摘要
【目的】广安市古树名木众多。研究古树名木地理分布特征、树龄与生长环境及生态因子间的关系,对古树名木保护具有重要意义。【方法】利用标准差椭圆了解古树名木分布特征,通过地理加权回归模型(GWR)和多元线性回归模型(MLR)模拟树高、胸围、平均冠幅、海拔和坡度对树龄的回归强度。【结果】①广安市古树名木沿水系、山脉、交通线呈线状分布;政府驻地、红色旅游区向外扩散呈圈层结构;乡村多于城市,平地占主导;正常株多于衰弱株,生长环境适中;高海拔区多于低海拔区,垂直差异明显。②地理加权回归模型优于普通最小二乘法模型(OLS),平均冠幅、胸围、树高是影响树龄的关键因素,坡度对树龄影响较小,海拔与树龄呈负相关关系。③多元线性回归模型相关系数比地理加权回归模型高0.297,各解释变量与回归变量的系数强度同地理加权回归模型高度一致,且对300 a以下的古树树龄估算精度较高。【结论】标准差椭圆可定量分析古树名木的空间分布特征,地理加权回归模型和多元线性回归模型可准确估算古树树龄。
[Objective]To better protect and preserve the ancient and famous trees in Guang’an City,this study is focused on the geographical distribution characteristics of them and the relationship between tree age,growth environment and ecological factors,which is of vital importance.[Method]First,the standard deviation ellipse was used to summarize the distribution characteristics of ancient and famous trees.Then,the geographically weighted regression model(GWR)and multivariable linear regression model(MLR)were employed to simulate the regression intensity of tree height,chest circumference,average crown width,altitude and slope to tree age.[Result](1)The ancient and famous trees in Guang’an are distributed in a linear pattern along the water system,mountains and traffic lines while in a circular structure along the government residence and the red tourist destinations;more are distributed in the countryside than in the city with flat land as the dominant habitat;there are more normal plants than weak ones with a moderately favorable growing environment;more are distributed in the high-altitude areas than in the low-altitude areas with a significant vertical difference.(2)GWR works better than the ordinary least squares model(OLS);the average crown width,chest circumference and tree height are the key factors that affect tree age;the slope has little effect on tree age and altitude has a negative correlation with tree age.(3)The correlation coefficient of MLR is 0.297 higher than that of GWR.The coefficient intensity of each explanatory variable and regression variable is highly consistent with that of GWR,and the accuracy of estimating the age of ancient trees under 300 a is higher.[Conclusion]With the employment of standard deviation ellipse,geographically weighted regression model and multivariable linear regression model,the distribution characteristics of ancient and famous trees are better summarized,providing decision-making basis for the estimation of the age of ancient trees and the protection of ancient and famous trees.
作者
张艳丽
杨家军
ZHANG Yanli;YANG Jiajun(Guang’an Bureau of Forestry,Guang’an 638500,Sichuan,China;Guang’an District Bureau of Natural Resources and Planning,Guang’an 638550,Sichuan,China)
出处
《浙江农林大学学报》
CAS
CSCD
北大核心
2020年第5期841-848,共8页
Journal of Zhejiang A&F University
基金
四川省科技计划资助项目(18YYJC0211)。
关键词
树木资源调查
古树名木
树龄
空间分布
地理加权回归模型
多元线性回归模型
survey of tree resource
ancient and famous trees
age
spatial distribution
geographically weighted regression model(GWR)
multivariable linear regression model(MLR)
