Search

Machine learning methods are increasingly being used in the field of air quality research to investigate the relationship between air pollutant levels, emissions, and meteorological changes over time. This research is used for both scientific investigation, and policy assessment and development. However, there is a lack of studies that have compared the performance of different machine learning methods. To address this gap, this paper employed various machine learning techniques, including decision tree, random forest (RF), support vector machine (SVM), support vector regression (SVR), k-nearest neighbor, neural network, and Gaussian process regression, to predict daily average PM 2.5 levels and the number of days with PM 2.5 exceedance in the South Coast Air Basin of California from 2000 to 2019. The models were trained using meteorological factors, estimated emissions, and large-scale climate indices as inputs. The SVR model demonstrated the highest predictive accuracy for PM 2.5 levels and the SVM model gave the most accurate results for predicting the number of days with PM 2.5 exceedances. Conversely, the decision tree model performed the least accurately. The results also showed that emissions have a greater impact on PM 2.5 levels over time compared to meteorological factors, though meteorology is responsible for daily variability. The most important meteorological factors were identified as surface relative humidity and relative humidity at 850 mbars, which are related to partitioning, cloud cover and wet deposition. We conducted sensitivity tests on the model's response to emissions and meteorological factors. The predicted PM 2.5 from RF and SVR showed large correlations with emissions at the early period (2000–2010). However, the changes were minimal in more recent years (2011–2019), implying that there are biases in machine learning models, in which the models consistently predict the minimum PM 2.5 levels at a baseline. [Display omitted] • SVR is more accurate on daily PM 2.5 , particularly PM 2.5 exceedances predictions. • Surface RH was the most important meteorological factor for PM 2.5 prediction. • The impact of emissions on PM 2.5 was significant before 2010 but reduced thereafter. • ML models predict past better than future; all the ML models are limited at extremes. [ABSTRACT FROM AUTHOR]

This study characterized nitrogen oxides (NOx) and ammonia (NH 3) emissions from five in-use goods movement vehicles, including one diesel vehicle without selective catalytic reduction (SCR), two diesel vehicles with SCR, and two ultra-low NOx natural gas vehicles equipped with three-way catalysts (TWCs). Emissions testing was performed under real-world driving conditions in typical goods movement routes in the South Coast Air Basin (SCAB) of California. NOx emissions varied depending on the vehicle and the route. The no-SCR diesel vehicle showed the highest NOx emissions over all the routes. One of the SCR-equipped vehicles showed NOx emission rates that were two times higher than the other SCR-equipped vehicle with similar engine model year and mileage, which may be attributed to catalyst deterioration. The SCR-equipped diesel vehicles were within two to three times 0.2 g/bhp-hr certification standard over the different routes. The natural gas vehicles showed average NOx emissions around or below the optional Low NOx emissions standard of 0.02 g/bhp-hr. The three-bin moving average window (MAW) method was utilized to show the NOx emissions across various modes of operation, including idle, low load, and medium/high load. The highest fraction of vehicle operation was found in either the low load or medium/high load bins, depending on the route. NOx emissions for each of the bins varied between different vehicles, and are further analyzed in this paper. NH 3 emissions formation was favored for the TWC-equipped natural gas vehicles, which produced about 20 times more NH 3 emissions than the SCR-equipped diesel vehicles. [Display omitted] • Significant NOx reductions for the natural gas vehicles compared to the SCR-equipped diesel vehicles. • NH 3 formation favored in the TWC for the natural gas vehicles and produced during acceleration events. • Low load and medium/high load conditions dominated vehicle operation and NOx emissions. [ABSTRACT FROM AUTHOR]

The Volkswagen Group of America (VW) was found by the US Environmental Protection Agency (EPA) and the California Air Resources Board (CARB) to have installed “defeat devices” and emit more oxides of nitrogen (NOx) than permitted under current EPA standards. In this paper, we quantify the hidden NOx emissions from this so-called VW scandal and the resulting public health impacts in California. The NOx emissions are calculated based on VW road test data and the CARB Emission Factors (EMFAC) model. Cumulative hidden NOx emissions from 2009 to 2015 were estimated to be over 3500 tons. Adult mortality changes were estimated based on ambient fine particulate matter (PM 2.5 ) change due to secondary nitrate formation and the related concentration-response functions. We estimated that hidden NOx emissions from 2009 to 2015 have resulted in a total of 12 PM 2.5 -associated adult mortality increases in California. Most of the mortality increase happened in metropolitan areas, due to their high population and vehicle density. [ABSTRACT FROM AUTHOR]

