基于有序Logistic回归的行道树健康评价体系^∗

Street Tree Health Assessment System Based on Ordinal Logistic Regression Analysis

摘要

【目的】构建简洁有效、实操方便的行道树健康评价体系，为城市行道树健康评价与养护管理提供决策参考。【方法】以济南市二球悬铃木、洋白蜡、国槐3种骨干行道树为研究对象，通过对指标变量的筛选分析，利用有序Logistic回归模型，分树种构建行道树特异性健康分级评价模型，建立融合地域特性、树种差异的轻量化评价体系。【结果】二球悬铃木健康评价模型为logit=2.42-1.32×病虫害+0.41×胸径年平均增长量+0.02×树冠完整度+0.01×树冠茂密度-0.62×光照，阈值为θ′₁=0.68和θ′₂=-1.42；洋白蜡健康评价模型为logit=1.93-1.18×病虫害+0.52×胸径年平均增长量+0.04×树冠完整度+0.15×树池面积，阈值为θ′₁=2.57和θ′₂=1.27；国槐健康评价模型为logit=2.15-1.25×病虫害+0.48×胸径年平均增长量+0.03×树冠完整度+0.02×树冠茂密度，阈值为θ′₁=2.45和θ′₂=0.65。3个树种评价模型均通过验证评估且准确率分别达到92.8%、94%、89.3%，并经过了初步适用性验证。【结论】首次利用有序Logistic回归模型，分树种构建了简洁有效的分级评价模型，支持在济南行道树健康分级评价中推广应用。胸径年平均增长量是行道树健康促进的核心指标，病虫害程度是影响所有树种健康的最强因子，树冠完整度是树冠结构健康的关键，树冠茂密度对二球悬铃木、国槐的健康影响较为明显；首次验证了树池面积对洋白蜡健康、光照条件对二球悬铃木健康的正向贡献。

Abstract

【Objective】This study develops a concise, effective and practical health evaluation system for street trees, aiming to provide a reference for decision-making in the health evaluation and management of urban street trees. 【Method】Taking three key street tree species in Jinan, including Platanus acerifolia, Fraxinus pennsylvanica and Styphnolobium japonicum as the research objects, the study screens and analyzes indicator variables, and constructs species-specific health grading evaluation formulas for each of the tree species using Ordinal Logistic Regression. Then a lightweight evaluation system that integrates regional characteristics and species-specific variations is established. 【Result】The health evaluation formula for P. acerifolia comprises logit=2.42-1.32×Pest&Disease+0.41×Mean Annual DBH Increment+0.02×Canopy Integrity+0.01×Canopy Density-0.62×Light Exposure, with thresholds θ′₁=0.68 and θ′₂=-1.42. The health evaluation formula for F. pennsylvanica is logit=1.93-1.18×Pest&Disease+0.52×Mean Annual DBH Increment+0.04×Canopy Integrity+0.15×Tree Pit Area, with thresholds θ′₁=2.57 and θ′₂=1.27. The health evaluation formula for S. japonicum is composed of logit=2.15-1.25×Pest&Disease+0.48×Mean Annual DBH Increment+0.03×Canopy Integrity+0.02×Canopy Density, with thresholds θ′₁=2.45 and θ′₂=0.65. All three species-specific evaluation formula are validated, achieving accuracy rates of 92.8%, 94% and 89.3%, respectively, and they have also undergone preliminary applicability verification. 【Conclusion】The ordinal logistic regression model is employed for the first time to construct a concise and effective species-specific grading evaluation formula, facilitating its widespread application in the health grading evaluation of street trees in Jinan. The Mean annual DBH increment is the core indicator for improving the health of street trees; the degree of Pest&Disease is the strongest factor affecting the health of all tree species; Canopy Integrity is the key to the healthy canopy structure; Canopy Density has a significant impact on the health of P. acerifolia and S. japonicum. The positive contributions of Tree Pit Area to the health of F. pennsylvanica and of Light Exposure to P. acerifolia are verified for the first time.

关键词

行道树 / 健康评价 / 有序Logistic回归 / 树种特异性模型

Key words

street tree / health assessment / Ordinal Logistic Regression / tree species specific model

参考文献

[1] 文超, 王春红, 黄珺, 等 . 济南市行道树现状调查及优化策略[J]. 现代园艺, 2023, 46(16): 15-18, 23.

