Water Quality Prediction of Water Sources Based on Meteorological Factors using the CA-NARX Approach

With the increasingly serious problem of surface water environmental safety, it is of great significance to study the changing trend of reservoir water quality, and it is necessary to establish a water quality prediction and early warning system for the management and maintenance of water resources. Aiming at the problem of water quality prediction in reservoirs, a CA-NARX algorithm is designed, which combines the improved dynamic clustering algorithm with the idea of machine learning and the forward dynamic regression neural network. The improved dynamic clustering algorithm is used to classify the eutrophication degree of waterbodies according to the total phosphorus and total nitrogen content. Considering four meteorological factors, air temperature, water temperature, water surface evaporation, and rainfall, synthetically for each water quality condition, the total phosphorus and total nitrogen in the waterbody are forecasted by an improved forward NARX dynamic regression neural network. Based on this, the CA-NARX prediction algorithm can realize short period water quality prediction. Compared with the traditional support vector regression machine model, improved GA-BP neural network, and exponential smoothing method, the CA-NARX model has the least prediction error.

     

Modelling geosmin concentrations in three sources of raw water in Quebec, Canada

The presence of off-flavour compounds such as geosmin, often found in raw water, significantly reduces the organoleptic quality of distributed water and diverts the consumer from its use. To adapt water treatment processes to eliminate these compounds, it is necessary to be able to identify them quickly. Routine analysis could be considered a solution, but it is expensive and delays associated with obtaining the results of analysis are often important, thereby constituting a serious disadvantage. The development of decision-making tools such as predictive models seems to be an economic and feasible solution to counterbalance the limitations of analytical methods. Among these tools, multi-linear regression and principal component regression are easy to implement. However, due to certain disadvantages inherent in these methods (multicollinearity or non-linearity of the processes), the use of emergent models involving artificial neurons networks such as multi-layer perceptron could prove to be an interesting alternative. In a previous paper (Parinet et al., Water Res 44: 5847-5856, 2010), the possible parameters that affect the variability of taste and odour compounds were investigated using principal component analysis. In the present study, we expand the research by comparing the performance of three tools using different modelling scenarios (multi-linear regression, principal component regression and multi-layer perceptron) to model geosmin in drinking water sources using 38 microbiological and physicochemical parameters. Three very different sources of water, in terms of quality, were selected for the study. These sources supply drinking water to the Québec City area (Canada) and its vicinity, and were monitored three times per month over a 1-year period. Seven different modelling methods were tested for predicting geosmin in these sources. The comparison of the seven different models showed that simple models based on multi-linear regression provide sufficient predictive capacity with performance levels comparable to those obtained with artificial neural networks. The multi-linear regression model (R ²?=?0.657, <0.001) used only four variables (phaeophytin, sum of green algae, chlorophyll-a and potential Redox) in comparison with ten variables (potassium, heterotrophic bacteria, organic nitrogen, total nitrogen, phaeophytin, total organic carbon, sum of green algae, potential Redox, UV absorbance at 254 nm and atypical bacteria) for the best model obtained with artificial neural networks (R ²?=?0.843).     

Using multivariate techniques as a strategy to guide optimization projects for the surface water quality network monitoring in the Velhas river basin,Brazil

Surface water quality monitoring networks are usually deployed and rarely re-evaluated with regard to their effectiveness. In this sense, this work sought to evaluate and to guide optimization projects for the water quality monitoring network of the Velhas river basin, using multivariate statistical methods. The cluster, principal components, and factorial analyses, associated with non-parametric tests and the analysis of violation to the standards set recommended by legislation, identified the most relevant water quality parameters and monitoring sites, and evaluated the sampling frequency. Thermotolerant coliforms, total arsenic, and total phosphorus were considered the most relevant parameters for characterization of water quality in the river basin. The monitoring sites BV156, BV141, BV142, BV150, BV137, and BV153 were considered priorities for maintenance of the network. The multivariate statistical analysis showed the importance of a monthly sampling frequency, specifically the parameters considered most important.     

