Gateway Access Monitor

In brief

Neopanamax draft 49.0 ft in force since 28 Sep (A-36); 32 slots/day to 14 Oct (A-29), 33/day from 15 Oct (A-36); August Canal Waters Time 34.90 h, +37.5% on July (A-34); record ≈US$5.3M auction premium (24 Aug, press-reported) stands.

What changed since the last edition

Direction by dimension (8 applicable) — Methodology §4 · what each measure means

• Nominal capacity — stable: The stated maximum sustainable capacity, 36–38/day, is restated unchanged in the ACP's August operations summary (A-34).
• Effective capacity — stable: In-force Neopanamax draft limit raised 48.0 to 49.0 ft on 28 Sep (A-36), a 2.1% move inside the 5% band; slots unchanged at 32/day to 14 Oct (33/day from 15 Oct announced); AIS 28.1 vs 29.0/day, inside the band.
• Reservable capacity — stable: Slots offered for September transit dates unchanged at 32/day (A-29; Neopanamax 63 a week, A-31); A-36's extra Neopanamax slot (33/day) is for booking dates from 15 Oct — announced, read once awarded.
• Waiting time — deteriorating (worse): The authority's latest monthly measure, August Canal Waters Time 34.90 h (A-34), is 37.5% above the July 25.38 h carried in September; the monthly high rose from 41.13 h to 69.41 h, in-transit time 11.27 vs 10.87 h.
• Cost of secured passage — stable: The record auction premium stands at ≈US$5.3M (24 Aug, press-reported; confirmed to press by the ACP), no higher outcome reported through September; the ACP publishes auction volumes and a bid distribution, not bids.
• Allocation rules — stable: A-35 (16 Sep) and A-36 (rules from 29 Sep for dates from 15 Oct) rewrite booking rules — consecutive dates, segment caps 63→70/week, Last-Minute at discretion — booking stays the only guarantee: rewritten, not eased.
• Access uniformity — stable: The rung (Rule-differentiated, published and observed) and the published criteria — class, segment, TTA range, booking tier, customer ranking, willingness-to-pay — are unchanged; A-36 restates them for dates from 15 Oct.
• Recovery flexibility — stable: Last-Minute Reservation re-admitted on paper only — A-36 rule 12 offers it per N-07 at the ACP's sole discretion when a slot frees, no offering observed; A-37's late-cancellation surcharge is read in November.

Daily capacity: stated, booking slots offered and ships observed

Nominal capacity

36–38/day

In force to 14 Oct (A-29)

32/day

From 15 Oct (A-36)

33/day

Observed transits (Sep, AIS)

28.1/day

Key figures and evidence

Each figure shows its evidence label; open it for the full text and its source, where linked.

• Canal Waters TimePrimary-published

  34.90 h average in August (July 25.38 h, +37.5%); monthly high 69.41 h (July 41.13 h) against a published In-Transit Time of 11.27 h, so the average queue component was ~23.6 h (July ~14.5 h). Oceangoing transits averaged 33.19/day (1,029 in the month; July 34.03) — the ACP's administrative count of all transits, a different population from the AIS series in the comparable table. Published 10 September in A-34-2026, after the September edition's data cut — the follow-up that edition promised; the September operations summary is expected around 10 October.

  Evidence label: Primary-published — The authority's own published figure.

• Booking regime (A-36, in force 29 Sep for booking dates from 15 Oct)Primary-published

  A-36-2026 (issued 28 September; rules effective 29 September for booking dates from 15 October) adds one Neopanamax booking slot — nine to ten a day, with the Booking Period 2 offer rising from one slot to two, the 'conditioned' slot included, and the slot offered at the 2nd Period Competition that opened 1 October — taking the daily offer to 33: 30 booked plus 3 auction (Neopanamax 9+1, Supers 17+1, Regulares 4+1; the Panamax 23 of A-29 unchanged). Neopanamax weekly slots by segment: full container 42, LPG 21, LNG 4, vehicle carriers/RoRo/bulkers/other 3 — 70 a week including the daily auction slot, from A-31's 63 — with minimums of six full-container and three LPG slots per date and four LNG a week; no per-customer limit and consecutive dates allowed, as A-31 already provided for the Neopanamax locks; auction slots sit outside the segment limits. To 14 October, A-29's 32/day offer and A-31's allocation govern. The ACP repeats that, while the water deficit in the watershed continues, the Booking System remains the only mechanism that guarantees a transit date.

