The aerosols in this region include sulfate, dust, soot and sea salt, in a highly mixed condition, producing a complex aerosol loading in the troposphere. Some studies have shown that aerosols might be an important factor for the regional cooling in the Sichuan basin in Southern China [Luo et al., 2001; Li et al., 1995] and drought in northern China [Menon et al., 2002]. Since there are few observations available over the oceans, satellite measurements, due to their large spatial coverage (e.g, polar-orbit satellite) and high temporal resolution (e.g., geostationary satellite), provide a unique tool for quantifying aerosol properties and spatial distributions. However, to reliably retrieve aerosol properties from satellite measurements, ground and aircraft measurements are needed to constrain the satellite retrieval processes and validate the satellite results. This study demonstrates such a strategy, with emphasis on estimating the day-time diurnal change of aerosol radiative forcing, by using geostationary satellite data and other measurements during the ACE-Asia Intensive Observation Period (IOP), April 01-April 30, 2001. ACE-Asia was conducted off the coast of East China, Korea, and Japan from late March to early May 2001. A detailed description of this campaign is given by Huebert et al [2003]. In contrast with the time-invariant aerosol models used in previous satellite retrieval studies [e.g.,Wang et al., 2003; Rao et al., 1989], aerosol properties in this study are calculated as a function of space and time (called the dynamic aerosol model) by incorporating an aerosol climatology for East Asia as well as observed aerosol properties from ground and in-situ measurements during ACE-Asia. The detailed description of the retrieval method and validations of the retrieval algorithms can be found in the reference listed above. The following is the case demo of our AOT product.

2. Hypothesis and Methodology
Aerosols with diameters around 1 ~2µm are efficient in scattering the visible light. During MODIS passing time (locally, 10:30AM for TERRA and 1:30 for AQUA) in clear sky conditions, the atmospheric boundary layer is well mixed. Hence, the MODIS visible reflectance and its column aerosol optical thickness (AOT) retrievals can be used as indicators of the PM2.5 mass at the surface. In this study, we compared MODIS AOT with the ground-based PM2.5 hourly measurements.  For each comparison, MODIS AOT time is centered around the PM2.5 observation time period. The final goal of this comparison is to evaluate the quality of MODIS AOTs in the context of air quality applications before they are assimilated into the air quality models.  This is important because evaluation of data quality is a critical step in the data assimilation processes.

3. Data and Study Area

• MODIS AOT from TERRA and AQUA, 2002.
• PM2.5 measured from Tapered-Element Oscillating Microbalance (TEOM) in Alabama and Texas.
• Sunphotometer data in Stennis, MS.
• EPA PM2.5 analysis and extinction analysis from IMRPOVE measurements.

4. A Case Demo. (more cases and analysis can be found in our paper and poster)
The following figures show a heavy haze event identified by the spatial distribution of MODIS AOT. Also shown is the linearly derived Air Quality Index (AQI) and the 700mb geopotential heights. Grey regions are areas where MODIS AOT is not available due to possible sun glint or cloud contamination.

5. Conclusions

• Using one year of the MODIS AOT from the TERRA/AQUA satellites collocated with hourly particular matter mass measured at about 40 ground stations over Alabama and Texas, we show that the MODIS AOT has a good positive correlation with PM2.5 mass (linear coefficient around 0.7).
• Through statistical analysis, the MODIS AOT product can be used to discern air quality categories such as good, moderate and unhealthy to a relatively high degree of confidence. However we would like to outline several factors that could affect the relationship between PM2.5 and satellite-derived AOT. These factors include vertical distribution, optical and hydroscopic properties of aerosols.
• Aerosol extinction profile from ground based lidars or from satellite measurements such as CALIPSO are highly important for further enhancing the use of satellite data for air quality studies. This study implies that assimilation of MODIS AOT has the potential to improve the air quality forecasts.

2. Why do we concern about the dust optical properties, especially in East Asian dust source region?
The effect of aerosols on climate is one of the largest uncertainties in current global climate models [Hansen et al., 1997]. Current understanding of the radiative forcing of dust aerosols is limited, especially over dust source regions where ground observations are sparse and polar orbiting, multi-spectral satellite retrievals at visible to near-infrared wavelengths are often difficult due to the high surface albedo [Kaufman et al., 2002]. While several studies and field experiments have been conducted to study Saharan dust aerosols [Tanré et al., 2003], the widely prevalent dust events ("yellow sand") from the Taklamakan and Gobi deserts in Northwest China, have only gained attention recently [e.g. Husar et al., 2001]. In addition, previous experiments such as ACE-Asia [Hubert et al., 2003] examined aerosol properties thousands of kilometers downwind from the Taklamakan and Gobi deserts. To our knowledge, no long-term or mid-term systematic observations of dust radiative properties in the dust source region over China have been presented.

