【摘要】 随着因健康食品获取不足而爆发肥胖症、糖尿病等慢性疾病的地区越来越多,美国等国家已经将"食品沙漠"问题上升到国家战略。在大数据背景的数据获取驱使下,"食品沙漠"研究由地块或街区等行政区划单元尺度转变到街道网络空间尺度下的食品可达性分析,但该研究将分析尺度固定在街道网络的弧段尺度,不利于街道网络事件的多尺度、多因素分析。考虑到上述问题,本文基于线性剖分模型,将街道网络弧段剖分成同质的线性单元,综合考虑加拿大埃德蒙顿地区精确定位的人口分布及出行方式,基于网络核密度估计方法计算该地区的街道网络空间食品可达性,并探测及评估"食品沙漠"现象分布。该地区所呈现的"中心—外围"、"多中心"的食品可达性分布结果表明其食品获取环境以及区域发展均存在失衡现象,"食品沙漠"现象集中分布在城市外围区域人口密集处。该模型精确界定了"食品沙漠"的分布范围,提供了一种可以灵活综合街道网络空间下的多因素约束的框架,同时也为街道网络空间下的城市研究提供了新思路。更多还原

【Abstract】 Since the incidence of obesity,diabetes and other chronic diseases becomes higher and higher due to the lack of healthy food in more and more regions,addressing the "food desert" problem has risen to the national strategy in the United States and other countries.However,due to the limitations in data and computation,traditional food desert research analyzes food accessibility at an aggregate scale,such as in municipalities and neighborhoods,ignoring the fact that human activities are in the street network framework.Driven by data acquisition in the big data movement,more and more researchers try to do urban research in the street network space.And the analysis scale of these studies is fixed on the network arc scale,failing to analyze issues in multi-scale and multi-factor.To solve these problems,we propose a multi-scale analysis method under the linear tessellation model.In this method,the street network,regarded as a linear reference space,is divided into homogeneous linear units.Considering the precise population distribution and the travel constraints of Edmonton,Canada,the food accessibility is assessed by the network kernel density estimation under the reference space.Then,the distribution of the "food desert" phenomenon can be discovered and evaluated.The experimental results show that the food accessibility of this region presents a core-periphery and multi-centered distribution pattern,which indicates that the food environment is unequal.And the relationship between food accessibility and regional planning reveals that disadvantaged areas are more susceptible to "food desert",especially in suburb regions with dense populations.This model is competent in detecting "food desert" and provides a framework which can integrate the multi-factor constraints in the street network space.Under this framework,we can precisely define the distribution range of "food desert",evaluate the severity of this phenomenon by food demand and supply.Consequently,the model provides the decision-making reference for the alleviation and resolution of "food desert",and provides a new idea for the urban research under the street network space.更多还原