Quantitative Assessment of the Uncertainty in Aerosol Optical Property Inversion due to Different Surface Reflection Models
Inaccurate estimation of surface reflectance characteristics is one of the substantial sources of uncertainty in satellite aerosol optical property retrieval algorithms. Bidirectional reflectance distribution function (BRDF) models are frequently employed by retrieval algorithms to obtain the reflec...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
American Association for the Advancement of Science (AAAS)
2025-01-01
|
| Series: | Journal of Remote Sensing |
| Online Access: | https://spj.science.org/doi/10.34133/remotesensing.0637 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Inaccurate estimation of surface reflectance characteristics is one of the substantial sources of uncertainty in satellite aerosol optical property retrieval algorithms. Bidirectional reflectance distribution function (BRDF) models are frequently employed by retrieval algorithms to obtain the reflective properties of diverse surface types. This study, aerosol optical depth (AOD), fine mode AOD (AODF), coarse mode AOD (AODC), absorbing AOD (AAOD), single scattering albedo (SSA), Ångström exponent (AE), absorbing AE (AAE), and scattering AE (SAE) are retrieved by the GRASP/Component algorithm with Ross_Li and Rahman–Pinty–Verstraete (RPV) models. Results indicate overall consistency in AOD, AODF, and AODC between the 2 models, with slightly better performance in North and Central America, Europe, and southern Africa. In northern Africa, the Middle East, and eastern Asia, there is slightly worse performance. Generally, the Ross_Li model yields slightly lower values for AOD and AODF, and the difference in AOD retrieval uncertainty seems primarily due to AODF. AODC retrieved by Ross_Li tends to be overestimated compared to RPV, with higher uncertainties. Uncertainties in AAOD and SSA retrievals are more complex, with the RPV model producing higher AAOD values in southern Europe and western North America. SSA retrievals are more scattered, with larger root mean square error, yet lower relative deviations due to generally larger values. The RPV model yields slightly lower SSA values, with good consistency in southern Europe and southern Asia. Uncertainties in AE, AAE, and SAE retrievals are significant, particularly in AAE, where the largest bias is observed. Improving the accuracy of AAOD inversion might reduce the uncertainty in AAE. |
|---|---|
| ISSN: | 2694-1589 |