基于Sentinel-2数据的山仔水库水华遥感监测  

Remote Sensing Monitoring of Shanzai Reservoir Algal Blooms Based on Sentinel-2 Data

作　　者：陈若薇 陈峰 陈增文 屈同 翁笑艳 陈文惠[2] 金致凡 雷少华 CHEN Ruowei;CHEN Feng;CHEN Zengwen;QU Tong;WENG Xiaoyan;CHEN Wenhui;JIN Zhifan;LEI Shaohua(School of Geographical Sciences&School of Carbon Neutrality Future Technology,Fujian Normal University,Fuzhou 350117,China;Institute of Geography,Fujian Normal University,Fuzhou 350117,China;Environmental Monitoring Center of Fujian Province,Fuzhou 350003,China;Fuzhou Environmental Monitoring Center of Fujian Province,Fuzhou 350011,China;National Key Laboratory of Water Disaster Prevention,Nanjing Hydraulic Research Institute,Nanjing 210029,China;Beijing Soil and Water Conservation Ecological Consulting Corporation,Beijing 100055,China)

机构地区：[1]福建师范大学地理科学学院、碳中和未来技术学院,福州350117 [2]福建师范大学地理研究所,福州350117 [3]福建省环境监测中心站,福州350003 [4]福建省福州环境监测中心站,福州350011 [5]南京水利科学研究院水灾害防御全国重点实验室,南京210029 [6]北京水保生态工程咨询有限公司,北京100055

出　　处：《亚热带资源与环境学报》2024年第3期179-188,共10页Journal of Subtropical Resources and Environment

基　　金：福建省环保科技计划项目(2022R001、2023R008、2024R004);国家重点研发计划(2023YFC3208903);国家自然科学基金项目(42101384);江苏省自然科学基金项目(BK20210043)。

摘　　要：山仔水库位于福建省福州市连江县,利用Sentinel-2高分辨率多光谱遥感影像,结合多种水体提取算法和藻类水华识别方法,对山仔水库2019—2023年间的水华进行了时空分析。研究结果表明:1)在水体提取方面,NDWI在研究区域内表现最佳,能够有效识别水体边界。2)在水华反演方面,通过NDVI和FAI的阈值法以及随机森林(RF)、支持向量机(SVM)、梯度提升树(GBT)等监督分类方法的对比分析,结果显示RF方法在分类准确性上表现最佳,提取精度达到了95.75%。3)山仔水库的水华爆发具有显著的季节性和年际波动特征,春季(3—4月)为主要爆发时段,秋季(9—10月)次之,而夏季和冬季的水华强度较低且分布较为分散。随着气温下降与降水量增加,水华的面积在2021—2023年间持续回落。4)在空间分布方面,水华高发区集中在水库北部狭窄水道和东南、西南部弯曲支流区域,而中部开阔水域因水体流动性强,水华频率相对较低。本研究为山仔水库的藻类水华监测提供了有效的遥感技术支持,可为水生态环境管理和水华预警提供科学依据。Shanzai Reservoir is located in Lianjiang County,Fuzhou City,Fujian Province.This study used Sentinel-2 high-resolution multispectral remote sensing images,combined with various water extraction algorithms and algal bloom identification methods,to conduct spatiotemporal analysis of algal blooms in Shanzai Reservoir from 2019 to 2023.The research results indicate that;1)in terms of water extraction,NDWI performs the best in the study area and can effectively identify water boundaries.2)In terms of algal bloom inversion,a comparative analysis was conducted using threshold methods such as NDVI and FAI,as well as supervised classification methods including Random Forest(RF),Support Vector Machine(SVM),and Gradient Boosting Tree(GBT).The results showed that the RF method performed the best in classification accuracy,with an extraction accuracy of 95.75%.3)The outbreak of algal blooms in Shanzai Reservoir has significant seasonal and interannual fluctuations,with spring(March April)being the main outbreak period,followed by autumn(September October),while the intensity and distribution of algal blooms in summer and winter are relatively low.With the decrease in temperature and increase in precipitation,the area of algal blooms will continue to decline from 2021 to 2023.4)In terms of spatial distribution,the areas with high incidence of algal blooms are concentrated in the narrow waterways in the north of the reservoir and the curved tributaries in the southeast and southwest,while the open water areas in the middle have relatively low frequency of algal blooms due to strong water mobility.This study provides effective remote sensing technology support for monitoring algal blooms in Shanzai Reservoir,which can provide scientific basis for water ecological environment management and algal bloom warning.

关 键 词：水华 监测 富营养化 水库 Sentinel-2 MSI 遥感 

分 类 号：X87[环境科学与工程—环境工程]