Spatial Variability of Rare Earth Elements in Groundwater in the Vicinity of a Coal-Fired Power Plant and Associated Health Risk.

This study investigated the occurrence and distribution of rare earth elements (REEs), including 14 lanthanoids, scandium (Sc), and yttrium (Y), in groundwater around a large coal-fired thermal power plant (TPP). The ICP-MS technique was used to analyze 16 REEs in groundwater samples collected from monitoring wells. REE concentrations ranged from 59.9 to 758 ng/L, with an average of 290 ng/L. The most abundant was Sc, followed by La, accounting for 54.2% and 21.4% of the total REE concentration, respectively. Geospatial analysis revealed the REE enrichment at several hotspots near the TPP. The highest REE concentrations were observed near the TPP and ash landfill, decreasing with the distance from the plant and the landfill. REE fractionation ratios and anomalies suggested the Light REE dominance, comprising over 78% of the total REEs. Correlation and principal component analyses indicated similar behavior and sources for most REEs. Health risk assessment found hazard indices (HI) of 1.36 × 10-3 and 1.98 × 10-3 for adults and children, respectively, which are far below the permissible limit (HI = 1). Likewise, incremental lifetime cancer risks (ILCR) were all below 1 × 10-6. Nevertheless, ongoing ash disposal and potential accumulation in the environment could elevate the REE exposure over time.

Ecological and Health Risks Attributed to Rare Earth Elements in Coal Fly Ash.

The occurrence and distribution of yttrium and rare earth elements (REYs), along with major elements and heavy metal(loid)s (HMs) in coal fly ash (CFA) from five coal-fired power plants (CFPPs), were analyzed, and the REY-associated ecological and health risks were assessed. The individual REYs in CFA were abundant in the following order: Ce > La > Nd > Y > Pr > Gd > Sm > Dy > Er > Yb > Eu > Ho > Tb > Tm > Lu. The total REY content ranged from 135 to 362 mg/kg, averaging 302 mg/kg. The mean light-to-heavy REY ratio was 4.1, indicating prevalent light REY enrichment in CFA. Significantly positive correlations between the REYs suggested that they coexist and share similar origins in CFA. REYs were estimated to pose low to moderate ecological risks, with risk index (RI) values ranging from 66 to 245. The hazard index (HI) and target cancer risk (TCR) of REYs from CFA, estimated to be higher for children (HIc = 0.15, TCRc = 8.4 × 10-16) than for adults (HIa = 0.017, TCRa = 3.6 × 10-16), were well below the safety limits (HI = 1, TCR = 1.0 × 10-6). However, the danger to human health posed by HMs in the same CFA samples (HIc = 5.74, TCRc = 2.6 × 10-4, TCRa = 1.1 × 10-4) exceeded the safe thresholds (excl. HIa = 0.63). The mean RI and HI attributed to REYs in CFA were 14% and 2.6%, respectively, of the total risks that include HMs.

Lignin Microspheres Modified with Magnetite Nanoparticles as a Selenate Highly Porous Adsorbent.

Highly porous lignin-based microspheres, modified by magnetite nanoparticles, were used for the first time for the removal of selenate anions, Se(VI), from spiked and real water samples. The influence of experimental conditions: selenate concentration, adsorbent dosage and contact time on the adsorption capacity was investigated in a batch experimental mode. The FTIR, XRD, SEM techniques were used to analyze the structural and morphological properties of the native and exhausted adsorbent. The maximum adsorption capacity was found to be 69.9 mg/g for Se(VI) anions at pH 6.46 from the simulated water samples. The modified natural polymer was efficient in Se(VI) removal from the real (potable) water samples, originated from six cities in the Republic of Serbia, with an overage efficacy of 20%. The regeneration capacity of 61% in one cycle of desorption (0.5 M NaOH as desorption solution) of bio-based adsorbent was gained in this investigation. The examined material demonstrated a significant affinity for Se(VI) oxyanion, but a low potential for multi-cycle material application; consequently, the loaded sorbent could be proposed to be used as a Se fertilizer.

A green approach to starch modification by solvent-free method with betaine hydrochloride.

