Impacts of thermal circulations induced by urbanization on ozone formation in the Pearl River Delta region, China

Atmospheric Environment 127 (2016) 382e392

Contents lists available at ScienceDirect

Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv

Impacts of thermal circulations induced by urbanization on ozone formation in the Pearl River Delta region, China Mengmeng Li a, b, Yu Song a, *, Zhichun Mao a, c, Mingxu Liu a, Xin Huang d a

State Key Joint Laboratory of Environmental Simulation and Pollution Control, Department of Environmental Science, Peking University, Beijing 100871, China b School of Atmospheric Sciences, Nanjing University, Nanjing 210023, China c 68028 Troops of the Lanzhou Military Area Command, Lanzhou 730058, China d Institute for Climate and Global Change Research & School of Atmospheric Sciences, Nanjing University, Nanjing, 210023, China

h i g h l i g h t s
The UHI effect enhances turbulent mixing and modifies thermal circulations.
The deeper urban ABL and intense updraft are favorable for pollutant dilution.
NOx dilution weakens O3 photochemical production and titration destruction.
Surface UHI breeze and intensified sea breeze trap O3 through advection.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 23 May 2015 Received in revised form 22 September 2015 Accepted 27 October 2015 Available online 30 October 2015

Thermal circulations induced by urbanization could exert important effects on regional ozone (O3) formation through regulating the chemical transformations and transport of O3 and its precursors. In this study, the Weather Research and Forecasting/Chemistry (WRF/Chem) model combined with remote sensing are used to investigate the impacts of urbanization-induced circulations on O3 formation in the Pearl River Delta (PRD) region, China. The urban heat island (UHI) effect in PRD significantly enhances turbulent mixing and modifies local circulations, i.e., initiates the UHI circulation and strengthens the sea breeze, which in turn cause a detectable decrease of daytime O3 concentration (1.3 ppb) and an increase of O3 (þ5.2 ppb) around the nocturnal rush-hours. The suppressed O3 titration destruction due to NOx dilution into the deeper urban boundary layer (200e400 m) is the main reason for elevated nocturnal O3 levels. In the daytime, however, the upward transport of O3 precursors weakens nearsurface O3 photochemical production and conversely enhances upper-level O3 generation. Furthermore, the surface UHI convergence flow and intensified sea breeze act to effectively trap O3 at the suburban and coastal regions. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Urbanization Thermal circulation Ozone PRD

1. Introduction Urbanization has been recognized as one of the most important aspects of global land use and land cover changes, particularly in the most rapidly developing countries (Lambin et al., 2003; Liu and Tian, 2010). In China, urban areas expanded by 35.7% from 1980 to 2005 due to the accelerated industrialization and economic growth (Liu and Tian, 2010), and is projected to reach 48e50% by 2020 (Chen, 2007).

* Corresponding author. E-mail address: [email protected] (Y. Song). http://dx.doi.org/10.1016/j.atmosenv.2015.10.075 1352-2310/© 2015 Elsevier Ltd. All rights reserved.

Modifications of the urban surface characteristics (e.g., albedo, emissivity and roughness etc.) can result in perceivable alterations of atmospheric thermal structures and circulations (de Foy et al., 2006; Li et al., 2014), boundary layer dynamics (Ludwig and Dabberdt, 1973) and precipitation (Yang et al., 2014). These changes in turn could profoundly impact the transport, chemical transformations and removal of pollutants in atmospheric boundary layer (ABL) (Ryu et al., 2013; Taha, 2008). Tropospheric ozone (O3), produced through complex photochemical reactions involving nitrogen oxides (NO and NO2, referred to collectively as NOx) and volatile organic compounds (VOCs), has continued to be a critical environmental problem in major urban centers of China (Ran et al., 2009; Zhang et al., 2008). Tropospheric

M. Li et al. / Atmospheric Environment 127 (2016) 382e392

383

O3 formation is closely related to urban climate. Tao et al. (2013) modeled the effects of land cover changes on some key atmospheric processes in U.S. and found that urbanization caused more high O3 (>70 ppb) occurrences. Yu et al. (2012) and Wang et al. (2007) reported that the O3 concentrations in China's three largest metropolitan areas increased by 5e20 ppb owing to increased temperature and turbulence, and reduced wind speed. Under calm conditions, thermal circulations that arise from surface heterogeneity are also critical to the transport of air pollutants (Fast et al., 2000; Jazcilevich et al., 2002). Urban heat island (UHI) circulation has been documented in numerous observations (Fan et al., 2011; Wong and Dirks, 1978) or modeling studies (Li et al., 2014; Lo et al., 2006; Ryu et al., 2013). Ryu et al. (2013) highlighted that urban breeze could transport high-reactivity precursors and O3-rich air to Seoul, leading to a 13e16 ppb increase in O3 concentrations. At coastal urban regions, recirculation of O3 and its precursors by UHI and sea/land breeze circulations is suggested as the major contributor to deteriorative O3 air quality (Levy et al., 2008; Martins et al., 2012). The Pearl River Delta (PRD) region, located in Guangdong province of southern China, is one of China's three largest urban agglomerations and most densely populated areas. In the recent decades, the PRD region suffered from dramatic urban expansion (Seto et al., 2002) and air quality deterioration (Huang et al., 2005; Wang et al., 2009), with surface O3 concentrations frequently exceeding national standards (93.3 ppb). Numerous studies have focused on the impacts of urbanization on O3 air quality in PRD (Lo et al., 2006; Wang et al., 2007). Yet, many critical processes that are related to the interactions between thermal circulations and O3 formation are still poorly understood. Moreover, Li et al. (2014) suggest that current mesoscale models lack correct representations of vegetation abundance in urban areas, which is critical to capture the major features of UHI and local thermal circulations. This study aims to investigate the impacts of thermal circulations induced by urbanization on O3 formation in PRD, using the Weather Research and Forecasting/Chemistry (WRF/Chem) model coupled with remote sensing. Model setup and experiment designs are described in Section 2. Section 3 discusses the impacts of urbanization on thermal circulations and O3 formation. Finally, a summary and conclusion are presented in Section 4.

