Analysis of Spatial Organization in Urban Network Based on Air Flows of People: Empirical Evidence for Iran

Introduction
Studying the spatial structure of urban systems and the inter-urban relationships in particular,
has occupied a special place in the context of urban and regional studies for a long time
(Meijers, 2007, 245). Since the late 1960s and the emergence of a system approach an urban
system is not only a set of physical instances- in this case urban nodes- but rather it includes the
spatial interurban linkages (Simmons, 1978) and is defined through relations and flows among
cities and their position in the outer complex network. Therefore, a change in any defining
components of an urban system (including nodes and linkages) can be associated with a change
in its spatial constellation including spatial structure and organization. In other words, increase
in accessibility between places are more a function of cross-border networks and are
increasingly determined by flows within system (Derudder and Witlox, 2005; Neal, 2010) rather
than what fixed within them (Smith, 2003).
As most of empirical evidences have been done so far was based on location attributes
derived by size-based approach (like population), a little attention has been paid to connections
between urban nodes. Apart from these short comings, few attempts are made to propose a 
methodology for assessment and evaluation of an urban system configuration based on the
analysis of exchanges flows. Therefore, in one hand, this article attempts to present a more
complete and advanced methodology of the study on the nature of the spatial constellation in
urban systems to advance this research field. This helps identification of different dimensions of
their configuration as derived from an interaction approach. On the other hand, it attempts to
identify urban spatial constellation by providing empirical evidence from Air Passenger Flows
(APF) in 2006 and compare this with the resulted urban hierarchy based on population in the
same year.
Methodology
In this study, data were used from long-distance personal travel. With regard to the purpose of
this research and within its defined theoretical framework, only O-D data were applied. The
required O-D data for APF were obtained from the statistical year book of Air Transport of Iran
published by the Iran Airlines Organization (IAO) in 2006. According to this report, there were
59 domestic airports in Iran in 2006, with the nearest city to each airport utilized as the relevant
transport network (city) node in the calculations (Statistical Annals of Aerial Transportation of
Iran, 2006).
The spatial scale of this study is national, but in respect of mentioned limitations in
providing geographical information on the origin and destination of long-distance travels, it is
preferable and plausible to use political division of provinces as delimitation of the study area.
For this reason, we employ thirty provincial main urban areas in 2006 as spatial units of
analysis. As a result, the spatial level has been restricted to inter urban relations through flows
of people.
The five dimensions of an urban systems spatial configuration investigated are including: 1.
centrality and the dominance of the vertices, 2. network cohesion, 3. network strength, 4.
network symmetry, and 5. communities and levels, which are described systematically and
made measurable due to the values of the indices. The position of the spatial configuration,
regardless of spatial scale or the type of flow, ranges across a continuum from purely
monocentric to completely networked.
Results and Discussion
In APF the total number of movement of passengers via inner aerial public transportation fleet
in 2006 was 12, 225, 183 According to the first dimension, centrality and power, out of 400
possible centrality points, Tehran had value of 271. Mashhad with a centrality of 166.17, which
is substantially less than that for Tehran, was ranked at second. Consequently, it is concluded
that the APF is topologically monocentric. The maximum value for network cohesion can be
700 and the APF gains only 457.70 scores and this network with low cohesion differs
significantly from a networked constellation and has greater similarity to mono-polar structures.
The maximum value of the third dimension, network strength, can be as high as 300 and the
APF is a more concentrated network with strength of 169.13. The range of the obtained values
for the fourth dimension, network symmetry, can be from 0 to 200 and according to obtained 
value (193.43), it can be claimed that the air flows of people is symmetrical. The results from
the fifth dimension, community and levels, shows that for APF, which has more in common
with the mono-centric model, because of limitation of city organization in superior levels, only
two communities of vertices and three levels were detectable.
Conclusion
The empirical findings expresses that Tehran is in the first rank and then Mashhad as having
significant differences in comparison with other cities at lower spatial levels in that it is the
prime city of network; hence the spatial constellation of the network is considerably different
from a polycentric. In regard to the second dimension, network cohesion, APF represents low
level of cohesion. This occurred because for APF, Mashhad, Bandar Abbas, Ahvaz and Shiraz
are all located at further geographical distances from the network central city (Tehran) and yet
are still able to achieve efficient interactions with Tehran. The third dimension, network
strength shows that APF is a concentrated network, so it is best described by a monocentric
constellation structure with regard to both topology and weight. According to the results of
network symmetry dimension, the flows of people are symmetrical nationally and annually. In
the fifth dimension, one of the reasons that clusters are not formed at this level is the absence of
a central city at the median level that can attract significant passenger flows. The expansion of
areal flows in long-distance travels does not allow the formation of median cities close to their
peripheral cities where are accessible by other terrestrial modes of transportation. Thus, urban
hierarchy derived by size-based approach because of observing distinguished dimensions is
different from network-based one.