中文摘要：

伴随着城市研究新数据源与方法论的兴起,从时空间整合的角度反思和修正传统的概念与方法,成为时空行为研究重要发展趋势。地理背景作为城市研究的核心变量,其界定方式不确定性如何影响对个体行为作用效应的分析结果,已成为地理学的新方法论问题。在已有文献的基础上,本文基于西宁市的实证研究,尝试验证地理背景不确定性对行为模式分析的影响,并进一步探讨多个行为变量与其相互关系以及时间维度差异。通过比较两类地理背景范围（仅考虑居住地建成环境的地理背景范围、综合考虑居住地与家外活动地建成环境的地理背景范围）,分析对个体行为作用效应结果的差异性。研究发现：如果仅以居住地作为地理背景范围,关注其中建成环境对行为影响时,可能会夸大＂家＂的作用,而综合考虑居住地与活动地的地理背景范围时,更为接近于＂真实＂的地理背景,建成环境对家外活动时长、出行时长和活动空间有一定的解释力,且能够捕捉到一些重要的影响因素并修正有悖于常识的结论。同时地理背景不确定性对个体行为作用效应的分析结果影响还表现出较明显的工作日/休息日差异,表明地理背景存在时间维度的不确定性。研究结论在一定程度上揭示了时空行为研究在重新审视传统城市地理学概念、探索日常生活视角的度量方法与研究范式方面的有效性。

英文摘要：

The emerging technologies for collection and analysis of massive spatiotemporal databases have largely benefitted the innovation of urban studies, and enabled us to revisit the traditional constructs in terms of space-time behavior. On specific case is the uncertain geographic context problem（UGCo P） was proposed by Kwan（2012）, which refers to the spatial uncertainty in the actual areas that exert contextual influences on the individuals being studied.This research makes an attempt to address the multivariable analysis as well as the temporal variations in the research of UGCo P.Two inter-related questions are explored in this research：First,how the contextual effects on different aspects of individual behaviors（and their inter-relationships）may be affected by the delineation of geographic context.Second,how the uncertainty of geographic context is evident in terms of the day-to-day variations.This exploration is based on an empirical study of activity-travel behavior in Xining,China.An activity-diary dataset collected in 2013 and including 1098 valid samples is applied.The research first delineates the contextual unit of the built environment variables in two different ways：one is merely based on the residential neighborhoods,and the other considers both residential neighborhoods and activity spaces.Then,the research testifies the hypothesis about the potential impact of uncertain geographic context on the space-time behavior analysis.Also,the results on the weekday and weekend are compared in order to testify the day-to-day variation.Based on ANOVA test,geo-visualization,and the structural equation modelling,the study yields several findings.First of all,the geographic context based on residential neighborhood may lead to misunderstanding results,such as overestimating the contextual effects of neighborhood areas,and underestimating those of other activity places.Secondly,the effects of geographic context on individual behaviors show significant differences between the weekday and weekend.These