玉田水库消落带植物群落特征及多样性研究  

Characteristics and Diversity of Plant Communities in the Yutian Reservoir Hydro-Fluctuation Belt

作　　者：夏国勋 张海娜 鲁向晖[1,2] 马露露 张利超 XIA Guo-xun;ZHANG Hai-na;LU Xiang-hui(College of Soil and Water Conservation,Nanchang Institute of Technology,Nanchang,Jiangxi 330099;Jiangxi Provincial Technology Innovation Center for Ecological Water Engineering in Poyang Lake Basin,Nan-chang,Jiangxi 330099)

机构地区：[1]南昌工程学院水土保持学院,江西南昌330099 [2]江西省鄱阳湖生态水利技术创新中心,江西南昌330099

出　　处：《安徽农业科学》2025年第5期49-55,63,共8页Journal of Anhui Agricultural Sciences

基　　金：国家自然科学基金项目(52269013);江西省水利厅科技项目(202224ZDKT17)

摘　　要：消落带水位周期性调节和人类活动导致植物群落结构变化频繁,及时准确获取消落带植物群落结构变化过程,可以为植物群落结构优化调整提供技术支持。以江西省景德镇玉田水库消落带植物群落为研究对象,无人机遥感影像为基础数据源,运用随机森林、决策树、支持向量机和最大似然方法对消落带植物进行物种分类,以实现为库区消落带植物物种遥感精细分类及植物多样性保护提供科学的参考依据。结果显示,利用随机森林方法识别库区消落带植物种类的分类精度可达86.48%,总体精度较其他分类方法提高2.50百分点~5.74百分点,Kappa系数提高0.03~0.07,且植物分类完整性强,准确性高。以野外调查和随机森林分类为基础的消落带植物群落Simpson优势度指数分别为0.61和0.70,Shannon-Wiener多样性指数分别为1.01和1.26。因此,利用无人机遥感影像及随机森林分类是分析库区消落带植物群落特征及保护植物多样性的有效方法。The cyclic adjustment of water level and human activities in the hydro-fluctuation belt lead to frequent changes in plant community structure,and timely and accurate access to the process of plant community structure change in the hydro-fluctuation belt can provide technical support for the optimisation and adjustment of plant community structure.In this paper,the plant community of Yutian Reservoir in Jing-dezhen,Jiangxi Province was taken as the object of study,and the UAV remote sensing images were used as the basic data source,and ran-dom forest,decision tree,support vector machine and maximum likelihood methods were applied to classify the species of the plants in the hydro-fluctuation belt,so as to achieve the fine classification of the plant species in the hydro-fluctuation belt in the reservoir area by remote sensing,and to provide a scientific reference for the protection of plant diversity.The results showed that the classification accuracy of identif-ying the plant species in the hydro-fluctuation belt of the reservoir area using the random forest method could reach 86.48%,with the overall accuracy increased by 2.50 percentage points-5.74 percentage points and the Kappa coefficient increased by 0.03-0.07 compared with the other classification methods,and that the plant classification had strong integrity and high accuracy.The Simpson’s dominance index of the plant communities in the hydro-fluctuation belt based on the field survey and random forest classification were 0.61 and 0.70,and the Shan-non-Wiener’s diversity index was 1.01 and 1.26,respectively.Therefore,the use of unmanned aerial remote sensing imagery and random forest classification is an effective method for analysing the characteristics of the plant communities in the hydro-fluctuation belt of the reservoir area and for conserving the plant diversity.

关 键 词：消落带 无人机遥感 随机森林分类 植物多样性 植物识别 

分 类 号：O948[理学]