330

ACTA METEOROLOGICA SINICA

VOL.26

Simulation and Validation of the Aerosol Optical Thickness over China in 2006 ZHANG Hua1∗ (

), ZHANG Min

2

(

Û

), CUI Zhenlei2 (

and XIN Jinyuan3 (

)

), WANG Yuesi3 (



),

1 Laboratory for Climate Studies, National Climate Center, China Meteorological Administration, Beijing 100081 2 Tianjin Municipal Meteorological Bureau, Tianjin 300074 3 State Key Laboratory of Atmospheric Boundary Layer Physics and Atmospheric Chemistry, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029 (Received February 16, 2011; in final form April 24, 2012)

ABSTRACT The Model of Atmospheric Transport and Chemistry (MATCH) developed by the US National Center for Atmospheric Research (NCAR) was used to calculate the aerosol optical thickness (AOT) over China in 2006, with emission source data of the Intercontinental Chemical Transport Experiment Phase B (INTEX-B) and NCEP/NCAR reanalysis data as inputs. The simulation results of AOT were then validated with observational data from the Moderate Resolution Imaging Spectroradiometer (MODIS), Chinese Sun Hazemeter Network (CSHNET), Aerosol Robotics Network (AERONET), and China Aerosol Remote Sensing Network (CARSNET) at more than 30 stations over China. The comparison results indicated that the high values of AOT in the areas such as the Sichuan basin and East and South China and the low values of AOT over the Tibetan Plateau and Northwest and Northeast China were reasonably simulated by the MATCH. This model tended to underestimate the AOT values in high-aerosol-loading areas but overestimate the AOT values in less polluted areas because there are still large uncertainties in the expression of emission sources, the description of the optical properties of aerosols, the treatment of cloud and precipitation, and the selection of grid resolution. The modeling results were consistent with the CSHNET, CARSNET, AERONET, and MODIS data in most parts of China, and the correlation coefficient of the monthly mean AOT between the model and the observation was 0.79 with CSHNET data at 23 stations, 0.51 with MODIS data, and 0.88 with data at 3 CARSNET stations and 2 other stations. All of them passed the significance test with α < 0.0001. The results demonstrated that the MATCH has the ability to simulate the characteristics of the AOT distribution and its seasonal variation over China. Key words: Model of Atmospheric Transport and Chemistry (MATCH), aerosol, aerosol optical thickness (AOT), Intercontinental Chemical Transport Experiment Phase B (INTEX-B), Aerosol Robotics Network (AERONET), Chinese Sun Hazemeter Network (CSHNET) Citation: Zhang Hua, Zhang Min, Cui Zhenlei, et al., 2012: Simulation and validation of the aerosol optical thickness over China in 2006. Acta Meteor. Sinica, 26(3), 330–344, doi: 10.1007/s13351-0120306-x.

1. Introduction Aerosol optical thickness (AOT), one of the most fundamental aerosol optical properties, can be used to evaluate atmospheric pollution and to study the effects of aerosols on climate and climate change. AOT data can be obtained directly by ground observation or indirectly by satellite remote sensing. The two methods

are important to obtain AOT. However, the former cannot provide high-resolution spatial and temporal datasets, and the latter cannot provide a high accuracy retrieval dataset in some regions, especially in regions with high surface albedo where satellite data are invalid (Mao et al., 2002b). Additionally, neither method can distinguish different aerosol species effectively or provide a detailed description of the AOT

Supported by the National Basic Research and Development (973) Program of China (2012CB955303 and 2011CB403405), National Science and Technology Support Program of China (2007BAC03A01), Chinese Academy of Meteorological Sciences Basic Research Project (2012Y003), and Tianjin Municipal Meteorological Bureau Research Program (201210). ∗ Corresponding author: [email protected]. ©The Chinese Meteorological Society and Springer-Verlag Berlin Heidelberg 2012

NO.3

ZHANG Hua, ZHANG Min, CUI Zhenlei, et al.

by aerosol species. Nevertheless, numerical models can be used to calculate the AOTs of all aerosol constituents and their proportions in the atmosphere; thus, they have become a complementary tool to the two aforementioned approaches. In recent years, many ground-based AOT measurements were carried out. Li and Lu (1997) and Qiu et al. (1997) calculated AOT by using solar radiation and visibility data, and analyzed their seasonal and annual variation. Luo et al. (2002) and Zhang Junhua et al. (2002) found two high AOT value areas of China located in the Sichuan basin and South Xinjiang basin after analyzing the radiation data during 1960–1994. Liu and Zhou (1999), Zhang Wenxing et al. (2002), and Duan and Mao (2007) observed and analyzed the AOT of China offshore, Beijing, and Yangtze River delta. Satellite-based datasets include solar radiation data (King et al., 1978), sun-photometer data (Xin et al., 2007), and the satellite data of Advanced Very High Resolution Radiometer (AVHRR; Higurashi et al., 2000), Total Ozone Mapping Spectrometer (TOMS; Torres et al., 2002), Microwave Integrated Retrieval System (MIRS; Kahn et al., 2005), Moderate Resolution Imaging Spectroradiometer (MODIS; Kovacs, 2006; Li et al., 2003; Mao et al., 2002a; Wang et al., 2007), and Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO; Thomason, 2007). Additionally, great progress has been made in AOT calculations using numerical models. Chin et al. (2002) used the GOCART (Goddard Chemistry Aerosol Radiation and Transport) model to calculate the global AOT distribution and compared it with the TOMS and AVHRR satellite data and the Aerosol Robotics Network (AERONET) observational data. Takemura et al. (2003) used the climate model coupled with the aerosol transport model to simulate the global monthly averaged AOT and compared their results with the AVHRR and AERONET data. Ginoux et al. (2006) also carried out a similar study using the MOZART (model for ozone and related chemical tracers) model. However, few numerical model studies on the simulation of the AOT distribution over China

