Article is available in full to IFST members and subscribers.

Register on the FST Journal website for free

Click the button to register to FST Journal online for free and gain access to the latest news


If you are an IFST member, please login through the Members Area of the IFST website. 













Border delays could cause congestion

Brexit - border delays

Ke Han, Daniel Graham and Washington Ochieng of Imperial College London report on their independent study, commissioned by the BBC’s Inside Out South East programme, to investigate the impact of increased check time for vehicles at the ports of Dover and Folkestone on highway congestion.

Lorry traffic passing through Dover saw a record year for freight vehicles in 2017, up 0.4% on 2016, according to figures issued by the port authority[1]. The figures show that Dover handled a record 2.6m freight vehicles in 2017, representing up to 17% of the UK’s trade in goods, worth up to an estimated £122bn last year. These statistics have fuelled concerns over delays and potential gridlock in Dover and other ports after Brexit. More checks could cause major disruption not just to traffic, but to the supply chains for goods, from food to the motor industry.

Research has been carried out by a team from Imperial College London, through Imperial Consultants, to quantify the congestion impact on the M20/ A20 of check time increase at Port of Dover and Eurotunnel (Folkestone) in a post-Brexit scenario. The study focused on a 40-mile segment of the M20/A20 motorway between Maidstone and Dover, with local access to Ashford and Folkestone.

Outbound lorries and passenger vehicles that use the ferry and tunnel to cross the Straight of Dover, as well as traffic with local origins and destinations were considered. Traffic simulations were conducted with assumptions regarding the check times at Dover and Eurotunnel for both current and post-Brexit scenarios. The impact of vehicle queuing at these locations was assessed in terms of queue length, travel time and disruption to local traffic.

Dover handled a record 2.6m freight vehicles in 2017, representing up to 17% of the UK’s trade in goods.


The study area included the M20/A20 corridor between Maidstone (east of Junction 7, near Thurnham Lane) and Dover (Port of Dover), measuring 38.2 miles in length. The road network considered for the simulation included the A20 as an alternative route between Maidstone and Folkestone, local streets near Ashford, Folkestone and Dover and part of the A2 approaching Dover from the north (a potential route shift towards the M2/A2 corridor for Dover-bound traffic).

The simulation spanned a 24- hour period when performance indicators, including queue length and travel time, were calculated hourly. The base scenario (normal traffic operation) was sampled from working days in October 2017.

The study considered both east- and west-bound traffic with local origins and destinations at Maidstone, Ashford, Eurotunnel, Folkestone and Dover. The traffic mix consisted of passenger vehicles, light goods vehicles and heavy goods vehicles. Taking into consideration different types of vehicles allowed more accurate modelling of traffic dynamics in free-flow, congested and queuing conditions, as they occupy different road space and have varying free-flow speeds.

State-of-the-art dynamic traffic assignment models and simulation platforms developed at Imperial College London were employed[2-5]. The dynamic traffic assignment model takes network characteristics (origin, destination, topology, link, node and path parameters) as well as time-varying origin-destination (OD) demand matrices as input. The algorithm assigns the OD demands within each hour to the corresponding set of feasible routes according to the dynamic user equilibrium principle[2,4,6]. The model captures the dynamic propagation of flow, queuing and vehicle spillback, while incorporating different mixes of traffic flows (e.g. passenger car, LGV, HGV).

To simulate the bottleneck effect at Dover and Eurotunnel caused by border checks, assumptions were made to calculate the corresponding bottleneck capacity (maximum number of vehicles that can go through in one hour). This allows the capture of the queue formation from these locations and their subsequent interaction with motorway and local traffic.


Under normal operational conditions at both Dover and Eurotunnel (before Brexit), the average time for a single vehicle to go through the checks is two minutes. Depending on the vehicle type, passport type and trip purpose, for which the data is quite scarce, the actual processes to go through may vary and the times taken are probabilistic. BBC Inside Out South East gathered information from the Union for Borders, Immigration and Customs and a Port of Dover study, which confirmed that the two-minute check time is a reasonable assumption for the current study.

Given that post-Brexit passport and custom checks are likely to bring extra delays at the borders, the study hypothesised that the check times at Dover and Eurotunnel are (A) two minutes on average (normal conditions), (B) three minutes on average (one minute additional delay to normal condition) and (C) four minutes on average (two minutes additional delay). These scenarios are not predictions of post-Brexit check times, as the outcome of the negotiation remains uncertain. Rather, they are hypothesised to understand the motorway congestion corresponding to different levels of border delays.


To ensure the accuracy of the simulation model, the simulation results were validated for the normal check time scenario using existing data sources. The main simulation results for the three scenarios are summarised in Table 1.

Under the three scenarios considered, queues concentrated on local streets (secondary or reactionary queues) resulting from motorway deadlock can reach up to 2.4 miles, 7.1 miles, and 10.3 miles respectively.

In addition, live traffic updates were used to validate the spatial distribution of congestion predicted by the simulation. Figure 1 compares the actual traffic queues with the simulated ones during the afternoon peak (15:00-17:00). At 15:09, the queue emerging from Dover approaches Folkestone and minor queuing occurs around Eurotunnel, which is consistent across real-world and simulated traffic. At 16:33, the queue on the A20 towards Dover persists and the queue from Eurotunnel has expanded to cover the A20 between Junctions 11A and 12 as well as M20 between Junctions 11 and 11A. The same queuing pattern is observed in the simulation.

Based on the above validation in terms of daily traffic volume and peak traffic congestion, we conclude that the simulation model under normal operational conditions provides a reasonably accurate estimation of the real-world traffic on the M20/A20.

The findings of the study show that even one or two minutes of extra check times at the borders are accompanied by a dramatic increase of congestion on the motorways as well as local streets, with queues extending up to 30 miles from Dover/ Eurotunnel towards Maidstone and travel time approaching five hours in peak times.

Table 1 Network performance indicators for different scenarios
Figure 1 Comparison of actual traffic queues (left column[7]) with those predicted by the simulation (right column)

Dr Ke Han

Senior Lecturer (Associate Professor) in Transport Operations and Logistics, Professor Daniel Graham and Professor Washington Ochieng

Centre for Transport Studies, Department of Civil and Environmental Engineering, Imperial College London





2. Han, K., Piccoli, B., Szeto, W.Y. (2016). Continuous-time link-based kinematic wave model: formulation, solution existence, and well-posedness. Transportmetrica B: Transport Dynamics, 4 (3), 187-222.

3. Han, K., Szeto, W.Y., Friesz, T.L. (2015). Formulation, existence, and computation of boundedly rational dynamic user equilibrium with fixed or endogenous user tolerance. Transportation Research Part B: Methodological, 79, 16-49.

4. Friesz, T.L., Han, K., Neto, P.A., Meimand, A., Yao, T. (2013). Dynamic user equilibrium based on a hydrodynamic model. Transportation Research Part B: Methodological, 47, 102-126.

5. Smits E., Bliemer M., van Arem B. (2011). Dynamic network loading of multiple user-classes with the link transmission model, Proceedings of the Second International Conference on Models and Technologies for ITS , Leuven, Belgium, 24th June 2011

6. Yperman, I., Logghe, S., Immers, L., 2005. The link transmission model: An efficient implementation of the kinematic wave theory in traffic networks. In: Advanced OR and AI Methods in Transportation, Proceedings of the 10th EWGT Meeting and 16th Mini-EURO Conference, Poznan, Poland. Publishing House of Poznan University of Technology, pp. 122–127.

7.Traffic England, a service from Highways England. URL:

View the latest digital issue of FS&T or browse the archive


Click here

Become a member of the Institute of Food Science and Technology