Combining deterministic and statistical approaches for PM10 forecasting in Europe

Well documented adverse health effects of airborne particulate matter (PM) stimulate intensive research aimed at understanding and forecasting its behaviour. Forecasting of PM levels is commonly performed with either statistical or deterministic chemistry-transport models (CTM). In this study, we investigate advantages of combining deterministic and statistical approaches for PM₁₀ forecasting over Europe one day ahead. The proposed procedure involves statistical postprocessing of deterministic forecasts by using PM₁₀ monitoring data. A series of experiments is performed using a state-of-the-art CTM (CHIMERE) and statistical models based on linear regressions. It is found that performance of both CTM simulations and “pure” statistical models is inferior to that of the combined models. In particular, the root mean squared error of the deterministic forecasts can be reduced, on the average, by up to 45 percent (specifically, from 12.8 to 6.9 μg/m³ at urban sites in summer) and the coefficient of determination can be almost doubled. Importantly, it is found that the combined models for rural sites in summer and for urban and suburban sites in both summer and winter are representative, on the average, not only for a given monitoring site used for their training, but also of territories of similar type of environment (rural, suburban or urban) within several hundreds of kilometers away.     

Accuracy and reliability of an automated air quality forecast system for ozone in seven Kentucky metropolitan areas

An automated forecast system for ozone in seven Kentucky metropolitan areas has been operational since 2004. The forecast system automatically downloads the required input data twice each day, produces next-day forecasts of metro area peak 8-h average ozone concentration using a computer coded hybrid nonlinear regression (NLR) model, and posts the results on a website. The automated models were similar to previous NLR models, first applied to forecasting ozone in the Louisville metro area. The forecast system operated reliably during the 2004 and 2005 O₃ seasons, producing at least one forecast per day better than 99% of the time. The forecast accuracy of the automated system was good. For all 2004 and 2005 forecasts, the mean absolute error was equal to 8.7 ppb, or 15.6% of the overall mean concentration. The overall detection rate of air quality standard exceedences was 56%, and the overall false alarm rate was 42%. In Louisville, the performance of the automated system was comparable to that of expert forecasters using the NLR model as a forecast tool.     

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.     

Neural networks and periodic components used in air quality forecasting

Forecasting of air quality parameters is one topic of air quality research today due to the health effects caused by airborne pollutants in urban areas. The work presented here aims at comparing two principally different neural network methods that have been considered as potential tools in that area and assessing them in relation to regression with periodic components. Self-organizing maps (SOM) represent a form of competitive learning in which a neural network learns the structure of the data. Multi-layer perceptrons (MLPs) have been shown to be able to learn complex relationships between input and output variables. In addition, the effect of removing periodic components is evaluated with respect to neural networks. The methods were evaluated using hourly time series of NO₂ and basic meteorological variables collected in the city of Stockholm in 1994–1998. The estimated values for forecasting were calculated in three ways: using the periodic components alone, applying neural network methods to the residual values after removing the periodic components, and applying only neural networks to the original data. The results showed that the best forecast estimates can be achieved by directly applying a MLP network to the original data, and thus, that a combination of the periodic regression method and neural algorithms does not give any advantage over a direct application of neural algorithms.     

Operational ozone forecasts for the region of Copenhagen by the Danish Meteorological Institute

The Danish Meteorological Institute (DMI) has developed an operational forecasting system for ozone concentrations in the Atmospheric Boundary Layer; this system is called the Danish Atmospheric Chemistry FOrecasting System (DACFOS). At specific sites where real-time ozone concentration measurements are available, a statistical after-treatment of DACFOS’ results adjusts the next 48 h ozone forecasts. This post-processing of DACFOS’ forecasts is based on an adaptive linear regression model using an optimal state estimator algorithm. The regression analysis uses different linear combinations of meteorological parameters (such as temperature, wind speed, surface heat flux and atmospheric boundary layer height) supplied by the Numerical Weather Prediction model DMI-HIRLAM. Several regressions have been tested for six monitoring stations in Denmark and in England, and four of the linear combinations have been selected to be employed in an automatic forecasting system. A statistical study comparing observations and forecasts shows that this system yields higher correlation coefficients as well as smaller biases and RMSE values than DACFOS; the present post-processing thus improves DACFOS’ forecasts. This system has been operational since June 1998 at the DMI's monitoring station in the north of Copenhagen, for which a new ozone forecast is presented every 6 h on the DMI's internet public homepage.     