331

have been conducted. The objective of this study is to obtain the AOT distribution over China and its seasonal variation using the MATCH (Model of Atmospheric Transport and Chemistry) model and to obtain the contributions from different constituents of atmospheric aerosols to the total AOT. The simulation results will be validated with ground- and space-based observations. In this study, the MATCH model (Rasch et al., 1997) is used to simulate the AOT distributions of major types of aerosols in the atmosphere, such as sulfate, black and organic carbon, sea salt, and dust. The emission source data of the Intercontinental Chemical Transport Experiment Phase B (INTEX-B; Zhang et al., 2009) are renewed in MATCH in this study. In Section 2, a brief description of the MATCH and schemes for aerosol and source emissions are given. In Section 3, the simulation results and their validation are presented. Finally, Section 4 gives summaries and conclusions of this paper. 2. Model description MATCH is an off-line model, which was developed at the National Center for Atmospheric Research (NCAR) and can be driven by meteorological fields from the NCEP/NCAR reanalysis data or output data of general circulation models (GCMs). The version used in this work, MATCH4.2-Aer1.2, was modified in February 2003. The model runs on T42 grids with a horizontal resolution of approximately 2.8◦ ×2.8◦ and 26 vertical layers. In this study, we use the NCEP/NCAR reanalysis data in 2006 to drive the MATCH model. 2.1 Aerosol scheme The AOT for five species of aerosols, including sulfate, mineral dust, sea salt (SS), black carbon (BC), and organic carbon (OC), was simulated by the MATCH. Heterogeneous aerosols were not included, and all aerosols were treated as external mixtures. The calculation method for mineral dust was based on Tegen and Fung (1994). The dust was divided into four size bins of effective radii of 0.01–1, 1–10, 10–20,

332

ACTA METEOROLOGICA SINICA

and 20–50 μm. Dust was removed from the atmosphere by wet and dry deposition. A detailed description of the sulfate, sea salt, and black and organic carbon was given by Collins et al. (2001). 2.2 Emission source The model included emission sources for sulfur and carbonaceous aerosols. The sources for DMS, SO2 , and SO4 2− were described by Barth et al. (2000) and Zhang et al. (2009). Anthropogenic sources of SO2 and SO4 2− were from the Global Emissions Inventory Activity (GEIA) emission dataset (Benkovitz et al., 1996) and the Intergovernmental Panel on Climate Change (IPCC) emission dataset (Smith et al., 2001). SO2 and SO4 2− emissions in East Asia were

VOL.26

from INTEX-B (Zhang et al., 2009). The sources for carbonaceous aerosol included contributions from biomass burning (Liousse et al., 1996), fossil fuel combustion (Penner et al., 1993), and natural organics. The emission sources of black and organic carbon in East Asia were based on INTEX-B. Table 1 shows INTEX-B emissions by regions in China for the year 2006. Emissions vary considerably from region to region, with the highest emissions mainly located in eastern and central China. Hebei, Henan, Jiangsu, Shandong, and Sichuan provinces are the five largest contributors for most species, where Shandong is the largest contributor for SO2 , NOx , non-methane volatile organic compounds (NMVOC), PM10, and PM2.5, and it is also the second largest

Table 1. Anthropogenic emissions (109 g yr−1 ) in China in 2006 from INTEX-B Region Anhui Beijing Chongqing Fujian Gansu Guangdong Guangxi Guizhou Hainan Hebei Heilongjiang Henan Hong Kong Hubei Hunan Jiangsu Jiangxi Jilin Liaoning Inner Mongolia Ningxia Qinghai Shaanxi Shandong Shanghai Shanxi Sichuan Tianjin Xinjiang Tibet Yunnan Zhejiang China

SO2 693 248 1211 460 338 1175 880 1952 76 2281 242 1591 118 2200 915 1697 533 357 1027 1171 380 18 907 3102 618 1804 2555 336 210 0 489 1434 31020

NOx 715 327 326 547 323 1493 435 485 83 1308 839 1197 148 930 563 1486 390 473 955 860 175 46 352 1759 631 934 873 365 356 5 344 1106 20830

CO 7986 2591 2928 3895 2688 8693 4258 4409 724 15505 4967 10957 127 7482 5124 11326 3963 3794 8105 5253 961 616 3528 14970 1958 5787 10945 1860 2775 94 3765 4857 166889

VOC 958 497 343 701 303 1780 640 481 117 1521 771 1289 109 875 641 1814 463 523 989 575 131 74 491 2093 594 627 1312 381 391 14 515 1233 23247

PM10 757 123 340 435 296 942 468 571 67 1371 579 1193 27 772 576 1200 586 395 710 574 134 70 474 1702 138 969 1068 161 257 9 454 806 18223

PM2.5 574 90 257 337 222 680 348 435 53 981 440 834 18 559 424 881 400 293 512 420 98 54 328 1212 91 669 845 109 194 6 343 556 13266

BC 84 19 34 44 35 55 40 90 7 137 72 133 1 73 54 87 39 45 64 71 11 8 49 132 10 139 133 15 37 1 56 36 1811

OC 173 19 75 127 55 120 94 162 18 200 144 197 1 137 105 186 76 82 111 113 18 11 81 213 8 158 318 18 55 1 97 45 3217

NO.3

333

ZHANG Hua, ZHANG Min, CUI Zhenlei, et al.

contributor for CO and OC. Emissions from western regions, e.g., Qinghai and Tibet, are much less than from eastern ones. The regional differences of emissions are mainly caused by differences of economic development, industry structure, and population. Table 2 gives monthly emissions of INTEX-B in China in 2006 by species. Strong seasonal variations are observed for CO, BC, and OC, where the residential sector contributes the largest portion of emissions. The ratios of monthly CO, BC, and OC emis-

sions between maxima and minima are 1.6, 2.1, and 2.8, respectively. In contrast, SO2 and NOx emissions have weaker seasonal variations, with ratios of 1.4 and 1.3 between maxima and minima, because they mainly come from industrial and transportation emissions that have less of a seasonal cycle. We also find that SO2 and NOx emissions in February are lower than in neighboring months because of reduced industrial activities during the Chinese Spring Festival holiday.

