Atmospheric deposition of nitrogen to the eastern China seas has been simulated using the MM5/CMAQ model with the 2004 national emission inventory of China. Dry and wet fluxes are 0.05–0.5 and 0.2–0.6 g m−2 yr−1, respectively, with the wet deposition accounting for 79% of the total. The total deposition of inorganic nitrogen species, including NO3−, NH4+, HNO3, NOx, and NH3, to the eastern China seas is 498 GgN yr−1 and accounts for 3.4% of the nitrogen emission by China. Deposition of NO3− and NH4+ dominates. The model results agree well with available in situ measurements. The deposition of NH4+ and NH3 to the East China Sea is up to 166 GgN yr−1, which nearly equals the total input of 184 GgN yr−1 from the mainland, including riverine discharge, industrial wastewater, and domestic wastewater. Deposition of atmospheric ammonium can account for 56% of the external total input, which is 1.1–1.5 times the input from the major rivers to all the eastern China seas. Ammonium deposition to the Yellow Sea accounts for as much as 87% of the total input. The annual total nitrogen deposition can be converted to new primary biological productivity of 100–200 mmol C m−2 yr−1, or 1.1–3.9% of the new productivity in the East China Sea. Our results suggest that atmospheric deposition has important impact on biological productivity in all the eastern China seas.
 Nitrogen compounds are key nutrients for the marine ecosystem. Atmospheric deposition plays an important role in supplying external nitrogen to the seas. For the Bight Sea in Germany, 30% of the total nitrogen came from the atmosphere [Beddig et al., 1997]. De Leeuw et al.  reported that the average atmospheric deposition to the southern part of the North Sea in August 1999 could potentially promote primary production of 2.0 mmol C m−2 day−1, that is, approximately 5.5% of the total production at that time of the year. In the East China Sea, the annual deposition of particulate NO3− and NH4+ was similar to the outflow from the Yangtze River [Nakamura et al., 2005; Uno et al., 2007; Uematsu et al., 2010]. Atmospheric deposition of nitrogen can enhance marine productivity. Duce et al.  pointed out that the increased quantities of atmospheric nitrogen entering the open ocean, which could account for up to a third of the ocean's external supply, would enhance the annual new marine biological production by up to approximately 3%.
 The eastern China seas, including the Bohai Sea, Yellow Sea and East China Sea, are located near areas of rapid population growth and economic development in China. Because the eastern China seas are usually downwind from mainland China, they are affected by the westerly jet stream at 5000 m elevation. In the early 1980s, Duce et al.  reported that desert dust aerosol from central Asia was often transported to the tropical North Pacific. Since then, many cruises and model simulations have shown that the dust storm aerosols from northwest China and pollutants from Asia during nondust periods can be transported over the coastal regions to the open oceans [Kotamathi and Carmichael, 1993; Zhang et al., 1993; Wang et al., 2002; Satake et al., 2004; Nakamura et al., 2005; Uematsu et al., 2010].
An et al.  simulated the nitrate concentration in Asian precipitation and found the highest concentrations in industrialized regions, that is, the coastal area of the mainland of China, the Bay of the Yellow Sea and the Bo Sea, Korea, and southern Japan. Nakamura et al.  reported that the depositional flux of particulate NO3− and NH4+ was 160 GgN yr−1 and 270 GgN yr−1 in the East China Sea on the basis of cruise data from autumn of 2002. Uno et al.  reported that the annual deposition of nitrate to the East China Sea was 140 GgN yr−1 on the basis of a simulation that used Modles3. Zhang et al.  estimated depositional fluxes of 34.4 and 68.4 mmolN m−2 yr−1 to the Yellow Sea and the East China Sea from observations of NO3−, NO2−, and NH4+ at Qianliyan Island and Shengsi Island. However, the reported fluxes to the eastern China Seas were mostly based on limited observations from cruises, from few sampling sites, or from model simulations with few nitrogen species. The comparison of the atmospheric deposition in the whole eastern China seas with other inputs, including the comprehensive riverine, direct industrial and domestic discharges, as well as the further impacts of atmospheric deposition on the biological productivity in the eastern China seas, were seldom studied. Here we have estimated the atmospheric deposition of inorganic nitrogen, including NO3−, NH4+, HNO3, NOx, and NH3, to the eastern China marginal seas by a regional model and the emission inventory in 2004 for all of China. The seasonal variation and the proportions of each nitrogen species to the total deposition were also determined. In addition, the atmospheric deposition was compared with other external inputs of nitrogen, and the effect of nitrogen deposition on new biological productivity in the eastern China seas was investigated.
