater, if present. The case study selected for this analysis was the lower stretch of the Parma River (Italy), and the forecast horizon was extended up to 9 h. The performances of three ML algorithms, namely Support Vector Regression (SVR), MultiLayer Perceptron (MLP), and Long Short-term Memory (LSTM), were compared herein in terms of accuracy and computational time. Up to 6 h ahead, all models provided sufficiently accurate predictions for practical purposes (e.g., Root Mean Square Error < 15 cm, and Nash-Sutcliffe Efficiency coefficient > 0.99), while peak levels were poorly predicted for longer lead times. Moreover, the results suggest that the LSTM model, despite requiring the longest training time, is the most robust and accurate in predicting peak values, and it should be preferred for setting up an operational forecasting system.

show abstract

Section: Introductionmentioning

confidence: 99%

Flood Stage Forecasting Using Machine-Learning Methods: A Case Study on the Parma River (Italy)

Dazzi

Vacondio

Mignosa

2021

Water

View full text Add to dashboard Cite

show abstract

“…The specific evacuation routes of the closest shelters to each population grid were identified and visualized by a limiting condition of 3 km. This step is meaningful since it has been revealed that the residents in Syracuse have no sense of where to go during a flood event [28]. Therefore, the results from this study indicate the nearest designated evacuation shelter and evacuation route.…”

Section: Flood Evacuation Designmentioning

confidence: 81%

“…Many scholars have studied the death rate of floods using inundation depth. Literature suggests that the death rate of a population is strongly related to the inundation depth [28]. This study produced an estimate of the loss of population based on the death rate and inundation depth determined using the equation expressed in Figure 7.…”

Section: Loss Of Population In a 500-year Flood Eventmentioning

confidence: 99%

“…The macroscopic modeling method is based on the evacuation starting point and traffic allocation. All of these methods involve calculating the evacuation routes, which may be achieved in two ways: (1) mathematical analysis and (2) computer simulation [28]. GIS network analysis is one of the computer simulation methods frequently used to select the optimal evacuation path.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Urban Flood Loss Estimation and Evacuation Design Based on a 500-Year Extreme Flood Event in Syracuse City

2022

Water

View full text Add to dashboard Cite

To understand the potential risk of flooding in Syracuse City, New York State, USA, this research attempted to accomplish the flood hazard assessment for a simulated 500-year flood event in the downstream floodplain of Onondaga Creek within Syracuse. Based on the commonly used category of flood damages, the flood damage of Syracuse was divided into loss of buildings and loss of population. The results showed that the city’s center would have the highest damage rate for buildings and a total of 1139 buildings would be inundated, of which 326 buildings would be severely damaged by more than 80%. Furthermore, about 7390 people would be directly affected by the flood event, among which approximately 900 people might lose their lives. Communities near Onondaga Creek were assigned designated evacuation shelters based on the accessibility and distance to the shelters. The shortest available evacuation routes were calculated. More shelters should be provided in the central downtown area, with its large population, and distributed along the western bank of Onondaga Creek. This research offered a first approximate flood loss estimation that might lead to more attention and studies concerning a potential flood hazard in the future. It also provided science-based guidelines for city authorities to refer to in practical flood hazard mitigation.

show abstract

“…Satellite forecasts of cloud covers - [2] easily provide an indication of the spatial extent of cloudy regions over a city, and meteorological forecasts of extreme weather events, including a forecast of heavy precipitation, are also possible from observational and modelling studies. Models rely on the successful prediction of the onset times of precipitation for better flood management, and any evacuation measure is invariably tied down to the time duration over which the precipitation event occurs [3][4][5]. Earlier studies have highlighted that precipitation onset times as well as the initial storage levels in local water tanks are crucial parameters that modulate the extent of flooding in a region [6].…”

Section: Introductionmentioning

confidence: 99%

Quick Predictions of Onset Times and Rain Amounts from Monsoon Showers over Urban Built Environments

Gumber

Ghosh

2022

Atmosphere

View full text Add to dashboard Cite

Predicting the onset times of precipitation over densely populated cities for the purposes of timely evacuation is a challenge. This paper explored a flooding event over an urban built environment in a South Asian mega city, Chennai, where extant urban planning models rely on predicted rainwater amounts for early warning and impact assessment studies. However, the time duration of flooding events is related to the nature of the urban sprawl in the built environment. Any evacuation measure is invariably tied down to the time duration over which the precipitation event occurs, and therefore to the expected time of a precipitation event to begin. In this context, a crucial parameter useful to municipal authorities is the onset time of precipitation. This study used optimised analytical formulations to predict this time, and the derived analytical expressions for the case study yielded comparable times estimated from a computer-intensive full-scale large eddy model within an accuracy of 2%. It is suggested that municipal authorities (who are non-experts in fluid mechanics) use this early prediction for the purposes of quick alerts to a congested city’s most vulnerable citizens within urban sprawls. However, for the procedure to work at its best, it involves a two-stage procedure. The first step involves the use of a parcel model to obtain the expected cloud droplet spectral spreads based on the prevailing dynamical characterisations. The second step involves an optimisation procedure involving cloud spectral properties from the first step to quantify both the auto-conversion rates and the threshold. Thereafter, an onset time calculation based on cloud properties is estimated. These new results are cast in closed form for easy incorporation into meteorological applications over a variety of urban scales. Rain mass amounts were also predicted analytically and used to configure Aeronautical Reconnaissance Coverage Geographic Information System (ARCGIS) to compute low drainage flow rates over the vulnerable parts of Chennai city. It was found that heavy precipitation over the North Chennai region yielded discharge rates to the tune of ~250 m3s−1 during a 24 h period, causing intense flooding in the low-lying areas around the Cooum River basin with a large population density, with estimates sufficiently corroborating observations.

show abstract

Flood Evacuation Routes Based on Spatiotemporal Inundation Risk Assessment

Cited by 18 publications

References 38 publications

Flood Stage Forecasting Using Machine-Learning Methods: A Case Study on the Parma River (Italy)

Flood Stage Forecasting Using Machine-Learning Methods: A Case Study on the Parma River (Italy)

Urban Flood Loss Estimation and Evacuation Design Based on a 500-Year Extreme Flood Event in Syracuse City

Quick Predictions of Onset Times and Rain Amounts from Monsoon Showers over Urban Built Environments

Contact Info

Product

Resources

About