Use of artificial neural network in the prediction of algal blooms

Prediction of Aureococcus anophageffens using machine learning and deep learning

The recurrent brown tide phenomenon, attributed to Aureococcus anophagefferens (A. anophagefferens), constitutes a significant threat to the Qinhuangdao sea area in China, leading to pronounced ecological degradation and substantial economic losses. This study utilized machine learning and deep learning techniques to predict A. anophagefferens population density, aiming to elucidate the occurrence mechanism and influencing factors of brown tide. Specifically, Random Forest (RF) algorithm was utilized to impute missing water quality data, facilitating its direct application in subsequent algal population prediction models. The results revealed that all four models-RF, Support Vector Regression (SVR), Multilayer Perceptron (MLP), and Convolutional Neural Network (CNN)-exhibited high accuracy in predicting A. anophagefferens population densities, with R2 values exceeding 0.75. RF, in particular, showed exceptional accuracy and reliability, with an R2 value surpassing 0.8. Additionally, the study ascertained five critical factors influencing A. anophagefferens population density: ammonia nitrogen, pH, total nitrogen, temperature, and silicate.

Prediction of hydrological and water quality data based on granular-ball rough set and k-nearest neighbor analysis

Hydrological and water quality datasets usually encompass a large number of characteristic variables, but not all of these significantly influence analytical outcomes. Therefore, by wisely selecting feature variables with rich information content and removing redundant features, it not only can the analysis efficiency be improved, but the model complexity can also be simplified. This paper considers introducing the granular-ball rough set algorithm for feature variable selection and combining it with the k-nearest neighbor method and back propagation network to analyze hydrological and water quality data, thus promoting overall and fused inspection. The results of hydrological water quality data analysis show that the proposed method produces better results compared to using a standalone k-nearest neighbor regressor.

Remote sensing for mapping algal blooms in freshwater lakes: a review

A large number of freshwater lakes around the world show recurring harmful algal blooms, particularly cyanobacterial blooms, that affect public health and ecosystem integrity. Prediction, early detection, and monitoring of algal blooms are inevitable for the mitigation and management of their negative impacts on the environment and human beings. Remote sensing provides an effective tool for detecting and spatiotemporal monitoring of these events. Various remote sensing platforms, such as ground-based, spaceborne, airborne, and UAV-based, have been used for mounting sensors for data acquisition and real-time monitoring of algal blooms in a cost-effective manner. This paper presents an updated review of various remote sensing platforms, data types, and algorithms for detecting and monitoring algal blooms in freshwater lakes. Recent studies on remote sensing using sophisticated sensors mounted on UAV platforms have revolutionized the detection and monitoring of water quality. Image processing algorithms based on Artificial Intelligence (AI) have been improved recently and predicting algal blooms based on such methods will have a key role in mitigating the negative impacts of eutrophication in the future.

Characterizing ocean surface contamination: Composition, film thickness, and rheology

Contamination of the air-sea interface contributes to wave dissipation, which in turn affects the exchange of momentum and energy between the fluids. In this work, the interfacial region between ocean and atmosphere is investigated to determine its composition, vertical extent, and rhelogical behavior. It is found that average thicknesses of interfacial layers on the ocean, such as those resulting from oil spills, could vary from a few hundreds of nanometers to a few millimeters. In addition to oil, interfacial layers also consist of emulsions, plankton, and chemical compounds, and are expected to be non-Newtonian due to their complex composition. To confirm this, fluid samples (mixtures of water-oil emulsions and skimmed off top layer of a month-old Phormidium sp. strain introduced in artificial seawater nutrient medium) that loosely model an interfacial layer are prepared in the laboratory, and their rheological characteristics are investigated using a rheometer. All the samples exhibit shear thinning behavior within the considered ranges for temperature and shear rate. It is recommended that models for wave dissipation account for both the finite thickness of an interfacial fluid layer and its non-Newtonian behavior, for consistency with the ocean-atmosphere system.

