Water quality modelling using artificial neural network and multivariate statistical techniques

Artificial neural network modeling of the water quality index for Kinta River (Malaysia) using water quality variables as predictors

Rivers face increasing pollution, requiring accurate water quality assessment tools. Existing indices like the Water Quality Index (WQI) often overlook the integration of oxygen-related parameters critical to aquatic health. Here, we develop a machine learning model using Support Vector Regression (SVR) to predict the Water Quality Index (WQIOIs) by integrating key oxygen-related parameters, including Biological Oxygen Demand (BOD), Chemical Oxygen Demand (COD), Dissolved Oxygen (DO), and the reaeration coefficients (K1, K2). Applied to three rivers in Iran, the model demonstrated high accuracy, with a cross-validated R² > 0.95 and root mean squared error (RMSE) of 0.92 for the Haraz River and 1.41 for the Simineh River. Predictions showed strong correlation (r = 0.98) with standard indices, and feature importance analysis revealed DO as the most influential parameter. The model’s generalizability was confirmed through validation on independent river datasets, highlighting its robustness across diverse hydrological conditions. This approach offers a scalable, interpretable framework for continuous water quality monitoring, enabling more precise and data-driven management of aquatic ecosystems, particularly in regions with varying environmental factors.

The study was conducted to spatiotemporally analyze the quality, location and critical water variables influencing water quality using water monitoring data from the Department of Environment and Natural Resources, Dong Thap province in 2019. The water quality parameters including turbidity, pH, temperature, dissolved oxygen (DO), total suspended solids (TSS), biological oxygen demand (BOD), chemical oxygen demand (COD), nitrite (N-NO2−), nitrate (N-NO3−), ammonium (N-NH4+), total nitrogen (TN), orthophosphate (P-PO43−), chloride (Cl−), oil and grease, sulfate (SO42−), coliforms, and Escherichia coli (E. coli) were collected at 58 locations with the frequency of four times per year (February, May, August, and November). These parameters were compared with national technical regulation on surface water quality—QCVN 08-MT: 2015/BTNMT. Water quality index (WQI) was calculated and spatially presented by geographical information system (GIS) tool. Pearson correlation analysis, cluster analysis (CA), and principal component analysis (PCA) were used to evaluate the correlation among water quality parameters, group and reduce the sampling sites, and identify key parameters and potential water pollution sources. The results showed that TSS, BOD, COD, N-NH4+, P-PO43−, coliforms, and E. coli were the significant concerns impairing the water quality. Water quality was assessed from poor to medium levels by WQI analysis. CA suggested that the current monitoring locations could be reduced from 58 sites to 43 sites which can be saved the total monitoring budget up to 25.85%. PCA showed that temperature, pH, TSS, DO, BOD, COD, N-NH4+, N-NO2−, TN, P-PO43−, coliforms, and E. coli were the key water parameters influencing water quality in Dong Thap province’s canals and rivers; thus, these parameters should be monitored annually. The water pollution sources were possibly hydrological conditions, water runoff, riverbank erosion, domestic and urban activities, and industrial and agricultural discharges. Significantly, the municipal and agricultural wastes could be decisive factors to the change of surface water quality in the study area. Further studies need to focus on identifying sources of water pollution for implementing appropriate water management strategies.

The river Ganga, revered as a lifeline of North India, is facing severe water quality degradation due to various man-made activities. The city of Kanpur, Uttar Pradesh, India, located on the west bank of the river Ganga, is known for its leather industrial area. The city produces approximately 450 million litres of municipal sewage and industrial effluent, most of this flowing directly into the holy river Ganga. Due to this, the Ganga River basin pollution is increasing daily, and to overcome this problem scientifically, there is a need to develop different models for water quality, which has been growing in recent years. The key focus of the study is to determine the Water Quality Index (WQI) along with developing a model to assess the changes in water quality over a decade in the Ganga River basin in Kanpur City. The dataset includes monthly measurements (2015-2023) of pH, dissolved oxygen (DO), biochemical oxygen demand (BOD), chemical oxygen demand (COD), and total coliform at seven distinct locations in Kanpur city, encompassing both upstream and downstream areas. Artificial Neural Networks (ANNs) were employed to construct a model to pretend, along with forecasting, the levels of dissolved oxygen (DO). The model has a notable level of accuracy in predicting dissolved oxygen (DO). The scatter plots comparing the actual and anticipated levels of dissolved oxygen (DO) during the training and testing periods exhibited a strong coefficient of determination (R2) with values of 0.90 and 0.92, respectively. Results show that the ANN model can predict water quality effects in advance, and thus, it helps to take protective actions to preserve the river free from pollution. Keywords: Water Quality, River Ganga, Artificial Neural Network, Dissolved Oxygen.

