Long-term exposure to traffic-related air pollution and non-accidental mortality: A systematic review and meta-analysis

BACKGROUND

The health effects of traffic-related air pollution (TRAP) continue to be of important public health interest across the globe. Following its 2010 review, the Health Effects Institute appointed a new expert Panel to systematically evaluate the epidemiological evidence regarding the associations between long-term exposure to TRAP and selected health outcomes. This paper describes the main findings of the systematic review on non-accidental mortality.

METHODS

The Panel used a systematic approach to conduct the review. An extensive search was conducted of literature published between 1980 and 2019. A new exposure framework was developed to determine whether a study was sufficiently specific to TRAP, which included studies beyond the near-roadway environment. We performed random-effects meta-analysis when at least three estimates were available of an association between a specific exposure and outcome. We evaluated confidence in the evidence using a modified Office of Health Assessment and Translation (OHAT) approach, supplemented with a broader narrative synthesis.

RESULTS

Thirty-six cohort studies were included. Virtually all studies adjusted for a large number of individual and area-level covariates-including smoking, body mass index, and individual and area-level socioeconomic status-and were judged at a low or moderate risk for bias. Most studies were conducted in North America and Europe, and a few were based in Asia and Australia. The meta-analytic summary estimates for nitrogen dioxide, elemental carbon and fine particulate matter-pollutants with more than 10 studies-were 1.04 (95% CI 1.01, 1.06), 1.02 (1.00, 1.04) and 1.03 (1.01, 1.05) per 10, 1 and 5 µg/m³, respectively. Effect estimates are interpreted as the relative risk of mortality when the exposure differs with the selected increment. The confidence in the evidence for these pollutants was judged as high, because of upgrades for monotonic exposure-response and consistency across populations. The consistent findings across geographical regions, exposure assessment methods and confounder adjustment resulted in a high confidence rating using a narrative approach as well.

CONCLUSIONS

The overall confidence in the evidence for a positive association between long-term exposure to TRAP and non-accidental mortality was high.

Black carbon in Xiamen, China: Temporal variations, transport pathways and impacts of synoptic circulation

Black carbon (BC) plays a vital role in atmospheric environment and climate change. Temporal variations and transport pathways of BC in Xiamen, China with the impacts of synoptic circulation were investigated in 2014 with Aethalometer. Annual mean BC concentration was 4270 ng m^-3. BC exhibited clear diurnal (seasonal) variations, with the maximum of 6182 (4755) ng m^-3 at 6:00 (in spring) and minimum of 2847 (3774) ng m^-3 at 13:00 (in summer). Conditional probability function analysis indicated that high BC concentrations were associated with northwesterly winds with low wind speed. Air masses originating from the East China Sea and passing along with East China Coast had the highest BC concentrations. Potential source contribution function and concentration weighted trajectory analysis suggested that major sources for BC included the surrounding region, southwestern Fujian and eastern Guangdong to the southwest, Hubei, Hunan and Jiangxi to the northwest, the East China Sea and the South China Sea. Of the nine synoptic circulation patterns, three cyclone-related patterns were associated with low BC concentrations and small biomass burning (BC_bb) contributions. Of the six anticyclone-related patterns, the three cold-high circulations around winter were associated with moderate BC concentrations and large BC_bb contributions. The two cold-high patterns in spring and autumn were associated with high BC concentrations and small BC_bb contributions, while the warm-high pattern was associated with moderate BC concentration and small BC_bb contribution. The findings provide insights into the transport mechanisms of BC with the impacts of synoptic pattern in China.

Toward a better understanding of the impact of mass transit air pollutants on human health

Globally, modern mass transport systems whether by road, rail, water, or air generate airborne pollutants in both developing and developed nations. Air pollution is the primary human health concern originating from modern transportation, particularly in densely-populated urban areas. This review will specifically focus on the origin and the health impacts of carbonaceous traffic-related air pollutants (TRAP), including particulate matter (PM), volatile organic compounds (VOCs), and elemental carbon (EC). We conclude that the greatest current challenge regarding urban TRAP is understanding and evaluating the human health impacts well enough to set appropriate pollution control measures. Furthermore, we provide a detailed discussion regarding the effects of TRAP on local environments and pedestrian health in low and high traffic-density environments.

