Deep learning methods for flood mapping: a review of existing applications and future research directions

Hydrology and earth system sciences, Год журнала: 2022, Номер 26(16), С. 4345 - 4378

Опубликована: Авг. 25, 2022

Abstract. Deep learning techniques have been increasingly used in flood management to overcome the limitations of accurate, yet slow, numerical models and improve results traditional methods for mapping. In this paper, we review 58 recent publications outline state art field, identify knowledge gaps, propose future research directions. The focuses on type deep various mapping applications, types considered, spatial scale studied events, data model development. show that based convolutional layers are usually more as they leverage inductive biases better process characteristics flooding events. Models fully connected layers, instead, provide accurate when coupled with other statistical models. showed increased accuracy compared approaches speed methods. While there exist several applications susceptibility, inundation, hazard mapping, work is needed understand how can assist real-time warning during an emergency it be employed estimate risk. A major challenge lies developing generalize unseen case studies. Furthermore, all reviewed their outputs deterministic, limited considerations uncertainties outcomes probabilistic predictions. authors argue these identified gaps addressed by exploiting fundamental advancements or taking inspiration from developments applied areas. graph neural networks operators arbitrarily structured thus should capable generalizing across different studies could account complex interactions natural built environment. Physics-based preserve underlying physical equations resulting reliable speed-up alternatives Similarly, resorting Gaussian processes Bayesian networks.

Язык: Английский

---

The future of coastal monitoring through satellite remote sensing

Cambridge Prisms Coastal Futures, Год журнала: 2022, Номер 1

Опубликована: Ноя. 28, 2022

Abstract Satellite remote sensing is transforming coastal science from a “data-poor” field into “data-rich” field. Sandy beaches are dynamic landscapes that change in response to long-term pressures, short-term pulses, and anthropogenic interventions. Until recently, the rate breadth of beach have outpaced our ability monitor those changes, due spatiotemporal limitations observational capacity. Over past several decades, only handful worldwide been regularly monitored with accurate yet expensive situ surveys. The coastal-change data these few well-monitored led in-depth understanding many site-specific processes. However, because best-monitored not representative all beaches, much remains unknown about processes fate other >99% unmonitored worldwide. fleet Earth-observing satellites has enabled multiscale monitoring for very first time, by providing imagery global coverage up daily frequency. long-standing ever-expanding archive satellite will enable scientists investigate at sites vulnerable future sea-level rise, is, (almost) everywhere. In decade, capability observe space grown substantially computing algorithmic power. Yet, further advances needed automating using machine learning, deep computer vision fully leverage this massive treasure trove data. Extensive investigation causes effects requisite scales provide managers additional, valuable information evaluate problems solutions, addressing potential widespread loss accelerated development, reduced sediment supply. Monitoring currently means seamless high resolution scale impending impacts climate on systems.

Язык: Английский

---

Recent Advances and New Frontiers in Riverine and Coastal Flood Modeling

Reviews of Geophysics, Год журнала: 2023, Номер 61(2)

Опубликована: Май 27, 2023

Abstract Over the past decades, scientific community has made significant efforts to simulate flooding conditions using a variety of complex physically based models. Despite all advances, these models still fall short in accuracy and reliability are often considered computationally intensive be fully operational. This could attributed insufficient comprehension causative mechanisms flood processes, assumptions model development inadequate consideration uncertainties. We suggest adopting an approach that accounts for influence human activities, soil saturation, snow topography, river morphology, land‐use type enhance our understanding generating mechanisms. also recommend transition innovative earth system modeling frameworks where interaction among components simultaneously modeled. Additionally, more nonselective rigorous studies should conducted provide detailed comparison physical simplified methods inundation mapping. Linking process‐based with data‐driven/statistical offers opportunities yet explored conveyed researchers emergency managers. The main contribution this paper is notify scientists practitioners latest developments characterization modeling, identify challenges associated uncertainties risks coupled hydrologic hydrodynamic forecasting mapping, potential use state‐of‐the‐art data assimilation machine learning tackle complexities involved transitioning such operation.

