Inundation Model Research Articles

Urban inundation rapid prediction method based on multi-machine learning algorithm and rain pattern analysis

Urban inundation disasters caused by extreme rainfall events are becoming increasingly severe. However, the numerical inundation model based on physical process has relatively slow computational speed and struggle to meet the demands of current forecasting and early warning systems. Addressing how to integrate machine learning technology into rapid urban inundation prediction, especially when dealing with limited training samples, and selecting the most appropriate machine learning algorithms for different rain patterns, is a crucial and pressing issue. To tackle this problem, this work combines a coupled hydrological-hydrodynamic model known for its high-computational accuracy with efficient machine learning algorithms. Taking different rain patterns into account, a rapid urban inundation prediction method is proposed. This method utilizes data-driven results from the coupled hydrological-hydrodynamic model and constructs multiple machine learning algorithms based on the conclusions drawn from the analysis regarding the most suitable machine learning algorithms for different rain patterns. The results show that for single-peak, double-peak, and uniform rain patterns, the Ridge algorithm, KNN algorithm, and RF algorithm should be respectively employed. The method achieves a mean absolute percentage error of 5.32 %, 7.73 %, and 2.49 % for single-peak, double-peak, and uniform rainfall scenarios and can predict inundation within a 3.68 km2 area in 14.07 s. This method offers high prediction accuracy and fast computational speed, not only meeting the requirements of daily forecasting and early warning but also providing a new direction for the application of machine learning technology in urban inundation prevention.

A rapid simplified method for determining tsunami inundation extent based on energy conservation

AbstractThis paper develops a tsunami inundation model, filling the current void between industry applied simplified methods (bathtub and attenuation) and comprehensive numerical modeling. The proposed model utilizes two‐dimensional equations established on hydraulic principles (energy conservation and friction loss) to produce the finite‐difference, two‐dimensional model. While the sophistication of depth‐averaged numerical modeling has not been entirely replicated, the incorporation of hydraulic roughness and gravitational effects (slope) through the Bernoulli energy equation, accompanied by suitable assumptions, significantly improves performance over simplified bathtub and attenuation methods. This improvement not only retains excellent efficiency (completing case studies in under 100 s) but also yields inundation results comparable to traditional numerical modeling, which typically takes multiple hours to days. This is the first application of such equations to a two‐dimensional tsunami inundation model, leading to simulations at Gisborne and Christchurch, New Zealand, for validation against the numerical model, COMCOT. F1 scores produced from inundation extent comparisons were upwards of 84%. Inundation depth discrepancies recorded 77% within a 1 m range, while over 93% within a 2 m range.

Current and future rainfall-driven flood risk from hurricanes in Puerto Rico under 1.5 and 2 °C climate change

Abstract. Flooding associated with Hurricane Maria in 2017 had devastating consequences for lives and livelihoods in Puerto Rico. Yet, an understanding of current and future flood risk on small islands like Puerto Rico is limited. Thus, efforts to build resilience to flooding associated with hurricanes remain constrained. Here, we take an event set of hurricane rainfall estimates from a synthetic hurricane rainfall simulator as the input to an event-based rainfall-driven flood inundation model using the hydrodynamic code LISFLOOD-FP. Validation of our model against high-water-mark data for Hurricane Maria demonstrates the suitability of this model for estimating flood hazard in Puerto Rico. We produce event-based flood hazard and population exposure estimates for the present day and the future under the 1.5 and 2 ∘C Paris Agreement goals. Population exposure to flooding from hurricane rainfall in Puerto Rico for the present-day climate is approximately 8 %–10 % of the current population for a 5-year return period, with an increase in population exposure to flooding by 2 %–15 % and 1 %–20 % under 1.5 and 2 ∘C futures (5-year return period). This research demonstrates the significance of the 1.5 ∘C Paris Agreement goal for Small Island Developing States, providing the first event-based estimates of flooding from hurricane rainfall under climate change for a small island.

