Aerosol layer height (ALH) retrievals from oxygen absorption bands: intercomparison and validation among different satellite platforms, GEMS, EPIC, and TROPOMI

Abstract

The vertical distribution of aerosols is crucial for assessing surface air quality and its impact on the climate. Although aerosol vertical structures can be complex, assuming a certain shape for the aerosol vertical profile allows for the retrieval of a single parameter - aerosol layer height (ALH) - from passive remote sensing measurements. In this study, we evaluate ALH products retrieved using oxygen absorption measurements from multiple satellite platforms: the Geostationary Environment Monitoring Spectrometer (GEMS) focusing on Asia, the Earth Polychromatic Imaging Camera (EPIC) in deep space, and the polar-orbiting TROPOspheric Monitoring Instrument (TROPOMI). We use the extinction-weighted aerosol optical centroid height (AOCH) derived from aerosol extinction profiles of Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) as the ground truth. The differences due to the inconsistent definitions of ALH in various retrieval algorithms are investigated and eliminated before comparison. We select multiple dust and smoke cases under ideal observational conditions, referred to as "golden days", for the evaluation. Given the significant role of aerosol optical depth (AOD) in ALH retrieval, we first evaluate the AOD from these retrievals against the ground-based AErosol RObotic NETwork (AERONET). Results show that the GEMS AOD at 440 nm has better agreement with the AERONET AOD of the similar to 0.9 correlation coefficient (R) than that at 680 nm, both of which underestimate with a negative bias. In contrast, EPIC and TROPOMI tend to overestimate AOD by 0.33 and 0.23 for dust cases, while the bias for smoke plumes is small. Evaluation of ALH against CALIOP demonstrates that the EPIC/TROPOMI ALH has good consistency (R > 0.7) with CALIOP but is overestimated by approximately 0.8 km. The GEMS ALH displays minimal bias (0.1 km) but a slightly lower correlation (R = 0.64). Intercomparisons between three passive retrievals indicate that GEMS retrievals have a limited consistency with EPIC and TROPOMI of 0.3-0.4 R, while GEMS underestimates with ALHs of similar to 0.3 and similar to 0.6 km compared with TROPOMI and EPIC, respectively. The correlations improve under conditions of higher absorbing aerosols (UVAI >= 3), as the signal in the oxygen absorption band (O-2-O-2 used by GEMS) is enhanced. Although the ALH diurnal cycle from EPIC and GEMS shows some differences, they both demonstrate ALH descent in the afternoon, which might be related to the boundary layer process. Case studies show that the EPIC ALH indicates a morning ascent to around 4.5 km, while the GEMS ALH remains stable before descending to below 3 km in the afternoon.