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 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.     

Carbon monoxide concentration forecasting in Santiago, Chile

In the city of Santiago, Chile, air quality is defined in terms of particulate matter with an aerodynamic diameter < or = 10 microm (PM10) concentrations. An air quality forecasting model based on past concentrations of PM10 and meteorological conditions currently is used by the metropolitan agency for the environment, which allows restrictions to emissions to be imposed in advance. This model, however, fails to forecast between 40 and 50% of the days considered to be harmful for the inhabitants every year. Given that a high correlation between particulate matter and carbon monoxide (CO) concentrations is observed at monitoring stations in the city, a model for CO concentration forecasting would be a useful tool to complement information about expected air quality in the city. Here, the results of a neural network-based model aimed to forecast maximum values of the 8-hr moving average of CO concentrations for the next day are presented. Forecasts from the neural network model are compared with those produced with linear regressions. The neural network model seems to leave more room to adjust free parameters with 1-yr data to predict the following year's values. We have worked with 3 yr of data measured at the monitoring station located in the zone with the worst air quality in the city of Santiago, Chile.     

Estimating national-scale ground-level PM25 concentration in China using geographically weighted regression based on MODIS and MISR AOD

Taking advantage of the continuous spatial coverage, satellite-derived aerosol optical depth (AOD) products have been widely used to assess the spatial and temporal characteristics of fine particulate matter (PM2.5) on the ground and their effects on human health. However, the national-scale ground-level PM2.5 estimation is still very limited because the lack of ground PM2.5 measurements to calibrate the model in China. In this study, a national-scale geographically weighted regression (GWR) model was developed to estimate ground-level PM2.5 concentration based on satellite AODs, newly released national-wide hourly PM2.5 concentrations, and meteorological parameters. The results showed good agreements between satellite-retrieved and ground-observed PM2.5 concentration at 943 stations in China. The overall cross-validation (CV) R ² is 0.76 and root mean squared prediction error (RMSE) is 22.26 μg/m³ for MODIS-derived AOD. The MISR-derived AOD also exhibits comparable performance with a CV R ² and RMSE are 0.81 and 27.46 μg/m³, respectively. Annual PM2.5 concentrations retrieved either by MODIS or MISR AOD indicated that most of the residential community areas exceeded the new annual Chinese PM2.5 National Standard level 2. These results suggest that this approach is useful for estimating large-scale ground-level PM2.5 distributions especially for the regions without PMs monitoring sites.     

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.     

The Steubenville comprehensive air monitoring program (SCAMP): associations among fine particulate matter, co-pollutants, and meteorological conditions

We determined 24-hr average ambient concentrations of PM2.5 and its ionic and carbonaceous components in Steubenville, OH, between May 2000 and May 2002. We also determined daily average gaseous co-pollutant concentrations, meteorological conditions, and pollen and mold spore counts. Data were analyzed graphically and by linear regression and time series models. Multiple-day episodes of elevated fine particulate matter (PM2.5) concentrations often occurred during periods of locally high temperature (especially during summer), high pressure, or low wind speed (especially during winter) and generally ended with the passage of a frontal system. After removing autocorrelation, we observed statistically significant positive associations between concentrations of PM2.5 and concentrations of CO, NOx, and SO2. Associations with NOx and CO exhibited significant seasonal dependencies, with the strongest correlations during fall and winter. NOx, CO, SO2, O3, temperature, relative humidity, and wind speed were all significant predictors of PM2.5 concentration in a time-series model with external regressors, which successfully accounted for 79% of the variance in log-transformed daily PM2.5 concentrations. Coefficient estimates for NOx and temperature varied significantly by season. The results provide insight that may be useful in the development of future PM2.5 reduction strategies for Steubenville. Additionally, they demonstrate the need for PM epidemiology studies in Steubenville (and elsewhere) to carefully consider the potential confounding effects of gaseous co-pollutants, such as CO and NOx, and their seasonally dependent associations with PM2.5.     