Hydrogen has been proposed as a low polluting alternative transportation fuel that could help improve urban air quality. This paper examines the potential impact of introducing a hydrogen-based transportation system on urban ambient ozone concentrations. This paper considers two scenarios, where significant numbers of new hydrogen vehicles are added to a constant number of gasoline vehicles. In our scenarios hydrogen fuel cell vehicles (HFCVs) are introduced in Sacramento, California at market penetrations of 9% and 20%. From a life cycle analysis (LCA) perspective, considering all the emissions involved in producing, transporting, and using hydrogen, this research compares three hypothetical natural gas to hydrogen pathways: (1) on-site hydrogen production; (2) central hydrogen production with pipeline delivery; and (3) central hydrogen production with liquid hydrogen truck delivery. Using a regression model, this research shows that the daily maximum temperature correlates well with atmospheric ozone formation. However, increases in initial VOC and NO x concentrations do not necessarily increase the peak ozone concentration, and may even cause it to decrease. It is found that ozone formation is generally limited by NO x in the summer and is mostly limited by VOC in the fall in Sacramento. Of the three hydrogen pathways, the truck delivery pathway contributes the most to ozone precursor emissions. Ozone precursor emissions from the truck pathway at 9% market penetration can cause additional 3-h average VOC (or NO x ) concentrations up to approximately 0.05% (or 1%) of current pollution levels, and at 20% market penetration up to approximately 0.1% (or 2%) of current pollution levels. However, all of the hydrogen pathways would result in very small (either negative or positive) changes in ozone air quality. In some cases they will result in worse ozone air quality (mostly in July, August, and September), and in some cases they will result in better ozone air quality (mostly in October). The truck pathway tends to cause a much wider fluctuation in degradation or improvement of ozone air quality: percentage changes in peak ozone concentrations are approximately −0.01% to 0.04% for the assumed 9% market penetration, and approximately −0.03% to 0.1% for the 20% market penetration. Moreover, the 20% on-site pathway occasionally results in a decrease of about −0.1% of baseline ozone pollution. Compared to the current ambient pollution level, all three hydrogen pathways are unlikely to cause a serious ozone problem for market penetration levels of HFCVs in the 9–20% range. [Copyright &y& Elsevier]

The fine particle (PM 2.5 ) air pollution monitoring network in California underwent major changes between 1988 and 2013, adding continuous beta attenuation monitors (BAM) to the network. This paper analyzes temporal patterns using PM 2.5 mass measurements from three major air basins. Regressions of BAM against filter measurements generally show BAM reads higher by 3–6 μg/m 3 for annual averages. Most monitoring sites show clear downward trends in PM 2.5 concentrations over time, especially in the Los Angeles area where concentrations have decreased by more than 50% since the 1990s. In most cases, PM 2.5 concentrations are elevated in winter compared to summer, with basin-wide average increases of 80 ± 25% for the San Francisco Bay area and 123 ± 28% for the San Joaquin Valley. The differences are more prominent at night due to wood-burning and stagnant atmospheric conditions. The reverse is true at inland sites in southern California, which show average reductions of 46 ± 12% in winter. Weekend concentrations are slightly lower than on weekdays. Average weekend reductions in PM 2.5 relative to mid-week values are 8.9 ± 5.9% for the San Francisco Bay area and 8.0 ± 3.0% for the San Joaquin Valley. Up to 25% reductions in PM 2.5 are observed on weekends at traffic-dominated sites such as West Oakland. [ABSTRACT FROM AUTHOR]

Abstract: Land use regression (LUR) models are widely employed in health studies to characterize chronic exposure to air pollution. The LUR is essentially an interpolation technique that employs the pollutant of interest as the dependent variable with proximate land use, traffic, and physical environmental variables used as independent predictors. Two major limitations with this method have not been addressed: (1) variable selection in the model building process, and (2) dealing with unbalanced repeated measures. In this paper, we address these issues with a modeling framework that implements the deletion/substitution/addition (DSA) machine learning algorithm that uses a generalized linear model to average over unbalanced temporal observations. Models were derived for fine particulate matter with aerodynamic diameter of 2.5 microns or less (PM2.5) and nitrogen dioxide (NO2) using monthly observations. We used 4119 observations at 108 sites and 15,301 observations at 138 sites for PM2.5 and NO2, respectively. We derived models with good predictive capacity (cross-validated-R 2 values were 0.65 and 0.71 for PM2.5 and NO2, respectively). By addressing these two shortcomings in current approaches to LUR modeling, we have developed a framework that minimizes arbitrary decisions during the model selection process. We have also demonstrated how to integrate temporally unbalanced data in a theoretically sound manner. These developments could have widespread applicability for future LUR modeling efforts. [Copyright &y& Elsevier]

