Second‐order variability of inherent optical properties of particles in Bohai Sea and Yellow Sea: Driving factor analysis and modeling

Abstract

This study explores second-order variability of inherent optical properties (IOPs) of particles caused by particle intrinsic attributes, including the mean particle size (D A ) calculated by a LISST measurement restricted to a size range of 2.5-500 lm, mean apparent density (q a ), bulk refractive index (ñ p ), and composition (using a ratio of chlorophyll a to total suspended matter concentrations, Chl a/TSM, as the delegate), obtained from five field cruise observations collected from 2013 to 2015 in Bohai Sea (BS) and Yellow Sea (YS). Large variations in magnitude were observed in terms of mass-specific beam attenuation (c à p ), scattering (b à p ), and backscattering (b à bp ), indicative of optically complex water conditions. c à p and b à p showed similarities in their change and driving mechanisms, in which D A explained the majority (35%), bothñ p and q a separately contributed variability levels of approximately 25%, and Chl a/TSM accounted for the lowest proportion (15%). Most (approximately 38%) of the variability in b à bp was found to be attributed toñ p when the other factors were responsible for the rest of the variation to different degrees of 9-29%. We also model c à p , b à p , and b à bp with direct linkages to D A , q a ,ñ p , and Chl a/TSM using a multivariate linear regression method. The established c à p , b à p , and b à bp models generated very high goodness of fit results, with determination coefficients (R 2 ) of above 0.970 (p < 0.001) and with favorable predictive errors (a mean absolute percentage error, MAPE, of approximately 30% for the c à p and b à p models and of approximately 40% for the b à bp model). Overall, the findings of this study further knowledge on the nature of the inherent optical properties in BS and YS, and are very helpful for understanding many marine biogeochemical processes.

Estimation of Particulate Backscattering Coefficient in Turbid Inland Water Using Sentinel 3A-OLCI Image

“…In brief, Rrs(λ) was obtained by a FieldSpec spectroradiometer [20]. Particulate backscattering coefficients in six bands (442 nm, 488 nm, 532 nm, 590 nm, 676 nm, and 852 nm) were measured by using a HOBI Labs Hydroscat-6P (HS-6P) [21]. Before cruising, HS-6 was calibrated to confirm its performance.…”

“…Based on the assumptions that the total absorption at 852 nm can be regarded as the absorption of pure water [36,37] and that the particulate backscattering at 852 nm is approximately equal to the particulate backscattering at 865 nm, the method proposed by Lyu [38] was adopted to estimate the particulate backscattering coefficient at 852 nm by using the relationship between remote sensing reflectance and inherent optical properties [14,39]. The estimation algorithm is given in (21).…”

“…The values of aw(865) and bw(865) are 4.6052 m -1 and 0.00014 m -1 , respectively [40]. Finally, the particulate backscattering in all six bands can be obtained based on the reference particulate backscattering coefficients at a wavelength of 852 nm using (21) and the particulate backscattering coefficient simulation function using ( 10), ( 14) and (15). The procedures of backscattering coefficients estimation at six bands was shown in Fig.…”

“…The concentration of suspended particles has been considered to be the main factor affecting the backscattering coefficient, and the essential characteristics of particles such as composition and size of particulate matter are the main contributors to the second-order changes in backscattering coefficient [21,67,68]. In previous studies, the ratio of particulate organic carbon to suspended particulate matter (POC/SPM) [69], the ratio of particulate organic matter to suspended particulate matter (POM/SPM) [70], and Chla/TSM [13,71] have often been used as proxies of the particle composition.…”