  Evidence label: Primary-published — The authority's own published figure.

• Cost of secured passagePress-reported

  Record ≈US$5.3M auction premium (auction 24 August; Bloomberg 25 August; confirmed to press as the highest ever by the ACP's then administrator-designate on 28 August) stands — no higher outcome was reported through September in the daily sweep of wire and trade press; the ACP publishes auction volumes and a registered-access bid distribution, not individual winning bids. Prior record US$4.6M (~12 August). Documented auction record below. Documented auction outcomes

  Evidence label: Press-reported — Reported in press; not published by the authority.

• Announced measuresPrimary-published

  A-36-2026 (28 September): maximum authorised Neopanamax draft raised to 14.94 m (49.0 ft), effective immediately and until further notice, from the 14.63 m (48.0 ft) TFW in force since 2 September — citing September rainfall in the watershed, the water-saving measures at the locks and the current and projected Gatun Lake level; the 14.48 m (47.5 ft) step that A-33 (4 September) had postponed until further notice is superseded on the same measure by A-36, which does not mention it. Slots: 32/day (A-29) for booking dates to 14 October; 33/day from 15 October. The ACP's same-day press release ties the slot step to 'the present and projected level of Gatun Lake'. Update, 8 October: the live advisory index listed A-37-2026 (5 October) as the latest advisory at 00:00 UTC on 8 October. A-37 offers long-term slot packages for transit dates 3 January–3 April 2027 — a Neopanamax competition on 17 November and, for the first time, advance Panamax slots by sealed-bid auction on 24 November, each averaging three slots a day — and applies the late-cancellation surcharge from 15 October (see Recovery flexibility).

  Evidence label: Primary-published — The authority's own published figure.

• Queue Planning Ratio (supplementary)Derived

  August reading 2.46 (bracket 2.39–2.50; July 2.09): the month's worst published day implied a queue about two and a half times the average day's — worst-day queue ≈58.1 h against an average ≈23.6 h, computed from the ACP's published Canal Waters Time (34.90 h avg, 69.41 h high) and In-Transit Time (11.27 h avg; 10.33–13.03 h). The ratio rose with the level: both the average queue and its worst day lengthened. All vessels (booked and non-booked); differences under 0.1 are not read as movement. QPR series (CSV)

  Evidence label: Derived — Computed by WRI from primary figures by a stated method.

• Booking-slot utilisation by class (administrative population, monthly)Primary-published

  New series from this edition, read from the ACP's monthly operations summaries: August booked-slot use Neopanamax 175 of 178 offered (98.3%), Supers 472 of 561 (84.1%), Regular 124 of 150 (82.7%), auctioned slots 204 of 227 (89.9%); July 183 of 163, 463 of 506, 120 of 137 and 247 of 306. The ACP's 'available' count excludes additional auctioned slots and its 'used' count covers booked transits only, so July's Neopanamax use exceeds its offer on those definitions — an administrative population, not a capacity-utilisation rate, and it carries no direction.

  Evidence label: Primary-published — The authority's own published figure.

• Waiting timePrimary-published

  The latest monthly authority measure, August Canal Waters Time (A-34), reads 34.90 h against the July 25.38 h this page carried in September — up 37.5%, with the monthly high up from 41.13 h to 69.41 h and in-transit time 11.27 h (July 10.87 h): the queue, not the lockage, lengthened. Between its monthly summaries the ACP publishes a daily count of the vessels in queue at 00:01, by lock class, in its daily customer report; this edition did not use it. The September summary (~10 October) is the next monthly Canal Waters Time measure. Press-reported waits for unreserved vessels are not used as a measure here.