Size-resolved carbonaceous components and water-soluble ions measurements of ambient aerosol in Beijing

A MOUDI-120 sampler was used in Beijing to collect multi-stage samples in the summer and winter of 2013 to 2015. Thirty-three sample sets were collected during the daytime, nighttime, and different pollution levels. The actual relative humidity in the impactors was calculated for the first time. The carbonaceous components (organic and elemental carbon, OC and EC, respectively) and water-soluble inorganic ions (Na⁺, NH₄⁺, K⁺, Mg²⁺, Ca²⁺, Cl^-, NO₃^-, and SO₄^2-) were analyzed in each sample. The characteristics of the mass concentration distribution and charge balance were discussed. On the basis of relative humidity in the impactors, aerosols less than 1.0μm were sampled under relatively dry conditions in most cases. The concentration levels for the chemical species were higher in the winter than in the summer. Three modes (condensation mode, droplet mode, and coarse mode) could be identified from the distributions of NH₄⁺, NO₃^-, SO₄^2-, Cl^-, K⁺, OC and EC. The distribution characteristics for the pollution dissipation process were different from the pollution accumulation process. NO₃^- and NO₂^- contributed most of the negative electric charges in the stage below 0.1μm. In the condensation mode, the cations were dominated by NH₄⁺, which was sufficient to balance the anions. In the droplet mode of the heavily polluted samples, the ammonium was not sufficient to balance the anions. In the coarse mode, the positive electric charges were primarily composed of metal cations. The analyzed anions were not sufficient to neutralize the measured cations.

Long-term variation of black carbon and PM2.5 in Beijing, China with respect to meteorological conditions and governmental measures

Black carbon (BC) and PM2.5 were studied for nine years from 2005 to 2013 in the Beijing urban area. The overall weekly average mass concentrations of BC and PM2.5 were 4.3 and 66.8 μg/m³. PM2.5 annual means of the nine years are around 2 times of the standard (GB3095-2012) in China, and are 5-7 times higher than the WHO standard. The Beijing Olympic Games in 2008 was a milestone to mitigate aerosol pollution. Temporal distribution of BC shows a distinct declining trend, and annual mean mass concentrations of PM2.5 after 2008 were lower than those before 2008 but increased from 2011 to 2013. Wind rose plots show that high BC concentrations are usually associated with low wind speed of northeastern or southwestern winds, generally causing poor visibility. Governmental mitigation measures such as traffic restriction despite increased motor vehicle numbers and gasoline consumption and industry relocation with declining consumption of coal and coke were successful in reducing BC emissions. Annual mean of BC was reduced by 38% in 2013 compared to 2005. However, BC contamination in Beijing is still severe when compared to other urban areas around the world.

Black carbon emissions in gasoline vehicle exhaust: a measurement and instrument comparison

A pilot study was conducted to evaluate the performance and agreement of several commercially available black carbon (BC) measurement instruments, when applied to the quantification of BC in light-duty vehicle (LDV) exhaust. Samples from six vehicles, three fuels, and three driving cycles were used. The pilot study included determinations of the method detection limit (MDL) and repeatability. With respect to the MDL, the real-time instruments outperformed the time-integrated instruments, with MDL = 0.12 mg/mi for the AE51 Aethalometer, and 0.15 mg/mi for the Micro Soot Sensor (MSS), versus 0.38 mg/mi for the IMPROVE_A thermal/ optical method, and 0.35 mg/mi for the OT21_T Optical Transmissometer. The real-time instruments had repeatability values ranging from 30% to 35%, which are somewhat better than those of the time-integrated instruments (40-41%). These results suggest that, despite being less resource intensive, real-time methods can be equivalent or superior to time-integrated methods in terms of sensitivity and repeatability. BC mass data, from the photoacoustic and light attenuation instruments, were compared against same-test EC data, determined using the IMPROVE_A method. The MSS BC data was well correlated with EC, with R2 = 0.85 for the composite results and R2 = 0.86 for the phase-by-phase (PBP) results. The correlation of BC, by the AE51, AE22, and OT21_T with EC was moderate to weak. The weaker correlation was driven by the inclusion of US06 test data in the linear regression analysis. We hypothesize that test-cycle-dependent BC:EC ratios are due to the different physicochemical properties of particulate matter (PM) in US06 and Federal Test Procedure (FTP) tests. Correlation amongst the real-time MSS, PASS-1, AE51, and AE22 instruments was excellent (R2 = 0.83-0.95), below 1 mg/mi levels. In the process of investigating these BC instruments, we learned that BC emissions at sub-1 mg/mi levels can be measured and are achievable by current-generation gasoline engines.

