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

International Journal of Remote Sensing, Journal Year: 2024, Volume and Issue: 45(4), P. 1059 - 1078

Published: Feb. 2, 2024

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

Language: Английский

---

A Comparison of Machine Learning Models for Predicting Flood Susceptibility Based on the Enhanced NHAND Method

Sustainability, Journal Year: 2023, Volume and Issue: 15(20), P. 14928 - 14928

Published: Oct. 16, 2023

A flood is a common and highly destructive natural disaster. Recently, machine learning methods have been widely used in susceptibility analysis. This paper proposes NHAND (New Height Above the Nearest Drainage) model as framework to evaluate effectiveness of both individual learners ensemble models addressing intricate flood-related challenges. The evaluation process encompasses critical dimensions such prediction accuracy, training duration, stability. Research findings reveal that, compared Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Lasso, Random Forest (RF), Extreme Gradient Boosting (XGBoost), Stacked Generalization (Stacking) outperforms terms predictive accuracy Meanwhile, XGBoost exhibits notable efficiency duration. Additionally, Shapley Additive Explanations (SHAP) method employed explain predictions made by XGBoost.

Language: Английский

---

EarthObsNet: A Comprehensive Benchmark Dataset for Data-Driven Earth Observation Image Synthesis

EarthArXiv (California Digital Library), Journal Year: 2024, Volume and Issue: unknown

Published: March 4, 2024

Remote Sensing imagery serves as an important data source for Earth surface monitoring and processes studies. It is highly likely that RS will become more indispensable in the future due to its high scalability compatibility with data-driven models ever-evolving software hardware increasingly good at processing large datasets. Although promising future, usage of observation imagery, such Landsat, Sentinel-2, Sentinel-1 images, has been largely limited retrospective studies, where those images serve mainly documentations past events. Recently, there are attempts expand current forward-looking applications support decision-making fast response against natural hazards. Unlike many well-defined well-studied topics change detection semantic segmentation which benchmark datasets openly available, so far, public image synthesis tasks prototyping comparison. To close this gap, we introduced a comprehensive dataset containing previous observations, precipitation, soil moisture, land cover, Height Above Nearest Drainage (HAND), DEM, slope collected during catastrophic 2019 Central US Flooding events lasted than two seasons Mississippi Missouri River tributaries. We also incorporated reference labels allow further investigation usefulness synthesized downstream applications, flood inundation mapping. hope provide essential goal attracting attention inspiring efforts broaden into applications.

Language: Английский

---

Geo-spatial analysis of built-environment exposure to flooding: Iowa case study

Discover Water, Journal Year: 2024, Volume and Issue: 4(1)

Published: May 23, 2024

Abstract Flooding is the most frequent type of natural disaster, inducing devastating damage at large and small spatial scales. Flood exposure analysis a critical part flood risk assessment. While studies analyze elements separately, it crucial to perform multi-parameter consider different types zones gain comprehensive understanding impact make informed mitigation decisions. This research analyzes population, properties, road networks potentially exposed 100, 200, 500-year events county level in State Iowa using geospatial analytics. We also propose index fuzzy overlay help find impacted county. During flooding, results indicate that county-level percentage displaced length can reach up 46%, 41%, 40%, respectively. found buildings roads are laid residential areas. Also, 25% counties designated as very high-exposure study many stakeholders identify vulnerable areas ensure equitable distribution investments resources toward projects.

Language: Английский

---

MultiRS Flood Mapper: A Google Earth Engine Application for Water Extent Mapping with Multimodal Remote Sensing and Quantile-Based Postprocessing

EarthArXiv (California Digital Library), Journal Year: 2023, Volume and Issue: unknown

Published: Dec. 11, 2023

Remote Sensing (RS) imagery is an important data source in surface water mapping applications thanks to its high spatial and temporal consistency scalability. The introduction of Google Earth Engine (GEE) has cleared some the major barriers fast large-scale RS-based geospatial analyses by providing easy open access most commonly used RS image products as well built-in functions designed for analysis. There a growing interest developing GEE that can work different regions time durations improve reusability scripts reduce manual effort during entire workflow water-body extraction. Despite all those advancements efforts, there still need creating are user-friendly serve both remote sensing experts students. These also expected be powerful comprehensive enough handle each step along lifecycle body extraction capable handling geomorphic discrepancies between under various configurations. Given these needs challenges, this study presents MultiRS Flood Mapper, application incorporates three (i.e., Sentinel-1 SAR, Landsat 8, Sentinel-2) integrates advanced dynamic thresholding algorithms postprocessing modules classification results influence dense vegetation cloud, with constrained hydraulic conditions. In addition, Mapper comes self-explanatory interface. Most functional processing require professional knowledge fully automated remaining function intuitive interactive way, which therefore enables have great potential broad audience backgrounds purposes.

