Cruise ships and road congestion in South America

The question

To what extent does the presence of cruise ships in port increase road congestion in the urban areas near the terminals?

The method

AIS data on calls (over 10 million observations) and traffic data from a crowd-sourced navigation platform (about 68 million observations) for the whole of 2019, within 2 km of the terminal in five cities; total delay estimated for the 8-13 and 15-20 time slots (3,960 observations). ANCOVA and fixed-effects panel regression with autoregressive error.

What the study finds

+12%congestion within 2 km, per shipAccording to the econometric model, each cruise ship arriving or departing is associated on average with roughly a 12% rise in congestion within 2 km of the terminal (2019 data).

+19%congestion with at least one ship in portLooking only at whether at least one ship is in port, congestion is on average 19% higher; the biggest rise is in Cartagena, where delay grows by almost 50%.

91%of ships berthing before noonArrivals cluster in the morning: 91% of ships berth before noon and about two thirds of arrivals happen in the early morning.

81%of ships leaving after 3 pmDepartures happen mostly in the afternoon: 81% of ships leave after 3 pm.

In Callao there is no significant difference in congestion between times with and without ships; the authors suggest this may reflect the terminal's location, far from central Lima (about 15 km).
In 2019 the researchers identified 2,120 calls across the five ports; over 30% were in Cartagena, which had at least one ship present in 56% of the time slots considered.

Why it matters for ports and cities

Using large-scale data, the study measures an effect often only perceived: road congestion near terminals when ships arrive and leave. The authors recommend pairing investment in berths and terminals with traffic management, better public transport to attractions and coordination with the port so that peak disembarkation does not coincide with rush hour.

Limitations

The number of passengers per ship is missing (tonnage is not significant) and the number of vehicles is estimated using modelling assumptions; the analysis covers only 2019 and a 2 km radius, and the main model has a low R² except in the specification including accidents and road closures.

Full reference and source

Calatayud, A., Sánchez González, S., Marquez, J. M. (2022). Using big data to estimate the impact of cruise activity on congestion in port cities. Maritime Economics & Logistics (2022) 24:566–583 (Springer Nature).

Read the original study ↗

Summary written by the ODYBOAT Observatory in our own words. Tables and figures belong to the authors.