Process analysis of regional ozone formation over the Yangtze River Delta , China using the Community Multi-scale Air Quality modeling system

A high O3 episode was detected in urban Shanghai, a typical city in the Yangtze River Delta (YRD) region in August 2010. The CMAQ integrated process rate method is applied to account for the contribution of different atmospheric processes during the high pollution episode. The analysis shows that the maximum concentration of ozone occurs due to transport phenomena, including vertical diffusion and horizontal advective transport. Gas-phase chemistry producing O3 mainly occurs at the height of 300–1500 m, causing a strong vertical O3 transport from upper levels to the surface layer. The gas-phase chemistry is an important sink for O3 in the surface layer, coupled with dry deposition. Cloud processes may contribute slightly to the increase of O 3 due to convective clouds or to the decrease of O 3 due to scavenging. The horizontal diffusion and heterogeneous chemistry contributions are negligible during the whole episode. Modeling results show that the O 3 pollution characteristics among the different cities in the YRD region have both similarities and differences. During the buildup period, the O 3 starts to appear in the city regions of the YRD and is then transported to the surrounding areas under the prevailing wind conditions. The O3 production from photochemical reaction in Shanghai and the surrounding area is most significant, due to the high emission intensity in the large city; this ozone is then transported out to sea by the westerly wind flow, and later diffuses to rural areas like Chongming island, Wuxi and even to Nanjing. The O3 concentrations start to decrease in the cities after sunset, due to titration of the NO emissions, but ozone can still be transported and maintain a significant concentration in rural areas and even regions outside the YRD region, where the NO emissions are very small.


Conclusions References
Tables Figures

Introduction
With the rapid economic development and significant increase of energy consumption, the anthropogenic emissions have been continuously growing in recent years in Eastern China, which have led to significant changes in atmospheric ozone, including the increase in tropospheric ozone.More and more concerns have been focused on the regional pollution in those leading economic regions, for example, the Yangtze River delta (Zhao et al., 2004;Wang et al., 2005;Li et al., 2011), the Pearl River delta (Wang et al., 2010;Shen et al., 2011), and the Bohai Bay region (Wang et al., 2006(Wang et al., , 2009(Wang et al., , 2010;;An et al., 2007;Chou et al., 2011).In recent years, high ozone concentrations over 90 ppb have been frequently observed by in-situ monitoring in Eastern China (Zhang et al., 2008;Ran et al., 2009;Shao et al., 2009;Tang et al., 2009).In many megacities Introduction

Conclusions References
Tables Figures

Back Close
Full and polluted areas in Eastern China, serious ozone pollution caused by large precursor emissions has concerned citizens and decision-makers (Shao et al., 2009).Ozone originates from in-situ photochemical production in the reactions from the mixture of reactive volatile organic compounds (VOCs) and nitrogen oxides (NO x ) and from vertical and horizontal transport.In the troposphere, photolysis of ozone by solar UV radiation to electronically excited O( 1 D) and the subsequent reaction with water vapor is the major source of hydroxyl radical (OH).OH is one of the key species for the chemical reactions in the atmosphere and its abundance is an important index of the oxidizing capacity of the atmosphere.Thus, the control of O 3 is a complicated problem due to the nature of the non-linear formation of O 3 (Seinfeld and Pandis, 2006).
The Yangtze River Delta (YRD), characterized by high population density and welldeveloped industry, is one of the largest economic regions in China.Many studies related to the ozone concentration in the YRD have been conducted in the past years.Observations made by Xu et al. (2006) show that high ozone concentrations in the YRD have occurred.Xu et al. (2007) analyzed the tropospheric ozone using satellite data and found that the tropospheric ozone concentration has kept on increasing in the YRD.Especially, Chan and Yao (2008) have given a general review on the ozone mass concentrations and formation studies done in Shanghai, a typical mega city in the YRD, showing that some O3 episodes have occurred in the region.However, since the YRD is a large area, the emissions rates and emission characteristics of the ozone precursors vary greatly.This means that the ozone formation process differs with areas in the region.Studies on the process analysis of typical ozone episodes over the YRD are quite limited up to now.
In this study, we first perform an observation analysis to identify a typical summertime O 3 episode over the YRD in 2010.Then the Community Multi-scale Air Quality Modeling System (CMAQ) (Dennis et al., 1996;Byun et al., 1998;Byun and Ching, 1999) is used to reproduce the high ozone case, and integrated process rate analysis (IPR), implemented within CMAQ, is applied to analyze the formation of ozone at typical sites in the YRD.This is undertaken to identify the dominant processes contributing Introduction

