Road traffic impact on urban water quality: a step towards integrated traffic, air and stormwater modelling.

Methods for simulating air pollution due to road traffic and the associated effects on stormwater runoff quality in an urban environment are examined with particular emphasis on the integration of the various simulation models into a consistent modelling chain. To that end, the models for traffic, pollutant emissions, atmospheric dispersion and deposition, and stormwater contamination are reviewed. The present study focuses on the implementation of a modelling chain for an actual urban case study, which is the contamination of water runoff by cadmium (Cd), lead (Pb), and zinc (Zn) in the Grigny urban catchment near Paris, France. First, traffic emissions are calculated with traffic inputs using the COPERT4 methodology. Next, the atmospheric dispersion of pollutants is simulated with the Polyphemus line source model and pollutant deposition fluxes in different subcatchment areas are calculated. Finally, the SWMM water quantity and quality model is used to estimate the concentrations of pollutants in stormwater runoff. The simulation results are compared to mass flow rates and concentrations of Cd, Pb and Zn measured at the catchment outlet. The contribution of local traffic to stormwater contamination is estimated to be significant for Pb and, to a lesser extent, for Zn and Cd; however, Pb is most likely overestimated due to outdated emissions factors. The results demonstrate the importance of treating distributed traffic emissions from major roadways explicitly since the impact of these sources on concentrations in the catchment outlet is underestimated when those traffic emissions are spatially averaged over the catchment area.

Investigating the impact of aqueous-phase chemistry and wet deposition on organic aerosol formation using a molecular surrogate modeling approach.

A molecular surrogate representation of secondary organic aerosol (SOA) formation is used to investigate the effect of aqueous-phase (in clouds and particles) chemical processing and wet deposition on SOA atmospheric concentrations. To that end, the hydrophilic/hydrophobic organic (H(2)O) model was augmented to account for several gas/aqueous-phase equilibria and aqueous-phase processes, including the formation of oxalic, glyoxilic and pyruvic acids, the oxidation of methyl vinyl ketone (MVK) and methacrolein (MACR), the formation of tetrols and organosulfates from epoxydiols (IEPOX), and further oxidation of water-soluble SOA (aging). Among those processes, SOA chemical aging and IEPOX reactions led to the most significant increases (up to 1 µg m(-3) in some areas) in SOA concentrations in a one-month summer simulation over Europe. However, large uncertainties remain in the gas/aqueous-phase partitioning of oxalic acid, MVK, and MACR. Below-cloud scavenging of SOA precursor gases and of gas-phase SVOC was found to affect SOA concentrations by up to 20%, which suggests that it should be taken into account in air quality models.

Comparison of different gas-phase mechanisms and aerosol modules for simulating particulate matter formation.

The effects of two gas-phase chemical kinetic mechanisms, Regional Atmospheric Chemistry Mechanism version 2 (RACM2) and Carbon-Bond 05 (CB05), and two secondary organic aerosol (SOA) modules, the Secondary Organic Aerosoi Model (SORGAM) and AER/EPRI/Caltech model (AEC), on fine (aerodynamic diameter < or =2.5 microm) particulate matter (PM2.5) formation is studied. The major sources of uncertainty in the chemistry of SOA formation are investigated. The use of all major SOA precursors and the treatment of SOA oligomerization are found to be the most important factors for SOA formation, leading to 66% and 60% more SOA, respectively. The explicit representation of high-NO, and low-NOx gas-phase chemical regimes is also important with increases in SOA of 30-120% depending on the approach used to implement the distinct SOA yields within the gas-phase chemical kinetic mechanism; further work is needed to develop gas-phase mechanisms that are fully compatible with SOA formation algorithms. The treatment of isoprene SOA as hydrophobic or hydrophilic leads to a significant difference, with more SOA being formed in the latter case. The activity coefficients may also be a major source of uncertainty, as they may differ significantly between atmospheric particles, which contain a myriad of SOA, primary organic aerosol (POA), and inorganic aerosol species, and particles formed in a smog chamber from a single precursor under dry conditions. Significant interactions exist between the uncertainties of the gas-phase chemistry and those of the SOA module.

A case study of the relative effects of power plant nitrogen oxides and sulfur dioxide emission reductions on atmospheric nitrogen deposition.