Remote sensing to detect harmful algal blooms in inland waterbodies

Harmful algal blooms (HABs) are an issue of concern for water management worldwide. As such, effective monitoring strategies of HAB spatio-temporal variability in waterbodies are needed. Remote sensing has become an increasingly important tool for HAB detection and monitoring in large lakes. However, accurate HAB detection in small-medium waterbodies via satellite data remains a challenge. Current barriers include the waterbody size, the limited freely available high resolution satellite data, and the lack of field calibration data. To test the applicability of remote sensing for detecting HABs in small-medium waterbodies, three satellites (Planetscope, Sentinel-2 and Landsat-8) were used to understand how spatial resolution, the availability of spectral bands, and the waterbody size itself effect HAB detection skill. Different algorithms and a non-parametric method, Self-Organizing Map (SOM), were tested. Curvature Around Red and NIR minus Red had the best HAB detection skill of the 20 existing algorithms that were tested. Landsat 8 and Sentinel 2 were the best satellites for HAB detection in small to medium waterbodies. The most critical attribute for detecting HABs were the available satellite bands, which determine the detection algorithms that can be used. Importantly, algorithm performance was mostly unrelated to waterbody size. However, there remain some barriers in utilizing satellite data for HAB detection, including algae dynamics, macrophyte cover within the waterbody, weather effects, and the correction models for satellite data. Moreover, it is important to consider the match time between satellite overpass and sampling activities for calibration. Given these challenges, integrating regular sampling activities and remote sensing is recommended for monitoring and managing small-medium waterbodies.

Prediction of Microcystis Occurrences and Analysis Using Machine Learning in High-Dimension, Low-Sample-Size and Imbalanced Water Quality Data

Machine learning, Deep learning, and water quality data have been used in recent years to predict the outbreak of harmful algae, especially Microcystis, and analyze outbreak causes. However, for various reasons, water quality data are often High-Dimension, Low-Sample- Size (HDLSS), meaning the sample size is lower than the number of dimensions. Moreover, imbalance problems may arise due to bias in the occurrence frequency of Microcystis. These problems make predicting the occurrence of Microcystis and analyzing its causes with machine learning difficult. In this study, a machine learning model that applies Feature Engineering (FE) and Feature Selection (FS) algorithms are used to predict outbreaks of Microcystis and analyze the outbreak factors from imbalanced HDLSS water quality data. The prediction performance was verified with binary classification to determine whether Microcystis would occur in the future by applying three machine learning models to four data patterns. The cause analysis of Microcystis occurrence was performed by visualizing the results of applying FE and FS. For the test data, the predictive performance of FE and FS methods was significantly better than that of the conventional method, with an accuracy of .108 points and an F-value of .691 points higher than the conventional method. A prediction performance increase was observed with a smaller model capacity. Data-driven analysis suggested that total nitrogen, chemical oxygen demand, chlorophyll-a, dissolved oxygen saturation, and water temperature are associated with Microcystis occurrences. The results also indicated that basic statistics of the water quality distribution (especially mean, standard deviation, and skewness) over a year, not the concentrations of water components, are related to the occurrence of Microcystis. These are new findings not found in previous studies and are expected to contribute significantly to future studies of algae. This study provides a method for analyzing water quality data with high-dimensionality and small samples, imbalance problems, or both.

Chlorophyll and phycocyanin in-situ fluorescence in mixed cyanobacterial species assemblages: Effects of morphology, cell size and growth phase

Cyanobacteria harmful blooms can represent a major risk for public health due to potential release of toxins and other noxious compounds in the water. A continuous and high-resolution monitoring of the cyanobacteria population is required due to their rapid dynamics, which has been increasingly done using in-situ fluorescence of phycocyanin (f-PC) and chlorophyll a (f-Chl a). Appropriate in-situ fluorometers calibration is essential because f-PC and f-Chl a are affected by biotic and abiotic factors, including species composition. Measurement of f-PC and f-Chl a in mixed species assemblages during different growth phases - representative of most field conditions - has received little attention. We hypothesized that f-PC and f-Chl a of mixed assemblages of cyanobacteria may be accurately estimated if taxa composition and fluorescence characteristics are known. We also hypothesized that species with different morphologies would have different fluorescence per unit cell and biomass. We tested these hypotheses in a controlled culture experiment in which photosynthetic pigment fluorescence, chemical pigment extraction, optical density and microscopic enumeration of four common cyanobacteria species (Aphanocapsa sp, Microcystis aeruginosa, Dolichospermum circinale and Raphidiopsis raciborskii) were quantified. Both monocultures and mixed cultures were monitored from exponential to late stationary growth phases. The sum of fluorescence of individual species calculated for mixed samples was not significantly different than measured fluorescence of mixed cultures. Estimated and measured f-PC and f-Chl a of mixed cultures had higher correlations and smaller absolute median errors when estimations were based on fluorescence per biomass instead of fluorescence per cell. Largest errors were overestimations of measured fluorescence for species with different morphologies. Fluorescence per cell was significantly different among most species, while fluorescence per unit biomass was not, indicating that conversion of fluorescence to biomass reduces species-specific bias. This study presents new information on the effect of species composition on cyanobacteria fluorescence. Best practices of deployment and operation of fluorometers, and data-driven models supporting in-situ fluorometers calibration are discussed as suitable solutions to minimize taxa-specific bias in fluorescence estimates.