Conclusions References
Tables Figures

Back Close
Full to the O 3 formation and determine the characteristics of the photochemical system at different locations or at a given location on different days.Geographic Information System (GIS) technology is applied in gridding the YRD regional emission inventory to the model domain.The newly calculated emissions for the YRD in 2007(Huang et al., 2011) are updated to the year 2010 based on energy consumption and are then inserted into the regional East Asian emission inventory provided by INTEX-B (Streets et al., 2003a,b;Fu et al., 2008;Zhang et al., 2009).For biogenic VOC emissions, this study uses the natural VOC emission inventory of the GEIA Global Emissions Inventory Activity 1990 (http://geiacenter.org).

Model evaluation protocol
Predicted meteorological parameters including wind speed, wind direction, temperature and humidity are compared with the hourly observational data obtained during August 2010.Performance statistics of MM5 are calculated with application of the Metstat statistical analysis package (Emery et al., 2001).
The simulation of O 3 formation during 5-31 August is evaluated against observations made at the supersite in Shanghai Academy of Environmental Sciences (SAES).The measurements are collected simultaneously at the surface site.The levels of O 3 and NO x were measured by Ecotech commercial instruments EC9811 and EC9841A, respectively.
The model performance is judged by statistical measures, including the normalized mean bias (NMB), index of agreement (I), correlation coefficient (R), and factor of two.Figures

Back Close
Full The NMB, I and Factor of two are calculated by Eqs. ( 1)-(3): • 100 % (1) where p i represents the predicted data and o i represents the observational data.N means the number of data pairs.o denotes the average observed concentration and a value of 1 indicates perfect agreement between predicted and observed values.R is the percentage of the ratios between 0.5 and 2; N [1/2,2] is the number of the ratios between 0.5 and 2; and N t is the total number of comparison pairs.The larger the R value, the better the model performs.R = 100 % means the model performance is perfect.

Integrated process rate analysis
Quantifying the contributions of individual processes to model predictions provides a fundamental explanation and shows the relative importance of each process.IPR analysis deals with the effects of all the physical processes and the net effect of chemistry on model predictions.The IPR analysis calculates hourly contributions of horizontal advection and diffusion, vertical advection and diffusion, dry deposition, gasphase chemistry, clouds process and aqueous chemistry, etc.The IPR method has Introduction

Conclusions References
Tables Figures

Episode Selected for process analysis
The 16-17 August 2010, is a typical summertime situation, with average surface temperature of 30.5 • C, maximum temperature of 36.7 • C, occurring at 12:00 LST on 16 August.The average relative humidity was 63.5 %, and the wind speed was lower than 1 m s −1 during the period.The observed maximum surface hourly O 3 concentration was 86 ppb, which is very rare in an urban site of Shanghai region.In this paper, 16-17 August is used for the IPR analysis because the high ozone episode was observed at the SAES observational site.
For the IPR analysis, we first assess the roles of various atmospheric processes in O 3 formation at the supersite located in Shanghai Academy of Environmental Sciences, Xuhui District, which represents a typical site in urban Shanghai, and in Jinshan District, a chemical industrial site with high VOC emissions, which reflects the influences of pollutants transported from upwind areas and local precursor emissions.We then investigate the influences of different processes on the formation and evolution of regional O 3 pollution over the YRD region as a whole.The sites we selected include two provincial capital cities, Nanjing and Hangzhou.Nanjing is located in Jiangsu province, northwest of Shanghai, and Hangzhou is located in Zhejiang province, southwest of Shanghai.Locations of the sites selected to do O 3 process analysis in this study are shown in Fig. 2.