The contrasting effects of point source nitrogen oxides (NOx) and sulfur dioxide (SO2) air emission reductions on regional atmospheric nitrogen deposition are analyzed for the case study of a coal-fired power plant in the southeastern United States. The effect of potential emission reductions at the plant on nitrogen deposition to Escambia Bay and its watershed on the Florida-Alabama border is simulated using the three-dimensional Eulerian Community Multiscale Air Quality (CMAQ) model. A method to quantify the relative and individual effects of NOx versus SO2 controls on nitrogen deposition using air quality modeling results obtained from the simultaneous application of NOx and SO2 emission controls is presented and discussed using the results from CMAQ simulations conducted with NOx-only and SO2-only emission reductions; the method applies only to cases in which ambient inorganic nitrate is present mostly in the gas phase; that is, in the form of gaseous nitric acid (HNO3). In such instances, the individual effects of NOx and SO2 controls on nitrogen deposition can be approximated by the effects of combined NOx + SO2 controls on the deposition of NOy, (the sum of oxidized nitrogen species) and reduced nitrogen species (NHx), respectively. The benefit of controls at the plant in terms of the decrease in nitrogen deposition to Escambia Bay and watershed is less than 6% of the overall benefit due to regional Clean Air Interstate Rule (CAIR) controls.

Current understanding of ultrafine particulate matter emitted from mobile sources.

The physics and chemistry of ultrafine particulate matter (UFPM) associated with mobile source emissions is reviewed. UFPM includes those particles that are less than 0.1 microm in diameter. Measurements of UFPM emitted from mobile sources have been conducted in the laboratory, on roadways, and downwind of roadways. In addition, UFPM formation and evolution have been modeled and the modeling results have been compared with available measurements. The results of those measurement programs and modeling studies are synthesized into a coherent description of the formation of UFPM in mobile source emissions and its subsequent evolution in the ambient atmosphere. Recommendations are provided to address the gaps in our current understanding of the processes leading to the formation of UFPM, its chemical composition, and its evolution in the atmosphere.

Organic aerosol spatial/temporal patterns: perspectives of measurements and model.

Ambient measurements from SEARCH and model results from CMAQ-MADRID are analyzed side by side for the southeastern United States to understand the strengths and weaknesses of an air quality model in reproducing key spatial and temporal patterns related to organic aerosol (OA), with inferences regarding secondary organic aerosol (SOA). The model predicts a larger difference in OA concentrations between an urban (JST) and a rural site (YRK) than indicated by measurements. Modeled OA concentrations at JST and YRK are more strongly correlated than measurements. On average, models may understate urban OA emissions, while overstating urban SOA production; measurements indicate that SOA production takes place on the regional scale. Modeled diurnal fluctuations for OA are stronger than measured, due partially to overestimations of the temperature dependence parameters (deltaH(vap)) for SOA in the model. Urban-rural differences in the composition of SOA, inferred from the variations of estimated deltaH(vap), are not properly captured by the model, which does not represent multiple generations of SOA or varied reaction pathways as a function of chemical regimes. Model results are hampered by day-of-the-week and diurnal allocation issues related to EC and OA emissions. Top quintile (20%) afternoon OA concentrations are observed in both warm and cold seasons at the urban site. The frequency of high OA in the cold season is overstated in the model. The model predicts the warm vs cold season frequency of elevated OA episodes better at YRK than at JST, suggesting that regional emissions, chemistry, and transport are better simulated than urban processes.

Response of atmospheric particulate matter to changes in precursor emissions: a comparison of three air quality models.

