Transshipment hub or gateway port: read it off the call pattern

A port schedule doesn't label itself. Singapore and Colombo show up on the same tracking feed as Long Beach and Rotterdam, but they're doing structurally different jobs in the supply chain, and that distinction matters if you're trying to nowcast trade flow from call activity. Mix the two up and you'll misread a busy hub as strong consumer demand, or miss a real demand signal buried under transshipment noise.

What actually separates a hub from a gateway

A gateway port moves boxes to or from the hinterland behind it. Cargo arriving at Long Beach is mostly headed to a warehouse in the Inland Empire or beyond; cargo leaving came from a factory or farm somewhere in the country. The vessel call there is tied to end demand.

A transshipment hub, think Singapore, Colombo, Tanjung Pelepas, Algeciras, exists to move boxes between vessels. A container lands on a mainline ship from Asia, sits in the yard for a few days, then goes back out on a feeder to a smaller regional port, never touching a local truck, railcar, or consumer. The port's throughput number can be enormous and still tell you almost nothing about demand in the country it sits in.

Most ports are a blend of both roles. The practical question for a nowcast is how much of a port's call volume is cargo actually entering or leaving the local economy, and how much is cargo just passing through on its way to a third country.

Reading it from vessel call patterns

You don't need the terminal operator's internal split to get a workable read. The call pattern itself gives it away.

Feeder vessel calls vs mainline calls. Mainline ships are the big ones, the 8,000-TEU-plus vessels running fixed Asia-Europe or transpacific strings. Feeder vessels are smaller, often under 2,000 TEU, and they run short-haul legs to and from a regional hub. A port with a high ratio of feeder calls to mainline calls is doing hub work: boxes are arriving on small ships to be consolidated onto big ones, or coming off big ships to be redistributed on small ones. A gateway port skews toward direct mainline calls with comparatively few feeder movements, because the cargo's final leg is by road or rail, not another ship.

Transshipment ratio. This is the rough share of total container moves at a port that are box transfers between vessels rather than moves in or out of the hinterland. You can't get this number exactly without carrier manifest data, but vessel call composition gets you a workable proxy: count daily vessel calls by size class over a stretch of weeks, and watch the feeder-to-mainline ratio. A hub holds that ratio high and fairly stable, because its business is the transfer itself. A gateway's ratio tracks seasonal import and export cycles more than vessel-class mix, since its real business sits upstream of the dock.

Dwell and queue behavior. Hub ports tend to run tighter, more predictable dwell times because transshipment boxes are scheduled against a known onward sailing. Gateway ports show more variable dwell and anchorage queueing tied to customs clearance, trucking capacity, and warehouse availability on the landside. If anchorage counts spike and stay elevated for days without a matching rise in vessel calls, that points to gateway-side congestion rather than hub throughput.

A port's feeder share can shift over a quarter as carriers rework alliance networks, and a nominal gateway can pick up transshipment volume when a hub nearby is congested. The read comes from watching the ratio move over time against your own baseline for that port, not from a single day's vessel count.

That's the part that's hard to do from AIS pings alone, since AIS tells you a vessel arrived but not what's stacked in the yard behind it. A daily satellite pass across the berths and container blocks fills that gap, which is the premise behind a daily read on container stacks, vessel calls, and anchorage queues at named ports. If you're trying to separate hub churn from real demand before the trade data posts, that's the kind of feed worth watching.