2. Numerical Modeling
2.1. Regional Model System
 The mesoscale meteorological model MM5, version 3.7 [Grell et al., 1994], and the chemical transport model CMAQ, version 4.5.1 [Byun and Ching, 1999], were used for estimating deposition. MM5 used two one-way nested domains with grid resolutions of 81 and 27 km, covering and aligned with the CMAQ domain (Figure 1). Inner and outer grids consisted of 24 vertical layers, 14 of them in the PBL. The modeling period covered January, April, July, and October 2004, to represent typical winter, spring, summer, and autumn conditions.
 The cumulus parameterization of Grell et al. , the planetary boundary layer model of Hong and Pan  and the moisture schemes and surface energy flux of Dudhia  were used in the meteorological model. Observations and simulations of meteorological fields was compared by using meteorological data from Lvsi station (121°36′E, 32°04′N) and Shengsi station (122°37′E, 30°44′N), including wind speed, direction, pressure, and temperature. Simulated and observed meteorological parameters were correlated to better than 0.95. The gas chemistry of the carbon bond IV mechanism and the liquid chemistry of AQ were chosen for the CMAQ simulations. The fourth-generation aerosol module was used, in which three aerosol modes were used: Aitken (0.01 μm < D < 1 μm), accumulation (1 μm < D < 2.5 μm), and coarse (2.5 μm < D < 10 μm). The initialization of concentration was from the reference. Three days before each simulation period, an initialization run was started with horizontally uniform and vertically varying clean conditions in the first domain.
2.2. Modification to the Dry Deposition Simulation
 The dry deposition velocity (Vd) was estimated according to the resistance analogy method of Wesely [1985, 1989]. The default dry deposition module in CMAQ fit the natural water surface well. The impact of the sea-surface height on the dry deposition velocities has not been considered in view of atmospheric deposition to the sea surface. So, in this study, the dry deposition scheme to the sea surface was modified by adjusting the sea-surface roughness length Z0s. In general, Z0s suffered the integrated impacts of both the coarse and smooth currents. The sea surface coarse current scheme from Taylor and Yelland  was used during the parameterization of Z0s:
Here, hs is the effective wave height, and lp is the wavelength. Details of the parameterization of hs and lp can be seen in the work of Taylor and Yelland .
 The dry deposition flux Fd(Z) was estimated by the following formula:
Here, Z is the reference height; Vd(g) and Vd(a) are the dry deposition velocities of the gaseous and aerosol nitrogen species; and C(g)(Z) and C(a)(Z) are the concentrations of the gaseous and particulate nitrogen species, respectively.
2.3. Emission Inventory
 The emission inventory of nitrogen species, including NOx, NH3, was derived from the Regional Emission Inventory in Asia (REAS) (http://www.jamstec.go.jp), which was further modified according to the emission of each province reported in the yearly report of Chinese environmental quality for 2004. The sea-salt emission was treated in the inner module of CMAQ. The distribution of nitrogen emission in 81 km grid resolution in China is shown in Figure 1. The N emission intensity was approximately 14.5 Tg yr−1 for the whole China and 5.04 Tg yr−1 for East China, as shown in Figure 1.
3. Simulations of Atmospheric Deposition
3.1. Regional Dry and Wet Deposition Fluxes to the Eastern China Seas
3.1.1. Dry Deposition Velocity
 The dry deposition velocities of each nitrogen species are estimated with the modified simulation previously described, and are shown in Table 1. The mean dry deposition velocity (Vd) of HNO3 to the sea surface is 0.51 cm s−1 in January, 0.58 cm s−1 in April, 0.65 cm s−1 in July, and 0.79 cm s−1 in October. The results for NH3 are very close to those for HNO3. The dry deposition velocity of NOx is much smaller, however. According to the simulation, the effective surf height over the China seas is estimated to be 0.3–2.8 m, and the sea-surface roughness length is 0.01–0.06 cm. Thus it appears that the dry deposition velocities of NH3 and HNO3 are influenced by the sea-surface roughness length. Generally, the average contribution of the sea-surface coarse current to Vd could be up to10% of that from the smooth current.