Three mathematical models of air quality (CMAQ, CMAQ-MADRID, and REMSAD) are applied to simulate the response of atmospheric fine particulate matter (PM2.5) concentrations to reductions in the emissions of gaseous precursors for a 10 day period of the July 1999 Southern Oxidants Study (SOS) in Nashville. The models are shown to predict similar directions of the changes in PM2.5 mass and component (sulfate, nitrate, ammonium, and organic compounds) concentrations in response to changes in emissions of sulfur dioxide (SO2), nitrogen oxides (NO(x)), and volatile organic compounds (VOC), except for the effect of SO2 reduction on nitrate and the effect of VOC reduction on PM2.5 mass. Furthermore, in many cases where the directional changes are consistent, the magnitude of the changes are significantly different among models. Examples are the effects of SO2 and NO(x) reductions on nitrate and PM2.5 mass and the effects of VOC reduction on organic compounds, sulfate and nitrate. The spatial resolution significantly influences the results in some cases. Operational model performance for a PM2.5 component appears to provide some useful indication on the reliability of the relative response factors (RRFs) for a change in emissions of a direct precursor, as well as for a change in emissions of a compound that affects this component in an indirect manner, such as via oxidant formation. However, these results need to be confirmed for other conditions and caution is still needed when applying air quality models for the design of emission control strategies. It is advisable to use more than one air quality model (or more than one configuration of a single air quality model) to span the full range of plausible scientific representations of atmospheric processes when investigating future air quality scenarios.

A synthesis of progress and uncertainties in attributing the sources of mercury in deposition.

A panel of international experts was convened in Madison, Wisconsin, in 2005, as part of the 8th International Conference on Mercury as a Global Pollutant. Our charge was to address the state of science pertinent to source attribution, specifically our key question was: "For a given location, can we ascertain with confidence the relative contributions of local, regional, and global sources, and of natural versus anthropogenic emissions to mercury deposition?" The panel synthesized new research pertinent to this question published over the past decade, with emphasis on four major research topics: long-term anthropogenic change, current emission and deposition trends, chemical transformations and cycling, and modeling and uncertainty. Within each topic, the panel drew a series of conclusions, which are presented in this paper. These conclusions led us to concur that the answer to our question is a "qualified yes," with the qualification being dependent upon the level of uncertainty one is willing to accept. We agreed that the uncertainty is strongly dependent upon scale and that our question as stated is answerable with greater confidence both very near and very far from major point sources, assuming that the "global pool" is a recognizable "source." Many regions of interest from an ecosystem-exposure standpoint lie in between, where source attribution carries the greatest degree of uncertainty.

Modeling secondary organic aerosol formation via multiphase partitioning with molecular data.

A new model for atmospheric secondary organic aerosol (SOA) is presented for biogenic compounds. It is based to the extent possible on experimental molecular SOA data, and it is compatible with any existing gas-phase chemical kinetic mechanism. Six SOA precursors or groups of precursors are used to represent biogenic monoterpenes and sesquiterpenes. SOA formation is modeled using five SOA surrogates to represent classes of compounds with different partitioning properties, e.g., hydrophobicity, aqueous solubility, acid dissociation, and saturation vapor pressure. Model simulations are evaluated against smog chamber data for SOA yields and some adjustments are made to uncertain stoichiometric coefficients and saturation vapor pressure parameters to improve model performance. The model is applied undertypical atmospheric conditions to exemplify the effect of relative humidity on SOA formation and the relative contributions of hydrophilic and hydrophobic SOA.

Modeling mercury in power plant plumes.

Measurements of speciated mercury (Hg) downwind of coal-fired power plants suggest that the Hg(II)/(Hg0 + HgII) ratio (where HgII is divalent gaseous Hg and Hg0 is elemental Hg) decreases significantly between the point of emission and the downwind ground-level measurement site, but that the SO2/(Hg0 + HgII) ratio is conserved. We simulated nine power plant plume events with the Reactive & Optics Model of Emissions (ROME), a reactive plume model that includes a comprehensive treatment of plume dispersion, transformation, and deposition. The model simulations fail to reproduce such a depletion in HgII. A sensitivity study of the impact of the HgII dry deposition velocity shows that a difference in dry deposition alone cannot explain the disparity. Similarly, a sensitivity study of the impact of cloud chemistry on results shows that the effect of clouds on Hg chemistry has only minimal impact. Possible explanations include HgII reduction to Hg0 in the plume, rapid reduction of HgII to Hg0 on ground surfaces, and/or an overestimation of the HgII fraction in the power plant emissions. We propose that a chemical reaction not included in current models of atmospheric mercury reduces HgII to Hg0 in coal-fired power plant plumes. The incorporation of two possible reduction pathways for HgII (pseudo-first-order decay and reaction with SO2) shows better agreement between the model simulations and the ambient measurements. These potential HgII to Hg0 reactions need to be studied in the laboratory to investigate this hypothesis. Because the speciation of Hg has a significant effect on Hg deposition, models of the fate and transport of atmospheric Hg may need to be modified to account for the reduction of HgII in coal-fired power plant plumes if such a reaction is confirmed in further experimental investigations.

