Propagation of uncertainty in hourly utility NOx emissions through a photochemical grid air quality model: a case study for the Charlotte, NC, modeling domain

One of the major hypothesized sources of uncertainties in air quality model inputs is the emission inventory. A probabilistic hourly NOx emission inventory for 32 units of nine coal-fired power plants in the Charlotte domain for the year 1995 was propagated through the Multiscale Air Quality Simulation Platform (MAQSIP). The inventory was developed using time series techniques. Time series for a 4-d episode were simulated and propagated through the air quality model 50 times in order to represent the ranges of uncertainty in hourly emissions and predicted ozone levels. Intra-unit autocorrelation in emissions and inter-unit dependence were accounted for. The range of uncertainty in predicted ozone was greater when inter-unit dependence was included as compared to when units were treated as statistically independent. Uncertainties in maximum ozone hourly or 8-h concentrations at a specific location could be attributed to a specific power plant based upon regression analysis. Out of 3969 grid cells in the modeling domain, there were 43 and 1654 grid cells with a probability greater than 0.9 of exceeding a 1-h 120 ppb standard and an 8-h 80 ppb standard, respectively. The time series of predicted ozone values had similar autocorrelation as compared to monitored data. The implications of these results for air quality management are addressed.     

Atmospheric chlorine chemistry in southeast Texas: impacts on ozone formation and control

Recent evidence has demonstrated that chlorine radical chemistry can enhance tropospheric hydrocarbon oxidation and has the potential to enhance ozone formation in urban atmospheres. To assess these effects quantitatively, an August-September 2000 photochemical episode in southeast Texas was simulated using the comprehensive air quality model, with extensions (CAMx). During this episode, ambient measurements of a unique marker of atmospheric chlorine chemistry, 1-chloro-3-methyl-3butene-2-one (CMBO), were made and model performance was assessed by comparing modeled and observed CMBO mixing ratios. The model predicted ambient CMBO mixing ratios within the uncertainty limits associated with the emissions inventory, so the model was used to assess the impacts of chlorine chemistry on ozone formation. Based on the current emissions inventory, chlorine emissions have the potential to enhance 1-h-averaged ozone mixing ratios by 70 ppb, in very localized areas, during morning hours. Over wider areas, and at times of day when peak ozone concentrations are observed, the impacts of chlorine emissions on ozone concentrations are typically less than 10 ppb. Chlorine emissions also influenced changes in ozone concentrations due to hydrocarbon and NOx emission controls.     

Nonlinear response of ozone to emissions: source apportionment and sensitivity analysis

For secondary air pollutants, precursor emissions may impact concentrations in nonlinear and interdependent manners. We explore the nonlinear responses of one such pollutant, ozone, to emissions of its precursors, nitrogen oxides (NOx) and volatile organic compounds. Modeling is conducted for a high ozone episode in the southeastern United States, applying a second-order direct sensitivity method in a regional air quality model. As applied here, the sensitivity method neglects most aerosol and aqueous chemistry processes. Inclusion of second-order sensitivities is shown to enable accurate characterization of response to large perturbations in emissions. An index is introduced to characterize the nonlinearity of ozone response to NOx emitted from each source region. Nonlinearity is found to increase with the tonnage and emission density of the source region. Interactions among the impacts of emission sources are shown to lead to discrepancies between source contribution attributed to an ensemble of emitters and the sum of the contributions attributed to each component. A method is introduced for applying these "cross-sensitivity" interactions to assess the uncertainty of sensitivity and source apportionment estimates arising from uncertainty in an emissions inventory. For ozone response to NOx, underestimates in emission rates lead to underprediction of total source contribution but overprediction of per-ton sensitivity.     

Influence of population density and temporal variations in emissions on the air duality benefits of NOx emission trading

Ozone formation is a complex function of local hydrocarbon and nitrogen oxide emissions. Therefore, trading of NOx emissions among geographically distributed facilities can lead to more or less ozone formation than across-the-board reductions. Monte Carlo simulations of trading scenarios involving 51 large NOx point sources in eastern Texas were used in a previous study by the authors to assess the effects of trading on air quality benefits, as measured by changes in ozone concentrations. The results indicated that 12% of trading scenarios would lead to greater than a 25% variation from conventional across-the-board reductions when air quality benefits are based only on changes in ozone concentration. The current study found that when benefits are based on a metric related to population exposure to ozone, two-thirds of the trading scenarios lead to changes in air quality benefits of approximately 25%. Variability in air quality benefits is not as strongly dependent on the temporal distribution of NOx emissions.     

