基于OLCI数据的洱海叶绿素a浓度估算  

Estimation of chlorophyll-a concentration in Lake Erhai based on OLCI data

作　　者：毕顺 李云梅[1,2] 吕恒 朱利[3] 牟蒙[1] 雷少华 徐杰 温爽[1] 丁潇蕾[1] BI Shun;LI Yunmei;LV Heng;ZHU Li;MU Meng;LEI Shaohua;XU Jie;WEN Shuang;DING Xiaolei(Key Laboratory of Virtual Geographic Environment of Education Ministry,Nanjing Normal University,Nanjing 210023,R.P.China;Jiangsu Centre for Collaboration Innovation in Geographical Information Resource Development and Application,Nanjing 210023,R.P.China;Satellite Environment Centre,Ministry of Environmental Protection,Beijing 100094,R.P.China)

机构地区：[1]南京师范大学虚拟地理环境教育部重点实验室,南京210023 [2]江苏省地理信息资源开发与利用协同创新中心,南京210023 [3]环境保护部卫星环境应用中心,北京100094

出　　处：《湖泊科学》2018年第3期701-712,共12页Journal of Lake Sciences

基　　金：国家重点研发计划项目(2017YFB0503902);国家自然科学基金项目(41671340)联合资助

摘　　要：海陆颜色仪(OLCI)是搭载在Sentinel-3上的新型水色遥感传感器,其对于内陆清洁水体水质遥感监测的适用性有待验证.本研究以评价水体富营养化程度的重要参数叶绿素a(Chl.a)浓度为指标,以高原湖泊洱海为研究区,基于2017年4月19日共20个星地同步实验数据,建立了3种可应用于OLCI数据的Chl.a浓度遥感估算模型(波段比值模型、三波段模型以及FLH模型),并估算了当日洱海Chl.a浓度的空间分布.结果表明:(1)选用波段Oa8(665 nm)、Oa11(708.75 nm)和Oa12(753.75 nm)构建的三波段模型最适用于洱海水域的Chl.a浓度估算,其平均绝对误差百分比为12.37%,低于波段比值模型的16.04%和FLH模型的13.50%;(2)对OLCI使用的大气校正方法中,基于去瑞利散射的暗像元法对估算模型的适用性要优于6S、FLAASH以及QUAC方法;(3)洱海OLCI影像中近岸水体受邻近效应影响严重,近红外波段Oa12(753.75 nm)受陆地邻近效应影响的距离为1~2个像元,而Oa8(665 nm)、Oa10(681.25 nm)和Oa11(708.75 nm)波段为1个像元;(4)2017年4月19日全湖Chl.a浓度均值为12.15±5.72μg/L,洱海中部水域Chl.a浓度最低(9.00~12.00μg/L),北部水域浓度最高(12.00~22.76μg/L),南部水域浓度稍高(12.00~14.00μg/L),阳南溪与波罗江入湖口受降雨径流的影响出现"羽流现象",导致Chl.a浓度偏低,约为8.33μg/L.Ocean and land colour instrument(OLCI)is a new ocean colour remote sensor mounted on Sentinel-3,and its applicability to water quality monitoring of inland clean water is to be verified.Chlorophyll-a(Chl.a)concentration is an important water quality parameter for measuring the eutrophication of Lake Erhai.Based on the in-situ samples taken from April 19,2017,the performance of three Chl.a estimation models(including Band-Ratio Model,Three-Band Model and FLH Model)was evaluated in this study and the spatial distribution of Chl.a concentrations in Lake Erhai was estimated.The results showed that:(1)The Three-Band Model by Oa8,Oa11 and Oa12 was most suitable for the estimation of Chl.a concentration in Lake Erhai,the mean absolute percent error was 12.37%,lower than the Band-Ratio Model(16.04%)and FLH Model(13.50%);(2)Among atmospheric correction methods for OLCI,the application of the dark pixel method based on the rayleigh-scattering correction was better than that of the 6S,FLAASH and QUAC methods;(3)The near shore water pixels in the OLCI image of Lake Erhai were affected seriously by the land adjacency effect,and the distance of the land adjacency effect at the near-infrared band(Oa12)was 1~2 pixels while the Oa8,Oa10 and Oa11 bands were 1 pixel;(4)On April 19,2017,the average Chl.a concentration of the whole Lake Erhai was 12.15±5.72μg/L with the lowest in the middle waters(9.00-12.00μg/L)and the highest in the northern waters(12.00-22.76μg/L).Although the Chl.a concentration of the southern waters(12.00-14.00μg/L)was a little higher than that of the middle waters,the lowest concentration in the estuary of the Yangnan River and the Boluo River was around 8.33μg/L due to the“River Plume”caused by rainfall.

关 键 词：叶绿素A 洱海 海陆颜色仪(OLCI) 富营养化 

分 类 号：X524[环境科学与工程—环境工程]