Table 2. Monthly anthropogenic emissions (109 g mon−1 ) of INTEX-B in China in 2006 Species SO2 NOx CO NMVOC PM10 PM2.5 BC OC

Jan 2853 1839 18051 2528 1808 1416 240 511

Feb 2416 1666 15123 2154 1516 1166 189 385

Mar 2628 1770 14677 2034 1575 1163 168 310

Apr 2368 1627 12194 1707 1356 963 120 193

May 2389 1631 12131 1702 1361 962 117 185

2.3 Aerosol optical thickness AOT in the MATCH is denoted as τm (Liou, 2004) and expressed as  ps  τm = g −1 χi(p) qi(p) dp, (1) 0

i

where i is an index for the aerosol species, χi is the optical extinction for each species (Penner et al., 1993), qi is the corresponding aerosol mixing ratio, g is gravitational acceleration, and ps is surface pressure. The dependence on lateral coordinates is omitted for simplicity. The value of τm can be represented by a sum over the layers in the vertical grid:  τm = τm(k) , (2) k

τm(k) = g −1



pk+1

pk



χi(p) qi(p) dp.

(3)

i

The effects of hygroscopic growth of aerosols are computed from the K¨ ohler curve (Kiehl et al., 2000). The optical extinctions can be specified as a product of a prescribed “dry” value and a hygroscopic growth factor that is a function of relative humidity (RH): χi = χi(RH<<1) fi(RH) ,

(4)

Jun 2430 1654 12382 1727 1401 986 118 183

Jul 2472 1667 11714 1663 1320 934 114 182

Aug 2485 1675 11911 1684 1346 951 116 183

Sep 2460 1667 12302 1720 1395 982 118 182

Oct 2503 1697 12806 1778 1445 1023 126 200

Nov 2794 1866 15041 2037 1673 1211 164 283

Dec 3220 2071 18552 2513 2026 1507 221 419

where fi is the hygroscopic growth factor (Collins et al., 2001). 3. Model simulation and validation First, we verify that MATCH is able to simulate the main characteristics of the global AOT distribution and AOT seasonal variations in comparison with MODIS data MOD08− M3 (the level 3 monthly product). Then, MATCH is used to simulate the AOT distribution and its seasonal variation over China in 2006. The simulation results will be compared with the AOT data from the MODIS, CSHNET (Chinese Sun Hazemeter Network), AERONET, and CARSNET (China Aerosol Remote Sensing Network) in 2006. Finally, the comparisons among the modeling results and the four kinds of observational data over China will be discussed in detail. 3.1 Comparison with global AOT distribution of MODIS The MODIS imaging sensor is an important instrument onboard the first satellite Terra in NASA’s (National Aeronautics and Space Administration) Earth Observing System (EOS) science plan. It scans

334

ACTA METEOROLOGICA SINICA

and covers the earth once a day to provide visible light and near infrared and infrared channels, totaling 36, for global observation with a scanning range of 2330 km (width)×10 km (length). In this study, the data used were the monthly averaged MOD08− M3 AOT data provided by NASA (wavelength λ = 550 nm) with a horizontal resolution of 1◦ ×1◦ . Li et al. (2003) and Mao et al. (2002a) validated the usability of the MODIS data over China and concluded that MODIS data were accurate to some extent; the error was relatively low (less than 20%) in South China, which is covered with dense vegetation, whereas in other areas, the error was higher. To validate the MATCH model simulation of the aerosols over China, the simulation results were compared with the global AOT distribution of MODIS data in Fig. 1. Figure 1 shows the comparative results between the MATCH and MODIS AOT data, representative of the monthly average AOT at wavelengths of 630 and 550 nm, respectively. Both the model and the satellite data show that the high-value range of AOT was mainly distributed in central and northern Africa, East Asia, Europe, the eastern part of North America, and Arabia in July. Relatively large optical thickness of sulfate aerosol formed because of greater SO2 emission due to the population density and well-developed industries in East Asia, Europe, and the eastern part of North America, whereas in North Africa and Arabia, much of the AOT was caused by dust aerosol because these regions are dry and have large desert areas. Especially in July, Arabia and the northwestern desert of India have relatively high summer temperatures, less precipitation, and dry weather, and are affected by the southwest monsoon. Thus, a lot of dust aerosols were transported to the superjacent air space and formed one of the local high-value AOT centers. There were differences in AOT values between the model results and the MODIS satellite data in some regions. For example, the simulation over East Asia was not detailed enough, which might be due to the low resolution of the emission source. East Asia is densely populated with well-developed industries; thus, the locality of aerosol emission sources was obvious, and the topography is complicated, leading to the