Abstract: A number of recent toxicity studies have highlighted the increased potency of oxygen analogs (oxons) of several organophosphorus (OP) pesticides. These findings were a major concern after environmental oxons were identified in environmental samples from air and surfaces following agricultural spray applications in California and Washington State. This paper reports on the validity of oxygen analog measurements in air samples for the OP pesticide, chlorpyrifos. Controlled environmental and laboratory experiments were used to examine artificial formation of chlorpyrifos-oxon using OSHA Versatile Sampling (OVS) tubes as recommended by NIOSH method 5600. Additionally, we compared expected chlorpyrifos-oxon attributable to artificial transformation to observed chlorpyrifos-oxon in field samples from a 2008 Washington State Department of Health air monitoring study using non-parametric statistical methods. The amount of artificially transformed oxon was then modeled to determine the amount of oxon present in the environment. Toxicity equivalency factors (TEFs) for chlorpyrifos-oxon were used to calculate chlorpyrifos-equivalent air concentrations. The results demonstrate that the NIOSH-recommended sampling matrix (OVS tubes with XAD-2 resin) was found to artificially transform up to 30% of chlorpyrifos to chlorpyrifos-oxon, with higher percentages at lower concentrations (<30 ng m−3) typical of ambient or residential levels. Overall, the 2008 study data had significantly greater oxon than expected by artificial transformation, but the exact amount of environmental oxon in air remains difficult to quantify with the current sampling method. Failure to conduct laboratory analysis for chlorpyrifos-oxon may result in underestimation of total pesticide concentration when using XAD-2 resin matrices for occupational or residential sampling. Alternative methods that can accurately measure both OP pesticides and their oxygen analogs should be used for air sampling, and a toxicity equivalent factor approach should be used to determine potential health risks from exposures. [Copyright &y& Elsevier]

Abstract: The growth of commercial aviation has fueled concerns over air quality around airports and the surrounding communities. Airports must expand their operations to meet the increase in air traffic, but expansion plans have been delayed or canceled due to concerns over local air quality. This paper presents the methodology for real-time measurements of aircraft engine specific Particulate Matter (PM) emissions and analysis of the associated high resolution data acquired during normal Landing and Take-Off (LTO) operations 100–300 m downwind of an active taxi-/runway at the Oakland International Airport. The airframe–engine combinations studied included B737-300 with CFM56-3B engines, B737-700/800 with CFM56-7B engines, A320 with V2500-A5 engines, MD-80 with JT-8D engines, A300 with CF6-80 engines, DC-10 with CF6-50 engines, and CRJ-100/200 with CF34-3B engines. For all engine types studied, the size distributions were typically bimodal in nature with a nucleation mode comprised of freshly nucleated PM and an accumulation mode comprised mostly of PM soot with some condensed volatile material. The PM number-based emission index observed ranged between 7 × 1015–3 × 1017 particles kg−1 fuel burned at idle/taxi and between 4 × 1015–2 × 1017 particles kg−1 fuel burned at take-off, and the associated PM mass-based emission index (EIm) ranged between 0.1 and 0.7 g kg−1 fuel burned at both the idle/taxi and take-off conditions. Older technology engines such as the CFM56-3B and JT8D engines were observed to have as much as 3× higher PM EIm values at take-off compared to newer engine technology such as the CFM56-7B engine. The results from this study provide information for better characterizing evolving PM emissions from in-service commercial aircraft under normal LTO operations and assessing their impact on local and regional air quality and health related impacts. [Copyright &y& Elsevier]

Abstract: This paper investigates snow albedo changes in the Sierra Nevada Mountain area associated with potential deposition of absorbing aerosols in spring by using the snow albedo, aerosol optical depth, land surface temperature, and other relevant parameters available from the Moderate-Resolution Imaging Spectroradiometer (MODIS) onboard the NASA Terra satellite during 2000–2009. Satellite pixels with 100% snow cover have been selected to derive the monthly mean snow albedo value, along with aerosol optical depth, surface temperature, and days after snowfall in March and April to perform multiple regression analysis. We show that aerosol optical depth, which generally includes dust and black carbon over the Sierra Nevada as a result of the transpacific transport from East Asia and local sources, represents a significant parameter affecting snow albedo variation, second only to the land surface temperature change. The regression analysis illustrates that a one standard deviation increase in land surface temperature (2.2 K) and aerosol optical depth (0.044) can lead to decreases in snow albedo by 0.038 and 0.026, respectively. This study also shows that approximately 26% of snow albedo reduction from March to April over the Sierra Nevada is caused by an increase in aerosol optical depth, which has a profound impact on available water resources in California. However, the results show that there are no significant trends for snow albedo, surface temperature, and aerosol optical depth of snow-covered areas over the Sierra Nevada Mountain area in this 10-year period. [Copyright &y& Elsevier]

Abstract: We can control asthma through proper clinical and environmental management and education. The U.S. population is growing, urbanizing and aging; seniors of low-to-middle income families are working and living longer. We conducted community-based participatory research in Visalia, Tulare County, California with a prospective, cross-sectional repeated measures design and quantitative and qualitative process; home environment and health-related outcomes data were collected. In this paper, we presented results of the air quality sampling—pollen, carbon dioxide (CO2) and particulate matter (PM) outdoors away from most major sources (agricultural fields, large pollinating trees, etc)—at a community central site close to a mobile line source and participant homes in the cooling season, July, 2009. Weather was hot and dry with light winds; diurnal variation ranged between 65–107°F (18–42°C) and 12–76% relative humidity at the study’s central site. Co-located active (reference) and passive (PAAS) samplers were used for pollen; passive monitoring for CO2 (Telaire 7001) and active sampling for PM were conducted. Overall, we observed spatial variability in CO2, fine PM (PM2.5), and pollen counts. Weekday and study week average CO2 and PM2.5 concentrations were higher near study homes compared to central site sampling points, but peak measures and overnight/pre-dawn time period averages were elevated at the central site. Pollen counts were typically lower at homes—even if grass, trees, flowers or potted plants were present—compared to the central site closer to and generally downwind from commercial agricultural tree production. Data are new; the nine-county San Joaquin Valley has one pollen count station in the national network, and two of four government outdoor air monitoring stations in the county are in national parks. We suggest—given poor air quality in large part due to PM—adding routine pollen counts to regional/state agency air monitoring sites and more CO2 and PM monitoring. [Copyright &y& Elsevier]

