Identifying the Impacts of Land-Use Spatial Patterns on Street-Network Accessibility Using Geospatial Methods

While the land use-street network nexus is well acknowledged, evidence for the one-way impacts of land-use patterns on street accessibility is still inadequate. The measurements of land-use patterns and street accessibility lack systematic knowledge. Their empirical correlations also lack geographical variability, constraining site-specific land-use practices. Therefore, this study overcame the aforementioned limitations by examining the two-level spatial models to formulate accessibility-oriented land plans, using a well-developed Chinese city as an example. Firstly, two landscape metrics—Euclidean Nearest-Neighbor Distance (ENN) and Similarity Index (SIMI)—were used to quantify the intra- and inter-land-use configurations, respectively. Both city-level and local accessibility were measured using spatial design network analysis. Performing both ordinary least squares (OLS) and geographically weighted regression (GWR) models, results identified the statistically significant effects of inter-land-use patterns on two-level street accessibility. An exception was that land-use configurations within residential and industrial regions were irrelevant to street accessibility. We also found GWR was a better-fitting model than OLS when estimating locally-varied accessibility, suggesting hierarchical multiscale land-use planning. Overall, locally heterogeneous evidence in this study can substantialize land use-street network interactions and support the decision-making and implementation of place-specific accessibility-oriented land use.