  Evidence label: Primary-published — The authority's own published figure.

• Access uniformityPrimary-published

  Rule-differentiated (both): the published rules differentiate by locks and vessel class, cargo segment (full container, LPG, LNG, vehicle carriers/RoRo/bulkers/other), TTA capacity range, booking period and tier, customer ranking and willingness-to-pay — A-36 restates the criteria for booking dates from 15 October; no published criterion names an identity. Gatekeepers: ACP; insurer collectives; OFAC (extraterritorial).

  Evidence label: Primary-published — The authority's own published figure.

• Recovery flexibilityPrimary-published

  Announced, not observed: A-36 provides that the Last-Minute Reservation service 'shall be made available' under Notice to Shipping N-07 when a slot becomes available and the ACP, at its sole discretion, decides to offer it — a re-admission on paper of the service suspended by A-28 (20 August), with no offering recorded. Update, 8 October: A-37-2026 (5 October) applies the late-cancellation surcharge (tariff 1050.IBC9) from 15 October until further notice — 250% of the reservation fee, in addition to the cancellation charges, for cancellations received four days or less before the required arrival time and for vessels that fail to arrive within seven days after that time. A-37 describes this as a reduction from the seven days announced in A-28, which A-32 (2 September) postponed until further notice; A-37's own table lists the current notice period as two days. A-37 does not refer to A-32. Vessel substitution and swaps at same-or-greater TTA capacity and booking-date changes within the weekly segment allocation continue (A-36 rules 10–11); A-35 (16 September, effective immediately) lets Panamax customers hold consecutive transit dates awarded in different booking periods. Published rules only; no privately arranged workaround is read as a direction.

  Evidence label: Primary-published — The authority's own published figure.

• Advisory count (automated title screen)Derived

  17 ACP advisories in 2026 to the 1 October 2026 archive retrieval whose titles carry booking, transit, draft, condition, surcharge, slot, auction, reservation, modification or cancellation terms (14 at the 3 September retrieval; A-33, A-35 and A-36 added since) — an activity count from an automated screen of advisory titles, not a count of rule changes; titles without those terms are not captured (A-29-2026 is not counted). 2025: 12; 2023: 32. Advisories counted (CSV)

  Evidence label: Derived — Computed by WRI from primary figures by a stated method.

• WatchPrimary-published

  Hydrology remains the binding constraint, with two signs in the ACP's own documents. Relief: the official Gatun Lake level is 85.0 ft for transits beginning 1 October (2026 low 83.80 ft on 26 August; +0.81 ft from 31 August, 84.07 ft, to 29 September, 84.88 ft), the ACP projects 85.8 ft by 15 October and 86.0 ft by 21 October, and watershed rainfall across the series' 31 August–30 September labels (rain to 29 September; ACP 20-station series, posted at the next day's label) was 99.5% of the 1991–2020 norm for the same window. Deficit — the ACP's own 'water deficit continues': 2026 rainfall at the series' 30 September label is 93.4% of climatology (August 74.6%), and the September lake mean of 84.34 ft sits 1.8–2.8 ft below the four non-drought Septembers of 2021–25 (2023: 79.75 ft). NOAA CPC (10 September) assigns a greater-than-90% chance of a very strong El Niño through the Northern Hemisphere fall and winter 2026–27 (next discussion 8 October), and Panama's IMHPA (via the ACP, 11 September) warns the dry season could begin in the second half of November. Also open: the Miraflores west-lane culvert outage listed for 22 September–3 October (A-34, 'Tentative'; the estimated Panamax capacity of 23 printed against it is the standing figure on every Panamax row from 15 September to 16 November, not a culvert-specific number) — no advisory confirms, postpones or completes it, the window ends 3 October and A-36 is silent on it; the ACP's September operations summary is expected around 10 October. PortWatch's full-September vintage (6 October) finalised this edition's transit figures by the dated data refresh of 8 October.

  Evidence label: Primary-published — The authority's own published figure.

