利用渤、黄海407组实测透明度(0.1~18.0 m)与遥感反射率光谱数据, 评估了3种透明度半分析遥感反演算法在渤、黄海的适用性, 包括基于709 nm波段、基于560 nm波段和基于固有光学量准分析算法的透明度半分析算法。结果表明: 3种半分析算法均不适用于渤、黄海水体, 透明度反演值的均方根误差(RMSE)都大于2.2 m, 平均相对误差(APD)都高于68%。其中, 基于560 nm波段的半分析算法反演误差相对较小, RMSE为2.2~4.6 m, APD为68.0~86.7%。

Transparency, usually characterized by the Secchi disk depth ZSD, is a basic parameter of optical properties of ocean water. Main semi-analytical algorithms for transparency estimation are validated based on 407 in-situ data in the Bohai Sea and Yellow Sea (the measured ZSD is in the range of 0.1~18.0 m). The semi-analytical algorithms include those based on remote sensing reflectance at 709 nm, at 560 nm and the quasi-analytical algorithm (QAA) of the inherent optical properties. Comparisons between Secchi depth retrievals from semi-analytical algorithms and in-situ measurements show that these three algorithms are not applicable to the Bohai Sea and Yellow Sea. The root mean square error (RMSE) is larger than 2.2 m and the average absolute percentage difference (APD) is larger than 68%. The semi-analytical algorithm based on remote sensing reflectance at 560 nm shows a better performance with RMSE of 2.2~4.6 m and APD of 68.0~86.7%, compared to the algorithms based on remote sensing reflectance at 709 nm (RMSE of 10.8~15.9 m and APD of 70.4~116.3%) and QAA (RMSE of 2.8~4.4 m and APD of 73.9~115.1%).