Язык: Английский

---

Fast Flood Extent Monitoring With SAR Change Detection Using Google Earth Engine

IEEE Transactions on Geoscience and Remote Sensing, Год журнала: 2023, Номер 61, С. 1 - 19

Опубликована: Янв. 1, 2023

Flooding is one of the most frequent and disastrous natural hazards triggered by extreme precipitation, high river runoff, hurricane storm surges, compounding effects various flood drivers. This study introduces a new multisource remote sensing approach that leverages both multispectral optical imagery weather- illumination-independent characteristics synthetic aperture radar (SAR) data to streamline, automate, map geographically reliable inundation extents. Utilizing near real-time cloud computing capabilities Google Earth Engine (GEE), this process facilitates acquisition enables large-scale monitoring in an expeditious manner. Two major hurricanes along U.S. Gulf Coast were evaluated: 1) 2021 Hurricane Ida south New Orleans, LA, USA, 2) 2017 Harvey east Houston, TX, USA. We devised change detection thresholding framework using multitemporal SAR validated results with extent maps derived from Landsat 8 Sentinel-2 imagery. demonstrate constant threshold values for extraction indices are not ubiquitously suitable all geographies; thus, we outline heuristic can be used select thresholds specific sites through fully automated sensitivity analysis. The indicated agreement between (77%–80%), providing benefit under-cloud detection. Furthermore, our contribute scaling produce rapid accurate information decision-makers emergency responders during time-sensitive events.

Язык: Английский

---

Between flood and drought: How cities are facing water surplus and scarcity

Journal of Environmental Management, Год журнала: 2023, Номер 345, С. 118557 - 118557

Опубликована: Июль 8, 2023

Язык: Английский

---

Flood Detection with SAR: A Review of Techniques and Datasets

Remote Sensing, Год журнала: 2024, Номер 16(4), С. 656 - 656

Опубликована: Фев. 10, 2024

Floods are among the most severe and impacting natural disasters. Their occurrence rate intensity have been significantly increasing worldwide in last years due to climate change urbanization, bringing unprecedented effects on human lives activities. Hence, providing a prompt response flooding events is of crucial relevance for humanitarian, social economic reasons. Satellite remote sensing using synthetic aperture radar (SAR) offers great deal support facing flood mitigating their global scale. As opposed multi-spectral sensors, SAR important advantages, as it enables Earth’s surface imaging regardless weather sunlight illumination conditions. In decade, availability data, even at no cost, thanks efforts international national space agencies, has deeply stimulating research activities every Earth observation field, including mapping monitoring, where advanced processing paradigms, e.g., fuzzy logic, machine learning, data fusion, applied, demonstrating superiority with respect traditional classification strategies. However, fair assessment performance reliability techniques key importance an efficient disasters and, hence, should be addressed carefully quantitative basis trough quality metrics high-quality reference data. To this end, recent development open datasets specifically covering related ground-truth can thorough objective validation well reproducibility results. Notwithstanding, SAR-based monitoring still suffers from limitations, especially vegetated urban areas, complex scattering mechanisms impair accurate extraction water regions. All such aspects, methodologies, datasets, strategies, challenges future perspectives described discussed.

Язык: Английский

---

Assessment of surrogate models for flood inundation: The physics-guided LSG model vs. state-of-the-art machine learning models

Water Research, Год журнала: 2024, Номер 252, С. 121202 - 121202

Опубликована: Янв. 24, 2024

Hydrodynamic models can accurately simulate flood inundation but are limited by their high computational demand that scales non-linearly with model complexity, resolution, and domain size. Therefore, it is often not feasible to use high-resolution hydrodynamic for real-time predictions or when a large number of needed probabilistic design. Computationally efficient surrogate have been developed address this issue. The recently Low-fidelity, Spatial analysis, Gaussian Process Learning (LSG) has shown strong performance in both efficiency simulation accuracy. LSG physics-guided simulates first using an extremely coarse simplified (i.e. low-fidelity) provide initial estimate inundation. Then, the low-fidelity upskilled via Empirical Orthogonal Functions (EOF) analysis Sparse accurate predictions. Despite promising results achieved thus far, benchmarked against other models. Such comparison fully understand value guidance future research efforts simulation. This study compares four state-of-the-art assessed ability temporal spatial evolution events within beyond range used training. evaluated three distinct case studies Australia United Kingdom. found be superior accuracy extent water depth, including applied outside training data used, while achieving efficiency. In addition, play crucial role overall model.