Photochemical modeling of emissions trading of highly reactive volatile organic compounds in Houston, Texas. 2. Incorporation of chlorine emissions

As part of the State Implementation Plan for attaining the National Ambient Air Quality Standard for ozone, the Texas Commission of Environmental Quality has created a Highly Reactive Volatile Organic Compounds (HRVOC) Emissions Cap and Trade Program for industrial point sources in the Houston/Galveston/Brazoria area. This series of papers examines the potential air quality impacts of this new emission trading program through photochemical modeling of potential trading scenarios; this paper examines the air quality impact of allowing facilities to trade chlorine emission reductions for HRVOC allocations on a reactivity weighted basis. The simulations indicate that trading of anthropogenic chlorine emission reductions for HRVOC allowances at a single facility or between facilities, in general, resulted in improvements in air quality. Decreases in peak 1-h averaged and 8-h averaged ozone concentrations associated with trading chlorine emissions for HRVOC allocations on a Maximum Incremental Reactivity (MIR) basis were up to 0.74 ppb (0.63%) and 0.56 ppb (0.61%), respectively. Air quality metrics based on population exposure decreased by up to 3.3% and 4.1% for 1-h and 8-h averaged concentrations. These changes are small compared to the maximum changes in ozone concentrations due to the VOC emissions from these sources (5-10 ppb for 8-h averages; up to 30 ppb for 1-h averages) and the chlorine emissions from the sources (5-10 ppb for maximum concentrations over wide areas and up to 70 ppb in localized areas). The simulations indicate that the inclusion of chlorine emissions in the trading program is likely to be beneficial to air quality and is unlikely to cause localized increases in ozone concentrations ("hot spots").     

Ammonia emission controls as a cost-effective strategy for reducing atmospheric particulate matter in the Eastern United States

Current regulation aimed at reducing inorganic atmospheric fine particulate matter (PM2.5) is focused on reductions in sulfur dioxide (SO2) and oxides of nitrogen (NO(x) = NO + NO2); however, controls on these pollutants are likely to increase in cost and decrease in effectiveness in the future. A supplementary strategy is reduction in ammonia (NH3) emissions, yet an evaluation of controls on ammonia has been limited by uncertainties in emission levels and in the cost of control technologies. We use state of the science emission inventories, an emission-based regional air quality model, and an explicit treatment of uncertainty to estimate the cost-effectiveness and uncertainty of ammonia emission reductions on inorganic particulate matter in the Eastern United States. Since a paucity of data on agricultural operations precludes a direct calculation of the costs of ammonia control, we calculate the "ammonia savings potential", defined as the minimum cost of applying SO2 and NO(x) emission controls in order to achieve the same reduction in ambient inorganic PM2.5 concentration as obtained from a 1 ton decrease in ammonia emissions. Using 250 scenarios of NH3, SO2, and NO(x) emission reductions, we calculate the least-cost SO2 and NO(x) control scenarios that achieve the same reduction in ambient inorganic PM2.5 concentration as a decrease in ammonia emissions. We find that the lower-bound ammonia savings potential in the winter is $8,000 per ton NH3; therefore, many currently available ammonia control technologies are cost-effective compared to current controls on SO2 and NO(x) sources. Larger reductions in winter inorganic particulate matter are available at lower cost through controls on ammonia emissions.     