Modelling of threats that affect Cyano-HABs in an eutrophicated reservoir: First phase towards water security and environmental governance in watersheds

Cyano-HABs are proliferating around the world due to anthropogenic nutrient enrichment of freshwater bodies. This study seeks to obtain a holistic vision over the various threats that affect the Cyano-HABs of Umia basin and especially of A Baxe reservoir (Galicia, NW Spain), through the method of Partial least squares path modelling (PLS-PM). The A Baxe reservoirs is a fundamental source of drinking water supply to surrounding dwellings. This study identifies and quantify the variables that increase contaminant concentration and decrease ecological integrity, as well as how this scenario evolved over various hydrologic years. In this regard, the PLS-PM equations will be robust and powerful tools to predict changes in eutrophication and ecological integrity, as response to measures implemented in the basin that can improve water quality. The dependent latent variables are "Eutrophication" (chlorophyl-a, Microcystis sp.) and "Ecological Integrity" (METI Bioindicator). The independent latent variables are "SWP", which represents surface water parameters (phosphorus, nitrogen and pH) and "Climatic Conditions" (temperature, precipitation). The PLS-PM results revealed that 51.0% of "Eutrophication" is predicted by the independent variables. The connections between latent variables are quantified through path coefficients (β). The "SWP" contributes by increasing "Eutrophication" (β = 0.235), the same occurring with the "Climatic Conditions" (β = -0.672). The variables "Eutrophication" (β = -0.217) and "SWP" (β = -0.483) lower the "Ecological Integrity". On the other hand, different trophic scenarios, adapted to the temperature increase predicted for the study area, were tested, and it was found that ecological integrity would improve by 46% if the oligotrophic state were reached. Therefore, it is recommended to prevent pollution by means of water control and governance plans, as well as corrective and preventive measures, which guarantee the water security of the river basins. Despite the complex mathematics behind the PLS-PM models, their user-friendly development and application through interactive graphical interfaces make them easily transposable to other eutrophic reservoirs, widening the readership of these studies focused on multiple-geosphere assessment of environmental impacts.

Machine learning based marine water quality prediction for coastal hydro-environment management

During the past three decades, harmful algal blooms (HAB) events have been frequently observed in marine waters around many coastal cities in the world including Hong Kong. The increasing occurrence of HAB has caused acute influences and damages on water environment and marine aquaculture with millions of monetary losses. For example, the Tolo Harbour is one of the most affected areas in Hong Kong, where more than 30% HAB occurred. In order to forewarn the potential HAB incidents, the machine learning (ML) methods have been increasingly resorted in modelling and forecasting water quality issues. In this study, two different ML methods - artificial neural networks (ANN) and support vector machine (SVM) - are implemented and improved by introducing different hybrid learning algorithms for the simulations and comparative analysis of more than 30-year measured data, so as to accurately forecast algal growth and eutrophication in Tolo Harbour in Hong Kong. The application results show the good applicability and accuracy of these two ML methods for the predictions of both trend and magnitude of the algal growth. Specifically, the results reveal that ANN is preferable to achieve satisfactory results with quick response, while the SVM is suitable to accurately identify the optimal model but taking longer training time. Moreover, it is demonstrated that the used ML methods could ensure robustness to learn complicated relationship between algal dynamics and different coastal environmental variables and thereby to identify significant variables accurately. The results analysis and discussion of this study also indicate the potentials and advantages of the applied ML models to provide useful information and implications for understanding the mechanism and process of HAB outbreak and evolution that is helpful to improving the water quality prediction for coastal hydro-environment management.

A systematic literature review of forecasting and predictive models for cyanobacteria blooms in freshwater lakes

Cyanobacteria harmful blooms (CyanoHABs) in lakes and reservoirs represent a major risk for water authorities globally due to their toxicity and economic impacts. Anticipating bloom occurrence and understanding the main drivers of CyanoHABs are needed to optimize water resources management. An extensive review of the application of CyanoHABs forecasting and predictive models was performed, and a summary of the current state of knowledge, limitations and research opportunities on this topic is provided through analysis of case studies. Two modelling approaches were used to achieve CyanoHABs anticipation; process-based (PB) and data-driven (DD) models. The objective of the model was a determining factor for the choice of modelling approach. PB models were more frequently used to predict future scenarios whereas DD models were employed for short-term forecasts. Each modelling approach presented multiple variations that may be applied for more specific, targeted purposes. Most models reviewed were site-specific. The monitoring methodologies, including data frequency, uncertainty and precision, were identified as a major limitation to improve model performance. A lack of standardization of both model output and performance metrics was observed. CyanoHAB modelling is an interdisciplinary topic and communication between disciplines should be improved to facilitate model comparisons. These shortcomings can hinder the adoption of modelling tools by practitioners. We suggest that water managers should focus on generalising models for lakes with similar characteristics and where possible use high frequency monitoring for model development and validation.