Analysis of Temporal and Spatial Dichotomous PM Air Samples in the El Paso-Cd. Juarez Air Quality Basin

ABSTRACT

This paper presents and discusses the results obtained from the gravimetric and chemical analyses of the 24-hr average dichotomous samples collected from five sites in the El Paso-Cd. Juarez air quality basin between August 1999 and March 2000. Gravimetric analysis was performed to determine the temporal and spatial variations of PM_2.5 (particulate matter less than 2.5 um in diameter) and PM2.5-10 (particulate matter less than 10 μm but greater than 2.5 μm in diameter) mass concentrations. The results indicate that ~25% of the PM10 (i.e., PM2.5 + PM2.5-10) concentration is composed of PM2.5. Concurrent measurements of hourly PM concentrations and wind speed showed strong diurnal patterns of the regional PM pollution. Results of X-ray fluorescence (XRF) elemental analyses were compared to similar but limited studies performed by the Texas Natural Resource Conservation Commission (TNRCC) in 1990 and 1997. Major elements from geologic sources—Al, Si, Ca, Na, K, Fe, and Ti—accounted for 35% of the total mass concentrations in the PM2.5-10 fraction, indicating that geologic sources in the area are the dominant PM sources. Levels of toxic trace elements, mainly considered as products of anthropogenic activities, have decreased significantly from those observed in 1990 and 1997.     

Analysis of PM10, PM2.5, and PM2 5-10 concentrations in Santiago, Chile, from 1989 to 2001

Daily particle samples were collected in Santiago, Chile, at four urban locations from January 1, 1989, through December 31, 2001. Both fine PM with da < 2.5 microm (PM2.5) and coarse PM with 2.5 < da < 10 microm (PM2.5-10) were collected using dichotomous samplers. The inhalable particle fraction, PM10, was determined as the sum of fine and coarse concentrations. Wind speed, temperature and relative humidity (RH) were also measured continuously. Average concentrations of PM2.5 for the 1989-2001 period ranged from 38.5 microg/m3 to 53 microg/m3. For PM2.5-10 levels ranged from 35.8-48.2 microg/m3 and for PM10 results were 74.4-101.2 microg/m3 across the four sites. Both annual and daily PM2.5 and PM10 concentration levels exceeded the U.S. National Ambient Air Quality Standards and the European Union concentration limits. Mean PM2.5 levels during the cold season (April through September) were more than twice as high as those observed in the warm season (October through March); whereas coarse particle levels were similar in both seasons. PM concentration trends were investigated using regression models, controlling for site, weekday, month, wind speed, temperature, and RH. Results showed that PM2.5 concentrations decreased substantially, 52% over the 12-year period (1989-2000), whereas PM2.5-10 concentrations increased by approximately 50% in the first 5 years and then decreased by a similar percentage over the following 7 years. These decreases were evident even after controlling for significant climatic effects. These results suggest that the pollution reduction programs developed and implemented by the Comisión Nacional del Medio Ambiente (CONAMA) have been effective in reducing particle levels in the Santiago Metropolitan region. However, particle levels remain high and it is thus imperative that efforts to improve air quality continue.     

An enhanced PM2.5 air quality forecast model based on nonlinear regression and back-trajectory concentrations

An enhanced PM_2.5 air quality forecast model based on nonlinear regression (NLR) and back-trajectory concentrations has been developed for use in the Louisville, Kentucky metropolitan area. The PM_2.5 air quality forecast model is designed for use in the warm season, from May through September, when PM_2.5 air quality is more likely to be critical for human health. The enhanced PM_2.5 model consists of a basic NLR model, developed for use with an automated air quality forecast system, and an additional parameter based on upwind PM_2.5 concentration, called PM24. The PM24 parameter is designed to be determined manually, by synthesizing backward air trajectory and regional air quality information to compute 24-h back-trajectory concentrations. The PM24 parameter may be used by air quality forecasters to adjust the forecast provided by the automated forecast system. In this study of the 2007 and 2008 forecast seasons, the enhanced model performed well using forecasted meteorological data and PM24 as input. The enhanced PM_2.5 model was compared with three alternative models, including the basic NLR model, the basic NLR model with a persistence parameter added, and the NLR model with persistence and PM24. The two models that included PM24 were of comparable accuracy. The two models incorporating back-trajectory concentrations had lower mean absolute errors and higher rates of detecting unhealthy PM2.5 concentrations compared to the other models.     