The 2008-year Multi-resolution Emission Inventory for China (MEIC) with 0.25  0.25 resolution is used for the outer domain. The PRD regional emission inventory in 2006 developed by Zheng et al. (2009) is incorporated into the inner domain. Inventories for emissions of open-field biomass burning are compiled by Song et al. (2009) and Huang et al. (2012). Biogenic VOC emissions are calculated online using Model of Emissions of Gases and Aerosols from Nature (MEGAN) (Guenther et al., 2006). The CBM-Z gas-phase chemical mechanism (Zaveri and Peters, 1999) and MOSAIC aerosol module (Zaveri et al., 2008) are used in WRF/Chem. The physical parameterized options contain the Noah land surface scheme (Ek et al., 2003), the Lin microphysics scheme (Lin et al., 1983), the Grell cumulus scheme (Grell and Devenyi, 2002), the Yonsei University (YSU) boundary layer scheme (Noh et al., 2003), the Goddard short-wave radiation scheme (Chou and Suarez, 1999), the Rapid Radiative Transfer Model (RRTM) long-wave radiation scheme (Gallus and Bresch, 2006) and the Monin-Obukhov surface similarity scheme (Monin and Obukhov, 1954). Urban canopy module is not considered for lack of urban geometry and energy parameters. To evaluate the contributions of individual process to changes in chemical species concentrations, a tool of tendency diagnostic analysis is provided (Jiang et al., 2012; Tao et al., 2015). The change rates of species concentrations are described by a set of mass continuity equations:

2. Methods and data

Two simulations that are designated as “URBAN1990” and “URBAN2006” are conducted, using two satellite data setsdAdvanced Very High Resolution Radiometer (AVHRR) and Moderate Resolution Imaging Spectroradiometer (MODIS)dto portray the land cover type, green vegetation fraction (GVF) and leaf area index (LAI) in the early 1990s and 2006, respectively. In both land-use experiments, identical parameterization schemes, initial conditions and emissions are used. In URBAN1990, the U.S. Geological Survey (USGS) global 1-km land cover map (Fig. 2a), which is derived from monthly AVHRR Normalized Difference Vegetation Index (NDVI) over April 1992 to March 1993 (Loveland et al., 2000), is used in WRF. GVF is determined using AVHRR NDVI collected from 1985 through 1990 at a resolution of 20 km (Gutman and Ignatov, 1998) (Fig. S1a). LAI is calculated online using the tabulated value assigned to each land/ vegetation type (Fig. S1c). The URBAN2006 case is conducted with updated land surface parameters (including land cover type, GVF and LAI) from MODIS observations in 2006 to generate urbanization-altered meteorological fields. The MODIS Land Cover Type Product (MCD12Q1) in 2006 and Water Mask Product (MOD44W) in 2000, both with a resolution of 500 m (Justice et al., 2002), are combined to construct a new land cover map in PRD (Fig. 2b). The monthly, 1-km MODIS NDVI Product (MOD13A2) is used to estimate GVF (Purevdorj et al.,

2.1. WRF/Chem modeling system WRF/Chem is an online three-dimensional, Eulerian chemical transport model that considers the complex atmospheric physical and chemical processes (Grell et al., 2005). It is configured with a two-nested grid system (Fig. 1) centered at 22.33 N, 114.05 E, with horizontal grid spacings of 12 and 4 km and grid points of 92  82 and 94  88, respectively. The outer domain comprises southern China and extends into the South China Sea (Fig. 1a). The inner domain focuses on the PRD region that contains nine cities, in addition to Hong Kong and Macao: Guangzhou, Shenzhen, Foshan, Dongguan, Huizhou, Jiangmen, Zhongshan, Zhaoqing and Zhuhai. 24 vertical layers are spaced unequally from the ground to 50 hPa, with 8 layers (model full-s levels are 1.0, 0.993, 0.983, 0.97, 0.954, 0.934, 0.909 and 0.88) located below 1 km to resolve the boundary layer process, as in many regional simulations (Hu et al., 2010; Shin and Hong, 2011; Wang et al., 2010). 12-hour spin-up time is allowed for each 60-h model cycle to minimize the influence of initial conditions (Berge et al., 2001). The National Centers for Environmental Prediction/Final (NCEP/FNL) meteorological data are used to initialize each model cycle. The chemical initial fields are provided with outputs from the previous model cycle.

 
 vC vC ¼ V$ VC adv þ V$ðKe VCÞdiff þ Pchem  Lchem þ E þ vt vt dry       vC vC vC þ þ þ vt wdep vt cloud vt aer (1) where C is the species concentration, ðVÞ is the velocity vector, Ke is the eddy diffusion coefficient. The right-hand process terms successively denote advection, turbulent diffusion, chemical production (Pchem) and loss (Lchem), emission (E), dry and wet deposition, in-cloud process and aerosol-phase process. 2.2. Experiment designs

384

M. Li et al. / Atmospheric Environment 127 (2016) 382e392

Fig. 1. (a) Two-nested WRF/Chem modeling domains. (b) Locations of meteorological (blue dots) and air quality (red dots) monitoring sites. The shaded contour in panel (b) represents topography height (m); the air quality monitoring sites are labeled as follows: S1-Chenzhong, S2-Shunde, S3-Huijingcheng, S4-Donghu, S5-Zimaling, S6-Tangjia, S7Jinguowan, S8-Xiabu, S9-Liyuan, S10-Luhu, S11-Wanqingsha, S12-Tianhu, S13-Tamen, S14-Dongyong and S15-Quanwan. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 2. Land cover maps in (a) URBAN1990 and (b) URBAN2006.