Fully Convolutional Neural Network for Detection and Counting of Diatoms on Coatings after Short-Term Field Exposure

While the use of deep learning is a valuable technology for automatic detection systems for medical data and images, the biofouling community is still lacking an analytical tool for the detection and counting of diatoms on samples after short-term field exposure. In this work, a fully convolutional neural network was implemented as a fast and simple approach to detect diatoms on two-channel (fluorescence and phase-contrast) microscopy images by predicting bounding boxes. The developed approach performs well with only a small number of trainable parameters and a F1 score of 0.82. Counting diatoms was evaluated on a data set of 600 microscopy images of three different surface chemistries (hydrophilic and hydrophobic) and is very similar to counting by humans while demanding only a fraction of the analysis time.

Spatiotemporal variability of phytoplankton functional groups in a shallow eutrophic lake from cold, arid regions

In the shallow eutrophic lakes in cold, arid regions, the phytoplankton functional groups and the factors that drive their spatiotemporal variabilities remain unclear. Samples were collected from Lake Ulansuhai in April, August, and October 2017 (wet season) and January 2018 (dry season). Based on the functional group classification method, 23 phytoplankton functional groups with 5 major ones were identified. During the wet season, high amounts of nutrients, elevated temperatures, and heavy rainfall produced spatiotemporal variabilities in phytoplankton communities, whereas during the dry season, the frozen period was the critical factor that determined the spatiotemporal variabilities in the phytoplankton communities. Through redundancy analyses, total nitrogen and total phosphorus concentrations were observed to directly affect the phytoplankton growth; algal growth affected the chemical oxygen demand, and pH and environmental factors interacted with the phytoplankton growth. These results highlight the complex feedbacks of shallow eutrophic lake ecosystems in arid regions. Group TC (represented by Lyngbya) was correlated with Huangtai algae. In August, a Huangtai algal bloom resulted in a relatively stable water column, which was conducive to group TC growth. Therefore, the presence of certain phytoplankton functional groups can indicate the current lake conditions by identifying the coverage of Huangtai algae, which provides a scientific basis for an early warning of a potential algal bloom.

Machine Learning Approaches for Predicting Health Risk of Cyanobacterial Blooms in Northern European Lakes

Cyanobacterial blooms are considered a major threat to global water security with documented impacts on lake ecosystems and public health. Given that cyanobacteria possess highly adaptive traits that favor them to prevail under different and often complicated stressor regimes, predicting their abundance is challenging. A dataset from 822 Northern European lakes is used to determine which variables better explain the variation of cyanobacteria biomass (CBB) by means of stepwise multiple linear regression. Chlorophyll-a (Chl-a) and total nitrogen (TN) provided the best modelling structure for the entire dataset, while for subsets of shallow and deep lakes, Chl-a, mean depth, TN and TN/TP explained part of the variance in CBB. Path analysis was performed and corroborated these findings. Finally, CBB was translated to a categorical variable according to risk levels for human health associated with the use of lakes for recreational activities. Several machine learning methods, namely Decision Tree, K-Nearest Neighbors, Support-vector Machine and Random Forest, were applied showing a remarkable ability to predict the risk, while Random Forest parameters were tuned and optimized, achieving a 95.81% accuracy, exceeding the performance of all other machine learning methods tested. A confusion matrix analysis is performed for all machine learning methods, identifying the potential of each method to correctly predict CBB risk levels and assessing the extent of false alarms; random forest clearly outperforms the other methods with very promising results.

Unsteady Multi-Element Time Series Analysis and Prediction Based on Spatial-Temporal Attention and Error Forecast Fusion

Harmful algal blooms (HABs) often cause great harm to fishery production and the safety of human lives. Therefore, the detection and prediction of HABs has become an important issue. Machine learning has been increasingly used to predict HABs at home and abroad. However, few of them can capture the sudden change of Chl-a in advance and handle the long-term dependencies appropriately. In order to address these challenges, the Long Short-Term Memory (LSTM) based spatial-temporal attentions model for Chlorophyll-a (Chl-a) concentration prediction is proposed, a model which can capture the correlation between various factors and Chl-a adaptively and catch dynamic temporal information from previous time intervals for making predictions. The model can also capture the stage of Chl-a when values soar as red tide breaks out in advance. Due to the instability of the current Chl-a concentration prediction model, the model is also applied to make a prediction about the forecast reliability, to have a basic understanding of the range and fluctuation of model errors and provide a reference to describe the range of marine disasters. The data used in the experiment is retrieved from Fujian Marine Forecasts Station from 2009 to 2011 and is combined into 8-dimension data. Results show that the proposed approach performs better than other Chl-a prediction algorithms (such as Attention LSTM and Seq2seq and back propagation). The result of error prediction also reveals that the error forecast method possesses established advantages for red tides prevention and control.