The Intergovernmental Panel on Climate Change (IPCC) Fifth Assessment Report (AR5) predicted that recent extreme hydrological events would affect water quality and aggravate various forms of water pollution. To analyze changes in water quality due to future climate change, input data (precipitation, average temperature, relative humidity, average wind speed and sunlight) were established using the Representative Concentration Pathways (RCP) 8.5 climate change scenario suggested by the AR5 and calculated the future runoff for each target period (Reference:1989-2015; I: 2016-2040; II: 2041-2070; and III: 2071-2099) using the semi-distributed land use-based runoff processes (SLURP) model. Meteorological factors that affect water quality (precipitation, temperature and runoff) were inputted into the multiple linear regression analysis (MLRA) and artificial neural network (ANN) models to analyze water quality data, dissolved oxygen (DO), biological oxygen demand (BOD), chemical oxygen demand (COD), suspended solids (SS), total nitrogen (T-N) and total phosphorus (T-P). Future water quality prediction of the Anseongcheon River basin shows that DO at Gongdo station in the river will drop by 35% in autumn by the end of the 21st century and that BOD, COD and SS will increase by 36%, 20% and 42%, respectively. Analysis revealed that the oxygen demand at Dongyeongyo station will decrease by 17% in summer and BOD, COD and SS will increase by 30%, 12% and 17%, respectively. This study suggests that there is a need to continuously monitor the water quality of the Anseongcheon River basin for long-term management. A more reliable prediction of future water quality will be achieved if various social scenarios and climate data are taken into consideration.

A study was initiated to predict water quality index (WQI) using artificial neural networks (ANNs) with respect to the concentrations of 16 groundwater quality variables collected from 47 wells and springs in Andimeshk during 2006–2013 by the Iran’s Ministry of Energy. Such a prediction has the potential to reduce the computation time and effort and the possibility of error in the calculations. For this purpose, three ANN’s algorithms including ANNs with early stopping, Ensemble of ANNs and ANNs with Bayesian regularization were utilized. The application of these algorithms for this purpose is the first study in its type in Iran. Comparison among the performance of different methods for WQI prediction shows that the minimum generalization ability has been obtained for the Bayesian regularization method (MSE = 7.71) and Ensemble averaging method (MSE = 9.25), respectively and these methods showed the minimum over-fitting problem compared with that of early stopping method. The correlation coefficients between the predicted and observed values of WQI were 0.94 and 0.77 for the test and training data sets, respectively indicating the successful prediction of WQI by ANNs through Bayesian regularization algorithm. A sensitivity analysis was implemented to show the importance of each parameter in the prediction of WQI during ANN’s modeling and showed that parameters like Phosphate and Fe are the most influential parameters in the prediction of WQI.

Rising solid waste production has caused high levels of environmental pollution. Population growth, economic patterns, and lifestyle patterns are major factors that have led to the alarming rate of solid waste production. Generally, solid wastes such as paper, wood, and plastic are disposed into landfills due to its low operation and maintenance costs. However, leachate discharged from landfills could be a problem in surfaces and groundwater if not adequately treated. This study investigated the patterns of the water quality index (WQI) and polycyclic aromatic hydrocarbons (PAH) along Johan River in Perak, Malaysia, which received treated leachate from a nearby landfill. An artificial neural network (ANN) was also applied to predict WQI and PAH concentration of the river. Seven sampling stations were chosen along the river. The stations represented the upstream of leachate discharge, point of leachate discharge, and five locations downstream of the landfill. Sampling was conducted for one year starting July 2018. Physicochemical parameters, namely pH, biological oxygen demand, chemical oxygen demand, ammoniacal nitrogen, total suspended solids, and dissolved oxygen, were used to compute the water quality index (WQI). PAH concentrations were determined by liquid–liquid extraction of water samples followed by an analysis using gas chromatography. Results showed that WQI of Johan River was under Class III where intensive treatment was required to make it suitable for drinking purposes. The highest recorded PAH concentrations were fluoranthene (333.4 ppb) in the dry season and benzo(a) pyrene (93.5 ppb) in the wet season. A correlation coefficient (Rp) for a model prediction based on WQI-ANN and TEC-ANN (toxicity equivalent concentration) in the wet and dry seasons was 0.9915, 0.9431, 0.9999, and 0.9999, respectively. ANN results showed good model performance with Rp ≈ 0.9. This study suggested that ANN is a useful tool for water quality studies.