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.     

Analysis of temporal and spatial dichotomous PM air samples in the El Paso-Cd. Juarez air quality basin.

This paper presents and discusses the results obtained from the gravimetric and chemical analyses of the 24-hr average dichotomous samples collected from five sites in the El Paso-Cd. Juarez air quality basin between August 1999 and March 2000. Gravimetric analysis was performed to determine the temporal and spatial variations of PM2.5 (particulate matter less than 2.5 microm in diameter) and PM25-10 (particulate matter less than 10 pm but greater than 2.5 microm in diameter) mass concentrations. The results indicate that approximately 25% of the PM10 (i.e., PM25 + PM25-10) concentration is composed of PM2.5. Concurrent measurements of hourly PM concentrations and wind speed showed strong diurnal patterns of the regional PM pollution. Results of X-ray fluorescence (XRF) elemental analyses were compared to similar but limited studies performed by the Texas Natural Resource Conservation Commission (TNRCC) in 1990 and 1997. Major elements from geologic sources-Al, Si, Ca, Na, K, Fe, and Ti-accounted for 35% of the total mass concentrations in the PM2.5-10 fraction, indicating that geologic sources in the area are the dominant PM sources. Levels of toxic trace elements, mainly considered as products of anthropogenic activities, have decreased significantly from those observed in 1990 and 1997.     

Near-road air pollutant concentrations of CO and PM2.5: A comparison of MOBILE6.2/CALINE4 and generalized additive models

The contribution of vehicular traffic to air pollutant concentrations is often difficult to establish. This paper utilizes both time-series and simulation models to estimate vehicle contributions to pollutant levels near roadways. The time-series model used generalized additive models (GAMs) and fitted pollutant observations to traffic counts and meteorological variables. A one year period (2004) was analyzed on a seasonal basis using hourly measurements of carbon monoxide (CO) and particulate matter less than 2.5 μm in diameter (PM_2.5) monitored near a major highway in Detroit, Michigan, along with hourly traffic counts and local meteorological data. Traffic counts showed statistically significant and approximately linear relationships with CO concentrations in fall, and piecewise linear relationships in spring, summer and winter. The same period was simulated using emission and dispersion models (Motor Vehicle Emissions Factor Model/MOBILE6.2; California Line Source Dispersion Model/CALINE4). CO emissions derived from the GAM were similar, on average, to those estimated by MOBILE6.2. The same analyses for PM_2.5 showed that GAM emission estimates were much higher (by 4–5 times) than the dispersion model results, and that the traffic-PM_2.5 relationship varied seasonally. This analysis suggests that the simulation model performed reasonably well for CO, but it significantly underestimated PM_2.5 concentrations, a likely result of underestimating PM_2.5 emission factors. Comparisons between statistical and simulation models can help identify model deficiencies and improve estimates of vehicle emissions and near-road air quality.     

Two-days ahead prediction of daily maximum concentrations of SO2, O3, PM10, NO2, CO in the urban area of Palermo,Italy

Artificial neural networks are functional alternative techniques in modelling the intricate vehicular exhaust emission dispersion phenomenon. Pollutant predictions are notoriously complex when using either deterministic or stochastic models, which explains why this model was developed using a neural network. Neural networks have the ability to learn about non-linear relationships between the used variables. In this paper a recurrent neural network (Elman model) based forecaster for the prediction of daily maximum concentrations of SO₂, O₃, PM10, NO₂, CO in the city of Palermo is proposed. The effectiveness of the presented forecaster was tested using a time series recorded between 1 January 2003 to 31 December 2004 in eight monitoring stations in urban area of Palermo (Italy). Experimental trials show that the developed and tuned model is appropriate, giving small values of root mean square error (RMSE) , mean absolute error (MAE) and mean square error (MSE). In addition, the related correlation coefficient ranges from 0.72 to 0.97 for each forecasted pollutant, underlying a small difference between the forecasted and the measured values. The above results make the proposed forecaster a powerful tool for pollution management systems.     

