Effects of tropospheric emissions on global tropospheric ozone distribution : A CTM simulation study

In this work, we examined the effect of tropospheric emissions on tropospheric 1 ozone (O3) by conducting three-dimensional (3D) chemistry transport model (CTM) 2 simulations. For the control run, the CTM model simulates tropospheric O3 levels with 3 a complete set of anthropogenic, biomass burning, and vegetation emissions [8]. For the 4 no-emission simulation, all anthropogenic, biomass burning, and vegetation emissions were 5 turned off. Comparisons of results from these two simulations exhibit the emission impacts 6 on the tropospheric O3. In the no-emission simulation, distinctive low surface O3 with 7 concentrations less than 5 ppbv prevail over the Amazon basin, tropical South America, 8 tropical South Africa, Southeast Asia. Transport of air from these land areas downwind 9 contributes to the low O3 over the remote marine boundary layer. In contrast, elevated 10 O3 levels over the extra-tropical remote marine boundary layer are less supported by the 11 anthropogenic and biomass burning emissions but more sustained by the downward transport 12 of O3 from the stratosphere. These results demonstrate that the northern hemisphere 13 continental areas (north of 30◦N), polar regions, and tropical continental regions are more 14 sensitive to the tropospheric emissions. The northern hemisphere winter is mostly dominated 15 by the stratospheric processes, while the tropospheric emissions dominate over the southern 16 hemisphere tropical continental areas from tropics to 30◦S latitudinal bands. The northern 17 hemisphere continental regions are increasingly dominated by tropospheric emissions from 18 spring, to reach maxima in summer, and started to reduce in autumn months. 19


Introduction
Atmospheric O 3 exhibit a near order two of magnitude change from the upper troposphere to the lower stratosphere.The omnipresence of the steep vertical gradient in O 3 and active stratospheric-tropospheric exchange process affect O 3 in the troposphere on a seasonal and global scales [1].
On the other hand, tropospheric emissions exert a direct impact on the tropospheric O 3 concentrations.
The emissions from man-made industrial sources such as nitrogen oxides (NOx), hydrocarbons, carbon monoxide (CO), biomass burning activities, and vegetation emissions [2;3;4;5;6;7;8] are the main sources of primary emissions for the photochemical productions of O 3 in the troposphere.These emissions are local in nature, mainly concentrated in the continental areas of both hemispheres.
How are these two effects, one from above the troposphere and the other one from below the troposphere, compete to dominate local tropospheric O 3 on a seasonal and global scale?[9;10].
There are two views regarding the source of the photochemical processes in the troposphere contributing to the tropospheric O 3 concentrations.First, wintertime accumulation of O 3 producing precursors over the continental industrial areas where anthropogenic emissions have a longer chemical lifetime than other seasons.Also, the major sinks for O 3 , species such as hydrogen oxides (HOx) and water (H 2 O), are low during the winter, hence more O 3 can accumulate in high latitudes than other seasons.The accumulation of O 3 precursors and O 3 from the stratosphere give rise to the northern hemisphere springtime O 3 maximum [11;12;7].Also, long-range transport of continental pollutants provides another source of elevated O 3 and other pollution over the remote marine boundary layer [13;12;14].
Second, the timing of the spring O 3 maximum coincides with the period of the most active biomass burning in the tropical and subtropical regions.Biomass burning emissions emit O 3 producing precursors.Long-range transport and cloud convective transport processes transport these O 3 precursors to the remote atmosphere.These are the major driving force of the southern hemisphere spring O 3 maximum [15;16;17;18;7].Two approaches are used to access the impact of the photochemical production of O 3 in the troposphere.First, the stratospheric tracer approach.By using a marked stratospheric O 3 tracer, the difference between the O 3 concentrations from full chemistry integrations and the O 3 from stratospheric tracer is calculated as the O 3 that originates from photochemistry in the troposphere [19;20;12;7;5].
Roelofs et al. [19] calculated that 30% of O 3 in the troposphere derived from anthropogenic emission of O 3 producing precursors and natural photochemical production, respectively, and the stratosphere contributed about 40% of O 3 in the troposphere.
Second, the emission approach.Lelieveld and Dentener [7] experimented various emission scenarios to access present day polluted O 3 levels to that of O 3 from a preindustrial period.Lelieveld and Dentener [7] estimated that industrial and fossil-fuel related emissions strongly affected tropospheric O 3 in the extra-tropical northern hemisphere, while natural emissions play a significant role in the tropics and southern hemisphere.They calculated that man-made biomass burning emissions contribute 10% to 15% to the tropospheric ozone column in the tropical latitudes.
In this work, we present a study on the impact of tropospheric O 3 from tropospheric emissions.We conducted and compare results from simulations based on various scenarios of tropospheric emissions and stratospheric influences against control run.These comparisons reveal the temporal and spatial distribution of tropospheric effects and stratospheric effect on tropospheric O 3 concentrations on a global scale.