Karacaören I Dam Lake, a potential source of drinking water, is under significant threat from human activities. This study aimed to determine the seasonal variations of certain physicochemical and nutrient parameters and to use the water quality index (WQI) to assess the seasonal water quality of the lake. Some physicochemical parameters, including dissolved oxygen (DO), temperature, salinity, total dissolved solids (TDS), chemical oxygen demand (COD), biological oxygen demand (BOD), electrical conductivity (EC), transparency, and nutrients such as phosphate, sulphate, nitrite, and nitrate, were monitored seasonally across seven sampling points. The mean values observed for pH, DO, temperature, TDS, EC, SO42-, NO2-, NO3-, PO43⁻, BOD and COD were 7.81±0.39, 8.9±1.62 mg/L, 21.99±6.28°C, 177.70±33.39 mg/L, 352.92±70.63 μS/cm, 17.36±2.70 mg/L, 0.06±0.10 mg/L, 4.60±3.41 mg/L, 0.30±0.94 mg/L, 3.19±0.54 mg/L and 15.50±2.51 mg/L, respectively. All examined parameters except for PO43⁻, BOD, and COD showed significant differences from season to season (P≤0.05). According to the water quality classification outlined in the Turkish Water Pollution Control Regulation, the nitrite levels were found to be at a critical level, rendering the water unsuitable for human consumption. The Water Quality Index (WQI) results indicated that the water in the dam lake is in good conduction, but it is not classified as excellent for human consumption. Therefore, necessary protective measures should be implemented regarding the planned use of the dam lake's water.

Rivers and lakes are primary sources of water for towns and villages. However, the quality of these water bodies is currently deteriorating due to urbanization, industrialization, farming, and other anthropogenic activities. This study was conducted within January 25th to November 26th 2017at designated points at Pantai River, Anak Air Batu River, and Varsity Lake, University of Malaya, Kuala Lumpur, Malaysia, where the physic-chemical and biological characteristics were determined in order to increase the statistical basis of these regions. In the physic-chemical evaluation of the water samples, diverse parameters were measured, such as the pH, turbidity, Dissolved Oxygen (DO), Biochemical Oxygen Demand (BOD), Chemical Oxygen Demand (COD), Suspended Solids (SS), ammonia, phosphate and nitrate attention, and biology. These parameters were compared with the Water Quality Index (WQI) standards of water quality. Overall, the result showed that the DO, BOD, COD, and SS fall within classes II, III, while Ammoniacal Nitrogen (NH3-N) falls within classes II, IV, and V, and pH and temperature fall within class I. The study area reported moderate levels of pollution between classes I and IIA due to urban activities. The results showed that the water collected from the study area was unsuitable for human consumption due to the significant presence of microbes.

Enhancing riverine load prediction of anthropogenic pollutants: Harnessing the potential of feed-forward backpropagation (FFBP) artificial neural network (ANN) models

This study proposes an innovative, intelligent, and scalable approach to water pollution monitoring by utilizing an Artificial Neural Network (ANN) to predict the Water Quality Index (WQI), thereby overcoming the limitations of traditional, labour-intensive monitoring methods. The analysis was conducted on River A in Johore, Malaysia, a representative freshwater system experiencing significant urban and industrial pressure. The dataset, comprising 90 samples collected across a nine-weeks sampling period, included measurements for eight key physicochemical parameters: pH, temperature, dissolved oxygen (DO), biochemical oxygen demand (BOD), chemical oxygen demand (COD), ammonia, total suspended solids (TSS), and turbidity. The developed ANN model demonstrated superior predictive capability, achieving a high coefficient of determination (R2) of 0.9902 on the training dataset and 0.9956 on the testing dataset, confirming its effectiveness in capturing the complex, nonlinear pollution dynamics. Application of the model revealed the river's overall status to be Class IV (Polluted) based on the Malaysian Water Quality Standards. Furthermore, the study identified that acute pollution loading, evidenced by excessive COD and Class V Ammonia Nitrogen (NH3-N) levels (up to 9.31 mg/L), is primarily driven by heavy rainfall events mobilizing surface runoff and contamination from nearby industrial activities. This validated ANN framework offers a cost-effective and reliable tool for real-time pollution forecasting, empowering decision-makers in the sustainable management and ecological restoration of polluted river basins.

Dissolved oxygen (DO) as one of the most fundamental parameters of water quality plays a vital role in aquatic life. This study was conducted to predict DO, biological oxygen demand (BOD), and chemical oxygen demand (COD) in an intensive rainbow trout rearing system with different biomass (B). The multilayer perceptron (MLP) and the radial basis function (RBF) neural networks were employed for evaluating the impacts of food parameters (crude protein (CP), consumed feed (CF)), fish parameters (different values of B, and weight gain (WG)), and water quality parameters including temperature (T) and flow rate (Q) on variation of DO, BOD, and COD concentrations. This study's results showed that although both MLP and RBF neural networks are capable to estimate DO, BOD, and COD concentrations, RBF neural network showed better performance compared to MLP neural network. The results of sensitivity analysis indicated that the parameter CF has the highest effect on DO concentration estimation. Independent variables CF, CP, WG, and B showed the highest to the lowest rank of impacts on BOD estimation, respectively. The results also illustrated a decreasing trend of the effects on the estimation error of COD changes simulation by all independent variables, including B, T, WG, CF, CP, and Q, respectively. RBF neural network based on better stability and generalization ability with average root mean square error (RMSE) and mean absolute percentage error (MAPE) values of less than 0.12 and 3% was superior to MLP in DO, BOD, and COD concentration prediction. Moreover, CF was identified as the most effective factor in estima12tion process. Based on the present study results, there are direct relationships between DO, BOD, and COD concentrations and water quality parameters, fish parameters, and food parameters. Food parameters relative to fish and water quality parameters imposed the greatest effects. Improvement in feeding process such as application of intelligence feeding methods and change in fish diet and feeding time can considerably reduce losses in production system. Graphical abstract.