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.     

The Steubenville Comprehensive Air Monitoring Program (SCAMP): analysis of short-term and episodic variations in PM2.5 concentrations using hourly air monitoring data

One-hour average ambient concentrations of particulate matter (PM) with an aerodynamic diameter < 2.5 microm (PM2.5) were determined in Steubenville, OH, between June 2000 and May 2002 with a tapered element oscillating microbalance (TEOM). Hourly average gaseous copollutant [carbon monoxide (CO), sulfur dioxide (SO2), nitrogen oxide (NOx), and ozone (O3)] concentrations and meteorological conditions also were measured. Although 75% of the 14,682 hourly PM2.5 concentrations measured during this period were < or = 17 microg/m3, concentrations > 65 microg/m3 were observed 76 times. On average, PM2.5 concentrations at Steubenville exhibited a diurnal pattern of higher early morning concentrations and lower afternoon concentrations, similar to the diurnal profiles of CO and NO(x). This pattern was highly variable; however, PM2.5 concentrations > 65 microg/m3 were never observed during the mid-afternoon between 1:00 p.m. and 5:00 p.m. EST. Twenty-two episodes centered on one or more of these elevated concentrations were identified. Five episodes occurred during the months June through August; the maximum PM2.5 concentration during these episodes was 76.6 microg/m3. Episodes occurring during climatologically cooler months often featured higher peak concentrations (five had maximum concentrations between 95.0 and 139.6 microg/m3), and many exhibited strong covariation between PM2.5 and CO, NO(x), or SO2. Case studies suggested that nocturnal surface-based temperature inversions were influential in driving high nighttime concentrations of these species during several cool season episodes, which typically had dramatically lower afternoon concentrations. These findings provide insights that may be useful in the development of PM2.5 reduction strategies for Steubenville, and suggest that studies assessing possible health effects of PM2.5 should carefully consider exposure issues related to the intraday timing of PM2.5 episodes, as well as the potential for toxicological interactions among PM2.5, and primary gaseous pollutants.     

Air quality at Santiago,Chile: a box modeling approach II. PM2.5, coarse and PM10 particulate matter fractions

Ambient monitored data at Santiago, Chile, are analyzed using box models with the goal of assessing contributions of different economic activities to air pollution levels. The box modeling approach was applied to PM₁₀, PM_2.5 and coarse (PM₁₀–PM_2.5) particulate matter (PM) fractions; the period analyzed is 1989–1999. A linear model for each PM fraction was obtained, having as independent variables CO and SO₂ concentrations, plus a term proportional to (wind speed)⁻¹ that lumps together non-combustion emissions and secondary generation terms; wet scavenging is included as another independent variable. Model identification results show good agreement for the different parameters across monitoring stations. The washout ratios and scavenging coefficients agree with data published in the literature, being higher for the coarse PM fraction. The CO and SO₂ coefficients fitted for 1989–1995 agree with a priori estimates for the same period. Background estimates for the PM fractions are in agreement with measurement campaigns in upwind sites. Results show that transportation sources have become the dominant contributors to ambient PM levels, while stationary sources have decreased their contributions in the last years. The relative importance of mobile sources to PM_2.5 ambient concentrations has doubled in the last 10 years, whereas stationary sources have reduced their relative contributions to half the value in the early 1990s. Model estimates of regional background of PM_2.5 and PM₁₀ have decreased 50% and 22% in the last decade, respectively; coarse background has shown no significant change. The final conclusion is that there is room and need for a more intensive emission reduction strategy for Santiago, focusing on mobile sources. The approach pursued in this work is feasible for cities or regions where comprehensive, transport and chemistry models are not available yet, but estimates of air quality contributions are needed for policy purposes. The methodology requires data on ambient air quality measurements and surface meteorology.     

