摘要
不同生活型水生植物对水环境的影响和碳固持能力不同,开展大尺度范围内不同生活型水生植物的时空分布和动态变化研究,是全面掌握湖泊水生态环境变化趋势、准确核算水生生态系统碳源/碳汇的前提。以长江中下游10 km^(2)以上(共131个)的湖泊为研究对象,基于野外调查和先验知识,通过光谱分析,研发了不同生活型水生植物遥感高精度机器学习识别算法,解析了长江中下游湖泊群不同生活型水生植物的时空变化规律。研究表明,长江中下游湖泊群不同生活型水生植物遥感监测精度为0.81,Kappa系数为0.74;1986—_(2020年)长江中下游湖泊群水生植物面积为2541.58~4571.42 km^(2),占湖泊总面积的15.99%~28.77%,沉水植物是优势类型(Max1995年=2649.21 km^(2),Min_(2005年)=921.38 km^(2)),其次是挺水植物(Max_(2005年)=1779.44 km^(2),Min_(2020年)=569.05 km^(2))和浮叶植物(Max_(2015年)=685.68 km^(2),Min_(2000年)=293.04km^(2));水生植物主要分布在长江干流流域湖泊群,其次是鄱阳湖流域、洞庭湖流域、太湖流域和汉江流域;变化趋势上,1986—_(2020年)长江中下游湖泊群水生植物面积呈现先增长(1986—1995年)、后下降(1995—2010年)、再增加(2010年后)的趋势。本研究可为长江中下游湖泊群生态环境调查及水环境管理提供重要参考。
Different life-form aquatic vegetation support critical ecosystem services in most shallow lakes by improving water environment,and sequestering carbon and nutrients.Accurate spatio-temporal distribution dynamics of aquatic vegetation is basis to comprehensively obtain the changing trend of lacustrine ecological environment.In this study,we first constructed a high-precision machine learning algorithm to identify different life-form aquatic vegetation.Then we applied this approach to achieve the spatio-temporal distribution dynamics of different life-form lacustrine aquatic vegetation in the middle and lower reaches of the Yangtze River Basin.Our study showed that the overall classification accuracy was 0.81,and the Kappa coefficient was 0.74.The total area of aquatic vegetation ranged from^(2)541.58 km^(2)to 4571.42 km^(2)during 1986-2020,covering 15.99%-28.77%of the total lake area,and submerged vegetation was the dominated types(Max 1995yr=2649.21 km^(2),Min_(2005yr)=921.38 km^(2)),followed by emergent vegetation(Max_(2005yr)=1779.44 km^(2),Min_(2020yr)=569.05 km^(2))and floating vegetation(Max_(2015yr)=685.68 km^(2),Min_(2000yr)=293.04 km^(2)).Aquatic vegetations were mostly located in the lakes of mainstream of the Yangtze River,and followed by Lake Poyang Basin,Lake Dongting Basin,Lake Taihu Basin and Han River Basin.It indicated that the area of aquatic vegetation showed a trend of increase(1986-1995),then decrease(1995-2010)and increase again(2010-2020).This study could provide a scientific basis for lake ecological research and management.
作者
刘兆敏
张玉超
关保华
曹振
来莱
杨启铎
Liu Zhaomin;Zhang Yuchao;Guan Baohua;Cao Zhen;Lai Lai;Yang Qiduo(State Key Laboratory of Lake Science and Environment,Nanjing Institute of Geography and Limnology,Chinese Academy of Sciences,Nanjing 210008,P.R.China;University of Chinese Academy of Sciences,Beijing 100049,P.R.China)
出处
《湖泊科学》
EI
CAS
CSCD
北大核心
2023年第6期2022-2035,共14页
Journal of Lake Sciences
基金
国家自然科学基金项目(42141015,42171359)
中国科学院科研仪器设备研制项目(YJKYYQ20200048)
江苏省水利科技项目(2021032)联合资助。
关键词
生活型
水生植物
遥感监测
机器学习
长江中下游湖泊
Life-form
aquatic vegetation
remote sensing monitoring
machine learning
lakes of the middle and lower reaches of the Yangtze River