**4. Aerosols monitored by satellites over the land**

#### **4.1. Data**

The AOT appears the highest value between 30°N and 40°N in the latitude direction, as there are main industry regions. The FMF also reaches the highest values in the same region. Both AOT and FMF get less towards the southern of the 30°N, due to lack of industry pollution. The human influence is very evident at longitude direction. China coast is around 120°E with the highest of the AOT and the FMF. Both the AOT and FMF gradually decrease away from the coast. The decrease tendency of AOT over the East China Sea has an affinity with the distance far from the continent, suggesting the AOT should be influenced by the continental sources.

The aerosol's transport depends on wind, so seasonal change of wind may lead to the spatial and the temporal distribution of aerosols over the China Sea. The mean wind field at 850 hp in 2001–2006 is shown in **Figure 10**. The patterns are similar between winter and spring, demonstrating that northwestern wind control the north of 20°N and eastern wind control the south of 20°N. The eastern wind prevails at most of region in summer and fall, while southwestern wind exists in the region between 30°N–40°N. Meanwhile, rainfall is the high in

**Figure 10.** Seasonal distribution of wind field at 850 hp over the China Sea. (a) winters; (b) spring; (c) summer; (d) fall.

**3.5. Correlation to meteorological conditions**

122 Aerosols - Science and Case Studies

summer and low in winter and spring.

Anhui province lies in the hinterland of Yangtze Delta in China between temperate zone and sub-tropical zone (**Figure 11**), one of the most important agricultural provinces covering 139,600 km2 . The agricultural activity and urbanization play important impacts on aerosols and air quality. For example, agricultural residues burning causes serious environment problem [32]. However, there is a little information related to aerosol characteristics with their influences of meteorological factors.

**Figure 11.** The geographical regions and the location of AERONET sites and meteorological stations.

One sample of SeaWiFS is selected and shown in **Figure 12**. It is a composite image of the satellite-received reflectance at the TOA on December 24, 2003. It covers a complicate structure of the ground reflectance including very turbid waters to oceanic waters, showing a mixture of reflectance from the Rayleigh, aerosols, ocean and land.

**Figure 12.** The image of the satellite-received reflectance at the TOA from SeaWiFS data on Dec. 24, 2003.

#### **4.2. Some results**

The UAC model uses the same computation method of the Rayleigh reflectance for the ocean and land, with the value is corrected by the elevation of the land surface. The aerosol reflectance obtained by the UAC model is shown in **Figure 13**, with significantly different structure from that of the ground reflectance. The image shows coherent spatial distributions extending from land into the coastal ocean, especially almost no difference in the aerosol distributions between lakes and nearby lands. The large spatial variations of the aerosol reflectance occur over the region around Shanghai and Lake Taihu. The spatial distribution of aerosols demonstrates that the UAC model has the advantage to estimate the aerosol reflectance without discontinuity between land and ocean.

#### **4.3. Validating the retrieval aerosols over the land**

There are more than 500 sites measured by AERONET around the world and four sites located in the region, which are EPA-NCU, Gosan\_SNU, Gwangju\_K-JIST and Taipei\_CWB. The AOT data are used to obtain the aerosol scattering reflectance. They are matched with cloud free SeaWiFS data under the time difference of ±2 h at the same locations. The reflectance in Band 2 of the Gosan\_SNU site is selected and shown in **Figure 14** according to the time of measurements.

of the ground reflectance including very turbid waters to oceanic waters, showing a mixture

**Figure 12.** The image of the satellite-received reflectance at the TOA from SeaWiFS data on Dec. 24, 2003.

The UAC model uses the same computation method of the Rayleigh reflectance for the ocean and land, with the value is corrected by the elevation of the land surface. The aerosol reflectance obtained by the UAC model is shown in **Figure 13**, with significantly different structure from that of the ground reflectance. The image shows coherent spatial distributions extending from land into the coastal ocean, especially almost no difference in the aerosol distributions between lakes and nearby lands. The large spatial variations of the aerosol reflectance occur over the region around Shanghai and Lake Taihu. The spatial distribution of aerosols demonstrates that the UAC model has the advantage to estimate the aerosol reflectance without discontinuity