Assessment of chlorophyll-a variations in high- and low-flow seasons in Apalachicola Bay by MODIS 250-m remote sensing

Chlorophyll-a (chl-a) is considered as a primary indicator for water quality and foods for oyster growth in Apalachicola estuarine ecosystem. Assessment of chl-a concentration variation in response to river inflow is important for estuarine environmental research and management. In this study, remote sensing analysis has been conducted to evaluate the effects of river inflow on chlorophyll concentrations in Apalachicola Bay of Florida in the northeast Gulf of Mexico. A remote sensing model for chl-a was improved and applied to map spatial distributions of chl-a by using Moderate Resolution Imaging Spectroradiometer (MODIS) 250-m resolution imageries in high-flow and low-flow seasons in 2001 and 2008. Chl-a values approximately ranged from the minimum 6 μg/l to the maximum 29 μg/l in the study period. Maximum chl-a concentration in high-flow season was almost twice above that in low-flow season. The averaged mean and minimum chl-a level in the high-flow season were approximately 42 and 28 % higher than those in low-flow season, respectively. The remote sensing mapping of chl-a was able to show spatial variations of chl-a in the entire bay under different flow conditions, which indicated its advantage over the traditional field data sampling for monitoring water quality over a large area of estuary. The MODIS 250-m remote sensing regression model presented from this study can be used to support monitoring and assessment of the spatial chl-a distribution in the bay for environmental research and management in Apalachicola Bay.     

Assessment of surface water quality using multivariate statistical techniques in red soil hilly region: a case study of Xiangjiang watershed, China

In the study, multivariate statistical methods including factor, principal component and cluster analysis were applied to analyze surface water quality data sets obtained from Xiangjiang watershed, and generated during 7 years (1994-2000) monitoring of 12 parameters at 34 different profiles. Hierarchical cluster analysis grouped 34 sampling sites into three clusters, including relatively less polluted (LP), medium polluted (MP) and highly polluted (HP) sites, and based on the similarity of water quality characteristics, the watershed was divided into three zones. Factor analysis/principal component analysis, applied to analyze the data sets of the three different groups obtained from cluster analysis, resulted in four latent factors accounting for 71.62%, 71.77% and 72.01% of the total variance in water quality data sets of LP, MP and HP areas, respectively. The PCs obtained from factor analysis indicate that the parameters for water quality variations are mainly related to dissolve heavy metals. Thus, these methods are believed to be valuable to help water resources managers understand complex nature of water quality issues and determine the priorities to improve water quality.     

Artificial neural network modeling of dissolved oxygen in reservoir

The water quality of reservoirs is one of the key factors in the operation and water quality management of reservoirs. Dissolved oxygen (DO) in water column is essential for microorganisms and a significant indicator of the state of aquatic ecosystems. In this study, two artificial neural network (ANN) models including back propagation neural network (BPNN) and adaptive neural-based fuzzy inference system (ANFIS) approaches and multilinear regression (MLR) model were developed to estimate the DO concentration in the Feitsui Reservoir of northern Taiwan. The input variables of the neural network are determined as water temperature, pH, conductivity, turbidity, suspended solids, total hardness, total alkalinity, and ammonium nitrogen. The performance of the ANN models and MLR model was assessed through the mean absolute error, root mean square error, and correlation coefficient computed from the measured and model-simulated DO values. The results reveal that ANN estimation performances were superior to those of MLR. Comparing to the BPNN and ANFIS models through the performance criteria, the ANFIS model is better than the BPNN model for predicting the DO values. Study results show that the neural network particularly using ANFIS model is able to predict the DO concentrations with reasonable accuracy, suggesting that the neural network is a valuable tool for reservoir management in Taiwan.     