VOL.26

dislocation of the high-value AOT center. In October 2006, a serious forest fire occurred in Indonesia, and a great amount of carbonaceous aerosols were emitted into the atmosphere, causing the continuously high AOT there (Fig. 1h). However, the MATCH aerosol emission source did not consider this abrupt event; thus, this high-value AOT range was not reflected in the model results. Note that few data satisfy the requirement for the AOT inversion in the MODIS data, and the precision of the obtained data was not high in the areas with a relatively high surface albedo such as areas under the topographic condition of snow or desert. Wang et al. (2007) showed that the MODIS AOT was representative for agrosystem areas, whereas over the Tibetan Plateau, the northern desert areas, urbanized areas, the clean area in Northeast China, and the tropical rainforests, the MODIS products had a relatively high error and were not applicable. Moreover, the MODIS AOT data were at a wavelength of 550 nm, whereas the model result was at a wavelength of 630 nm, indicating that there were differences in the AOT obtained at different wavelengths. These factors combined to cause the differences between the model results and the satellite data. In this paper, the simulation results are qualitatively comparable with the MODIS AOT and its seasonal variation, but a quantitative comparison between them should use the ground observation data as a standard. Overall, although the MATCH results differed from the MODIS data in some areas, the characteristics of the global AOT distribution and its seasonal variation in 2006 were well simulated. 3.2 Comparison with CSHNET The CSHNET (Xin et al., 2007) was established and placed in operation in August 2004. The CSHNET includes 19 stations of the Chinese Ecosystem Research Network (CERN), 4 urban stations, and 2 calibration centers. The geographical locations of these stations are shown in Fig. 2. The new generation of the portable LED sun-photometer, i.e., Sun Hazemeter, was used in the CSHNET. The observational data used in this study were the monthly mean AOT at wavelengths of 500 and 650 nm at 23 observa-

NO.3

ZHANG Hua, ZHANG Min, CUI Zhenlei, et al.

335

Fig. 1. Global AOT distribution in January (a, e), April (b, f), July (c, g), and October (d, h) 2006 from the MATCH model (left panels) and the MODIS data (right panels). Model results are at a wavelength of 630 nm, and the MODIS data are at 550 nm.

tional stations of the CSHNET in 2006. The AOT at 630 nm was then calculated by linearly interpolating the AOT at 650 and 500 nm. We calculated the optical thickness for five species of aerosols, i.e., sulfate, dust, black and organic car-

bon, and sea salt, with the MATCH model. The total AOT is the summation of the AOT values for these five aerosols. Figure 3 compares the AOT values among the model simulation, CSHNET, and MODIS at 23 stations over China in 2006, and the results are discus-

336

ACTA METEOROLOGICA SINICA

VOL.26

Fig. 2. Geographical locations of the CSHNET stations.

sed below. Yanting station, located in the Sichuan basin, has a complicated topographical condition and a large local SO2 emission, and the diffusion of pollutants is difficult. Thus, there is a large amount of local sulfate aerosols annually, leading to a large AOT with insignificant seasonal variation. The simulation results at this station were consistent with the observations. Fengqiu station is located in the north of Henan Province, north of the Yellow River, and Jiaozhou Bay station is located in Shandong Peninsular, near the Yellow Sea. The areas surrounding both stations are densely populated with relatively well-developed industries and serious pollution. Thus, the AOT values were higher, about 0.45. The AOT values in April and May were greater due to the influence of dust in spring, whereas the AOT values due to sulfate aerosol were larger in summer because of the higher transformation rate of SO2 into sulfate. Therefore, the AOT values at the two stations were higher in spring and summer and lower in autumn and winter. The MATCH simulated the seasonal variation of AOT at the two stations well.

Shanghai and Tai Lake stations are near each other. The observational results at both stations were similar, with higher local AOT in spring and winter, and lower AOT in summer and autumn. The AOT values were affected greatly by the local sub-tropical marine climate, i.e., there is much more precipitation in summer and autumn, accompanied by frequent rainfall due to typhoons, which plays an important role in wet removal of sulfate aerosol. Therefore, much lower AOT values occurred in summer and autumn. In spring and winter, lower levels of precipitation resulted in less wet removal of sulfate aerosol; thus, the AOT values were higher. These simulation results were consistent with the CSHNET data. Beijing station is located in the downtown area of Beijing city and is a typical urban station. The area is characterized by a variety of emission sources of aerosol precursors and great emission intensity due to the large population, developed industries, and increasing number of family cars in Beijing. Thus, the AOT values at this station were higher. Comparison results indicate that MATCH simulated the monthly

NO.3

ZHANG Hua, ZHANG Min, CUI Zhenlei, et al.

337

Fig. 3. Comparison of total AOT calculated in the model (thick line) with that measured (vertical bar) and MODIS data (fine line) at 23 CSHNET stations in 2006.

338

ACTA METEOROLOGICA SINICA

variation of AOT at Beijing station well. The simulation results and CSHNET data both showed that the AOT values at Beijing station were higher in autumn and winter and lower in spring and summer. This is because North China is dry and rainless in spring, and dust storms occur readily in the desert of Mongolia. A large amount of dust can be transported eastward into Beijing, and it is most significant in April, leading to the high AOT values in spring in Beijing. In July and August, the temperature of Beijing is higher than in other months, so the efficiency of the oxidization of SO2 into sulfate increases greatly. Since the increased precipitation in summer in Beijing contributes little to the wet deposition of aerosol due to the intensive source emission of SO2 , the net production of sulfate aerosol is still high. The concentration of sulfate aerosol is higher in summer than in winter. Wang and Shi (2000) gave a detailed explanation of this phenomenon. The AOT values at Beijing Forest station were lower than those at Beijing station, but the seasonal variations of AOT at both stations were similar. The AOT values at Xianghe station were high. This is because many local industries are located there, the air pollution is very serious in Xianghe, and it is easily influenced by the pollution from Beijing. Xianghe station is located in the east of Beijing, whereas Beijing Forest station is located at Duling Mountain in the west of Beijing. These two stations are less than 100 km from Beijing station. The comparison showed that the AOT values at Beijing station were higher than those at Beijing Forest station, and most AOT values were higher than those at Xianghe station. However, the AOT values at Xianghe station in September and October were higher than those at Beijing station. MODIS data also showed the same results. This was due to the regional differences in emission sources. There were no significant differences in AOT values among simulation results at these three stations, although we used low-resolution MATCH model. The model could not simulate the high values at Xianghe station in September and October. Shenyang station is located in Liaoning Province and is another typical urban station. The local heavy industries are well developed, and the AOT values are

