Flood Inundation Mapping Research Articles

Flood risk evaluation of the coastal city by the EWM-TOPSIS and machine learning hybrid method

The frequent occurrence of floods and waterlogging has significantly impacted coastal cities. Effective mapping of flood risk can enhance the precision of disaster risk reduction strategies. However, there is room for improvement in fields such as flood inundation data, evaluation objectivity, and classification refinement. This study, using Xiamen as a case study, advanced flood inundation data accuracy by generating a frequency-based flood inundation map from multi-year remote sensing imagery. This study integrated the EWM-TOPSIS model with a neural network model to evaluate the flood risk. The EWM-TOPSIS model was employed to assess the flood vulnerability and flood exposure, while the multi-class neural network model to simulate the flood hazard. This combined approach reduced subjectivity in the flood risk assessment of Xiamen and achieved a finer level of risk classification compared to traditional binary neural networks. The results indicated that the flood vulnerable areas of Xiamen are concentrated along waterways distant from roads, with high flood exposure in Tong'an and Xiang'an districts. The flood-prone areas in Xiamen are primarily located along the coastal areas characterized by extensive impermeable surfaces. High flood risk areas are mainly distributed in Tong'an and Xiang'an subdistricts, particularly in Xiangping and Xindian. The method developed offers support for accurate flood risk identification and decision-making in cities with limited data resources.

Effects of high-quality elevation data and explanatory variables on the accuracy of flood inundation mapping via Height Above Nearest Drainage

Abstract. Given the availability of high-quality and high-spatial-resolution digital elevation maps (DEMs) from the United States Geological Survey's 3D Elevation Program (3DEP), derived mostly from light detection and ranging (lidar) sensors, we examined the effects of these DEMs at various spatial resolutions on the quality of flood inundation map (FIM) extents derived from a terrain index known as Height Above Nearest Drainage (HAND). We found that using these DEMs improved the quality of resulting FIM extents at around 80 % of the catchments analyzed when compared to using DEMs from the National Hydrography Dataset Plus High Resolution (NHDPlusHR) program. Additionally, we varied the spatial resolution of the 3DEP DEMs at 3, 5, 10, 15, and 20 m (meters), and the results showed no significant overall effect on FIM extent quality across resolutions. However, further analysis at coarser resolutions of 60 and 90 m revealed a significant degradation in FIM skill, highlighting the limitations of using extremely coarse-resolution DEMs. Our experiments demonstrated a significant burden in terms of the computational time required to produce HAND and related data at finer resolutions. We fit a multiple linear regression model to help explain catchment-scale variations in the four metrics employed and found that the lack of reservoir flooding or inundation upstream of river retention systems was a significant factor in our analysis. For validation, we used Interagency Flood Risk Management (InFRM) Base Level Engineering (BLE)-produced FIM extents and streamflows at the 100- and 500-year event magnitudes in a sub-region in eastern Texas.

Dam break risk analysis and flood inundation mapping: a case study of Wadi Al-Arab Dam

Dam break risk analysis and flood inundation mapping: a case study of Wadi Al-Arab Dam

Impacts of DEM type and resolution on deep learning-based flood inundation mapping

Impacts of DEM type and resolution on deep learning-based flood inundation mapping

Dam breach analysis and flood inundation mapping of Dire Dam, using HEC-HMS and HEC-RAS models

Dam breach analysis and flood inundation mapping of Dire Dam, using HEC-HMS and HEC-RAS models

Flood sequence mapping with multimodal remote sensing under the influence of dense vegetation

