A Review of Machine Learning Models for Harmful Algal Bloom Monitoring in Freshwater Systems DOI Open Access
Ibrahim Busari,

Debabrata Sahoo,

R. Daren Harmel

et al.

Journal of Natural Resources and Agricultural Ecosystems, Journal Year: 2023, Volume and Issue: 1(2), P. 63 - 76

Published: Jan. 1, 2023

Highlights Machine Learning (ML) models are identified, reviewed, and analyzed for HAB predictions. Data preprocessing is vital efficient ML model development. toxin production monitoring limited. Abstract. Harmful algal blooms (HABs) detrimental to livestock, humans, pets, the environment, global economy, which calls a robust approach their management. While process-based can inform practitioners about enabling conditions, they have inherent limitations in accurately predicting harmful blooms. To address these limitations, potentially leverage large volumes of IoT data aid near real-time evolved as tools understanding patterns relationships between water quality parameters expansion. This review describes currently used forecasting HABs freshwater ecosystems presents structures application related toxins. The revealed that regression trees, random forest, Artificial Neural Network (ANN), Support Vector Regression (SVR), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU) most frequently monitoring. shows models' prowess identifying significant variables influencing growth, drivers, multistep prediction. Hybrid also improve prediction algal-related through improved optimization techniques variable selection algorithms. often focus on biomass prediction, few studies apply limitation be associated with lack high-frequency datasets development, exploring this domain encouraged. serves guide policymakers researchers implement reveals potential decision support early Keywords: Cyanobacteria, Freshwater, blooms, learning, Water quality.

Language: Английский

Harmful Algal Blooms in Eutrophic Marine Environments: Causes, Monitoring, and Treatment DOI Open Access

Jiaxin Lan,

Pengfei Liu,

Xi Hu

et al.

Water, Journal Year: 2024, Volume and Issue: 16(17), P. 2525 - 2525

Published: Sept. 5, 2024

Marine eutrophication, primarily driven by nutrient over input from agricultural runoff, wastewater discharge, and atmospheric deposition, leads to harmful algal blooms (HABs) that pose a severe threat marine ecosystems. This review explores the causes, monitoring methods, control strategies for eutrophication in environments. Monitoring techniques include remote sensing, automated situ sensors, modeling, forecasting, metagenomics. Remote sensing provides large-scale temporal spatial data, while sensors offer real-time, high-resolution monitoring. Modeling forecasting use historical data environmental variables predict blooms, metagenomics insights into microbial community dynamics. Control treatments encompass physical, chemical, biological treatments, as well advanced technologies like nanotechnology, electrocoagulation, ultrasonic treatment. Physical such aeration mixing, are effective but costly energy-intensive. Chemical including phosphorus precipitation, quickly reduce levels may have ecological side effects. Biological biomanipulation bioaugmentation, sustainable require careful management of interactions. Advanced innovative solutions with varying costs sustainability profiles. Comparing these methods highlights trade-offs between efficacy, cost, impact, emphasizing need integrated approaches tailored specific conditions. underscores importance combining mitigate adverse effects on

Language: Английский

Citations

24
Recent advances in algal bloom detection and prediction technology using machine learning DOI
Jungsu Park,

Keval K. Patel,

Woo Hyoung Lee

et al.

The Science of The Total Environment, Journal Year: 2024, Volume and Issue: 938, P. 173546 - 173546

Published: May 27, 2024

Language: Английский

Citations

16
Current status and prospects of algal bloom early warning technologies: A Review DOI
X.L. Xiao, Yazhou Peng, Wei Zhang

et al.

Journal of Environmental Management, Journal Year: 2023, Volume and Issue: 349, P. 119510 - 119510

Published: Nov. 9, 2023

Language: Английский

Citations

30
Spatiotemporal-aware machine learning approaches for dissolved oxygen prediction in coastal waters DOI
Wenzhao Liang, Tongcun Liu, Yuntao Wang

et al.

The Science of The Total Environment, Journal Year: 2023, Volume and Issue: 905, P. 167138 - 167138

Published: Sept. 19, 2023

Language: Английский

Citations

26
Real-Time Chlorophyll-a Forecasting using Machine Learning Framework with Dimension Reduction and Hyperspectral Data DOI
Doyun Kim,

KyoungJin Lee,

SeungMyeong Jeong

et al.

Environmental Research, Journal Year: 2024, Volume and Issue: 262, P. 119823 - 119823

Published: Aug. 22, 2024

Language: Английский

Citations

13
The Intelligent Path Planning System of Agricultural Robot via Reinforcement Learning DOI Creative Commons
Jiachen Yang,

Jingfei Ni,

Yang Li

et al.