Reported auction prices

Documented auction outcomes (record, not a series — carries no direction)

Outlook

In brief

~54% of pre-crisis traffic (September, full month); Cape diversion +67% over reference; Gemini AE5/AE11/AE12/ME2 return on the water — first cohort Suez passage 28 Sep (tracker, observed); Yanbu loadings observed resumed 27–28 Sep (tracker); war-risk differentiated by affiliation (press-reported).

What changed since the last edition

Direction by dimension (7 applicable) — Methodology §4 · what each measure means

Not applicable here, so not counted: reservable capacity.

• Nominal capacity — stable: Stated capacity and channel specifications are unchanged; the design figure near 97/day remains a cited, not SCA-published, number.
• Effective capacity — stable: 40.4/day in September (full month) against 41.0 in August — a 1.5% move inside the 5% stability band, with 0 of 30 days below 80% of the August level.
• Waiting time — stable: Convoy structure and the ~24h missed-convoy interval are unchanged; the EU's 28 Sep statement that commercial ships queue for Aspides escorts at Bab el-Mandeb gives no measure, so no movement is read.
• Cost of secured passage — deteriorating (worse): Reuters 24 Sep (four industry sources): Saudi-linked Yanbu calls ~3% of vessel value from under 1% in early July, up to 7% south of Yanbu — the exposed user's price rose; MSC's 19 Oct extension announced.
• Allocation rules — stable: SCA convoy rules are unchanged; MARAD 2026-013 (23 Sep) renewed the US advisory line by superseding 2026-006 with guidance continuous; Italy's escort (reported 18 Sep) ran within Aspides' existing architecture.
• Access uniformity — deteriorating (worse): The Houthi-run HOCC's 10 Sep communication (via Reuters/Splash247) names Saudi vessels a banned class — a fourth named class on this page's record; attributed, single-source, and called “previously imposed” by the HOCC.
• Recovery flexibility — improving (better): Observed reversions (tracker, no authority record): Maastricht Maersk's southbound Suez passage 28 Sep, first of the Gemini AE5/AE11/AE12/ME2 return; AE15 Salalah calls 26–28 Sep; four more vessels in passage.

Key figures and evidence

Each figure shows its evidence label; open it for the full text.

• TransitsDerived

  40.4/day in September (IMF PortWatch, full month; finalised 8 October on the full-September vintage) against 41.0 in August and 75.3 at the June-2023 reference — about 54% of normal, the plateau near half of pre-crisis traffic now running for more than two years; a design capacity near 97/day is widely cited, though the SCA's public pages carry channel specifications rather than a vessels-per-day figure. Movement against August (−1.5%) is inside the edition's 5% stability band and 0 of 30 days sat below 80% of the August level: read as stable. The SCA's navigation-statistics page carried no September or third-quarter 2026 figures at 1 October — its quarterly series ends at Q2 2026 (3,580 transits: 1,807 southbound, 1,773 northbound; 405 container ships), the latest authority figure.

  Evidence label: Derived — Computed by WRI from primary figures by a stated method.

• The diversion mirrorDerived

  Cape of Good Hope traffic 86.1/day in September vs 51.6 at the June-2023 reference (+67%) — the rerouted cohort, visible daily; 4.7% below August's 90.3, inside the band.

  Evidence label: Derived — Computed by WRI from primary figures by a stated method.

• Bab el-MandebDerived

  26.6/day in September — the lowest monthly mean on the PortWatch series since January 2023, 2.9% below August's 27.4 and inside the 5% band; about 35% of the 76.5 reference, with 0 of 30 days below 80% of the August level.

  Evidence label: Derived — Computed by WRI from primary figures by a stated method.

• Reservable capacityPrimary-published

  Not applicable — the SCA operates convoy scheduling (two daily convoys with published cutoff times), not slot booking; no reservation product exists.

  Evidence label: Primary-published — The authority's own published figure.

