SAGA: a decision support system for air pollution management around a coal-fired power plant

Air quality models are currently feasible approaches to prevent air pollution episodes. From one of the first source-oriented modelling approaches for air pollution forecasting (Souto et al., 1994, 1996, 1998), a new decision support system for air quality management, SAGA, was developed to provide support to As Pontes Power Plant (APPP) staff. SAGA can provide air pollution forecasts and manage meteorological and air quality measurements. Power plant decisions are supported by the results of a non-hydrostatic meteorological model (ARPS, Xue et al., 2001) to produce Meteorological Forecasts (MFs), and to be coupled to different Lagrangian dispersion models.     

Forecasting daily pollen concentrations using data-driven modeling methods in Thessaloniki,Greece

Airborne pollen have been associated with allergic symptoms in sensitized individuals, having a direct impact on the overall quality of life of a considerable fraction of the population. Therefore, forecasting elevated airborne pollen concentrations and communicating this piece of information to the public are key issues in prophylaxis and safeguarding the quality of life of the overall population. In this study, we adopt a data-oriented approach in order to develop operational forecasting models (1–7 days ahead) of daily average airborne pollen concentrations of the highly allergenic taxa: Poaceae, Oleaceae and Urticaceae. The models are developed using a representative dataset consisting of pollen and meteorological time-series recorded during the years 1987–2002, in the city of Thessaloniki, Greece. The input variables (features) of the models have been optimized by making use of genetic algorithms, whereas we evaluate the performance of three algorithms: i) multi-Layer Perceptron, ii) support vector regression and iii) regression trees originating from distinct domains of Computational Intelligence (CI), and compare the resulting models with traditional multiple linear regression models. Results show the superiority of CI methods, especially when forecasting several days ahead, compared to traditional multiple linear regression models. Furthermore, the CI models complement each other, resulting to a combined model that performs better than each one separately. The overall performance ranges, in terms of the index of agreement, from 0.85 to 0.93 clearly suggesting the potential operational use of the models. The latter ones can be utilized in provision of personalized and on-time information services, which can improve quality of life of sensitized citizens.     

Deep learning architecture for air quality predictions

With the rapid development of urbanization and industrialization, many developing countries are suffering from heavy air pollution. Governments and citizens have expressed increasing concern regarding air pollution because it affects human health and sustainable development worldwide. Current air quality prediction methods mainly use shallow models; however, these methods produce unsatisfactory results, which inspired us to investigate methods of predicting air quality based on deep architecture models. In this paper, a novel spatiotemporal deep learning (STDL)-based air quality prediction method that inherently considers spatial and temporal correlations is proposed. A stacked autoencoder (SAE) model is used to extract inherent air quality features, and it is trained in a greedy layer-wise manner. Compared with traditional time series prediction models, our model can predict the air quality of all stations simultaneously and shows the temporal stability in all seasons. Moreover, a comparison with the spatiotemporal artificial neural network (STANN), auto regression moving average (ARMA), and support vector regression (SVR) models demonstrates that the proposed method of performing air quality predictions has a superior performance.     

Potential assessment of the "support vector machine" method in forecasting ambient air pollutant trends

Monitoring and forecasting of air quality parameters are popular and important topics of atmospheric and environmental research today due to the health impact caused by exposing to air pollutants existing in urban air. The accurate models for air pollutant prediction are needed because such models would allow forecasting and diagnosing potential compliance or non-compliance in both short- and long-term aspects. Artificial neural networks (ANN) are regarded as reliable and cost-effective method to achieve such tasks and have produced some promising results to date. Although ANN has addressed more attentions to environmental researchers, its inherent drawbacks, e.g., local minima, over-fitting training, poor generalization performance, determination of the appropriate network architecture, etc., impede the practical application of ANN. Support vector machine (SVM), a novel type of learning machine based on statistical learning theory, can be used for regression and time series prediction and have been reported to perform well by some promising results. The work presented in this paper aims to examine the feasibility of applying SVM to predict air pollutant levels in advancing time series based on the monitored air pollutant database in Hong Kong downtown area. At the same time, the functional characteristics of SVM are investigated in the study. The experimental comparisons between the SVM model and the classical radial basis function (RBF) network demonstrate that the SVM is superior to the conventional RBF network in predicting air quality parameters with different time series and of better generalization performance than the RBF model.     