Air quality modeling of hazardous pollutants: current status and future directions.

The current requirements and status of air quality modeling of hazardous pollutants are reviewed. Many applications require the ability to predict the local impacts from industrial sources or large roadways as needed for community health characterization and evaluating environmental justice concerns. Such local-scale modeling assessments can be performed by using Gaussian dispersion models. However, these models have a limited ability to handle chemical transformations. A new generation of Eulerian grid-based models is now capable of comprehensively treating transport and chemical transformations of air toxics. However, they typically have coarse spatial resolution, and their computational requirements increase dramatically with finer spatial resolution. The authors present and discuss possible advanced approaches that can combine the grid-based models with local-scale information.

Modeling atmospheric mercury deposition in the vicinity of power plants.

Two mathematical models of the atmospheric fate and transport of mercury (Hg), an Eulerian grid-based model and a Gaussian plume model, are used to calculate the atmospheric deposition of Hg in the vicinity (i.e., within 50 km) of five coal-fired power plants. The former is applied using two different horizontal resolutions: coarse (84 km) and fine (16.7 km). More than 96% of the power plant Hg emissions are calculated with the plume model to be transported beyond 50 km from the plants. The grid-based model predicts a lower fraction to be transported beyond 50 km: >91% with a coarse resolution and >95% with a fine resolution. The contribution of the power plant emissions to total Hg deposition within a radius of 50 km from the plants is calculated to be <8% with the plume model, <14% with the Eulerian model with a coarse resolution, and <10% with the Eulerian model with a fine resolution. The Eulerian grid-based model predicts greater local impacts than the plume model because of artificially enhanced vertical dispersion; the former predicts about twice as much Hg deposition as the latter when the area considered is commensurate with the resolution of the grid-based model. If one compares the local impacts for an area that is significantly less than the grid-based model resolution, then the grid-based model may predict lower local deposition than the plume model, because two compensating errors affect the results obtained with the grid-based model: initial dilution of the power plant emissions within one or more grid cells and enhanced vertical mixing to the ground.

Global source attribution for mercury deposition in the United States.

A multiscale modeling system that consists of a global chemical transport model (CTM) and a nested continental CTM was used to simulate the global atmospheric fate and transport of mercury and its deposition over the contiguous United States. The performance of the CTMs was evaluated against available data. The coefficient of determination (r2) for observed versus simulated annual mercury wet deposition fluxes over North America was 0.50 with average normalized error and bias of 25% and 11%, respectively. The CTMs were used to conduct a global source attribution for selected receptor areas. Three global emission scenarios were used that differed in their distribution of background emissions among direct natural emissions and re-emissions of natural and anthropogenic mercury. North American anthropogenic sources were calculated to contribute only from 25 to 32% to the total mercury deposition over the continental United States. At selected receptors, the contribution of North American anthropogenic emissions ranges from 9 to 81%; Asian anthropogenic emissions were calculated to contribute from 5 to 36%; natural emissions were calculated to contribute from 6 to 59%.

Application and evaluation of two air quality models for particulate matter for a southeastern U.S. episode.