The IMS CTM Model
The Integrated Modeling System (IMS) three-dimensional (3D) tropospheric chemistry transport model (CTM) is used in this work [8;21;22].The model uses a general gas-phase reaction mechanism for NOx, methane, non-methane hydrocarbon (NMHC), biogenic volatile organic compounds (VOCs), and some sulfur and halogen reactions [22].The IMS model used specified surface emissions for anthropogenic sources from Emission Database for Global Atmospheric Research (EDGAR) and Global Emission InitiAtive (GEIA) [22].The model was driven by analyzed winds from European Centre for Medium-Range Weather Forecast (ECMWF), and it contains 19 vertical layers which extend from surface to about 10 hPa.The horizontal resolution for the model is 7.5 by 4.5-degree longitude-latitude.
The concentrations of O 3 , reactive nitrogen (NOy), and nitric acid (HNO 3 ) in the stratosphere are prescribed similarly to the method of Berntsen and Isaksen [23].

List of CTM Simulation Experiments
To study the effect of tropospheric emissions on the levels of O 3 in the troposphere, we conduct simulations with various emission scenarios as shown in Table 1.
The industrial emissions include emissions of CO, hydrocarbons, and NOx.We note that the NOx emissions include fossil fuel-related emissions, lightning, and soil process.Hence, there is no major O 3 photochemical sources (e.g., biomass burning, fossil fuel burning, lightning, soil process) included in the simulation.Biogenic sources paly a vital role in affecting tropospheric O 3 via the emissions of biogenic hydrocarbons and NOx [25;18;22].
For comparison of the stratospheric effects, we performed simulations with different stratospheric fluxes on tropospheric O 3 concentrations.We test the stratospheric influxes of O 3 and NOy [26] on O 3 levels in the troposphere.

Simulation Setup and Platform
The IMS model run two years of simulations for each scenario shown in Table 1.The model used the 6-hourly meteorological analysis from the ECMWF analysis for the period 1991-1992, and output The Cray J90 is a shared memory supercomputer, which enables codes to run on an ensemble of central processing unit (CPU) simultaneously with greater efficiency than traditional single CPU simulation method [24].We specifically designed the IMS codes to match and benefit from the unique hardware and software mechanisms provided by the Cray J90.As such, the model run parallelly on the Cray J90 since 1995.

Results Analysis
We use O 3 ratios to exam the effect of a scenario simulation.For example, we divided tropospheric O 3 from a particular situation to the tropospheric O 3 from control run (the full model simulation).Hence, the O 3 ratios measure the effects that particular scenario.
For the scenario without tropospheric emissions, the variations of the O 3 ratios indicate the effect of O 3 in the troposphere from primary photochemical sources within the troposphere.An O 3 ratio with unity indicates that tropospheric photochemical sources produce no impact at all on the levels of O 3 in the troposphere.The smaller the O 3 ratios, the bigger the impact derived from the emission sources in the troposphere.
The comparisons were made at the surface using a time-series plot, and above the surface using a series of the zonal mean cross-sections for different seasons.We also compare different model simulations with surface and ozonesonde measurements to highlight the spatial and temporal distribution of the impact of O 3 from combustion-derived photochemical sources in the troposphere.There are agreements and discrepancies in the levels of O 3 simulated between these two simulations.

