Validation of MODIS 3 km Land Aerosol Optical Depth

The study conducted by Gupta et al. (2018) focuses on the validation of the MODIS (Moderate Resolution Imaging Spectroradiometer) 3 km aerosol optical depth (AOD) product derived from NASA's Terra and Aqua satellites. This higher resolution product was introduced in the Collection 6 data release to enhance the quality of aerosol retrievals over land surfaces.

This research compares MODIS AOD data against AERONET (AErosol RObotic NETwork) measurements to evaluate the accuracy and reliability of the new 3 km product. The analysis includes 161,410 high-confidence AOD pairs from Terra (2000-2015) and 90,162 pairs from Aqua (2003-2015), revealing that 62.5% and 68.4% of the AOD values for Terra and Aqua, respectively, fall within the predefined error bounds of ±(0.05 + 0.2 × AOD), indicating a high correlation (R = 0.87) between these two datasets.

Key Findings

AOD Bias: The study identifies a mean positive bias in AOD retrievals, approximately 0.06 for Terra and 0.03 for Aqua. Notably, the bias for the 3 km product is about 0.03 larger than that of the 10 km product.
Factors Affecting Validation: The validation results showed dependencies on various factors like quality flags, region, viewing geometry, and spatial variability of aerosols, rather than a direct correlation with aerosol loading.
Time Series Analysis: Over time, the validation accuracy has shifted, particularly for Terra MODIS, likely due to sensor degradation and changes in AERONET station distribution.

The MODIS DT (Dark Target) algorithm employs spectral reflectances at specific wavelengths to retrieve aerosol parameters over land. For the 3 km product, reflectances are aggregated from a set of native pixels, with stringent filtering to obtain high-quality data. The product generation is highly selective, allowing the algorithm to discard up to 90% of pixels that do not meet retrieval criteria, thereby enhancing the accuracy of the AOD measurements.

Collocation and Methodology

To validate the AOD products, a collocated dataset (CDS) was created, matching MODIS retrievals with AERONET observations within a specified spatial and temporal framework. The collocation statistics were derived from pairs of AOD measurements to compute correlation coefficients, root mean square errors (RMSE), and mean biases, providing a comprehensive overview of the product's performance.

Comparison Across Regions

Regional statistics revealed variability in retrieval accuracy, with northern mid-latitude areas demonstrating high correlations with AERONET data. However, results were less reliable in regions with complex aerosol characteristics, such as West Africa where heavy dust and smoke intersect make retrieval models challenging.

The study concludes that the MODIS 3 km AOD product offers significant promise for various geophysical research applications, but emphasizes using the product cautiously with attention to quality assurance flags and regional characteristics to derive accurate atmospheric insights.

Recommendations

For global assessments and climate studies, it is recommended to continue using the 10 km AOD products due to their established accuracy and reliability.
The 3 km product should be preferred for local applications that require finer resolution in aerosol monitoring and analysis, taking into account the observed biases and error profiles.

The results from Gupta et al. inform researchers on the strengths and limitations of the MODIS AOD products while providing a foundation for future studies and refinements in aerosol retrieval methodologies.

Validation of MODIS 3 km Land AOD Summary

Validation of MODIS 3 km Land AOD Summary