IMPLICATIONS

Most comparison studies of black carbon (BC) measurement methods were carried out in the ambient air. This study assesses the agreement among various BC measurement instrument in emissions from light-duty gasoline vehicles (LDGVs) on standard test cycles, and evaluates applicability of these methods under various fuel types, driving cycles, and engine combustion technologies. This research helps to fill in the knowledge gap of BC method standardization as stated in the U.S. Environmental Protection Agency (EPA) 2011 Report to Congress on Black Carbon, and these results demonstrate the feasibility of quantification of BC at the 1 mg/mi PM standard in California Low Emission Vehicle III regulations.

Filter light attenuation as a surrogate for elemental carbon

Light attenuation (b(att)) measured from filter light transmission is compared with elemental carbon (EC) measurements for more than 180,000 collocated PM2.5 (particulate matter [PM] < or = 2.5 microm in aerodynamic diameter) and PM10 (PM < or = 10 microm in aerodynamic diameter) samples from nearly 200 U.S. locations during the past 2 decades. Although there are theoretical reasons for expecting highly variable relationships between b(att) and EC (such as the effects of "brown carbon" and iron oxides in PM2.5), reasonable correlations are found. These correlations are not a strong function of season or location (e.g., rural vs. urban). Median EC concentrations can be predicted from filter transmittance measurements to within +/- 15-30%. Although EC predicted from b(att) shows larger uncertainties (30-60%), especially at concentrations less than 0.3 microg/m3, the consistent mass absorption efficiency (sigm(att)) derived from the regression analysis demonstrates the feasibility of using b(att) as a surrogate for EC. This study demonstrates that a constant factor of 0.1 g/m2 (equivalent to the 10 m2/g sigma(att) used in the Interagency Monitoring of Protected Visual Environments chemical extinction formula) can be used to estimate EC concentrations from b(att) through a Teflon-membrane filter sample. Greater accuracy is achieved with site-specific sigma(att) derived from a period with collocated EC measurements.

Positive sampling artifacts in particulate organic carbon measurements in roadside environment

The positive artifacts in particulate organic carbon measurements in a roadside environment were characterized using two filters in tandem. The experiments were performed for PM(1.0), PM(2.5), and PM(10) at 24-h interval using a URG sampler, followed by organic carbon (OC)/elemental carbon (EC) analysis by the Interagency Monitoring of Protected Visual Environments thermal/optical reflectance carbon analysis protocol. The OC concentrations, derived from the quartz filter behind a front quartz filter, were quite similar for PM(1.0), PM(2.5), and PM(10), ranging from 0.6 to 2.7 microg C m(-3) for PM(1.0), from 0.7 to 2.7 microg C m(-3) for PM(2.5), and from 1.1 to 2.7 microg C m(-3) for PM(10). They were respectively approximately 2.8%, approximately 2.4%, and approximately 1.6% of the particulate mass. The most abundant species on the backup quartz filters were OC2 (250 degrees C) and OC3 (450 degrees C), accounting for approximately 80% of measured organic carbon on the backup quartz filters. It indicates the filter artifacts are mainly composed of adsorbed semi-volatile organics (below the analysis temperature of 450 degrees C) including gaseous and particulate phase; the loading of artifacts depends on the nature of vapor and its interaction with filter substrate, rather than particle sizes. The uncorrected OC/EC ratios on the front quartz filters were approximately 10% higher than the corrected OC/EC ratios by positive organic artifacts in winter, and it is approximately 20% higher in summer. Another finding is that the separation distance of the front and backup filters influence the level of artifacts assessed by the backup filter.

Absorption coefficient and site-specific mass absorption efficiency of elemental carbon in aerosols over urban, rural, and high-altitude sites in India