Comparison of Model with Surface Measurements
Close agreement between the simulations indicates that anthropogenic and pyrogenic emissions are of less importance, while the disagreements between the two simulations provide a measure of the contribution from photochemical production on O 3 in the troposphere.For example, the no-emission simulation at Westman Island shows close agreement between the model and the daily minimum measurements (Figure 1a).This indicates that background O 3 concentrations from the stratosphere mostly controls the variation of the daily observed lowest O 3 levels at this location.
The disagreements between the two model simulations occur from mid-spring to late summer.The timing of the maximum discrepancies occur in summer is consistent with the period when photochemical production in the troposphere is active.These characteristics between the two model calculations, close agreement in winter and differences in spring to summer, also occur at other locations in Figure 1.
Notice that in the no-emission simulation the O 3 maximum peaks in spring, instead of winter [20].
Notice also that the timing of peak O 3 maximum in the simulations occurs during April to June, slightly lags the no-emission simulation at these two sites (Westman Island, Figure 1a; and Bermuda, Figure 1b).
These delays in the timing of peak spring O 3 maximum, due to anthropogenic emissions in the northern hemisphere industrial latitudes, can be used as a measure for assessing the impact of future NOx emission scenarios [7].
At Mauna Loa (Figure 1c), the no-emission model follows closely but slightly lower than the O 3 concentrations simulated when emissions were included.The higher O 3 calculated from the no-emission model compared with the observed daily minimum values indicates that the model overestimates O 3 from the stratosphere to the troposphere at this location.At Samoa (Figure 1d), the model without combustion emissions included follows closely the observed daily maximum O 3 levels.Additional O 3 production from photochemical processes, e.g., due to pyrogenic emissions from tropical continents, further enhances O 3 levels in the model at this site.
Notice that the higher surface O 3 from the no-emission calculation in late August (late winter in the southern hemisphere) than the emission included simulation.This phenomenon is due to the lower net O 3 photochemical loss in the former (without emissions) than the later (with emissions) simulations during this short period at this site.This stratospheric influence supported the inter-hemispheric asymmetry in O 3 even if there are no combustion emissions included in the model.

Meridional Cross-Section
Figure 4 compares the zonal mean cross-section of the O 3 ratios, simulated without to those with emissions, in four seasons.During the northern hemisphere summer (Figure 4b), zonal mean O 3 ratios less than 60% − 65% can be seen in the tropical lower troposphere and the northern hemisphere lower troposphere at latitudes north of 30 • N, respectively.While low O 3 ratios in tropical regions indicate the influence of O 3 production from biomass burning emissions, low O 3 ratios in the mid to high latitudes illustrates the importance of photochemical O 3 production derived from anthropogenic emissions.These summertime low O 3 ratios in the northern hemisphere lower troposphere gradually increase as the season moves through to autumn (Figure 4c) to reach their highest values (≥ 85% − 95%)in the northern hemisphere winter troposphere (Figure 4d).The O 3 ratios greater than 95% in the latitudes north of 30 • N and throughout the troposphere indicate the dominant contribution of O 3 from the stratosphere to the troposphere during the northern hemisphere winter [19].
Similar seasonal movements occur for the O 3 ratios in the southern hemisphere through summer (Figure 4d) to winter (Figure 4b).The vertical extent of the stratospheric dominance in the austral winter hemisphere is about 95% at altitudes between the tropopause and the mid-troposphere (500 hPa), and about 85% − 95% below the mid troposphere.A smaller stratospheric dominance in the austral winter lower troposphere compared with the boreal winter lower troposphere is consistent with less great winter mass influx from the stratosphere to the troposphere in the southern hemisphere than in northern hemisphere [27;6].
Hence, the stratosphere dominated the tropospheric O 3 in the winter hemisphere, while a mixture of both photochemical O 3 production and O 3 from the stratosphere controlled O 3 in the summer hemisphere.
As the northern hemisphere enters its spring season, Figure 4a shows a symmetry of the O 3 ratios between northern hemisphere and southern hemisphere.In the lowest part of the troposphere and at latitudes between 30 • N and 45 • N where low O 3 ratio indicates the gradual development of photochemical O 3 production from the low troposphere as the sun gradually returns to the northern hemisphere.In the meantime, the full photochemical production in the southern hemisphere summer has slowly faded away as the southern hemisphere enters its autumn.Therefore a longer-lived O 3 accumulated in the previous winter plus the return of the photochemical processes produce a tropospheric O 3 maximum in spring [20].
For the southern hemisphere spring (Figure 4c), the seasonal pyrogenic emissions of O 3 precursors in the southern hemisphere tropical continents fueled photochemical production.at Bermuda (Figure 5a).These characters are similar to the measurements (Figure 5b).Vertical distribution of the O 3 ratios at this site (Figure 5c) show that about 85%−95% of the northern hemisphere spring O 3 maximum can be reproduced by the model without considering combustion emission of O 3 precursors from continental areas [28;31].Photochemical O 3 production becomes important in the lower troposphere during late spring to late summer seasons.