Likelihood of achieving air quality targets under model uncertainties

Regulatory attainment demonstrations in the United States typically apply a bright-line test to predict whether a control strategy is sufficient to attain an air quality standard. Photochemical models are the best tools available to project future pollutant levels and are a critical part of regulatory attainment demonstrations. However, because photochemical models are uncertain and future meteorology is unknowable, future pollutant levels cannot be predicted perfectly and attainment cannot be guaranteed. This paper introduces a computationally efficient methodology for estimating the likelihood that an emission control strategy will achieve an air quality objective in light of uncertainties in photochemical model input parameters (e.g., uncertain emission and reaction rates, deposition velocities, and boundary conditions). The method incorporates Monte Carlo simulations of a reduced form model representing pollutant-precursor response under parametric uncertainty to probabilistically predict the improvement in air quality due to emission control. The method is applied to recent 8-h ozone attainment modeling for Atlanta, Georgia, to assess the likelihood that additional controls would achieve fixed (well-defined) or flexible (due to meteorological variability and uncertain emission trends) targets of air pollution reduction. The results show that in certain instances ranking of the predicted effectiveness of control strategies may differ between probabilistic and deterministic analyses.     

Reformulated and alternative fuels: modeled impacts on regional air quality with special emphasis on surface ozone concentration

The comprehensive European Air Pollution and Dispersion model system was used to estimate the impacts of the usage of reformulated and alternative fuels on regional air quality with special emphasis on surface ozone concentrations. A severe western European summer smog episode in July 1994 has been used as a reference, and the model predictions have been evaluated for this episode. A forecast simulation for the year 2005 (TREND) has been performed, including the future emission development based on the current legislation and technologies available. The results of the scenario TREND are used as a baseline for the other 2005 fuel scenarios, including fuel reformulation, fuel sulfur content, and compressed natural gas (CNG) as an alternative fuel. Compared to the year 1994, significant reductions in episode peak ozone concentrations and ozone grid hours are predicted for the TREND scenario. These reductions are even more pronounced within the investigated alternative and reformulated fuel scenarios. Especially, low sulfur fuels are appropriate for an immediate improvement in air quality, because they effect the emissions of the whole fleet. Furthermore, the simulation results indicate that the introduction of CNG vehicles would also enhance air quality with respect to ozone.     

Photochemical modeling of emissions trading of highly reactive volatile organic compounds in Houston, Texas. 1. Reactivity based trading and potential for ozone hot spot formation

As part of the State Implementation Plan for attaining the National Ambient Air Quality Standard for ozone, the Texas Commission of Environmental Quality has created a Highly Reactive Volatile Organic Compounds (HRVOC) Emissions Cap and Trade Program for industrial point sources in the Houston/Galveston/Brazoria area. This program has a number of unique features, including its focus on a limited group of ozone precursors and its provisions for trading emissions based on atmospheric reactivity. This series of papers examines the potential air quality impacts of this new emission trading program through photochemical modeling of potential trading scenarios; this first paper in the series describes the air quality modeling methods used to assess potential trades, the potential for localized increases in ozone concentrations (ozone "hot spots") due to HRVOC emission trading, and the use of reactivity scales in the trading. When HRVOC emissions are traded on a mass basis, the simulations indicate that trading of HRVOC allowances between facilities resulted in less than 0.15 ppb (<0.13%) and 0.06 ppb (<0.06%) increases in predicted maximum, area-wide 1-h averaged and 8-h averaged ozone concentrations, respectively. Maximum decreases in ozone concentrations associated with trading, as opposed to across-the-board reductions, were larger than the increases. All of these changes are small compared to the maximum changes in ozone concentrations due to the VOC emissions from these sources (up to 5-10 ppb for 8 h averages; up to 30 ppb for 1-h averages). When emissions of HRVOCs are traded for other, less reactive emissions, on a reactivity weighted basis, air quality simulations indicate that daily maximum ozone concentrations increased by less than 0.3%. Because these relatively small changes (< 1%) are for unlikely trading scenarios designed to produce a maximum change in ozone concentrations (all emissions traded into localized regions), the simulations indicate that the implementation of the trading program, as currently configured and possibly expanded, is unlikely to cause localized increases in ozone concentrations ("hot spots").     