ABSTRACT Remote sensing (RS) imagery is becoming increasingly popular in surface water extent extraction thanks to the increasing availability of RS data and advancements in image processing algorithms, software, and hardware. Many studies proved that RS imagery can work independently or along with other approaches to identify flood extent. However, due to the insufficiency in the number of images from single-sourced RS and independent references for validation, most studies just depicted the inundation status near the peak of the inundation. The potential of those images to document flood events at different stages (e.g. rising and receding stages) has not been well investigated. To close that gap, this study investigated the efficacy of RS-based multi-spatiotemporal flood inundation mapping using multimodal RS images captured on different dates to describe the entire flooding process. Additionally, a Quantile-based Filling & Refining (QFR) workflow was proposed to resolve the blocking effects of dense vegetation in study areas. We tested the multimodal flood mapping workflow plus the QFR correction in four lock and dam sites on the Mississippi River by comparing the RS-based flood maps with HEC-RAS simulations. Our results demonstrated the usefulness of multimodal RS images in describing flooding events at different stages and showcased the potential of those images to serve as a reliable reference source in data-scarce areas. In addition, results showed that the standard water extraction plus post-processing will not guarantee accurate flood maps in densely vegetated areas. In contrast, map processed with QFR were noticeably more consistent with HEC-RAS maps, especially for flood maps generated with PlanetScope images, for which the median accuracy has been improved from below 0.5 to above 0.94 after the QFR postprocessing. Thanks to the simple structure, the proposed multimodal RS flood mapping workflow plus QFR correction procedures can be fully automated and can thus benefit near-real-time applications.

Extension of the Geomorphic Flood Index classifier to predict flood inundation maps for uncalibrated rainfall depths in arid regions

Flash floods are a rapid hydrological response that occurs within a short time with rapidly rising water levels and could lead to massive structural, social and economic damages. Therefore, generating flood inundation maps becomes necessary to distinguish areas exposed to floods. Hydrodynamic models are commonly used to generate inundation maps; however, they require high computational power and time, depending on the complexity of the model. For that, researchers developed effective, fast and simplified models. Among the simplified models, the Geomorphic Flood Index (GFI) is one of the most useful classifiers to generate inundation maps. Three main objectives are addressed in this study: (1) extend the GFI classifier to predict flood extent maps for uncalibrated rainfall depths, which will enhance early warning models for better risk assessments of extreme events; (2) enhance the accuracy of the simulated inundation maps using different calibration methods; and (3) investigate the performance of the GFI in various terrains with different resolutions. Three case studies in arid regions in Saudi Arabia were examined with different topographies, using terrains of high resolutions of 1 m and resampled low resolutions, as well as various rainfall depths corresponding to 5–100-yr return periods. The HEC-RAS 2D model was used to generate reference flood inundation maps. The obtained flood extent maps show high similarity compared to the reference maps with accuracy above 80%. Strong relationships between rainfall depths and the threshold GFI parameter were developed which allow producing inundation maps for any rainfall event.

Suitability of the height above nearest drainage (HAND) model for flood inundation mapping in data-scarce regions: a comparative analysis with hydrodynamic models

Unprecedented floods from extreme rainfall events worldwide emphasize the need for flood inundation mapping for floodplain management and risk reduction. Access to flood inundation maps and risk evaluation tools remains challenging in most parts of the world, particularly in rural regions, leading to decreased flood resilience. The use of hydraulic and hydrodynamic models in rural areas has been hindered by excessive data and computational requirements. In this study, we mapped the flood inundation in Huron Creek watershed, Michigan, USA for an extreme rainfall event (1000-year return period) that occurred in 2018 (Father’s Day Flood) using the Height Above Nearest Drainage (HAND) model and a synthetic rating curve developed from LIDAR DEM. We compared the flood inundation extent and depth modeled by the HAND with flood inundation characteristics predicted by two hydrodynamic models, viz., HEC-RAS 2D and SMS-SRH 2D. The flood discharge of the event was simulated using the HEC-HMS hydrologic model. Results suggest that, in different channel segments, the HAND model produces different degrees of concurrence in both flood inundation extent and depth when compared to the hydrodynamic models. The differences in flood inundation characteristics produced by the HAND model are primarily due to the uncertainties associated with optimal parameter estimation of the synthetic rating curve. Analyzing the differences between the HAND and hydrodynamic models also highlights the significance of terrain characteristics in model predictions. Based on the comparable predictive capability of the HAND model to map flood inundation areas during extreme rainfall events, we demonstrate the suitability of the HAND-based approach for mitigating flood risk in data-scarce, rural regions.

