|
|||||
|
|
| 二类水体水色遥感的主要进展与发展前景 | |
| 引用本文： | 任敬萍,赵进平.二类水体水色遥感的主要进展与发展前景[J].地球科学进展,2002,17(3):363-371. |
| 作者姓名： | 任敬萍 赵进平 |
| 作者单位： | 1. 中国科学院海洋研究所,山东,青岛,266071 2. 国家海洋局第二海洋研究所,浙江,杭州,310012 |
| 基金项目： | 山东省青年基金项目“SeaWiFS资料监测海岸带生态环境的应用研究”(编号 :Q98E0 113 8)，中国科学院知识创新工程项目“近海海洋动力过程的遥感分析与应用研究”(编号 :KZCX2 2 0 2 )资助 |
| 摘 要： | Ⅱ类水体水色遥感是海洋水色遥感的难点和热点。针对Ⅱ类水体的光谱特性和海洋现象的特点，综述了水色卫星传感器在光谱波段配置、辐射探测性能和时空分辨率等方面的设计要求与技术进步。从水色遥感资料反演的两大关键技术——大气校正和生物光学算法两个方面，概述了Ⅱ类水体水色反演算法的研究现状和发展方向。根据我国近海的水体特点，提出了我国水色遥感研究需要解决的关键问题。 |
| 关 键 词： | 水色遥感 Ⅱ类水体 反演算法 |
| 文章编号： | 1001-8166(2002)03-0363-09 |
| 收稿时间： | 2001-05-08 |
| 修稿时间： | 2001年5月8日 |
Progress and Prospect of Ocean Color Remote Sensing in Case 2 Waters |
|
| REN Jing ping ,ZHAO Jin ping.Progress and Prospect of Ocean Color Remote Sensing in Case 2 Waters[J].Advance in Earth Sciences,2002,17(3):363-371. | |
| Authors: | REN Jing ping ZHAO Jin ping |
| Institution: | 1. Institute of Oceanology, CAS, Qingdao 266071,China; 2. Second Institute of Oceanography, SOA, Hangzhou 310012,China |
| Abstract: | case 1 waters and case 2 waters are different water types defined by optical characteristics. case 1 water is clear, open-ocean water, and case 2 is generally coastal, higher productivity, turbid water. Ocean color in case 2 waters is influenced by three major components of the water, namely phytoplankton pigment, suspended sediment and yellow substance. case 2 waters are more complex than case 1 waters in their composition and optical properties. To date, remote sensing of ocean color has focused largely on case 1 waters. It has been demonstrated that the standard algorithms in use today for chlorophyll retrieval from satellite data work well in case 1 waters, but they often break down in case 2 waters. With the advent of the new sensors and the emergence of the new algorithm in parallel, better interpretation of ocean color in case 2 waters are under intensive investigation. Technical requirements for ocean color measurements are reviewed first according to the spectral signatures and ocean processes in case 2 waters. The minimum requirements for ocean color sensors designed for case 1 applications are introduced. Ocean color sensors for case 2 waters must meet all the requirements for case 1 waters, as well as the special requirements for case 2 waters. ①In the visible domain, additional spectral channels are required for the measurement of chlorophyll fluorescence, suspended sediment, yellow substance and shallow bottom reflectance. In the near infrared region, one or more additional channels are required for atmospheric correction over shallow or turbid coastal waters because of the non-zero water leaving radiance beyond 700 nm. ②Because the range of remote sensing reflectance in case 2 waters is larger than in case 1 waters, the sensitivity and signal-to-noise ratio must be increased. In addition, ocean color sensors must not saturate over clouds or the coast, so very high dynamic range is required. ③More temporal resolution and more spatial resolution are required to monitor the dynamical processes of the coastal zone. No single existing or planned satellite sensors meet all those requirements. Monitoring of coastal waters must involve sensors aboard various platforms, whether they are spaceborne, airborne, in situ or land-based. Development of retrieval algorithms is then elucidated. With at least three groups of different color producing components, all varying independently with local and seasonal variations, remote sensing in case 2 waters is a non-linear, multivariate problem, and the algorithms must be designed accordingly. The algorithms are being developed toward treating the ocean-atmosphere system as a coupled system, retrieving aquatic properties based on theoretical models and introducing new and powerful mathematical and statistical approaches to solve non-linear, multivariate problem. Atmospheric correction in turbid coastal waters is complicated by the occurrence of non-zero water leaving radiance beyond 700 nm. There are two approaches to correcting for the effect of the near infrared contribution of water leaving radiance to the atmospheric correction. One is to use a coupled hydrological atmospheric model to calculate the atmospheric path radiance iteratively. Another is to apply the aerosol type observed over adjacent, less turbid waters to the turbid water pixels. Inverse techniques can also be used to estimate simultaneously in-water constituents and aerosols. Bio-optical algorithms, including empirical approaches and model-based approaches, are reviewed. Empirical algorithm is successful in case 1 waters, but its accuracy is usually relatively-low in Case 2 waters. Model-based algorithms use bio-optical models to describe the relationship between water constituents and spectra of water leaving radiance and reflectance, and use radiative transfer models to simulate the light propagation through the water and the atmosphere. There are four major groups of algorithms developed to date, the algebraic methods, the non-linear optimization techniques, the principal component approach and the neural network approach. The algebraic method is a semi-analytical model, and is the simplest of the model-based approaches. The non-linear optimization directly invert the forward model to estimate simultaneously all the concentrations of the aquatic constituents by minimizing the differences between the calculated values and the measured radiances using such minimization techniques as the simplex algorithm, Levenberg-Marquardt method and the Gauss-Newton algorithm, etc. It doesn't depend on a pre-defined, simulated data set. The complexity level and retrieval accuracy are the highest among the four approaches. The principal component analysis of simulated data is introduced to deal with the high correlation between signals from different wavebands in case 2 waters, determining the spectral dimensionality of the data and the weighting coefficients of each spectral channel. Atmospheric correction is unnecessary and the model can be implemented at high computation speed. The neural network is a powerful approach to the retrieval of water constituents and also to atmospheric correction over case 2 waters. The disadvantage is that the design of the training and test data set and the training procedure require extensive experience. In summary, the algorithms are still far from being operational, and significant improvements are needed. Finally, key issues for ocean color remote sensing in China seas are discussed based on the specific characteristics of the coastal waters. Most of the East China Seas belong to case 2 waters, and the optical properties of the shallow, turbid waters are very complex. It is well recognized worldwide that the turbid atmospheric correction over East China Seas is a difficult task, and new algorithms are needed to address the challenge. Up to now, we know little about the Inherent Optical Properties in case 2 waters, and in situ measurements are necessary both in the development of algorithms, as well as for subsequent validation of the retrieval results from satellite data. Bottom reflectance, bottom characterization and bathymetry must be taken into account when developing retrieval algorithms in optically-shallow waters. The optical properties of coastal waters vary greatly over time and space. Regional algorithms optimized for local conditions are required to better interpret ocean color in China Seas. |
| Keywords: | Ocean color remote sensing Case 2 waters Retrieval algorithm |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ | |
| 点击此处可从《地球科学进展》浏览原始摘要信息 | |
| 点击此处可从《地球科学进展》下载免费的PDF全文 | |