Abstract:Harmful algal blooms (HABs) can have dire repercussions on aquatic wildlife and human health, and may negatively affect recreational uses, aesthetics, taste, and odor in drinking water. The factors that influence the occurrence and magnitude of harmful algal blooms and toxin production vary in space and time and remain poorly understood. It is within this context that we use machine learning (ML) and two 14-year (2005‚Äď2018) data sets on water quality and meteorological conditions of China's lakes and reservoirs to shed light on the magnitude and associated drivers of HAB events. General regression neural network (GRNN) models are developed to predict chlorophyll a concentrations for each lake and reservoir during two study periods (2005‚Äď2010 and 2011‚Äď2018). The developed models with an acceptable model fit are then analyzed by two indices to determine the areal HAB magnitudes and associated drivers. Our national assessment suggests that HAB magnitudes for China's lakes and reservoirs displayed a decreasing trend from 2006 (1363.3‚ÄĮkm2) to 2013 (665.2‚ÄĮkm2), and a slightly increasing trend from 2013 to 2018 (775.4‚ÄĮkm2). Among the 142 studied lakes and reservoirs, most severe HABs were found in Lakes Taihu, Dianchi and Chaohu with their contribution to the total HAB magnitude varying from 89.2% (2013) to 62.6% (2018). HABs in Lakes Taihu and Chaohu were strongly associated with both total phosphorus and nitrogen concentrations, while our results were inconclusive with respect to the predominant environmental factors shaping the eutrophication phenomena in Lake Dianchi. The present study provides evidence that effective HAB mitigation may require both nitrogen and phosphorus reductions and longer recovery times; especially in view of the current climate-change projections. ML represents a robust strategy to elucidate water quality patterns in lakes, where the available information is sufficient to train the constructed algorithms. Our mapping of HAB magnitudes and associated environmental/meteorological drivers can help managers to delineate the hot-spots at a national scale, and comprehensively design the best management practices for mitigating the eutrophication severity in China's lakes and reservoirs.

Machine learning predicts the growth of cyanobacterial genera in river systems and reveals their different environmental responses

Comparative Analysis of Machine Learning Methods for Prediction of Chlorophyll-a in a River with Different Hydrology Characteristics: A Case Study in Fuchun River, China

Algal community structure prediction by machine learning

Revealing Physiochemical Factors and Zooplankton Influencing Microcystis Bloom Toxicity in a Large-Shallow Lake Using Bayesian Machine Learning

Two machine learning approaches for predicting cyanobacteria abundance in aquaculture ponds

Exploring the response and prediction of phytoplankton to environmental factors in eutrophic marine areas using interpretable machine learning methods

Deep-Learning-Based Approach for Prediction of Algal Blooms

Improved Predictive Performance of Cyanobacterial Blooms Using a Hybrid Statistical and Deep-Learning Method

Predicting and analyzing the algal population dynamics of a grass-type lake with explainable machine learning

Employing hybrid deep learning for near-real-time forecasts of sensor-based algal parameters in a Microcystis bloom-dominated lake

Supervised machine learning for understanding and predicting the status of bistable eukaryotic plankton community in urbanized rivers

Predicting coastal algal blooms with environmental factors by machine learning methods

Prediction of chlorophyll a and risk assessment of water blooms in Poyang Lake based on a machine learning method

Causality Analysis and Prediction of Riverine Algal Blooms by Combining Empirical Dynamic Modeling and Machine Learning Techniques

Projected response of algal blooms in global lakes to future climatic and land use changes: machine learning approaches

Multi-class secondary metabolites in cyanobacterial blooms from a tropical water body: Distribution patterns and real-time prediction

The relative importance of environmental factors in predicting phytoplankton shifting and cyanobacteria abundance in regulated shallow lakes

Microcystis abundance is predictable through ambient bacterial communities: A data-oriented approach

Toward a Predictive Understanding of Cyanobacterial Harmful Algal Blooms through AI Integration of Physical, Chemical, and Biological Data

Machine learning approach identifies water sample source based on microbial abundance.

The magnitude and drivers of harmful algal blooms in China’s lakes and reservoirs: A national-scale characterization