| 引用本文： |
-
谢明媚,孙德勇,丘仲锋,王胜强,路颖,吴晨颖,叶之翩,岳小媛.长江口水质MERIS卫星数据遥感反演研究[J].广西科学,2016,23(6):520-527. [点击复制]
- XIE Mingmei,SUN Deyong,QIU Zhongfeng,WANG Shengqiang,LU Ying,WU Chenying,YE Zhipian,YUE Xiaoyuan.Water Quality Retrievals from MERIS Satellite Data in Yangtze Estuary[J].Guangxi Sciences,2016,23(6):520-527. [点击复制]
|
|
| |
|
|
| 本文已被：浏览 427次 下载 461次 |
码上扫一扫！ |
| 长江口水质MERIS卫星数据遥感反演研究 |
|
谢明媚, 孙德勇, 丘仲锋, 王胜强, 路颖, 吴晨颖, 叶之翩, 岳小媛
|
|
|
| (南京信息工程大学海洋科学学院, 江苏南京 210044) |
|
| 摘要: |
| [目的]获取准确的水质参数分布情况,进而对水质参数与动力作用复杂的河口水域环境进行综合评价。[方法]利用2011年5月30组长江口水域的遥感反射率数据,在尝试多种波段组合以及不同函数形式后,针对叶绿素a浓度和总悬浮物浓度分别建立最优经验反演模型。[结果]对总悬浮物浓度,波段差值(634~644 nm)的二次函数形式最优,决定系数R2为0.837,均方根误差(RMSE)为 0.226 mg·L-1,利用独立的验证样本得到平均绝对百分比误差(MAPE)为58.2%。对叶绿素a浓度,波段比值(650 nm/644 nm)的二次函数形式最优,R2为0.552,RMSE为0.486 mg·m-3,利用独立的验证样本得到MAPE为66.2%。将模型运用于2011年5月MERIS卫星数据,反演出长江口水域叶绿素a浓度与总悬浮物浓度空间分布图,叶绿素a浓度呈现出从河口向外海逐渐减少的趋势,最大值出现在舟山群岛附近。总悬浮物浓度呈阶梯状向外海减少。[结论]通过评价参数可看出,总悬浮物浓度反演模型对总悬浮物浓度反演效果较为准确,而叶绿素a浓度反演模型显示对叶绿素a的反演浓度较低。 |
| 关键词: 叶绿素a浓度 总悬浮物浓度 MERIS 卫星遥感算法 |
| DOI：10.13656/j.cnki.gxkx.20161230.002 |
| 投稿时间：2016-09-26 |
| 基金项目:国家自然科学基金项目(41276186,41576172,41506200),江苏省自然科学基金项目(BK20151526,BK20150914),江苏省高校自然科学基金项目(15KJB170015),"全球变化与海气相互作用"专项(GASI03030101),江苏省"青蓝工程"优秀青年骨干教师项目和南京信息工程大学大学生实践创新训练计划项目(201510300256,201510300073,201610300035,201610300064)资助。 |
|
| Water Quality Retrievals from MERIS Satellite Data in Yangtze Estuary |
|
XIE Mingmei, SUN Deyong, QIU Zhongfeng, WANG Shengqiang, LU Ying, WU Chenying, YE Zhipian, YUE Xiaoyuan
|
| (School of Marine Science, Nanjing University of Information Science & Technolgy, Nanjing, Jiangsu, 210044, China) |
| Abstract: |
| [Objective] In order to grasp the distribution map of water quality parameters more accurately,we evaluate the estuary which has complex dynamical system.[Methods] After using 30 in situ remote sensing reflectance data collected in May 2011 and trying many kinds of band combinations and function forms,the optimal empirical models for chlorophyll-a concentration and total suspended matter concentration were established.[Results] For total suspended matter concentration,Quadratic function model by band difference(634~644 nm) performs best,of which coefficient of determination (R2) is 0.837 and root mean square error(RMSE) is 0.226 mg·L-1.In the model validation,the mean absolute percentage error(MAPE) by using the independent dataset shows a value of 58.2%.For chlorophyll-a concentration,quadratic function model by band ratio (650 nm/644 nm) performs best,of which coefficient of determination (R2) is 0.552 and root mean square error(RMSE) is 0.486 mg·m-3.In the model validation,the mean absolute percentage error(MAPE) by using the independent dataset shows a value of 66.2%.The MERIS satellite data in May 2011 were evaluated by the models,and the distribution map was obtained for spatial concentration of chlorophyll-a and total suspended matter of the Changjiang Estuary through simulation.There was a trend that the concentration of chlorophyll-a decreased gradually from the estuary to the sea,and the maximum value appeared in the Zhoushan Islands.The concentration of total suspended matter reduced stepwisely to the sea.[Conclusion] As is shown in the evaluation parameters,the total suspended matter concentration model has accurate effect in retrieve water quality while the chlorophyll-a concentration model has lower effect. |
| Key words: chlorophyll-a concentration total suspended matter concentration MERIS remote sensing algorithm |
|
|
|
|
|