In this study, a novel simple and eco-efficient, semi-dry method with a spray system for starch modification has been developed. Compared to conventional semi-dry methods, this method does not use solvents so that no slurry or semi-liquid mixture is obtained, the material is in a moisted/semi-moisted state. The modification of starch was performed using betaine hydrochloride (BHC) as the cationic reagent, and the characteristics of such starch derivates were compared with cationic starches obtained using glycidyltrimethylammonium chloride (GTMAC). Due to the instability, toxicity, and high cost of the most commonly used GTMAC, it should be replaced with more eco-friendly reagents, such as BHC, which is derived from betaine found in most green plants (e.g., spinach - Spinacia oleracea, beets - Beta vulgaris). The influence of processing conditions such as temperature, concentration of cationic reagents, presence and concentration of natural plasticizers/catalyst on physico-chemical and structural properties of cationic starches have also been studied. The cationic degree varied from 0.045-0.204 for the starch-BHC samples and within the range of 0.066-0.245 for the starch-GTMAC samples. The modification of starch with cationic reagents resulted in an increased solubility and swelling capacity, followed by decreased viscosity of the modified starches.

Urban population exposure to tropospheric ozone: A multi-country forecasting of SOMO35 using artificial neural networks.

Urban population exposure to tropospheric ozone is a serious health concern in Europe countries. Although there are insufficient evidence to derive a level below which ozone has no effect on mortality WHO (World Health Organization) uses SOMO35 (sum of means over 35 ppb) in their health impact assessments. Is this paper, the artificial neural network (ANN) approach was used to forecast SOMO35 at the national level for a set of 24 European countries, mostly EU members. Available ozone precursors' emissions, population and climate data for the period 2003-2013 were used as inputs. Trend analysis had been performed using the linear regression of SOMO35 over time, and it has demonstrated that majority of the studied countries have a decreasing trend of SOMO35 values. The created models have made majority of predictions (≈60%) with satisfactory accuracy (relative error <20%) on testing, while the best performing model had R2 = 0.87 and overall relative error of 33.6%. The domain of applicability of the created models was analyzed using slope/mean ratio derivate from the trend analysis, which was successful in distinguishing countries with high from countries with low prediction errors. The overall relative error was reduced to <14%, after the pool of countries was reduced based on the abovementioned criterion.

Virtual water quality monitoring at inactive monitoring sites using Monte Carlo optimized artificial neural networks: A case study of Danube River (Serbia).

Rationalization of water quality monitoring stations nowadays is applied in many countries. In some cases, missing data from abandoned/inactive stations, spatial and temporal, could be very important, hence the use of artificial neural networks (ANNs) for virtual water quality monitoring at inactive monitoring sites was investigated. The aim was to develop single-output and simultaneous ANNs for the spatial interpolation of 18 water quality parameters at single- and multi-inactive monitoring sites on Danube River course through Serbia. Those different modeling approaches were considered in order to determine the most suitable combination of models. The variable selection and sensitivity analysis in the case of simultaneous models were performed using a modified procedure based on Monte Carlo Simulations (MCS). In general, the multi-target models tend to be more accurate than single target ones, while single output models outperform the simultaneous ones. Hence, for particular monitoring network and set of water quality parameters the optimal combination of models must be defined based on model's accuracy and computational effort needed. The MCS selection procedure has proved to be efficient only in the case of simultaneous multi-target model. MCS based analysis of input-output interactions has shown all significant interactions in the case of simultaneous single-target are grouped as a complex cluster of interactions, where majority of inputs influence on several outputs. In the case multi-target model those interactions were portioned in five separate clusters, there majority of them mimic the input-output interactions that are present in single output models. The modeling strategy for study area was proposed on the basis of the performance of created models (mean average percentage error < 10%): simultaneous multi-target model for pH, alkalinity, conductivity, hardness, dissolved oxygen, HCO3-, SO42- and Ca, single-output multi-target models for temperature and Cl-, simultaneous single-target models for Mg and CO2, single output single target models for NO3-.

Application of experimental design for the optimization of artificial neural network-based water quality model: a case study of dissolved oxygen prediction.