Using GA-Ridge regression to select hydro-geological parameters influencing groundwater pollution vulnerability

For groundwater conservation and management, it is important to accurately assess groundwater pollution vulnerability. This study proposed an integrated model using ridge regression and a genetic algorithm (GA) to effectively select the major hydro-geological parameters influencing groundwater pollution vulnerability in an aquifer. The GA-Ridge regression method determined that depth to water, net recharge, topography, and the impact of vadose zone media were the hydro-geological parameters that influenced trichloroethene pollution vulnerability in a Korean aquifer. When using these selected hydro-geological parameters, the accuracy was improved for various statistical nonlinear and artificial intelligence (AI) techniques, such as multinomial logistic regression, decision trees, artificial neural networks, and case-based reasoning. These results provide a proof of concept that the GA-Ridge regression is effective at determining influential hydro-geological parameters for the pollution vulnerability of an aquifer, and in turn, improves the AI performance in assessing groundwater pollution vulnerability.     

Formation and modeling of disinfection by-products in drinking water of six cities in China

Water quality parameters including TOC, UV(254), pH, chlorine dosage, bromide concentration and disinfection by-products were measured in water samples from 41 water treatment plants of six selected cities in China. Chloroform, bromodichloromethane, dibromochloromethane, dichloroacetic acid and trichloroacetic acid were the major disinfection by-products in the drinking water of China. Bromoform and dibromoacetic acid were also detected in many water samples. Higher concentrations of trihalomethanes and haloacetic acids were measured in summer compared to winter. The geographical variations in DBPs showed that TTHM levels were higher in Zhengzhou and Tianjin than other selected cities. And the HAA5 levels were highest in Changsha and Tianjin. The modeling procedure that predicts disinfection by-products formation was studied and developed using artificial neural networks. The performance of the artificial neural networks model was excellent (r > 0.84).     

Water quality of a tributary of the Pearl River, the Beijiang, Southern China: implications from multivariate statistical analyses

Water quality information of Beijiang River, a tributary of Pearl River in Guangdong, China, was analyzed to provide an overview of the hydrochemical functioning of a major agricultural/rural area and an industrial/urban area. Eighteen water quality parameters were surveyed at 13 sites from 2005 to 2006 on a monthly basis. A bivariate correlation analysis was carried out to evaluate the regional correlations of the water quality parameters, while the principal component analysis (PCA) technique was used to extract the most influential variables for regional variations of river water quality. Six principal components were extracted in PCA which explained more than 78% and 84% of the total variance for agricultural/rural and industrial/urban areas, respectively. Physicochemical factor, organic pollution, sewage pollution, geogenic factor, agricultural nonpoint source pollution, and accumulated pesticide usage were identified as potential pollution sources for agricultural/rural area, whereas industrial wastewaters pollution, mineral pollution, geogenic factor, urban sewage pollution, chemical industrial pollution, and water traffic pollution were the latent pollution sources for industrial/urban area. A multivariate linear regression of absolute principal component scores (MLR-APCS) technique was used to estimate contributions of all identified pollution sources to each water quality parameter. High coefficients of determination of the regression equations suggested that the MLR-APCS model was applicable for estimation of sources of most water quality parameters in the Beijiang River Basin.     

The use of multivariate statistical methods for optimization of the surface water quality network monitoring in the Paraopeba river basin,Brazil

This study sought to evaluate and propose adjustments to the water quality monitoring network of surface freshwaters in the Paraopeba river basin (Minas Gerais, Brazil), using multivariate statistical methods. A total of 13,560 valid data were analyzed for 19 water quality parameters at 30 monitoring sites, over a period of 5 years (2008–2013). The cluster analysis grouped the monitoring sites in eight groups based on similarities of water quality characteristics. This analysis made it possible to detect the most relevant monitoring stations in the river basin. The principal components analysis associated with non-parametric tests and the analysis of violation of the standards prescribed by law, allowed for identifying the most relevant parameters which must be maintained in the network (thermotolerant coliforms, total manganese, and total phosphorus). The discharge of domestic sewage and industrial wastewater, that from mining activities and diffuse pollution from agriculture and pasture areas are the main sources of pollution responsible for the surface water quality deterioration in this basin. The BP073 monitoring site presents the most degraded water quality in the Paropeba river basin. The monitoring sites BP094 and BP092 are located geographically close and they measure similar water quality, so a possible assessment of the need to maintain only one of the two in the monitoring network is suggested. Therefore, multivariate analyses were efficient to assess the adequacy of the water quality monitoring network of the Paraopeba river basin, and it can be used in other watersheds.     