Fast-Processing DEM-Based Urban and Rural Inundation Scenarios from Point-Source Flood Volumes

Flooding has always been a huge threat to human society. Global climate change coupled with unsustainable regional planning and urban development may cause more frequent inundations and, consequently, higher societal and economic losses. In order to characterize floods and reduce flood risk, flood simulation tools have been developed and widely applied. Hydrodynamic models for inundation simulation are generally sophisticated, yet they normally require massive setup and computational costs. In contrast, simplified conceptual models may be more easily applied and efficient. Based on the Hierarchical Filling-and-Spilling or Puddle-to-Puddle Dynamic Filling-and-Spilling Algorithms (i.e., HFSAs), Safer_RAIN has been developed as a fast-processing DEM-based model for modelling pluvial flooding over large areas. This study assesses Safer_RAIN applicability outside the context for which it was originally developed by looking at two different inundation problems with point-source flooding volumes: (1) rural inundation modelling associated with levee breaching/overtopping; (2) urban flooding caused by drainage systems outflow volumes.

Tsunami Inundation Modelling in a Built-In Coastal Environment with Adaptive Mesh Refinement: The Onagawa Benchmark Test

In tsunami studies, understanding the intricate dynamics in the swash area, characterised by the shoaling effect, remains a challenge. In this study, we employed the adaptive mesh refinement (AMR) method to model tsunami inundation and propagation in the Onagawa town physical flume experiment. Using the open-source flow solver Basilisk, we implemented the Saint-Venant (SV) equations, Serre–Green–Naghdi (SGN) equations, and a nonhydrostatic multilayer (ML) extension of the SGN equations. A hydraulic bore tsunami-like wave was used as the input boundary condition. The objective was to assess the efficiency of the AMR method with nonhydrostatic tsunami models in overcoming limitations in 2D and quasi-3D models in flume experiments, particularly with respect to improving accuracy in arrival time and run-up detection. The results indicate improved performance of the SGN and SV models in determining tsunami arrival times. The ML model demonstrated enhanced wave run-up simulations on complex built-in terrain. The refined roughness coefficient determined using the ML solver captured the arrival time well in the northern section of the Onagawa model, albeit with a 1 s delay. The AMR method offered a computationally stable solution with an 86.3% reduction in computational time compared to a constant grid. While effective, the nonhydrostatic models entail the use of a great deal of computational resources.

FloodGNN-GRU: a spatio-temporal graph neural network for flood prediction

Abstract Classical approaches for flood prediction apply numerical methods for the solution of partial differential equations that capture the physics of inundation processes (e.g., the 2D Shallow Water equations). However, traditional inundation models are still unable to satisfy the requirements of many relevant applications, including early-warning systems, high-resolution (or large spatial domain) simulations, and robust inference over distributions of inputs (e.g., rainfall events). Machine learning (ML) approaches are a promising alternative to physics-based models due to their ability to efficiently capture correlations between relevant inputs and outputs in a data-driven fashion. In particular, once trained, ML models can be tested/deployed much more efficiently than classical approaches. Yet, few ML-based solutions for spatio-temporal flood prediction have been developed, and their reliability/accuracy is poorly understood. In this paper, we propose FloodGNN-GRU, a spatio-temporal flood prediction model that combines a graph neural network (GNN) and a gated recurrent unit (GRU) architecture. Compared to existing approaches, FloodGNN-GRU (i) employs a graph-based model (GNN); (ii) operates on both spatial and temporal dimensions; and (iii) processes the water flow velocities as vector features, instead of scalar features. We evaluate FloodGNN-GRU using a LISFLOOD-FP simulation of Hurricane Harvey (2017) in Houston, Texas. Our results, based on several metrics, show that FloodGNN-GRU outperforms several data-driven alternatives in terms of accuracy. Moreover, our approach can be trained 100x faster and tested 1000x faster than the time required to run a comparable simulation. These findings illustrate the potential of ML-based methods to efficiently emulate physics-based inundation models, especially for short-term predictions.