This paper presents an application of experimental design for the optimization of artificial neural network (ANN) for the prediction of dissolved oxygen (DO) content in the Danube River. The aim of this research was to obtain a more reliable ANN model that uses fewer monitoring records, by simultaneous optimization of the following model parameters: number of monitoring sites, number of historical monitoring data (expressed in years), and number of input water quality parameters used. Box-Behnken three-factor at three levels experimental design was applied for simultaneous spatial, temporal, and input variables optimization of the ANN model. The prediction of DO was performed using a feed-forward back-propagation neural network (BPNN), while the selection of most important inputs was done off-model using multi-filter approach that combines a chi-square ranking in the first step with a correlation-based elimination in the second step. The contour plots of absolute and relative error response surfaces were utilized to determine the optimal values of design factors. From the contour plots, two BPNN models that cover entire Danube flow through Serbia are proposed: an upstream model (BPNN-UP) that covers 8 monitoring sites prior to Belgrade and uses 12 inputs measured in the 7-year period and a downstream model (BPNN-DOWN) which covers 9 monitoring sites and uses 11 input parameters measured in the 6-year period. The main difference between the two models is that BPNN-UP utilizes inputs such as BOD, P, and PO43-, which is in accordance with the fact that this model covers northern part of Serbia (Vojvodina Autonomous Province) which is well-known for agricultural production and extensive use of fertilizers. Both models have shown very good agreement between measured and predicted DO (with R2 ≥ 0.86) and demonstrated that they can effectively forecast DO content in the Danube River.

A linear and non-linear polynomial neural network modeling of dissolved oxygen content in surface water: Inter- and extrapolation performance with inputs' significance analysis.

Accurate prediction of water quality parameters (WQPs) is an important task in the management of water resources. Artificial neural networks (ANNs) are frequently applied for dissolved oxygen (DO) prediction, but often only their interpolation performance is checked. The aims of this research, beside interpolation, were the determination of extrapolation performance of ANN model, which was developed for the prediction of DO content in the Danube River, and the assessment of relationship between the significance of inputs and prediction error in the presence of values which were of out of the range of training. The applied ANN is a polynomial neural network (PNN) which performs embedded selection of most important inputs during learning, and provides a model in the form of linear and non-linear polynomial functions, which can then be used for a detailed analysis of the significance of inputs. Available dataset that contained 1912 monitoring records for 17 water quality parameters was split into a "regular" subset that contains normally distributed and low variability data, and an "extreme" subset that contains monitoring records with outlier values. The results revealed that the non-linear PNN model has good interpolation performance (R2=0.82), but it was not robust in extrapolation (R2=0.63). The analysis of extrapolation results has shown that the prediction errors are correlated with the significance of inputs. Namely, the out-of-training range values of the inputs with low importance do not affect significantly the PNN model performance, but their influence can be biased by the presence of multi-outlier monitoring records. Subsequently, linear PNN models were successfully applied to study the effect of water quality parameters on DO content. It was observed that DO level is mostly affected by temperature, pH, biological oxygen demand (BOD) and phosphorus concentration, while in extreme conditions the importance of alkalinity and bicarbonates rises over pH and BOD.

Prediction of municipal solid waste generation using artificial neural network approach enhanced by structural break analysis.

This paper presents the development of a general regression neural network (GRNN) model for the prediction of annual municipal solid waste (MSW) generation at the national level for 44 countries of different size, population and economic development level. Proper modelling of MSW generation is essential for the planning of MSW management system as well as for the simulation of various environmental impact scenarios. The main objective of this work was to examine the potential influence of economy crisis (global or local) on the forecast of MSW generation obtained by the GRNN model. The existence of the so-called structural breaks that occur because of the economic crisis in the studied period (2000-2012) for each country was determined and confirmed using the Chow test and Quandt-Andrews test. Two GRNN models, one which did not take into account the influence of the economic crisis (GRNN) and another one which did (SB-GRNN), were developed. The novelty of the applied method is that it uses broadly available social, economic and demographic indicators and indicators of sustainability, together with GRNN and structural break testing for the prediction of MSW generation at the national level. The obtained results demonstrate that the SB-GRNN model provide more accurate predictions than the model which neglected structural breaks, with a mean absolute percentage error (MAPE) of 4.0 % compared to 6.7 % generated by the GRNN model. The proposed model enhanced with structural breaks can be a viable alternative for a more accurate prediction of MSW generation at the national level, especially for developing countries for which a lack of MSW data is notable.