1998) (Fig. S1b). The LAI distribution is directly substituted with MODIS LAI (MCD15A2), with a spatial resolution of 1 km and an 8day temporal resolution (Knyazikhin et al., 1998) (Fig. S1d). The assessment by Schneider et al. (2009) confirms that MODIS urban land cover has high accuracy (93%) at both pixel and city levels. From the early 1990s to 2006, an unprecedented urban expansion (~27 times) and cropland loss occurred in PRD, especially in the central regions along the Pearl River Estuary (PRE) (Fig. 2). Correspondingly, less vegetation cover (<0.1) and lower LAI (1.5e2.0) are observed in urban areas (Fig. S1). This study uses a month-long simulation lasting from 1 to 31 July in 2006, and focuses on a high-O3 pollution event during 19e24 July. Examination of the Korean Meteorological Administration (KMA) synoptic chart (Fig. 3) shows that during this episode, the PRD region was controlled by a stagnant high-pressure system, caused by the approaching tropical cyclone “KAMEI” originating in the South China Sea. The weak winds, strong solar radiation and high temperatures associated with the high-pressure system result in this severe and prolonged high-O3 episode. Moreover, the weak synoptic forcing and calm conditions in PRD also provide optimal conditions for the development of thermal circulations.

3. Results and discussions In the following sections, model evaluations and comprehensive impacts of urbanization on O3 air quality will first be presented for the 1-month simulation. Detailed discussions on the role of urbanization-induced thermal circulations in O3 formation, however, will focus on the 19e24 July episode. 3.1. Model validation 3.1.1. Meteorology Meteorological outputs from URBAN2006 are examined against hourly observations at 72 surface automatic stations throughout the inner domain (Fig. 1b). The statistical terms include mean bias (MB), root mean square error (RMSE), fractional bias (FB), fractional error (FE) and index of agreement (IOA) (Table 1). The WRF model performance was well within the typical ranges of mesoscale numerical models (Hanna and Yang, 2001; Li et al., 2014) and U.S. EPA benchmarks recommended by Emery et al. (2001) (e.g., MB: ±0.5  C, IOA: 0.8 for temperature, MB: ±0.5 m s1, IOA: 0.6 for wind speed; MB: ±10.0 for wind

M. Li et al. / Atmospheric Environment 127 (2016) 382e392

385

Fig. 3. 500 hPa KMA weather map at 08:00 LT on 22 July.

Table 1 Performance statistics for URBAN2006 meteorological simulation. Index

Variable

a

MB RMSEb FBc FEd IOAe a

b c

d e

MB ¼ N1

Temperature ( C)

Relative humidity (%)

Wind speed (m s

0.55 2.21 0.02 0.06 0.85

7.04 12.84 0.09 0.14 0.77

0.95 2.30 0.26 0.61 0.54

1

)

Wind direction ( ) 14.79 76.87 e e e

PN

i¼1 ðsimi  obsi Þ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi PN 2 RMSE ¼ i¼1 ðsimi  obsi Þ =N pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi FB ¼ 2 ðsimi  obsi Þ=ðsimi þ obsi Þ=N qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi FE ¼ jsimi  obsi j=ðsimi þ obsi Þ2 =N

 2  , where sim and obs refer to the simulated results and observations, respectively; obs represents the averaged observation values; N IOA ¼ 1  PN  NRMSE ð obsi obsþsimi obsÞ2 i¼1

represents the number of data pairs.

direction). 2-m temperature was slightly underestimated, with a cold bias of 0.55  C and RMSE of 2.21  C. In terms of 2-m relative humidity, the simulation has high moist bias (7.04%) and RMSE (12.84%). Generally, model overestimation of wind speed is evident (0.95 m s1) during the modeling period. A high positive bias in wind speed has also been reported by several other studies using WRF (Li et al., 2014; Matsui et al., 2009; Molders et al., 2012; Tuccella et al., 2012), which may be because that it's difficult to resolve the weak winds and complex topographical features realistically in current generations of mesoscale models (Cheng and Steenburgh, 2005; Zhang et al., 2015). Mean angular errors for wind directions are within 80 and a MB of 14.79 is seen. 3.1.2. O3 and NO2 Hourly ground-level O3 and NO2 concentrations collected at 15 air quality monitoring stations (Fig. 1b) are used for model

evaluations of chemical composition. The magnitudes and diurnal/daily trends for O3 concentrations were reasonably predicted at most sites (Fig. S2), with an overall correlation coefficient of 0.72 and normalized mean bias (NMB) of 23%. At some sites (e.g., Donghu, Chengzhong, and Quanwan), the O3 concentrations were overestimated by 10e30%. Overestimation of summertime O3 is thought to be a common problem over East Asia due to the model inability to accurately describe cloud cover and monsoon rainfall, as well as southerly inflow of clean marine air masses (Chatani and Sudo, 2011; Lin et al., 2009). It is also noted that the O3 peak concentrations tended to be under-predicted by about 20% at a few upwind rural sites in eastern PRD (e.g., Jinguowan and Xiabu), which may be caused by the uncertainties in precursor emissions, chemical mechanisms in low-NOx regimes and wind errors. As an important O3 precursor, model evaluations of NO2 concentrations further demonstrate that O3 formation is reasonably

386

M. Li et al. / Atmospheric Environment 127 (2016) 382e392

captured, with a domain-average NMB of 9%. Fig. S3 presents the time series of NO2 concentrations at several representative sites. NO2 mixing ratios are influenced by local effects, such as emissions and diffusion conditions. Thus, the model overestimates NO2 within 10 ppb around the source regions (e.g., Shunde and Tangjia) and underestimates NO2 at suburban sites (e.g., Jinguowan and Xiabu) (Fig. S3). 3.2. Thermal circulations induced by urbanization Distributions of modeled land surface temperature (LST) (Fig. 4a, b) and MODIS measurements (Fig. 4c) during 19e24 July imply that as the urban size expands, a typical UHI is generated in URBAN2006. The averaged LST over urban grids in URBAN2006 is 3.7  C higher than that of the URBAN1990 case (Fig. 4a, b). The comparisons of 2-m air temperature between URBAN2006 and URBAN1990 also exhibit a remarkable warming effect within cities, that ranges from 0 to 2.3  C (Fig. 5a). The urban warming displays a marked diurnal variation, which maximizes (~2.3  C) around sunset (18:00e20:00 local time (LT)) due to the massive release of daytime

