Spatial interdependence of fractal dimension and topological parameters of road network: a geographically weighted regression approach

Abstract

Road networks are spatial entities that display several geometric, topological variations and fractal nature. The study of the varied properties of road networks is of paramount importance in the field of urban and transport planning. This exploratory research paper aims to study the topological parameters of the road network of Tiruchirappalli city in Tamil Nadu, India based on Graph theory at different scales, i.e., by considering the road network of the study area as 65 wards, 4 zones and the city as a whole entity. The parameters considered in this study include connectivity indices, coverage indices and fractal dimension. Correlation analysis followed by regression analysis is carried out to find the relationship between the variables. The interrelationship between the deduced fractal dimension and other topological parameters of the road network are explored using analytical modelling technique alternatives like Ordinary Least Square (OLS) regression and Geographically Weighted Regression (GWR) by using Geographical Information System (GIS). Results prove that fractal dimension is an indicator of coverage and the fractal dimension of the road network in Tiruchirappalli corporation has a value 1.549 when considered as a whole. It is found that the Geographically Weighted Regression model provides a better fit compared to the ordinary least square regression model. The superiority of the GWR model over the OLS model is because it takes into account the variation of entities over a spatial region. Accessibility analysis for the 65 wards is performed using Shimbel index and it was ascertained that the central wards showed greater accessibility.