VOL.26

high, with an average value of 0.4 by observation. The simulation result was about 0.15 lower than the observation, which might be due to the insufficient intensity of the emission sources or the neglect of unknown types of aerosol emission sources expressed in the model. Changbai Mountain station is located inside the Changbai Mountain Natural Reserve, where the air is clean, and the AOT values are lower than those of urban stations. The simulation results were consistent with the observations for this station. Lanzhou station is located in Lanzhou City of Gansu Province in a typical basin in the northwestern plateau of China and has been seriously polluted because of the special topography and a great amount industry emission. The simulation showed that the AOT values were high in summer, which could be attributed to high temperatures that increase the transformation rate of SO2 into sulfate aerosol. Ansai station is located in the central part of the Loess Plateau. The simulation values were lower than the observed values from February to June at Ansai station. This may be due to an underestimation of the dust aerosol. Ordos station is located in the central part of Inner Mongolia near a dust source area; thus, dust aerosol makes a large contribution to the total AOT at this station. The local AOT peak in April–May in the observational data was also captured by the model. The last five stations are located in the remote areas of China and have lower AOT values than urban stations because they are less affected by anthropogenic activities. Haibei and Lhasa stations are located over the Tibetan Plateau, whereas Fukang station is located in the central part of Xinjiang region. The model showed that sulfate aerosol was a major component at these stations, followed by dust aerosol, and the total AOT peaked in June and July. The simulation results were a little higher at Haibei station than the observation, especially in June–September. This was partly attributed to an overestimation of sulfate aerosol. The scatterplot for the comparison between the modeling results and the observations at the 23 stations is shown in Fig. 4. It is seen that most of the simulation results were in the range of an uncertainty

NO.3

ZHANG Hua, ZHANG Min, CUI Zhenlei, et al.

factor of 2 (Myhre et al., 2005), compared with the observational results. The correlation coefficient of the AOT between model and observation is 0.79, at the significance level of α < 0.0001. Overall, the simulated AOT values were higher in areas like the Sichuan basin and eastern and southern China, where there are well-developed industries and dense populations. In the Tibetan Plateau, Northwest China, and Northeast China, the AOT values were lower. Additionally, the AOT values at urban stations were higher than those in the remote areas. The simulations indicated that AOT values for most stations of the CSHNET were mainly accounted for by sulfate and dust aerosol. Organic and black carbon and sea salt aerosol made relatively small contributions to the total AOT. The model tended to overestimate the AOT values in regions like Changbai Mountain and Ordos area, which are located in the remote areas of China with lower AOT values. However, in seriously polluted areas like Lanzhou and Shenyang stations, the model tended to underestimate the AOT values, which is consistent with the results from Chin et al. (2002). This might be attributable to uncertainties in surface emission sources because urban pollution is local and

339

complicated. The model results are representative of the mean values within the model grid, whereas the observational data are representative of single points at each of the stations, and the geographical locations and environmental and meteorological factors affected the observational values greatly. Therefore, the differences between the simulation and the observation might have been caused by many factors in addition to emission sources, as suggested by Chin et al. (2002). 3.3 Comparison with AERONET The AERONET program is a federation of the ground-based sun photometer measurement network (Holben et al., 1998), which started in 1993 at more than a dozen sites and has grown rapidly to over 100 sites worldwide (Holben et al., 2001). The AERONET has been providing column-integrated aerosol optical properties at 8 wavelengths from 340 to 1020 nm. Data from AERONET measurements have been used for satellite and model validation. The AERONET AOT at 630 nm used in this work was calculated by linear interpolation of the AOT at 675 and 440 nm. Figure 5 shows the comparison between the simulation results in this work and the AERONET observational data. There were no long-term and representative AOT data in 2006 for some AERONET stations over China. Therefore, we used available AERONET data at some stations to verify the simulation results in this paper. At Beijing, Xianghe, and SACOL (Semi-Arid Climate and Environment Observatory of Lanzhou University) stations, the simulation results were lower than the AERONET values, which might be due to uncertainties in emission sources and complicated aerosol types in the atmosphere. The simulated AOT values were slightly higher than the observational results in August and October at SACOL station. This was possibly caused by the lower wet removal of aerosol due to an underestimation of precipitation in the model.

Fig. 4. Comparison of AOT between the simulation and the CSHNET data in 2006. Solid line is the fitting line, and dotted lines identify the region of uncertainty with a factor of 2.