Abstract: This paper presents an analysis of the aerosol chemical composition, optical properties and size distributions for a range of conditions encountered during a field measurement campaign conducted between July 7–29, 2005 at Point Reyes National Seashore, north of San Francisco, CA. The fractional mass loading derived from hourly measurements of an Aerodyne Mass Spectrometer (AMS) during this period are compared with filter-pack measurements from the Pt. Reyes IMPROVE station with good agreement found between the two if it assumed that chloride is primarily from large sea-salt particles (not measured by the AMS). During the first half of the campaign (July 7–17), conditions at the site were largely maritime while flow during the second half of the campaigns (July 18–29) was influenced by a thermal trough that added a cyclonic twist to the incoming marine air, bringing it from the south with a more extensive over-land trajectory. Neither flow regime was associated with air coming from the San Francisco Bay area to the south. The AMS measurements are partitioned into clear and foggy conditions which are then used to calculate the equivalent molar ratio of ammonium to the sum of sulfate, nitrate and chloride. Ratios calculated from measurements made before the onset of the thermal trough on July 18th were associated with acidic or near-neutral particles. Measurements made after July 18th yield ratios that appear to have excess ammonium. Model calculations of the equilibrium gas-phase mixing ratio of NH3 suggest very high values which we attribute to agricultural practices within the park. Reported as an incidental finding is evidence for the cloud droplet activation of large particles (DP > 0.2 μm) with a corresponding reduction in the single scattering albedo of the non-activated particles, followed by a return in the particle size spectrum to the pre-fog conditions immediately afterwards. [Copyright &y& Elsevier]

Abstract: Oxygenated additives in gasoline are designed to decrease the ozone-forming hydrocarbons and total air toxics, yet they can increase the emissions of aldehydes and thus increase human exposure to these toxic compounds. This paper describes a study conducted to characterize targeted aldehydes in microenvironments in Sacramento, CA, and Milwaukee, WI, and to improve our understanding of the impact of the urban environment on human exposure to air toxics. Data were obtained from microenvironmental concentration measurements, integrated, 24-h personal measurements, indoor and outdoor pollutant monitors at the participants'' residences, from ambient pollutant monitors at fixed-site locations in each city, and from real-time diaries and questionnaires completed by the technicians and participants. As part of this study, a model to predict personal exposures based on individual time/activity data was developed for comparison to measured concentrations. Predicted concentrations were generally within 25% of the measured concentrations. The microenvironments that people encounter daily provide for widely varying exposures to aldehydes. The activities that occur in those microenvironments can modulate the aldehyde concentrations dramatically, especially for environments such as “indoor at home.” By considering personal activity, location (microenvironment), duration in the microenvironment, and a knowledge of the general concentrations of aldehydes in the various microenvironments, a simple model can do a reasonably good job of predicting the time-averaged personal exposures to aldehydes, even in the absence of monitoring data. Although concentrations of aldehydes measured indoors at the participants'' homes tracked well with personal exposure, there were instances where personal exposures and indoor concentrations differed significantly. Key to the ability to predict exposure based on time/activity data is the quality and completeness of the microenvironmental characterizations for the chemicals of interest. Consistent with many earlier studies, personal exposures are difficult to predict using data from regional outdoor monitors. [Copyright &y& Elsevier]

Abstract: This paper is concerned with the spatiotemporal mapping of monthly 8-h average ozone (O 3) concentrations over California during a 15-years period. The basic methodology of our analysis is based on the spatiotemporal random field (S/TRF) theory. We use a S/TRF decomposition model with a dominant seasonal O 3 component that may change significantly from site to site. O 3 seasonal patterns are estimated and separated from stochastic fluctuations. By means of Bayesian Maximum Entropy (BME) analysis, physically meaningful and sufficiently detailed space–time maps of the seasonal O 3 patterns are generated across space and time. During the summer and winter months the seasonal O 3 concentration maps exhibit clear and progressively changing geographical patterns over time, suggesting the existence of relationships in accordance with the typical physiographic and climatologic features of California. BME mapping accuracy can be superior to that of other techniques commonly used by EPA; its framework can rigorously assimilate useful data sources that were previously unaccounted for; the generated maps offer valuable assessments of the spatiotemporal O 3 patterns that can be helpful in the identification of physical mechanisms and their interrelations, the design of human exposure and population health models, and in risk assessment. As they focus on the seasonal patterns, the maps are not contingent on short-time and locally prevalent weather conditions, which are of no interest in a global and non-forecasting framework. Moreover, the maps offer valuable insight about the space–time O 3 concentration patterns and are, thus, helpful for disentangling the influence of explanatory factors or even for identifying some influential ones that could have been otherwise overlooked. [Copyright &y& Elsevier]