Evaluation of model performance
Table 1 shows comparisons between observed and modeled meteorological parameters including surface temperature, wind speed, wind direction and relative humidity Introduction

Conclusions References
Tables Figures

Back Close
Full during the period of 5-31 August 2010 at four surface stations in Shanghai, namely Baoshan (BS), Jinshan (JS), Nanhui (NH) and Qingpu (QP), shown in Fig. 2. Figure 3 gives average data for the four meteorology monitoring sites.The average bias of wind speed, wind direction, temperature and humidity are 0.65, 2.5, 0.13 and −0.81, respectively.Figure 3 shows the daily comparisons of the meteorological parameters, which indicates that MM5 can reflect the variation trends of the major meteorological conditions.The selected parameters adopted in MM5 can be used in the pollutant concentration simulation.
Figure 4 shows the comparisons between model results and observational data for O 3 , NO 2 , NO x and NO y hourly concentrations at Shanghai during 5-31 August 2010.
Results show that CMAQ can reproduce the variation trends of the O 3 , with a correlation coefficient of 0.78, NMB of 30.2 %, NME of 55.8 %, and I of 0.91, comparable to the performance of other CMAQ applications (Goncalves et al., 2009;Shen et al., 2011).The model also reproduces the daily change of O 3 concentration.With increase of solar radiation early in the day, the O 3 concentration rises; while in the afternoon, with decrease of radiation, the O 3 concentration gradually declines.Comparisons of precursor concentrations including NO 2 , NO x and NO y at the monitoring site further demonstrate that the O 3 formation is captured reasonably well over the domain and throughout the period.The index of agreement for NO 2 , NO x and NO y are 0.91, 0.63 and 0.64, respectively, as shown in Table 2.

Urban area of Shanghai
The contributions of different atmospheric processes to the evolution of O 3 in the urban Shanghai area (Xuhui) from 16 to 17 August, 2010 at different layers are shown in well-developed photochemistry in the urban plume, thereby further enhancing the O 3 concentrations in urban Shanghai area and leading to another high O 3 episode, approaching 86 ppb.

Suburban and industrial area of Shanghai
The Jinshan District (JS) is located in the southwest of urban Shanghai area.It belongs to an oil and chemistry industrial region.Lots of petrochemical enterprises, including Shanghai Jinshan Photochemical Company are located in this region.Compared to urban Shanghai area with lots of NO x emissions from motor vehicles, the emissions of volatile organic compounds (VOCs) are much more significant.As shown in Fig. 6, The major processes controlling the surface ozone production in the JS site during the daytime on both days include photochemical reaction, vertical diffusion and horizontal advective transport, while dry deposition and vertical advective transport are the most significant sink of O 3 concentrations.During the simulation period, the average positive contributions of vertical diffusion and horizontal transport are 25.8 ppb h −1 , 10.3 ppb h −1 , accounting for 25.9 % and 11.The daily change of O 3 concentrations are most significant in surface layers, however, the diurnal change becomes less with vertical height.At the 900-1400 m height, the ozone concentrations retain around 50 ppb.

Nanjing, the capital city of Jiangsu province
The surface O 3 concentrations in Nanjing, the capital city of Jiangsu province, have been modeled and the IPR analysis was applied into the process contribution calculation.As shown in Fig. 7, The major processes controlling the surface ozone production at the Nanjing site during the daytime on both days include vertical diffusion and horizontal transport, while photochemical reaction, vertical advective transport and dry deposition are the most significant sinks of O 3 .During the simulation period, the average O 3 production rates contributed by vertical diffusion and horizontal transport are 25.7 (350-500 m) and 8 (500-900 m) are generally higher than the ground layer, which is around 50 ppb during the whole simulation period.As shown in Fig. 8, The major processes controlling the surface ozone production in the Hangzhou site during the daytime on both days include vertical diffusion, while dry deposition are the most significant sink of O 3 concentrations.During the simulation period, the average positive contribution of vertical diffusion is 21.4 ppb h −1 , accounting for 28.9 %.During the buildup of daytime maximum O 3 from 10:00 to 14:00 LST on both 16 and 17 August, gas-phase chemistry also plays an important role in the formation of net surface O 3 production.The average positive contributions to O 3 are 12.6 and 7.0 ppb h −1 , accounting for 10.7 % and 6.2 % during this time period on 16 and 17 August, 2010, respectively.
The process contributions to net surface O 3 concentrations assessed in Hangzhou site during this period indicates that net transport accounts for 9 % and chemical reaction for −9 %.The O 3 concentrations in most urban cities during the night time (0:00-6:00) are very low, because the high NO emissions in the urban area gradually titrate ozone, which causes the ozone concentration to decrease.However, the NO emissions in the rural area are much lower, which preserves the high ozone concentration.

