Ports in a World Maritime System: A Multilevel Analysis - Methods for Multilevel Analysis and Visualisation of Geographical Networks - page 116

and 2006. Based on these preliminary results, the third section provides a multilevel
analysis based on the application of strength clustering. The conclusion presents
implications for policy decisions and some possibilities for further research.

8.2 The Network Analysis of Maritime Transport

8.2.1 A Lack of Network Analysis on Maritime Transport

Transport geographers have faced many obstacles in the study of maritime networks.
One obstacle is contextual. Recent technological innovations in maritime transport
that displace port functions from urban areas have eroded the importance of
maritime transport in public representations. While these circumstances have likely
reinforced the influence of the central place theory as stated by Bird ( 1970 ), they
have paradoxically renewed scholars' interest in the transformation of sea transport
itself, notably through its impact on port development ( Slack , 1993 ). Shipping
networks have become more footloose, partly due to shrinking transport costs
and reduced trade barriers in general ( Clark, Dollar, & Micco , 2004 )butalso
due to the capacity of shipping lines to reorganize their networks globally with
increased bargaining power and to integrate land transport through vertical integra-
tion ( Robinson , 2002 ). However, the spatial complexity realized by such changes
worldwide seems to have discouraged scholars from mapping the new organization
of sea transport. Instead, port and maritime specialists have concentrated their
efforts on the changes occurring within ports ( Slack & Frémont , 2005 ), between
ports of a given region ( Ducruet, Roussin, & Jo , 2009 ), and across port hinterlands
( Notteboom & Rodrigue , 2005 ), where various actors intervene in a territorial
context that is more visible than across the oceans. In addition, containerization has
permitted the integration of maritime transport with land-based logistics through the
actions of giant companies ( Robinson , 2002 ), while the land leg has remained the
most expensive portion and has witnessed the most concentrated scholarly efforts.
Therefore, the dynamics occurring across the maritime space between ports are not
well known, although the current evolution may lead us toward an “ultimate system
of maritime transportation [ ] whereby every port node can theoretically be linked

to every other port node” ( Bird , 1984 , p. 26).
A second obstacle in the current study of maritime networks is economic and
technical. Detailed data on maritime traffic between ports are often difficult to obtain
because of their rarity and their very high cost. Thus, most port and maritime spe-
cialists use aggregated measures of port throughput for comparative analysis (e.g.,
traffic volume, growth, and concentration) because these are simplified and easily
accessible measures that indirectly analyze the spatial changes in sea transport such
as the impact of port selection and the hub-and-spoke strategies of ocean carriers.
Some researchers have used data from meteorological offices to obtain a snapshot of
the worldwide location

...

of sea-going vessels ( Brocard, Joly, & Steck , 1995 ), while