EUVIMEDTutustu EUVIMEDiin
Crossref

Satellite-derived benthic cover of coral reef from multispectral data: comparison of different bottom reflectance parameterizations

Yuye Huang, Shilin Tang, Yongming Liu, Hongqiang Yang, Yupeng Liu, Yangbin Huang, Sufen Wang, Song Gao

Applied Optics · 2026-08-01 · 65 (22) · 7645

DOI: 10.1364/ao.608653 ↗

At a glance

Journal article in Applied Optics, published online on 2026-07-29. Written by Yuye Huang and 7 co-authors. 70 references. Funded by: Science & Technology Fundamental Resources Investigation Program, National Natural Science Foundation of China, Self-Deployed Project of South China Sea Institute of Oceanology, Chinese Academy of Sciences.

TiivistelmäBibliografiset tiedotCiteRelated studies

Abstract (original language)

Coral reef ecosystems are deteriorating extensively under the dual pressures of anthropogenic activities and global warming. Benthic cover generally serves as a critical proxy for reef health. Although existing studies have proved that satellite remote sensing technology can derive benthic cover, satellite-derived benthic cover (SDBC) faces uncertainties due to high heterogeneity and complex reef structural characteristics, which lead to bottom reflectance mixing and nonlinear scattering effects (e.g., multiple scattering and bidirectional reflectance effects). The purpose of this paper was then to evaluate the impacts of different bottom reflectance parameterizations on SDBC retrieval using three spectral matching-based models, namely the pure pixel-based spectral matching (SM-P) model, the linear spectral unmixing-based spectral matching (USM-L) model, and the polynomial post-nonlinear spectral unmixing-based spectral matching (USM-PPNM) model. Results from HydroLight-simulated Sentinel-2 multispectral datasets indicated that spectral unmixing-based models enhance retrieval accuracy by effectively decoupling mixed bottom reflectance signatures. Specifically, the root mean square error (RMSE) and the median absolute percentage difference (MAPD) of the USM-PPNM and USM-L models are both lower than those of the SM-P model at depths of 0.5–20 m. Additionally, by explicitly quantifying nonlinear scattering effects, the nonlinear unmixing model slightly outperforms the linear counterpart, with the USM-PPNM model achieving a MAPD of 26.96% versus 27.21% for USM-L. Comparison results from Sentinel-2 imagery of selected reefs in the Xisha Islands showed that the USM-PPNM and USM-L models exhibited comparable performance as validated by in situ benthic cover measurements, and their benthic cover spatial patterns align with those of the Allen Coral Atlas. This work reveals the significant impacts of bottom reflectance configurations, offering optimized strategies for accurate coral reef benthic cover retrieval in complex shallow-water environments.

Funding

  • Science & Technology Fundamental Resources Investigation Program
  • National Natural Science Foundation of China
  • Self-Deployed Project of South China Sea Institute of Oceanology, Chinese Academy of Sciences
  • State Key Laboratory of Tropical Oceanography
  • Guangdong Basic and Applied Basic Research Foundation
  • Project of State Key Laboratory of Tropical Oceanography
  • Hainan Provincial Natural Science Foundation of China

Cite

APA

Huang, Y., Tang, S., Liu, Y., Yang, H., Liu, Y., Huang, Y., Wang, S., & Gao, S. (2026). Satellite-derived benthic cover of coral reef from multispectral data: comparison of different bottom reflectance parameterizations. Applied Optics, 65(22), 7645. https://doi.org/10.1364/ao.608653

Vancouver

Huang Y, Tang S, Liu Y, Yang H, Liu Y, Huang Y, et al. Satellite-derived benthic cover of coral reef from multispectral data: comparison of different bottom reflectance parameterizations. Applied Optics. 2026;65(22):7645. doi:10.1364/ao.608653

BibTeX
@article{Huang2026,
  author = {Huang, Yuye and Tang, Shilin and Liu, Yongming and Yang, Hongqiang and Liu, Yupeng and Huang, Yangbin and Wang, Sufen and Gao, Song},
  title = {Satellite-derived benthic cover of coral reef from multispectral data: comparison of different bottom reflectance parameterizations},
  journal = {Applied Optics},
  year = {2026},
  volume = {65},
  number = {22},
  pages = {7645},
  doi = {10.1364/ao.608653}
}
Start for free →