“Flood inundation mapping of upstream region in the Adyar River basin: Integrating hydrologic engineering centre's river analysis system (HEC-RAS) approach with groundwater considerations”

Chennai, the capital of Tamil Nadu, has grappled with recurrent devastating floods, notably in 2015 and 2021, originating from the upstream of the Adyar River basin. These events have profoundly impacted the city's growth sectors. The regions both upstream and downstream of the Adyar basin, covering Chennai and Chengalpattu districts, face increased vulnerability to severe riverine flooding due to rapid urbanization. The interplay of extensive urban development and river overflow during monsoon seasons has created conditions conducive to substantial flooding, necessitating the development of an effective flood model. This study utilizes digital elevation data and unit hydrograph analysis to estimate flood return probability water levels for various recurrence intervals, including the 10-year and 100-year events. Moreover, it investigates flood inundation, specifically in the Somangalam area, providing insights into the extent of water spread. To comprehensively address the flood dynamics, the study incorporates considerations for groundwater interactions. The overarching objective remains to identify the most vulnerable zones within the Adyar River region prone to flooding. Accurate identification of these vulnerable areas is crucial for policymakers and stakeholders to devise proactive planning and early warning measures, enhancing disaster preparedness and resilience in the face of potential flood disasters. This approach aims to safeguard the well-being of Chennai's residents and promote sustainable development in the region.

Risk Reduction and Recovery from Flood Hazard- A Case of Human Adjustment in Lower Damodar Basin, West Bengal

The study investigates the risk of flood hazards in the lower Damodar basin, focusing on human perception and adjustment strategies to mitigate and recover from the societal risk posed by these hazards. The river Damodar is known as ‘the sorrow of Bengal’ due to its flood ravages in the entire Damodar valley, which caused much unhappiness and distress in the lower Damodar region. In this area, the main flood-causing agents are rainfall and discharge from upland reservoirs. In the study, the annual peak flow is used to determine the flood frequency analysis. Flood inundation mapping analysis revealed that Amta and Khanakul blocks of Hooghly district are most vulnerable to maximum flood inundation during monsoon months. The damage due to floods in terms of economic loss and the probability of flood occurrences are used for assessing flood risk and the magnitude of vulnerability in the study area. The adopted structural measures in this area have not yielded sufficient results to mitigate the chronic flood problems. Therefore, some non-structural measures, particularly human adjustment strategies, are taken into consideration, like flood forecasting, alternative cropping arrangements, choices for other livelihood options, and capacity building through self-help- groups.

Fast high-fidelity flood inundation map generation by super-resolution techniques

Abstract Flooding is one of the most frequent natural hazards and causes more economic loss than all the other natural hazards. Fast and accurate flood prediction has significance in preserving lives, minimizing economic damage, and reducing public health risks. However, current methods cannot achieve speed and accuracy simultaneously. Numerical methods can provide high-fidelity results, but they are time-consuming, particularly when pursuing high accuracy. Conversely, neural networks can provide results in a matter of seconds, but they have shown low accuracy in flood map generation by all existing methods. This work combines the strengths of numerical methods and neural networks and builds a framework that can quickly and accurately model the high-fidelity flood inundation map with detailed water depth information. In this paper, we employ the U-Net and generative adversarial network (GAN) models to recover the lost physics and information from ultra-fast, low-resolution numerical simulations, ultimately presenting high-resolution, high-fidelity flood maps as the end results. In this study, both the U-Net and GAN models have proven their ability to reduce the computation time for generating high-fidelity results, reducing it from 7–8 h down to 1 min. Furthermore, the accuracy of both models is notably high.