There are more than 500 sites measured by AERONET around the world and four sites located in the region, which are EPA-NCU, Gosan\_SNU, Gwangju\_K-JIST and Taipei\_CWB. The AOT data are used to obtain the aerosol scattering reflectance. They are matched with cloud free SeaWiFS data under the time difference of ±2 h at the same locations. The reflectance in Band 2 of the Gosan\_SNU site is selected and shown in **Figure 14** according to the

**4.2. Some results**

124 Aerosols - Science and Case Studies

between land and ocean.

time of measurements.

**4.3. Validating the retrieval aerosols over the land**

of reflectance from the Rayleigh, aerosols, ocean and land.

**Figure 13.** The aerosol reflectance obtained by the UAC model from SeaWiFS data on December. 24, 2003 shown with the RGB channels from Bands 6, 7 and 2.

**Figure 14.** Comparison between the aerosol scattering reflectance from SeaWiFS data using the UAC model and that from the AERONET measurements in Band 2.

A total of 68 pairs of match-up data were established during 2006 and 2007 at the Gosan\_SNU site. The aerosol scattering reflectance was obtained from SeaWiFS data using the UAC model and from the AOT data. The two reflectances in Band 2 vary consistently with each other, with the relative error of 21.9%. The reflectance in other four bands also displays similar results with the relative errors of 22.9, 23.8, 28.3 and 33.7%, respectively.

The aerosol scattering reflectance obtained from the AOT at other AERONET sites is also used to validate the accuracy of the reflectance from SeaWiFS data using the UAC model, shown in **Table 2**. Most of the relative errors are around 35%, with the mean values of 32.5, 33.3 and 33.6% corresponding to the sites of EPA-NCU, Gwangju\_K-JIST and Taipei\_CWB, respectively. The mean error of match-up pairs at the four sites is 31.4%. The standard deviations (STD) are almost the same, with only 2.8% difference. The two mean differences are the same, at 0.00085 sr−1.


**Table 2.** The relative error of the SeaWiFS retrieval aerosol reflectance using the UAC model with respect to values calculated from the AERONET measured AOT data.

In order to take into account both spatial and temporal variations in aerosol distribution in Anhui, MODIS AOT at 10 km × 10 km resolution and AERONET AOT at 15-m intervals need to be co-located in space and time. In order to perform direct comparison and validation AOT at 550 nm, the values of AERONET AOT at 550 nm are retrieved by using Angstrom exponent calculated between 440 and 870 nm. The MODIS AOTs agree generally well with the AERO-NET AOTs. Most of dots overlay with the error range of △τ = ±0.05 ± 0.20 τ. The linear fit slopes are less than 1, correlation coefficients are larger than 0.8, and RMS are less than 0.21 at four AERONET sites, which indicate that MODIS aerosol inversions have systematic biases in Anhui. The deviation from the unity of the slope of correlation plot represents systematic biases, whereas the intercept represents the errors due to surface reflectance assumptions [33]. The correlation coefficient results are all consistent.

#### **4.4. Spatial and temporal variations of aerosols over the land**

The spatial distribution and seasonal variations in AOT in Anhui averaged over the period 2001–2009 are presented in **Figure 15**. From the feature of aerosol spatial distribution, higher values of the AOT are observed in a large plain region at the north and middle of Anhui, while smaller values can been found over the mountainous region of southern Anhui centred over Huangshan and Dabieshan Mountain. We can find that the AOT is obviously influenced by topography with lower AOT over the mountainous region. Big particle aerosols are removed by gravitational settling with dominant small particle aerosols when terrain altitudes increase. The AOT is higher over the plain region, due to complex aerosol sources from dust, industry emissions, biomass burning and traffic sources.