Table 1. Dry Deposition Velocities of Gaseous Nitrogen Species Over the China Seas
NO2 × 10 (cm s−1)
NO × 100 (cm s−1)
NH3 (cm s−1)
HNO3 (cm s−1)
0.0529 ± 4.2E-4
0.0034 ± 3.5E-4
0.501 ± 0.0621
0.505 ± 0.0579
0.0535 ± 3.4E-4
0.0042 ± 1.2E-4
0.581 ± 0.0799
0.584 ± 0.0806
0.0536 ± 2.3E-4
0.0042 ± 2.2E-4
0.644 ± 0.0664
0.648 ± 0.0673
0.0535 ± 2.2E-4
0.0037 ± 2.5E-4
0.779 ± 0.1820
0.785 ± 0.1842
0.0534 ± 3.0E-4
0.0039 ± 2.4E-4
0.626 ± 0.0976
0.630 ± 0.0975
3.1.2. Dry Deposition Flux
 The total annual dry deposition flux of nitrogen to the inland and the eastern China seas has been simulated, as shown in Figure 2. Species included NOx, HNO3, NH3, NO3− and NH4+. By comparing to the emissions in different areas, it can be seen that the distribution of dry deposition depends on the locations of the emission. For instance, in Shandong Province the flux is as high as 2–3 gN m−2 yr−1 because of its high emissions there. The flux to eastern China is 0.5–3.0 gN m−2 yr−1. Wang et al.  reported that the nitrogen deposition (excluding ammonia) to eastern China of 0.4–1.6 gN m−2 yr−1 by simulating with RegADMS. If our flux of NH4+ is added to Wang's simulation, both results are comparable. At the juncture between land and sea, the dry deposition flux of nitrogen is ∼0.5 gN m−2 yr−1. It is evident that there is a gradient from inland to sea. As the distance from shore increased, the deposition decreases gradually, and reaches 20% of the inland deposition at ∼150 km offshore, and 10% at ∼300 km. The dry deposition flux in the eastern China seas is 0.05–0.5 gN m−2 yr−1, which is higher than values reported for coastal New England (0.08), the North Sea (0.16), and the Chesapeake Bay (0.30), but close to those in the Delaware Bay (0.3–0.48) and the east coast of Spain (0.39–0.98) [Smith et al., 2007; De Leeuw et al., 2003; Sheeder et al., 2002; Scudlark and Church, 1993; Sanz et al., 2002].
3.1.3. Wet Deposition Flux
 The simulated wet deposition flux is shown in Figure 3. It depends on the amount of precipitation and the concentration of the species in it. The wet deposition flux to the eastern China seas is 0.2–0.6 gN m−2 yr−1, which declines to 0.2–0.4 gN m−2 yr−1 at 300 km offshore. Unlike dry deposition, there is no big gradient from land to sea. The wet deposition flux over the eastern China seas is 0.4–0.6 gN m−2 yr−1, which is similar to that to the North Sea (0.75 gN m−2 yr−1) and the Arcachon Bay in France (0.52 gN m−2 yr−1) [De Leeuw et al., 2003; Rimmelin et al., 1999]. This indicates that atmospheric deposition to the China seas is as important as to the European seas, where input to coastal zones has been of great concern in recent years (the ANICE project, etc.).
3.2. Annual Mean Nitrogen Deposition Over the Eastern China Seas
 Nitrogen deposition to the eastern China seas (model domain 2, including the Yellow Sea and the East China Sea, but excluding the Bo Sea) from this model simulation is listed in Table 2. The area this domain is 1,272,105 km2. The results show that total nitrogen deposition for the region is 166 GgN for the four typical months (January, April, spring and autumn) and ∼498 GgN for the whole year, which accounts for 18% of the total deposition (2.7 TgN) in the full model domain, including the land of eastern China and part of the eastern China seas. Dry and wet depositions account for 21% and 79% of the total deposition, respectively. Compared with the nitrogen emission from the land of East China (5.04 TgN yr−1) in domain 2 (Figure 1) and from the whole China region (14.52 TgN yr−1), the total deposition over the eastern China seas accounts for ∼10% and ∼3.4% of the anthropogenic nitrogen emission from East China and from the whole China region, respectively.
