北京市核心区洋白蜡行道树健康评价及影响因素  

Health assessment and influencing factors of Fraxinus pennsylvanica in Beijing core area

作　　者：王琰[1] 余韵 刘勇[1] 王开勇 周晓杰 王洋 WANG Yan;YU Yun;LIU Yong;WANG Kaiyong;ZHOU Xiaojie;WANG Yang(Key Laboratory for Silviculture and Conservation of Ministry of Education,College of Forestry,Beijing Forestry University,Beijing 100083,China;Beijing Jingcaiyanyuan Landscape Technology Co.,Ltd,Beijing 101111,China;Management Office,Beijing Da’an Forest Farm,Beijing 102419,China)

机构地区：[1]北京林业大学林学院省部共建森林培育与保护教育部重点实验室,北京100083 [2]北京市京彩燕园园林科技有限公司,北京101111 [3]北京市大安山林场管理处,北京102419

出　　处：《浙江农林大学学报》2022年第6期1340-1349,共10页Journal of Zhejiang A&F University

基　　金：北京市园林绿化局项目(CEG-2018-01)。

摘　　要：【目的】针对城市树木健康评价的快速化、准确化及生产化,调查并分析北京市首都功能核心区主要绿化树种洋白蜡Fraxinus pennsylvanica的生长、健康及管护现状,为构建精炼、全面的洋白蜡行道树健康评价体系及管护技术措施提供科学依据。【方法】以北京市核心区11条主要街道的洋白蜡行道树为研究对象,采用主成分分析、K-means聚类构建健康评价模型,采用判别分析进行结果验证,将评价结果与评价指标结合建立洋白蜡BP神经网络健康预测模型;同时从生长状况、管护措施以及环境因子3个方面,分析胸径、树高、修枝留茬数量、防踩铺装等因素对洋白蜡行道树健康状况的影响。【结果】研究区域中,健康、亚健康、不健康及濒死单株分别占总数的39.20%、41.26%、16.78%和2.76%;建立的BP神经网络健康预测模型中,训练集、验证集、测试集与总集的期望值与预测值之间的相关系数分别为0.9997、0.9720、0.9976及0.9953,均大于0.9500,表明此模型能准确地反映洋白蜡行道树12个评价指标与健康评价类别的关系,可用于对同一地区其他白蜡行道树健康状况预测;方差分析和多重比较表明:胸径、修枝留茬、防踩铺装、株距、树池面积、车道数及车道走向均对洋白蜡行道树的健康状况存在显著影响(P<0.05),树高影响较小。当洋白蜡胸径为60~70 cm、树体无修枝留茬、树池铺装透水性好、株距为5~10 m、树池面积为1~2 m^(2)、道路宽度适宜且为南北走向时,洋白蜡健康状况最好。【结论】研究区域内,洋白蜡行道树整体呈亚健康状态,具有良好的保育保养潜力,部分受害严重,亟需处理。在洋白蜡行道树栽植及后期管护中,确保其修剪规范、铺装材料透水强、株距5~10 m、树池面积1~2 m^(2)以及适宜的道路情况,是保证洋白蜡行道树健康状况良好的重要措施。图3表13参26。[Objective]For the rapid,accurate,and productive assessment of urban tree health,this study aims to investigate and analyze the growth,health,and management status of Fraxinus pennsylvanica in Beijing core functional area,so as to provide a scientific basis for comprehensive health assessment and technical measures for management and protection.[Method]Taking street trees of 11 main streets in the core area of Beijing as the research object,the health assessment model of F.pennsylvanica was constructed by principal component analysis and K-means clustering.The results were verified by discriminant analysis,and the BP neural network health prediction model was established by combining the assessment results with the assessment indexes.Meanwhile,the effects of DBH,tree height,the number of pruning stubble,anti-trampling pavement etc.on the health status of F.pennsylvanica were analyzed.[Result]The results indicated that healthy,sub-healthy,unhealthy,and dying plants accounted for 39.20%,41.26%,16.78%,and 2.76%,respectively.In the BP neural network health prediction model,the correlation coefficients between expected value and predicted value of the training set,verification set,test set,and total set were 0.9997,0.9720,0.9976 and 0.9953 respectively,all greater than 0.9500,indicating that the model could accurately reflect the relationship between 12 evaluation indexes and health assessment categories of F.pennsylvanica and could be used to predict the health status of other F.pennsylvanica in the same area.Variance analysis and multiple comparisons showed that DBH,pruning stubble,anti-trampling pavement,plant spacing,tree pool area,number of lanes and lane direction all had significant effects on the health status of F.pennsylvanica(P<0.05),but tree height had little effect.The health condition of F.pennsylvanica was the best under the following conditions:60−70 cm DBH,tree body without pruning stubble,tree pool pavement with good water permeability,5−10 m tree spacing,1−2 m^(2) tree pool area,suitable road

关 键 词：洋白蜡 首都功能核心区 树木健康评价 主成分分析 BP神经网络 

分 类 号：S731[农业科学—林学]