Climatology and forecast modeling of ambient carbon monoxide in Phoenix, Arizona

We perform a climatology of factors influencing ambient carbon monoxide (CO), in which we examine the relationships between meteorology, traffic patterns, and CO at seasonal, weekly, and diurnal time scales in Phoenix, Arizona. From this analysis we identify a range of potentially important variables for statistical CO modeling. Using stepwise multivariate regression, we create a suite of models for hourly and 8-h ambient CO designed for daily operational forecasting purposes. The resulting models include variables and interaction terms related to anticipated nocturnal atmospheric stability as well as antecedent and climatological CO behavior. The models are evaluated using a range of error statistics and skill measures. The most successful approach employs a two-stage modeling strategy in which an initial prediction is made that may, depending on the forecast value, be followed by a second prediction that improves upon the first. The best models provide accurate daily forecasts of CO, with explained variances approaching 0.9 and errors under 1 ppm.     

A semiparametric statistical approach for forecasting SO2 and NO x concentrations

This study develops a new semiparametric statistical approach for urban air quality forecasting. Compared to conventional approaches, the semiparametric approach allows the model users to benefit from the positive aspects and alleviate the negative ones of parametric and nonparametric approaches. Two advantages of the approach lie in (1) the interpretation of the data set being easily decoded and used by the model and (2) its capability in dependence on prior assumption. To illustrate the performance of the proposed approach, three semiparametric regression models (i.e., linear-, quadratic-, and interactive-based semiparametric regression) are applied to an air quality forecasting problem in the city of Xiamen, China, and satisfactory training and prediction performance are obtained. The three models are also compared to three parametric and two nonparametric regression models. The results indicate that the predictive accuracy of semiparametric regression models is higher than those obtained from the parametric and stepwise cluster analysis models. However, the proposed three semiparametric regression models could be much favored, since they can be achieved more easily and rapidly than the artificial neural network model.     

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.     

Neural network and multiple regression models for PM10 prediction in Athens: a comparative assessment

Particulate atmospheric pollution in urban areas is considered to have significant impact on human health. Therefore, the ability to make accurate predictions of particulate ambient concentrations is important to improve public awareness and air quality management. This study examines the possibility of using neural network methods as tools for daily average particulate matter with aerodynamic diameter <10 microm (PM10) concentration forecasting, providing an alternative to statistical models widely used up to this day. Based on a data inventory, in a fixed central site in Athens, Greece, ranging over a two-year period, and using mainly meteorological variables as inputs, neural network models and multiple linear regression models were developed and evaluated. Comparison statistics used indicate that the neural network approach has an edge over regression models, expressed both in terms of prediction error (root mean square error values lower by 8.2-9.4%) and of episodic prediction ability (false alarm rate values lower by 7-13%). The results demonstrate that artificial neural networks (ANNs), if properly trained and formed, can provide adequate solutions to particulate pollution prognostic demands.     

Neural Network and Multiple Regression Models for PM10 Prediction in Athens: A Comparative Assessment

Abstract

Particulate atmospheric pollution in urban areas is considered to have significant impact on human health. Therefore, the ability to make accurate predictions of particulate ambient concentrations is important to improve public awareness and air quality management. This study examines the possibility of using neural network methods as tools for daily average particulate matter with aerodynamic diameter <10 µm (PM₁₀) concentration forecasting, providing an alternative to statistical models widely used up to this day. Based on a data inventory, in a fixed central site in Athens, Greece, ranging over a two-year period, and using mainly meteorological variables as inputs, neural network models and multiple linear regression models were developed and evaluated. Comparison statistics used indicate that the neural network approach has an edge over regression models, expressed both in terms of prediction error (root mean square error values lower by 8.2–9.4%) and of episodic prediction ability (false alarm rate values lower by 7–13%). The results demonstrate that artificial neural networks (ANNs), if properly trained and formed, can provide adequate solutions to particulate pollution prognostic demands.     

