渤黄海水体漫衰减系数的遥感反演

A Novel Remote Sensing Algorithm for Estimating Diffuse Attenuation Coefficient in the Bohai Sea and Yellow Sea

关键词:  海洋光学  经验算法  半分析算法  联合算法

Key words:  marine optics  empirical algorithm  semi-analytical algorithm  combined algorithm

摘要:

[目的]针对中国渤黄海水体类型复杂的特性,建立一个同时适用于浑浊水域与清澈水域的漫衰减系数反演算法。[方法]利用2000年7月至2004年2月在中国近海海域现场测量的238组光谱数据,针对浑浊水域与清澈水域分别建立漫衰减系数的半分析算法和经验算法,并基于加权方法将两算法结合,构建水体联合算法。[结果]新建立的联合算法精度较高且较为稳定:算法的决定系数(R²)达到0.891,均方根误差(RMSE)与平均绝对误差(MAPE)分别为0.543 m^-1,26.77%。在误差敏感性分析中,半分析算法与经验算法的MAPE与RMSE在6%和0.06 m^-1以内。[结论]新的反演算法适用于我国渤黄海水体漫衰减系数K_d(490)的反演。

Abstract:

[Objective] Considering the complex characteristics of the Chinese Seas, a novel,rather high accuracy and stable algorithm was built to suit for both the clear water body and the turbid water body.[Methods] Based on in situ data around the coastal regions of China from January 2000 to February 2004,semi-analytical and empirical algorithms were developed for deriving diffuse attenuation coefficient in turbid and clear waters,respectively.Based on the method of weighting,these two algorithms were combined as a new algorithm.[Results] The retrieval result of the new combined algorithm shows that the correlation coefficient(R²) of the algorithm reaches 0.891,and the RMSE and the MAPE are 0.543 m^-1 and 26.77%,respectively.Furthermore,the precision and the stability of the algorithm are relatively good in model validation.[Conclusion] The new algorithm is suitable for the retrieval of K_d(490) in Bohai Sea and Yellow Sea.