The Atlantic Basin
At Irene, the no-emission model shows a distinctive O 3 maximum in the troposphere during the southern hemisphere winter to spring seasons (Figure 5d).These patterns are similar to that of the measurements (Figure 5e).About 85% − 95% of O 3 in the troposphere at this site during the southern hemisphere winter and early spring can be maintained by the downward transport of O 3 from the stratosphere (Figure 5f).Low O 3 ratios (less than 80%) occur at this site during the southern hemisphere spring to summer months, from October to December and from January to March (Figure 5f).
These low O 3 ratios indicate that pyrogenic emissions play a major role in the lower troposphere.
These results are consistent with previous analyses.Transport of photochemically produced O 3 , due to biomass burning emissions and/or O 3 producing precursors from continents, is the major factor for elevated O 3 observed at this region [2;3;18].The no-emission model at Tahiti (Figure 6d) shows downward transport of O 3 from the stratosphere during the southern hemisphere winter.These modeled results are comparable with the measurements (Figure 6e).Ratios of O 3 close to unity at this site indicate that photochemical production in the troposphere is not a significant contributor to O 3 at this site (Figure 6f).

Northern Hemisphere Higher Latitudes
For locations at higher latitudes, the non-emission model shows the distinctive northern hemisphere spring O 3 maximum in the troposphere at Trinidad Head (Figure 7a) and Boulder (Figure 7d).These model results are close to the measurements (Figures 7b and e).Occurrences of high O 3 ratios (about 0.85-0.95; Figure 7c and f this region during the northern hemisphere spring.As the season enters summer, the dramatic reduction in the values of the O 3 ratios in the lower troposphere illustrate the importance of the tropospheric photochemical processes (Figure 7c and d).For example, the model without considering combustion emissions of O 3 precursors in the lower troposphere produced less than 50% of the surface O 3 at Boulder.

Comaprisons of Scenarios and Seasonal Variations
Previous sections showed that downward transport of elevated ozone from the upper troposphere and lower stratosphere regions, and upward transport of primary emissions from the surface affect ozone concentrations in the troposphere.In this section, we further test these two main effects, and exam seasonal variations of these effects on tropospheric ozone concentrations.For all these experiments of stratospheric effects, the winter northern hemisphere is mostly dominated by the stratospheric processes, while the tropospheric emissions dominate over the southern hemisphere from tropics to 30 • S regions.The northern hemisphere continental regions are increasingly dominated by tropospheric emissions from spring, to reach maxima in summer, and started to reduce in autumn months.
For the experiments with no industrial, biomass burning, and vegetation emissions (Figure 10); and no vegetation emissions (Figure 11), the northern hemisphere continental areas all exhibit consistent effects dominated by tropospheric emissions from spring to autumn months.The tropospheric emissions consistently dominate over the southern hemisphere continental regions from tropics to 30 • S.