Flood Inundation and Streamflow Changes in the Kabul River Basin under Climate Change

The Kabul basin yields around 16% of the total annual water availability in Pakistan. Changing climate will alter the precipitation regime in terms of intensity and frequency, which will affect the water yield and cause flood hazards. Against this background, this study aims to quantify the impacts of changing climate on the water yield, its timings, and, more importantly, the associated flood hazards in the transboundary Kabul basin. For this, we used a rainfall-runoff inundation (RRI) model coupled with the snow and glacier melt routines and drove it for historical and future climates simulated by the atmosphere-only general circulation model (AGCM) at 20 km spatial resolution. The model simulations reveal that rainfall runoff contributes around 50% of the annual flows, and the rest is contributed by glaciers and snow melts. Annual precipitation is projected to increase by 14% from 535 mm, whereas temperatures will rise by 4.7 °C. In turn, the Kabul River flows will only increase by 4% to 1158 m3s−1 from 1117 m3s−1, mainly due to an increase in winter flows. In contrast to a minute increase in the mean river flows, the maximum flood inundation area is projected to increase by 37%, whereas its depth will rise between 5 and 20 cm.

Flow-tub model: A modified bathtub flood model with hydraulic connectivity and path-based attenuation

Global climate change and sea level rise are increasing the risks of flooding for coastal communities. Probabilistic coastal flood risk analysis at regional or global scales requires flood models with relatively low data requirements and low computational costs. Bathtub inundation models, which compute flood depth as the difference between water level and ground elevation, are well-suited for large-scale flood risk analysis. However, these models may overestimate floods because they do not capture some of the relevant underlying hydrodynamic processes that govern flood propagation on land.We present Flow-Tub, a modified bathtub inundation model that integrates two hydrodynamic processes to improve the accuracy of the bathtub inundation model while retaining computational efficiency: hydraulic connectivity and path-based attenuation.1.Hydraulic connectivity ensures that inundation is restricted to areas connected to the water source.2.Path-based attenuation ensures that the modeled flood water depths are reduced along the flow paths to represent the effects of surface friction and the temporary nature of storm surges.We validate the Flow-tub model against a hydrodynamic model. We also compare results of the bathtub model and the Flow-Tub model, highlighting the improved accuracy in the estimation of flood depths in the latter.

Microsoft building footprint application To detect human exposure due to tsunami

Tsunami events at night are more prone to causing fatalities because humans are resting in residential buildings (houses). In this study, residential buildings were extracted using the Microsoft Building Footprint (MBF), which resulted from applying artificial intelligence technology. This study aims to analyze the number of people exposed to tsunamis at night using MBF. The tsunami modeling was carried out using the Berryman method. Sentinel 2-A Image extracted from Google Earth Engine. The results of the inundation modeling analysis show that the total inundated area is 717 Ha or 17.34% of the total area. The results of the MBF accuracy analysis on the entire data are a Precision of 99.02%, Recall of 98.40%, and F1 score of 98.71%. The results of the MBF error analysis are False Positive 0.97%, False Negative 1.60%, and Intersection of Union 0.12%. The number of people exposed is 2,749, or 6.32% of the total population.

Impact of land use changes on the tsunami hazards in part of coastal Kebumen

This research was conducted to analyze land use change from 2016 to 2022 and their impact on the tsunami hazard zone on the coast of Kebumen Regency. For land use change analysis, remote sensing, geographic information systems (GIS), and statistical tests were applied to Sentinel-2A satellite imagery. The tsunami hazard was simulated using tsunami inundation modeling based on land use spatial data and DEMNAS processed using GIS. Land use changes occurred significantly in the study area, especially in Mirit District. The Southern Cross Road (JJLS) and coastal morphological conditions influence land use change patterns. Land use changes impact changes in the tsunami hazard zone, especially in the use of fir forests and shrimp ponds. The research findings can be used as input for developing a tsunami disaster mitigation plan and detailed spatial planning on the coast of Kebumen Regency.