Contributions of natural emissions to ozone and PM2.5 as simulated by the Community Multiscale Air Quality (CMAQ) model

The relative roles of natural and anthropogenic sources in determining ozone and fine particle concentrations over the continental United States (U.S.) are investigated using an expanded emissions inventory of natural sources and an updated version of the Community Multiscale Air Quality (CMAQ) model. Various 12-month CMAQ simulations for the year 2002 using different sets of input emissions data are combined to delineate the contributions of background pollutants (i.e., model boundary conditions), natural emissions, anthropogenic emissions, as well as the specific impacts of lightning and wildfires. Results are compared with observations and previous air quality model simulations. Wildfires and lightning are both identified as contributing significantly to ozone levels with lightning NO(x) adding as much as 25-30 ppbV (or up to about 50%) to surface 8-h average natural O(3) mixing ratios in the southeastern U.S. Simulated wildfire emissions added more than 50 ppbV (in some cases >90%) to 8-h natural O(3) at several locations in the west. Modeling also indicates that natural emissions (including biogenic, oceanic, geogenic and fires) contributed ≤ 40% to the annual average of total simulated fine particle mass over the eastern two-thirds of the U.S. and >40% across most of the western U.S. Biogenic emissions are the dominant source of particulate mass over the entire U.S. and wildfire emissions are secondary. Averaged over the entire modeling domain, background and natural ozone are dominant with anthropogenically derived ozone contributing up to a third of the total only during summer. Background contributions to fine particle levels are relatively insignificant in comparison. Model results are also contrasted with the U.S. Environmental Protection Agency (EPA) default values for natural light scattering particle concentrations to be used for regional haze regulatory decision-making. Regional differences in EPA guidance are not supported by the modeling and EPA uncertainty estimates for default values are far smaller than the modeled variability in natural particle concentrations.     

Adjoint sensitivity analysis for a three-dimensional photochemical model: application to Southern California

An adjoint method was used to investigate the sensitivity of peak ozone at selected sites in Southern California to nearly 900 model inputs including surface emissions, reaction rate coefficients, dry deposition velocities, boundary conditions, and initial conditions. Simulations showed large changes in ozone and ozone sensitivities at three sites investigated between summers 1987 and 1997 due to emission reductions. However, only small changes in ozone and ozone sensitivities were predicted between 1997 and 2010. Sensitivities of the differences in ozone between simulations with different emission scenarios were calculated and compared to sensitivities of ozone in each simulation. In some cases, the sensitivities of ozone differences were smaller than those of ozone itself, but in other cases, such as when the sensitivityto NOx emissions changed sign, sensitivities of differences were larger. The adjoint method was most useful for determining when and where model inputs affect, or have the potential to affect, an ozone response. For example, the method was used to plot the spatial distribution of important emission source regions to 1-hour versus 8-hour peak ozone. Changes in the distribution and sign of the adjoint function for emitted species revealed changes in the area of influence of pollutant emissions on peak ozone due to emission controls. The adjoint method provides useful information complementary to that obtained from forward sensitivity analysis methods.     

The impacts of urbanization on emissions and air quality: comparison of four visions of Austin, Texas

The impacts of alternative regional development patterns on emissions, dry deposition, and air quality were examined using four visions of future land use in Austin, Texas associated with a doubling of the population in 20-40 years from 2001. Emissions and their spatial allocation were determined based on the development pattern and used to predict hourly ozone concentrations. Differences in hourly ozone concentrations due to changes in anthropogenic emissions between the future case scenarios and a 2007 base case ranged from -14 to 22 ppb and were primarily associated with the implementation of federal mobile source standards; differences due to biogenic emissions and dry deposition due to urbanization ranged from only -1.4 to 0.7 ppb. These differences in the magnitude of emissions produced greater changes in air quality than differences in regional development patterns between the four scenarios. Differences in hourly ozone concentrations between the future development scenarios and a 2007 base case ranged from -14 to 22 ppb, in contrast to differences of -3 to 5 ppb between the future scenarios. The results imply that although the effects of urbanization patterns are non-negligible, the pattern of urban development is not as significant as reductions in emissions per capita.     