Using Machine-Learning Algorithms for Eutrophication Modeling: Case Study of Mar Menor Lagoon (Spain)

The Mar Menor is a hypersaline coastal lagoon with high environmental value and a characteristic example of a highly anthropized hydro-ecosystem located in the southeast of Spain. An unprecedented eutrophication crisis in 2016 and 2019 with abrupt changes in the quality of its waters caused a great social alarm. Understanding and modeling the level of a eutrophication indicator, such as chlorophyll-a (Chl-a), benefits the management of this complex system. In this study, we investigate the potential machine learning (ML) methods to predict the level of Chl-a. Particularly, Multilayer Neural Networks (MLNNs) and Support Vector Regressions (SVRs) are evaluated using as a target dataset information of up to nine different water quality parameters. The most relevant input combinations were extracted using wrapper feature selection methods which simplified the structure of the model, resulting in a more accurate and efficient procedure. Although the performance in the validation phase showed that SVR models obtained better results than MLNNs, experimental results indicated that both ML algorithms provide satisfactory results in the prediction of Chl-a concentration, reaching up to 0.7 R²_CV (cross-validated coefficient of determination) for the best-fit models.

Transfer learning for neural network model in chlorophyll-a dynamics prediction

Neural network models have been used to predict chlorophyll-a concentration dynamics. However, as model generalization ability decreases, (i) the performance of the models gradually decreases over time; (ii) the accuracy and performance of the models need to be improved. In this study, Transfer learning (TL) is employed to optimize neural network models (including feedforward neural networks (FNN), recurrent neural networks (RNN) and long short-term memory (LTSM)) and overcome these problems. Models using TL are able to reduce the influence of mutable data distribution and enhance generalization ability. Thus, it can improve the accuracy of prediction and maintain high performance in long-term applications. Also, TL is compared with parameter norm penalties (PNP) and dropout-two other methods used to improve model generalization ability. In general, TL has a better prediction effect than PNP and dropout. All the models, including FNN with different architectures, RNN and LSTM, as well as models optimized by PNP, dropout, and TL, are applied to an estuary reservoir in eastern China to predict chlorophyll-a dynamics at 5-min intervals. According to the results of this study, (i) models with TL produce the best prediction results; (ii) the original models and the models with PNP and dropout lose their ability to predict within 3 months, while TL models retain a high prediction accuracy.

Characterising and predicting cyanobacterial blooms in an 8-year amplicon sequencing time course

Cyanobacterial blooms occur in lakes worldwide, producing toxins that pose a serious public health threat. Eutrophication caused by human activities and warmer temperatures both contribute to blooms, but it is still difficult to predict precisely when and where blooms will occur. One reason that prediction is so difficult is that blooms can be caused by different species or genera of cyanobacteria, which may interact with other bacteria and respond to a variety of environmental cues. Here we used a deep 16S amplicon sequencing approach to profile the bacterial community in eutrophic Lake Champlain over time, to characterise the composition and repeatability of cyanobacterial blooms, and to determine the potential for blooms to be predicted based on time course sequence data. Our analysis, based on 135 samples between 2006 and 2013, spans multiple bloom events. We found that bloom events significantly alter the bacterial community without reducing overall diversity, suggesting that a distinct microbial community-including non-cyanobacteria-prospers during the bloom. We also observed that the community changes cyclically over the course of a year, with a repeatable pattern from year to year. This suggests that, in principle, bloom events are predictable. We used probabilistic assemblages of OTUs to characterise the bloom-associated community, and to classify samples into bloom or non-bloom categories, achieving up to 92% classification accuracy (86% after excluding cyanobacterial sequences). Finally, using symbolic regression, we were able to predict the start date of a bloom with 78-92% accuracy (depending on the data used for model training), and found that sequence data was a better predictor than environmental variables.