Temporal and spatial variability in the absorption coefficient (b(abs), Mm(-1)) and mass absorption efficiency (MAE, sigma(abs), m(2)g(-1)) of elemental carbon (EC) in atmospheric aerosols studied from urban, rural, and high-altitude sites is reported here. Ambient aerosols, collected on tissuquartz filters, are analyzed for EC mass concentration using thermo-optical EC-OC analyzer, wherein simultaneously measured optical-attenuation (ATN, equivalent to initial transmittance) of 678 nm laser source has been used for the determination of MAE and absorption coefficient. At high-altitude sites, measured ATN and surface EC loading (EC(s), microg cm(-2)) on the filters exhibit linear positive relationship (R(2) = 0.86-0.96), suggesting EC as a principal absorbing component. However, relatively large scatter in regression analyses for the data from urban sites suggests contribution from other species. The representative MAE of EC, during wintertime (Dec 2004), at a rural site (Jaduguda) is 6.1 +/- 2.0 m(2)g(-1). In contrast, MAE at the two high-altitude sites is 14.5 +/- 1.1 (Manora Peak) and 10.4 +/- 1.4 (Mt. Abu); and that at urban sites is 11.1 +/- 2.6 (Allahabad) and 11.3 +/- 2.2 m(2)g(-1) (Hisar). The long-term average MAE at Manora Peak (February 2005 to June 2007) is 12.8 +/- 2.9 m(2)g(-1) (range: 6.1-19.1 m(2)g(-1)). These results are unlike the constant conversion factor used for MAE in optical instruments for the determination of BC mass concentration. The absorption coefficient also shows large spatiotemporal variability; the lower values are typical of the high-altitude sites and higher values for the urban and rural atmosphere. Such large variability documented for the absorption parameters suggests the need for their suitable parametrization in the assessment of direct aerosol radiative forcing on a regional scale.

Methods to assess carbonaceous aerosol sampling artifacts for IMPROVE and other long-term networks

Volatile organic compounds (VOCs) and semi-volatile organic compounds (SVOCs) adsorb to quartz fiber filters during fine and coarse particulate matter (PM2.5 and PM10, respectively) sampling for thermal/optical carbon analysis that measures organic carbon (OC) and elemental carbon (EC). Particulate SVOCs can evaporate after collection, with a small portion adsorbed within the filter. Adsorbed organic gases are measured as particulate OC, so passive field blanks, backup filters, prefilter organic denuders, and regression methods have been applied to compensate for positive OC artifacts in several long-term chemical speciation networks. Average backup filter OC levels from the Interagency Monitoring of Protected Visual Environments (IMPROVE) network were approximately 19% higher than field blank values. This difference is within the standard deviation of the average and likely results from low SVOC concentrations in the rural to remote environments of most IMPROVE sites. Backup filters from an urban (Fort Meade, MD) site showed twice the OC levels of field blanks. Sectioning backup filters from top to bottom showed nonuniform OC densities within the filter, contrary to the assumption that VOCs and SVOCs on a backup filter equal those on the front filter. This nonuniformity may be partially explained by evaporation and readsorption of vapors in different parts of the front and backup quartz fiber filter owing to temperature, relative humidity, and ambient concentration changes throughout a 24-hr sample duration. OC-PM2.5 regression analysis and organic denuder approaches demonstrate negative sampling artifact from both Teflon membrane and quartz fiber filters.

Continuous and filter-based measurements of PM 2.5 nitrate and sulfate at the Fresno Supersite

PM(2.5) nitrate (NO-(3)) and sulfate (SO=(4)) were measured continuously with R&P8400N and R&P8400S instruments, respectively, and compared with filter-based measurements at the Fresno Supersite from October, 2000 through December, 2005. NO-(3) concentrations were higher in winter than summer with a long-term decreasing trend. Correlations between 24-h average continuous and filter-based NO-(3) were greater than 0.96 in 4 out of 5 years. Continuous NO-(3) was generally lower than filter-based NO-(3) although the difference decreased over time, from -52% in 2001 to +13% in 2005. These differences were similar in winter (-23%) and summer (-19%) while the corresponding differences between ambient and instrument temperature were -12 and 0.7 degrees C, respectively. Neither seasonal nor long-term trends in NO-(3) can be explained by variations in ambient temperature, the difference between ambient and instrument temperature, or changes in aerosol chemical composition. There were no seasonal or long-term trends in SO=(4) concentrations, partially due to low concentrations observed in Fresno. Long-term variability in the performance of R&P8400 NO-(3) and SO=(4) instruments suggest that collocation with filter measurements is needed for long-term measurements.