Regional ozone transport
The O 3 concentration in the Yangtze River Delta started to accumulate from 8:00 on 16 August, 2010.As shown in Fig. 10, from 8:00 on 16 August, the O 3 concentrations started to rise in the cities of Suzhou, Hangzhou and Ningbo.Under the southwest wind direction, this ozone started to diffuse to Shanghai and the surrounding area.From 10:00 on 16 August, the O 3 production from photochemistry in major cities like Hangzhou, Ningbo, Shanghai and Suzhou is high.The maximum ozone production rates from photochemical reaction reached 45.4 ppb h −1 , occurring at 12:00 on 16 August, with the contribution of 20.4 % to the total O 3 .Later on, the O 3 is transported to the ocean under the westerly wind, and then flows back to North Shanghai and South Jiangsu area due to the northeast wind off the sea.After sunset, the O 3 concentrations started to decrease.
On the next day, the O 3 was produced from chemical reactions in the cities of Shanghai, Hangzhou, Suzhou, Wuxi and Nanjing from 10:00 on 17 August 2010.The O 3 transported during 10:00-15:00 is mainly surrounding the Shanghai area, since the wind speed is relatively low, as shown from both Figs. 9 and 10.Later on, it spreads to the northwest part of the region, including Nanjing, under the easterly wind.On this day, although the solar radiation and the highest air temperature are both lower than the previous day, the highest O 3 concentration occurs in Shanghai.This is because the wind speed is low, and the horizontal transport from the surrounding area is greater than the previous day, while diffusion to the other regions is low.Introduction

Conclusions References
Tables Figures

Back Close
Full

Conclusions
The Community Multi-scale Air Quality modeling system (CMAQ) was applied to investigate the high O 3 pollution episodes over the YRD region during August, 2010.The modeling results agree well with the observational data, and the model reproduced the concentration levels and variations.Model performance evaluation shows that the model is suitable for the simulation and prediction of photochemical pollution.
The Integrated Process Rate implemented in the CMAQ model was applied to obtain quantitative information about atmospheric processes affecting the ozone concentration in typical cities located in the Yangtze River Delta area, including Shanghai, Nanjing and Hangzhou.A representative summertime photochemical pollution episode (16-17 August, 2010) was selected.Applying the Integrated Process Rate tool to the first vertical layer simulated provides information about the surface concentration of pollutants estimated by the model.In order to perform a deeper study of the contributions of the main atmospheric processes leading to the levels of these pollutants, the vertical ozone production in layer 1 (0-40 m), layer 7 (350-500 m), layer 8 (500-900 m) and layer 10 (1400-2000 m) have been examined.
Process analysis indicates that the maximum concentration of photochemical pollutants occur due to transport phenomena, including vertical transport and horizontal transport.The gas-phase chemistry producing O 3 mainly occurs in the height of 300-1500 m, making a strong vertical O 3 transportation from upper level to the surface layer.In the downwind area, the high surface O 3 levels are not produced in situ, but come from horizontally advected flows during the morning and gas-phase chemical contributions occurring aloft.The urban Shanghai domain behavior slightly differs: the horizontal advection is also the main contributor to O 3 surface concentrations, but the chemical formation takes place in the whole vertical column below the PBL.
The gas-phase chemistry is an important sink for O 3 in the lowest layer, coupled with vertical advection flows and dry deposition.The horizontal diffusive processes contributions to net O3 concentrations under the PBL are relatively low, which is negligible Introduction

