地貌分带下基于ICESat-2与GF-1的珊瑚礁水深遥感优化反演方法  

Optimal bathymetry retrieval method for coral reef habitats based on ICESat-2 and GF-1 under geomorphic zoning

作　　者：段正贤 左秀玲 余克服 苏奋振[4] 臧声琳 覃宾妮 吴会胜[5] Duan Zhengxian;Zuo Xiuling;Yu Kefu;Su Fenzhen;Zang Shenglin;Qin Binni;Wu Huisheng(School of Marine Sciences,Guangxi University,Nanning 530004,Guangxi,China;Guangxi Laboratory on the Study of Coral Reefs in the South China Sea,Guangxi University,Nanning 530004,Guangxi,China;Southern Marine Science and Engineering Guangdong Laboratory(Zhuhai),Zhuhai 519080,Guangdong China;State Key Laboratory of Resources and Environmental Information System,Institute of Geographic Sciences and Natural Resources Research,Chinese Academy of Sciences,Beijing 100101,China;College of Oceanography and Space Informatics,China University of Petroleum(East China),Qingdao 266580,Shandong,China)

机构地区：[1]广西大学海洋学院,广西南宁530004 [2]广西南海珊瑚礁研究重点实验室,广西南宁530004 [3]南方海洋科学与工程广东省实验室(珠海),广东珠海519080 [4]中国科学院地理科学与资源研究所资源与环境信息系统国家重点实验室,北京100101 [5]中国石油大学(华东)海洋与空间信息学院,山东青岛266580

出　　处：《地理科学》2023年第9期1640-1648,共9页Scientia Geographica Sinica

基　　金：广西自然科学基金项目(2022GXNSFAA035548);国家自然科学基金项目(42276182,41801341);中国科学院战略性先导科技专项(类)(XDA13010400)资助。

摘　　要：研究以中国南海西沙群岛华光礁西部剖面为试验区,采用高分一号(GF-1)遥感影像及“冰、云和陆地高程2号”(ICESat-2)星载激光测高数据,基于珊瑚礁地貌分带建模优选策略构建了试验区25 m以浅水深遥感反演模型,并将模型外推应用至西沙群岛有实地调查水深数据的珊瑚礁剖面研究区。建模结果显示,礁坡和深潟湖地貌采用波段比值二项式回归模型,礁坪、浅潟湖和点礁地貌均采用多波段比值回归模型,可以使得地貌单元内基于GF-1影像的水深遥感反演达到最优效果。模型反演水深数据与实地调查水深数据相关系数为0.95;实地调查水深数据与ICESat-2深水区数据综合验证结果显示,模型均方根误差为1.34 m,平均绝对误差为1.08 m,平均相对误差为17.10%。研究表明,本文提出的基于地貌的主被动遥感水深反演模型显著改进了珊瑚礁区中分辨率遥感水深反演模型的精度,可为大范围珊瑚礁区高精度的水深反演提供方法支持。In this study,the western profile of Huaguang Reef,Xisha Islands,South China Sea was used as the test area,and the remote sensing images of Gaofen-1(GF-1)and the"Ice,Cloud and Land Elevation 2"(ICESat-2)spaceborne laser altimetry satellite data were used.Based on the optimal strategy for modeling cor-al reef geomorphological zoning,a remote sensing inversion model for coral reef area shallower than 25 m was constructed.The extrapolation of the model was applied to the coral reef profile study areas in the Xisha Is-lands with field survey data.Modeling results show that the optimal models of reef slope and deep lagoon geo-morphic zones were the band ratio binomial regression models,and they were the multi-band ratio regression models on the geomorphic zones of reef flat,shallow lagoon and patch reef.The correlation coefficient(R2)of the model inverted bathymetry data and field survey data is 0.95;The comprehensive verification by the field survey data and the ICESat-2 deep water data shows that the root mean square error of the model is 1.34 m,and the average absolute error is 1.08 m,and the average relative error is 17.10%.The research results show that the geomorphology-based active and passive remote sensing bathymetry inversion model proposed in this study significantly improves the accuracy of the medium-resolution remote sensing bathymetry inversion models in coral reef areas,and can provide method support for high-precision bathymetry retrieval in large-scale coral reef areas.

关 键 词：珊瑚礁 地貌 水深反演 高分一号 ICESat-2 

分 类 号：P237[天文地球—摄影测量与遥感]