3.4 Comparison with MODIS at 29 observational stations To validate the MATCH simulation of the aerosol

340

ACTA METEOROLOGICA SINICA

VOL.26

Fig. 5. Comparison of total AOT calculated in the model (thick line) with that measured (vertical bar) and MODIS data (fine line) at 6 AERONET stations in 2006.

distribution over China, a comparison of the monthly mean AOT given by CSHNET at 23 stations and by AERONET at 6 stations with the simulation results and MODIS data was conducted. The results are presented in Figs. 4 and 5. Figures 3 and 5 show the monthly averaged AOT of MODIS from January to December at all the stations described above. It is seen that there was a large difference between the MATCH simulation and the MODIS product at the stations in the northwestern desert and the Tibetan Plateau (Shapotou, Erdos, Fukang, Lahsa, and Haibei stations) due to less precipitation, sparse vegetation, and long durations of snow cover. Note that MODIS data for these stations did not satisfy the requirement of land surface albedo in the retrieval process. Therefore, the MODIS data were not representative for these regions. Figure 6 shows the scatterplot of AOT simulation results and MODIS data. Their correlation coefficient is 0.51 at the significance level of α < 0.0001. There were almost no MODIS AOT data available in winter for Northeast China, Northwest China, and the Tibetan Plateau because of higher surface albedo. In these cases, the obtained data were limited, and MODIS data in these regions were less reliable. The vegetation was well developed in South China, Southwest China, and North China; thus, the surface albedo

was lower, and the error of the MODIS data in these areas was small. Thus, the data were highly reliable. Wang et al. (2007) also analyzed the errors of MODIS AOT data from China. 3.5 Comparison with other observations The China Aerosol Remote Sensing Network (CARSNET) is a ground-based aerosol monitoring network that uses the same CE-318 sun photometers as AERONET. CARSNET was established by the

Fig. 6. As in Fig. 4, but between the simulation and MODIS data.

NO.3

341

ZHANG Hua, ZHANG Min, CUI Zhenlei, et al.

China Meteorological Administration for the aerosol optical property study in 2002 (Che et al., 2009a). This network, including about 37 stations throughout China, is a routine operation network for the study of aerosol optical property over different areas in China. For validating the AOT simulation results of MATCH, Shangdianzi, Longfengshan, and Lin’an stations of CARSNET (Che et al., 2009b) and two other stations (Pan et al., 2010) were selected. Longfengshan, Shangdianzi, and Lin’an stations are regional background stations of China. Lin’an station is located in northwestern Zhejiang Province and the station data represent the characteristics of aerosol optical properties in the Yangtze River Delta region (Tang et al., 2007). Shangdianzi station is located in the southeast of Beijing and the station data represent the basic characteristics of aerosol optical properties in

northern China, especially the Beijing-Tianjin-Hebei Metropoltitan region (Tang et al., 2007). Longfengshan station is located in the southeastern part of Heilongjiang Province and the station data represent the basic situation of aerosol optical properties in the Northeast Plain of China. Table 3 shows the comparison of monthly mean AOT between the MATCH simulations and observations at the 5 stations. It is seen that there is little difference between the simulation and observation. At Longfengshan, observational AOT values were lower than other stations as its location is far from cities. The model simulated this phenomenon well. The highest observation value of 0.94 in July of Tai Lake station is not reliable because 7 observation days are not enough. Therefore, this value was deleted in Fig. 7, a scatterplot of simulation and observation results. All

Table 3. Monthly mean AOT of simulation and observation at some regional background stations

Shangdianzi 40.39◦ N, 117.07◦ E, 293.3 m Jan Apr Jul Oct Longfengshan 44.44◦ N, 127.36◦ E, 330.5 m Jan Apr Jul Oct Lin’an 30.18◦ N, 119.44◦ E, 138.6 m Jan Apr Jul Oct Pudong 31.22◦ N, 121.55◦ E, 4.2 m Jan Apr Jul Oct Tai Lake 31.40◦ N, 120.23◦ E, 2.1 m Jan Apr Jul Oct

Simulation

Observation

Number of observation

Observation

AOT630nm

AOT670nm

days

period

0.27 0.38 0.49 0.34

0.26 0.43 0.51 0.32

2006–2007 2006–2007 2006–2007 2006–2007

Che Che Che Che

et et et et

al. al. al. al.

(2009) (2009) (2009) (2009)

0.18 0.29 0.31 0.17

0.21 0.34 0.25 0.19

2006–2007 2006–2007 2006–2007 2006–2007

Che Che Che Che

et et et et

al. al. al. al.

(2009) (2009) (2009) (2009)

0.53 0.57 0.41 0.48

0.52±0.27 0.58±0.41 0.50±0.37 0.60±0.23

22 16 49 34

2007–2008 2007–2008 2007–2008 2007–2008

Pan Pan Pan Pan

et et et et

al. al. al. al.

(2010) (2010) (2010) (2010)

0.42 0.57 0.39 0.37

0.28±0.16 0.60±0.23 0.52±0.35 0.38±0.21

18 13 41 42

2007–2008 2007–2008 2007–2008 2007–2008

Pan Pan Pan Pan

et et et et

al. al. al. al.

(2010) (2010) (2010) (2010)

0.54 0.64 0.45 0.25

0.49±0.27 0.59±0.37 0.94±0.47 0.35±0.26

21 35 7 29

2007–2008 2007–2008 2007–2008 2007–2008

Pan Pan Pan Pan

et et et et

al. al. al. al.