A comparison of statistical methods for deriving freshwater quality criteria for the protection of aquatic organisms

Species sensitivity distributions (SSDs) are increasingly used in both ecological risk assessment and derivation of water quality criteria. However, there has been debate about the choice of an appropriate approach for derivation of water quality criteria based on SSDs because the various methods can generate different values. The objective of this study was to compare the differences among various methods. Data sets of acute toxicities of 12 substances to aquatic organisms, representing a range of classes with different modes of action, were studied. Nine typical statistical approaches, including parametric and nonparametric methods, were used to construct SSDs for 12 chemicals. Water quality criteria, expressed as hazardous concentration for 5 % of species (HC₅), were derived by use of several approaches. All approaches produced comparable results, and the data generated by the different approaches were significantly correlated. Variability among estimates of HC₅ of all inclusive species decreased with increasing sample size, and variability was similar among the statistical methods applied. Of the statistical methods selected, the bootstrap method represented the best-fitting model for all chemicals, while log-triangle and Weibull were the best models among the parametric methods evaluated. The bootstrap method was the primary choice to derive water quality criteria when data points are sufficient (more than 20). If the available data are few, all other methods should be constructed, and that which best describes the distribution of the data was selected.     

Neural network forecasting of air pollutants hourly concentrations using optimised temporal averages of meteorological variables and pollutant concentrations

The new method for the forecasting hourly concentrations of air pollutants is presented in the paper. The method was developed for a site in urban residential area in city of Zagreb, Croatia, for four air pollutants (NO₂, O₃, CO and PM₁₀). Meteorological variables and concentrations of the respective pollutant were taken as predictors. A novel approach, based on families of univariate regression models, was employed in selecting the averaging intervals for input variables. For each variable and each averaging period between 1 and 97 h, a separate model was built. By inspecting values of the coefficient of correlation between measured and modelled concentrations, optimal averaging periods for each variable were selected. A new dataset for building the forecasting model was then calculated as temporal moving averages (running means) of former variables. A multi-layer perceptron type of neural networks is used as the forecasting model. Index of agreement, calculated for the entire dataset including the data for model building, ranged from 0.91 to 0.97 for the respective pollutants. As suggested by the analysis of the relative importance of the input variables, different agreements for different pollutants are likely due to different sources and production mechanisms of investigated pollutants. A comparison of the new method with more traditional method, which takes hourly averages of the forecast hour as input variables, showed similar or better performance. The model was developed for the purpose of public-health-oriented air quality forecasting, aiming to use a numerical weather forecast model for the prediction of the part of input data yet unknown at the forecasting time. It is to expect that longer term averages used as inputs in the proposed method will contribute to smaller input errors and the greater accuracy of the model.     

Model for forecasting expressway fine particulate matter and carbon monoxide concentration: application of regression and neural network models

The Borman Expressway is a heavily traveled 16-mi segment of the Interstate 80/94 freeway through Northwestern Indiana. The Lake and Porter counties through which this expressway passes are designated as particulate matter < 2.5 microm (PM2.5) and ozone 8-hr standard nonattainment areas. The Purdue University air quality group has been collecting PM2.5, carbon monoxide (CO), wind speed, wind direction, pressure, and temperature data since September 1999. In this work, regression and neural network models were developed for forecasting hourly PM2.5 and CO concentrations. Time series of PM2.5 and CO concentrations, traffic data, and meteorological parameters were used for developing the neural network and regression models. The models were compared using a number of statistical quality indicators. Both models had reasonable accuracy in predicting hourly PM2.5 concentration with coefficient of determination -0.80, root mean square error (RMSE) <4 microg/m3, and index of agreement (IA) > 0.90. For CO prediction, both models showed moderate forecasting performance with a coefficient of determination -0.55, RMSE < 0.50 ppm, and IA -0.85. These models are computationally less cumbersome and require less number of predictors as compared with the deterministic models. The availability of real time PM2.5 and CO forecasts will help highway managers to identify air pollution episodic events beforehand and to determine mitigation strategies.     