Comparison of models for predicting the changes in phytoplankton community composition in the receiving water system of an inter-basin water transfer project

Inter-basin water transfer projects might cause complex hydro-chemical and biological variation in the receiving aquatic ecosystems. Whether machine learning models can be used to predict changes in phytoplankton community composition caused by water transfer projects have rarely been studied. In the present study, we used machine learning models to predict the total algal cell densities and changes in phytoplankton community composition in Miyun reservoir caused by the middle route of the South-to-North Water Transfer Project (SNWTP). The model performances of four machine learning models, including regression trees (RT), random forest (RF), support vector machine (SVM), and artificial neural network (ANN) were evaluated and the best model was selected for further prediction. The results showed that the predictive accuracies (Pearson's correlation coefficient) of the models were RF (0.974), ANN (0.951), SVM (0.860), and RT (0.817) in the training step and RF (0.806), ANN (0.734), SVM (0.730), and RT (0.692) in the testing step. Therefore, the RF model was the best method for estimating total algal cell densities. Furthermore, the predicted accuracies of the RF model for dominant phytoplankton phyla (Cyanophyta, Chlorophyta, and Bacillariophyta) in Miyun reservoir ranged from 0.824 to 0.869 in the testing step. The predicted proportions with water transfer of the different phytoplankton phyla ranged from -8.88% to 9.93%, and the predicted dominant phyla with water transfer in each season remained unchanged compared to the phytoplankton succession without water transfer. The results of the present study provide a useful tool for predicting the changes in phytoplankton community caused by water transfer. The method is transferrable to other locations via establishment of models with relevant data to a particular area. Our findings help better understanding the possible changes in aquatic ecosystems influenced by inter-basin water transfer.

Monitoring studies should consider temporal variability to reveal relations between cyanobacterial abundance and environmental variables

One of the main goals of monitoring cyanobacteria blooms in aquatic environments is to reveal the relationship between cyanobacterial abundance and environmental variables. Studies typically correlate data that were simultaneously sampled. However, samplings occur sparsely over time and may not reveal the short-term responses of cyanobacterial abundance to environmental changes. In this study, we tested the hypothesis that stronger cyanobacteria x environment relationships in monitoring are found when the temporal variability of sampling points is incorporated in the statistical analyses. To this end, we investigated relationships between cyanobacteria and seven environmental variables that were sampled twice yearly for three years across 11 reservoirs, and data from an intensive monitoring in one of these reservoirs. Poor correlations were obtained when correlating data simultaneously sampled. In fact, the 'highly recurrent' role of phosphorus in cyanobacteria blooms is not properly observed in all sampling periods. On the other hand, the strongest correlation values for the total phosphorus x cyanobacteria relationship were observed when we used the variation of sampling points. We have also shown that environment variables better explain cyanobacteria when a time lag is considered. We conclude that, in cyanobacteria monitoring, the best approach to reveal determinants of cyanobacteria blooms is to consider environmental variability.

Estimating the risk of cyanobacterial occurrence using an index integrating meteorological factors: application to drinking water production

The sudden appearance of toxic cyanobacteria (CB) blooms is still largely unpredictable in waters worldwide. Many post-hoc explanations for CB bloom occurrence relating to physical and biochemical conditions in lakes have been developed. As potentially toxic CB can accumulate in drinking water treatment plants and disrupt water treatment, there is a need for water treatment operators to determine whether conditions are favourable for the proliferation and accumulation of CB in source waters in order to adjust drinking water treatment accordingly. Thus, a new methodology with locally adaptable variables is proposed in order to have a single index, f(p), related to various environmental factors such as temperature, wind speed and direction. The index is used in conjunction with real time monitoring data to determine the probability of CB occurrence in relation to meteorological factors, and was tested at a drinking water intake in Missisquoi Bay, a shallow transboundary bay in Lake Champlain, Québec, Canada. These environmental factors alone were able to explain a maximum probability of 68% that a CB bloom would occur at the drinking water treatment plant. Nutrient limitation also influences CB blooms and intense blooms only occurred when the dissolved inorganic nitrogen (DIN) to total phosphorus (TP) mass ratio was below 3. Additional monitoring of DIN and TP could be considered for these source waters prone to cyanobacterial blooms to determine periods of favourable growth. Real time monitoring and the use of the index could permit an adequate and timely response to CB blooms in drinking water sources.