Real-time flood inundation monitoring in Capital of India using Google Earth Engine and Sentinel database

This study focuses on researching flood inundation and vulnerable areas in Delhi NCT using Remote Sensing (RS) and GIS techniques during flood from July 8 to July 15, 2023, with the entire analysis conducted through satellite and cloud-based processing methods, specifically employing the Google Earth Engine (GEE). Leveraging high-temporal satellites that gather data enables the identification of flooded zones in real-time, during floods, and in the aftermath. Analyzing data collected at different stages of a flood provides valuable insights for pinpointing affected areas. Water naturally flows from high to low elevation areas, and based on elevation data, lowest elevation regions, particularly along the Yamuna riverbank under Delhi NCT, are highly susceptible to flooding, are considered flood-prone areas and flood water inundated. A thorough comprehension and flood-prone area mapping, along with a map illustrating highly inundated zones. After obtaining flood inundation maps and overlaying them with Land Use and Land Cover (LULC) classified maps, the study identified specific areas that experienced flood inundation. The analysis generated a flood zone map, indicating that the flooded area encompasses approximately 110 km², within a total study area of 1488.4 km². The affected areas have elevations ranging from 200 to 210 meters, whereas the maximum elevation in the study area is approximately 326 meters. The GEE platform is employed for processing, utilizing a Supervised classification algorithm for LULC mapping, and an Inverse Distance Weight method for mapping temperature and rainfall. This study utilized the GEE platform to create pre- and post-flood maps based on Sentinel 1 satellite datasets. Generated DEM, and employed it to create various surface estimation maps, including a stream order map. The GIS is employed to enhance the efficiency of monitoring and managing flood disasters, with the high temporal and spatial resolution data playing a pivotal role in flood monitoring.

Selection of Elevation Models for Flood Inundation Map Generation in Small Urban Stream: Case Study of Anyang Stream

To reduce flood damages, the Ministry of Environment in Korea has provided a flood inundation map so that people can expediently identify flood-prone areas. However, the current flood inundation maps have been produced based on the DEM which makes it difficult to represent realistic situations due to the lack of reproduction of land surface conditions. This study aims to provide more accurate and detailed flood inundation maps for flooding events due to river overflow in small urban areas. In this study, flood inundation analysis is performed using the river analysis system, HEC-RAS 2D, with the DSM and the DEM of urban areas in the Anyang Stream Basin, Korea to examine the differences in terms of terrain data and flooded area. Finally, for urban areas with dense buildings and congested road networks, the flood inundation analysis based on DSM can represent a more realistic flood situation and create an appropriate flood inundation map.

Innovative methods for rapid flood inundation mapping in Pul-e-Alam and Khoshi districts of Afghanistan using Landsat 9 images: spectral indices vs. machine learning models

Innovative methods for rapid flood inundation mapping in Pul-e-Alam and Khoshi districts of Afghanistan using Landsat 9 images: spectral indices vs. machine learning models

Flood Impact Assessment in Koton Karfe Using Sentinel-1 Synthetic Aperture Radar (SAR) Data

Flood has proven to be an incessant menace in Nigeria more threatening to riverine areas. The most recent flood ensued in 2022 as a result of heavy rainfall and the release of water from Lagdo Dam in Cameroon which became very devastating in many areas notably the Koton Karfe area in Kogi State, causing business shutdowns and the loss of lives and properties. In this work, Sentinel-1 Synthetic Aperture Radar (SAR) imagery was used for flood inundation mapping, and the accompanying damages were investigated using Landsat derived Land cover maps of Koton Karfe during the 2022 devastating flood. Overall, the results obtained in this study show that the regions that felt the impact of the flood the most were the southern and western areas, which must have experienced such an impact due to their proximity to the rivers Niger and Benue and also the water coming from the upper stream part of Cameroon. Further findings revealed that during the flood period on October 13, 2022, the total inundated area in Koton Karfe was estimated to be 198.255 sq. km. In terms of damage assessment, the urban areas had reduced from 220.902 sq. km in May 2022 to 87.473 sq. km in November 2022. This shows that over 133 sq. km of the urban settlement have been lost, indicating that lives must have been lost, properties too, and humans must have been displaced. This research will assist in the space of flood emergency response and disaster management.