Table 2. Dry and Wet Deposition of Each Nitrogen Species in Four Typical Months
January (106 gN)
April (106 gN)
July (106 gN)
October (106 gN)
Deposition for Spec. (106 gN)
Total (106 gN)
dry + wet
 In addition, the atmospheric depositions show obvious seasonal variations, as shown in Table 2. The greatest deposition of the four typical months, ∼74 GgN mon−1, was for January, and accounts for 45% of the sum of the four months. During winter, all of China is dominated by the westerly or northwesterly winds, mainly caused by the Siberian high-pressure area and the East Asian trough. The atmospheric circulation of January favored the transport of atmospheric nitrogen from the terrestrial sources to the eastern China seas. This increased the airborne nitrogen content in the marine PBL relative to the other seasons. Because of the favorable amounts of precipitation, the wet deposition flux is greatest in winter.
Table 2 and Figure 4 also show that NO3− and NH4+ has the highest deposition of the nitrogen species. Ammonium is the greatest contributor to the eastern China seas, and reached 300 GgN yr−1, 96% of which is wet deposition by precipitation. Ammonium deposition accounts for more than 50% of the total nitrogen deposition in each of the seasons, with a maximum near 70% in July. Nitrate is the next greatest contributor, and accounts for ∼20% in each seasons. Total nitrate deposition in the eastern China seas is 106 GgN yr−1, with the same order of magnitude as the 140 GgN yr−1 simulated by Uno et al.  for the western North Pacific in 2002. Deposition of gaseous nitrogen, including NOx, NH3, and HNO3, accounts for 10–30% of the total nitrogen deposition. The deposition of gaseous HNO3 in autumn is obviously high, and similar to nitrate. This effect is probably caused by higher HNO3 from westerly transport and favorable temperatures in autumn, which allows NH4NO3 transported to the marine atmosphere to regenerated gaseous HNO3.
3.3. Comparison of Simulated Deposition Flux With in Situ Observation
 To get perspective on the simulations, some observations of dry and wet deposition in the model domain are compared with the simulations. As shown in Table 3, the dry nitrogen flux in Fujian Province is simulated to be 0.50 gN m−2 yr−1, versus 0.18–0.48 gN m−2 yr−1 from observations in the drainage area of the Jiulongjiang River in Fujian Province in 2004 by Chen . The simulated dry deposition flux of 1.5–2.0 gN m−2 yr−1 in Sichuan Province is very near to the observed of 1.8–2.0 gN m−2 yr−1 in Chongqing by Jin et al. . In both terrestrial cases, observation and simulation agree well. For the Yellow Sea and East China Sea, the dry nitrogen deposition observed at Qianliyan Island in the Yellow Sea and Shengsi Island in the East China Sea during 1999–2003 was 0.14 and 0.20 gN m−2 yr−1, respectively [Zhang et al., 2007], and the dry deposition simulated here is 0.15–0.30 gN m−2 yr−1. The agreement of simulation and measurement indicates, as over the land, that the dry deposition flux in eastern China seas from this study is reasonable.
Table 3. Comparison of the Simulated Dry Deposition With the Observation From Other Works
Table 4 lists the wet nitrogen deposition flux simulated here and from measurements reported in the literature. The wet deposition in terrestrial eastern China is 0.4–2.0 gN m−2 yr−1. The measured wet deposition of nitrogen in Changshu, Jiangsu Province, and the Jiulongjiang River region, Fujian Province, was 0.75 and 0.85–1.12 gN m−2 yr−1 [Su et al., 2003; Chen, 2006], respectively. Both are in the range of 0.4–2.0 gN m−2 yr−1 from the simulation. The simulated nitrogen wet deposition in Fujian Province, 0.6 gN m−2 yr−1, agrees especially well with the measured 0.85 gN m−2 yr−1 in the Jiulongjiang River region of Fujian Province [Chen, 2006]. The simulated deposition for Shandong Province, 1.5–2.0 gN m−2 yr−1, also matches the measurements there, 1.8 gN m−2 yr−1 [Zhang et al., 2006]. The measured 2.7 gN m−2 yr−1 of wet deposition of total nitrogen [Wang et al., 2004] exceeds the simulated 2.2 gN m−2 yr−1, probably because the flux observed in the Taihu Lake region included both inorganic and organic nitrogen, whereas the simulations only cover inorganic species. Regarding the nitrogen wet deposition into the seas, the measurements by Zhang et al.  for the Yellow Sea and East China Sea were 0.34 and 0.76 gN m−2 yr−1, respectively, and the simulations in the eastern China seas are 0.2–1.0 gN m−2 yr−1. Again, measurement and the simulation were similar.