ground heat storage, remains nearly constant at 1.8  C overnight, and grows weaker around 07:00e12:00 LT as the mixed layer builds (Fig. 5a). These changes in thermal structures can profoundly alter ABL evolution (Fig. 5a), another crucial variable controlling pollutant dispersion. As expected, a deeper urban boundary layer is generated over the megacities in URBAN2006 (Fig. 5a), which could be explained by the enhanced surface heating buoyancy and lessstabilized air. The maximum increase (200e400 m) in ABL height appears in the afternoon and around sunset (11:00e19:00 LT). Not only the ABL structures, but also the flow patterns are altered, as reported in our previous work (Li et al., 2014). As a result of elevated urban/rural temperature and pressure gradients in URBAN2006, obvious urban breeze and wind convergence (112.9 E113.4 E), that are characterized by faster inflow of cooler rural air toward cities, turn to be dominant across the urban/rural borders in the afternoon and evening (Fig. 6b). Moreover, intense thermal updraft (0.2e0.6 m s1) is present above urban centers, e.g., Guangzhou-Foshan (Fig. 6b) and Dongguan (Fig. 6d). Particularly, it is seen that around sunset, a closed UHI circulation is

Fig. 4. Simulated and observed LST ( C) during 19e24 July: (a) URBAN1990, (b) URBAN2006 and (c) MODIS.

M. Li et al. / Atmospheric Environment 127 (2016) 382e392

387

Fig. 5. Diurnal variations for differences of (a) 2-m temperature ( C) and ABL height (m), and (b) NOx and O3 concentrations (ppb) over urban grids between URBAN2006 and URBAN1990 during 19e24 July.

Fig. 6. Cross sections of wind and temperature ( C) (shaded contour) in URBAN1990 at (a) 19:00 LT along AA0 , and (c) 15:00 LT along BB0 in Fig. 4b on 22 July. The (b) and (d) panels are the same as (a) and (c), respectively, but for URBAN2006. The green and blue bars at the bottom represent urban areas and oceans, respectively; the black-shaded areas represent terrain; the blue boxes in panel (b) and (d) mark UHI and sea breeze circulations; the reference magnitude for vertical wind speed is 1 m s1. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

initiated in central PRD (Fig. 6b). Contrast between the two simulations (Fig. S4) also shows an increased wind speed around cities, which is consistent with the features of UHI circulation. The evolution of sea breeze circulation is also modified by

coastal urbanization. For example, at 15:00 LT on 22 July, the sea breeze dominates PRE (113.5 E113.9 E) and penetrates on land into Dongguan and Foshan in both simulations (Fig. 6c, d and Fig. S5). In URBAN2006, however, the sea breeze towards Dongguan

388

M. Li et al. / Atmospheric Environment 127 (2016) 382e392

Fig. 7. Spatial differences of (a) O3 and (b) NOx (ppb) between URBAN2006 and URBAN1990 at 18:00e24:00 LT during 19e24 July.

is stronger due to increased land/sea temperature contrast induced by coastal urban heating. In the contrast map (Fig. S4a), stronger wind speed (>0.6 m s1) can also be seen over the PRE, indicating the enhancement of daytime sea breeze. 3.3. Impacts of urbanization-induced thermal circulations on O3 formation 3.3.1. Comparisons of O3 formation process in both urban scenarios Urbanization has considerable impacts on O3 formation. Fig. S6 presents the monthly differences for simulated NOx and O3

concentrations over urban grids between URBAN2006 and URBAN1990. The O3 differences are more substantial in urban areas (Fig. S6b). Overall, urbanization causes an increase of O3 concentrations (þ1.8 ppb) and a reduction of NOx concentrations (3.1 ppb) over urban grids (Fig. S6a). Noticeably, the maximum O3 increase (þ5.8 ppb) and NOx reduction (9.8 ppb) occur around the nocturnal rush-hours. Similar with the monthly-average results (Fig. S6), impacts of urbanization on pollutant redistribution during the 19e24 July episode are more significant in cities and their downwind areas (Fig. 7). In the metropolitan regions, the deeper ABL (Fig. 5a) and

Fig. 8. (a) Process contribution (ppb hr1) to O3 over urban grids in URBAN2006 and (b) its differences (ppb hr1) with URBAN1990. NETP represents the net contribution of chemical, advection and diffusion processes.

M. Li et al. / Atmospheric Environment 127 (2016) 382e392

intense upward flow (Fig. 6) are favorable for pollutant dilution. Thus, the URBAN2006 run generates remarkably lower NOx concentrations by 6.6 ppb on average (Fig. 5b), as also noticed by Sarrat et al. (2006) in Paris and Ryu et al. (2013) in Seoul. The NOx decrease is particularly large (10e15 ppb) around nocturnal rushhours (18:00e22:00 LT), when the increase of ABL height (Fig. 5a) and vertical velocity (Fig. 6b) remain large, and simultaneously anthropogenic emissions in conjunction with rush-hour traffic are strong. The aforementioned meteorological changes in URBAN2006 collectively cause a 1.3 ppb reduction of O3 concentration from 12:00 to 17:00 LT over urban grids, and an O3 enhancement by 5.2 ppb in the late afternoon and early evening (18:00e24:00 LT) (Fig. 5b). Tendency diagnostic analyses for both experiments are performed over urban grids at the lowest two layers (below 100 m) (Fig. 8). Near-surface O3 formation is influenced by three major processes, namely chemical reactions (CHEM), vertical mixing coupled with dry deposition into a single process term (VMIX), and advection (ADVT, including horizontal and vertical components). The timing (Fig. 5b) and area (Fig. 7) with maximum O3 increase (>10 ppb) coincide with that of sharp NOx reduction around sunset. Thus, the elevated ground-level O3 concentration is presumably attributed to suppressed NOx titration destruction in the nighttime, which is also confirmed in the contrast map of chemical

