Publications

Yu, XL; Lee, ZP; Lai, WD (2023). Global distribution of the spectral power coefficient of particulate backscattering coefficient obtained by a neural network scheme. REMOTE SENSING OF ENVIRONMENT, 296, 113750.

Abstract
The spectral power coefficient (eta) of the particulate backscattering coefficient (b(bp)(lambda)) is directly estimated from the remote sensing reflectance (R-rs(lambda)) with a neural network-based scheme (NN eta) in this study. Evaluations with both synthetic dataset and in-situ measurements show that NN eta could significantly improve the accuracy of estimated eta compared to several conventional schemes reported in the literature that are based on chlorophyll-a concentration (Chl), band ratios of R-rs(lambda), or remotely sensed b(bp)(lambda). Demonstrations with measurements from MODerate resolution Imaging Spectroradiometer (MODIS) Aqua further confirm the robustness of NN eta, where reasonable spatial distribution and seasonality of eta in the global oceans can be acquired by NN eta. High and low eta values are observed in the oligotrophic gyres and the coastal zones, respectively, which are consistent with the current understanding of eta distribution concluded from theoretic analysis and repeated field measurements in the global ocean. Implementation of NN eta to 19-year MODIS monthly composite measurements from 2003 to 2021 reveals strong seasonal variations of eta in most of the global ocean, but the decadal changes of eta are insignificant in the majority (similar to 82.2%) of the global ocean. Similar to any empirical algorithms, the performance of NN eta is dependent on the training dataset, particularly its range, here a proper upper limit of eta for natural waters is provided.

DOI:
10.1016/j.rse.2023.113750

ISSN:
1879-0704