Skip to Main Content
Scientific implications and practical applications of spectral particulate backscattering (bbp(λ)) in oceanography are wide ranging, particularly in optical remote sensing as the light backscattered from various seawater constituents provides possibility to derive information about the particle properties of the water under investigation. Several inversion models have been previously developed for use with remote sensing reflectance (Rrs) data over open ocean waters; however, when applied to coastal waters, these models return bbp having large differences with in situ bbp values primarily because of the improper definitions of parameters of the functions describing the spectra of bbp(λ). The present study is aimed to develop a new inversion model with appropriate definitions of parameters of the functions to provide reliable retrievals of bbp in a variety of waters covering both the coastal and ocean environments. The new model is tested using large independent in situ data sets (NOMAD and Carder data sets) and simulated data provided by the IOCCG working group. When applied to these data sets, the new model outperformed the currently existing inversion models (e.g., GSM, LM and QAA models). The percent mean relative error (MRE) and root mean square error (RMSE) were found to be MRE -5.16 ~ 0.35% and RMSE 0.114 ~ 0.146 (in the 412-555 nm range) for the simulated data, MRE -0.14 ~ 2.42% and RMSE 0.125 ~ 0.157 for the NOMAD data, MRE -0.77% ~ 2.23% and RMSE 0.124 ~ 0.15 for the Korean regional data, and MRE 6.88% and RMSE 0.218 for the Carder data (for 490 nm only). Slopes close to unity, high R2 and low intercept values also indicated that the new model provides better performance over other models. The results further suggest that the new model is more robust and can be effectively applied to satellite ocean color data to retrieve the particulate backscattering coefficients in a variety of waters commonly found in the coastal and offshore domains.