Comparison of artificial neural network and regression models in the prediction of urban stormwater quality.

Urban stormwater quality is influenced by many interrelated processes. However, the site-specific nature of these complex processes makes stormwater quality difficult to predict using physically based process models. This has resulted in the need for more empirical techniques. In this study, artificial neural networks (ANN) were used to model urban stormwater quality. A total of 5 different constituents were analyzed-chemical oxygen demand, lead, suspended solids, total Kjeldahl nitrogen, and total phosphorus. Input variables were selected using stepwise linear regression models, calibrated on logarithmically transformed data. Artificial neural networks models were then developed and compared with the regression models. The results from the analyses indicate that multiple linear regression models were more applicable for predicting urban stormwater quality than ANN models.     

Using continuous PM2.5 monitoring data to report an air quality index

As stated in 40 CFR 58, Appendix G (2000), statistical linear regression models can be applied to relate PM2.5 continuous monitoring (CM) measurements with federal reference method (FRM) measurements, collocated or otherwise, for the purpose of reporting the air quality index (AQI). The CM measurements can then be transformed via the model to remove any bias relative to FRM measurements. The resulting FRM-like modeled measurements may be used to provide more timely reporting of a metropolitan statistical area's (MSA's) AQI. Of considerable importance is the quality of the model used to relate the CM and FRM measurements. The use of a poor model could result in misleading AQI reporting in the form of incorrectly claiming either good or bad air quality. This paper describes a measure of adequacy for deciding whether a statistical linear regression model that relates FRM and continuous PM2.5 measurements is sufficient for use in AQI reporting. The approach is the U.S. Environmental Protection Agency's (EPA's) data quality objectives (DQO) process, a seven-step strategic planning approach to determine the most appropriate data type, quality, quantity, and synthesis for a given activity. The chosen measure of model adequacy is r2, the square of the correlation coefficient between FRM measurements and their modeled counterparts. The paper concludes by developing regression models that meet this desired level of adequacy for the MSAs of Greensboro/Winston-Salem/High Point, NC; and Davenport/Moline/Rock Island, IA/IL. In both cases, a log transformation of the data appeared most appropriate. For the data from the Greensboro/Winston-Salem/High Point MSA, a simple linear regression model of the FRM and CM measurements had an r2 of 0.96, based on 227 paired observations. For the data from the Davenport/Moline/Rock Island MSA, due to seasonal differences between CM and FRM measurements, the simple linear regression model had to be expanded to include a temperature dependency, resulting in an r2 of 0.86, based on 214 paired observations.     

Using Continuous PM2.5 Monitoring Data to Report an Air Quality Index

Abstract

As stated in 40 CFR 58, Appendix G (2000), statistical linear regression models can be applied to relate PM_2.5 continuous monitoring (CM) measurements with federal reference method (FRM) measurements, collocated or otherwise, for the purpose of reporting the air quality index (AQI). The CM measurements can then be transformed via the model to remove any bias relative to FRM measurements. The resulting FRM-like modeled measurements may be used to provide more timely reporting of a metropolitan statistical area’s (MSA’s) AQI.¹ Of considerable importance is the quality of the model used to relate the CM and FRM measurements. The use of a poor model could result in misleading AQI reporting in the form of incorrectly claiming either good or bad air quality.

This paper describes a measure of adequacy for deciding whether a statistical linear regression model that relates FRM and continuous PM_2.5 measurements is sufficient for use in AQI reporting. The approach is the U.S. Environmental Protection Agency’s (EPA’s) data quality objectives (DQO) process, a seven-step strategic planning approach to determine the most appropriate data type, quality, quantity, and synthesis for a given activity.² The chosen measure of model adequacy is r², the square of the correlation coefficient between FRM measurements and their modeled counterparts. The paper concludes by developing regression models that meet this desired level of adequacy for the MSAs of Greensboro/Winston-Salem/High Point, NC; and Davenport/Moline/Rock Island, IA/IL. In both cases, a log transformation of the data appeared most appropriate. For the data from the Greens-boro/Winston-Salem/High Point MSA, a simple linear regression model of the FRM and CM measurements had an r² of 0.96, based on 227 paired observations. For the data from the Davenport/Moline/Rock Island MSA, due to seasonal differences between CM and FRM measurements, the simple linear regression model had to be expanded to include a temperature dependency, resulting in an r² of 0.86, based on 214 paired observations.     

