TL;DR
This paper adapts a satellite-based aerosol retrieval algorithm (PARASOL LUT) to the complex urban environment of Beijing. By employing a spectral reflectance shape invariance principle, the algorithm successfully retrieves aerosol optical depth (AOD) and fine mode volume weighting (FMVW) with good agreement to ground-based AERONET data, highlighting the importance of data quality for accurate results.
What Problem Does This Paper Address?
Accurate monitoring of aerosol properties is crucial for understanding their impact on climate and air quality, especially in densely populated megacities like Beijing, which experiences high aerosol loading. Existing satellite retrieval methods often struggle with the complex interplay of fine and coarse mode aerosols and the influence of heterogeneous urban surfaces. This paper aims to improve the retrieval of aerosol properties, specifically for fine/coarse aerosol mixtures in an urban setting, by adapting and testing an extended Look-Up Table (LUT) algorithm using PARASOL satellite data. The goal is to achieve higher spatial resolution monitoring for better air quality assessment.
Methods
The study utilizes an extended version of the Look-Up Table (LUT) aerosol retrieval algorithm originally developed for PARASOL (Polarization and Anisotropy of Reflectances for Atmospheric Science coupled with Observations from a Lidar) measurements. A key methodological innovation is the adoption of the spectral reflectance shape invariance principle. This principle assumes that the spectral shape of the surface's Hemispherical Directional Reflectance Factor (HDRF) is relatively constant across wavelengths. This assumption allows for the separation of surface and atmospheric (aerosol) contributions to the total radiance measured by the satellite.
Instead of relying on a pre-defined surface reflectance model (as used in the GRASP algorithm), this method uses the spectral invariance to constrain aerosol properties. The algorithm simulates polarized reflectances for various aerosol models and optical depths, then compares these simulations to the PARASOL measurements. The best-fitting aerosol model and optical depth are selected based on minimizing the difference between simulated and observed polarized radiances. The retrieval process involves cloud screening, preliminary selection of aerosol models and AODs, and a final decision based on minimizing derivation between simulated and measured polarized radiance.
Datasets
- PARASOL (Polarization and Anisotropy of Reflectances for Atmospheric Science coupled with Observations from a Lidar) satellite data: This dataset provides multi-spectral, multi-directional, and polarized measurements, crucial for the retrieval algorithm.
- AERONET (AErosol RObotic NETwork) ground-based measurements: AERONET data from Beijing and Xianghe sites were used for validation. Specifically, AOD (Aerosol Optical Depth) at 550 nm and Fine Mode Volume Weighting (FMVW) were used. Level 1.5 AERONET retrievals were used due to the unavailability of Level 2.0 data close to satellite overpass times.
Key Insights and Ideas
- Spectral Reflectance Shape Invariance for Urban Surfaces: The paper demonstrates that the spectral reflectance shape invariance principle, while developed for land surfaces, can be effectively applied even in a complex urban environment like Beijing. This is a significant finding as it allows for the separation of surface and aerosol signals without needing detailed prior knowledge of urban surface spectral properties.
- Fine/Coarse Mode Aerosol Retrieval: The extended LUT algorithm is capable of retrieving properties for both fine and coarse mode aerosols, which is crucial for understanding the complex aerosol mixtures found in urban areas.
- Data Quality is Paramount: The study strongly emphasizes that the accuracy of aerosol retrieval is highly dependent on the data quality index. Retrievals with a higher quality index (Qmean > 0.75) show significantly lower errors for both AOD and FMVW compared to those with lower quality indices. This highlights the need for robust quality control in satellite-based aerosol products.
- PARASOL's Potential for Urban Aerosol Monitoring: The successful application of the algorithm to Beijing demonstrates the potential of PARASOL data for detailed, fine-resolution aerosol monitoring in urban megacities, contributing to better air quality management.
Key Results
- High Correlation with AERONET: The retrieved AODs from PARASOL showed a high correlation (R = 0.913) with AERONET measurements over Beijing.
- Low RMSE: The root-mean-square-error (RMSE) for AOD retrieval was low at 0.098.
- Promising Linear Regression: The linear regression analysis yielded a slope of 1.083 and an intercept of 0.0074, suggesting the algorithm is promising for aerosol retrieval in urban areas.
- High Fraction of Good Retrievals (Gfrac): The algorithm achieved a high Gfrac of 78%, indicating that a large proportion of retrievals fell within acceptable error thresholds.
- Impact of Quality Index:
- For cases with Qmean > 0.75: Average AOD retrieval error was 7.0% with an RMSE of 0.092. Average FMVW retrieval error was 7.3%.
- For cases with Qmean < 0.75: Average AOD retrieval error was 26.2% with an RMSE of 0.109. Average FMVW retrieval error was 58.2%.
Limitations and Future Work
- Dependence on Prior Aerosol Models: The accuracy of the derived FMVW estimates showed less resemblance to AERONET values for higher quality indices, potentially due to discrepancies between the pre-assumed aerosol models used in the LUT and the actual regional aerosol models.
- Future Work: The authors suggest including more typical aerosol modes in the aerosol mode database to improve retrieval accuracy. Future work will also focus on evaluating the retrieval accuracy of other microphysical properties, such as Single Scattering Albedo (SSA) and refractive index. The paper also notes that the retrieval accuracy of microphysical properties will be evaluated in subsequent work.
Summary updated