Modeling the BOD of Danube River in Serbia using spatial, temporal, and input variables optimized artificial neural network models.

This paper describes the application of artificial neural network models for the prediction of biological oxygen demand (BOD) levels in the Danube River. Eighteen regularly monitored water quality parameters at 17 stations on the river stretch passing through Serbia were used as input variables. The optimization of the model was performed in three consecutive steps: firstly, the spatial influence of a monitoring station was examined; secondly, the monitoring period necessary to reach satisfactory performance was determined; and lastly, correlation analysis was applied to evaluate the relationship among water quality parameters. Root-mean-square error (RMSE) was used to evaluate model performance in the first two steps, whereas in the last step, multiple statistical indicators of performance were utilized. As a result, two optimized models were developed, a general regression neural network model (labeled GRNN-1) that covers the monitoring stations from the Danube inflow to the city of Novi Sad and a GRNN model (labeled GRNN-2) that covers the stations from the city of Novi Sad to the border with Romania. Both models demonstrated good agreement between the predicted and actually observed BOD values.

Estimation of NMVOC emissions using artificial neural networks and economical and sustainability indicators as inputs.

This paper describes the development of an artificial neural network (ANN) model based on economical and sustainability indicators for the prediction of annual non-methane volatile organic compounds (NMVOCs) emissions in China for the period 2005-2011 and its comparison with inventory emission factor models. The NMVOCs emissions in China were estimated using ANN model which was created using available data for nine European countries, which NMVOC emission per capita approximately correspond to the Chinese emissions, for the period 2004-2012. The forward input selection strategy was used to compare the significance of particular inputs for the prediction of NMVOC emissions in the nine selected EU countries and China. The final ANN model was trained using only five input variables, and it has demonstrated similar accuracy in predicting NMVOC emissions for the selected EU countries that were used for the development of the model and then for China for which the input dataset was previously unknown to the ANN model. The obtained mean absolute percentage error (MAPE) values were 8 % for EU countries and 5 % for China. Also, the temporal trend of NMVOC emissions predicted in this study is generally consistent with the trend obtained using inventory emission models. The proposed ANN approach can represent a viable alternative for the prediction of NMVOC emissions at the national level, in particular for developing countries which are usually lacking emission data.

Chemometrics in biomonitoring: Distribution and correlation of trace elements in tree leaves.

The concentrations of 15 elements were measured in the leaf samples of Aesculus hippocastanum, Tilia spp., Betula pendula and Acer platanoides collected in May and September of 2014 from four different locations in Belgrade, Serbia. The objective was to assess the chemical characterization of leaf surface and in-wax fractions, as well as the leaf tissue element content, by analyzing untreated, washed with water and washed with chloroform leaf samples, respectively. The combined approach of self-organizing networks (SON) and Preference Ranking Organization Method for Enrichment Evaluation (PROMETHEE) aided by Geometrical Analysis for Interactive Aid (GAIA) was used in the interpretation of multiple element loads on/in the tree leaves. The morphological characteristics of the leaf surfaces and the elemental composition of particulate matter (PM) deposited on tree leaves were studied by using scanning electron microscopy (SEM) with energy dispersive spectroscopy (EDS) detector. The results showed that the amounts of retained and accumulated element concentrations depend on several parameters, such as chemical properties of the element and morphological properties of the leaves. Among the studied species, Tilia spp. was found to be the most effective in the accumulation of elements in leaf tissue (70% of the total element concentration), while A. hippocastanum had the lowest accumulation (54%). After water and chloroform washing, the highest percentages of removal were observed for Al, V, Cr, Cu, Zn, As, Cd and Sb (>40%). The PROMETHEE/SON ranking/classifying results were in accordance with the results obtained from the GAIA clustering techniques. The combination of the techniques enabled extraction of additional information from datasets. Therefore, the use of both the ranking and clustering methods could be a useful tool to be applied in biomonitoring studies of trace elements.

The novel approach to the biomonitor survey using one- and two-dimensional Kohonen networks.

