Nature of optical products inverted semi-analytically from remote sensing reflectance of stratified waters

从分层水体的遥感反射率半解析反演光学产品的性质

[Abstract]

Concentrations of phytoplankton chlorophyll (Chl，mg/m^-3) and the inherent optical properties (IOPs) (Preisendorfer 1976) of water can be derived (IOCCG 2000, 2006; Lee et al., 2002) from the spectrum of remote sensing reflectance (R_rs，sr^-1), which is defined as the ratio of water-leaving radiance to downwelling irradiance just above the surface. Through measurements from ocean color satellites, such products are critical for evaluating the spatial-temporal variations of the oceans and large lakes under a changing climate (Gregg and Conkright 2002; Boyce et al. 2014; Signorini et al. 2015). For stratified waters, Gordon and Clark (1980) proposed a weighted-average formula to evaluate remotely sensed Chl vs measured Chl, which has been perceived as the standard for evaluating remote sensing products of stratified waters. However, this formula was not developed for IOPs retrieved from R_rs by a semi-analytical algorithm (SAA), thus not necessarily applicable to SAA-derived IOPs.

To fill the gap in understanding SAA-derived Chl or IOPs of stratified waters, a team lead by Prof. Lee and Prof. Shang developed theoretical expressions of the analytically inverted IOPs for stratified waters, which show that SAA-inverted IOPs are far more complex than the classic model (Gordon and Clark 1980; Gordon 1992), and the absorption and backscattering coefficients are associated with different ways of “weighted averaging”. Especially, it is found that in general the SAA-derived IOPs will be lower than its vertical average. The results will not only provide a comprehensive understanding of IOPs (and Chl) inverted from R_rs of stratified waters, but also reduce uncertainties when comparing remotely sensed products with measurements for such waters.

浮游植物叶绿素的浓度（Chl，mg/m^-3）或水的固有光学特性（IOPs）(Preisendorfer 1976)可以从遥感反射率（R_rs，sr^-1）光谱得到(IOCCG 2000, 2006; Lee et al., 2002)，其定义为水表面上方的离水辐亮度与下行辐照度之比。通过海洋水色卫星的测量，这些产品对于评估气候变化下海洋和大型湖泊的时空变化至关重要（Gregg and Conkright 2002; Boyce et al. 2014; Signorini et al. 2015）。对于分层的水体，Gordon和Clark（1980）提出了一套加权平均模式，用来对遥感产品进行有效比较。过去40年来，该模式一直是对遥感产品进行检验的“黄金标准”。然而，该模式并不是基于辐射传输的解析算法（SAA），因此并不一定适用于SAA获得的遥感产品。

为了填补对分层水体SAA反演的Chl或IOPs的理解的空白，李忠平和商少凌教授带领的团队推导出了解析算法获得的分层水体IOP的理论表达式，表明SAA反演的IOPs比经典模型（Gordon and Clark 1980; Gordon 1992）复杂得多，其中吸收系数和后向散射系数具有不同的“加权平均”关系。特别是，总体来说，SAA反演的IOPs比其深度的平均值要低。这些结果不仅可以帮助全面理解SAA获得的分层水体 的IOP和Chl特性，还将减少对此类水体遥感产品进行检验时的不确定性。

Fig. 1. (right) Spectra of b_bp at 0.5 m, weighted-average following Zaneveld et al. (2005), and that obtained from QAA and HOPE, respectively, for a synthesized Chl profile (left, zmax = 20 m).

图 1 （右）b_bp在0.5 m处的光谱，Zaneveld et al. (2005)的加权平均模型，以及分别从QAA和HOPE获得的光谱。用于模拟的Chl剖面数据（左，zmax = 20 m）。

Fig. 2. (right) Spectra of a at 0.5 m, weighted-average following Zaneveld et al. (2005), and that obtained from QAA and HOPE, respectively, for a synthesized Chl profile (left, zmax = 20 m).

图2 （右）a在0.5 m处的光谱，权重因子采用Zaneveld et al. (2005)模型，以及根据QAA和HOPE获得的光谱。用作模拟的Chl剖面图（图左，zmax = 20 m）。

Lee, Z; Shang, SL; Wang, YC; Wei, JW; Ishizaka, J. Nature of optical products inverted semianalytically from remote sensing reflectance of stratified waters. Limnology and Oceanography, 2020, 65(2): 387-400.