A neural network model forecasting for prediction of daily maximum ozone concentration in an industrialized urban area

Prediction of ambient ozone concentrations in urban areas would allow evaluation of such factors as compliance and noncompliance with EPA requirements. Though ozone prediction models exist, there is still a need for more accurate models. Development of these models is difficult because the meteorological variables and photochemical reactions involved in ozone formation are complex. In this study, we developed a neural network model for forecasting daily maximum ozone levels. We then compared the neural network's performance with those of two traditional statistical models, regression, and Box-Jenkins ARIMA. The neural network model for forecasting daily maximum ozone levels is different from the two statistical models because it employs a pattern recognition approach. Such an approach does not require specification of the structural form of the model. The results show that the neural network model is superior to the regression and Box-Jenkins ARIMA models we tested.     

Conjunctive Use of Models to Design Cost-Effective Ozone Control Strategies

Abstract

The management of tropospheric ozone (O₃) is particularly difficult. The formulation of emission control strategies requires considerable information including: (1) emission inventories, (2) available control technologies, (3) meteorological data for critical design episodes, and (4) computer models that simulate atmospheric transport and chemistry. The simultaneous consideration of this information during control strategy design can be exceedingly difficult for a decision-maker. Traditional management approaches do not explicitly address cost minimization. This study presents a new approach for designing air quality management strategies; a simple air quality model is used conjunctively with a complex air quality model to obtain low-cost management strategies. A simple air quality model is used to identify potentially good solutions, and two heuristic methods are used to identify cost-effective control strategies using only a small number of simple air quality model simulations. Subsequently, the resulting strategies are verified and refined using a complex air quality model. The use of this approach may greatly reduce the number of complex air quality model runs that are required. An important component of this heuristic design framework is the use of the simple air quality model as a screening and exploratory tool. To achieve similar results with the simple and complex air quality models, it may be necessary to “tweak” or calibrate the simple model. A genetic algorithm-based optimization procedure is used to automate this tweaking process. These methods are demonstrated to be computationally practical using two realistic case studies, which are based on data from a metropolitan region in the United States.     

Investigating hydrochemistry of groundwater in Indo-Gangetic alluvial plain using multivariate chemometric approaches

Groundwater hydrochemistry of an urban industrial region in Indo-Gangetic plains of north India was investigated. Groundwater samples were collected both from the industrial and non-industrial areas of Kanpur. The hydrochemical data were analyzed using various water quality indices and nonparametric statistical methods. Principal components analysis (PCA) was performed to identify the factors responsible for groundwater contamination. Ensemble learning-based decision treeboost (DTB) models were constructed to develop discriminating and regression functions to differentiate the groundwater hydrochemistry of the three different areas, to identify the responsible factors, and to predict the groundwater quality using selected measured variables. The results indicated non-normal distribution and wide variability of water quality variables in all the study areas, suggesting for nonhomogenous distribution of sources in the region. PCA results showed contaminants of industrial origin dominating in the region. DBT classification model identified pH, redox potential, total-Cr, and λ ₂₅₄ as the discriminating variables in water quality of the three areas with the average accuracy of 99.51 % in complete data. The regression model predicted the groundwater chemical oxygen demand values exhibiting high correlation with measured values (0.962 in training; 0.918 in test) and the respective low root mean-squared error of 2.24 and 2.01 in training and test arrays. The statistical and chemometric approaches used here suggest that groundwater hydrochemistry differs in the three areas and is dominated by different variables. The proposed methods can be used as effective tools in groundwater management.     

A new methodology development for the regulatory forecasting of PM10. Application in the Greater Athens Area,Greece

The paper introduces a new methodology for the prediction of daily PM10 concentrations, in line with the regulatory framework introduced through the EU Directive 2008/50/EC. The proposed approach is based on the efficient utilisation of the data collected over short time intervals (hourly) rather than the daily values used to derive the daily regulatory threshold. It is sufficiently simple and easily applicable in operational forecasting systems with the ability to accept as inputs both historical data and exogenous paraeters, such as meteorological variables. The application of the proposed methodology is demonstrated using data from five monitoring stations of air pollutants located in Athens, over a five year period (2000–2004) as well as compatible meteorological data from the NCEP (National Centers for Environmental Protection). A set of different models have been tested at the same time to reveal the effectiveness of the proposed approach, both univariate and multivariate, and linear and non-linear models. The analysis of all examined datasets has shown conclusive evidence that the introduction of the newly developed procedure which utilises data collected over a shorter horizon can significantly increase the forecasting ability of any developed model using daily historic PM10 data, under all examined metrics.     

