Combining Neural Network Models to Predict Spatial Patterns of Airborne Pollutant Accumulation in Soils around an Industrial Point Emission Source

Abstract

Neural networks (NNs) have the ability to model a wide range of complex nonlinearities. A major disadvantage of NNs, however, is their instability, especially under conditions of sparse, noisy, and limited data sets. In this paper, different combining network methods are used to benefit from the existence of local minima and from the instabilities of NNs. A nonlinear k-fold cross-validation method is used to test the performance of the various networks and also to develop and select a set of networks that exhibits a low correlation of errors. The various NN models are applied to estimate the spatial patterns of atmospherically transported and deposited lead (Pb) in soils around an historical industrial air emission point source. It is shown that the resulting ensemble networks consistently give superior predictions compared with the individual networks because, for the ensemble networks, R² values were found to be higher than 0.9 while, for the contributing individual networks, values for R² ranged between 0.35 and 0.85. It is concluded that combining networks can be adopted as an important component in the application of artificial NN techniques in applied air quality studies.     

A Comparison of Nonlinear Regression and Neural Network Models for Ground-Level Ozone Forecasting

ABSTRACT

A hybrid nonlinear regression (NLR) model and a neural network (NN) model, each designed to forecast next-day maximum 1-hr average ground-level O₃ concentrations in Louisville, KY, were compared for two O₃ seasons—1998 and 1999. The model predictions were compared for the forecast mode, using forecasted meteorological data as input, and for the hindcast mode, using observed meteorological data as input. The two models performed nearly the same in the forecast mode. For the two seasons combined, the mean absolute forecast error was 12.5 ppb for the NLR model and 12.3 ppb for the NN model. The detection rate of 120 ppb threshold exceedances was 42% for each model in the forecast mode. In the hindcast mode, the NLR model performed marginally better than the NN     

Forecasting Daily Source Air Quality Using Multivariate Statistical Analysis and Radial Basis Function Networks

Abstract

It is vital to forecast gas and particle matter concentrations and emission rates (GPCER) from livestock production facilities to assess the impact of airborne pollutants on human health, ecological environment, and global warming. Modeling source air quality is a complex process because of abundant nonlinear interactions between GPCER and other factors. The objective of this study was to introduce statistical methods and radial basis function (RBF) neural network to predict daily source air quality in Iowa swine deep-pit finishing buildings. The results show that four variables (outdoor and indoor temperature, animal units, and ventilation rates) were identified as relative important model inputs using statistical methods. It can be further demonstrated that only two factors, the environment factor and the animal factor, were capable of explaining more than 94% of the total variability after performing principal component analysis. The introduction of fewer uncorrelated variables to the neural network would result in the reduction of the model structure complexity, minimize computation cost, and eliminate model overfitting problems. The obtained results of RBF network prediction were in good agreement with the actual measurements, with values of the correlation coefficient between 0.741 and 0.995 and very low values of systemic performance indexes for all the models. The good results indicated the RBF network could be trained to model these highly nonlinear relationships. Thus, the RBF neural network technology combined with multivariate statistical methods is a promising tool for air pollutant emissions modeling.     

The estimation of historical CO2 trajectories is indeterminate: Comment on “A new look at atmospheric carbon dioxide”

A paper by Hofmann et al. (2009, this journal) is critiqued. It is shown that their exponential model for characterizing CO₂ trajectories for historical data is not estimated properly. An exponential model is properly estimated and is shown to fit over the entire 51 year period of available data. Further, the entire problem of estimating models for the CO₂ historical data is shown to be ill-posed because alternate model forms fit the data equally well. To illustrate this point the past 51 years of CO₂ data were analyzed using three different time-dependent models that capture the historical pattern of CO₂ increase. All three fit with R² > 0.98, are visually indistinguishable when overlaid, and match each other during the calibration period with R² > 0.999. Projecting the models forward to 2100, the exponential model comes quite close to the Intergovernmental Panel on Climate Change (IPCC) best estimate of 836 ppmv. The other two models project values far below the IPCC low estimates. The problem of characterizing historical CO₂ levels is thus indeterminate, because multiple models fit the data equally well but forecast very different future trajectories.     