A comprehensive air quality modeling project was carried out to simulate size and composition resolved airborne particulate matter concentrations in northern and central California using the pollutant concentration and meteorological data collected during the California Regional PM10/PM2.5 Air Quality Study (CRPAQS) from December 15, 2000 to January 7, 2001. Measured 24-h average PM2.5 concentrations during this time period exceeded 180μgm−3 at Bakersfield, making it the most severe PM2.5 air quality episode ever recorded in the United States with a rigorous measurement database to support modeling. In this paper, the UCD/CIT source-oriented air quality model is used to predict the concentrations of O3, NO, NO2, CO, elemental carbon (EC), organic compounds (OC), nitrate and PM2.5 mass concentration over a 24-day period using a horizontal resolution of 4km×4km to cover all of central California. This is the first extensive evaluation of an air quality model in central California using the fine spatial resolution appropriate for the mountain–valley topography of the region combined with the relatively long multi-week time scales associated with winter stagnation events. Fractional bias (FB) values were calculated at all sites on each day of the study to quantify model performance. The CO (FB=−0.5 to +0.3), O3 (FB=−0.5 to +0.25), NO (FB=−0.9 to −0.1) and NO2 (FB=0 to +0.4) concentrations predicted by the UCD/CIT model are in general agreement with observations at most monitoring stations throughout the Valley. The predicted PM2.5 concentrations (FB=−0.5 to +0.75) generally agree with observations at Bethel Island, Sacramento, Fresno and Bakersfield spanning the entire length of the model domain. PM2.5 concentrations are over-predicted at the remote monitoring site Angiola in the central portion of the domain. Part of the over-prediction is due to excess fugitive dust emissions. CO, NO, EC and OC were all under-predicted at Angiola, indicating possible missing combustion sources in the emission inventory. The regional nitrate (FB=−1.5 to +1.25) formation dynamics were correctly reproduced by the model simulation but imperfect wind fields cause differences between the predicted vs. measured spatial distribution of nitrate during the last several days of simulation leading to the broader range of fraction bias. Overall, the results of the current study confirm the ability of the air quality model to capture the major features of a severe particulate air pollution event in northern and central California providing a foundation for future studies on source apportionment and emissions control. [Copyright &y& Elsevier]

Emission control technologies designed to meet the 2007 and 2010 emission standards for heavy-duty diesel vehicles (HDDV) remove effectively the non-volatile fraction of particles, but are comparatively less efficient at controlling the semi-volatile components. A collaborative study between the California Air Resources Board (CARB) and the University of Southern California was initiated to investigate the physicochemical and toxicological characteristics of the semi-volatile and non-volatile particulate matter (PM) fractions from HDDV emissions. This paper reports the physical properties, including size distribution, volatility (in terms of number and mass), surface diameter, and agglomeration of particles emitted from HDDV retrofitted with advanced emission control devices. Four vehicles in combination with six after-treatment devices (V-SCRT®, Z-SCRT®, CRT®, DPX, Hybrid-CCRT®, EPF) were tested under three driving cycles: steady state (cruise), transient (urban dynamometer driving schedule, UDDS), and idle. An HDDV without any control device is served as the baseline vehicle. Substantial reduction of PM mass emissions (>90%) was accomplished for the HDDV operating with advanced emission control technologies. This reduction was not observed for particle number concentrations under cruise conditions, with the exceptions of the Hybrid-CCRT® and EPF vehicles, which were efficient in controlling both—mass and number emissions. In general, significant nucleation mode particles (<50nm) were formed during cruise cycles in comparison with the UDDS cycles, which emit higher PM mass in the accumulation mode. The nucleation mode particles (<50nm) were mainly internally mixed, and evaporated considerably between 150 and 230°C. Compared to the baseline vehicle, particles from vehicles with controls (except of the Hybrid-CCRT®) had a higher mass specific surface area. [Copyright &y& Elsevier]

