Runoff predictions in new-gauged basins using two transformer-based models

Journal of Hydrology, Год журнала: 2023, Номер 622, С. 129684 - 129684

Опубликована: Май 18, 2023

Язык: Английский

---

Urban flood susceptibility mapping using remote sensing, social sensing and an ensemble machine learning model

Sustainable Cities and Society, Год журнала: 2024, Номер 108, С. 105508 - 105508

Опубликована: Май 5, 2024

Язык: Английский

---

Deep learning rapid flood risk predictions for climate resilience planning

Journal of Hydrology, Год журнала: 2024, Номер 631, С. 130817 - 130817

Опубликована: Фев. 3, 2024

Язык: Английский

---

Streamflow prediction in ungauged catchments through use of catchment classification and deep learning

Journal of Hydrology, Год журнала: 2024, Номер 639, С. 131638 - 131638

Опубликована: Июль 3, 2024

Язык: Английский

---

Advancing Hydrology through Machine Learning: Insights, Challenges, and Future Directions Using the CAMELS, Caravan, GRDC, CHIRPS, PERSIANN, NLDAS, GLDAS, and GRACE Datasets

Water, Год журнала: 2024, Номер 16(13), С. 1904 - 1904

Опубликована: Июль 3, 2024

Machine learning (ML) applications in hydrology are revolutionizing our understanding and prediction of hydrological processes, driven by advancements artificial intelligence the availability large, high-quality datasets. This review explores current state ML hydrology, emphasizing utilization extensive datasets such as CAMELS, Caravan, GRDC, CHIRPS, NLDAS, GLDAS, PERSIANN, GRACE. These provide critical data for modeling various parameters, including streamflow, precipitation, groundwater levels, flood frequency, particularly data-scarce regions. We discuss type methods used significant successes achieved through those models, highlighting their enhanced predictive accuracy integration diverse sources. The also addresses challenges inherent applications, heterogeneity, spatial temporal inconsistencies, issues regarding downscaling LSH, need incorporating human activities. In addition to discussing limitations, this article highlights benefits utilizing high-resolution compared traditional ones. Additionally, we examine emerging trends future directions, real-time quantification uncertainties improve model reliability. place a strong emphasis on citizen science IoT collection hydrology. By synthesizing latest research, paper aims guide efforts leveraging large techniques advance enhance water resource management practices.

Язык: Английский

---

Advancing rapid urban flood prediction: a spatiotemporal deep learning approach with uneven rainfall and attention mechanism

Journal of Hydroinformatics, Год журнала: 2024, Номер 26(6), С. 1409 - 1424

Опубликована: Май 28, 2024

ABSTRACT Urban floods pose a significant threat to human communities, making its prediction essential for comprehensive flood risk assessment and the formulation of effective resource allocation strategies. Data-driven deep learning approaches have gained traction in urban emergency prediction, addressing efficiency constraints physical models. However, spatial structure rainfall, which has profound influence on flooding, is often overlooked many investigations. In this study, we introduce novel model known as CRU-Net equipped with an attention mechanism predict inundation depths terrains based spatiotemporal rainfall patterns. This method utilizes eight topographic parameters related height waterlogging, combined data inputs model. Comparative evaluations between developed two other models, U-Net ResU-Net, reveal that adeptly interprets traits accurately estimates depths, emphasizing flood-vulnerable regions. The demonstrates exceptional accuracy, evidenced by root mean square error 0.054 m Nash–Sutcliffe 0.975. also predicts over 80% locations exceeding 0.3 m. Remarkably, delivers predictions 3 million grids 2.9 s, showcasing efficiency.

Язык: Английский

---

Protocols for Water and Environmental Modeling Using Machine Learning in California

Hydrology, Год журнала: 2025, Номер 12(3), С. 59 - 59

Опубликована: Март 14, 2025

The recent surge in popularity of generative artificial intelligence (GenAI) tools like ChatGPT has reignited global interest AI, a technology with well-established history spanning several decades. California Department Water Resources (DWR) been at the forefront this field, leveraging Artificial Neural Networks (ANNs), core technique machine learning (ML), which is subfield for water and environmental modeling (WEM) since early 1990s. While protocols WEM exist California, they were designed primarily traditional statistical or process-based models that rely on predefined equations physical principles. In contrast, ML learn patterns from data require different development methodologies, existing do not address. This study, drawing DWR’s extensive experience ML, addresses gap by developing standardized implementation California. proposed cover four key phases implementation: (1) problem definition, ensuring clear objectives contextual understanding; (2) preparation, emphasizing collection, quality control, accessibility; (3) model development, advocating progression simple to hybrid ensemble approaches while integrating domain knowledge improved accuracy; (4) deployment, highlighting documentation, training, open-source practices enhance transparency collaboration. A case study provided demonstrate practical application these step step. Once implemented, can help achieve standardization, assurance, interoperability, using

Язык: Английский

---

On the relation between antecedent basin conditions and runoff coefficient for European floods

Journal of Hydrology, Год журнала: 2023, Номер 625, С. 130012 - 130012

Опубликована: Авг. 21, 2023

The event runoff coefficient (i.e. the ratio between and precipitation that originated runoff) is a key factor for understanding basin response to events. Runoff depends on intensity duration but also specific geohydrology attributes (including soil type, geology, land cover, topography ) last not least, antecedent (or pre-storm) conditions (i.e., amount of water stored in different hydrological compartments, like river, groundwater, snowpack). relation pre-storm critical flood forecasting, yet, where, when how much control coefficients still an open question. Here, we tested 60620 events across 284 basins Europe. To do so, derived from proxies, namely: precipitation; surface root zone moisture models, reanalyses models ingesting satellite observations; river discharge, and, total storage anomalies. We evaluated coupling strength proxies five classes European basins, defined based use type (as indexed by Soil Conservation Service curve number CN), topography, hydrology long-term climate their ability explain stormflow volume variability. found explains relatively well volumes both small large very peak especially floods. shows distributions correlates with deep storages (such as root-zone anomalies), discharge snow equivalent. Overall, these correlations depend class. Poor are against index despite its wide community. Seasonal interannual variability exert role inducing sharp changes correlation season climate. These results increase our coefficient. This will aid model calibration data assimilation. Furthermore, findings can help us better interpret future projections Europe expected long short-term climatic drivers.

Язык: Английский

---

Enhancing urban flood forecasting in drainage systems using dynamic ensemble-based data mining

Water Research, Год журнала: 2023, Номер 247, С. 120791 - 120791

Опубликована: Окт. 27, 2023

This study presents a novel approach for urban flood forecasting in drainage systems using dynamic ensemble-based data mining model which has yet to be utilised properly this context. The proposed method incorporates an event identification technique and rainfall feature extraction develop weak learner models. These models are then stacked create time-series ensemble decision tree algorithm confusion matrix-based blending method. was compared other commonly used real-world system the UK. results show that achieves higher hit rate benchmark models, with of around 85% vs 70 % next 3 h forecasting. Additionally, smart can accurately classify various timesteps or non-flood events without significant lag times, resulting fewer false alarms, reduced unnecessary risk management actions, lower costs real-time early warning applications. findings also demonstrate two features, "antecedent precipitation history" "seasonal time occurrence rainfall," significantly enhance accuracy ranging from 60 10 lead 15 min h.

Язык: Английский

---

A framework on utilizing of publicly availability stream gauges datasets and deep learning in estimating monthly basin-scale runoff in ungauged regions

Advances in Water Resources, Год журнала: 2024, Номер 188, С. 104694 - 104694

Опубликована: Апрель 9, 2024

Язык: Английский

---