Table 4. Comparison of the Simulated Wet Deposition With the Observation From Other Works
 The ratio of NH4+ to NO3− (NH4+/NO3−) simulated for wet deposition was compared with the measurements from the EANET project (http://www.eanet.cc), as shown in Figure 5. In general, the flux for NH4+ is greater than for NO3−, and the ratio NH4+/NO3− reaches 3.0–5.0 at a certain site. The values for the NH4+/NO3− ratio at the 9 EANET sampling sites during January, April, July, and October agree well with the simulated values. In addition, the value of 1.3–3.4 observed in rainwater over the Yellow Sea and the East China Sea by Zhang et al.  matched very well with the simulated value of 2.8 for the eastern China seas.
 The simulated annual nitrogen deposition to the eastern China seas was also compared with measurements. From cruise sampling in the eastern China seas from September to October in 2002 [Nakamura et al., 2005], the annual deposition of particulate NH4+ and NO3− was at 270 and 160 GgN, as opposed to the 208 and 130 gN m−2 simulated here for October. Again, simulation and measurement agreed well.
 We note that uncertainties remain for the simulated deposition owing to biases in nitrogen emission, dry deposition velocities, precipitation, etc. For instance, the simulated amount of precipitation and the concentration of atmospheric nitrogen in some grids are smaller than the measurements from during EANET, which probably reduce the simulated wet deposition. The limited number of stations in the present observational net makes it difficult to verify the simulations over the eastern China Sea. From the available observational data, we conclude that the simulations reasonably represent to evaluate the depositional fluxes to the eastern China seas.
4. Comparison of Deposited Ammonium With Other Inputs to the China Seas
 The simulated nitrogen deposition over the eastern China seas is compared with other inputs of nitrogen. Figure 6 shows the ocean regions in the model domain1 for this study. Regions A, B, and C roughly represent the Bo Sea (BS), the Yellow Sea (YS), and the East China Sea (ECS), according to the definition of marginal seas by Chen . These regions are not exactly the same as the conventional geographical regions for these seas. Only the grids with more than 50% sea are considered as sea area here. The atmospheric deposition of ammonia to land and the eastern China seas are listed in Table 5. Domestic and industrial wastewater shown there is direct influx to the seas, independent of the riverine input. The total terrestrial inputs include riverine input, domestic wastewater, and industrial wastewater. The inputs from major rivers are also listed in Table 6.
Table 5. Ammonia Nitrogen Deposition From Atmospheric and Terrestrial Inputsa
Values in parentheses denote estimations based on the riverine input from Yangtze River, which was cited from Zhang , while values not in parentheses are based on data from the Bulletin of Marine Environmental Quality of China.
Sum of values for the Bo, Yellow, and East China seas could be used to represent the eastern China Seas.
Major riverine input
Sea area (103 km2)
Annual dry deposition
Annual wet deposition
Atmospheric (terrestrial + atmospheric) (%)
Table 6. Ammonia Nitrogen of Major Riverine Inputs to the Eastern China Seasa
 According to the China Environment Yearbook and Bulletin of Marine Environmental Quality of China (http://www.coi.gov.cn), the total terrestrial ammonium inputs to the BS, YS, and ECS are 122, 23, and 140 GgN yr−1, respectively, among which the riverine input is dominant except for YS. In BS, the influx of ammonium from the Yellow River is 102 GgN yr−1, and accounted for ∼84% of the terrestrial input. In ECS, the riverine input, including the Yangtze River, Min River, etc. reaches 151 GgN yr−1, and accounts for 82% of the terrestrial input. Because no large rivers feed YS, its riverine input is only 11 GgN yr−1, and domestic and industrial wastewater contributes 60% of the terrestrial input.
 The annual riverine inputs of ammonia nitrogen to BS, YS, and ECS (mostly from the Yangtze River) are 108, 106, and 151 GgN, respectively. However, the annual deposition of atmospheric ammonium to the BS, YS, and ECS is 79, 171, and 166 GgN, respectively. Thus, the input is 0.73, 16, and 1.1 times of the riverine input into BS, YS, and ECS, respectively. The deposition of ammonium could account for 39%, 87%, and 47% of the input to BS, YS, and ECS, respectively. For the input to the Yellow Sea, where there was little river influx, the atmospheric deposition surpasses the terrestrial input. These results indicate that atmospheric deposition could play a major role in the total, which is verified by the results of Zhang . For the total eastern China seas, the annual atmospheric nitrogen input is estimated to be ∼416 GgN yr−1, or 1.3 times of the total terrestrial input.