Water quality change in dam reservoir and shallow aquifer: analysis on trend,seasonal variability and data reduction

Change of water quality in dam reservoir and aquifer complicates safe drinking water supply. Few parameters are monitored to control water quality in these sources. Adequate knowledge on the correlation structure, interaction effect, trends and seasonal variability of these parameters is essential to control water quality. This study applied time series and multivariate analyses on 15 water quality parameters, collected from the King Fahd dam reservoir (L1) and aquifer (L2) in Saudi Arabia during April 2010 to February 2012. Moderate to strong correlations were observed between sulfate, hardness, fluoride, chloride, magnesium, conductivity, turbidity and total dissolved solids (TDS), while separate clusters were visible for TDS-chloride-magnesium-conductivity; fluoride-turbidity; chloride-hardness; ammonia-nitrate; and calcium-magnesium-hardness. Four major principal components explained 81.1 % and 83.2 % of the overall variances in L1 and L2, respectively. The factor analysis showed that 53 % and 67 % of the data were necessary to explain 81.3 % and 83.2 % of total variances for L1 and L2, respectively, indicating the possibility of data reduction. Possible degradation of water quality in these sources was highlighted, while such degradation may require enhanced treatment for producing drinking water in future.     

Forest cover change and fragmentation using Landsat data in Maçka State Forest Enterprise in Turkey

In this study, Grey model (GM) and artificial neural network (ANN) were employed to predict suspended solids (SSeff) and chemical oxygen demand (CODeff) in the effluent from a wastewater treatment plant in industrial park of Taiwan. When constructing model or predicting, the influent quality or online monitoring parameters were adopted as the input variables. ANN was also adopted for comparison. The results indicated that the minimum MAPEs of 16.13 and 9.85% for SSeff and CODeff could be achieved using GMs when online monitoring parameters were taken as the input variables. Although a good fitness could be achieved using ANN, they required a large quantity of data. Contrarily, GM only required a small amount of data (at least four data) and the prediction results were even better than those of ANN. Therefore, GM could be applied successfully in predicting effluent when the information was not sufficient. The results also indicated that these simple online monitoring parameters could be applied on prediction of effluent quality well.     

Assessment of seasonal variations of surface water quality characteristics for Porsuk Stream

In this study, the factor analysis technique is applied to surface water quality data sets obtained from Porsuk stream in Turkey, generated during 10 years (1995-2005) monitoring of 29 parameters at one site (Esenkara) for all four seasons. The varifactors obtained from factor analysis indicate that the parameters responsible for water quality variations are mainly related to mineral and inorganic nutrients, organic pollution, microbiological pollution in winter and spring; mineral and nutrients in summer; microbiological and nutrient pollution in fall. This study presents the necessity and usefulness of multivariate statistical assessment of large and complex databases in order to get better information about the quality of surface water.     

A statistical approach for the assessment and redesign of the Nile Delta drainage system water-quality-monitoring locations

