【作者】 雷惠；

【导师】 潘德炉；

【作者基本信息】 浙江大学 ， 地图学与地理信息系统， 2011， 博士

【摘要】 赤潮是东海主要海洋灾害之一,建立准确的遥感赤潮提取算法是对其及时有效的监测,降低危害的重要前提。目前海洋赤潮的遥感提取算法主要是叶绿素浓度异常或直接的光谱(表观光学量)算法。东海赤潮高发区水体光学性质复杂,受水体有色溶解有机物、悬浮颗粒等的干扰,已有的赤潮光谱提取算法在该海域的应用往往失效。为有效提取该海域的赤潮,需要结合赤潮藻类粒径、生理特性等,从固有光学量入手对水体性质进行分析,以实现对该海域赤潮水体的有效提取。对于赤潮的防范和治理,人们不仅需要对其发生范围的有效监测,还更加关注有害赤潮类型的准确识别与预警,目前国内在这方面的研究工作还较少展开。固有光学量能够直接反映水体物质的生理生化特性,随水体物质的变化而表现出差异,在有效排除近岸水体物质干扰,实现赤潮水体与非赤潮水体的有效区分和赤潮藻种识别方面有着巨大的潜力。因此,本文从水体固有光学量入手,对东海的赤潮与非赤潮水体,以及甲藻和硅藻赤潮进行分析,建立适用于该海区的赤潮提取算法。主要研究内容和成果为：(1)论文首先利用2009年东海海域夏季和冬季两个大型调查航次实测数据,系统分析了海区水体吸收和散射系数等固有光学要素在季节上和空间上的分布与变化情况,获得了东海赤潮的固有光学量背景场。发现在中等浑浊水体中,陆源输入的有色溶解有机物和非藻类碎屑颗粒对色素吸收光谱存在明显干扰。(2)对东海近十年来发生的赤潮事件进行分析,发现东海典型赤潮藻种以甲藻(东海原甲藻、米氏凯伦藻)和硅藻(中肋骨条藻)等类型为主,其中甲藻类常形成有毒赤潮。论文分析了长江口与温州南麂列岛海域观测到的硅藻与甲藻赤潮事件水体固有光学性质参数,获得了东海赤潮与非赤潮环境水体吸收系数与散射系数的固有光学性质差异,发现甲藻类与硅藻类赤潮水体在实测固有光学性质上可以实现区分。(3)论文分别从实测数据和遥感赤潮数据对东海典型赤潮事件固有光学量参数进行对比分析,建立了用以识别赤潮水体以及区分甲藻与硅藻类不同赤潮的算法模型,包括色素吸收比重,叶绿素比吸收系数,散射-吸收比值和后向散射比率四种吸收和散射固有光学性质赤潮模型,并开发了基于遥感反射率的赤潮水体光谱高度法识别模型。在此基础上,建立了综合的赤潮卫星遥感识别算法,可用于对东海赤潮水体的提取和甲藻与硅藻赤潮的判别。(4)将建立的基于固有光学量赤潮识别算法模型应用于近年来东海大型赤潮事件卫星遥感数据集,实现了对赤潮水体发生范围和具体藻种门类的准确识别,总体获得了良好的效果。在试验的十六次样本中,对赤潮藻种门类的判别准确率达到了92%。本文研究的主要创新点为：(1)在深化认识赤潮水体区别于正常水体的固有光学特性的基础上,发现了东海甲藻与硅藻赤潮水体的固有光学性质差异,对相关赤潮发生、发展和消亡等的科学研究有重要意义。(2)基于赤潮水体以及甲藻与硅藻间的固有光学特征,建立了光谱高度法、色素比重法和后向散射比值法三种卫星遥感监测赤潮模型以及两藻种的分类模型,验证表明该模型有实用性,对自动化监测赤潮有应用价值。更多还原

【Abstract】 Harmful Algal Bloom (HAB) is one of the most common disasters in the East China Sea (ECS), and the establishment of an accurate remote sensing algorithm of HAB extraction is the primary concern of timely and effective monitoring and predicting to minimize the harm. The Ocean Color remote sensing algorithms of HAB extractions now mainly concentrate on chlorophyll concentration anormaly or spectral (Apparent Optical Properties) extraction algorithms. The seawater properties in the HAB highly occur area in ECS is very complicated, which is under severe influences of colored dissolved organic matter as well as suspended particles, and thus the spectral extraction algorithms are usually prone to be invalid. In order to extract the HAB effectively, the seawater Inherent Optical Properties (IOPs) including algal cell size and physiological properties need to be analyzed. Typically, researchers are more willing to develop a targetable algorithm for a typical algal type, which usually tends to be more harmful and popular, so as to prevent and govern the disasters more effectively. Currently, research in this area is little yet.IOPs can reflect the physiological and biochemical properties of the water body and reflect differences due to the substance changes. It plays significant role in developing proper extraction algorithms, and possesses huge potentials in HAB monitoring and identification. Accordingly, the thesis starts with the IOPs of seawater of dinoflagellates and diatom type HAB and the normal environment in ECS to establish the extraction algorithm. The content and achievement of this thesis is as follows:(1) Based on the research and analyses of seawater IOPs parameters from the two large scale investigations conducted in ECS in summer and winter,2009. the seasonal and spatial distributions and variations of seawater absorption and scattering coefficients are analyzed, and it forms the background of HAB in ECS. It is found that, in moderate turbid seawaters, the terrestrial colored dissolved organic matter as well as non-algal detritus particles will apparently interfere the pigment absorption spectra.(2) By analyzing the HAB events occurred in the last ten years in ECS, it is found that the main type of HAB is dinoflagellates (i.e. Prorocentrum donghaiense and karenia mikimotoi) and diatom (i.e. Skeletonema costatum). and the dinoflagellates type usually forms toxic HAB. And accordingly, the IOPs of these type HAB events measured from the Changjiang Estuary and Nanji Islands respectively are analyzed, and the differences of seawater absorption and scattering coefficient are gained for discrimination.(3) The IOPs data from insitu and remote sensing of HAB waters in ECS are analyzed, and a model for eliminating the HAB region and identifying the algal type is established, based on the pigment absorption ratio, the chlorophyll specific phytoplankton absorption coefficient, the back scattering-to-total absorption coefficient ratio, and the backscattering coefficient ratio, respectively. With aid of reflectance spectral properties, the extraction and identification of different algal classes such as diatom and dionflagellates are achieved.(4) The established model is applied to the big HAB events in ECS and the final result of extraction and identification is satisfied. The accuracy rate of identification for the test of the 16 samples is up to 92%.The main innovations of this work are:(1) The characteristics and discrimination of IOPs between HAB water and normal seawater as well as dinoflagellates and diatom type HAB seawater in ECS are eliminated, which is of significance to the scientific research on HAB development.(2) A series of extraction models including the spectra relative height, the pigment absorption proportion and the backscattering ratio based on IOPs of seawater absorption and backscattering coefficient are established and applied for HAB extraction and class type discrimination, and it is proved to be practical and of reference value for automatic HAB detection.更多还原