Conclusions References
Tables Figures

Back Close
Full As shown by the modeling results, the O 3 pollution characteristics among the different cities in the YRD region have both similarities and differences.During the buildup period (usually from 8:00 in the morning after sunrise), the O 3 starts to appear in the city regions like Shanghai, Hangzhou, Ningbo and Nanjing and is then transported to the surrounding areas under the prevailing wind conditions.On both days, the O 3 production from photochemical reaction in Shanghai and the surrounding area are most significant, due to the high emission intensity in the large city; this ozone is then transported out to sea by the westerly wind flow, and later diffuses to rural areas like Chongming island, Wuxi and even to Nanjing.The O 3 concentrations start to decrease in the cities after sunset, due to titration of the NO emissions, but ozone can still be transported and maintain a significant concentration in rural areas and even regions outside the YRD region, where the NO emissions are very small.This work explores the possibilities of applying process analysis to high ozone pollution episodes, proving that it is useful not only to better evaluate the simulation results, but also to perform more accurate source apportionment of pollutants over a region.
Results show that to control O 3 pollution, the measures should be taken locally and regionally as well.Introduction

Conclusions References
Tables Figures

Regional ozone transport
Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | N).The model domain is shown in Fig. 1.The MM5 domain is larger than the CMAQ domain, with three grids more than the CMAQ domain on each boundary.The time period chosen for simulation is 1-31 August 2010, when the YRD experienced high ozone concentrations.The initial conditions of CMAQ are prepared by running the model three days ahead of each start date with clean initial conditions (IC).The boundary condition (BC) used for the largest domain of CMAQ is clean air, while the BCs for the nested domains are extracted from the CMAQ Chemical Transport Model (CCTM) concentration files of the larger domain.Both the MM5 and CMAQ employ 14 vertical layers of varying thickness with denser layers in the lower atmosphere to better resolve the mixing height.The driving meteorological inputs for CMAQ are provided by MM5, and the meteorology-chemistry interface processor (MCIP) is used to transfer MM5 output into gridded meteorological field data as the input to CMAQ.The inputs for MM5 are NCEP FNL (Final) Operational Global Analysis data, which are available on 1.0 • × 1.0 • grids continuously for every 6 h (http://dss.ucar.edu/datasets/ds083.2/).The Carbon Bond 05 chemical mechanism (CB05) is used in the CMAQ model (Sarwar, 2008).Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper |

Fig. 5 .
photochemistry was stronger at 300-1500 m height than in the surface layer, making a strong vertical O 3 transport from upper levels to the surface layer.This indicates that the strong vertical O 3 import at surface layer is initiated by the urban plume arrival.On 17 August, high O 3 input by transport occurred later in the afternoon, arising from 3 % of net O 3 change, respectively.The average O 3 production rate contributed by dry deposition and vertical advective transport are −24.1, and −11.9 ppb h −1 , accounting for −20.5 % and −11.2 % of net O 3 change, respectively.The O 3 production rates from chemistry during 10:00-14:00 LST on 16-17 August, 2010 are between 8.2-45.4ppb h −1 .The maximum ozone production rates from photochemical reaction were 45.4 and 23.3 ppb h −1 , occurring at 12:00 on 16 August and 14:00 on 17 August, with the contribution of 20.4 % and 12.9 %, respectively.The processes contributions to net surface O 3 concentrations assessed in JS site during this period indicates that net transport (ZADV + HADV + HDIF + VDIF) accounts for 26.3 % and chemical reaction for −11.1 %. Discussion Paper | Discussion Paper | Discussion Paper | 3 ppb h −1 , 15.3 ppb h −1 , accounting for 25.8 % and 18.5 % of net O 3 change, respectively.The average contributions of photochemistry, vertical advective transport and dry deposition to O 3 change are −30.6,−6.2 and −3.7 ppb h −1 , accounting for −31.8 %, −7.3 % and −4.7 %, respectively.The maximum surface O 3 concentration was 88.7 ppb, occurring at 12:00 LST on 16 August, 2010.At this time, the O 3 production rate from chemistry is 46.7 ppb h −1 , accounting for 59.4 % of the net O 3 concentration change.The processes contributions to net surface O 3 concentrations assessed at the Nanjing site during this period indicate that net transport accounts for 9 % and chemical reaction for −32 %.However, if we look at the O 3 concentration change in the height of 350-500 m and 500-900 m, we can see that during the time period of 10:00-15:00 LST, photochemistry plays the most important role in net O 3 production.The highest positive contributions from gas-phase chemistry to net O 3 production in the height of 350-500 m and 500-900 m reached 87.3 % and 68.6 %, respectively, and making a strong vertical O 3 transportation from the upper level to the surface layer., the capital city of Zhejiang province Although during 16-17 August 2010, Shanghai and Nanjing experienced high O 3 concentrations, a similar situation did not occur in Hangzhou.Modeling results show that the highest O 3 concentration in the cell of Hangzhou during this period was only 59.0 ppb, occurring at 15:00 LST on 17 August, 2010.The O 3 concentrations in layer