389

contribution (Fig. 8b). Beyond that, under the urbanized conditions, more O3-rich air aloft is brought downward to the ground (~10 ppb h1) via enhanced turbulence mixing (Fig. 8b). Last, dry deposition is a major path for O3 removal (Seinfeld and Pandis, 2006). The dry deposition velocity of O3 over less-vegetated urban surface (e.g., 0.25 cm s1 in the daytime) is almost half of that in URBAN1990 (0.49 cm s1), which efficiently reduces O3 removal. Opposite to these positive contributions, more surface O3 is vented out by the vigorous upward motions, leading to a net O3 advective loss (Fig. 8b). All these processes combine to cause a daytime O3 reduction and a nocturnal O3 accumulation in urban areas (Fig. 5b). Impacts of urbanization on O3 vertical distributions are discussed in the supplementary material (Text S1). 3.3.2. Impacts on O3 transport by urbanization-induced thermal circulations The role of UHI circulation in the horizontal and vertical transport fluxes of O3 is discussed in the supplementary material (Text S2). As pointed out, UHI circulation causes a net O3 advective loss over urban areas (Fig. 8b), indicating that the upward export of O3 by urban updraft surpasses the horizontal input induced by urban breeze. Over the suburban regions, however, the enhanced urban breeze may act to effectively accumulate O3. Fig. 9 shows the

Fig. 9. Cross sections of (a, b) O3 concentration (ppb) and wind along AA0 at 19:00 LT on 22 July. Differences of (c) advection contributions and (d) chemical contributions to O3 (ppb hr1) between URBAN2006 and URBAN1990. The red bar represents the extent of urban area.

390

M. Li et al. / Atmospheric Environment 127 (2016) 382e392

Fig. 10. Cross sections of (a) O3 concentration (ppb) and (c) chemical contribution to O3 (ppb hr1) in URBAN1990 along BB0 at 15:00 LT on 22 July. The (b) panel is the same as (a), but for URBAN2006. The (d) and (e) panels are differences of chemical and advection contributions to O3 (ppb hr1) between URBAN2006 and URBAN1990. The red and blue bars represent the extent of urban areas and oceans, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

cross sections of O3 concentrations and process contributions along AA’ (depicted in Fig. 4b) at 19:00 LT on 22 July, when a complete UHI circulation develops and the maximum increase of O3 (~30 ppb) occurs (Fig. 9a, b). In the late afternoon, as the

prevailing wind turns to urban breeze, O3-rich air is transported back from the countryside and accumulated around the urban/ rural borders, as can be inferred from the largely increased advection contributions near city boundaries (112.8 E113.6 E)

M. Li et al. / Atmospheric Environment 127 (2016) 382e392

within 0.5 km (Fig. 9c and Fig. S9a). The enhancement of sea breeze is also critical for the photochemical production and cross-city transport of O3. Fig. 10 presents the cross sections of O3 concentrations and process contributions along BB’ (depicted in Fig. 4b) at 15:00 LT on 22 July. Along the northeastern coast of PRE (113.8 E114.0 E), plentiful NOx is accumulated in the shallow marine boundary layer, which inhibited O3 production (Fig. 10c). The stronger mixing embedded in the sea breeze inflow induced by coastal urbanization leads to markedly lower NOx (Fig. S10) and reduced O3 chemical destruction along the coastlines, as can be inferred from the contrast map of chemical contributions (Fig. 10d). Examination of the advection contribution (Fig. 10e) also implies that the intensified sea breeze towards Dongguan traps O3-laden marine air near the coastal areas, leading to a positive advection contribution across land/sea boundaries. Consequently, the modified sea breeze over PRE may exacerbate coastal O3 pollution by 10e20 ppb (Fig. 10a, b).

391

canopy processes are not considered in this study. Feng et al. (2014) find that anthropogenic heat release causes a 0.5e1  C temperature increase in China's three urban clusters. Using multiple urban canopy schemes, Liao et al. (2014) suggest that O3 concentrations in the Yangtze River Delta region increase by 10% in January and decrease by 15% in July. All these studies highlight a necessary use of advanced urban modules. Additionally, as suggested by Chen et al. (2014), in addition to meteorological processes, the effects of increasing anthropogenic emissions induced by urbanization are also worthy of further investigation. It is predicted that 60% of China's population will live in cities by 2030 (Chen, 2007) and China's urban clusters will expand at a rate of 3.0e5.1%/decade from 2010 to 2100 (Yuan et al., 2013). Therefore, this work not only helps understand the role of thermal circulations in O3 formation, but also sets a stage for better consideration of future improvement of urban environments. Acknowledgments

3.3.3. Impacts on O3-NOx-VOC chemistry by UHI circulations Nighttime O3 chemistry is governed by NOx titration loss (NO þ O3 / NO2 þ O2) (Seinfeld and Pandis, 2006). Comparisons between the two simulations imply that, following the enhanced turbulent mixing and updraft in URBAN2006, the NOx plume is transported deeper into the free atmosphere (>3 km) and surface NOx concentration reduces by over 10 ppb (Fig. S10a, b). The lower NOx budget weakens O3 titration destruction, showing a chemical enhancement to O3 (þ10 ppb h1) over most urban grids in the vertical (Fig. 9d) and horizontal contrast maps (Fig. S9b). The daytime O3 photochemistry represents a more complex process, depending on the relative balance between NOx and VOC. In the afternoon, O3 is produced via photochemical reactions over most inland areas, e.g., Dongguan and Zhongshan in URBAN1990 (Fig. 10c). Owing to the enhanced urban turbulent mixing and upward transport by the rising branch of UHI circulation, NOx concentrations significantly reduce around urban centers (Fig. S10c, d). As a result, the photochemical generation of O3 over the inland regions is reduced in the lower ABL; otherwise, O3 production is markedly enhanced in the upper ABL (around 1.0e2.8 km), as can be inferred from the chemical comparisons (Fig. 10d). Along the coastlines, however, NOx dilution tends to favor O3 photochemical generation, as has been discussed in Section 3.3.2. 4. Conclusions The interactions between urbanization and thermal circulations play an important role in regional O3 formation. In this study, the WRF/Chem modeling system coupled with remote sensing are used to investigate the impacts of thermal circulations induced by urbanization on O3 formation in the PRD region. A typical UHI is generated in PRD. This UHI effect enhances turbulent mixing and modifies local circulations, i.e., initiates the UHI circulation and strengthens the sea breeze over PRE. Overall, urbanization causes a detectable decrease of daytime O3 concentration by 1.3 ppb, and an increase of O3 by 5.2 ppb around the nocturnal rush-hours. The deeper urban boundary layer (200e400 m) and rising branch (0.2e0.6 m s1) of UHI circulation are favorable for NOx dilution, thus generating lower NOx concentrations (6.6 ppb) over urban regions. The dilution of O3 precursors contributes to the elevated nocturnal O3 levels through suppressed NOx titration destruction, and conversely weakens low-level O3 photochemical production in the daytime. Furthermore, the surface UHI convergence and intensified sea breeze at coastal urbanized regions may act to exacerbate O3 pollution through pollutant recirculation. Some issues still exist in this study. At present, detailed urban