Modelling of multi-nutrient interactions in growth of the dinoflagellate microalga Protoceratium reticulatum using artificial neural networks

This study examines the use of artificial neural networks as predictive tools for the growth of the dinoflagellate microalga Protoceratium reticulatum. Feed-forward back-propagation neural networks (FBN), using Levenberg-Marquardt back-propagation or Bayesian regularization as training functions, offered the best results in terms of representing the nonlinear interactions among all nutrients in a culture medium containing 26 different components. A FBN configuration of 26-14-1 layers was selected. The FBN model was trained using more than 500 culture experiments on a shake flask scale. Garson's algorithm provided a valuable means of evaluating the relative importance of nutrients in terms of microalgal growth. Microelements and vitamins had a significant importance (approximately 70%) in relation to macronutrients (nearly 25%), despite their concentrations in the culture medium being various orders of magnitude smaller. The approach presented here may be useful for modelling multi-nutrient interactions in photobioreactors.

Prevention of Phormidium tenue Biofilm Formation by TiO(2) Photocatalysis

We showed that the photocatalytic effect of a coating of TiO(2) greatly reduces the formation of a biofilm by Phormidium tenue (P. tenue), a filamentous cyanobacterium, on glass plates. Sample plates were immersed in P. tenue culture solution (OD(730)=0.3) under concurrent illumination with white fluorescent (WF) and UV light (0.3 mW cm(-2), each) for 11 days. TiO(2)-coated glass plates showed greatly reduced adhesion of P. tenue over 11 days compared to bare plates. The number of P. tenue adhering to bare glass plates increased to over 10(6) cells cm(-2) in 6 days. The photocatalytic anti-biofilm effect was also observed under WF light, although it was small and lasted only a few days. The addition of 1 mM mannitol, a scavenger for the hydroxyl radical (·OH), suppressed the effect. The surface of TiO(2)-coated plates was maintained in a highly hydrophilic state for 11 days, regardless of the addition of mannitol. Therefore, we conclude that the photocatalytic oxidation of P. tenue is effective in preventing the formation of a biofilm.

Identification of spatiotemporal nutrient patterns in a coastal bay via an integrated k-means clustering and gravity model

This study presents an integrated k-means clustering and gravity model (IKCGM) for investigating the spatiotemporal patterns of nutrient and associated dissolved oxygen levels in Tampa Bay, Florida. By using a k-means clustering analysis to first partition the nutrient data into a user-specified number of subsets, it is possible to discover the spatiotemporal patterns of nutrient distribution in the bay and capture the inherent linkages of hydrodynamic and biogeochemical features. Such patterns may then be combined with a gravity model to link the nutrient source contribution from each coastal watershed to the generated clusters in the bay to aid in the source proportion analysis for environmental management. The clustering analysis was carried out based on 1 year (2008) water quality data composed of 55 sample stations throughout Tampa Bay collected by the Environmental Protection Commission of Hillsborough County. In addition, hydrological and river water quality data of the same year were acquired from the United States Geological Survey's National Water Information System to support the gravity modeling analysis. The results show that the k-means model with 8 clusters is the optimal choice, in which cluster 2 at Lower Tampa Bay had the minimum values of total nitrogen (TN) concentrations, chlorophyll a (Chl-a) concentrations, and ocean color values in every season as well as the minimum concentration of total phosphorus (TP) in three consecutive seasons in 2008. The datasets indicate that Lower Tampa Bay is an area with limited nutrient input throughout the year. Cluster 5, located in Middle Tampa Bay, displayed elevated TN concentrations, ocean color values, and Chl-a concentrations, suggesting that high values of colored dissolved organic matter are linked with some nutrient sources. The data presented by the gravity modeling analysis indicate that the Alafia River Basin is the major contributor of nutrients in terms of both TP and TN values in all seasons. With this new integration, improvements for environmental monitoring and assessment were achieved to advance our understanding of sea-land interactions and nutrient cycling in a critical coastal bay, the Gulf of Mexico.

EVALUATION OF ENVIRONMENTAL FACTORS ON CYANOBACTERIAL BLOOM IN EUTROPHIC RESERVOIR USING ARTIFICIAL NEURAL NETWORKS1

Cyanobacterial blooms are a common issue in eutrophic freshwaters, and some cyanobacteria produce toxins, threatening the health of humans and livestock. Microcystin, a representative cyanobacterial hepatotoxin, is frequently detected in most Korean lakes and reservoirs. This study developed predictive models for cyanobacterial bloom using artificial neural networks (ANNs; self-organizing map [SOM] and multilayer perceptron [MLP]), including an evaluation of related environmental factors. Fourteen environmental factors, as independent variables for predicting the cyanobacteria density, were measured weekly in the Daechung Reservoir from spring to autumn over 5 years (2001, 2003-2006). Cyanobacterial density was highly associated with environmental factors measured 3 weeks earlier. The SOM model was efficient in visualizing the relationships between cyanobacteria and environmental factors, and also for tracing temporal change patterns in the environmental condition of the reservoir. And the MLP model exhibited a good predictive power for the cyanobacterial density, based on the environmental factors of 3 weeks earlier. The water temperature and total dissolved nitrogen were the major determinants for cyanobacteria. The water temperature had a stronger influence on cyanobacterial growth than the nutrient concentrations in eutrophic waters. Contrary to general expectations, the nitrogen compounds played a more important role in bloom formation than the phosphorus compounds.