Figure 9
Figure 9 shows the time series of meteorological conditions and O 3 and NO x mass concentrations observed at the site of SAES during the high O 3 pollution episodes on 16-17 August, 2010.During the pollution episode, the NO x , O 3 and O x (NO 2 + O 3 ) continued to increase.Daily average concentrations of NO 2 and O x were 23.4 ppb and 57.5 ppb on 16 August, and increased to 35.0 ppb and 66.5 ppb on 17 August 2010, respectively.The maximum hourly concentration of O 3 increased from 75.1 ppb on 16 August to 86.5 ppb on 17 August 2010.During the period, the air temperature was around 30 • C, with the highest temperature of 36.7 • C at 12:00 on 16 August and 15062 Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper |compared to other atmospheric processes.The diffusive processes contributions to net O 3 concentrations under the PBL are relatively low, and in particular the horizontal diffusion is negligible compared to other atmospheric processes.Vertical diffusion compensates the loss of O 3 in surface layers due to NO titration, contributing positively to net O 3 concentrations in urban areas.The O 3 peaks at surface level are higher in the suburban industrial region (Jinshan), mainly due to the much simpler transport pattern compared to the downtown region, together with the more significant photochemistry.The cloud processes, wet deposition and heterogeneous chemistry contributions are negligible during the whole episode, characterized by high solar radiation and no precipitation or cloudiness.
Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Technology Commission of Shanghai Municipality Fund Project via grants No. 11231200500, and No. 10231203802.The authors appreciate the suggestions made by the reviewers that helped greatly to improve this paper.Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper |

Figure 1
Figure 1 One-way nested CMAQ model domain riving meteorological inputs for CMAQ are provided by MM5, and the -chemistry interface processor (MCIP) is used to transfer MM5 output into gridded cal field data as the input to CMAQ.The inputs for MM5 are NCEP FNL (Final)

Fig. 3 .
Fig.3.Time series of simulated surface wind speed, wind direction, temperature and relative humidity compared with observations at four monitoring sites during 5--31 August 2010.

Figure 4
Figure 4 shows the comparisons between model results and observational data for O 3 , NO 2 , NO x and NO y hourly concentrations at Shanghai during 5--31 August 2010.Results show that CMAQ can reproduce the variation trends of the O 3 , with a correlation coefficient of 0.78, NMB of 30.2%,NME of 55.8%, and I of 0.91, comparable to the performance of other CMAQ applications (Goncalves et al.,

Fig. 3 .Fig. 4 .
Fig. 3. Time series of simulated surface wind speed, wind direction, temperature and relative humidity compared with observations at four monitoring sites during 5-31 August 2010.

Fig. 4 .
Fig. 4. Time series of simulated surface O 3 , NO 2 , NO x and NO y against that observed at SAES monitoring site during 5-31 August 2010.