The results of two recent vehicle emission studies are described in this paper, along with a statistical analysis of the changes in tailpipe emissions due to the use of ethanol that includes the results from these two studies in combination with results from other literature reports. The first study evaluates the effect of two low blend ethanol gasolines (E10, E20) on tailpipe and evaporative emissions from three multi-port fuel injection vehicles and one gasoline direct injection vehicle at two different test temperatures. The second study evaluates the differences in tailpipe emissions and fuel consumptions of paired flexible fuel and conventional gasoline vehicles operating on California RFG Phase 2 and/or E85 fuels at 20°C. The vehicles were tested over the four-phase FTP or UDDS and US06 driving cycles. Tailpipe emissions were characterized for criteria pollutants (CO, NO X , NMHC, NMOG), greenhouse gases (CO2, CH4, N2O), and a suite of unregulated emissions including important air toxics (benzene, 1,3-butadiene, formaldehyde, acetaldehyde, acrolein), and ozone reactivity. In the low blend ethanol study, evaporative emissions were quantified and characterized for NMHC. While contradicting, results can be seen among the various literature reports and with these two new studies, the statistical analyses of the aggregated data offers much clearer pictures of the changes in tailpipe emissions that may be expected using either low blend ethanol gasoline (E10) or E85. The results of the statistical analysis suggest that the use of E10 results in statistically significant decreases in CO emissions (−16%); statistically significant increases in emissions of NMHC (9%), NMOG (14%), acetaldehyde (108%), 1,3-butadiene (16%), and benzene (15%); and no statistically significant changes in NO X , CO2, CH4, N2O or formaldehyde emissions. The statistical analysis suggests that the use of E85 results in statistically significant decreases in emissions of NO X (−45%), NMHC (−48%), 1,3-butadiene (−77%), and benzene (−76%); statistically significant increases in emissions of formaldehyde (73%) and acetaldehyde (2540%), and no statistically significant change in CO, CO2, and NMOG emissions. [Copyright &y& Elsevier]

Emission factors of various particle species from light- and heavy-duty vehicles (LDVs and HDVs, respectively), including organic and elemental carbon (OC and EC), sulfate, polycyclic aromatic hydrocarbons (PAHs), hopanes, steranes, trace metals, elements, and particle number (PN), were estimated based on roadway measurements. Sampling campaigns were conducted at two different roadways: the CA-110 highway (where only gasoline-powered vehicles are allowed), and the I-710 freeway (where about 20% of the total number of vehicles are diesel-powered trucks). The particulate matter (PM) emission factors determined in these roadways were compared to those reconstructed from recent source emission data from the Caldecott tunnel [Phuleria, H.C., Geller, M.D., Fine, P.M., Sioutas, C., 2006. Size-resolved emissions of organic tracers from light- and heavy-duty vehicles measured in a California roadway tunnel. Environmental Science and Technology 40 (13), 4109–4118], and those from previous tunnel and chassis dynamometer studies. Very good agreement between estimated and reconstructed emission factors was found for PN, EC, sulfate, high-molecular-weight (MW) PAHs, hopanes and steranes. This suggests that PM-speciated chemical data collected at roadsides can be used to calculate reliable emission factors for several important particle species at other locations characterized by a similar mix of on-road motor vehicles. The agreement between our results and other studies in the emission factors of trace elements and metals varied from very good (for species such as Cu, Mo, Ba, Pb) to poor (for species such as Mg, Fe, Ca), probably because the atmospheric concentrations of the latter elements are associated with both traffic and non-traffic sources, and the relative abundances of Mg, Ca, and Fe in road dust varies considerably across locations. The emission factors of OC and EC were clearly the highest for HDVs, and those of PAHs, hopanes, and steranes from our roadway measurements were well within the range of values reported in the literature from tunnel and dynamometer studies. The approach presented in this paper allows for a straight-forward estimation of PM emission factors from ambient, near-freeway measurements. Although the uncertainties inherent in the method proposed here must be acknowledged (e.g. assumptions were made to estimate the average fleet composition and the total carbon content in the vehicles’ exhaust), our results are generally in very good agreement with those in the available literature for most non-labile PM species. [Copyright &y& Elsevier]

In previous work, we showed that the intake fraction (iF) for nonreactive primary air pollutants was 20 times higher in central tendency for small-scale, urban-sited distributed electricity generation (DG) sources than for large-scale, central station (CS) power plants in California [Heath, G.A., Granvold, P.W., Hoats, A.S., Nazaroff, W.W., 2006. Intake fraction assessment of the air pollutant exposure implications of a shift toward distributed electricity generation. Atmospheric Environment 40, 7164–7177]. The present paper builds on that study, exploring pollutant- and technology-specific aspects of population inhalation exposure from electricity generation. We compare California''s existing CS-based system to one that is more reliant on DG units sited in urban areas. We use Gaussian plume modeling and a GIS-based exposure analysis to assess 25 existing CSs and 11 DG sources hypothetically located in the downtowns of California''s most populous cities. We consider population intake of three pollutants—PM2.5, NO x and formaldehyde—directly emitted by five DG technologies—natural gas (NG)-fired turbines, NG internal combustion engines (ICE), NG microturbines, diesel ICEs, and fuel cells with on-site NG reformers. We also consider intake of these pollutants from existing CS facilities, most of which use large NG turbines, as well as from hypothetical facilities located at these same sites but meeting California''s best-available control technology standards. After systematically exploring the sensitivity of iF to pollutant decay rate, the iFs for each of the three pollutants for all DG and CS cases are estimated. To efficiently compare the pollutant- and technology-specific exposure potential on an appropriate common basis, a new metric is introduced and evaluated: the intake-to-delivered-energy ratio (IDER). The IDER expresses the mass of pollutant inhaled by an exposed population owing to emissions from an electricity generation unit per quantity of electric energy delivered to the place of use. We find that the central tendency of IDER is much greater for almost every DG technology evaluated than for existing CS facilities in California. [Copyright &y& Elsevier]