To compare the applicability of the leaves of horse chestnut (Aesculus hippocastanum) and linden (Tilia spp.) as biomonitors of trace element concentrations, a coupled approach of one- and two-dimensional Kohonen networks was applied for the first time. The self-organizing networks (SONs) and the self-organizing maps (SOMs) were applied on the database obtained for the element accumulation (Cr, Fe, Ni, Cu, Zn, Pb, V, As, Cd) and the SOM for the Pb isotopes in the leaves for a multiyear period (2002-2006). A. hippocastanum seems to be a more appropriate biomonitor since it showed more consistent results in the analysis of trace elements and Pb isotopes. The SOM proved to be a suitable and sensitive tool for assessing differences in trace element concentrations and for the Pb isotopic composition in leaves of different species. In addition, the SON provided more clear data on seasonal and temporal accumulation of trace elements in the leaves and could be recommended complementary to the SOM analysis of trace elements in biomonitoring studies.

Modeling of ammonia emission in the USA and EU countries using an artificial neural network approach.

Ammonia emissions at the national level are frequently estimated by applying the emission inventory approach, which includes the use of emission factors, which are difficult and expensive to determine. Emission factors are therefore the subject of estimation, and as such they contribute to inherent uncertainties in the estimation of ammonia emissions. This paper presents an alternative approach for the prediction of ammonia emissions at the national level based on artificial neural networks and broadly available sustainability and economical/agricultural indicators as model inputs. The Multilayer Perceptron (MLP) architecture was optimized using a trial-and-error procedure, including the number of hidden neurons, activation function, and a back-propagation algorithm. Principal component analysis (PCA) was applied to reduce mutual correlation between the inputs. The obtained results demonstrate that the MLP model created using the PCA transformed inputs (PCA-MLP) provides a more accurate prediction than the MLP model based on the original inputs. In the validation stage, the MLP and PCA-MLP models were tested for ammonia emission predictions for up to 2 years and compared with a principal component regression model. Among the three models, the PCA-MLP demonstrated the best performance, providing predictions for the USA and the majority of EU countries with a relative error of less than 20%.

Artificial neural network modelling of biological oxygen demand in rivers at the national level with input selection based on Monte Carlo simulations.

Biological oxygen demand (BOD) is the most significant water quality parameter and indicates water pollution with respect to the present biodegradable organic matter content. European countries are therefore obliged to report annual BOD values to Eurostat; however, BOD data at the national level is only available for 28 of 35 listed European countries for the period prior to 2008, among which 46% of data is missing. This paper describes the development of an artificial neural network model for the forecasting of annual BOD values at the national level, using widely available sustainability and economical/industrial parameters as inputs. The initial general regression neural network (GRNN) model was trained, validated and tested utilizing 20 inputs. The number of inputs was reduced to 15 using the Monte Carlo simulation technique as the input selection method. The best results were achieved with the GRNN model utilizing 25% less inputs than the initial model and a comparison with a multiple linear regression model trained and tested using the same input variables using multiple statistical performance indicators confirmed the advantage of the GRNN model. Sensitivity analysis has shown that inputs with the greatest effect on the GRNN model were (in descending order) precipitation, rural population with access to improved water sources, treatment capacity of wastewater treatment plants (urban) and treatment of municipal waste, with the last two having an equal effect. Finally, it was concluded that the developed GRNN model can be useful as a tool to support the decision-making process on sustainable development at a regional, national and international level.

Ultrasonic assisted arsenate adsorption on solvothermally synthesized calcite modified by goethite, α-MnO2 and goethite/α-MnO2.

A highly porous calcium carbonate (calcite; sorbent 1) was used as a support for modification with α-FeOOH (calcite/goethite; sorbent 2), α-MnO2 (calcite/α-MnO2; sorbent 3) and α-FeOOH/α-MnO2 (calcite/goethite/α-MnO2; sorbent 4) in order to obtain a cheap hybrid materials for simple and effective arsenate removal from aqueous solutions. The adsorption ability of synthesized adsorbents was studied as a function of functionalization methods, pH, contact time, temperature and ultrasonic treatment. Comparison of the adsorptive effectiveness of synthesized adsorbents for arsenate removal, under ultrasound treatment and classical stirring method, has shown better performance of the former one reaching maximum adsorption capacities of 1.73, 21.00, 10.36 and 41.94 mg g(-1), for sorbents 1-4, respectively. Visual MINTEQ equilibrium speciation modeling was used for prediction of pH and interfering ion influences on arsenate adsorption.