Conclusions
In this work, we examined the effect of tropospheric emissions on tropospheric ozone (O 3 ) by

Figure 1
Figure 1 compares time-series plots of modeled O 3 from simulations with and without combustion emissions included, respectively, with observed O 3 levels at four sites located in the remote marine boundary layer.The simulation without the combustion emissions can reproduce the observed spring O 3 maximum at the surface similar to that shown in the measurements.

O 3
Figure 2a shows surface O 3 distribution simulated without combustion emissions included in the troposphere during the southern hemisphere spring months.The high surface O 3 (highlighted by the O 3 =35 ppbv contour) is well maintained but with less widespread coverage over the southern hemisphere oceans in the model without emissions compared with the model with emissions included (Figure 2b).The no-emission simulation presents the inter-hemispheric asymmetry in O 3 over the marine boundary layer.This pattern indicates downward transport of O 3 from the stratosphere mostly maintained the inter-hemispheric asymmetric of O 3 .In the northern hemisphere emission areas such as eastern U.S., Europe, and East Asia the model without emissions included shows that O 3 over these regions has reduced.For the tropical and the southern hemisphere regions such as South America, Central Africa, and Southeast Asia, these areas exhibit low O 3 concentrations in the model without emissions included.For example, less than 5 ppbv of O 3 occur over the Amazon basin, tropical Africa, and Southeast Asia.

Figure
Figure2cshows surface O 3 concentrations from the run without to the run with emissions included for the southern hemisphere spring.Without photochemical source emissions in the troposphere, surface O 3 levels in the tropical and the southern hemisphere continental regions (close to intense biomass burning activities) are dramatically reduced to as low as about 20% in South America and Africa, and about 30% in Southeast Asia.

Figure 3
Figure 3 shows mean O 3 at surface without (Figure 3a) and with (Figure 3b) emissions for the March-April period.Both model simulations produce the distinctive patterns of inter-hemispheric asymmetry in the surface O 3 over the marine boundary layer.Surface O 3 concentrations from the model without emissions included are about 80% − 90% to the O 3 levels from the model with combustion emissions added over the northern hemisphere marine boundary layer(Figure 3c).Continental regions such as the Amazon basin and South America, equatorial Africa, and Southeast Asia exhibit low levels of the surface O 3 in the model without combustion emissions included.The no-emission simulation calculated less than 5 ppbv of O 3 concentrations in some of these areas.In addition to the lack of O 3 producing precursors in the no-combustion emissions model, high density terrestrial vegetation, which emits a large amount of volatile biogenic organic compounds such as isoprene and monoterpenes, reacts with O 3 in these area and causing a sharp reduction in O 3 concentrations [8].In addition to the tropical and southern hemisphere continental areas where low surface O 3 exist, low surface O 3 (≤ 15-20 ppbv) prevails over the tropical South Atlantic, the Indian, and the Pacific.As shown from the no-combustion model simulation (Figure 3a), low surface O 3 extends from tropical and southern hemisphere continental areas to the remote tropical and extra-tropical marine boundary layer.The surface distribution of the O 3 ratios (Figure 3c) shows a long narrow band of low values (≤ 60 − 70%) over the remote tropical marine boundary layer, extending from continental areas where as low as less than 20% of the O 3 ratios exist.These results exhibit the influence of tropical and southern hemisphere continental air on the remote tropical and extra-tropical marine boundary layer.We find that surface O 3 with concentrations less than 5 ppbv prevail over the Amazon basin, tropical South America, equatorial and tropical Africa, and Southeast Asia in the model without emissions included.Low O 3 concentrations over the remote marine boundary layer are associated with air low in O 3 transported from the clean tropical continental regions.On the other hand, high O 3 concentrations can be maintained over the extra-tropical remote marine boundary layer due to the downward transport of O 3 from the stratosphere.