This paper explores the range of CALINE4''s PM2.5 modeling capabilities by comparing previously collected PM2.5 data with CALINE4 predicted values. Two sampling sites, a suburban site located at an intersection in Sacramento, CA, and an urban site located in London, were used. Predicted concentrations are graphed against observed concentrations and evaluated against the criterion that 75% of the points fall within the factor-of-two prediction envelope. For the suburban site, data estimated by CALINE4 produced results that fell within the acceptable factor-of-two percentage envelope. A reverse dispersion test was also conducted for the suburban site using observed and calculated emission factors, and although it showed correlations between the observed values and CALINE4 predicted values, it could not conclusively prove that the model is accurate at predicting PM2.5 concentrations. Although the results suggest that CALINE4 PM2.5 predictions may be reasonably close to observed values, the number of observations used to verify the model was small and consequently, findings from the suburban site should be considered exploratory. For the urban site, a much larger data set was evaluated; however, the CALINE4 results for this site did not fall 75% within the factor-of-two envelope. Several factors, including street canyon effects, likely contributed to an inaccuracy of the emission factors used in CALINE4, and therefore, to the overall CALINE4 predictions. In summary, CALINE4 does not appear to perform well in densely populated areas and differences in topography may be a decisive factor in determining when CALINE4 may be applicable to modeling PM2.5. For critical transportation projects requiring PM2.5 analysis, use of CALINE4 may not be optimal because of its inability to produce reasonable estimates for highly trafficked areas. Additional data sets for CALINE4 analysis, particularly in urban environments, are required to fully understand CALINE4''s PM2.5 modeling capabilities. [Copyright &y& Elsevier]

A summer air quality monitoring campaign focusing on the evolution of ultrafine (<180nm in diameter) particle concentrations was conducted at an urban site in Los Angeles during June–July 2006. Previous observations suggest that ultrafine aerosol at this site are generally representative of the Los Angeles urban environment. Continuous and intermittent gas and aerosol measurements were made over 4 weeks with consistent daily meteorological conditions. Monthly averages of the data suggest the strong influence of commute traffic emissions on morning observations of ultrafine particle concentrations. By contrast, in the afternoon our measurements provide evidence of secondary photochemical reactions becoming the predominant formation mechanism of ultrafine aerosols. The ultrafine number concentration peak occurs in the early afternoon, before the maximum ozone concentration is observed. The source of this offset is unknown and requires further investigation. It is possible that the chemical mechanisms responsible for secondary organic aerosol formation evolve as atmospheric conditions change and/or secondary semi-volatile components of the aerosol re-volatilize due to the elevated peak temperatures observed (ca. 30–35°C) combined with the increased atmospheric dilution during that time. Measurements of the volatility of the ultrafine aerosol are consistent with this interpretation as overall volatility increases in the afternoon and there is less evidence of external mixing. Composition data presented in the companion paper support these conclusions [Ning et al., 2007. Daily variation in chemical characteristics of urban ultrafine aerosols and inference of their sources. Environmental Science and Technology, in press]. [Copyright &y& Elsevier]

Abstract: Recent health studies have reported that ultrafine particles (UFP) (&#60;0.1μm in diameter) may be responsible for some of the adverse health effects broadly attributed to particulate matter. In urban areas, UFP are produced by combustion sources, such as vehicle exhaust, and by secondary formation in the atmosphere. While UFP can be monitored, few studies have explored the impact of local primary sources in urban areas (including mobile sources on freeways) on the temporal and spatial distribution of UFP. This paper describes the integration of multiple monitoring technologies on a mobile platform designed to characterize UFP and associated pollutants, and the application of this platform in a study of UFP number concentrations and size distributions in Los Angeles. Monitoring technologies included two condensation particle counters (TSI Model 3007 and TSI 3022A) and scanning mobility particle sizers for UFP. Real-time measurements made of NO x (by chemiluminesence), black carbon (BC) (by light absorption), particulate matter-phase PAH (by UV ionization), and particle length (by diffusional charging) showed high correlations with UFP numbers, ( for NO, 0.76 for BC, 0.69 for PAH, and 0.88 for particle length). Average concentrations of UFP and related pollutants varied strongly by location, road type, and truck traffic volumes, suggesting a relationship between these concentrations and truck traffic density. [Copyright &y& Elsevier]

