Explainable AI approach with original vegetation data classifies spatio-temporal nitrogen in flows from ungauged catchments to the Great Barrier Reef DOI Creative Commons

Cherie M. O’Sullivan,

Ravinesh C. Deo, Afshin Ghahramani

et al.

Scientific Reports, Journal Year: 2023, Volume and Issue: 13(1)

Published: Oct. 24, 2023

Transfer of processed data and parameters to ungauged catchments from the most similar gauged counterpart is a common technique in water quality modelling. But catchment similarities for Dissolved Inorganic Nitrogen (DIN) are ill posed, which affects predictive capability models reliant on such methods simulating DIN. Spatial proxies classify DIN responses demonstrated solution, yet their applicability unexplored. We adopted neural network pattern recognition model (ANN-PR) explainable artificial intelligence approach (SHAP-XAI) match all that flow Great Barrier Reef ones based proxy spatial data. Catchment suitability was verified using (ANN-WQ) simulator trained datasets, tested by matched unsupervised learning scenarios. show discriminating training regime benefits ANN-WQ simulation performance scenarios ( p< 0.05). This phenomenon demonstrates useful tool with regimes. Catchments lacking similarity identified as priority monitoring areas gain observed regimes Reef, Australia.

Language: Английский

Deep learning insights into suspended sediment concentrations across the conterminous United States: Strengths and limitations DOI
Yalan Song,

Piyaphat Chaemchuen,

Farshid Rahmani

et al.

Journal of Hydrology, Journal Year: 2024, Volume and Issue: 639, P. 131573 - 131573

Published: June 24, 2024

Language: Английский

Citations

4
A Hybrid Machine Learning Model for Modeling Nitrate Concentration in Water Sources DOI

Adnan Mazraeh,

Meysam Bagherifar,

Saeid Shabanlou

et al.

Water Air & Soil Pollution, Journal Year: 2023, Volume and Issue: 234(11)

Published: Nov. 1, 2023

Language: Английский

Citations

10
Enhancing prediction and inference of daily in-stream nutrient and sediment concentrations using an extreme gradient boosting based water quality estimation tool - XGBest DOI Creative Commons
Shubham Jain, Arun Bawa,

Katie Mendoza

et al.

The Science of The Total Environment, Journal Year: 2025, Volume and Issue: 963, P. 178517 - 178517

Published: Jan. 20, 2025

Language: Английский

Citations

0
Artificial Intelligence in Aquatic Biodiversity Research: A PRISMA-Based Systematic Review DOI Creative Commons
Tymoteusz Miller, Grzegorz Michoński, Irmina Durlik

et al.

Biology, Journal Year: 2025, Volume and Issue: 14(5), P. 520 - 520

Published: May 8, 2025

Freshwater ecosystems are increasingly threatened by climate change and anthropogenic activities, necessitating innovative scalable monitoring solutions. Artificial intelligence (AI) has emerged as a transformative tool in aquatic biodiversity research, enabling automated species identification, predictive habitat modeling, conservation planning. This systematic review follows the PRISMA framework to analyze AI applications freshwater studies. Using structured literature search across Scopus, Web of Science, Google Scholar, we identified 312 relevant studies published between 2010 2024. categorizes into assessment, ecological risk evaluation, strategies. A bias assessment was conducted using QUADAS-2 RoB 2 frameworks, highlighting methodological challenges, such measurement inconsistencies model validation. The citation trends demonstrate exponential growth AI-driven with leading contributions from China, United States, India. Despite growing use this field, also reveals several persistent including limited data availability, regional imbalances, concerns related generalizability transparency. Our findings underscore AI’s potential revolutionizing but emphasize need for standardized methodologies, improved integration, interdisciplinary collaboration enhance insights efforts.

Language: Английский

Citations

0
Performance of LSTM over SWAT in Rainfall-Runoff Modeling in a Small, Forested Watershed: A Case Study of Cork Brook, RI DOI Open Access

Shiva Gopal Shrestha,

Soni M. Pradhanang

Water, Journal Year: 2023, Volume and Issue: 15(23), P. 4194 - 4194

Published: Dec. 4, 2023

The general practice of rainfall-runoff model development towards physically based and spatially explicit representations hydrological processes is data-intensive computationally expensive. Physically models such as the Soil Water Assessment tool (SWAT) demand spatio-temporal data expert knowledge. Also, difficulty complexity compounded in smaller watershed due to constraint models’ inability generalize hydrologic processes. Data-driven can bridge this gap with their mathematical formulation. Long Short-Term Memory (LSTM) a data-driven Recurrent Neural Network (RNN) architecture, which better suited solve time series problems. Studies have shown that LSTM competitive performance hydrology studies. In study, comparative analysis SWAT Cork Brook shows results from were flow prediction NSE 0.6 against 0.63, respectively, given limited availability data. do not overestimate high flows like SWAT. However, both these struggle low values estimation. Although interpretability, explainability, use across different datasets or events outside training may be challenging, are robust efficient.

Language: Английский

Citations

9
Estimating soil mineral nitrogen from data-sparse field experiments using crop model-guided deep learning approach DOI
Rishabh Gupta,

Satya Krishna Pothapragada,

Weihuang Xu

et al.