A total of 68 pairs of match-up data were established during 2006 and 2007 at the Gosan\_SNU site. The aerosol scattering reflectance was obtained from SeaWiFS data using the UAC model and from the AOT data. The two reflectances in Band 2 vary consistently with each other, with the relative error of 21.9%. The reflectance in other four bands also displays similar results with

The aerosol scattering reflectance obtained from the AOT at other AERONET sites is also used to validate the accuracy of the reflectance from SeaWiFS data using the UAC model, shown in **Table 2**. Most of the relative errors are around 35%, with the mean values of 32.5, 33.3 and 33.6% corresponding to the sites of EPA-NCU, Gwangju\_K-JIST and Taipei\_CWB, respectively. The mean error of match-up pairs at the four sites is 31.4%. The standard deviations (STD) are almost the same, with only 2.8% difference. The two mean differences are the same, at 0.00085

**SeaWiFS (nm) 412 443 510 670 865** AERONET (nm) 380 440 500 675 865 Gosan\_SNU (%) 22.88 21.94 23.84 28.29 33.74 EPA-NCU (%) 29.45 30.68 29.60 32.78 39.87 Gwangju\_K-JIST (%) 28.30 28.51 30.62 36.14 43.16 Taipei CWB (%) 32.31 33.25 29.70 32.64 40.05

**Table 2.** The relative error of the SeaWiFS retrieval aerosol reflectance using the UAC model with respect to values

In order to take into account both spatial and temporal variations in aerosol distribution in Anhui, MODIS AOT at 10 km × 10 km resolution and AERONET AOT at 15-m intervals need to be co-located in space and time. In order to perform direct comparison and validation AOT at 550 nm, the values of AERONET AOT at 550 nm are retrieved by using Angstrom exponent calculated between 440 and 870 nm. The MODIS AOTs agree generally well with the AERO-NET AOTs. Most of dots overlay with the error range of △τ = ±0.05 ± 0.20 τ. The linear fit slopes are less than 1, correlation coefficients are larger than 0.8, and RMS are less than 0.21 at four AERONET sites, which indicate that MODIS aerosol inversions have systematic biases in Anhui. The deviation from the unity of the slope of correlation plot represents systematic biases, whereas the intercept represents the errors due to surface reflectance assumptions [33].

The spatial distribution and seasonal variations in AOT in Anhui averaged over the period 2001–2009 are presented in **Figure 15**. From the feature of aerosol spatial distribution, higher values of the AOT are observed in a large plain region at the north and middle of Anhui, while smaller values can been found over the mountainous region of southern Anhui centred over Huangshan and Dabieshan Mountain. We can find that the AOT is obviously influenced by

the relative errors of 22.9, 23.8, 28.3 and 33.7%, respectively.

calculated from the AERONET measured AOT data.

The correlation coefficient results are all consistent.

**4.4. Spatial and temporal variations of aerosols over the land**

sr−1.

126 Aerosols - Science and Case Studies

**Figure 15.** The seasonal mean of the AOT over Anhui province in China: (a) spring, (b) summer, (c) autumn and (d) winter.

The temporal variation of aerosol in Anhui can be well seen in **Figure 16**. It can be seen that there are obvious periodic variations in AOT, and the monthly variations are in agreement with the aforesaid result of seasonal changes. We find that the lowest AOT values are observed during the autumn with the highest during spring and summer. The high values of the aerosols are affected by dust originating from the northwestern China in spring. Despite that the amount of rainfall is maximum in summer, but the month of the highest AOT is June. The decrease in the AOT during autumn and winter is related with meteorological conditions.