Язык: Английский

---

Residual wave vision U-Net for flood mapping using dual polarization Sentinel-1 SAR imagery

International Journal of Applied Earth Observation and Geoinformation, Год журнала: 2024, Номер 127, С. 103662 - 103662

Опубликована: Янв. 21, 2024

The increasing severity, duration, and frequency of destructive floods can be attributed to shifts in climate, infrastructure, land use, population demographics. Obtaining precise timely data about the extent floodwaters is crucial for effective emergency preparedness mitigation efforts. Deep convolutional neural networks (CNNs) have shown astonishing effectiveness various remote sensing applications, including flood mapping. One key limitations CNNs that they only predict whether a desired feature will appear an image, not where it recognized. To address this limitation, incorporation self-attention mechanisms deployed vision transformers (ViTs) particularly effective. However, modules ViTs are complex computationally expensive, require wealth ground attain their full capability image classification/segmentation. Thus, paper, we develop Residual Wave Vision U-Net (WVResU-Net), deep learning segmentation architecture utilizes advanced Multi-Layer Perceptrons (MLPs) ResU-Net accurate reliable mapping using Sentinel-1 SAR's dual polarization data. Results showed significant superiority developed WVResU-Net algorithms over several well-known CNN ViT models, Swin U-Net, U-Net+++, Attention R2U-Net, ResU-Net, TransU-Net TransU-Net++. For example, accuracy TransU-Net++, SwinU-Net, TransU-Net, was significantly improved by approximately 5, 12, 13, 16, 19, 23 percentage points, respectively terms recall obtained with value 69.67%. code made publicly available at https://github.com/aj1365/RWVUNet.

Язык: Английский

---

Deep artificial intelligence applications for natural disaster management systems: A methodological review

Ecological Indicators, Год журнала: 2024, Номер 163, С. 112067 - 112067

Опубликована: Май 6, 2024

Deep learning techniques through semantic segmentation networks have been widely used for natural disaster analysis and response. The underlying base of these implementations relies on convolutional neural (CNNs) that can accurately precisely identify locate the respective areas interest within satellite imagery or other forms remote sensing data, thereby assisting in evaluation, rescue planning, restoration endeavours. Most CNN-based deep-learning models encounter challenges related to loss spatial information insufficient feature representation. This issue be attributed their suboptimal design layers capture multiscale-context failure include optimal during pooling procedures. In early CNNs, network encodes elementary representations, such as edges corners, whereas, progresses toward later layers, it more intricate characteristics, complicated geometric shapes. theory, is advantageous a extract features from several levels because generally yield improved results when both simple maps are employed together. study comprehensively reviews current developments deep methodologies segment images associated with disasters. Several popular models, SegNet U-Net, FCNs, FCDenseNet, PSPNet, HRNet, DeepLab, exhibited notable achievements various applications, including forest fire delineation, flood mapping, earthquake damage assessment. These demonstrate high level efficacy distinguishing between different land cover types, detecting infrastructure has compromised damaged, identifying regions fire-susceptible further dangers.

Язык: Английский

---

Review article: A comprehensive review of compound flooding literature with a focus on coastal and estuarine regions

Natural hazards and earth system sciences, Год журнала: 2025, Номер 25(2), С. 747 - 816

Опубликована: Фев. 20, 2025

Abstract. Compound flooding, where the combination or successive occurrence of two more flood drivers leads to a greater impact, can exacerbate adverse consequences particularly in coastal–estuarine regions. This paper reviews practices and trends compound research synthesizes regional global findings. A systematic review is employed construct literature database 279 studies relevant flooding context. explores types events their mechanistic processes, it terminology throughout literature. Considered are six (fluvial, pluvial, coastal, groundwater, damming/dam failure, tsunami) five precursor environmental conditions (soil moisture, snow, temp/heat, fire, drought). Furthermore, this summarizes methodology study application trends, as well considers influences climate change urban environments. Finally, highlights knowledge gaps discusses implications on future practices. Our recommendations for (1) adopt consistent approaches, (2) expand geographic coverage research, (3) pursue inter-comparison projects, (4) develop modelling frameworks that better couple dynamic Earth systems, (5) design coastal infrastructure with compounding mind.

Язык: Английский

---