Adaptive rerouting reshapes impacts of maritime chokepoint disruptions

This study utilizes a large-scale agent-based model to demonstrate that adaptive rerouting in response to maritime chokepoint disruptions creates complex, dynamic loss patterns characterized by delayed vessel cycles and uneven regional impacts, proving that the true economic risk is driven by routing behavior and timing rather than static network topology.

The Gist

Imagine the global shipping network as a massive, bustling highway system where thousands of trucks (ships) are constantly moving goods between cities (ports). Most of these trucks funnel through a few critical bridges and tunnels (chokepoints) like the Suez Canal, the Panama Canal, and the Strait of Malacca.

This paper builds a giant, high-tech simulation of this entire highway system to answer a simple question: What happens when one of these critical bridges gets blocked, and how do the truck drivers react?

Here is the breakdown of their findings using everyday analogies:

1. The "Smart Driver" vs. The "Static Map"

Previous studies often looked at shipping like a static map: "If Bridge A is closed, 100 trucks are stuck." They assumed the trucks would just sit there or stop moving.

This paper introduces a smart simulation where every single ship is an "agent" with a brain. When a bridge closes, these ships don't just freeze; they look at their GPS, calculate a new route, and drive around the blockage.

• The Analogy: Imagine a traffic jam on a main highway. A static map says, "Everyone is stuck." A smart driver says, "I'll take the back roads." The paper shows that while taking the back roads helps the individual driver, it creates new problems for the whole system.

2. The Two Types of Pain: The "Spike" and the "Drip"

The researchers found that disruptions cause two different kinds of damage, and they behave differently:

• The "Spike" (Peak Shortfall): This is the immediate chaos when the bridge first closes. Ships scramble to find new routes. This causes a sudden, sharp drop in deliveries on the worst day.
  • The Finding: Once the ships figure out the detour, this "spike" doesn't get much worse, even if the bridge stays closed for a month. The system adapts quickly to the new reality.
• The "Drip" (Cumulative Loss): This is the slow, steady leak of time. Because the new detour routes are longer, ships take more days to complete their trips.
  • The Finding: Every single day the bridge stays closed, the ships are stuck on these longer routes a little bit longer. This creates a "drip" of lost time that adds up. Even after the ships have adapted, they are still losing efficiency. If the bridge is closed for 50 days, the total loss is roughly 50 times the daily "drip."

3. The "Domino Effect" of Delays

The paper highlights that the damage isn't just where the bridge is closed.

• The Analogy: Imagine a delivery truck that takes a detour. It arrives at the first city late. Because it's late, it misses its scheduled departure for the next city. That delay ripples down the line, causing missed deliveries in cities thousands of miles away that had nothing to do with the original bridge.
• The Finding: While rerouting saves some ports from immediate disaster, it often shifts the pain to later stops in the journey or to different regions entirely. The "loss" spreads out like a ripple in a pond.

4. The Power of Knowing the "End Time"

One of the most interesting findings is about information. The researchers tested what happens if the ship captains know when the bridge will reopen.

• The "Wait vs. Detour" Dilemma: If a captain knows the bridge will be open in 2 days, they might just wait at the entrance. If they know it will be closed for 20 days, they take the long detour immediately.
• The Finding: Knowing the end date is the most valuable piece of information. It prevents ships from making bad decisions, like taking a long detour for a short closure or waiting too long for a long closure.
  • The Result: If ships know when the bridge will reopen, they avoid "avoidable losses." They don't waste time on unnecessary detours or waiting in line. It doesn't fix the long routes, but it stops the system from panicking and making inefficient choices.

5. The "Unavoidable" Bridge (Strait of Hormuz)

The paper also looked at a bridge that cannot be detoured around (like the Strait of Hormuz, which is the only way out for oil tankers in that region).

• The Finding: Here, the rules change. Because there is no "back road," the losses don't just add up linearly; they explode. If you block a route with no alternative, the system breaks down much faster and harder than if you block a route where ships can go around.

Summary

The paper concludes that the risk of a maritime chokepoint isn't just about how many ships usually go through it (the static map). It is a dynamic problem of timing and behavior.

• Rerouting helps but doesn't fix everything; it just moves the delay.
• Long closures hurt more because of the accumulated "drip" of lost time, not just the initial shock.
• Knowing the end date is crucial for minimizing waste.

In short: When a global shipping bridge breaks, the ships are smart enough to find a new way, but that new way is slower, and the delay keeps piling up until the bridge is fixed. Knowing when it will be fixed helps everyone drive more efficiently.

Technical Summary

Technical Summary: Adaptive Rerouting Reshapes Impacts of Maritime Chokepoint Disruptions

Problem Statement
Maritime chokepoints (e.g., Suez Canal, Panama Canal, Strait of Malacca) concentrate global shipping traffic, making them critical nodes in the global supply chain. While previous research has extensively mapped the structural vulnerability of these networks, it has largely treated the maritime system as static or relied on simplified routing assumptions. Consequently, existing models fail to capture the adaptive behavior of the shipping fleet—specifically, how individual vessels dynamically reroute, reschedule, and interact following a network shock. This gap limits the ability to predict the true system-wide impacts of disruptions, as topology alone cannot account for how rerouting mitigates direct losses while potentially propagating delays to downstream ports and dependent regions.