Sensors, Journal Year: 2022, Volume and Issue: 22(12), P. 4316 - 4316

Published: June 7, 2022

Agricultural robots are one of the important means to promote agricultural modernization and improve efficiency. With development artificial intelligence technology maturity Internet Things (IoT) technology, people put forward higher requirements for robots. must have intelligent control functions in scenarios be able autonomously decide paths complete tasks. In response this requirement, paper proposes a Residual-like Soft Actor Critic (R-SAC) algorithm realize safe obstacle avoidance path planning addition, order alleviate time-consuming problem exploration process reinforcement learning, an offline expert experience pre-training method, which improves training efficiency learning. Moreover, optimizes reward mechanism by using multi-step TD-error, solves probable dilemma during training. Experiments verify that our proposed method has stable performance both static dynamic environments, is superior other learning algorithms. It efficient visible application potential

Language: Английский

Citations

33
Data and domain knowledge dual‐driven artificial intelligence: Survey, applications, and challenges DOI Open Access
Jing Nie, Jiachen Jiang, Yang Li

et al.

Expert Systems, Journal Year: 2023, Volume and Issue: 42(1)

Published: Aug. 14, 2023

Abstract At present, the mainstream mode of machine learning algorithms is data‐driven method, which mainly relies on self‐learning ability deep neural networks and continuously evolving models in training. However, pure method has some critical problems, such as high data collection cost, poor interpretability easy to be disturbed by noise. Although knowledge‐driven stability, it lacks evolution face comprehensive complex problems. In recent years, convergence domain knowledge combined advantages both paradigms. One typical way embed into model improve model, then use explore knowledge, iterate form a closed loop. The data‐knowledge dual‐driven methods have brought transformative innovations learning. This review first introduced necessity field artificial intelligence. Then, applications smart marine were introduced. Finally, challenges trends intelligence are anticipated.

Language: Английский

Citations

21
Modeling of algal blooms: Advances, applications and prospects DOI

Yichong Wang,

Chao Xu, Qianru Lin

et al.

Ocean & Coastal Management, Journal Year: 2024, Volume and Issue: 255, P. 107250 - 107250

Published: June 24, 2024

Language: Английский

Citations

8
Monitoring Harmful Algal Blooms and Water Quality Using Sentinel-3 OLCI Satellite Imagery with Machine Learning DOI Creative Commons

Neha Joshi,

Jongmin Park, Kaiguang Zhao

et al.

Remote Sensing, Journal Year: 2024, Volume and Issue: 16(13), P. 2444 - 2444

Published: July 3, 2024

Cyanobacterial harmful algal blooms release toxins and form thick blanket layers on the water surface causing widespread problems, including serious threats to human health, ecosystem, economics, recreation. To identify potential drivers for bloom, there is a need extensive observations of sources with bloom occurrences. However, traditional methods monitoring sources, such as collection point ground samples, have proven limited due spatial temporal variability resources, cost associated collecting samples that accurately represent this variability. These limitations can be addressed through use high-frequency satellite data. In study, we explored Random Forest (RF), which one widely used machine learning architectures, evaluate performance Sentinel-3 OLCI (Ocean Land Color Imager) images in predicting proxies western region Lake Erie. The sixteen available bands were predictor variables, while four cyanobacterial masses, Chlorophyll-a, Microcystin, Phycocyanin, Secchi-depth, considered response variables RF models, model per proxy. Each comes unique set traits help detection. Among Chlorophyll-a performed best R2 = 0.55 RMSE 20.84 µg/L, rest other was less than 0.5. This because most dominant optically active pigment water, strong indicator present low concentrations. Additionally, responsible toxicity, has spectral sensitivity, Secchi-depth could influenced by various factors besides blooms, colored dissolved organic inorganic matter. On further examining relationship between proxies, Microcystin significantly correlated enhances usefulness identifying presence blooms.

Language: Английский

Citations

5
IOT based prediction of rainfall forecast in coastal regions using deep reinforcement model DOI Creative Commons

J. Nithyashri,

Ravi Kumar Poluru, Subramanian Balakrishnan

et al.

Measurement Sensors, Journal Year: 2023, Volume and Issue: 29, P. 100877 - 100877

Published: Aug. 8, 2023

This research proposes an IoT based technique for predicting rainfall forecast in coastal regions using a deep reinforcement learning model. The proposed utilizes Long Short-Term Memory (LSTM) networks to capture the temporal dependencies between data collected from and prediction model parameters. is evaluated on dataset of India compared traditional methods forecasting. accuracy reliability these models are by comparing them prior models. Precipitation locations may be predicted with average 89% suggested model, as shown results. framework computationally efficient can trained little input. results this give strong evidence that effective tool precipitation

Language: Английский

Citations

12