Case study · Strait of Hormuz · January–October 2026

The strait all but emptied.
Outside it, the anchorages filled.

Ship-tracking shows the crossings of the Strait of Hormuz collapsing. Outside it, radar satellites count far more ships at anchor off Fujairah, Khor Fakkan and Sohar than before the war. We counted them on every pass, and dated and sourced every attack on a named ship or site.

Strait of Hormuz · Fujairah · SoharSentinel-1 radar · IMF PortWatchEvidence: …
01 · Four numbers

Few ships through. Many at anchor outside.

Two independent counts tell the same story. Ship-tracking signals (AIS) show the crossings collapsing. Radar, which sees a steel hull whether or not it sends a signal, shows the anchorages on the Gulf of Oman side filling up.

…
 
…
 
…
 
…
 

The war began on 28 February with US and Israeli strikes on Iran; «before the war» is 1 Jan–27 Feb 2026. A typical pass is the mean of the evening passes (Sentinel-1 relative orbit 57, about 18:16 local, 14:16 UTC) in each period. The numbers are the 3D globe's own, read from its data as this page opens; the method and every source are in the globe.

02 · What the radar sees

The anchorages outside the strait filled.

Radar sees a steel hull whether or not the ship sends its tracking signal. Counted pass after pass, the anchorages outside the strait hold far more ships than before the war. Radar does not say why a ship is there: Fujairah is also a refuelling hub and the outlet of a pipeline that bypasses the strait.

…

03 · The attacks

Ships and sites hit, each dated and named.

…

A claim is a statement by a party to the war only: it is shown as a claim, never as a fact.

04 · The strait in 3D

Fly the strait. Count the ships yourself.

The radar passes from January to October, ship by ship; every attack on its date with its sources; the ship-tracking crossings week by week; and who depends on Gulf fuel. Follow the tour stop by stop, or explore on your own.

Open the 3D globe

Contains modified Copernicus Sentinel-1 and Sentinel-2 data 2026 (ESA) · terrain, Bing imagery and Google Photorealistic 3D Tiles via Cesium ion; the Google 3D model's date is unknown: it is not the current condition · © OpenStreetMap contributors. Best on a computer; on a phone, the tour works with a tap.

05 · What this cannot tell

The limits, said plainly.

Ship-tracking (AIS) misses ships, so the crossings PortWatch counts are likely an undercount (ships can switch tracking off; false positions also occur). Radar sees steel hulls, not names or cargo: it does not identify the type of ship, and lengths from 20 m pixels are good to about ±20–40 m. Small boats and dhows are under-counted. Ships moored side by side can merge into one echo, so in a dense anchorage the count is, if anything, low. Ships at berth are not counted: land and a 300 m strip along every coast are left out. The trend uses the evening orbit only (about 18:16 local), the one that sees every area in full. Attack reports are kept in three kinds, confirmed, single source and claims, and a claim is never counted as a fact.

Watchfloor

From a strait all but empty to the anchorages outside it.

Every analysis starts from a multi-source gather: ship tracking, satellite imagery, public reports and our own measurements. Ask for one on the crisis you follow.

Sources: IMF PortWatch (daily chokepoint transits from AIS, downloaded 19 October 2026) · contains modified Copernicus Sentinel data 2026 (Sentinel-1 radar, Sentinel-2) · ship counts on radar: Watchfloor · attacks and seizures compiled by us from public reports, each linked to its source: the IMO's list of confirmed incidents, UKMTO warnings, Reuters, AP, AFP, Lloyd's List, Maritime Executive and others · © OpenStreetMap contributors, ODbL · UN Comtrade · World Bank commodity prices · 3D: Google Photorealistic 3D Tiles, Cesium World Terrain and Bing imagery via Cesium ion. Method and limits, with every source, in the 3D globe.