Application of a subgrid-scale urban inundation model for a storm surge simulation: Case study of typhoon Haiyan

This study evaluated the individual drag force model (iDFM), a subgrid-scale model for simulating tsunami inundation in coastal urban areas. The iDFM was then applied to characterize Typhoon Haiyan-induced storm surge inundations in 2013. Using a post-typhoon survey dataset containing inundation depths at different survey points with inundation limits, the performance of iDFM was validated, and the results confirmed that iDFM accurately estimates the storm surge inundation depths but overestimates the limit of inundation. In addition, the results revealed that the iDFM flow velocity was a dominant factor, and the inclusion of buildings in the simulated affected the arrival of the leading edge of inundation. Furthermore, the reduction in flow velocity was sensitive to the subgrid-scale parameters, and the spatial distribution pattern of buildings located in the inundated area affected the relative difference in the maximum flow velocity with the regional roughness model. The inundation characteristics of the Typhoon Haiyan storm surge models were compared with a historical tsunami event. The inundation characteristics were similar and model differences were observed mainly in the fluid velocity and momentum flux rather than the inundation depth. However, the magnitudes of relative differences observed in Tacloban were larger for fluid velocity and momentum flux, respectively, than in the tsunami event. The effect of building drag on advection was also substantial in Tacloban due to the differences in the inflow mass and momentum fluxes.

Efficient solutions for fractional Tsunami shallow-water mathematical model: A comparative study via semi analytical techniques

In this study, we delve into the effective solution of the fractional tsunami shallow-water mathematical model. Tsunami propagation and inundation modeling is of paramount importance for hazard assessment and disaster preparedness. By employing two distinct semi-analytical techniques: the residual power series method and the modified Taylor series method, we obtain approximate solutions for the tsunami mathematical model. Solutions for varying coastal inclinations and ocean depths have been acquired for the given model taking into consideration the changes in coastal gradient and ocean depth and how to impact tsunami wave velocity and wave amplification at different values of the fractional order derivative and different steps of time. The utilized semi analytical techniques generate the solution in the form of power series expansion, so the convergence study is implemented. In order to prove the efficiency of the proposed methods, several tables for comparisons between our results and other numerical results from literature are presented. Also graphs of the obtained solutions introduced in two and three dimensions.

Modified geomorphic flood index (GFI) method on backwater problem in tidal tropical river for rapid flood assessment

The Geomorphic Flood Index (GFI) model can show flood inundation in a watershed based on flood events. Flood inundation can occur because of the backwater effect in the river channel. The different water levels from the river and tide that occur downstream may temporarily disturb the water depth and flow velocity. Because of this disturbance, a backwater effect might occur. The case study is located at the Plumbon River in Semarang City, which has the potential for a backwater effect in the river downstream. Backwater is a hydraulic phenomenon that occurs in rivers and can be simulated using a hydrodynamic model, as it can show the relationship between river discharge and tides. The conventional GFI method does not consider hydraulic effects, such as backwater effects, as part of the calculation. Therefore, modifications will be inserted into the calculation of the hydraulic effect on the modelling of flood inundation. The main purpose of this study is to build a modified GFI method that includes backwater analysis in its modelling to predict the backwater effect that occurs downstream of rivers and its impact on inundation. This study will help rapidly predict flood inundation.

Coastal Inundation Hazard Assessment in Australian Tropical Cyclone Prone Regions

One of the hazards associated with tropical cyclones (TCs) is a storm surge, which leads to coastal inundation and often results in loss of life and damage to infrastructure. In this study, we used GIS-based bathtub models and tide-gauge-derived water levels to assess coastal inundation scenarios for the landfall region of TC Debbie. The three scenarios modelled what could have happened if the TC’s maximum storm surge had coincided with the maximum storm tide for that day, month, or TC season, where the water levels were determined through analysis of tide gauge data, using a new method called the variable enhanced Bathtub Model. Additionally, this study analysed the impact of excluding the correction of water levels with the Australian Height Datum. Our study found that between the least and most severe scenarios, with the input water-level difference for the model along the coastline being 0.43 m, the observed inundation depth of the analysed populated region increased from 0.25 m to 1 m. Ultimately, it was found that in the worst-case scenario, the study region could have experienced coastal inundation 0.63 m higher than it did, inundating 72.53 km2 of the coast. The results of this study support the consensus that coastal inundation is highly dependent on the characteristics of the terrain, and that coastal inundation modelling, such as that completed in this study, needs to be performed to better inform decision makers and communities of the potential impacts of TC-induced storm surges.