Figure 5
Figure 5 compares time-height cross sections of modeled and measured O 3 profiles at a northern (Bermuda; Figures5a and b) and a southern (Irene; Figures5d and e) Atlantic sites.For the simulations ) and a southern (Irene; Figures5d and e) Atlantic sites.For the simulations without combustion emissions, distinctive spring maximum occur in the upper to middle troposphere Preprints (www.preprints.org)| NOT PEER-REVIEWED | Posted: 20 June 2017 doi:10.20944/preprints201706.0091.v1

Figure
Figure6ashows time-height cross sections of vertical ozone profiles simulated without combustion emissions in the model at Taiwan, which is located in the subtropical western North Pacific.The model shows downward intrusion of elevated O 3 from the stratosphere to upper and the middle troposphere from February to May.The vertical extent of the elevated O 3 in the troposphere at this site is not as extensive as O 3 shown from the measurements during the same period (Figure6b).High O 3 ratios (greater than 85%) occur during winter (December, January, and February; Figure6c), indicating that the no-emission model can produce O 3 in the troposphere comparable with the model with emissions included.On the other hand, development of the low O 3 ratios from spring to late summer and early autumn, and extends from the lower troposphere in spring to the whole troposphere in September.These patterns indicate that tropospheric photochemical production of O 3 plays a significant role in contributing to the observed O 3 maximum in spring and summer at this location.These characteristics are consistent with previous studies[29;30].

Figure 8
Figure 8 shows test results of ground-level O 3 distribution when the tropospheric chemistry (including photochemical sources and sinks, and deposition removal of O 3 ) turn off (experiment 7).As no photochemical processes exit for O 3 , this is a long-lived O 3 tracer test.The simulation results show that O 3 from the tropopause regions fill the entire troposphere.However, there exist spatial and temporal variabilities in the amount and distribution of elevated ozone been transported downward from the tropopause regions.For the northern hemisphere winter months (DJF), the maximum ground-level O 3 concentrations occur around 30 • N. In the southern hemisphere, the highest ground-level O 3 concentrations occur around 20 • S − 30 • S. The western North America, western North Africa, and the Himalaya area are three hot-spot areas in the northern hemisphere.The eastern South Pacific, South Atlantic, South Africa, and Australia are hot spot areas in the southern hemisphere.These hot spot areas are also geographically consistent with the global distribution of desert areas.During the northern hemisphere spring months (MAM), both 30 • N and 20 • S −30 • S latitudinal bands exhibit maximum ground-level O 3 concentrations.During the northern hemisphere summer months (JJA), the 30 • N and 20 • S − 30 • S latitudinal bands again present maximum ground-level O 3 concentrations.However, the ground-level O 3 patterns over the North Atlantic and the northwestern Pacific show the northward intrusion of low O 3 from tropical latitudes, and southward transport of elevated O 3 from mid-latitudes.The atmospheric transport processes associated with the subtropical high circulations over the North Pacific and the North Atlantic, respectively, are responsible for the generation of these patterns during the northern hemisphere summer season.During the northern hemisphere autumn months (SOP), both 30 • N and 20 • S − 30 • S latitudinal bands remain to exhibit elevated O 3 concentrations on the ground.The above results show that both 30 • N and 20 • S − 30 • S latitudinal bands exhibit high impact from the tropopause regions.As such, these latitudinal bands are the areas which are less responsive to the tropospheric emissions and more susceptible to the transport processes in the tropopause regions.Indeed, a followed up short-lived tracer experiment (experiment 28, Figure9) confirms that 30 • N and 20 • S − 30 • S latitudinal bands are more susceptible to the tropopause regions than other latitudinal bands.
conducting three-dimensional (3D) chemistry transport model (CTM) simulations.For the control run, the CTM model simulates tropospheric O 3 levels with a complete set of anthropogenic, biomass burning, and vegetation emissions.The pyrogenic emissions (biomass burning emissions from tropical and southern continents such as South America, Africa, and Southeast Asia) and anthropogenic emissions from fossil-fuel related combustions.For the no-emission simulation, all anthropogenic and biomass burning emissions were turned off.Comparisons of results from these two simulations exhibit the emission impacts on the tropospheric O 3 .In the no-emission simulation, distinctive low surface O 3 with concentrations less than 5 ppbv prevail over the Amazon basin, tropical South America, tropical South Africa, Southeast Asia.Transport of air from these land areas downwind contributes to the low O 3 over the remote marine boundary layer.In contrast, elevated O 3 levels over the extra-tropical remote marine boundary layer are less supported by the anthropogenic and biomass burning emissions but more sustained by the downward transport of O 3 from the stratosphere.These results demonstrate that the northern hemisphere continental areas (north of 30 • N), polar regions, and tropical continental regions are more sensitive to the tropospheric emissions.The northern hemisphere winter is mostly dominated by the stratospheric processes, while the tropospheric emissions dominate over the southern hemisphere tropical continental areas from tropics to 30 • S latitudinal bands.The northern hemisphere continental regions are increasingly dominated by tropospheric emissions from spring, to reach maxima in summer, and started to reduce in autumn months.