#### **4.5. The influence for aerosol by agricultural residues burning**

The analysis shows that the month of the highest AOT is June. June is the beginning month of rainy season, in which AOT generally decreases by wet deposition. **Figure 17** shows a large increase (21%) in the AOT value from 0.77 in May to 0.93 in June and a prominent decrease (32%) in AOT value from June (0.93) to July (0.63). The great change of AOT is likely from other sources of aerosols in June compared to in May and July. June is the time of agricultural residues burning after harvest periods in Anhui. Anhui province is a large agriculture area with about 30,000,000 tons of the grain output every year. The field burning is a general practice to clear the land for the next crop, controlling weeds and enriching soil nutrients. Thus, large amounts of aerosol particles are emitted by burning agricultural residues [34]. The fire pixels almost dominate the whole northern Anhui province and the fire pixel counts peak in June. Compared to the counts in other months, the fire pixel count is quite large in June and clearly matches with the high AOT loads during the burning month. As a result, the high AOT values observed in June in Anhui correspond to the burning agricultural.

**Figure 16.** Monthly mean AOT and Angstrom exponent over Anhui province in China: (a) AOT and (b) Angstrom exponent.

#### **4.6. Back trajectory analysis**

The 3-day back trajectory analysis on 850 hPa was calculated to examine the aerosol sources of different AOT by using the hybrid single-particle Lagrangian integrated trajectory (HYSPLIT) model. In order to obtain the influence for different aerosol particle size ranges by trajectories, two categories of back trajectories are defined based on values of the Angstrom exponent: coarse mode aerosol (CMA) with Angstrom exponent less than 1 and fine mode aerosol (FMA) with Angstrom exponent over 1, in order to obtain the influence for (**Figure 18**). A similar analysis for AOT degrees and the Angstrom exponent levels has been conducted.

(32%) in AOT value from June (0.93) to July (0.63). The great change of AOT is likely from other sources of aerosols in June compared to in May and July. June is the time of agricultural residues burning after harvest periods in Anhui. Anhui province is a large agriculture area with about 30,000,000 tons of the grain output every year. The field burning is a general practice to clear the land for the next crop, controlling weeds and enriching soil nutrients. Thus, large amounts of aerosol particles are emitted by burning agricultural residues [34]. The fire pixels almost dominate the whole northern Anhui province and the fire pixel counts peak in June. Compared to the counts in other months, the fire pixel count is quite large in June and clearly matches with the high AOT loads during the burning month. As a result, the high AOT values observed

**Figure 16.** Monthly mean AOT and Angstrom exponent over Anhui province in China: (a) AOT and (b) Angstrom ex-

The 3-day back trajectory analysis on 850 hPa was calculated to examine the aerosol sources of different AOT by using the hybrid single-particle Lagrangian integrated trajectory

in June in Anhui correspond to the burning agricultural.

128 Aerosols - Science and Case Studies

ponent.

**4.6. Back trajectory analysis**

**Figure 17.** Monthly variation of fire pixels in 2007 in Anhui province from MODIS global monthly fire location product (MCD14ML). Fire pixels are shown in black dots.

For CMA category (**Figure 18a**), 8 clusters were obtained. The northwest transport pathway (Clusters 1, 2, 5–8) contributed 70% of all the trajectories, appearing with coarse mode aerosol. Long northwest clusters stand out with abundance of dusts, and dusts are carried by the northwest trajectories from dust sources to Fuyang. Therefore, aerosol particle sizes are mainly influenced by the northwest flows.

For FMA category (**Figure 18b**), six clusters were observed. Clusters 2 (13%) and 3 (8%) represent the northwest flows, while Cluster 1 represents the flow of the southwest sector (17%). Clusters 4 (23%), 5 (19%) and 6 (20%) have westerly, easterly and north easterly origins, respectively. The most notable difference of trajectories between CMA and FMA is the distance of transport flows. The long northwest trajectories (70%) are dominant for CMA, while the short transport flows (62%) are the most for FMA. This result is attributed to different aerosol types under different air masses. Therefore, anthropogenic aerosols mainly lead to fine mode, and natural aerosols dominate the long northwest trajectories.

**Figure 18.** Mean HYSPLIT 72-h backward trajectories of 1500 m altitude at 02:30 GMT by cluster analysis during 2001– 2009 at Fuyang station. (a) CMA, Angstrom exponent <1 and (b) FMA, Angstrom exponent > 1.