The Models-3 Community Multiscale Air Quality (CMAQ) Modeling System and the Particulate Matter Comprehensive Air Quality Model with extensions (PMCAMx) were applied to simulate the period June 29-July 10, 1999, of the Southern Oxidants Study episode with two nested horizontal grid sizes: a coarse resolution of 32 km and a fine resolution of 8 km. The predicted spatial variations of ozone (O3), particulate matter with an aerodynamic diameter less than or equal to 2.5 microm (PM2.5), and particulate matter with an aerodynamic diameter less than or equal to 10 microm (PM10) by both models are similar in rural areas but differ from one another significantly over some urban/suburban areas in the eastern and southern United States, where PMCAMx tends to predict higher values of O3 and PM than CMAQ. Both models tend to predict O3 values that are higher than those observed. For observed O3 values above 60 ppb, O3 performance meets the U.S. Environmental Protection Agency's criteria for CMAQ with both grids and for PMCAMx with the fine grid only. It becomes unsatisfactory for PMCAMx and marginally satisfactory for CMAQ for observed O3 values above 40 ppb. Both models predict similar amounts of sulfate (SO4(2-)) and organic matter, and both predict SO4(2-) to be the largest contributor to PM2.5. PMCAMx generally predicts higher amounts of ammonium (NH4+), nitrate (NO3-), and black carbon (BC) than does CMAQ. PM performance for CMAQ is generally consistent with that of other PM models, whereas PMCAMx predicts higher concentrations of NO3-, NH4+, and BC than observed, which degrades its performance. For PM10 and PM2.5 predictions over the southeastern U.S. domain, the ranges of mean normalized gross errors (MNGEs) and mean normalized bias are 37-43% and -33-4% for CMAQ and 50-59% and 7-30% for PMCAMx. Both models predict the largest MNGEs for NO3- (98-104% for CMAQ 138-338% for PMCAMx). The inaccurate NO3- predictions by both models may be caused by the inaccuracies in the ammonia emission inventory and the uncertainties in the gas/particle partitioning under some conditions. In addition to these uncertainties, the significant PM overpredictions by PMCAMx may be attributed to the lack of wet removal for PM and a likely underprediction in the vertical mixing during the daytime.

Regional modeling of the atmospheric fate and transport of benzene and diesel particles.

The Community Multiscale Air Quality model (CMAQ) was modified to simulate the atmospheric fate and transport of benzene and diesel particles. We simulated the July 11-15, 1995 period over a domain covering the eastern United States with a 12-km horizontal resolution and a finer (4 km) resolution over a part of the northeastern United States that includes Washington, DC and New York City. The meteorological fields were obtained from a simulation conducted earlier with the mesoscale model MM5. Gridded emission files for benzene and diesel particles were developed using the SMOKE modeling system. The results of the model simulations showed that benzene concentrations were commensurate with available measurements. Over the 4-km resolution domain, a comparison between simulated and measured 24-h average concentrations showed a fractional error of 0.46, a fractional bias of 0.14, and a coefficient of determination (r2) of 0.25. A comparison between simulated benzene hourly concentrations in New York City and in the Brigantine Wilderness Area, NJ, showed that urban concentrations were greater than the remote area concentrations by a factor of 2-5. The results of the diesel particle simulations showed spatial and temporal patterns that were similar to those obtained for benzene. However, because of the lesser contribution of on-road mobile sources to diesel particle emissions compared to benzene emissions, diesel particle concentrations showed stronger gradients between urban areas and remote areas. A comparison between diesel particle concentrations in New York City and in the Brigantine Wilderness Area, NJ, showed that the urban concentrations were greater than the remote area concentrations by a factor of 2-10. Assuming that diesel particles consist of 50% "elemental" carbon (EC), the simulated EC concentrations were in close agreement (within 10%) with the measured concentration in the urban area (Washington, DC) but were significantly lower than the measured EC concentrations in the remote area (Brigantine Wilderness Area). This result suggests that other sources beside diesel fuel engines contribute to atmospheric EC concentrations and that EC may not be a good surrogate for diesel particles. A comparison of both benzene and diesel particle simulated concentrations between an urban area (New York City) and a remote area (Brigantine Wilderness Area) shows that, at a spatial resolution of 4 km, the regional background may contribute from 10 to 20% to the peak concentrations. These results suggest that the regional background may not be negligible and should be taken into account in urban air toxics studies.

Uncertainties in modeling secondary organic aerosols: three-dimensional modeling studies in Nashville/western Tennessee.