Language: Английский

---

Improving Sentinel-1 Flood Maps Using a Topographic Index as Prior in Bayesian Inference

Water, Journal Year: 2023, Volume and Issue: 15(23), P. 4034 - 4034

Published: Nov. 21, 2023

Sentinel-1-based flood mapping works well but with well-known issues over rugged terrain. Applying exclusion masks to improve the results is common practice in unsupervised and global applications. One such mask height above nearest drainage (HAND), which uses terrain information reduce lookalikes SAR images. The TU Wien algorithm one operational workflow using this mask. Being a Bayesian method, can integrate auxiliary as prior probabilities classifications. This study improves by introducing HAND function instead of it We estimate optimal parameters observe performance flooded non-flooded scenarios six sites. compare maps generated (baseline) non-informed priors reference CEMS rapid extents. Our show enhanced decreasing false negatives at cost slightly increasing positives. In utilizing single parametrization, improved shows potential for implementation.

Language: Английский

---

EarthObsNet: A Comprehensive Benchmark Dataset for Data-Driven Earth Observation Image Synthesis

Environmental Modelling & Software, Journal Year: 2024, Volume and Issue: unknown, P. 106292 - 106292

Published: Dec. 1, 2024

Language: Английский

---

A New Data Processing Model for Distributed Urban Stagnant Analysis Based on Improved Yolov5 and Opencv

Published: May 25, 2023

In recent years, since flood disasters have brought immeasurable losses to the city, it is urgent prevent and solve of stagnant water. Considering shortage real-time accuracy hydrological analysis, Opencv technology used in this paper process obtained data real time. For improved Yolov5, BoTNet GAMAttention Transformer are improve Yolov5 enhance its ability recognition prediction better identify surface gathered The rate 7.1% higher than that Yolov7 1.7% Yolov5.After that, contour preprocessing image carried out through cropping identification frame eliminate relatively unstable factors. principle binocular distance measurement measure three-dimensional coordinates actual distance, constrain proportion picture, then get outline water, HSV combined with color processing pictures for water generation, area correspond corresponding parameters provide important help prevention storm drainage.

Language: Английский

---

Effects of High-Quality Elevation Data and Explanatory Variables on the Accuracy of Flood Inundation Mapping via Height Above Nearest Drainage

Published: July 10, 2023

Abstract. Given the availability of high quality and spatial resolution digital elevation models (DEMs) from United States Geological Survey’s 3-Dimensional Elevation Program (3DEP) derived mostly Light Detection Ranging sensors, we examined effects these DEMs at various resolutions on flood inundation map (FIM) extents a terrain index known as Height Above Nearest Drainage (HAND). We found that using improved resulting FIMs around 80 % catchments analyzed when compared to National Hydrography Dataset Plus High Resolution program. Additionally, varied 3DEP 3, 5, 10, 15, 20 meters results showed no significant overall effect FIM extent across resolutions. However, our experiments demonstrated burden computational time produce HAND. fit multiple linear regression model help explain catchment scale variation in four metrics employed lack reservoir flooding, or upstream river retention systems, was factor analysis. For validation, used Interagency Flood Risk Management Base Level Engineering produced streamflows 100 500 year event magnitudes sub-region Eastern Texas.

Language: Английский

---

A Novel Methodology for Enhancing Flood Risk Communication: The Nines of Safety

EarthArXiv (California Digital Library), Journal Year: 2024, Volume and Issue: unknown

Published: Jan. 12, 2024

Flood risk communication helps people plan for and recover from disasters, especially in flood-prone areas. The Nines of Safety (NoS) concept described this study provides a new perspective flood assessment. NoS method can help analyze comprehensively support decision-makers the public understand their vulnerability under various conditions. This novel approach considers physical parameters, socioeconomic factors, demographics to assess risk. analysis demonstrates that water characteristics are crucial determining safety. parameters deal with how income, age, population density affect flooding shows these factors scale. These variations highlight importance community-specific strategy. Explaining complexity assessment makes it more accessible. Given its quantitative qualitative effects, strategy could empower communities make sensible decisions adapt changing scenarios. better risks. Information on vulnerable individuals land use different profiles. discusses technique transform perceptions strengthen communities. By integrating into management, stakeholders may tailor responses each community, making them robust flooding.

Language: Английский

---