Characterisation of solvent extractable organic constituents in atmospheric particulate matter: an overview

In spite of accounting for 10-70% of the atmospheric aerosol mass, particulate-phase organic compounds are not well characterised, and many aspects of aerosol formation and evolution are still unknown. The growing awareness of the impact of particulate aerosols on climate, and the incompletely recognised but serious effects of anthropogenic constituents on air quality and human health, have conducted to several scientific studies. These investigations have provided information about the behaviour of atmospheric particulate matter and the description of the character of its carbonaceous content. The compilation of such results is important as they append to the emergent global-wide dataset of the organic composition of atmospheric aerosols. The contribution of the major emission sources to regional particulate pollution can be diagnosed by using specific molecular markers. This overview is mainly focused on results obtained with gas chromatography coupled with mass spectrometry, since it is the analytical method of choice in elucidating the solvent-extractable organic compounds in atmospheric particulate matter. A synopsis of the selection of organic tracers and the application of geochemical parameters to the analysis of organic constituents as a tool for source apportionment is shown here. Besides the assessment of current knowledge, this paper also presents the identification of further areas of concern.

Transport and dispersion during wintertime particulate matter episodes in the San Joaquin Valley, California

Data analysis and modeling were performed to characterize the spatial and temporal variability of wintertime transport and dispersion processes and the impact of these processes on particulate matter (PM) concentrations in the California San Joaquin Valley (SJV). Radar wind profiler (RWP) and radio acoustic sounding system (RASS) data collected from 18 sites throughout Central California were used to estimate hourly mixing heights for a 3-month period and to create case studies of high-resolution diagnostic wind fields, which were used for trajectory and dispersion analyses. Data analyses show that PM episodes were characterized by an upper-level ridge of high pressure that generally produced light winds through the entire depth of the atmospheric boundary layer and low mixing heights compared with nonepisode days. Peak daytime mixing heights during episodes were -400 m above ground level (agl) compared with -800 m agl during nonepisodes. These episode/nonepisode differences were observed throughout the SJV. Dispersion modeling indicates that the range of influence of primary PM emitted in major population centers within the SJV ranged from -15 to 50 km. Trajectory analyses revealed that little intrabasin pollutant transport occurred among major population centers in the SJV; however, interbasin transport from the northern SJV and Sacramento regions into the San Francisco Bay Area (SFBA) was often observed. In addition, this analysis demonstrates the usefulness of integrating RWP/RASS measurements into data analyses and modeling to improve the understanding of meteorological processes that impact pollution, such as aloft transport and boundary layer evolution.

Nanoparticle and ultrafine particle events at the Fresno supersite

Continuous measurements of particle size distributions of 3-407 nm were collected from August 2002 to July 2004 at the Fresno Supersite to understand their number concentrations, size distributions, and formation processes. Measurements for fine particulate matter (PM2.5) mass, sulfate (SO4(2-)), nitrate (NO3-), black carbon (BC), particle-bound polycyclic aromatic hydrocarbons (PAHs), nitrogen oxides (NOx), carbon monoxide (CO), ozone (O3), and meteorological data (wind speed, wind direction, temperature [T], relative humidity [RH], and solar radiation) were used to determine the causes of nanoparticle (3-10 nm) and ultrafine (10-100 nm) particle events. These events were found to be divided into four types: (1) 3- to 10-nm morning nucleation; (2) 10- to 30-nm morning traffic; (3) 10- to 30-nm afternoon photochemical; and (4) 50- to 84-nm evening home heating, including residential wood combustion. Intense examples of the first type (>10(4) number [#]/cm3) were observed on 29 days, nearly always during the summer. The second type of event was observed on more than 73 days and occurred throughout the year. The third type was observed on 36 days, from spring through summer. The fourth type was found on 109 days, all of them during the winter. Although sulfur dioxide (SO2) emissions in Central California are low, the small residual amounts in gasoline and diesel fuel are apparently sufficient to initiate nucleation events. These were measured in the morning, soon after the shallow surface inversion coupled with layers aloft where nucleation probably was initiated. PM2.5 concentrations were poorly correlated with nanoparticle number.