Variation of phytoplankton community before an induced cyanobacterial (Arthrospira platensis) bloom

A cyanobacterial (Arthrospira platensis) bloom was induced in situ by nutrient manipulation in an enclosure. The succession of the phytoplankton community and the water chemistry variations before the appearance of bloom, as well as their relationship, were investigated. The cell pigment variations were studied simultaneously. The Pearson's correlation analysis showed that there was no significant correlation between water chemistry and green algal or cyanobacterial composition, indicating that water chemistry variations were not suitable to be used as indicators for cyanobacterial-bloom early-warning. However, the diversity index of the phytoplankton community decreased sharply before the bloom appeared. Therefore, the dynamics of phytoplankton community was put forward to be an indicator for cyanobacterial-bloom early-warning. In addition, the cell pigment variations represented the changes of community structure, which should be useful for studying the dynamics of phytoplankton community.

Water quality prediction of marine recreational beaches receiving watershed baseflow and stormwater runoff in southern California, USA

Beach advisories are issued to the public in California when the concentration of fecal indicator bacteria (FIB), including total coliform, fecal coliform (or Escherichia coli), and Enterococcus, exceed their recreational water health standards, or when the amount of a rainfall event is above the pre-determined threshold. However, it is not fully understood about how and to what degree stormwater runoff or baseflow exerts impacts on beach water quality. Furthermore, current laboratory methods used to determine the FIB levels take 18-96 h, which is too slow to keep pace with changes in FIB levels in water. Thus, a beach may not be posted when it is contaminated, and may be posted under advisory when bacterial levels have already decreased to within water quality standards. The study was designed to address the above critical issues. There were large temporal and spatial variations in FIB concentrations along two popular State Beaches in San Diego, CA, USA. The rainstorm-induced runoff from the watersheds exerts significant impacts on the marine recreational water quality of the beaches adjacent to lagoons during the first 24-48 h after a rain event. The large volume of stormwater runoff discharging to beaches caused high FIB concentrations in beach water not only at the lagoon outlet channel and the mixing zone, but also at the locations 90 m away from the channel northward or southward along the shoreline. The geomorphology of beach shoreline, distance from the outlet channel, wind strength, wind direction, tide height, wave height, rainfall, time lapse after a rainstorm, or channel flow rate played a role in affecting the distribution of FIB concentrations in beach water. Despite the great temporal and spatial variability of FIB concentrations along a shoreline, the artificial neural network-based models developed in this study are capable of successfully predicting FIB concentrations at different beaches, different locations, and different times under baseflow or rainstorm conditions. The models are based on readily measurable variables including temperature, conductivity, pH, turbidity, channel water flow, rainfall, and/or time lapse after a rainstorm. The established models will help fill the current gap between beach posting and actual water quality and make more meaningful and effective decisions on beach closures and advisories.

Time series forecasting of cyanobacteria blooms in the Crestuma Reservoir (Douro River, Portugal) using artificial neural networks

In this work, time series neural networks were used to predict the occurrence of toxic cyanobacterial blooms in Crestuma Reservoir, which is an important potable water supply for the Porto region, located in the north of Portugal. These models can potentially be used to provide water treatment plant operators with an early warning for developing cyanobacteria blooms. Physical, chemical, and biological parameters were collected at Crestuma Reservoir from 1999 to 2002. The data set was then divided into three independent time series, each with a fortnightly periodicity. One training series was used to "teach" the neural networks to predict results. Another series was used to verify the results, and to avoid over-fitting of the data. An additional independently collected data series was then used to test the efficacy of the model for predicting the abundance of cyanobacteria. All of the models tested in this study incorporated a prediction time (look-ahead parameter) equal to the sampling interval (two weeks). Various lag periods, from 2 to 52 weeks, were also investigated. The best model produced in this study provided the following correlations between the target and forecast values in the training, verification, and validation series: 1.000 (P = 0.000), 0.802 (P = 0.000), and 0.773 (P = 0.001), respectively. By applying this model to the three-year data set, we were able to predict fluctuations in cyanobacteria abundance in the Crestuma Reservoir, with a high level of precision. By incorporating a lag-period of eight weeks, we were able to detect secondary fluctuations in cyanobacterial abundance over the annual cycle.