The aim of this study is to analyze the water quality of the Way Kuripan River based on the Water Quality Index (WQI) calculation method that is developed by the Malaysian Department of Environment (DOE). Water samples were taken from five sample points (SK01, SK02, SK03, SK04 and SK05) in January 2017. WQI was calculated on the basis of six parameters: dissolved oxygen (DO), biochemical oxygen demand (BOD), chemical oxygen demand (COD), pH, total suspended solid (TSS) and ammoniac-nitrogen (NH3-N). The calculation procedure consists of three stages. Firstly, identifying the equation of the sub-index (SI) based on the parameter value. Secondly, calculate the sub-index (SI) of each parameter. Last is the calculation of the water quality index. The results show that SK01 and SK04 have WQI values of 70.3 and 70.11. Those values show that water quality of the Way Kuripan river is class III so the water is slightly polluted. Sample points, SK02 (WQI = 55.8) and SK03 (WQI=53.8) are highly polluted. The lowest WQI of the Way Kuripan river is SK05 = 38.3, so it is classified as,Class V (highly polluted). In conclusion, this data confirms that the water quality in the Sungai Kuripan River has been polluted.Keywords: Way Kuripan’ s River, Water Quality Index (WQI), water quality parameter

With the development of industry and agriculture, nitrogen, phosphorus and other nutrients in the Hanshui River greatly increase and eutrophication has become an important threat to the water quality of the Hanshui River, especially in the middle and lower reaches. The primary objective of this study was to establish the water quality model for the middle and lower reaches of the Hanshui River based on the model of MIKE 11. The main pollutants migration and transformation process could be simulated using the water quality model. The rainfall-runoff model, hydrodynamic model and water quality model were established using MIKE 11. The pollutants, such as chemical oxygen demand (COD), biochemical oxygen demand (BOD), ammonia nitrogen, nitrate nitrogen, phosphorus, dissolved oxygen (DO), were simulated and predicted using the above three models. A set of methods computing non-point source pollution load of the Hanshui River Basin was proposed in this study. The simulated and observed values of COD, BOD5, ammonia, nitrate, DO, and total phosphorus were compared after the parameter calibration of the water quality model. The simulated and observed results match better, thus the model can be used to predict water quality in the future for the Hanshui River. The pollution trend could be predicted using the water quality model according pollution load generation. It is helpful for government to take effective measures to prevent the water bloom and protect water quality in the river.

This work proposed a novel procedure of Water Quality Index (WQI) development that could be used for practical applications on a local or regional scale, based on available monitoring data. Principal component analysis (PCA) was applied to the monthly data of 11 water quality parameters (pH, conductivity (EC), total suspended solid (TSS), dissolved oxygen (DO), five -day biological oxygen demand (BOD), chemical oxygen demand (COD), ammonia (N-NH4), nitrate (N-NO3), phosphate (P-PO4), total coliform, and total dissolved iron monitored at 11 sites at Huong river in the years 2014–2016. From the PCA, the three extracted principal components explained 67% of the total variance of original variables. From the set of communality values, the weight (wi) for each parameter was determined. Linear sub-index functions were established based on the permissible limits from the National Technical Regulations on Surface Water Quality set up by the Vietnam Environment Agency (VEA) to derive the sub-index (qi) for each parameter. The multiplicative formula that is the product of the sub-indices (qi) raised to the respective weights (wi), was used for calculation of the final WQI values. The proposed index (WQI) was then applied to the river with quarterly data of the 11 parameters monitored at ten sites in the years 2017–2020. The WQI representatively reflected the actual status of the river overall water quality, of which 97.8% of the WQI values belonged to grades of EXCELLENT and GOOD, and 2.2% of grade MODERATE. Comparison between the river water quality evaluations resulting from the developed WQI with the WQI adopted by National Sanitation Foundation (NSF-WQI) and the index issued by Vietnam Environment Agency (VN-WQI) indicated that the proposed WQI was more suitable for river quality assessment.