This study is supported by National Natural Science Foundation of China (41275155 and 41121004), the Public Welfare Project for Environmental Protection (201309009), the project 2010CB428501 supported by Chinese Ministry of Science and Technology, and the special fund of State Key Joint Laboratory of Environment Simulation and Pollution Control (14Y01ESPCP). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.atmosenv.2015.10.075. References Berge, E., Huang, H.C., Chang, J., Liu, T.H., 2001. A study of the importance of initial conditions for photochemical oxidant modeling. J. Geophys. Res. Atmos. 106, 1347e1363. Chatani, S., Sudo, K., 2011. Influences of the variation in inflow to East Asia on surface ozone over Japan during 1996-2005. Atmos. Chem. Phys. 11, 8745e8758. Chen, B., Yang, S., Xu, X.D., Zhang, W., 2014. The impacts of urbanization on air quality over the Pearl River Delta in winter: roles of urban land use and emission distribution. Theor. Appl. Climatol. 117, 29e39. Chen, J., 2007. Rapid urbanization in China: a real challenge to soil protection and food security. Catena 69, 1e15. Cheng, W.Y.Y., Steenburgh, W.J., 2005. Evaluation of surface sensible weather forecasts by the WRF and the Eta models over the western United States. Weather Forecast 20, 812e821. Chou, M.D., Suarez, M.J., 1999. NASA/TM-104606. A Shortwave Radiation Parameterization for Atmospheric Studies, vol. 15, p. 40. de Foy, B., Molina, L.T., Molina, M.J., 2006. Satellite-derived land surface parameters for mesoscale modelling of the Mexico City basin. Atmos. Chem. Phys. 6, 1315e1330. Ek, M.B., Mitchell, K.E., Lin, Y., Rogers, E., Grunmann, P., Koren, V., Gayno, G., Tarpley, J.D., 2003. Implementation of Noah land surface model advances in the National Centers for Environmental Prediction operational mesoscale Eta model. J. Geophys. Res. Atmos. 108, 8851. http://dx.doi.org/10.1029/ 2002jd003296. Emery, C., Tai, E., Yarwood, G., 2001. Enhanced Meteorological Modeling and Performance Evaluation for Two Texas Episodes. A report prepared for the Texas Natural Resources Conservation Commission (TNRCC), ENVIRON International Corporation, Novato, CA. Fan, S.J., Fan, Q., Yu, W., Luo, X.Y., Wang, B.M., Song, L.L., Leong, K.L., 2011. Atmospheric boundary layer characteristics over the Pearl River Delta, China, during the summer of 2006: measurement and model results. Atmos. Chem. Phys. 11, 6297e6310. Fast, J.D., Doran, J.C., Shaw, W.J., Coulter, R.L., Martin, T.J., 2000. The evolution of the boundary layer and its effect on air chemistry in the Phoenix area. J. Geophys. Res. Atmos. 105, 22833e22848. Feng, J.M., Wang, J., Yan, Z.W., 2014. Impact of anthropogenic heat Release on Regional Climate in three vast urban agglomerations in China. Adv. Atmos. Sci. 31, 363e373. Gallus, W.A., Bresch, J.F., 2006. Comparison of impacts of WRF dynamic core, physics package, and initial conditions on warm season rainfall forecasts. Mon. Weather Rev. 134, 2632e2641. Grell, G.A., Devenyi, D., 2002. A generalized approach to parameterizing convection combining ensemble and data assimilation techniques. Geophys. Res. Lett. 29,