(2010) (2010) (2010) (2010)

References

342

ACTA METEOROLOGICA SINICA

Fig. 7. Comparison of AOT between the simulation and observation at Shangdianzi, Longfengshan, Lin’an, Pudong, and Tai Lake stations. Solid line is the fitting line, and dotted lines identify the line 1:1 and the region of uncertainty with a factor of 2.

the simulation results are in the range of an uncertainty factor of 2, compared with the observations. The correlation coefficient of the AOT between model and observation was 0.88, at the significance level of α < 0.0001. Table 3 and Fig. 7 indicate that the MATCH model has a good ability in simulating monthly mean AOT at the regional background stations like Longfengshan, Shangdianzi, Lin’an, Pudong, and Tai Lake stations. 4. Conclusions The MATCH model was used in this paper to calculate the AOT distribution and its seasonal variation over China in 2006, and the simulation results were validated by observational data of the CSHNET, AERONET, CARSNET, and MODIS satellite data over China. The main conclusions are summarized as follows. The high values of AOT in the areas such as the Sichuan basin and East and South China, and the low values of AOT over the Tibetan Plateau and Northwest and Northeast China were well simulated by the MATCH. The modeling results were consistent with the CSHNET data and MODIS data in most parts of China, and the correlation coefficient of the monthly

VOL.26

mean AOT between the model and observational data was 0.79 with CSHNET data at 23 stations, 0.88 with data at 3 CARSNET stations and 2 other stations, and 0.51 with MODIS data. All of them passed the significance test with α < 0.0001. It was concluded that MATCH is able to simulate the AOT distribution and its seasonal variation over China. Detailed comparisons showed that the model tended to underestimate the AOT values in high-aerosol-loading areas but overestimate the AOT values in less polluted areas because there are still large uncertainties in the expression of emission sources, the description of the optical properties of aerosols, the treatment of cloud and precipitation, as well as the selection of grid resolution, in the model. Thus, improvements in numerical models are required to reduce the differences between simulation and observation. Additionally, certain assumptions were made in obtaining the satellite products; thus, the differences between simulation and observation were the result of complicated factors. Challenges remain in quantitative comparison of modeling results and satellite data, as discussed in great detail by Chin et al. (2002). Acknowledgments. The authors would like to thank Philip J. Rasch and Danielle Coleman of NCAR for providing the code of MATCH; the Chinese Ecosystem Research Network (CERN); the Institute of Atmospheric Physics, Chinese Academy of Sciences for providing CSHNET data; Dr. Che Huizheng for providing the CARSNET data and David Streets for providing the updated aerosol source emission data. The authors would also like to thank three anonymous reviewers for their constructive suggestions and comments.

REFERENCES Barth, M. C., P. J. Rasch, J. T. Kiehl, et al., 2000: Sulfur chemistry in the National Center for Atmospheric Research Community Climate Model: Description, evaluation, features and sensitivity to aqueous chemistry. J. Geophys. Res., 105, 1387–1415. Benkovitz, C. M., M. T. Scholz, J. Pacyna, et al., 1996: Global gridded inventories of anthropogenic emissions of sulfur and nitrogen. J. Geophys. Res., 101,

NO.3

ZHANG Hua, ZHANG Min, CUI Zhenlei, et al.

29239–29253. Che Huizheng, Yang Zhifeng, Zhang Xiaoye, et al., 2009a: Study on the aerosol optical properties and their relationship with aerosol chemical compositions over three regional background stations in China. Atmos. Environ., 43, 1093–1099. —–, Zhang Xiaoye, Chen Hongbin, et al., 2009b: Instrument calibration and aerosol optical depth validation of the China Aerosol Remote Sensing Network. J. Geophys. Res., 114, D03206, doi: 10.1029/2008JD011030. Chin, M., P. Ginoux, S. Kinne, et al., 2002: Tropospheric aerosol optical thickness from the GOCART model and comparisons with satellite and sun photometer measurements. J. Atmos. Sci., 59, 461–483. Collins, W. D., P. J. Rasch, B. E. Eaton, et al., 2001: Simulating aerosols using a chemical transport model with assimilation of satellite aerosol retrievals: Methodology for INDOEX. J. Geophys. Res., 106, 7313–7336. Duan Jing and Mao Jietai, 2007: Study on the distribution and variation trends of atmospheric aerosol optical depth over the Yangtze River delta. Acta Scientiae Circumstantiae, 27(4), 537–543. (in Chinese) Ginoux, P., L. W. Horowitz, V. Ramaswamy, et al., 2006: Evaluation of aerosol distribution and optical depth in the Geophysical Fluid Dynamics Laboratory coupled model CM2.1 for present climate. J. Geophys. Res., 111, D22210, doi: 10.1029/2005JD006707. Higurashi, A., T. Nakajima, and B. N. Holben, 2000: A study of global aerosol optical climatology with two-channel AVHRR remote sensing. J. Climate, 13, 2011–2027. Holben, B. N., and Coauthors, 1998: AERONET—A federated instrument network and data archive for aerosol characterization. Remote Sens. Environ., 66, 1–16. —–, D. Tanre, A. Smirnov, et al., 2001: An emerging ground-based aerosol climatology: Aerosol optical thickness from AERONET. J. Geophys. Res., 106, 12067–12098. Kahn, R. A., B. J. Gaitley, and J. V. Martonchik, 2005: Multiangle Imaging Spectroradiometer (MISR) global aerosol optical thickness validation based on 2 years of coincident Aerosol Robotic Network (AERONET) observations. J. Geophys. Res., 110, D10S04, doi: 10.1029/2004JD004706.