Assessing extreme flood inundation and demographic exposure in climate change using large ensemble climate simulation data in the Lower Chao Phraya River Basin of Thailand

Study regionLower Chao Phraya River Basin (LCPRB), Thailand. Study focusThis study aims to make a robust assessment based on a large ensemble d4PDF dataset and a flood-inundation model Inundation Model Coupling Rainfall-runoff (IMCR) for the LCPRB. Inundation area and depth for 100-year flooding are evaluated for the flood volume of capacity greater than 2000 m3/s for both past (1951–2010) and future (2051–2100) climates. This study also evaluates the affected population exposure in the region for both past and future climate scenarios. New hydrological insights for the regionThe IMCR inundation simulation findings indicate that compared to the historical climate, the inundation area increases by an average of 1.0–1.4 times, and the critical area (depth >3 m) increases by an average of 1.1–1.3 times. On the other hand, the exposed population in the future, with respect to the SSP5 scenario "Taking the Highway," is expected to decrease on average by 0.7–0.9 times in comparison to the past climate for depth > 0 m. However, keeping the population constant as in the past, the exposed population is likely to increase on average by 1.3–1.4 times in comparison to the past climate for depth > 0 m.

A Novel Method to Determine Probabilistic Tsunami Hazard Using a Physics‐Based Synthetic Earthquake Catalog: A New Zealand Case Study

AbstractTsunamis can have devastating consequences for coastal communities. Yet hazards from future tsunamis are difficult to quantify due to their rarity in the instrumental record. Statistical earthquake catalogs have previously been used to quantify tsunami hazards. For the first time, we use a physics‐based synthetic earthquake catalog to assess probabilistic tsunami hazard in a local region. We analyze a 30,000‐year catalog to capture a wide range of the source complexity evident in tsunamigenic earthquakes. We produce models of land‐surface displacements and subsequent tsunamis from 2,585 earthquakes with magnitude > MW 7.0. Modeled slip of the Hikurangi and Tonga‐Kermadec subduction thrusts generated maximum wave heights at the coast of up to 28 m and these earthquakes pose the greatest tsunami hazard along New Zealand's coastline. The results provide a “proof of concept” for using earthquake simulators for probabilistic tsunami hazard models. They further present a platform upon which next‐generation probabilistic tsunami inundation models may be constructed.

A novel conceptual flood inundation model for large scale data-scarce regions

We propose a novel conceptual flood inundation model that can be coupled with any hydrological model to generate the probabilistic flood extent maps for a given time step. We generate the runoff rasters using the dynamic Budyko hydrological model that uses the regional IMD/CPC precipitation datasets and the global JAXA GSMaP gauge-precipitation (V7) product over Kerala/Louisiana flood events during August 2018/2016. A freely available global dataset is used for obtaining river geometry information and topographic information is obtained from ASTER/SRTM DEM and localised LiDAR DEM. We evaluated the proposed model using the ground observed flood points from the state disaster management agency for Kerala and flood maps published by USGS for Louisiana. The proposed model captured ∼ 45% of flooding during peak flood days while taking ∼ 15 min for each run on a daily time step on a desktop computer.