A model for predicting maximum and 8 h average ozone in Houston

The objective of this research was to develop a statistical model to predict one day in advance both the maximum and 8 h (10 am–5 pm) average ozone for Houston (TX). A loess/generalized additive model (GAM) approach was taken to model development. Ozone data (1983–1991) from ten stations in the immediate Houston area were used in the study. The meteorological data came from the Houston International Airport. The models were developed using data for April through October for 1983–1987 and 1989–1990. Forecasts were developed for 1988 and 1991. The final model, which was multiplicative in nature, contained three interaction terms for the west/east and south/north wind components (average of hourly values from 8 pm to 5 am, 6 am to 9 am, and 10 am to 5 pm). Opaque cloud cover (averaged over the period 10  am to 5 pm), yesterday’s maximum ozone, today’s maximum temperature and morning mixing depth were also important variables in the model.Individual forecasts were generated for all ten stations in the Houston area using observed meteorology. In addition forecasts were produced for three measures of the network as a whole. The root-mean-square prediction error for the 8 h average forecasts ranged from 13.2 to 16.3 ppb (with R² ranging from 0.66 to 0.73) for the individual stations and from 18.5 to 22.0 ppb (with R² ranging from 0.61 to 0.68) for maximum ozone. A detailed examination was undertaken for a day on which the forecast was much too low.     

Testing the capability of the chemistry transport model LOTOS-EUROS to forecast PM10 levels in the Netherlands

Since particulate matter has a direct and adverse impact on public health, a good air quality forecast is important. Several European countries presently use statistical forecasting models, which have their limitations, especially for PM10. An alternative approach is to use a chemistry transport model. Here, the ability of the chemical transport model LOTOS-EUROS to forecast PM10 concentrations in the Netherlands was investigated. LOTOS-EUROS models several PM10 components individually. For sulphate, nitrate and ammonium aerosol the evaluation against observations shows that the modelled annual mean concentrations are within 20% of the measured concentration and that the temporal correlation is reasonably good (R > 0.6). For sea salt the model tended to overestimate the measured concentrations. For elemental carbon the correspondence with black smoke observations was reasonable. However, total PM10 is seriously underestimated, due to unmodelled components (secondary organic aerosols, mineral dust) and missing sources. Therefore, a simple bias correction for four seasons was derived based on the years 2004–2006. The model was compared with the Dutch operational statistical model PROPART and ground-level observations. With bias correction, LOTOS-EUROS performed better than PROPART regarding the timing of events. The major flaw of LOTOS-EUROS was that high values (>50 μg m^?3) were still underestimated. Another advantage of LOTOS-EUROS over the statistical model was the more detailed information in space and time, which facilitates communication of the forecast to the general public.     

Comparison of hybrid machine learning methods for the prediction of short-term meteorological droughts of Sakarya Meteorological Station in Turkey