Methodology
The authors developed a full-scale, empirically calibrated agent-based model (ABM) of the global commercial shipping fleet to simulate adaptive responses to chokepoint disruptions.

• Scale and Scope: The model represents 35,954 active ships moving among 1,651 ports, covering cargo, dry bulk, and tanker vessels.
• Behavioral Core: Ship routing decisions are driven by a two-step Markov chain model, calibrated on historical port call sequences (2019–2025). This allows ships to choose destinations based on their two most recent ports, preserving empirical routing regularities while enabling adaptation.
• Simulation Mechanics:
  • Network: A weighted marine network based on SeaRoute, with edge weights representing traversal time.
  • Adaptation: Ships recompute routes using the A* algorithm immediately upon network changes (chokepoint closure) and periodically (every ~20 days) while en route.
  • Port Operations: Ports have shared queues and service capacities calibrated from historical data (90th percentile of daily ship presence). Delays occur when demand exceeds capacity.
  • Scenarios: The study simulates closures of the Suez, Panama, and Malacca chokepoints individually and simultaneously, varying durations from 1 to 50 days.
  • Information Regimes: The model tests four scenarios regarding ship knowledge of the disruption: (1) no warning/no forecast, (2) warning only, (3) reopening forecast only, and (4) full-window information (start and end dates known).
• Validation: The model was validated against observed mean daily port calls (Pearson correlations $r > 0.95$) and the 2021 Ever Given Suez Canal blockage, showing strong correlation ($r=0.65$) in relative arrival changes.

Key Contributions

1. Dynamic vs. Static Modeling: The paper introduces a one-to-one agent-based approach that moves beyond static structural exposure metrics. It demonstrates that realized losses differ significantly from static predictions because ships actively reroute, altering the spatial and temporal distribution of delays.
2. Decoupling Peak vs. Cumulative Loss: The study distinguishes between maximum arrival shortfall (the worst single day of disruption) and net cumulative shipping-day losses (the total deficit over time). It reveals that while peak shortfalls are often limited and occur early, cumulative losses grow persistently as ships remain on longer alternative routes.
3. Information Value: The research quantifies the operational value of information, specifically showing that knowledge of a disruption's end date is more critical than the start date for minimizing cumulative losses, as it allows ships to avoid unnecessary detours or waiting periods.
4. Non-Linear vs. Additive Effects: The model analyzes whether simultaneous closures create non-linear cascades. It finds that for reroutable chokepoints, global losses are approximately additive, though regional impacts can be dominated by the most critical corridor.

Key Results

• Adaptive Rerouting Redistributes Loss: Rerouting reduces direct losses at exposed ports but shifts deficits to later port calls and dependent regions. For example, a Panama closure causes immediate losses in the Americas but delayed losses in European and Asian ports due to vessel cycle delays.
• Duration Sensitivity:
  • Peak Shortfall: Once the fleet adapts, extending the closure duration does not significantly worsen the worst day of arrivals.
  • Cumulative Loss: Losses grow linearly with duration. Each additional closure day reduces global shipping arrivals by approximately 3.0% for Suez, 2.2% for Panama, and 2.1% for Malacca. A simultaneous closure of all three results in a 7.7% daily reduction.
• Information Impact:
  • Knowing the reopening date significantly reduces cumulative losses (e.g., reducing 50-day losses from 4.62 to 3.78 shipping days) by preventing avoidable waiting or detours.
  • Start-date information primarily helps in avoiding peak shortfalls for long-duration closures where waiting is no longer a viable strategy.
• Static vs. Dynamic Discrepancy: Static structural exposure models often overstate losses for reroutable chokepoints (like Malacca) because they ignore detours. Conversely, they understate losses for chokepoints where delays propagate to later legs of a journey (like Suez and Panama), as static models miss the "missing vessel cycle" effects.
• Non-Reroutable Chokepoints: In the Strait of Hormuz scenario (where rerouting is geographically impossible), losses scale superlinearly with duration, contrasting with the linear scaling observed in reroutable scenarios.

Significance and Claims
The paper claims that maritime chokepoint risk is not a static property of network topology but a dynamic problem involving routing, timing, and regional exposure. The authors argue that:

• Static exposure metrics are insufficient for predicting realized disruption impacts because they fail to capture adaptive rerouting and the propagation of delays through vessel cycles.
• Operational resilience depends heavily on the information available to shipping agents; specifically, knowing the duration of a disruption allows for more efficient decision-making (waiting vs. rerouting), thereby reducing avoidable economic losses.
• Systemic risk is characterized by the persistence of delays rather than just the immediate blockage. Even after the fleet adapts, the system continues to lose efficiency due to the inherent length of alternative routes.

The study concludes that while adaptive capacity limits the severity of immediate arrival shortfalls, it cannot eliminate the cumulative cost of longer routes, and that the distribution of these losses is highly heterogeneous across regions and vessel types.