392

M. Li et al. / Atmospheric Environment 127 (2016) 382e392

1693. http://dx.doi.org/10.1029/2002gl015311. Grell, G.A., Peckham, S.E., Schmitz, R., McKeen, S.A., Frost, G., Skamarock, W.C., Eder, B., 2005. Fully coupled “online” chemistry within the WRF model. Atmos. Environ. 39, 6957e6975. Guenther, A., Karl, T., Harley, P., Wiedinmyer, C., Palmer, P.I., Geron, C., 2006. Estimates of global terrestrial isoprene emissions using MEGAN (Model of Emissions of Gases and Aerosols from Nature). Atmos. Chem. Phys. 6, 3181e3210. Gutman, G., Ignatov, A., 1998. The derivation of the green vegetation fraction from NOAA/AVHRR data for use in numerical weather prediction models. Int. J. Remote Sens. 19, 1533e1543. Hanna, S.R., Yang, R.X., 2001. Evaluations of mesoscale models' simulations of nearsurface winds, temperature gradients, and mixing depths. J. Appl. Meteorol. 40, 1095e1104. Hu, X.M., Nielsen-Gammon, J.W., Zhang, F.Q., 2010. Evaluation of three planetary boundary layer schemes in the WRF model. J. Appl. Meteorol. Clim. 49, 1831e1844. Huang, J.P., Fung, J.C.H., Lau, A.K.H., Qin, Y., 2005. Numerical simulation and process analysis of typhoon-related ozone episodes in Hong Kong. J. Geophys. Res. Atmos. 110, D05301. http://dx.doi.org/10.1029/2004jd004914. Huang, X., Li, M.M., Li, J.F., Song, Y., 2012. A high-resolution emission inventory of crop burning in fields in China based on MODIS Thermal Anomalies/Fire products. Atmos. Environ. 50, 9e15. Jazcilevich, A.D., Garcia, A.R., Ruiz-Suarez, L.G., 2002. A modeling study of air pollution modulation through land-use change in the Valley of Mexico. Atmos. Environ. 36, 2297e2307. Jiang, F., Zhou, P., Liu, Q., Wang, T.J., Zhuang, B.L., Wang, X.Y., 2012. Modeling tropospheric ozone formation over East China in springtime. J. Atmos. Chem. 69, 303e319. Justice, C.O., Townshend, J.R.G., Vermote, E.F., Masuoka, E., Wolfe, R.E., Saleous, N., Roy, D.P., Morisette, J.T., 2002. An overview of MODIS land data processing and product status. Remote Sens. Environ. 83, 3e15. Knyazikhin, Y., Martonchik, J.V., Myneni, R.B., Diner, D.J., Running, S.W., 1998. Synergistic algorithm for estimating vegetation canopy leaf area index and fraction of absorbed photosynthetically active radiation from MODIS and MISR data. J. Geophys. Res. Atmos. 103, 32257e32275. Lambin, E.F., Geist, H.J., Lepers, E., 2003. Dynamics of land-use and land-cover change in tropical regions. Annu. Rev. Env. Resour. 28, 205e241. Levy, I., Dayan, U., Mahrer, Y., 2008. A five-year study of coastal recirculation and its effect on air pollutants over the East Mediterranean region. J. Geophys. Res. Atmos 113, D16121. http://dx.doi.org/10.1029/2007jd009529. Li, M.M., Song, Y., Huang, X., Li, J.F., Mao, Y., Zhu, T., Cai, X.H., Liu, B., 2014. Improving mesoscale modeling using satellite-derived land surface parameters in the Pearl River Delta region, China. J. Geophys. Res. Atmos. 119, 6325e6346. Liao, J.B., Wang, T.J., Wang, X.M., Xie, M., Jiang, Z.Q., Huang, X.X., Zhu, J.L., 2014. Impacts of different urban canopy schemes in WRF/Chem on regional climate and air quality in Yangtze River Delta, China. Atmos. Res. 145, 226e243. Lin, M., Holloway, T., Oki, T., Streets, D.G., Richter, A., 2009. Multi-scale model analysis of boundary layer ozone over East Asia. Atmos. Chem. Phys. 9, 3277e3301. Lin, Y.L., Farley, R.D., Orville, H.D., 1983. Bulk parameterization of the snow field in a Cloud model. J. Clim. Appl. Meteorol. 22, 1065e1092. Liu, M.L., Tian, H.Q., 2010. China's land cover and land use change from 1700 to 2005: estimations from high-resolution satellite data and historical archives. Glob. Biogeochem. Cy. 24, Gb3003. http://dx.doi.org/10.1029/2009gb003687. Lo, J.C.F., Lau, A.K.H., Fung, J.C.H., Chen, F., 2006. Investigation of enhanced cross-city transport and trapping of air pollutants by coastal and urban land-sea breeze circulations. J. Geophys. Res. Atmos. 111, D14104. http://dx.doi.org/10.1029/ 2005jd006837. Loveland, T.R., Reed, B.C., Brown, J.F., Ohlen, D.O., Zhu, Z., Yang, L., Merchant, J.W., 2000. Development of a global land cover characteristics database and IGBP DISCover from 1 km AVHRR data. Int. J. Remote Sens. 21, 1303e1330. Ludwig, F.L., Dabberdt, W.F., 1973. Effects of urbanization on turbulent diffusion and mixing depth. Int. J. Biometeorol. 17, 1e11. Martins, D.K., Stauffer, R.M., Thompson, A.M., Knepp, T.N., Pippin, M., 2012. Surface ozone at a coastal suburban site in 2009 and 2010: relationships to chemical and meteorological processes. J. Geophys. Res. Atmos. 117, D05306. http:// dx.doi.org/10.1029/2011jd016828. Matsui, H., Koike, M., Kondo, Y., Takegawa, N., Kita, K., Miyazaki, Y., Hu, M., Chang, S.Y., Blake, D.R., Fast, J.D., Zaveri, R.A., Streets, D.G., Zhang, Q., Zhu, T., 2009. Spatial and temporal variations of aerosols around Beijing in summer 2006: model evaluation and source apportionment. J. Geophys. Res. Atmos. 114, D00g13. http://dx.doi.org/10.1029/2008jd010906. Molders, N., Tran, H.N.Q., Cahill, C.F., Leelasakultum, K., Tran, T.T., 2012. Assessment of WRF/Chem PM2.5 forecasts using mobile and fixed location data from the Fairbanks, Alaska winter 2008/09 field campaign. Atmos. Pollut. Res. 3, 180e191. Monin, A.S., Obukhov, A.M., 1954. Basic laws of turbulent mixing in the surface layer of the atmosphere. Tr. Akad. Nauk. SSSR Geophiz. Inst. 24, 163e187.