Drought is a harmful natural disaster with various negative effects on many aspects of life. In this research, short-term meteorological droughts were predicted with hybrid machine learning models using monthly precipitation data (1960–2020 period) of Sakarya Meteorological Station, located in the northwest of Turkey. Standardized precipitation index (SPI), depending only on precipitation data, was used as the drought index, and 1-, 3-, and 6-month time scales for short-term droughts were considered. In the prediction models, drought index was predicted at t?+?1 output variable by using t, t???1, t???2, and t???3 input variables. Artificial neural networks (ANNs), adaptive neuro-fuzzy inference system (ANFIS), Gaussian process regression (GPR), support vector machine regression (SVMR), k-nearest neighbors (KNN) algorithms were employed as stand-alone machine learning methods. Variation mode decomposition (VMD), discrete wavelet transform (DWT), and empirical mode decomposition (EMD) were utilized as pre-processing techniques to create hybrid models. Six different performance criteria were used to assess model performance. The hybrid models used together with the pre-processing techniques were found to be more successful than the stand-alone models. Hybrid VMD-GPR model yielded the best results (NSE?=?0.9345, OI?=?0.9438, R²?=?0.9367) for 1-month time scale, hybrid VMD-GPR model (NSE?=?0.9528, OI?=?0.9559, R²?=?0.9565) for 3-month time scale, and hybrid DWT-ANN model (NSE?=?0.9398, OI?=?0.9483, R²?=?0.9450) for 6-month time scale. Considering the entire performance criteria, it was determined that the decomposition success of VMD was higher than DWT and EMD.

     

Modelling of silica diffusion experiments with Si in Boom Clay

A mathematical model describing the dissolution of nuclear glass directly disposed in clay combines a first-order dissolution rate law with the diffusion of dissolved silica in clay. According to this model, the main parameters describing the long-term dissolution of the glass are ηR, the product of the diffusion accessible porosity η and the retardation factor R, and the apparent diffusion coefficient D_app of dissolved silica in clay.For determining the migration parameters needed for long-term predictions, four Through-Diffusion (T-D) experiments and one percolation test have been performed on undisturbed clay cores. In the Through-Diffusion experiments, the concentration decrease after injection of ³²Si (radioactive labelled silica) was measured in the inlet compartment. At the end of the T-D experiments, the clay cores were cut in thin slices and the activity of labelled silica in each slice was determined. The measured activity profiles for these four clay cores are well reproducible.Since no labelled silica could be detected in the outlet compartments, the Through-Diffusion experiments are fitted by two In-Diffusion models: one model assuming linear and reversible sorption equilibrium and a second model taking into account sorption kinetics. Although the kinetic model provides better fits, due to the sufficiently long duration of the experiments, both models give approximately similar values for the fit parameters. The single percolation test leads to an apparent diffusion coefficient value about two to three times lower than those of the Through-Diffusion tests.Therefore, dissolved silica appears to be strongly retarded in Boom Clay. A retardation factor R between 100 and 300 was determined. The corresponding in situ distribution coefficient K_d is in the range 25–75 cm³ g⁻¹. The apparent diffusion coefficient of dissolved silica in Boom Clay is estimated between 2×10⁻¹³ and 7×10⁻¹³ m² s⁻¹. The pore diffusion coefficient is in the range from 6×10⁻¹¹ to 1×10⁻¹⁰ m² s⁻¹.     

Ferrate(VI) enhanced photocatalytic oxidation of pollutants in aqueous TiO<Subscript>2</Subscript> suspensions

Background, aim and scope  

Photocatalytic oxidation using UV irradiation of TiO₂ has been studied extensively and has many potential industrial applications, including the degradation of recalcitrant contaminants in water and wastewater treatment. A limiting factor in the oxidation process is the recombination of conduction band electrons (e ⁻ _cb) with electron holes (h_vb⁺) on the irradiated TiO₂ surface; thus, in aqueous conditions, the presence of an effective electron scavenger will be beneficial to the efficiency of the oxidation process. Ferrate (FeO₄²⁻) has received much recent attention as a water treatment chemical since it behaves simultaneously as an oxidant and coagulant. The combination of ferrate [Fe(VI)] with UV/TiO₂ photocatalysis offers an oxidation synergism arising from the Fe(VI) scavenging of e ⁻ _cb and the corresponding beneficial formation of Fe(V) from the Fe(VI) reduction. This paper reviews recent studies concerning the photocatalytic oxidation of problematic pollutants with and without ferrate.     