Characterization of flood and drought hazards on the Gereb-Geba water supply dam in the semi-arid northern Ethiopian highlands

Evaluating the flood and drought hazards provides vital information for sustainable water resources management, particularly in semi-arid, water-deficit environments. Most prior studies were limited in exploring the flood and drought hazards, which are important for early warning systems and preparedness. This study characterized the hydrological extreme hazards on the Gereb-Geba reservoir, namely the Suluh, Genfel, and Agula rivers. Flood frequency analysis was performed using the fitted flood frequency distribution in MATLAB. The 2D hydrodynamic model HEC-RAS was implemented to produce a flood-inundation map. Meteorological, agricultural, and hydrological droughts were analyzed using the Standardized Precipitation Index (SPI), Vegetation Condition Index (VCI), and Streamflow Drought Index (SDI), respectively. Using the Generalized Extreme Value (GEV), the estimated flood magnitude showed an increasing tendency in all the rivers across all the return periods (2-, 5-, 10-, 20-, 50-, and 100-years). The reservoir inundated an area of 12.8 km2 at an elevation of 1830 m.a.s.l. with a water depth of 80 m at the outlet. Suluh experienced more severe to extreme hydrological drought episodes than the Agula and Genfel rivers. Severe to extreme meteorological droughts were also observed in the respective catchments. Moreover, severe agricultural drought prevalence was also detected across all the river catchments. This study provides vital and comprehensive flood and drought information for water resources planning, management, and development.

XFIMNet: an Explainable deep learning architecture for versatile flood inundation mapping with synthetic aperture radar and multi-spectral optical images

ABSTRACT Floods are among the most dangerous natural disasters, and their frequency and severity are rapidly increasing due to climate change. Flood inundation mapping is critical for better understanding which areas are prone to flooding and identifying emerging floods. Recently, deep learning algorithms have been developed to improve the accuracy and timeliness of the production of these maps. In this study a new architecture, named Explainable Flood Inundation Mapping Network (XFIMNet) is presented, which is an explainable deep learning segmentation model for producing flood inundation maps from both synthetic aperture radar (SAR) and multi-spectral optical images. Comparative analysis with state-of-the-art (SOTA) models is performed, where the versatility of the model is demonstrated, in that it achieves superior performance across both image types, while the SOTA models are less consistent. The models are further evaluated with only the red, green, and blue (RGB) bands of the Sentinel-2 images, demonstrating that the full 13-band image is the superior choice for training a deep learning model. XFIMNet achieves an Intersection Over Union (IOU) of 0.59 with SAR images, 0.7 with optical, and 0.47 with the RGB bands, outperforming each of the SOTA models, as well as providing the most detailed and accurate visual segmentation mask, with minimal increase in the computation time for training or inference. The findings of this study highlight that XFIMNet is not only able to provide the most accurate segmentation mask but also prioritizes the most appropriate features when making decisions across both image types.

Flood Hazard Mapping and Risk Assessment in Narayani River Basin, Nepal

Floods, earthquakes, forest fires, landslides, and other natural hazard are common in Nepal. Among them, the flood is one of the natural disasters and that occurred in the Narayani River basin. The Narayani River Basin's exposures and vulnerabilities are at danger from a flood catastrophe of this kind, hence this studies attempts to reduce and control the risk of flooding in order to better manage disasters. The factors of flood hazard, flood exposures, and flood vulnerability are investigated as part of the ward-level flood risk assessment, which aims to prevent and manage the flood disaster. The hydraulic model (HEC-RAS) for the 2018 flood event was used in this study to create the GIS- based modeling of the flood inundation maps. Additionally, it calculated the various return periods for floods in the river basin—5, 10, 25, 50, 75, and 100 years. Furthermore, the flood extent was confirmed using the flood map produced by the Google Earth Engine (GEE) using remotely sensed techniques. For Hydraulic modelling, ALOS/PALSAR Digital Elevation Model (12.5m.) spatial resolution was used. In addition, RAS-MAPPER generated the geometric data for the hydraulic modeling, which was then transferred into HEC-RAS. This data included the cross-section, flow route, streamline, and bank lines. On both sides of the river, the necessary Manning value "n" values were computed for every cross-section. The steady-flow models of the anticipated flood hydrographs were created using the hydraulic model. Google Earth Engine (GEE) flood maps generated from Sentinel-1C radar satellite data were used to validate the results for the 2018 flood events. When comparing the simulated result's flood inundation area with the remote sensing data's flood area, the overlap area for the 2018 flood event is 65%. Additionally, the flood area is verified. In addition, the hydraulic model generated flow conditions for 5, 10, 25, 50, 75, and 100 year return periods. The river basin's surface water level and flood extent are both progressively rising. In order to effectively manage and prepare for potential flood hazards in the study area, an analysis of the flood risk assessment was conducted by taking into consideration three primary factors: the flood hazard map, flood exposure, and flood vulnerability. The layers of the population, crops, schools, hospitals, and road network were all exposed to flooding, and the factors that determined flood vulnerabilities were literacy, urban area, and agecomposition (less than 14 and more than 65). The higher flood risk area was found in Ward number 1, 3, 4,16,18,26 of Bharatpur metropolitan, Gaindakot municipality ward number 1 and 12, 15 of Madhyabindu Municipality.