The Cost of Imperfect Knowledge: How Epistemic Uncertainties Influence Flood Hazard Assessments

AbstractClassical approaches to flood hazard are obtained by the concatenation of a recurrence model for the events (i.e., an extreme river discharge) and an inundation model that propagates the discharge into a flood extent. The classical approach, however, uses “best‐fit” models that do not include uncertainty from incomplete knowledge or limited data availability. The inclusion of these, so called epistemic uncertainties, can significantly impact flood hazard estimates and the corresponding decision‐making process. We propose a simulation approach to robustly account for uncertainty in model's parameters, while developing a useful probabilistic output of flood hazard for further risk assessments via the Bayesian predictive posterior distribution of water depths. A Peaks‐Over‐Threshold Bayesian analysis is performed for future events simulation, and a pseudo‐likelihood probabilistic approach for the calibration of the inundation model is used to compute uncertain water depths. The annual probability averaged over all possible models’ parameters is used to develop hazard maps that account for epistemic uncertainties. Results are compared to traditional hazard maps, showing that not including epistemic uncertainties can underestimate the hazard and lead to non‐conservative designs, and that this trend increases with return period. Results also show that the influence of the uncertainty in the future occurrence of discharge events is predominant over the inundation simulator uncertainties for the case study.

Combining Landsat, VIIRS Night-time Light, and Sentinel-1 SAR for Spatial Flood Risk Assessment in Coastal Area: an Earth Engine Cloud Computing Process

The intensity of development in coastal areas stimulates various potential issues such as flood disasters. This study aims to demonstrate the importance of latest methods and geospatial data as inputs for coastal spatial planning policies in efforts to reduce flood disaster risks. Leveraging spatial analysis with cloud computing through Google Earth Engine (GEE), this research assesses flood risk components—hazards, vulnerability, and capacity. The method involves processing SAR Sentinel-1 data to map flood inundation as a representation of hazards, analyzing Landsat and WorldPop data to evaluate vulnerability, and assessing capacity by utilizing VIIRS nighttime light level imagery to determine economic activities. The chosen research study location is the coastal area of Pekalongan due to frequent flood disasters throughout the year. The results demonstrate that cloud computing is capable of assessing flood risks. The flood inundation model using SAR data covers an area of 2,780 hectares with an accuracy of 96.75%. The analysis also reveals the highest vulnerability level, reaching 15.7% (946.32 hectares) of the total area. The capacity analysis indicates a medium to high level of 15% (913.6 hectares). The assessment of flood risks in the coastal area is dominated by the medium to very high-risk class, covering 43% (2,631.84 hectares) of the area. In conclusion, integrating cloud-based flood risk modeling into spatial planning is crucial, considering disaster resilience for sustainable human habitats.

Dairy farming exposure and impacts from coastal flooding and sea level rise in Aotearoa-New Zealand

Globally, episodic coastal flooding is expected to have increasingly severe impacts on agriculture in low-lying coastal areas, as climate change continues to cause relative sea level rise (RSLR). This study analyses dairy farm exposure and impacts from extreme sea levels (ESL) and RSLR in Aotearoa-New Zealand (A-NZ). Spatio-temporal modelling is undertaken using: 1) a dairy production land layer; 2) a static coastal inundation model from 10, 100, 200 and 500-year annual recurrence interval (ARI) ESL events at 0–2 m RSLR (0.5 m increments); and 3) an Impact State (IS) scheme relating percentage of farmland inundated and depth of flooding at milking shed locations to categorise the severity of damage. Through this analysis we highlight the increasing exposure of the dairy industry to coastal flooding, with 472 farms potentially impacted (at or greater than IS 1) by a 10-year ARI event and 1276 farms exposed to a 500-year ARI event at current sea levels, rising to 1276 (10-year ARI) and 1344 farms (500-year ARI) with 2 m RSLR. RSLR is identified as the factor driving this increase in exposure and severity of impacts to dairying, with the increase in the number of farms impacted in a 500-year compared to 10-year event only 5 % greater when considered with 1 m of RSLR, compared to 37 % greater at 0 m RSLR. This has significant implications for A-NZ where the dairy industry is of high economic and social importance, demonstrating the need for industry targeted climate change adaptation and disaster risk reduction measures.