There are several deficiencies in the statistical approaches proposed in the literature for the assessment and redesign of surface water-quality-monitoring locations. These deficiencies vary from one approach to another, but generally include: (i) ignoring the attributes of the basin being monitored; (ii) handling multivariate water quality data sequentially rather than simultaneously; (iii) focusing mainly on locations to be discontinued; and (iv) ignoring the reconstitution of information at discontinued locations. In this paper, a methodology that overcomes these deficiencies is proposed. In the proposed methodology, the basin being monitored is divided into sub-basins, and a hybrid-cluster analysis is employed to identify groups of sub-basins with similar attributes. A stratified optimum sampling strategy is then employed to identify the optimum number of monitoring locations at each of the sub-basin groups. An aggregate information index is employed to identify the optimal combination of locations to be discontinued. The proposed approach is applied for the assessment and redesign of the Nile Delta drainage water quality monitoring locations in Egypt. Results indicate that the proposed methodology allows the identification of (i) the optimal combination of locations to be discontinued, (ii) the locations to be continuously measured and (iii) the sub-basins where monitoring locations should be added. To reconstitute information about the water quality variables at discontinued locations, regression, artificial neural network (ANN) and maintenance of variance extension (MOVE) techniques are employed. The MOVE record extension technique is shown to result in a better performance than regression or ANN for the estimation of information about water quality variables at discontinued locations.     

Characterization and source apportionment of water pollution in Jinjiang River, China

Characterizing water quality and identifying potential pollution sources could greatly improve our knowledge about human impacts on the river ecosystem. In this study, fuzzy comprehensive assessment (FCA), pollution index (PI), principal component analysis (PCA), and absolute principal component score–multiple linear regression (APCS–MLR) were combined to obtain a deeper understanding of temporal–spatial characterization and sources of water pollution with a case study of the Jinjiang River, China. Measurement data were obtained with 17 water quality variables from 20 sampling sites in the December 2010 (withered water period) and June 2011 (high flow period). FCA and PI were used to comprehensively estimate the water quality variables and compare temporal–spatial variations, respectively. Rotated PCA and receptor model (APCS–MLR) revealed potential pollution sources and their corresponding contributions. Application results showed that comprehensive application of various multivariate methods were effective for water quality assessment and management. In the withered water period, most sampling sites were assessed as low or moderate pollution with characteristics pollutants of permanganate index and total nitrogen (TN), whereas 90 % sites were classified as high pollution in the high flow period with higher TN and total phosphorus. Agricultural non-point sources, industrial wastewater discharge, and domestic sewage were identified as major pollution sources. Apportionment results revealed that most variables were complicatedly influenced by industrial wastewater discharge and agricultural activities in withered water period and primarily dominated by agricultural runoff in high flow period.     

Assessment of surface water quality using multivariate statistical techniques: a case study of Behrimaz Stream, Turkey

Multivariate statistical techniques, such as cluster analysis (CA), principal component analysis, and factor analysis, were applied for the evaluation of temporal/spatial variations and for the interpretation of a water quality data set of the Behrimaz Stream, obtained during 1 year of monitoring of 20 parameters at four different sites. Hierarchical CA grouped 12 months into two periods (the first and second periods) and classified four monitoring sites into two groups (group A and group B), i.e., relatively less polluted (LP) and medium polluted (MP) sites, based on similarities of water quality characteristics. Factor analysis/principal component analysis, applied to the data sets of the two different groups obtained from cluster analysis, resulted in five latent factors amounting to 88.32% and 88.93% of the total variance in water quality data sets of LP and MP areas, respectively. Varifactors obtained from factor analysis indicate that the parameters responsible for water quality variations are mainly related to discharge, temperature, and soluble minerals (natural) and nutrients (nonpoint sources: agricultural activities) in relatively less polluted areas; and organic pollution (point source: domestic wastewater) and nutrients (nonpoint sources: agricultural activities and surface runoff from villages) in medium polluted areas in the basin. Thus, this study illustrates the utility of multivariate statistical techniques for analysis and interpretation of data sets and, in water quality assessment, identification of pollution sources/factors and understanding temporal/spatial variations in water quality for effective stream water quality management.     