Noh, Y., Cheon, W.G., Hong, S.Y., Raasch, S., 2003. Improvement of the K-profile model for the planetary boundary layer based on large eddy simulation data. Bound Lay Meteorol. 107, 401e427. Purevdorj, T., Tateishi, R., Ishiyama, T., Honda, Y., 1998. Relationships between percent vegetation cover and vegetation indices. Int. J. Remote Sens. 19, 3519e3535. Ran, L., Zhao, C.S., Geng, F.H., Tie, X.X., Tang, X., Peng, L., Zhou, G.Q., Yu, Q., Xu, J.M., Guenther, A., 2009. Ozone photochemical production in urban Shanghai, China: analysis based on ground level observations. J. Geophys. Res. Atmos. 114, D15301. http://dx.doi.org/10.1029/2008jd010752. Ryu, Y.H., Baik, J.J., Kwak, K.H., Kim, S., Moon, N., 2013. Impacts of urban landsurface forcing on ozone air quality in the Seoul metropolitan area. Atmos. Chem. Phys. 13, 2177e2194. Sarrat, C., Lemonsu, A., Masson, V., Guedalia, D., 2006. Impact of urban heat island on regional atmospheric pollution. Atmos. Environ. 40, 1743e1758. Schneider, A., Friedl, M.A., Potere, D., 2009. A new map of global urban extent from MODIS satellite data. Environ. Res. Lett. 4, 044003. http://dx.doi.org/10.1088/ 1748-9326/4/4/044003. Seinfeld, J.H., Pandis, S.N., 2006. Atmospheric Chemistry and Physics: from Air Pollution to Climate Change. Seto, K.C., Woodcock, C.E., Song, C., Huang, X., Lu, J., Kaufmann, R.K., 2002. Monitoring land-use change in the Pearl River Delta using Landsat TM. Int. J. Remote Sens. 23, 1985e2004. Shin, H.H., Hong, S.Y., 2011. Intercomparison of planetary boundary-layer parametrizations in the WRF model for a single day from CASES-99. Bound Lay Meteorol. 139, 261e281. Song, Y., Liu, B., Miao, W.J., Chang, D., Zhang, Y.H., 2009. Spatiotemporal variation in nonagricultural open fire emissions in China from 2000 to 2007. Glob. Biogeochem. Cy. 23, Gb2008. http://dx.doi.org/10.1029/2008gb003344. Taha, H., 2008. Urban surface modification as a potential ozone air-quality improvement strategy in california: a mesoscale modelling study. Bound Lay Meteorol. 127, 219e239. Tao, W., Liu, J., Ban-Weiss, G.A., Hauglustaine, D.A., Zhang, L., Zhang, Q., Cheng, Y., Yu, Y., Tao, S., 2015. Effects of urban land expansion on the regional meteorology and air quality of eastern China. Atmos. Chem. Phys. 15, 8597e8614. Tao, Z., Santanello, J.A., Chin, M., Zhou, S., Tan, Q., Kemp, E.M., Peters-Lidard, C.D., 2013. Effect of land cover on atmospheric processes and air quality over the continental United States - a NASA Unified WRF (NU-WRF) model study. Atmos. Chem. Phys. 13, 6207e6226. Tuccella, P., Curci, G., Visconti, G., Bessagnet, B., Menut, L., Park, R.J., 2012. Modeling of gas and aerosol with WRF/Chem over Europe: evaluation and sensitivity study. J. Geophys Res-Atmos 117, D03303. http://dx.doi.org/10.1029/ 2011jd016302. Wang, T., Wei, X.L., Ding, A.J., Poon, C.N., Lam, K.S., Li, Y.S., Chan, L.Y., Anson, M., 2009. Increasing surface ozone concentrations in the background atmosphere of Southern China, 1994-2007. Atmos. Chem. Phys. 9, 6217e6227. Wang, X., Liang, X., Jiang, W., Tao, Z., Wang, J., Liu, H., Han, Z., Liu, S., Zhang, Y., Grell, G., Peckham, S., 2010. WRF-Chem simulation of East Asian air quality: sensitivity to temporal and vertical emissions distributions. Atmos. Environ. 44, 660e669. Wang, X.M., Lin, W.S., Yang, L.M., Deng, R.R., Lin, H., 2007. A numerical study of influences of urban land-use change on ozone distribution over the Pearl River Delta region, China. Tellus B 59, 633e641. Wong, K.K., Dirks, R.A., 1978. Mesoscale perturbations on airflow in the urban mixing layer. J. Appl. Meteorol. 17, 677e688. Yang, L., Smith, J.A., Baeck, M.L., Bou-Zeid, E., Jessup, S.M., Tian, F.Q., Hu, H.P., 2014. Impact of urbanization on heavy convective precipitation under Strong largescale forcing: a Case study over the Milwaukee-Lake Michigan Region. J. Hydrometeorol. 15, 261e278. Yu, M., Carmichael, G.R., Zhu, T., Cheng, Y.F., 2012. Sensitivity of predicted pollutant levels to urbanization in China. Atmos. Environ. 60, 544e554. Yuan, Y.W., Zhao, T., Wang, W.M., Chen, S.H., Wu, F., 2013. Projection of the spatially explicit land Use/Cover Changes in China, 2010-2100. Adv. Meteorol. 908307. http://dx.doi.org/10.1155/2013/908307. Zaveri, R.A., Easter, R.C., Fast, J.D., Peters, L.K., 2008. Model for simulating aerosol interactions and Chemistry (MOSAIC). J. Geophys. Res. Atmos. 113, D13204. http://dx.doi.org/10.1029/2007jd008782. Zaveri, R.A., Peters, L.K., 1999. A new lumped structure photochemical mechanism for large-scale applications. J. Geophys. Res. Atmos. 104, 30387e30415. Zhang, B., Wang, Y., Hao, J., 2015. Simulating aerosol-radiation-cloud feedbacks on meteorology and air quality over eastern China under severe haze conditions in winter. Atmos. Chem. Phys. 15, 2387e2404. Zhang, Y.H., Hu, M., Zhong, L.J., Wiedensohler, A., Liu, S.C., Andreae, M.O., Wang, W., Fan, S.J., 2008. Regional integrated experiments on air quality over Pearl River Delta 2004 (PRIDE-PRD2004): overview. Atmos. Environ. 42, 6157e6173. Zheng, J.Y., Zhang, L.J., Che, W.W., Zheng, Z.Y., Yin, S.S., 2009. A highly resolved temporal and spatial air pollutant emission inventory for the Pearl River Delta region, China and its uncertainty assessment. Atmos. Environ. 43, 5112e5122.