Computers and Electronics in Agriculture, Journal Year: 2024, Volume and Issue: 225, P. 109355 - 109355

Published: Aug. 22, 2024

Language: Английский

Citations

3
Comparative efficiency of the SWAT model and a deep learning model in estimating nitrate loads at the Tuckahoe creek watershed, Maryland DOI
Jiye Lee, Dongho Kim,

Seokmin Hong

et al.

The Science of The Total Environment, Journal Year: 2024, Volume and Issue: 954, P. 176256 - 176256

Published: Sept. 20, 2024

Language: Английский

Citations

3
Deep Dive into Global Hydrologic Simulations: Harnessing the Power of Deep Learning and Physics-informed Differentiable Models (δHBV-globe1.0-hydroDL) DOI Creative Commons
Dapeng Feng, Hylke E. Beck, Jens de Bruijn

et al.

Published: Oct. 5, 2023

Abstract. Accurate hydrological modeling is vital to characterizing how the terrestrial water cycle responds climate change. Pure deep learning (DL) models have shown outperform process-based ones while remaining difficult interpret. More recently, differentiable, physics-informed machine with a physical backbone can systematically integrate equations and DL, predicting untrained variables processes high performance. However, it was unclear if such are competitive for global-scale applications simple backbone. Therefore, we use – first time at this scale differentiable hydrologic (fullname δHBV-globe1.0-hydroDL shorthanded δHBV) simulate rainfall-runoff 3753 basins around world. Moreover, compare δHBV purely data-driven long short-term memory (LSTM) model examine their strengths limitations. Both LSTM provide competent daily simulation capabilities in global basins, median Kling-Gupta efficiency values close or higher than 0.7 (and 0.78 subset of 1675 long-term records), significantly outperforming traditional models. regionalized demonstrated stronger spatial generalization ability (median KGE 0.64) parameter regionalization approach 0.46) even ungauged region tests Europe South America. Nevertheless, relative LSTM, hampered by structural deficiencies cold polar regions, highly arid significant human impacts. This study also sets benchmark estimates world builds foundations improving simulations.

Language: Английский

Citations

5
Global Daily Discharge Estimation Based on Grid-Scale Long Short-Term Memory (LSTM) Model and River Routing DOI Open Access
Yuan Yang, Dapeng Feng, Hylke E. Beck

et al.

Authorea (Authorea), Journal Year: 2023, Volume and Issue: unknown

Published: Oct. 14, 2023

Accurate global river discharge estimation is crucial for advancing our scientific understanding of the water cycle and supporting various downstream applications. In recent years, data-driven machine learning models, particularly Long Short-Term Memory (LSTM) model, have shown significant promise in estimating discharge. Despite this, applicability LSTM models remains largely unexplored. this study, we diverge from conventional basin-lumped modeling limited basins. For first time, apply an on a 0.25° grid, coupling it with routing model to estimate every reach worldwide. We rigorously evaluate performance over 5332 evaluation gauges globally period 2000-2020, separate training basins period. The grid-scale effectively captures rainfall-runoff behavior, reproducing high accuracy achieving median Kling-Gupta Efficiency (KGE) 0.563. It outperforms extensively bias-corrected calibrated benchmark simulation based Variable Infiltration Capacity (VIC) which achieved KGE 0.466. Using develop improved reach-level daily dataset spanning 1980 2020, named GRADES-hydroDL. This anticipated be useful myriad applications, including providing prior information Surface Water Ocean Topography (SWOT) satellite mission. openly available via Globus.

Language: Английский

Citations

5
Explainable deep learning identifies patterns and drivers of freshwater harmful algal blooms DOI Creative Commons
Shengyue Chen, Jinliang Huang, Jiacong Huang

et al.

Environmental Science and Ecotechnology, Journal Year: 2024, Volume and Issue: 23, P. 100522 - 100522

Published: Dec. 27, 2024

The escalating magnitude, frequency, and duration of harmful algal blooms (HABs) pose significant challenges to freshwater ecosystems worldwide. However, the mechanisms driving HABs remain poorly understood, in part due strong regional specificity processes uneven data availability. These complexities make it difficult generalize HAB dynamics effectively predict their occurrence using traditional models. To address these challenges, we developed an explainable deep learning approach long short-term memory (LSTM) models combined with explanation techniques that can capture complex patterns provide insights into key drivers. We applied this for density modeling at 102 sites China's lakes reservoirs over three years. LSTMs captured daily dynamics, achieving mean maximum Nash-Sutcliffe efficiency coefficients 0.48 0.95 during testing phase. Moreover, water temperature emerged as primary driver both nationally 30% localities, stronger sensitivity observed mid-to low-latitudes. also identified similarities allow successful transferability dynamics. Specifically, fine-tuned transfer learning, improved prediction accuracy 75% gauged areas. Overall, LSTM-based addresses by tackling limitations. By accurately predicting identifying critical drivers, provides actionable HABs, ultimately aids implementation effective mitigation measures nationwide ecosystems.

Language: Английский

Citations

1