近岸二类水体生物光学模型参数优化  

Optimization of Bio-Optical Algorithm Parameters in Coastal CASE-II Waters

作　　者：韩留生[1,2,3] 陈水森[3] 李丹[3] 陈修治[4] 李勇[3] 陈燕乔[3] 方立刚[5] 

机构地区：[1]中国科学院广州地球化学研究所,广州510640 [2]中国科学院大学,北京100049 [3]广州地理研究所,广州510070 [4]中国科学院华南植物研究园,广州510650 [5]苏州市职业大学,苏州215104

出　　处：《热带地理》2014年第3期351-358,共8页Tropical Geography

基　　金：广东省科技计划重点项目(2011B031100003);国家自然科学基金项目(41201338)

摘　　要：为了优化生物光学模型中的悬浮物后向散射模型系数(bbx)和双向反射因子(f/Q),基于珠江口、韩江河口、徐闻珊瑚礁保护区实测数据(遥感反射率、叶绿素a质量浓度、悬浮物质量浓度、CDOM吸收系数),采用模拟退火算法同时优化了bbx以及f/Q(N=43),并开发了生物光学正演模型(N=22)。文章优化的f/Q为0.104 9,悬浮物在参考波长531 nm处后向散射系数bbx为0.268×[TSS]0.295。通过输入实测的水质数据模拟出531 nm处的遥感反射率,并外推到其他波段(412、443、490、551、667 nm)。模拟得到的531 nm处遥感反射率与实测值的均方根误差RMSE=0.000 8,N=22,除了模拟的667 nm处遥感反射率误差较大外(RMSE=0.003 6,N=22),其他波段都具有较好的精度(RMSE<0.002 3)。结果表明,该研究区与其他研究区(如黄、东海)的bbx和f/Q不一致,这2个参数需要根据研究区的数据进行优化。验证结果说明迭代优化的bbx以及f/Q能够适用于广东近岸河口生物光学参数模拟。The optimization of bio-optical algorithm parameters in remote sensing of CASE-II water is a necessary step before they can be applied, because these parameters vary with the changes of study area. Biochemical parameters （Chlorophyll a, Total suspended solids, CDOM absorptioncoefficient）and above-water surface reflectance collected inthePearl River Estuary, Hanjiang Estuary and Xuwen Coral Reef Zone were used to iterate the back scattering coefficient model of total suspended solids （bbx）and the bidirectional reflectance （f/Q） by simulated annealing algorithm （N=43）. The calibration results indicated that the f/Q was 0.1049, and thebbx at the wavelength of 531nm was 0.268×[TSS]0.295. Then, the forward model was developed to test the optimized results, i.e. the remote sensing reflectance was simulated by inputting water quality parameters （N=22）. The validation results showed that reflectance obtained from the optimized algorithm agreed well with field observations（RMSE=0.0008,N=22）. The remote sensing reflectance at the wavelength of 667nm extrapolated from the Bio-optical Algorithm had poor accuracy （RMSE=0.0036,N=22）compared with the field data , while all the other band （412,443,490,551 nm）had good precision （RMSE〈0.0023,N=22）. It showed thatbbx andf/Q varied with different study areas, and they should be optimized according to the measured data. Thebbx and f/Q values obtained from the optimized algorithm can be applied to bio-optical algorithm.

关 键 词：二类水体 生物光学模型 模拟退火算法 水质参数 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]