MODELING AND SIMULATING GLOBAL TRADE GEOGRAPHIES
ANALYSIS · CONTAINERIZED TRADE

From global trade to maritime routes

Containerized sports-footwear trade under a Bab el-Mandeb closure scenario

International trade data connect countries. Deep Frontiers translates those flows into a maritime geography by linking them to ports and computing the routes between them.

This case study applies the framework to global containerized trade in sports footwear in 2024, then tests how that geography changes under a hypothetical Bab el-Mandeb closure.

2024 · SITC S4 85125 · harmonized bilateral trade · computed maritime routing

Mapping trade onto the maritime network

431 ktmapped physical trade volume

$16.46 bntrade value

75,082directed port-to-port relations

To move from country-to-country trade to maritime geography, trade flows are assigned to container gateways, including foreign maritime gateways where relevant to a country’s access to the sea.

Deep Frontiers then computes a maritime route for each port-to-port relation under a reference network configuration. The map therefore represents modeled maritime trade flows, not observed AIS trajectories.

Modeled global maritime routes associated with containerized sport-shoe trade in 2024, showing the intensity of mapped trade flows between port gateways. — Reference configuration — Modeled maritime flows associated with global sport-shoe trade in 2024.

Once this reference geography is established, the network can be disrupted while keeping the underlying trade relations unchanged.

Testing a network disruption

23,567affected port-to-port relations

31.39%of modeled relations pass through Bab el-Mandeb in the reference configuration

+3,666 nmweighted mean detour

+40.8%weighted additional distance

In this scenario, Bab el-Mandeb becomes unavailable. The same port-to-port relations and trade quantities are retained; only the maritime routes are recomputed.

Of the 75,082 modeled port-to-port relations, 23,567 transit Bab el-Mandeb in the reference configuration. Closing the passage triggers large-scale rerouting, primarily around the Cape of Good Hope.

World map comparing affected sport-shoe trade routes before and after a hypothetical Bab el-Mandeb closure, with nominal routes in cyan and recomputed routes in yellow. — Before and after — Affected nominal routes are shown in cyan; recomputed routes after closure are shown in yellow.

The impact is highly concentrated

Just 2.59% of affected port-to-port relations account for 80.03% of the additional transport work.

4.5% account for 90%.

The impact of the disruption is far from evenly distributed. When affected relations are ranked by the additional transport work they generate — measured in tonne-nautical miles — a small share accounts for most of the increase.

This concentration identifies the relations where trade volume and additional sailing distance combine to generate the largest increases in transport work.

Cumulative concentration curve showing that a small proportion of rerouted port-to-port relations accounts for most additional tonne-nautical miles after the closure scenario. — Cumulative concentration curve of additional transport work after the closure scenario.

Structure of the concentration

The observed concentration results from the combination of affected volume and additional detour. Of the 23,567 port-to-port relations affected by the Bab el-Mandeb closure, 610 relations — 2.59% — account for 80.03% of additional transport work.

The figure decomposes this concentration relation by relation: the horizontal axis represents additional route distance, the vertical axis affected tonnage, and bubble area their product, expressed in tonne-nautical miles. Cyan-ringed relations correspond to the twelve largest contributions.

The dominant contributions arise from different combinations of tonnage and detour: neither high volume nor a large additional distance, taken alone, explains their weight in the overall result.

Bubble chart relating affected tonnage to additional route distance for port-to-port relations affected by the Bab el-Mandeb closure. Bubble area represents additional transport work; 610 relations, or 2.59% of the total, account for 80.03% of that work. — Figure 4 — Decomposition of additional transport work. The 610 highlighted relations represent 2.59% of the 23,567 affected relations and account for 80.03% of additional transport work. Bubble area is proportional to additional tonne-nautical miles; dashed lines indicate constant levels of transport work.

One case study, a broader framework

Trade

Bilateral product-level trade

Ports

Maritime gateways

Routes

Computed maritime routes

Configurations

Alternative network configurations

Time

Trade geographies since 1964

Sports footwear is used here as a product-level case study. It brings together the full analytical chain: trade, maritime gateways, computed routes, and network reconfiguration.

The same framework can be applied to other products, regions, time periods, and network configurations — from closing a passage to modifying a constraint or a cost function.

Deep Frontiers also draws on a harmonized UN Comtrade database extending back to 1964, making it possible to examine how maritime trade geographies change over time.

Scope of the analysis

The flows shown here are modeled port-level trade flows, not observations of individual cargo movements.

The Bab el-Mandeb closure is a counterfactual network experiment. The results measure changes in routing, distance, and transport work; they are not direct estimates of freight rates, operational delays, or carrier responses.

Deep Frontiers is an active research project. Detailed methods for graph construction, port allocation, and validation will be documented in forthcoming scientific publications.

Explore other configurations

Products, time periods, regions, corridors, and network configurations can be combined to build new experiments on Deep Frontiers.

Discuss a collaboration Explore the platform