生物天敌暴发导致珊瑚礁退化的高分遥感监测与分析——以南海太平岛为例  

High Resolution Remote Sensing Monitoring and Analysis of Coral Reef Degradation Caused by Outbreaks of Biological Natural Enemies:A Case Study of the Taiping Island in the South China Sea

作　　者：郑金辉 任广波 胡亚斌 张飞飞 马毅 李明杰 王瑞富[1] Jinhui Zheng;Guangbo Ren;Yabin Hu;Feifei Zhang;Yi Ma;Mingjie Li;Ruifu Wang(College of Geodesy and Geomatics,ShanDong University of Science and Technology,Qingdao 266590,China;Key Laboratory of Marine Environment Detection Technology and Application,Ministry of Natural Resources,Guangzhou 510300,China;First Institute of Oceanography,Ministry of Natural Resources,Qingdao 266061,China;Technology Innovation Center for Ocean Telemetry,Ministry of Natural Resources,Qingdao 266061,China;South China Sea Development Research Institute,Ministry of Natural Resources,Guangzhou 510310,China;Nansha Islands Coral Reef Ecosystem National Observation and Research Station,Guangzhou 510399,China)

机构地区：[1]山东科技大学测绘与空间信息学院,山东青岛266590 [2]自然资源部海洋环境探测技术与应用重点实验室,广州510300 [3]自然资源部第一海洋研究所,山东青岛266061 [4]自然资源部海洋遥测技术创新中心,山东青岛266061 [5]自然资源部南海发展研究院,广州510310 [6]海南南沙珊瑚礁生态系统国家野外科学观测研究站,广州510399

出　　处：《热带地理》2023年第10期1856-1873,共18页Tropical Geography

基　　金：自然资源部海洋环境探测技术与应用重点实验室开放基金课题(MESTA-2022-B008);中国高分辨率对地观测专项项目(41-Y30F07-9001-20/22);海洋领域融合应用示范项目(RHYJKF02)。

摘　　要：珊瑚生物天敌黑皮海绵、长棘海星在世界范围内暴发严重威胁珊瑚礁生态健康,然而目前珊瑚礁领域尚缺乏生物天敌暴发后对珊瑚礁地貌类型影响的定量研究。文章利用太平岛珊瑚生物天敌黑皮海绵、长棘海星2次暴发事件前后,覆盖太平岛2016—2022年的26期Sentinel-2遥感影像,结合高分辨率GF-2(PMS)遥感影像和卫星遥感影像地理信息系统Google Earth平台中的数据,开展中国南海太平岛珊瑚礁地貌类型遥感影像分类实验,对密集珊瑚沉积区、稀疏珊瑚沉积区、珊瑚丛生区、沙坪、浅礁前斜坡等珊瑚礁地貌类型演变进行特征分析。结果表明:1)结合专家解译知识和支持向量机(Support Vector Machine,SVM)分类算法开展的太平岛珊瑚礁地貌类型分类,最高总体精度和Kappa系数分别为96.46%和0.94。2)在2种珊瑚生物天敌暴发期间,太平岛的珊瑚礁丛生区、密集珊瑚沉积区、稀疏珊瑚沉积区等珊瑚礁地貌类型面积有显著下降;黑皮海绵暴发后对密集珊瑚沉积区影响最大,其面积减少72.92%;长棘海星暴发后对珊瑚丛生区影响最大,其面积减少59.17%。3)2016—2022年,太平岛珊瑚礁退化率高于恢复率,其中2017年3—6月珊瑚礁退化率最高,为23.88%;在2017年6—9月珊瑚礁恢复率最高,为18.03%。The global proliferation of natural predators,including Terpios hoshinota sponge and crown-of-thorns starfish,poses a significant threat to the ecological vitality of coral reefs.However,coral reef research currently lacks comprehensive quantitative studies on the ramifications of predator outbreaks in various coral reef landform types.To bridge this knowledge gap,this study utilized a dataset comprising 26 Sentinel-2 remote sensing images spanning the years 2016 to 2022,both pre-and post-two distinct episodes of natural predator outbreaks involving the Terpios hoshinota sponge and crown-of-thorns starfish,within the vicinity of Taiping Island.High-resolution GF-2(PMS)remote-sensing images and satellite data were used to augment the analysis.Based on data obtained from the Google Earth platform of the Geographic Information System for remote sensing imagery,we conducted an experiment involving the classification of coral reef landform types in Taiping Island in the South China Sea.During the Support Vector Machine(SVM)classification experiments,we performed a meticulous examination of the image interpretation outcomes,making comparisons and rectifications in consultation with domain experts.Classification was performed by analyzing remote sensing images,and the identified landform types included areas characterized by dense coral sedimentation,sparse coral sedimentation,coral clusters,sand flats,and shallow reef front slopes.Subsequent analyses focused on discerning distinctive characteristics and tracing their evolutionary patterns.The findings reveal that(1)this study effectively integrates expert interpretation knowledge and an SVM classification algorithm to classify the various landform types present in Taiping Island's coral reef ecosystem.Remarkably,the classification achieved an unprecedented level of accuracy,with an overall accuracy rate of 96.46%and a kappa coefficient of 0.94.(2)The analysis showed that during the two natural predator outbreaks,the coral reef landforms on Taiping Island,which included

关 键 词：珊瑚礁 黑皮海绵 长棘海星 高分遥感 太平岛 

分 类 号：P715.7[天文地球—海洋科学] P737.2