QSAR modeling for predicting mutagenic toxicity of diverse chemicals for regulatory purposes

The safety assessment process of chemicals requires information on their mutagenic potential. The experimental determination of mutagenicity of a large number of chemicals is tedious and time and cost intensive, thus compelling for alternative methods. We have established local and global QSAR models for discriminating low and high mutagenic compounds and predicting their mutagenic activity in a quantitative manner in Salmonella typhimurium (TA) bacterial strains (TA98 and TA100). The decision treeboost (DTB)-based classification QSAR models discriminated among two categories with accuracies of >96% and the regression QSAR models precisely predicted the mutagenic activity of diverse chemicals yielding high correlations (R ²) between the experimental and model-predicted values in the respective training (>0.96) and test (>0.94) sets. The test set root mean squared error (RMSE) and mean absolute error (MAE) values emphasized the usefulness of the developed models for predicting new compounds. Relevant structural features of diverse chemicals that were responsible and influence the mutagenic activity were identified. The applicability domains of the developed models were defined. The developed models can be used as tools for screening new chemicals for their mutagenicity assessment for regulatory purpose.

     

Consultant Guide/1986

ABSTRACT

To obtain annual odor emission profiles from intensive swine operations, odor concentrations and emission rates were measured monthly from swine nursery, farrowing, and gestation rooms for a year. Large annual variations in odor concentrations and emissions were found in all the rooms and the impact of the seasonal factor (month) was significant (P < 0.05). Odor concentration was low in summer when ventilation rate was high but high in winter when ventilation rate was low, ranging from 362 (farrowing room in July) to 8934 (nursery room in December) olfactory unit (OU) m^?3. This indicates that the air quality regarding odor was significantly better in summer than that in winter. Odor emission rate did not show obvious seasonal pattern as odor concentration did, ranging from 2 (gestation room in November) to 90 (nursery room in April) OU m^?2 sec^?1; this explains why the odor complaints for swine barns have occurred all year round. The annual geometric mean odor concentration and emission rate of the nursery room was significantly higher than the other rooms (P < 0.05). In order to obtain the representative annual emission rate, measurements have to be taken at least monthly, and then the geometric mean of the monthly values will represent the annual emission rate. Incorporating odor control technologies in the nursery area will be the most efficient in reducing odor emission from the farm considering its emission rate was 2 to 3 times of the other areas. The swine grower-finisher area was the major odor source contributing 53% of odor emission of the farm and should also be targeted for odor control. Relatively positive correlations between odor concentration and both H₂S and CO₂ concentrations (R ² = 0.58) means that high level of these two gases might likely indicate high odor concentration in swine barns.

IMPLICATIONS The emissions of air pollutants including odors, greenhouse gases, and toxic gases have become a major environmental issue facing animal farms in the U.S.A. and Canada. To ensure the air quality in the vicinity of intensive livestock farms, air dispersion models have been used to determine setback distances between livestock facilities and neighboring residences based on certain air quality requirement on odor and gases. Due to the limited odor emission data available, none of the existing models can take account of seasonal variations of odor emissions, which may result in great uncertainties in setback distance calculations. Therefore, the obtained seasonal odor and gas emission rates by this study can be used by the government regulatory organizations and researchers in air dispersion modeling to get improved calculation of setback distances.     

稻壳活性炭制备及其对磷的吸附

利用农业废弃物稻壳经炭化、活化、酸洗、水洗和干燥等工艺制备出一种富含微孔和中孔结构的稻壳活性炭,其BET比表面积达886.3 m2/g。通过正交实验优化了稻壳活性炭对磷吸附条件,并在该条件下进行了吸附等温和吸附动力学实验研究。结果表明,稻壳活性炭对磷的吸附等温曲线能较好符合Langmuir模型(R2=0.9284)和Freundlich模型(R2=0.9208),由Langmuir线性拟合方程可得稻壳活性炭对磷饱和吸附量达6.93 mg/g;稻壳活性炭对磷的吸附过程可用准二级动力学方程描述(R2=0.9968),吸附速度较快,颗粒内扩散为该过程控速阶段。稻壳活性炭作为一种易得、廉价、高效的填料,在农村分散型污水生态处理技术中,具有良好的应用前景。     