遥感技术在湖泊叶绿素ａ 监测中的应用研究———以太湖为例

水质遥感技术在湖泊水质监测领域内的应用具有十分积极的意义。在总结现有水质遥感反演方法的基础上，选取了遥感指数法和神经网络法两种理论完全不同的反演方法，构建太湖叶绿素a与MODIS影像波段间的函数关系，并从反演能力和反演精度两个角度对上述方法进行了比较研究。结果表明，神经网络模型的非线性特征能够敏感地把握住叶绿素a浓度变化在反射波谱信息上的微小响应，较为成功地反演出叶绿素a与反射光谱信息间的非线性关系。神经网络模型的反演能力和反演精度均优于遥感指数方法，具有较好的应用前景。     

A data-mining approach to predict influent quality

In wastewater treatment plants, predicting influent water quality is important for energy management. The influent water quality is measured by metrics such as carbonaceous biochemical oxygen demand (CBOD), potential of hydrogen, and total suspended solid. In this paper, a data-driven approach for time-ahead prediction of CBOD is presented. Due to limitations in the industrial data acquisition system, CBOD is not recorded at regular time intervals, which causes gaps in the time–series data. Numerous experiments have been performed to approximate the functional relationship between the input and output parameters and thereby fill in the missing CBOD data. Models incorporating seasonality effects are investigated. Four data-mining algorithms—multilayered perceptron, classification and regression tree, multivariate adaptive regression spline, and random forest—are employed to construct prediction models with the maximum prediction horizon of 5 days.     

Artificial neural network modeling of dissolved oxygen in the Heihe River, Northwestern China

Identification and quantification of dissolved oxygen (DO) profiles of river is one of the primary concerns for water resources managers. In this research, an artificial neural network (ANN) was developed to simulate the DO concentrations in the Heihe River, Northwestern China. A three-layer back-propagation ANN was used with the Bayesian regularization training algorithm. The input variables of the neural network were pH, electrical conductivity, chloride (Cl^?), calcium (Ca²⁺), total alkalinity, total hardness, nitrate nitrogen (NO₃-N), and ammonical nitrogen (NH₄-N). The ANN structure with 14 hidden neurons obtained the best selection. By making comparison between the results of the ANN model and the measured data on the basis of correlation coefficient (r) and root mean square error (RMSE), a good model-fitting DO values indicated the effectiveness of neural network model. It is found that the coefficient of correlation (r) values for the training, validation, and test sets were 0.9654, 0.9841, and 0.9680, respectively, and the respective values of RMSE for the training, validation, and test sets were 0.4272, 0.3667, and 0.4570, respectively. Sensitivity analysis was used to determine the influence of input variables on the dependent variable. The most effective inputs were determined as pH, NO₃-N, NH₄-N, and Ca²⁺. Cl^? was found to be least effective variables on the proposed model. The identified ANN model can be used to simulate the water quality parameters.     

Developing an environmental water quality monitoring program for Haraz River in Northern Iran

Water quality management plans are an indispensable strategy for conservation and utilization of water resources in a sustainable manner. One common industrial use of water is aquaculture. The present study is an attempt to use statistical analyses in order to prepare an environmental water quality monitoring program for Haraz River, in Northern Iran. For this purpose, the analysis of a total number of 18 physicochemical parameters was performed at 15 stations during a 1-year sampling period. According to the results of the multivariate statistical methods, the optimal monitoring would be possible by only 3 stations and 12 parameters, including NH₃, EC, BOD, TSS, DO, PO₄, NO₃, TDS, temperature, turbidity, coliform, and discharge. In other words, newly designed network, with a total number of 36 measurements (3 stations × 12 parameters = 36 parameters), could achieve exactly the same performance as the former network, designed based on 234 measurements (13 stations × 18 parameters = 234 parameters). Based on the results of cluster, principal component, and factor analyses, the stations were divided into three groups of high pollution (HP), medium pollution (MP), and low pollution (LP). By clustering the stations, it would be possible to track the water quality of Haraz River, only by one station at each cluster, which facilitates rapid assessment of the water quality in the river basin. Emphasizing on three main axes of monitoring program, including measurement parameters, sampling frequency, and spatial pattern of sampling points, the water quality monitoring program was optimized for the river basin based on natural conditions of the study area, monitoring objectives, and required financial resources (a total annual cost of about US $2625, excluding the overhead costs).