Current and future linked responses of ozone and PM2.5 to emission controls

Responses of ozone and PM2.5 to emission changes are coupled because of interactions between their precursors. Here we show the interdependencies of ozone and PM2.5 responses to emission changes in 2001 and 2050, with the future case accounting for both currently planned emission controls and climate change. Current responses of ozone and PM2.5 to emissions are quantified and linked on a daily basis for five cities in the continental United States: Atlanta, Chicago, Houston, Los Angeles, and NewYork. Reductions in anthropogenic NO(x) emissions decrease 24-h average PM2.5 levels but may either increase or decrease daily maximum 8-h average ozone levels. Regional ozone maxima for all the cities are more sensitive to NO(x) reductions than at the city center, particularly in New York and Chicago. Planned controls of anthropogenic NO(x) emissions lead to more positive responses to NO(x) reductions in the future. Sensitivities of ozone and PM2.5 to anthropogenic VOC emissions are predicted to decrease between 2001 and 2050. Ammonium nitrate formation is predicted to be less ammonia-sensitive in 2050 than 2001 while the opposite is true for ammonium sulfate. Sensitivity of PM2.5 to SO2 and NO(x) emissions changes little between 2001 and 2050. Both ammonium sulfate and ammonium nitrate are predicted to decrease in sensitivity to SO2 and NO(x) emissions between 2001 and 2050. The complexities, linkages, and daily changes in the pollutant responses to emission changes suggest that strategies developed to meet specific air quality standards should consider other air quality impacts as well.     

Regional air quality: local and interstate impacts of NO(x) and SO2 emissions on ozone and fine particulate matter in the eastern United States

While the U.S. air quality management system is largely designed and managed on a state level, many critical air quality problems are now recognized as regional. In particular, concentrations of two secondary pollutants, ozone and particulate matter, are often above regulated levels and can be dependent on emissions from upwind states. Here, impacts of statewide emissions on concentrations of local and downwind states' ozone and fine particulate matter are simulated for three seasonal periods in the eastern United States using a regional Eulerian photochemical model. Impacts of ground level NO(x) (e.g., mobile and area sources), elevated NO(x) (e.g., power plants and large industrial sources), and SO2 emissions are examined. An average of 77% of each state's ozone and PM(2.5) concentrations that are sensitive to the emissions evaluated here are found to be caused by emissions from other states. Delaware, Maryland, New Jersey, Virginia, Kentucky, and West Virginia are shown to have high concentrations of ozone and PM(2.5) caused by interstate emissions. When weighted by population, New York receives increased interstate contributions to these pollutants and contributions to ozone from local emissions are generally higher. When accounting for emission rates, combined states from the western side of the modeling domain and individual states such as Illinois, Tennessee, Indiana, Kentucky, and Georgia are major contributors to interstate ozone. Ohio, Indiana, Tennessee, Kentucky, and Illinois are the major contributors to interstate PM(2.5). When accounting for an equivalent mass of emissions, Tennessee, Kentucky, West Virginia, Virginia, and Alabama contribute large fractions of these pollutants to other states.     

Ozone formation potential of organic compounds in the Eastern United States: a comparison of episodes, inventories, and domains

Direct sensitivity analysis is applied for 3-D assessment of ozone reactivity (or ozone formation potential) in the Eastern United States. A detailed chemical mechanism (SAPRC-99) is implemented in a multiscale air quality model to calculate the reactivity of 32 explicit and 9 lumped compounds. Simulations are carried out for two different episodes and two different emission scenarios. While absolute reactivities of VOCs show a great deal of spatial variability, relative reactivities (normalized to the reactivity of a base mixture) produce a significantly more homogeneous field. Three types of domain-wide relative reactivity metrics are formed for 1-h and 8-h averaging intervals. In general, ozone reactivity metrics (with the exception of those based on daily peak ozone) are fairly robust and consistent between different episodes or emission scenarios. The 3-D metrics also show fairly similar rankings for VOC reactivity when compared to the box model scales. However, the 3-D metrics have a noticeably narrower range for species reactivities, as they result in lower reactivity for some of the more reactive, radical-producing VOCs (especially aldehydes). As expected, episodes and emission scenarios with less radical availability have higher absolute reactivities for all species and higher relative reactivities for the more radical-producing species. Finally, comparing the results with those from a different domain (central California) shows that relative reactivity metrics are comparable over these two significantly different domains.     