 The input from the Yangtze River (89,582 tons) came from the Bulletin of Marine Environmental Quality of China. This figure is much lower than that reported in the literature (190,000 tons), which was calculated from the ammonium concentration of 14.6 μM [Zhang, 1996] and the flow rate of 979 km3 yr−1 [Milliman and Jin, 1985]. If this value was used, the riverine input and the total terrestrial input to the East China Sea would become 252 and 284 GgN, respectively, and the atmospheric ammonium deposition would become 66% of the total riverine input. For the total eastern China seas, the annual atmospheric nitrogen input is also comparable to the total terrestrial input. Therefore, deposition of nitrogen to the eastern China seas would play an indispensable role in the marine ecosystem, since atmospheric transport can spread the nutrients and compounds harmful to marine organisms over large ocean areas.
5. Impact of Deposited Nitrogen Deposition on New Primary Productivity
 Assuming that all the nitrogen transported and deposited into the seas are bioavailable to phytoplankton, the input can be converted to carbon uptake by using a Redfield C:N ratio of 6.625. The deposition of nitrogen shown in Figures 7a and 7b would then create a primary productivity of 50–200 mmol C m−2 yr−1 in the China seas (including BS, YS, ECS, and the northern South China Sea) and 100–200 mmol C m−2 yr−1 in the eastern China seas. The carbon fixation in the marine biological productivity would then be estimated at 1599 Gg yr−1 in the East China Sea (assuming its area to be 630,000 km2). Nitrogen deposited to the eastern China seas can support 1.1–3.9% of the new production (180–626 mgC m−2 d−1, or 41–144 Tg yr−1) based on the result of Chen and Chen . Marine primary production depends on many complicated factors composed of the external factors including atmospheric, riverine, and industrial influx and the internal factors such as the upwelling nutrient elements. The internal upwelling such as the year-round Kuroshio constituted an important fraction of the new production in East China Sea [Chen and Chen, 2003]. However, the external nutrient sources have accumulative effect on the nutrients into marine ecosystem, which cannot be replaced by the internal sources. The estimated contribution to the new production from atmospheric deposited nitrogen was indeed nonnegligible because atmospheric input was an important external source of nutrients into the eastern China seas. This result is a further indication that human activities can greatly influence the marine ecosystem through the atmospheric input, since nearly all the nitrogen species transported from the continent were anthropogenic.
 Atmospheric deposition of nitrogen to the eastern China seas (mainly the Yellow Sea and the East China Sea) is 498 GgN yr−1, of which NH4+ and NO3− together account for ∼80%. Deposition over the eastern China seas accounts for ∼10% and ∼3.4% of the anthropogenic nitrogen emission from East China and from China nationwide, respectively. Deposition of atmospheric ammonium can account for 39%, 87%, and 47% of the external input to the Bo Sea, the Yellow Sea, and the East China Sea, respectively. Generally, for the whole eastern China seas, the annual input of atmospheric ammonium nitrogen is 1.3 times of the terrestrial input. Our result suggest that atmospheric deposition plays a much greater role than previously thought in the nutrient input compared with the total input. Atmospheric deposition can strongly enhance the new productivity of marine system, by 100–200 mmolC m−2 yr−1, or approximately 1.1–3.9% of the new primary productivity in the East China Sea. With the increase of nitrogen emission and deposition, human activity would increase its influence on the ecosystem of the eastern China Seas.
 Helpful comments were provided by Guoshun Zhuang, Zhigang Guo, and Ying Chen of the Center for Atmospheric Chemistry Study at Fudan University. Great thanks are given to Kenneth A. Rahn at URI for improvements in the English language and to NHPCC at Fudan University for computational resources. This work was supported by the National Programs for High Technology Research and Development of China (863 program, grants 2008AA06Z401 and 2007AA06A401), the National Key Project of Basic Research of China (973 program, grant 2006CB403704), and the Foundation for New Teachers by the Ministry of Education (grant 2008024610).