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.     

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.     

湖泊富营养化模型的研究进展

湖泊的富营养化是全球普遍关注的环境问题之一.湖泊的富营养化模型是防治、修复和治理湖泊富营养化的重要决策工具.按研究的侧重点不同,将湖泊富营养化模型分为简单回归模型、水质模型、生态模型和生态-水动力水质模型,并分别回顾了四类模型的研究进展.最后指出湖泊富营养化模型的发展趋势,强调不确定理论、3S技术、耦合模型是今后湖泊富营养化模型研究的重点,应在此基础上建立通用的模拟、预测、评价和优化模型,为湖泊富营养化管理提供科学依据.     

湖泊富营养化模型的研究进展

湖泊的富营养化是全球普遍关注的环境问题之一.湖泊的富营养化模型是防治、修复和治理湖泊富营养化的重要决策工具.按研究的侧重点不同,将湖泊富营养化模型分为简单回归模型、水质模型、生态模型和生态-水动力水质模型,并分别回顾了四类模型的研究进展.最后指出湖泊富营养化模型的发展趋势,强调不确定理论、3S技术、耦合模型是今后湖泊富营养化模型研究的重点,应在此基础上建立通用的模拟、预测、评价和优化模型,为湖泊富营养化管理提供科学依据.     

A Pilot Indoor-Outdoor Study of Organic Particulate Matter and Particulate Mutagenicity

Abstract

Many large metropolitan areas experience elevated concentrations of ground-level ozone pollution during the summertime “smog season”. Local environmental or health agencies often need to make daily air pollution forecasts for public advisories and for input into decisions regarding abatement measures and air quality management. Such forecasts are usually based on statistical relationships between weather conditions and ambient air pollution concentrations. Multivariate linear regression models have been widely used for this purpose, and well-specified regressions can provide reasonable results. However, pollution-weather relationships are typically complex and nonlinear—especially for ozone—properties that might be better captured by neural networks. This study investigates the potential for using neural networks to forecast ozone pollution, as compared to traditional regression models. Multiple regression models and neural networks are examined for a range of cities under different climate and ozone regimes, enabling a comparative study of the two approaches. Model comparison statistics indicate that neural network techniques are somewhat (but not dramatically) better than regression models for daily ozone prediction, and that all types of models are sensitive to different weather-ozone regimes and the role of persistence in aiding predictions.     

The use of wind fields in a land use regression model to predict air pollution concentrations for health exposure studies

A methodology is developed to include wind flow effects in land use regression (LUR) models for predicting nitrogen dioxide (NO₂) concentrations for health exposure studies. NO₂ is widely used in health studies as an indicator of traffic-generated air pollution in urban areas. Incorporation of high-resolution interpolated observed wind direction from a network of 38 weather stations in a LUR model improved NO₂ concentration estimates in densely populated, high traffic and industrial/business areas in Toronto-Hamilton urban airshed (THUA) of Ontario, Canada. These small-area variations in air pollution concentrations that are probably more important for health exposure studies may not be detected by sparse continuous air pollution monitoring network or conventional interpolation methods. Observed wind fields were also compared with wind fields generated by Global Environmental Multiscale-High resolution Model Application Project (GEM-HiMAP) to explore the feasibility of using regional weather forecasting model simulated wind fields in LUR models when observed data are either sparse or not available. While GEM-HiMAP predicted wind fields well at large scales, it was unable to resolve wind flow patterns at smaller scales. These results suggest caution and careful evaluation of regional weather forecasting model simulated wind fields before incorporating into human exposure models for health studies. This study has demonstrated that wind fields may be integrated into the land use regression framework. Such integration has a discernable influence on both the overall model prediction and perhaps more importantly for health effects assessment on the relative spatial distribution of traffic pollution throughout the THUA. Methodology developed in this study may be applied in other large urban areas across the world.