Fig. 5 .
Fig.5.Atmospheric processes contribution to net O3 density at Xuhui site during 16--17 August, 2010.Suburban and industrial area of ShanghaiThe Jinshan District is located in the southwest of urban Shanghai area.It belongs to an oil and chemistry industrial region.Lots of petrochemical enterprises, including Shanghai Jinshan Photochemical Company are located in this region.Compared to urban Shanghai area with lots of NOx emissions from motor vehicles, the emissions of volatile organic compounds (VOCs) are much more significant.As shown in Fig.6, The major processes controlling the surface ozone production in the JS site during the daytime on both days include photochemical reaction, vertical diffusion and horizontal advective transport, while dry deposition and vertical advective transport are the most significant sink

Fig. 7 .
Fig.7.Atmospheric processes contribution to net O3 density at Nanjing site during 16-17 August, 2010.Hangzhou, the capital city of Zhejiang province Although during 16-17, August, 2010, Shanghai and Nanjing experienced high O3 concentrations, a similar situation did not occur in Hangzhou.Modeling results show that the highest O3 concentration in the cell of Hangzhou during this period was only 59.0 ppb, occurring at 15:00 LST on 17 August, 2010.The O3 concentrations in layer 6 (350-500m) and 8 (500-900m) are generally higher than the ground layer, which is around 50 ppb during the whole simulation period.As shown in Fig.8, The major processes controlling the surface ozone production in the Hangzhou site during the daytime on both days include vertical diffusion, while dry deposition are the most significant sink of O3 concentrations.During the simulation period, the average positive contribution of vertical diffusion is

Figure 8
shows the time series of meteorological conditions and O3 and NOx mass concentrations observed at the site of SAES during the high O3 pollution episodes on 16-17 August, 2010.During the pollution episode, the NOx, O3 and Ox (NO2+O3) continued to increase.Daily average concentrations of NO2 and Ox were 23.4 ppb and 57.5 ppb on August 16, and increased to 35.0 ppb and 66.5 ppb on August 17, 2010, respectively.The maximum hourly concentration of O3 increased from 75.1 ppb on August 16 to 86.5 ppb on August 17, 2010.During the period, the air temperature was around 30℃,
Fig.8.Time series of meteorological conditions and mass concentrations that observed at SAES monitoring site during the high O 3 pollution episodes on 16-17 August, 2010.(RS: solar radiation intensity; WS: wind speed; AT: air temperature; RH: relative humidity)The O 3 concentrations in most urban cities during the night time (0:00-6:00) are very low, because high NO emissions in the urban area gradually titrate ozone, which causes the ozone concentration ecrease.However, the NO emissions in the rural area are much lower, which preserves the high ne concentration.The O concentration in the Yangtze River Delta started to accumulate from 8:00 on August 16,

Fig. 9 .
Fig. 9. Time series of meteorological conditions and mass concentrations that observed at SAES monitoring site during the high O 3 pollution episodes on 16-17 August 2010 (RS: solar radiation intensity; WS: wind speed; AT: air temperature; RH: relative humidity).

2 Methodology 2.1 Model setup and input data
In this paper, the CMAQ version 4.6 is used with the Carbon Bond 05 (CB05) chemical mechanism to simulate the high ozone episode in the YRD in 2010.The CMAQ model domain is based on a Lambert Conformal map projection, using a one-way nested mode with grid resolutions of 81 km (covering all China, Japan, Korea, parts of India and Southeast Asia); 27 km (covering Eastern China) and 9 km (covering major cityclusters including Shandong province, the YRD and the Pearl River Delta (PRD)).The large domain is centered at (118• E, 32 34.7 • C at 12:00 on 17 August 2010.The highest solar radiation levels were 786 and 571 W m −2 on 16 and 17 August, respectively.Under the high atmospheric oxidation conditions, the NO was rapidly oxidized to NO 2 , which is later converted to O 3 by photolytic destruction.

Table 1 .
Statistical results between MM5 model and observation data at surface stations in Shanghai.

Table 2 .
Statistical results between CMAQ model and observation data during 5-31 August 2010.

Table 3 .
Comparisons between the modeled hourly, max and min data against observations of surface O 3 , NO 2 , NO x and NO y during 5-31 August 2010.

Table 2 .
Statistical results between CMAQ model and observation data during 5--31 August 2010.