顾及尺度效应的安徽省传统村落空间异质性分析  

An Analysis of Spatial Heterogeneity in Traditional Villages in AnhuiProvince Considering Scale Effects

作　　者：殷玉忠 刘瑞 YIN Yuzhong;LIU Rui(Anhui Institute of Geo-Environment Monitoring(Anhui Geological disaster emergency technical guidance center),Hefei 230001,China;Anhui Xinhuai Electronics Co.,Ltd.,Huaibei 235000,China)

机构地区：[1]安徽省地质环境监测总站(安徽省地质灾害应急技术指导中心),安徽合肥230001 [2]安徽芯淮电子有限公司,安徽淮北235000

出　　处：《河南科技》2024年第14期93-98,共6页Henan Science and Technology

基　　金：国家自然科学基金(42071085);淮北市重大科技专项(HK2021004)。

摘　　要：【目的】分析不同尺度的影响因素对安徽省传统村落空间分布的影响。【方法】利用普通最小二乘(Ordinary Least Square,OLS)、地理加权回归(Geographically Weighted Regression,GWR)和多尺度地理加权回归(Multi-scale Geographically Weighted Regression,MGWR)并结合气候、地形、人文经济等影响因素,对安徽省传统村落分布的空间异质性进行分析,探讨影响因素的尺度效应对传统村落空间分布的影响。【结果】结果表明:(1)OLS、GWR和MGWR的R2分别为0.28、0.75和0.93,RAdjust2值分别为0.28、0.67和0.87。相较于普通最小二乘法和地理加权回归模型,MGWR的R2值分别高出0.47、0.18,RAdjust2值分别高出0.39、0.20,MGWR的拟合效果最优、对变量的解释能力最强;(2)MGWR结果显示,影响因素中气温、人口密度、GDP的带宽值最小、空间异质性最大、尺度效应显著,对传统村落空间分布有较强影响。【结论】研究结果可以为传统村落的研究与保护提供参考。[Purposes]This paper aims to analyze the impact of factors at different scales on the spatial dis-tribution of traditional villages.[Methods]Ordinary Least Square(OLS),Geographically Weighted Re-gression(GWR),and Multi-scale Geographically Weighted Regression(MGWR)were used.These mod-els were combined with factors such as climate,topography,and socio-economic aspects to analyze the spatial heterogeneity of traditional village distribution in Anhui Province and explore the scale effects of influencing factors on the spatial distribution of traditional villages.[Findings]The research results indi-cate:①The R2 values for OLS,GWR,and MGWR are 0.28,0.75,and 0.93,respectively,with correspond-ing RAdjusted2 values of 0.28,0.67,and 0.87.Compared to Ordinary Least Squares and Geographically Weighted Regression models,MGWR demonstrates significantly better performance with R2 values that are 0.47 and 0.18 higher,and RAdjusted2 values that are 0.39 and 0.20 higher.MGWR has the best fit and the strongest explanatory power for the variables.②The results of MGWR indicate that the band-width values for temperature,population density,and GDP among the influencing factors are the small-est,and the spatial heterogeneity is the greatest.The scale effects are significant,and these factors have a strong impact on the spatial distribution of traditional villages.[Conclusions]The research findings can provide valuable insights for the study and conservation of traditional villages.

关 键 词：传统村落 空间异质性 多尺度地理加权回归(MGWR) 尺度效应 

分 类 号：TU982.29[建筑科学—城市规划与设计]