Neural network modelling and prediction of hourly NOx and NO2 concentrations in urban air in London

Multilayer perceptron (MLP) neural networks were trained to model hourly NO_x and NO₂ pollutant concentrations in Central London from basic hourly meteorological data. Results have shown that the models perform well when compared to previous attempts to model the same pollutants using regression based models. This work also illustrates that MLP neural networks are capable of resolving complex patterns of source emissions without any explicit external guidance.     

A comparison of nonlinear regression and neural network models for ground-level ozone forecasting

A hybrid nonlinear regression (NLR) model and a neural network (NN) model, each designed to forecast next-day maximum 1-hr average ground-level O3 concentrations in Louisville, KY, were compared for two O3 seasons--1998 and 1999. The model predictions were compared for the forecast mode, using forecasted meteorological data as input, and for the hindcast mode, using observed meteorological data as input. The two models performed nearly the same in the forecast mode. For the two seasons combined, the mean absolute forecast error was 12.5 ppb for the NLR model and 12.3 ppb for the NN model. The detection rate of 120 ppb threshold exceedances was 42% for each model in the forecast mode. In the hindcast mode, the NLR model performed marginally better than the NN model. The mean absolute hindcast error was 11.1 ppb for the NLR model and 12.9 ppb for the NN model. The hindcast detection rate was 92% for the NLR model and 75% for the NN model.     

Evaluation of three regional air quality models

This paper summarizes the results of a thorough assessment of existing regional air quality models. Forty-two candidate models were reviewed and three repsesentative models were selected for rigorous and comprehensive assessment. The underlying scientific theories used in the models were evaluated, revealing many limitations. For example, the techniques used in the preparation of meteorological fields that drive the models give insufficient consideration to the physical basis of the relevant atmospheric processes. The primary operational evaluation of each of the models was performed by comparing calculated values with observations from the EPRI Sulfate Regional Experiment (SURE). Both short-term (6-h averages) and long-term (annual averages) comparisons reveal poor correlations for both SO₂ and SO²⁻₄ for the three models evaluated ranging from 0.05 to 0.32 for 3- to 6-h SO₂ concentration to 0.03 to 0.59 for 24-4 and monthly averages; in some cases, the correlations are negative. The results also show that calculated concentrations are generally characterized by high biases for 3- to 6-h concentration predictions. Biases tend to be somewhat smaller for monthly averages. All three models underpredicted wet deposition with average normalized residuals of approximately 0.2 for ENAMAP-2, and 0.5 for RTM-II and ACID.     

载钴活性焦对As(V)的吸附性能与去除效果

使用沉淀负载法制备了载钴活性焦,并研究了溶液pH值、反应时间、As(V)初始浓度以及共存阴离子等对载钴活性焦吸附去除水环境中As(V)的影响。结果表明,(1)载钴后活性焦的比表面积和孔容积分别提高了20.87%和43.47%;(2)载钴活性焦对As(V)最佳吸附pH值为4.0,当As(V)的初始浓度为10 mg/L时,As(V)去除率可达97%;(3)吸附过程符合准二级动力学模型(k2=0.66,R2=0.96),吸附等温线为Freundlich型(kF=8.227,1/n=0.396,R2=0.97);(4)稳定性实验验证了载钴活性焦的稳定性,钴不易脱附,最大脱附率仅为0.145%。     