Modelling of dissolved oxygen content using artificial neural networks: Danube River, North Serbia, case study.

The aims of this study are to create an artificial neural network (ANN) model using non-specific water quality parameters and to examine the accuracy of three different ANN architectures: General Regression Neural Network (GRNN), Backpropagation Neural Network (BPNN) and Recurrent Neural Network (RNN), for prediction of dissolved oxygen (DO) concentration in the Danube River. The neural network model has been developed using measured data collected from the Bezdan monitoring station on the Danube River. The input variables used for the ANN model are water flow, temperature, pH and electrical conductivity. The model was trained and validated using available data from 2004 to 2008 and tested using the data from 2009. The order of performance for the created architectures based on their comparison with the test data is RNN > GRNN > BPNN. The ANN results are compared with multiple linear regression (MLR) model using multiple statistical indicators. The comparison of the RNN model with the MLR model indicates that the RNN model performs much better, since all predictions of the RNN model for the test data were within the error of less than ± 10 %. In case of the MLR, only 55 % of predictions were within the error of less than ± 10 %. The developed RNN model can be used as a tool for the prediction of DO in river waters.

Bioreactor validation and biocompatibility of Ag/poly(N-vinyl-2-pyrrolidone) hydrogel nanocomposites.

Silver/poly(N-vinyl-2-pyrrolidone) (Ag/PVP) nanocomposites containing Ag nanoparticles at different concentrations were synthesized using γ-irradiation. Cytotoxicity of the obtained nanocomposites was determined by MTT assay in monolayer cultures of normal human immunocompetent peripheral blood mononuclear cells (PBMC) that were either non-stimulated or stimulated to proliferate by mitogen phytohemagglutinin (PHA), as well as in human cervix adenocarcinoma cell (HeLa) cultures. Silver release kinetics and mechanical properties of nanocomposites were investigated under bioreactor conditions in the simulated body fluid (SBF) at 37°C. The release of silver was monitored under static conditions, and in two types of bioreactors: perfusion bioreactors and a bioreactor with dynamic compression coupled with SBF perfusion simulating in vivo conditions in articular cartilage. Ag/PVP nanocomposites exhibited slight cytotoxic effects against PBMC at the estimated concentration of 0.4 µmol dm(-3), with negligible variations observed amongst different cell cultures investigated. Studies of the silver release kinetics indicated internal diffusion as the rate limiting step, determined by statistically comparable results obtained at all investigated conditions. However, silver release rate was slightly higher in the bioreactor with dynamic compression coupled with SBF perfusion as compared to the other two systems indicating the influence of dynamic compression. Modelling of silver release kinetics revealed potentials for optimization of Ag/PVP nanocomposites for particular applications as wound dressings or soft tissue implants.

PM(10) emission forecasting using artificial neural networks and genetic algorithm input variable optimization.

This paper describes the development of an artificial neural network (ANN) model for the forecasting of annual PM(10) emissions at the national level, using widely available sustainability and economical/industrial parameters as inputs. The inputs for the model were selected and optimized using a genetic algorithm and the ANN was trained using the following variables: gross domestic product, gross inland energy consumption, incineration of wood, motorization rate, production of paper and paperboard, sawn wood production, production of refined copper, production of aluminum, production of pig iron and production of crude steel. The wide availability of the input parameters used in this model can overcome a lack of data and basic environmental indicators in many countries, which can prevent or seriously impede PM emission forecasting. The model was trained and validated with the data for 26 EU countries for the period from 1999 to 2006. PM(10) emission data, collected through the Convention on Long-range Transboundary Air Pollution - CLRTAP and the EMEP Programme or as emission estimations by the Regional Air Pollution Information and Simulation (RAINS) model, were obtained from Eurostat. The ANN model has shown very good performance and demonstrated that the forecast of PM(10) emission up to two years can be made successfully and accurately. The mean absolute error for two-year PM(10) emission prediction was only 10%, which is more than three times better than the predictions obtained from the conventional multi-linear regression and principal component regression models that were trained and tested using the same datasets and input variables.