The formation of secondary organic aerosols (SOA) is simulated for the Nashville/western Tennessee domain using three recent SOA modules incorporated into the three-dimensional air quality model, CMAQ. The Odum/Griffin et al. and CMU/STI modules represent SOA absorptive partitioning into a mixture of primary and secondary particulate organic compounds (OC), with some differences in the formulation of the absorption process and the selection of SOA species and their precursors. Empirical representations based on measured laboratory SOA yields are used for condensable organic products in both these modules. The AEC module simulates SOA absorption into organic and aqueous particulate phases, and a representation based on an explicit gas-phase mechanism is used in the AEC module. Predicted SOA concentrations can vary by a factor of 10 or more. In general, the gas-phase mechanistic approach predicts a higher yield of SOA than those based on laboratory yields. There exist some differences in the two empirical modules despite their similar basis on experimental data. All three modules predict a dominance of SOA of biogenic origin as compared to SOA of anthropogenic origin. The causes for differences among the three SOA modules include the representation of terpenes, the mechanistic versus empirical representation of SOA-forming reactions, the identities of SOA, and the parameters used in the gas/particle partitioning calculations. Two sensitivity studies show that formation of water-soluble SOA and temperature dependence may be areas of key uncertainties affecting current models.

Day-of-Week behavior of atmospheric ozone in three U.S. cities.

The weekly cycles of atmospheric ozone (O3) are of interest because they provide information about the response of O3 to changes in anthropogenic emissions from weekdays to weekends. The weekly behavior of O3 in Chicago, IL; Philadelphia, PA; and Atlanta, GA, is contrasted. In Chicago and Philadelphia, maximum 1-hr average O3 increases on weekends. In Atlanta, O3 builds up from Mondays to Fridays and declines during weekends. In all three areas, volatile organic compound (VOC)/nitrogen oxides (NOx) ratios are higher during weekends, resulting from greater than proportionate decreases in NOx relative to VOC emissions. The VOC/NOx ratios correlate with maximum 1-hr O3 concentrations in Chicago, a response consistent with a VOC-sensitive airshed. A weak correlation between O3 concentrations and VOC/NOx ratios in Philadelphia suggests the impact of transported O3, which is formed in upwind VOC-sensitive locations that may be hundreds of kilometers away. Ozone concentrations in Atlanta do not correlate with VOC/NOx ratios but with concentrations of NOx and total reactive nitrogen (NOy) carried over from the previous day. When data from 1986-1990 and 1995-1999 are compared, only small differences in the weekly behavior of O3 are observed in Chicago and Philadelphia. The day-of-week differences in O3 are amplified in the more recent period in Atlanta, a possible result of urban growth.

Contributions of global and regional sources to mercury deposition in New York State.

A modeling system that includes a global chemical transport model (CTM) and a nested continental CTM (TEAM) was used to simulate the atmospheric transport, transformations and deposition of mercury (Hg). Three scenarios were used: (1) a nominal scenario, (2) a scenario conducive to local deposition and (3) a scenario conducive to long-range transport. Deposition fluxes of Hg were analyzed at three receptor locations in New York State. For the nominal scenario, the anthropogenic emission sources (including re-emission of deposited Hg) in New York State, the rest of the contiguous United States, Asia, Europe, and Canada contributed 11-1, 25-9, 13-19, 5-7, and 2-5%, respectively to total Hg deposition at these three receptors. Natural sources contributed 16-4%. The results from the local deposition and long-range transport scenarios varied only slightly from these results. However, there are still uncertainties in our understanding of the atmospheric chemistry of Hg that are likely to affect these estimates of local, regional and global contributions. Comparison of model simulation results with data from the Mercury Deposition Network suggests that local and regional contributions may currently be overestimated.

On the effect of spatial resolution on atmospheric mercury modeling.

Mathematical modeling of the atmospheric fate and transport of mercury (Hg) was conducted using three nested domains covering global, continental and regional scales with horizontal resolutions of approximately 1000, 100 and 20 km, respectively. Comparisons of modeling results with wet deposition fluxes show a coefficient of determination (r(2)) of 0.45 for the continental simulation and 0.14 for the continental/regional simulation. The poor correlation obtained in the regional simulation results to a large extent from the fact that the model predicts an increasing gradient in Hg wet deposition from Minnesota to Pennsylvania, which is not observed in the monitoring network. The use of a finer spatial resolution (20 km) improves model performance in Minnesota and Wisconsin (upwind of major Hg emission sources) but degrades model performance in Pennsylvania (downwind of major Hg emission sources). We suggest the hypothesis that some key Hg chemical transformations are likely missing in current models of atmospheric Hg.