Modeling of Carbon Monoxide Hot Spots

This paper describes the results of a measurement and modeling study of carbon monoxide (CO) concentrations In the proximity of intersections. Analysis for model performance of paired observed and predicted CO concentrations are presented. Two methodologies of pollutant prediction were used: the Intersection Midblock Model (IMM) and a statistical multiple linear regression. The results showed that both methods underpredicted frequently and dispensed results that were site specific. In addition, correlations of IMM predicted concentrations to observed concentrations were poor (typically r² values <0.25). Various explanations for this observation are proposed. The statistical approach exhibited an improved accuracy over that of IMM. However, some of the independent variables used might be difficult to obtain as a routine measurement, and use of a one or two independent parameter model yielded adjusted R² values comparable to the r² values observed with IMM. Based on these results, an Intersection model applicable under a wide range of conditions of traffic, meteorology, and geometry is not available. Research Is needed to develop one, since its use would often be called on in the development of air quality sections of Environmental Assessments or Environmental Impact Statements.     

Study on the removal of elemental mercury from simulated flue gas by Fe<Subscript>2</Subscript>O<Subscript>3</Subscript>-CeO<Subscript>2</Subscript>/AC at low temperature

Fe₂O₃ and CeO₂ modified activated coke (AC) synthesized by the equivalent-volume impregnation were employed to remove elemental mercury (Hg⁰) from simulated flue gas at a low temperature. Effects of the mass ratio of Fe₂O₃ and CeO₂, reaction temperature, and individual flue gas components including O₂, NO, SO₂, and H₂O (g) on Hg⁰ removal efficiency of impregnated AC were investigated. The samples were characterized by Brunauer–Emmett–Teller (BET), X-ray diffraction (XRD), scanning electron microscopy (SEM), and X-ray photoelectron spectroscopy (XPS). Results showed that with optimal mass percentage of 3 % Fe₂O₃ and 3 % CeO₂ on Fe3Ce3/AC, the Hg⁰ removal efficiency could reach an average of 88.29 % at 110 °C. Besides, it was observed that O₂ and NO exhibited a promotional effect on Hg⁰ removal, H₂O (g) exerted a suppressive effect, and SO₂ showed an insignificant inhibition without O₂ to some extent. The analysis of XPS indicated that the main species of mercury on used Fe3Ce3/AC was HgO, which implied that adsorption and catalytic oxidation were both included in Hg⁰ removal. Furthermore, the lattice oxygen, chemisorbed oxygen, and/or weakly bonded oxygen species made a contribution to Hg⁰ oxidation.     

Optimizing stabilization of waste-activated sludge using Fered-Fenton process and artificial neural network modeling (KSOFM,MLP)

Sludge management is a fundamental activity in accordance with wastewater treatment aims. Sludge stabilization is always considered as a significant step of wastewater sludge handling. There has been a progressive development observed in the approach to the novel solutions in this regard. In this research, based on own initially experimental results in lab-scale regarding Fered-Fenton processes in view of organic loading (volatile-suspended solids, VSS) removal efficiency, a combination of both methods towards proper improving of excess biological sludge stabilization was investigated. Firstly, VSS removal efficiency has been experimentally studied in lab-scale under different operational conditions taking into consideration pH [Fe²⁺]/[H₂O₂], detention time [H₂O₂], and current density parameters. Therefore, the correlations of the same parameters have been determined by utilizing Kohonen self-organizing feature maps (KSOFM). In addition, multi-layer perceptron (MLP) has been employed afterwards for a comprehensive evaluation of investigating parameters correlation and prediction aims. The findings indicated that the best proportion of iron to hydrogen peroxide and the optimum pH were 0.58 and 3.1, respectively. Furthermore, maximum retention time about 6 h with a hydrogen peroxide concentration of 1,568 mg/l and a current density of 650–750 mA results to the optimum VSS removal (efficiency equals to 81 %). The performance of KSOFM and MLP models is found to be magnificent, with correlation ranging (R) from 0.873 to 0.998 for the process simulation and prediction. Finally, it can be concluded that the Fered-Fenton reactor is a suitable efficient process to reduce considerably sludge organic load and mathematical modeling tools as artificial neural networks are impressive methods of process simulation and prediction accordingly.