Building occupants, including cleaning personnel, are exposed to a wide variety of airborne chemicals when cleaning agents and air fresheners are used in buildings. Certain of these chemicals are listed by the state of California as toxic air contaminants (TACs) and a subset of these are regulated by the US federal government as hazardous air pollutants (HAPs). California''s Proposition 65 list of species recognized as carcinogens or reproductive toxicants also includes constituents of certain cleaning products and air fresheners. In addition, many cleaning agents and air fresheners contain chemicals that can react with other air contaminants to yield potentially harmful secondary products. For example, terpenes can react rapidly with ozone in indoor air generating many secondary pollutants, including TACs such as formaldehyde. Furthermore, ozone–terpene reactions produce the hydroxyl radical, which reacts rapidly with organics, leading to the formation of other potentially toxic air pollutants. Indoor reactive chemistry involving the nitrate radical and cleaning-product constituents is also of concern, since it produces organic nitrates as well as some of the same oxidation products generated by ozone and hydroxyl radicals.Few studies have directly addressed the indoor concentrations of TACs that might result from primary emissions or secondary pollutant formation following the use of cleaning agents and air fresheners. In this paper, we combine direct empirical evidence with the basic principles of indoor pollutant behavior and with information from relevant studies, to analyze and critically assess air pollutant exposures resulting from the use of cleaning products and air fresheners. Attention is focused on compounds that are listed as HAPs, TACs or Proposition 65 carcinogens/reproductive toxicants and compounds that can readily react to generate secondary pollutants. The toxicity of many of these secondary pollutants has yet to be evaluated. The inhalation intake of airborne organic compounds from cleaning product use is estimated to be of the order of 10 mg d−1 person−1 in California. More than two dozen research articles present evidence of adverse health effects from inhalation exposure associated with cleaning or cleaning products. Exposure to primary and secondary pollutants depends on the complex interplay of many sets of factors and processes, including cleaning product composition, usage, building occupancy, emission dynamics, transport and mixing, building ventilation, sorptive interactions with building surfaces, and reactive chemistry. Current understanding is sufficient to describe the influence of these variables qualitatively in most cases and quantitatively in a few. [Copyright &y& Elsevier]

The East Bay Children''s Respiratory Health Study is examining associations between traffic-related pollutant exposures and respiratory health among children who reside and attend schools at varied proximity to northern California freeways. Chronic exposures are being inferred from outdoor pollutant concentrations at neighborhood schools. This paper reports primarily weeklong integrated NO2 and NOX concentrations measured with passive samplers placed outside at 10 elementary schools during 14 weeks in spring and 8 weeks in fall 2001. Measurements were also made outside selected student residences to examine spatial variability within three school neighborhoods. Regional concentrations of NO2 and NOX varied widely from week to week. School site data were normalized to measurements at a nearby regional monitoring station to facilitate analysis of relative pollutant exposures at the neighborhood schools. Normalized concentrations were consistent at each school throughout the study. Schools located upwind or far downwind of freeways were generally indistinguishable from one another and regional pollution levels. For school and neighborhood sites within 350 m downwind of a freeway, concentrations increased with decreasing downwind distance. The highest normalized concentrations occurred at a school located directly adjacent to a major freeway and a shopping center. In this case, normalized NO2 and NOX were ∼60% and ∼100% higher than regional background levels. At three schools within 130–230 m downwind of a freeway, normalized NO2 and NOX were ∼20–30% and ∼50–80% higher than regional levels. Validation testing of the passive samplers indicated precision of better than 5% for both NO2 and NOX when samplers were deployed outside for 1-week periods. Passive sampler results agreed with co-located chemiluminescence measurements to within 8% for NO2 and 3% for NOX. [Copyright &y& Elsevier]

Current modeling methods are inadequate for producing gridded mobile source inventories suitable for simulating fine-scale changes in emissions for transportation planning or regulatory control purposes. In this paper, we present a new model that more precisely simulates small-scale emission changes that can arise from, for example, implementation of traffic control measures in transportation networks. The new model, UCDrive, directly interfaces with most travel demand models and with California''s regulatory emission factor model EMFAC2001/2002. The model improves the sensitivity of emission estimates to variations in vehicle speeds and uses a theoretically consistent approach to spatially disaggregating mobile emissions. To demonstrate application of the model we use the Sacramento area network to perform a comparison study. We show that regional emissions with the new model replicate regulatory inventories, which is critical for using the model for regulatory purposes. We also show that the new model''s results better replicate regulatory totals than those derived with a frequently used gridded emissions inventory model. The new model should improve the ability to better simulate changes in mobile emissions from transportation projects and help to bridge the gap between gridded mobile emission inventories and regional mobile emission inventories. [Copyright &y& Elsevier]

This paper reviews sea level and elevated surface observations from Point Mugu and nearby Laguna Peak in southern Ventura County, rawin/ozonesonde data from Point Mugu, and aircraft measurements in the coastal zone and offshore obtained in support of the SCOS97-NARSTO effort. Streamline analyses of SCOS97-NARSTO wind data are also presented, which reveal synoptic and mesoscale influences on air flow over the southern California area. Complex diurnal, mesoscale and synoptic wind patterns redistribute pollutants across air basins. The data show that southeasterly flow, often present from morning into early afternoon, and coastal eddies and southerly surges convey ozone and precursors westward and northward where they may appear aloft as layers with several times the concentration of ozone at the surface. This flow can transport South Coast Air Basin (SCAB) pollutants via mountain passes to the desert where mixing with pollution from the San Joaquin Valley can occur. Climatological evidence shows this tendency for southerly flow in the earlier part of the day is a recurrent feature of southern California wind patterns, and a major mechanism for transport of ozone and precursors from the SCAB to coastal, inland and desert regions to the north. [Copyright &y& Elsevier]