Adjoint sensitivity analysis of ozone nonattainment over the continental United States

An application of the adjoint method in air quality management is demonstrated. We use a continental scale chemical transport model (STEM) to calculate the sensitivities of a nationwide U.S. ozone national ambient air quality standard (NAAQS) nonattainment metric to precursor emissions for the period July 1 to August 15, 2004. The model shows low bias and error (-4 and 24%, respectively), particularly for areas with high ozone concentrations. The nonattainment metric accounts for both 1-h and 8-h ozone standards, but is dominated by the 8-h exceedances (97% of the combined metric). Largest values of sensitivities are found to be with respect to emissions in the south and southeast U.S., Ohio River Valley, and California. When nonattainment sensitivities are integrated over the entire U.S., NOx emissions account for the largest contribution (62% of the total), followed by biogenic and anthropogenic VOCs (24% and 14%, respectively). For NOx emissions, point/area and mobile sources account for 54% and 46% of the total sensitivities, respectively. We also provide a state-by-state comparison for the nonattainment magnitude, nonattainment sensitivity, and emission magnitudes to explore the influence of interstate transport of ozone and its precursors, and policy implications of the results. Our analysis of the nationwide ozone nonattainment metric suggests that simple cap-and-trade programs may prove inadequate in achieving sought-after air quality objectives.     

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.     

Tracking petroleum refinery emission events using lanthanum and lanthanides as elemental markers for PM2.5

Fine particulate matter levels at four air sampling stations in the Houston, TX area are apportioned to quantify the impact of emissions from a local refinery during a reported emission event. Through quantification of lanthanum and lanthanides using a recently developed analytical technique, the impacts of emissions from fluidized-bed catalytic cracking (FCC) units are quantitatively tracked across the Houston region. The results show a significant (33-106-fold) increase in contributions of FCC emissions to PM2.5 compared with background levels associated with routine operation. This impact from industrial emissions to ambient air quality occurs simultaneously with a larger, regional haze episode that lead to elevated PM2.5 concentrations throughout the entire region. By focusing on detailed chemical analysis of unique maker metals (lanthanum and lanthanides), the impact of emissions from the FCC unit was tracked from the local refinery that reported the emission event to a site approximately 50 km downwind, illustrating the strength of the analytical method to isolate an important source during a regional haze episode not related to the emission event. While this source apportionment technique could separate contributions from FCC emissions, improved time-resolved sampling is proposed to more precisely quantify the impacts of transient emission events on ambient PM2.5.     

The global atmospheric environment for the next generation

Air quality, ecosystem exposure to nitrogen deposition, and climate change are intimately coupled problems: we assess changes in the global atmospheric environment between 2000 and 2030 using 26 state-of-the-art global atmospheric chemistry models and three different emissions scenarios. The first (CLE) scenario reflects implementation of current air quality legislation around the world, while the second (MFR) represents a more optimistic case in which all currently feasible technologies are applied to achieve maximum emission reductions. We contrast these scenarios with the more pessimistic IPCC SRES A2 scenario. Ensemble simulations for the year 2000 are consistent among models and show a reasonable agreement with surface ozone, wet deposition, and NO2 satellite observations. Large parts of the world are currently exposed to high ozone concentrations and high deposition of nitrogen to ecosystems. By 2030, global surface ozone is calculated to increase globally by 1.5 +/- 1.2 ppb (CLE) and 4.3 +/- 2.2 ppb (A2), using the ensemble mean model results and associated +/-1 sigma standard deviations. Only the progressive MFR scenario will reduce ozone, by -2.3 +/- 1.1 ppb. Climate change is expected to modify surface ozone by -0.8 +/- 0.6 ppb, with larger decreases over sea than over land. Radiative forcing by ozone increases by 63 +/- 15 and 155 +/- 37 mW m(-2) for CLE and A2, respectively, and decreases by -45 +/- 15 mW m(-2) for MFR. We compute that at present 10.1% of the global natural terrestrial ecosystems are exposed to nitrogen deposition above a critical load of 1 g N m(-2) yr(-1). These percentages increase by 2030 to 15.8% (CLE), 10.5% (MFR), and 25% (A2). This study shows the importance of enforcing current worldwide air quality legislation and the major benefits of going further. Nonattainment of these air quality policy objectives, such as expressed by the SRES-A2 scenario, would further degrade the global atmospheric environment.