When Ship-to-Ship Transfers Replace Berths: The Bonded Storage Demand Hidden in Houston's 51% STS Share

The Overwatch atlas mapped where AIS gaps cluster. Filtering for ship-to-ship rendezvous reveals which port-adjacent CRE absorbs the cargo that never docks. Houston, Rotterdam, and Singapore each tell a different story.

The Setup

Of 397 dark-event detections in the Houston Ship Channel box over the last 30 days, 202 — about 51% — carry the ship-to-ship rendezvous fingerprint. That share is high enough to matter for any landlord whose tenant assumes terminal-side throughput. Comparable shares run 65% in the Singapore Strait, 59% in the Rotterdam approach, and 74% in the Persian Gulf (small sample). Mean transfer dwell at Houston runs 27.9 hours; at Rotterdam it is 15.6 hours. Those two numbers describe two different real-estate problems.

The Chain

A ship-to-ship transfer offshore replaces a berth-side loading or discharge. Tonnage that physically lands in the harbor area still lands somewhere — but where it lands shifts. Short-dwell STS at Rotterdam (15.6 hours, lower mean risk) reads as bunkering and cargo lightering: the volumes never enter shore-side custody, and the inventory question stays at sea. Long-dwell STS at Houston (27.9 hours) and Singapore (30 hours) reads differently — it is consistent with a hold-and-trade pattern where cargo waits offshore for a buyer, a clearance, or a shore-side capacity opening, and then lands at a downstream facility rather than a berth assigned at the start of the voyage.

The CRE consequence is sector-specific. At Houston, the Ship Channel petrochemical corridor — Pasadena, Galena Park, Channelview, Deer Park — runs on bonded and Foreign-Trade-Zone storage that holds product before customs clearance and downstream sale. STS that lands as a late-bound cargo discharge raises bonded utilization and shortens the predictability window for terminal scheduling. At Rotterdam, the same fingerprint does not produce that demand because dwell is short and the transfer is operational, not commercial. At Singapore, the storage pressure absorbs into Jurong Island's chemical cluster and the OPL anchorage fleet itself, where leased floating storage solves what shore-side leasing cannot.

A 51% STS share at Houston with a near-30-hour transfer dwell implies a measurable share of Gulf throughput is reaching shore through a flexible discharge path that bypasses fixed berth allocations. The slack is absorbed by bonded warehouse and FTZ-designated tank space — not by adding berths.

The Implication

For investors and tenants in Houston Ship Channel petrochemical CRE, the directional read is that bonded-storage utilization is partially decoupled from berth scheduling and tracks the offshore STS profile. A move in Houston's STS share toward Singapore's 65% level would tighten bonded inventory days-on-hand. Conversely, a Rotterdam-like profile (shorter dwell, lower risk score) would not produce the same demand pull because the cargo is operationally transferred rather than commercially staged. For Rotterdam hinterland chemical storage, the implication is the opposite — short-dwell STS does not signal demand for bonded land-side storage and should not be priced as such. For Jurong Island, the right benchmark is leased floating capacity in the OPL belt, not the shore-side asset class.

What to Watch

• Houston Ship Channel bonded warehouse occupancy via FTZ Board annual reports and the Texas Comptroller's ad valorem inventory filings — STS share trending toward 60% should precede tightening.
• Rotterdam tank terminal availability via VTTI and Vopak quarterly disclosures — short-dwell STS should not move these.
• Singapore floating-storage rates and OPL anchorage occupancy via Vortexa and Argus — the structural baseline against which Houston is measured.
• CBRE and JLL Gulf Coast industrial cap-rate compression in submarkets contiguous to FTZ #84 (Harris County) and #149 (Freeport) — slow-moving but the right backstop for the directional thesis.

Limitations

The STS rendezvous fingerprint is a heuristic, not a confirmed transfer; mean risk and mean dwell describe pipeline outputs, not customs-cleared cargo. The translation from STS share to bonded-storage demand assumes a stable share of long-dwell transfers result in shore-side discharge — a reasonable working assumption, not a measured pass-through. Persian Gulf coverage is too thin to support a CRE read on Jubail or Ras Tanura. Bonded warehouse and FTZ occupancy data lag by at least a quarter, so this signal precedes the confirming series rather than coincides with it.

--- Data current as of 2026-04-27. Source: Overwatch dark_events pipeline, n = 5,780 dark events at the four chokepoints over 30 days. Cross-references the Overwatch chokepoint atlas (/blog/dark-events-chokepoint-atlas-overwatch-2026-04-24) and port-adjacent CRE atlas (/blog/dark-events-port-adjacent-cre-atlas-locus-2026-04-24).