343

Kiehl, J. T., T. L. Schneider, P. J. Rasch, et al., 2000: Radiative forcing due to sulfate aerosols from simulations with the National Center for Atmospheric Research Community Climate Model, version 3. J. Geophys. Res., 105, 1441–1457. King, M. D., D. M. Byrne, B. M. Herman, et al., 1978: Aerosol size distributions obtained by inversion of spectral optical thickness measurements. J. Atmos. Sci., 35, 2153–2167. Kovacs, T., 2006: Comparing MODIS and AERONET aerosol optical thickness at varying separation distances to assess ground-based validation strategies for spaceborne lidar. J. Geophys. Res., 111, D24203, doi: 10.1029/2006JD007349. Li Chengcai, Mao Jietai, A. K. Lau, et al., 2003: Characteristics of distribution and seasonal variation of aerosol optical depth in eastern China with MODIS products. Chinese Sci. Bull., 48, 2094–2100. (in Chinese) Li Fang and Lu Daren, 1997: Features of aerosol optical depth with visibility grade over Beijing. Atmos. Environ., 31(20), 3413–3419. (in Chinese) Liou, K. N., 2004: The Introduction to Atmospheric Radiation (2nd Edition). Translated by Guo Caili and Zhou Shijian. China Meteorological Press, 362–367. (in Chinese) Liousse, C., J. E. Penner, C. Chuang, et al., 1996: A global three-dimensional model study of carbonaceous aerosols. J. Geophys. Res., 101, 19411–19432. Liu Yi and Zhou Mingyu, 1999: Temporal and spatial characteristics of aerosols over the East China Sea. Acta Oceanologica Sinica, 21(1), 46–52. (in Chinese) Luo Yunfeng, Lu Daren, Zhou Xiuji, et al., 2002: Analyses on the spatial distribution of aerosol optical depth over China in recent 30 years. Chinese J. Atmos. Sci., 26(6), 721–730. (in Chinese) Mao Jietai, Liu Xiaoyang, Li Chengcai, et al., 2002a: The comparison of remote sensing aerosol optical depth from MODIS data and ground sun-photometer observations. J. Appl. Meteor. Sci., 13(Sl), 127–135. (in Chinese) —–, Zhang Junhua, and Wang Meihua, 2002b: Summary comment on research of atmospheric aerosol in China. Acta Meteor. Sinica, 60(5), 625–634. (in Chinese) Myhre, G., F. Stordal, M. Johnsrud, et al., 2005: Intercomparison of satellite retrieved aerosol optical depth over ocean during the period September 1997 to December 2000. Atmos. Chem. Phys., 5, 1697– 1719.

344

ACTA METEOROLOGICA SINICA

Pan Liang, Che Huizheng, Geng Fuhai, et al., 2010: Aerosol optical properties based on ground measurements over the Yangtze Delta region. Atmos. Environ., 44, 2587–2596. Penner, J. E., H. Eddleman, and T. Novakov, 1993: The development of a global inventory for black carbon emissions. Atmos. Environ., 27, 1277–1295. Qiu Jinhuan, Pan Jidong, Yang Liquan, et al., 1997: Variation of atmospheric aerosols over 10 sites in China during 1980–1994. Chinese J. Atmos. Sci., 21(6), 725–733. (in Chinese) Rasch, P. J., N. W. Mahowald, and B. E. Eaton, 1997: Representations of transport, convection, and the hydrologic cycle in chemical transport models: Implications for the modeling of short lived and soluble species. J. Geophys. Res., 102, 28127–28138. Smith, S. J., H. Pitcher, and T. M. L. Wigley, 2001: Global and regional anthropogenic sulfur dioxide emissions. Glob. Biogeochem. Cycles, 29, 99–119. Takemura, T., T. Nakajima, A. Higurashi, et al., 2003: Aerosol distributions and radiative forcing over the Asian Pacific region simulated by Spectral Radiation-Transport Model for Aerosol Species (SPRINTARS). J. Geophys. Res., 108(D23), 8659, doi: 10.1029/2002JD003210. Tang Jie, Wang Mulin, Cheng Hongbing, et al., 2007: Variation characteristics of ambient NMHCs at Shangdianzi and Lin’an regional GAW sites. Acta Meteor. Sinica, 21(3), 334–341. Tegen, I., and I. Fung, 1994: Modeling of mineral dust in the atmosphere: Sources, transport, and optical thickness. J. Geophys. Res., 99, 22897–22914. Thomason, L. W., M. C. Pitts, and D. M. Winker, 2007: CALIPSO observations of stratospheric aerosols: A

VOL.26

preliminary assessment. Atmos. Chem. Phys., 7, 5283–5290. Torres, O., P. K. Bhartia, J. R. Herman, et al., 2002: A long-term record of aerosol optical thickness from TOMS observations and comparison to aeronet measurements. J. Atmos. Sci., 59(01), 398–413. Wang Lili, Xin Jinyuan, Wang Yuesi, et al., 2007: Validation of MODIS aerosol products by CSHNET over China. Chinese Sci. Bull., 52(4), 477–486. (in Chinese) Wang Xihong and Shi Guangyu, 2000: A numerical study on temporal and spatial variations of anthropogenic sulfate column burden over eastern Asia. Chinese J. Atmos. Sci., 5(1), 58–66. (in Chinese) Xin Jinyuan, Wang Yuesi, Li Zhanqing, et al., 2007: Aerosol optical depth (AOD) and Angstrom exponent of aerosols observed by the Chinese Sun Hazemeter Network from August 2004 to September 2005. J. Geophys. Res., 112, D05203, doi: 10.1029/2006JD007075. Zhang Junhua, Mao Jietai, and Wang Meihua, 2002: Analysis of the aerosol extinction characteristics in different area in China. Adv. Atmos. Sci., 19(1), 136–150. (in Chinese) Zhang, Q., D. G. Streets, G. R. Carmichael, et al., 2009: Asian emissions in 2006 for the NASA INTEX-B mission. Atmos. Chem. Phys. Discussion, 9, 4081– 4139. Zhang Wenxing, Lu Daren, Wang Pucai, et al., 2002: The observation and analysis of atmospheric aerosol optical thickness over Beijing area. China Environmental Science, 22(6), 495–500. (in Chinese)