Particulate carbon measurements in California's San Joaquin Valley

Aerosol carbon sampling methods and biases were evaluated during the California Regional PM10/PM2.5 Air Quality Study (CRPAQS) and Fresno Supersite programs. PM2.5 sampling was conducted using Desert Research Institute (DRI) sequential filter samplers (SFS) from December 1999 through February 2001 at two urban sites (Fresno and Bakersfield), one regional transport site (Angiola), and two boundary sites (Bethel Island and Sierra Nevada Foothills) during CRPAQS in the San Joaquin Valley (SJV). Additional filter-based sampling was done in Fresno as part of the US Environmental Protection Agency (EPA) Supersites program. Organic carbon (OC) and elemental carbon (EC) concentrations were higher during winter (December-February) than summer (June-August) and this trend was most pronounced at Fresno and Bakersfield. OC and EC displayed similar diurnal trends during winter and summer at Fresno and during winter at Angiola. The diurnal pattern at Angiola reflected the transport of secondary pollutants to the site. Collocated measurements of OC and EC on undenuded quartz-fiber filters were made at Fresno with the DRI SFS and the Andersen FRM and RAAS samplers. All average differences in OC between samplers were less than their respective measurement uncertainties. Positive and negative OC biases were evaluated at Fresno using the Andersen RAAS sampler with carbon-denuded and undenuded channels with Teflon-membrane and quartz-fiber filter pairs. Differences between the denuded particle OC and that obtained by subtracting the quartz-behind-Teflon or quartz-behind-quartz OC from the undenuded quartz-fiber front filter were less than twice their measurement uncertainties in most cases. Particulate OC in the denuded channel agreed most closely with the difference between undenuded front and backup quartz-fiber OC.

Comparison of PM2.5 carbon measurement methods in Hong Kong, China

Samples from Hong Kong, China, were analyzed for organic carbon (OC), elemental carbon (EC), and total carbon (TC) by three thermal protocols (low-temperature IMPROVE and high-temperature STN and NIOSH) and two optical monitoring methods: reflectance and transmittance. Good agreement (+/-10%) for TC among the three protocols was observed for sample loadings of 1-55 microg m(-3). The two protocols using a reflectance pyrolysis correction showed best agreement for EC, with <20% differences found for approximately 80% of the samples. Hong Kong has a large diesel fleet, and for some heavily loaded samples the light transmittance was too low for quantitative detection, resulting in large uncertainties in the OC/EC split based on transmittance. Hong Kong experienced OC levels similar to those at US sites, but has much higher EC concentrations. OC/EC ratios range from 2 to 5 at two US sites and from 0.2 to 1.2 at three Hong Kong sites.

PM2.5 chemical composition in Hong Kong: urban and regional variations

Chemically speciated PM2.5 measurements were made at roadside, urban, and rural background sites in Hong Kong for 1 year during 2000/2001 to determine the spatial and temporal variations of PM2.5 mass and chemical composition in this highly populated region. Annual average PM2.5 concentrations at the urban and rural sites were 34.1 and 23.7 microg m(-3), respectively, approximately 50-100% higher than the United States' annual average National Ambient Air Quality Standard (NAAQS) of 15 microg m(-3). Daily PM2.5 concentrations exceeded the U.S. 24-h NAAQS of 65 microg m(-3) on 19 days, reaching 131+/-8 microg m(-3) at the roadside site on 02/28/2001. Carbonaceous aerosol is the largest contributor to PM2.5 mass (explaining 52-75% of PM2.5 mass at the two urban sites and 32% at the background site), followed by ammonium sulfate (ranging from 23% to 37% at the two urban sites and 51% at the background site). Ammonium sulfate and crustal concentrations showed more uniform spatial distributions, while the largest urban-rural contrasts found in carbonaceous aerosol (likely due to emissions from on-road gasoline and diesel vehicles). Marine influences accounted for 7% of the mass at the background site (more than twice as much as at the two urban sites). Ternary diagrams are utilized to illustrate the different spatial patterns.

Visibility: science and regulation

The 1999 Regional Haze Rule provides a context for this review of visibility, the science that describes it, and the use of that science in regulatory guidance. The scientific basis for the 1999 regulation is adequate. The deciview metric that tracks progress is an imperfect but objective measure of what people see near the prevailing visual range. The definition of natural visibility conditions is adequate for current planning, but it will need to be refined as visibility improves. Emissions from other countries will set achievable levels above those produced by natural sources. Some natural events, notably dust storms and wildfires, are episodic and cannot be represented by annual average background values or emission estimates. Sulfur dioxide (SO2) emission reductions correspond with lower sulfate (SO4(2-)) concentrations and visibility improvements in the regions where these have occurred. Non-road emissions have been growing more rapidly than emissions from other sources, which have remained stable or decreased since 1970. Simpler models representing transport, limiting precursor pollutants, and gas-to-particle equilibrium should be used to understand where and when emission reductions will be effective, rather than large complex models that have insufficient input and validation measurements. Examples of model-based source attribution show large differences among estimates from various modeling systems and with ambient measurements.