Figure 1 .
Figure 1.Comparison of two modeled seasonal cycles of O 3 (ppbv) calculated with (solid thin lines) and without (bold dashed lines) emissions included in the model at (a) Westman, Iceland (63.4 • N, 20.3 • W ), (b) Bermuda (32 • N, 65 • W ), (c) Mauna Loa (19.5 • N, 155.6 • W ), and (d) Samoa (12.3 • S, 170.6 • W ) with the measurements (thin dashed lines).Two measured O 3 levels for the period 1988-1992 (except at Westman where the 1992-1997 data were used) are shown here, one for the daily maximum, while the other one for the daily minimum.

Figure 2 .
Figure 2. September to October mean ozone distributions (ppbv) simulated for (a) without and (b) with combustion emissions included in the model [1]; and (c) ratio of O 3 between (a) and (b).

Figure 3 .
Figure 3. March to April mean ozone distributions (ppbv) simulated for (a) without and (b) with combustion emissions included in the model [1]; and (c) ratio of O 3 between (a) and (b).

Figure 4 .
Figure 4. Zonal mean cross-section of ratio of O 3 simulated without to those with combustion emissions included in the model for (a) March, April, and May, (b) June, July, and August, (c) September, October, and November, (d) December, January, and February.

Figure 7 .
Figure 7. Time-height cross sections of O 3 (ppbv) simulated without combustion emissions included at Trinidad Head (41.1 • N, 124.2 • W ) (a) and Boulder (40 • N, 105 • W ) (d); (b) and (e) show ozonesonde measurements; (c) and (f) show ratio of O 3 simulated without to those with combustion emissions included in the model.

Figure 8 .
Figure 8. Seasonal mean ground-level O 3 concentrations (a-d) from long-lived tracer experiment; and ground-level O 3 ratios (O 3 from long-lived simulation to those from control run; e-h).(a) and (e) are Winter DJF (December, January, February) months.(b) and (f) are spring MAM (March, April, May) months.(c) and (g) are summer JJA (June, July, August) months.(d) and (h) are autumn SON (September, October, November) months.

Figure 9 .
Figure 9. Seasonal mean ground-level O 3 concentrations (a-d) from short-lived tracer experiment; and ground-level O 3 ratios (O 3 from long-lived simulation to those from control run; e-h).(a) and (e) are DJF months.(b) and (f) are MAM months.(c) and (g) are JJA (June, July, August) months.(d) and (h) are SON months.

Figure 10 .Figure 11 .
Figure 10.Seasonal mean ground-level O 3 ratios from a simulation without stratospheric O 3 and NOy fluxes to control run (a-d); and ground-level O 3 ratios from a simulation without tropospheric industrial, biomass burning, and vegetation emissions to control run (e-h).(a) and (e) are DJF months.(b) and (f) are MAM months.(c) and (g) are JJA (June, July, August) months.(d) and (h) are SON months.

Table 1 .
List of CTM Scenarios Simulated in This Work.