Ozone predictions in Atlanta, Georgia: analysis of the 1999 ozone season

Ozone prediction has become an important activity in many U.S. ozone nonattainment areas. In this study, we describe the ozone prediction program in the Atlanta metropolitan area and analyze the performance of this program during the 1999 ozone-forecasting season. From May to September, a team of 10 air quality regulators, meteorologists, and atmospheric scientists made a daily prediction of the next-day maximum 8-hr average ozone concentration. The daily forecast was made aided by two linear regression models, a 3-dimensional air quality model, and the no-skill ozone persistence model. The team's performance is compared with the numerical models using several numerical indicators. Our analysis indicated that (1) the team correctly predicted next-day peak ozone concentrations 84% of the time, (2) the two linear regression models had a better performance than a 3-dimensional air quality model, (3) persistence was a strong predictor of ozone concentrations with a performance of 78%, and (4) about half of the team's wrong predictions could be prevented with improved meteorological predictions.     

Ozone Predictions in Atlanta,Georgia: Analysis of the 1999 Ozone Season

ABSTRACT

Ozone prediction has become an important activity in many U.S. ozone nonattainment areas. In this study, we describe the ozone prediction program in the Atlanta metropolitan area and analyze the performance of this program during the 1999 ozone-forecasting season. From May to September, a team of 10 air quality regulators, meteorologists, and atmospheric scientists made a daily prediction of the next-day maximum 8-hr average ozone concentration. The daily forecast was made aided by two linear regression models, a 3-dimensional air quality model, and the no-skill ozone persistence model. The team's performance is compared with the numerical models using several numerical indicators. Our analysis indicated that (1) the team correctly predicted next-day peak ozone concentrations 84% of the time, (2) the two linear regression models had a better performance than a 3-dimensional air quality model, (3) persistence was a strong predictor of ozone concentrations with a performance of 78%, and (4) about half of the team's wrong predictions could be prevented with improved meteorological predictions.     

Forecasting hourly PM<Subscript>10</Subscript> concentration in Cyprus through artificial neural networks and multiple regression models: implications to local environmental management

In the present work, two types of artificial neural network (NN) models using the multilayer perceptron (MLP) and the radial basis function (RBF) techniques, as well as a model based on principal component regression analysis (PCRA), are employed to forecast hourly PM₁₀ concentrations in four urban areas (Larnaca, Limassol, Nicosia and Paphos) in Cyprus. The model development is based on a variety of meteorological and pollutant parameters corresponding to the 2-year period between July 2006 and June 2008, and the model evaluation is achieved through the use of a series of well-established evaluation instruments and methodologies. The evaluation reveals that the MLP NN models display the best forecasting performance with R ² values ranging between 0.65 and 0.76, whereas the RBF NNs and the PCRA models reveal a rather weak performance with R ² values between 0.37-0.43 and 0.33-0.38, respectively. The derived MLP models are also used to forecast Saharan dust episodes with remarkable success (probability of detection ranging between 0.68 and 0.71). On the whole, the analysis shows that the models introduced here could provide local authorities with reliable and precise predictions and alarms about air quality if used on an operational basis.     

Evaluation of artificial neural networks for fine particulate pollution (PM10 and PM2.5) forecasting

Multi-layer perceptron (MLP) artificial neural network (ANN) models are compared with traditional multiple regression (MLR) models for daily maximum and average O3 and particulate matter (PM10 and PM2.5) forecasting. MLP particulate forecasting models show little if any improvement over MLR models and exhibit less skill than do O3 forecasting models. Meteorological variables (precipitation, wind, and temperature), persistence, and co-pollutant data are shown to be useful PM predictors. If MLP approaches are adopted for PM forecasting, training methods that improve extreme value prediction are recommended.     

Support vector machine―an alternative to artificial neuron network for water quality forecasting in an agricultural nonpoint source polluted river?