Air quality measurements from the Fresno Supersite

The Fresno Supersite intends to 1) evaluate non-routine monitoring methods, establishing their comparability with existing methods and their applicability to air quality planning, exposure assessment, and health effects studies; 2) provide a better understanding of aerosol characteristics, behavior, and sources to assist regulatory agencies in developing standards and strategies that protect public health; and 3) support studies that evaluate relationships between aerosol properties, co-factors, and observed health end-points. Supersite observables include in-situ, continuous, short-duration measurements of 1) PM2.5, PM10, and coarse (PM10 minus PM2.5) mass; 2) PM2.5 SO4(-2), NO3-, carbon, light absorption, and light extinction; 3) numbers of particles in discrete size bins ranging from 0.01 to approximately 10 microns; 4) criteria pollutant gases (O3, CO, NOx); 5) reactive gases (NO2, NOy, HNO3, peroxyacetyl nitrate [PAN], NH3); and 6) single particle characterization by time-of-flight mass spectrometry. Field sampling and laboratory analysis are applied for gaseous and particulate organic compounds (light hydrocarbons, heavy hydrocarbons, carbonyls, polycyclic aromatic hydrocarbons [PAH], and other semi-volatiles), and PM2.5 mass, elements, ions, and carbon. Observables common to other Supersites are 1) daily PM2.5 24-hr average mass with Federal Reference Method (FRM) samplers; 2) continuous hourly and 5-min average PM2.5 and PM10 mass with beta attenuation monitors (BAM) and tapered element oscillating microbalances (TEOM); 3) PM2.5 chemical speciation with a U.S. Environmental Protection Agency (EPA) speciation monitor and protocol; 4) coarse particle mass by dichotomous sampler and difference between PM10 and PM2.5 BAM and TEOM measurements; 5) coarse particle chemical composition; and 6) high sensitivity and time resolution scalar and vector wind speed, wind direction, temperature, relative humidity, barometric pressure, and solar radiation. The Fresno Supersite is coordinated with health and toxicological studies that will use these data in establishing relationships with asthma, other respiratory disease, and cardiovascular changes in human and animal subjects.     

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.     

Estimation of ambient PM2.5 concentrations in Maryland and verification by measured values

In 1997, Maryland had no available ambient Federal Reference Method data on particulate matter less than 2.5 microm in aerodynamic diameter (PM23), but did have annual ambient data for PM smaller than 10 microm (PM10) at 24 sites. The PM10 data were analyzed in conjunction with local annual and seasonal zip-code-level emission inventories and with speciated PM2.5 data from four nearby monitors in the IMPROVE network (located in the national parks, wildlife refuges, and wilderness areas) in an effort to estimate annual average and seasonal high PM2.5 concentrations at the 24 PM10 monitor sites operating from 1992 to 1996. All seasonal high concentrations were estimated to be below the 24-hr PM2.5 National Ambient Air Quality Standards (NAAQS) at the sites operating in Maryland between 1992 and 1996. The estimates also indicated that 12 monitor sites might exceed the 3-year annual average PM2.5 NAAQS of 15 microg/m3, but Maryland's air quality shows signs that it has been improving since 1992. The estimates also were compared with actual measurements after the PM2.5 monitor network was installed. The estimates were adequate for describing the chemical composition of the PM2.5, forecasting compliance status with the 24-hr and annual standards, and determining the spatial variations in PM2.5 across central Maryland.     

Model for estimation of traffic pollutant levels in northern communities

Using models to estimate the contribution of traffic to air pollution levels from known traffic data typically requires the knowledge of model parameters such as emission factors and meteorological conditions. This paper presents a state-space model analysis method that does not require the knowledge of model parameters; these parameters are identified from measured traffic and ambient air quality data. This method was used to analyze carbon monoxide (CO) in downtown Fairbanks, AK, which is the community of focus for this paper. It was found that traffic contributed, on average, 53% to the total CO levels over the last six winters. The correlation coefficient between the measured and model-predicted daily profiles of the CO concentration was 0.98, and the results were in good agreement with earlier findings obtained via a thorough CO emission inventory. This justified the usability of the method and it was further used to analyze fine particulate matter (PM2.5) in downtown Fairbanks. It was found that traffic contributed, on average, approximately 30% to the total PM2.5 levels over the last six winters. The correlation coefficient between the measured and model-predicted daily profiles of the PM2.5 concentration was 0.98.