Resilience-Focused Analysis of the United States Maritime Transportation System Using Automatic Identification System Data

Abstract

The U.S. maritime transportation system is a foundational element of the national supply chain, and its operators are under immense pressure to ensure that its ports can reliably and efficiently support supply chain demands despite their exposure to potential disruptions. Understanding the connectivity of these ports is critical to describing the robustness of port network regions in the face of disruption. In this work, Marine Cadastre Automatic Identification System data are utilized to describe a network of 62 interconnected ports (“nodes”) within the U.S. maritime transportation system. This network is analyzed with community detection via label propagation to quantitatively identify regions of the U.S. port network based on shared vessel traffic and examined with the PageRank algorithm to identify ports that are critical to regional traffic flow. The detected communities of the U.S. port network were driven by physical proximity in many instances. However, the Mississippi River and Gulf ports were split based on the predominant vessel type of each port. Additionally, ports along the Gulf often received lower PageRank scores than West coast ports of similar size. Quantitatively identifying port network regions and critical ports are valuable capabilities for discussion of regional robustness or investments to increase region-wide resilience.

US port connectivity and ramifications for maintenance of South Atlantic Division ports

“…The present study represents an effort to expand upon the port network description supplied by Young et al (2022); the current methodology follows that work closely. Like Young et al's study, this study used AIS data to quantify the vessels' transits between ports.…”

“…For example, a ship that transits from Charleston (South Carolina) to Hamburg (Germany) and then to Jacksonville (Florida) will appear in the Marine Cadastre AIS record as a port visit in Charleston followed by a multipleweek period in which the ship disappears before finally reappearing in Jacksonville. In a previous iteration of the US MTS description (Young et al 2022), this would have been treated as a direct trip from Charleston to Jacksonville, potentially inflating the importance of the Charleston-Jacksonville linkage within the overall network. The present study implemented a new methodology for inserting visits to unidentified international ports into the AIS data set prior to calculating network connectivity metrics.…”

“…To apply PageRank to a port network, the port areas were conceptualized as websites, and the vessel traffic between ports was treated as an analog for links between websites (Scully and Chambers 2019; Young et al 2022).…”

“…Metrics describing how central ports are to traffic flow across a network are referred to as centrality metrics; one of the more advanced centrality metrics is the PageRank algorithm (Page et al 1999). In the context of port maintenance, PageRank may be used along with tonnage to identify ports that are not only commercially critical but, due to their position in the network, would have an outsized impact on regional or nationwide commerce if they were to be disrupted (Young et al 2022). The present study proposes that the PageRank score could augment the traditional cargo tonnage utilized for dredge maintenance prioritization of USACE navigation channels in the approach to these critical ports.…”

“…The full description of nationwide commercial vessel traffic provided by this data set has additional benefits for understanding waterborne commerce and improving supply chain resilience. A prior characterization of the US MTS (Young et al 2022) weighted the network connections based on the raw count of vessels transiting that route, but this method may inflate the importance of routes transited by a large quantity of small vessels.…”