[2] 翁殊斐, 黎彩敏, 庞瑞君. 用层次分析法构建园林树木健康评价体系[J]. 西北林学院学报, 2009, 24(1): 177-181.

[3] GARY W H, ED P, RICHARD E. Validation of a tree failure evaluation system[J]. Journal of Arboriculture, 1995, 21(5): 233-234.

[4] BOONE R, WESTWOOD R. An assessment of tree health and trace element accumulation near a coal—fired generating station, Manitoba, Canada[J]. Environmental Monitoring and Assessment, 2006, 121: 151-172.

[5] TAN X Y, SHIBATA S. Factors influencing street tree health in constrained planting spaces: evidence from Kyoto City, Japan[J]. Urban Forestry & Urban Greening, 2022, 67: 127416.

[6] IAKOVOGLOU V, THOMPSON J, BURRAS L, et al. Factors related to tree growth across urban—rural gradients in the Midwest, USA[J]. Urban Ecosystems, 2001, 5(1): 71-85.

[7] 陈文静, 曹雨秋, 张慧会, 等 . 法桐行道树生长健康评价: 以南京市热河路为例[J]. 中南林业科技大学学报, 2022, 42(8): 176-190, 201.

[8] 黎彩敏, 翁殊斐, 庞瑞君. 广州市园林绿地树木健康调查与评价[J]. 佛山科学技术学院学报(自然科学版), 2012, 30(6): 8-13.

[9] 黎彩敏, 翁殊斐, 庞瑞君. 广州市14种常用园林树木健康评价[J]. 西北林学院学报, 2010, 25(2): 203-207.

[10] 王新茹, 南歆格, 晏海, 等 . 杭州西湖环湖道路行道树调查研究[J]. 浙江农业科学, 2022, 63(1): 105-113.

[11] 李月. 西安地区行道树的健康和安全性评价[D]. 杨凌: 西北农林科技大学, 2022.

[12] 周景斌, 韩东锋, 王彦平. 杨凌主城区行道树健康评价与分析[J]. 西北林学院学报, 2015, 30(2): 278—282, 292.

[13] 邵鹏. 北京城市核心区国槐行道树健康评价研究[D]. 北京: 北京林业大学, 2020.

[14] 王琰, 余韵, 刘勇, 等 . 北京市核心区洋白蜡行道树健康评价及影响因素[J]. 浙江农林大学学报, 2022, 39(6): 1340-1349.

[15] 李娟, 韩姣, 王晓娜, 等 . 行道树树种结构及健康评价的研究[J]. 吉林林业科技, 2015, 44(1): 16-20.

[16] 吴悠绿, 肖鹏峰, 刘豪, 等 . 基于因子分析法与k—means聚类的行道树安全风险评价[J]. 西南林业大学学报(自然科学), 2023, 43(1): 137-145.

[17] 沈大刚. 安康市中心城区行道树综合评价研究[J]. 陕西农业科学, 2011, 57(3): 70-72.

[18] 陈文静, 圣倩倩, 张慧会, 等 . 南京市老城区法桐行道树健康状况研究[J]. 西北林学院学报, 2023, 38(3): 251-258.

[19] 余韵. 北京城市核心区白蜡行道树健康评价研究[D]. 北京: 北京林业大学, 2020.

[20] 樊兰英, 闫丽娟, 刘随存. 太原市不同类型城市绿地土壤理化性质[J]. 中国城市林业, 2020, 18(3): 30-33.

[21] 路嘉丽, 沈光, 王琼, 等 . 树种差异对哈尔滨市土壤理化性质影响及造林启示[J]. 植物研究, 2016, 36(4): 549-555.

[22] 朱济友, 于强, 刘亚培, 等 . 植物功能性状及其叶经济谱对城市热环境的响应[J]. 北京林业大学学报, 2018, 40(9): 72-81.

[23] 汪炳根, 卢立华. 同一立地营造不同树种林木生长与土壤理化性质变化的研究[J]. 林业科学研究, 1995(3): 334-339.

[24] 杨瑞卿, 王本耀, 彭红玲, 等 . 不同栽植基质对上海市4种行道树表型生长和根系发育的影响[J]. 林业科学研究, 2017, 30(4): 659-666.

基金资助

*山东省住房和城乡建设厅科技计划项目(2023-100)