Water quality forecasting in agricultural drainage river basins is difficult because of the complicated nonpoint source (NPS) pollution transport processes and river self-purification processes involved in highly nonlinear problems. Artificial neural network (ANN) and support vector model (SVM) were developed to predict total nitrogen (TN) and total phosphorus (TP) concentrations for any location of the river polluted by agricultural NPS pollution in eastern China. River flow, water temperature, flow travel time, rainfall, dissolved oxygen, and upstream TN or TP concentrations were selected as initial inputs of the two models. Monthly, bimonthly, and trimonthly datasets were selected to train the two models, respectively, and the same monthly dataset which had not been used for training was chosen to test the models in order to compare their generalization performance. Trial and error analysis and genetic algorisms (GA) were employed to optimize the parameters of ANN and SVM models, respectively. The results indicated that the proposed SVM models performed better generalization ability due to avoiding the occurrence of overtraining and optimizing fewer parameters based on structural risk minimization (SRM) principle. Furthermore, both TN and TP SVM models trained by trimonthly datasets achieved greater forecasting accuracy than corresponding ANN models. Thus, SVM models will be a powerful alternative method because it is an efficient and economic tool to accurately predict water quality with low risk. The sensitivity analyses of two models indicated that decreasing upstream input concentrations during the dry season and NPS emission along the reach during average or flood season should be an effective way to improve Changle River water quality. If the necessary water quality and hydrology data and even trimonthly data are available, the SVM methodology developed here can easily be applied to other NPS-polluted rivers.     

Evaluation of Episode Control Schemes Through Air Quality Data Analysis

Prediction performance of various air pollution episode models are first compared with that of a persistence model which is based on the assumption that present concentrations persist to a future time. The comparisons are made by computing a correlation coefficient for different lead times between the observed and predicted values, and an auto-correlation function of the air quality data to which the episode model is applied. The persistence of high levels of air pollution is next examined, using existing air quality data, by constructing frequency distributions of air pollution episode duration for various concentration thresholds. Based on the results of persistence analysis, the flaws of currently used episode management schemes are discussed and some alternative episode management schemes are presented. Methodologies and parameters to evaluate the anticipated performances of episode management schemes are developed and some examples are worked out. In conclusion, it is suggested that a combination of episode persistence analysis and air pollution meteorological forecasting could lead to a workable air pollution episode management scheme.     

Evaluation of land-use regression models used to predict air quality concentrations in an urban area

Cohort studies designed to estimate human health effects of exposures to urban pollutants require accurate determination of ambient concentrations in order to minimize exposure misclassification errors. However, it is often difficult to collect concentration information at each study subject location. In the absence of complete subject-specific measurements, land-use regression (LUR) models have frequently been used for estimating individual levels of exposures to ambient air pollution. The LUR models, however, have several limitations mainly dealing with extensive monitoring data needs and challenges involved in their broader applicability to other locations. In contrast, air quality models can provide high-resolution source–concentration linkages for multiple pollutants, but require detailed emissions and meteorological information. In this study, first we predicted air quality concentrations of PM_2.5, NO_x, and benzene in New Haven, CT using hybrid modeling techniques based on CMAQ and AERMOD model results. Next, we used these values as pseudo-observations to develop and evaluate the different LUR models built using alternative numbers of (training) sites (ranging from 25 to 285 locations out of the total 318 receptors). We then evaluated the fitted LUR models using various approaches, including: 1) internal “Leave-One-Out-Cross-Validation” (LOOCV) procedure within the “training” sites selected; and 2) “Hold-Out” evaluation procedure, where we set aside 33–293 tests sites as independent datasets for external model evaluation. LUR models appeared to perform well in the training datasets. However, when these LUR models were tested against independent hold out (test) datasets, their performance diminished considerably. Our results confirm the challenges facing the LUR community in attempting to fit empirical response surfaces to spatially- and temporally-varying pollution levels using LUR techniques that are site dependent. These results also illustrate the potential benefits of enhancing basic LUR models by utilizing air quality modeling tools or concepts in order to improve their reliability or transferability.     

Forecasts using neural network versus Box-Jenkins methodology for ambient air quality monitoring data

This study explores ambient air quality forecasts using the conventional time-series approach and a neural network. Sulfur dioxide and ozone monitoring data collected from two background stations and an industrial station are used. Various learning methods and varied numbers of hidden layer processing units of the neural network model are tested. Results obtained from the time-series and neural network models are discussed and compared on the basis of their performance for 1-step-ahead and 24-step-ahead forecasts. Although both models perform well for 1-step-ahead prediction, some neural network results reveal a slightly better forecast without manually adjusting model parameters, according to the results. For a 24-step-ahead forecast, most neural network results are as good as or superior to those of the time-series model. With the advantages of self-learning, self-adaptation, and parallel processing, the neural network approach is a promising technique for developing an automated short-term ambient air quality forecast system.