Using AOT data from the Aerosol Robotic Network (AERONET), recent studies have indicated that the diurnal variability of dust AOT is small (<5%) over various observation sites such as Cape Verde [Kaufman et al., 2000; Smirnov et al., 2002] that is about 500km downwind from the Saharan dust source region. On the other hand, several cases studies from both ground observations [e.g., Levin et al., 1980] and geostationary satellite retrievals [e.g., Wang et al., 2003] indicate that the diurnal variation of dust aerosols is relatively larger (>15%) and are important for dust radiative forcing calculations [Christopher et al., 2003]. Our current study is different from previous research since we use Sunphotometer inferred aerosol optical thickness (SP AOT) data collected near the Chinese dust source regions over a longer time period to analyze the diurnal and seasonal variations of dust AOT and Angstrom exponent.

3. What makes our observation site and this study unique?

Details can be found in our paper and recent poster:

References:
Wang, J., X. Liu, S.A. Christopher, J.S. Reid, E.A. Reid, and H. Maring, The effects of non-sphericity on geostationary satellite retrievals of dust aerosols, Geophys. Res. Lett., 30, doi:10.1029/2003GL018697, 2003. (pdf file)

Wang, J., S A. Christopher, Xiong Liu, Jeffrey S. Reid,Elizabeth Reid, Hal Maring. The effect of non-sphericity on GOES-8 dust aerosol retrievals during PRIDE, Eos Trans. AGU, 84(46), Fall Meet. Suppl., Abstract A11E-0011, 2003. (pdf file, download first and then open it)

1. Background

• Dust affects visibility, human health, and the Earth energy budget. However, modeling of dust distribution and quantification of its radiative effects are difficult, simply because ground-based measurements for dust aerosols are limited in both space and time.
• The satellite measurements have been considered as one of the best tools to characterize the high spatial-temporal variations of aerosols. However, the current dust retrievals from satellite measurements have large uncertainties, mainly because dust particles are non-spherical, and their phase functions can not be calculated/treated properly.
• It has been shown theoretically that such uncertainties can be easily larger than 2. In the context of practical applications, few quantitative evaluations of non-spherical effect on satellite retrievals have been made, either due to the lack of in situ aerosol characterization measurements, or because most satellite measurements lack the capability to monitor the same dust layer from different angles with high temporal resolutions.

2. Objectives
Will consideration of non-spherical effects improve the satellite retrievals, if all the required data to characterize aerosol optical properties are given in the same temporal-spatial domain?

3. Methods/Strategies

• Using all possible in situ data collected during the Puerto Rico Dust Experiment (PRIDE).
• Measured dust size distribution (from 3 different sizers) and aerosol light scattering/extinction coefficients (from 3 Nephelometers) are combined together to infer the aerosol effective refractive index, and constrain aerosol properties in the retrievals. Inferred refractive index for dust particles is 1.53-0.0015i, and single scattering albedo is about 0.97~0.98.
• Dust samples were collected from the aircraft measurements and then analyzed through the scanning electronic monograph (SEM). A statistical mode for dust morphologies is created based on the SEM analysis of 60, 500 particles. The model uses 6 size intervals and 15 aspect ratios (1.2~10) to describe the dust size and shape. In this study, we assume dust particles are oblate spheroid. The aerosol optical properties are then computed through the T-matrix calculations.

4. Results and Conclusion
The retrieval results using different phase functions are shown in the following figure. Here listed some conclusions.

• Using SEM data and T-matrix calculations, the computed non-spherical phase function agrees well with synthetic phase function derived from independent measurements [Liu et al., 2003].
• Applying purely non-spherical phase functions into the satellite retrieval algorithms only shows slightly improvement at certain scattering angles. However, using composite phase function by considering both spherical and non-spherical particles greatly improves the retrievals.
• Further efforts are needed to combine the use of multi-angle, multi-channel, and polarization data sets to retrieve the morphologies of particles and to apply them in satellite retrievals.