FUNCTIONAL REGIONS BASED ON INTERREGIONAL AUTOMOBILE FLOWS IN THE NAGOYA METROPOLITAN AREA

This paper examines the problem of measuring the regional pattern of automobile flows within the Nagoya Metropolitan Area in order to reveal the basic structure of the functional regions in this area. The close relationship between movement patterns and the location of socio-economic activities would be assumed to hold within urban areas, in particular within large metropolitan areas. However, at present the measurement of the interdependence involved creates numerous conceptual, empirical and technical problems. In addition to this, it is very difficult to collect data on the flows of person, commodity and various kinds of information such as the commuting movement or telephone call and though detailed information can be collected on the distribution of activities, an appropriate breakdown of movement patterns would require extensive surveys because of their extreme complexity.<br> Most previous studies have examined several kinds of commodity flows assuming the hierarchy of regions or considering only the largest flow in origin/destination flow matrix. These have been apt to lead to difficulties in determining the underlying self-contained functional regions indicated by the flow system within the metropolitan area. An analysis considering only the largest flow is likely to show that this metropolitan area is dominated almost exclusively by Nagoya, but it may possibly assumed that there exist several functional regions with strong internal linkages but with weaker connections to other sub-systems in the metropolitan area. In this connection the method through which the analysis here has been carried out has proved to be useful for investigating the structure of complex flow systems so as to specify the principal groups of observations and the linkages they have in common.<br> In this study the data on automobile flows which are expressed in the form of 59 × 59 origin/destination automobile flow matrix are used as a single indicator of the basic structure of functional regions within the metropolitan area. In the first stage of the analysis, each column or destination vector in the matrix mentioned above is correlated with every other, yielding a 59 × 59 matrix of coefficients. Linking zones which have highly significant similarities in the way, they assemble their incoming trips suggests that most of the zones in the center of the metropolitan area are very similar to each other and that a neighborhood influence is at work in the system. A set of data which are assumed to contain several different variables can be divided objectively into meaningful groups by the factor analysis. So the next stage is the application of the factor analysis to the correlation matrix. Varimax method which is one of the procedure rotation of the factor analysis is used here to facilitate the interpretation of the latent factors. Geographically interpreted, the high factor loading extracted by this analysis may show groups of zones receiving trips from common origins, and the high factor scores specify these origins. In this study the first twelve factors each with more than 1. 69 percent of the total variance are thought to correspond to the functional sub-systems and only the factor loadings and factor scores greater than 0.2 and 0.1 respectively are adopted as the relevant indices.<br>Groups of origins and destinations corresponding with the twelve factors, when linked together, reveal twelve functional regions as illustrated in Figs. 5_??_8. Factor 1 accounting for 27. 54 percent of the total variance, represents the dominant nodal sub-system, centered in the heart of this metropolitan area, including Nagoya and its surrounding regions. In previous studies, it has not been easy to extract this main sub-system in such a compact form because of difficulties in distinguishing the dominant influence of Nagoya from that of other cities under its influence.