Optimal Fusion of Multispectral Optical and SAR Images for Flood Inundation Mapping through Explainable Deep Learning

In the face of increasing flood risks intensified by climate change, accurate flood inundation mapping is pivotal for effective disaster management. This study introduces a novel explainable deep learning architecture designed to generate precise flood inundation maps from diverse satellite data sources. A comprehensive evaluation of the proposed model is conducted, comparing it with state-of-the-art models across various fusion configurations of Multispectral Optical and Synthetic Aperture Radar (SAR) images. The proposed model consistently outperforms other models across both Sentinel-1 and Sentinel-2 images, achieving an Intersection Over Union (IOU) of 0.5862 and 0.7031, respectively. Furthermore, analysis of the different fusion combinations reveals that the use of Sentinel-1 in combination with RGB, NIR, and SWIR achieves the highest IOU of 0.7053 and that the inclusion of the SWIR band has the greatest positive impact on the results. Gradient-weighted class activation mapping is employed to provide insights into its decision-making processes, enhancing transparency and interpretability. This research contributes significantly to the field of flood inundation mapping, offering an efficient model suitable for diverse applications. This study not only advances flood inundation mapping but also provides a valuable tool for improved understanding of deep learning decision-making in this area, ultimately contributing to improved disaster management strategies.

Analysis of Flood Inundation Mapping Using the Weighting Method in Urban Floods

Floods are one of the most destructive natural disasters globally and are projected to increase in many regions of the world. One of the areas affected by the floods is Barito Kuala Regency in South Kalimantan. Flood hazard maps and flood inundation maps are one way to identify areas prone to flooding and to find out the area of inundated areas due to flooding. Methods of scoring and weighting the parameters that cause flood inundation to map an area that has the potential for flooding. Flood inundation maps can be generated using the Normalized Difference Water Index (NDWI) method. The purpose of these two analyzes is to determine the distribution of flood hazard levels in Barito Kuala District, to evaluate flood-prone areas with flood-prone maps, and to determine the effectiveness of the two methods in identifying flood-prone areas and to determine the accuracy of the resulting flood inundation maps. Flood-prone areas were identified using the scoring method and parameter weighting to determine the distribution of flood vulnerability levels in Barito Kuala District. Parameters of flood vulnerability level used include land cover, slope, rainfall, drainage density, and soil type. Flood vulnerability level parameters were obtained from NDWI analysis on ArcGIS and ENVI. Image data during the dry season and rainy season are thresholded at <0.3 to eliminate data that is considered non-water resulting in flood inundation. The results of processing the level of flood hazard in Barito Kuala Regency are divided into 4 classes, namely the safe class with 0.5%, the less prone class is 10.11%, the moderate class is 89.37 and the vulnerable class is 0.00% of the area of Barito Kuala Regency with the majority of the areas in the moderate class. Savings District. The results of the analysis of the flood inundation map and the flood hazard map show that the moderate class of flood inundation is 72.06% of the total area of flood inundation