• Allocation rulesPrimary-published

  Two daily convoys with published cutoffs (SCA) — the governing rule text is unchanged. The US advisory line for the Red Sea, Bab el-Mandeb, Gulf of Aden, Arabian Sea and Somali Basin is continuous: MARAD Advisory 2026-013 (effective 23 September, expires 22 March 2027) supersedes and cancels 2026-006, which MARAD's advisory page listed under Cancelled at 1 October; it carries the first placement in a US authority document of the reported 24 August Houthi missile strikes on two tankers off Yanbu and AIS guidance for US-flagged vessels. Escort capacity, not the canal, binds protected transits at Bab el-Mandeb: the EU High Representative said on 28 September that Operation Aspides does not have the ten ships its establishment foresaw, that commercial ships are “waiting in the queue to receive the escort”, and that two member states were “hinting” at additional vessels with no concrete commitment (EEAS press conference and arrival remarks, Foreign Affairs Council (Defence)).

  Evidence label: Primary-published — The authority's own published figure.

• Cost of secured passagePress-reported

  The access price is war-risk insurance, differentiated by affiliation. The latest press-reported quotations (Reuters, 24 September; four industry sources declining to be named, and one named broker) split by the vessel's Saudi connection rather than by stretch of water: Saudi-linked tankers calling at Yanbu around 3% of a vessel's value (under 1% in early July); calls at ports south of Yanbu, Jizan included, quoted as high as 7% (around 1% in early July; 'certainly north of 7%' per the head of marine at McGill and Partners); tankers transiting the Red Sea with no Saudi connection typically 0.2–0.3%, with discounts available. Reuters notes there is no public data for agreed rates, which can differ from brokers' and underwriters' quotes. These are a different population from the September edition's press-reported 'around 0.5% of hull value for Bab el-Mandeb cover and 1–2% southern Red Sea' and are not read against them; the affiliation split this page records is now quoted explicitly. MSC's Piracy Risk Surcharge (US$55/TEU) and Suez Canal Surcharge (US$36/TEU) have applied since 15 September by proforma date on Asia–East Mediterranean/Black Sea trades (carrier advisory of 3 September) — unchanged since entering application; on 30 September MSC announced their extension at the same rates to Asia-origin cargo for Derince and Port Said West and 43 West Mediterranean, Adriatic and North African ports from 19 October (proforma date) until further notice — announced, not yet in application; a charging rule rather than a measured realised cost. SCA transit-dues circulars adjust the published tariff (No. 2/2026, containerships, June 2026); rates sit in agents' schedules, not an open SCA table.

  Evidence label: Press-reported — Reported in press; not published by the authority.

• Access uniformityPress-reported

  Identity-conditional (both): Israeli-linked vessels de facto barred; elevated targeting of US-linked tonnage; underwriters screen affiliation. Houthi published criteria name Israel-, US- and UK-affiliated vessels and advise against switching off AIS; in a 10 September communication after the capture of Mocha, the Houthi-run Humanitarian Operations Coordination Center stated that navigation in the Red Sea remains safe with the only restriction a “previously imposed ban on Saudi vessels” (via Reuters and Splash247) — the first attributed statement on this page's record naming Saudi vessels as a class, read this edition as a widening of the identity gate; the ban's date of imposition is not established. No attack on shipping has been reported to UKMTO in the Red Sea or Gulf of Aden since 24 August (VRA Overview Report No. 8, 25 September; six Red Sea attacks year to date, revised from seven). Naval protection is organised by flag and operator: the Italian frigate Carlo Bergamini escorted the Italian-flag Jolly Oro northbound through Bab el-Mandeb within Operation Aspides (Marina Militare release of 18 September, via Report Difesa and RID), and Hapag-Lloyd states each Gemini transit runs under naval protection after a security review (customer advisory, 28 September). Gatekeepers: Houthi forces (non-state), Lloyd's JWC/LMA, EU Aspides (escort allocation), SCA (formally neutral), US/EU/UK sanctions authorities.

  Evidence label: Press-reported — Reported in press; not published by the authority.

• Recovery flexibilityPress-reported

  Convoy rescheduling only within the canal — a missed convoy waits ~24h for the next; press-reported cancellation charges run US$1,000–5,000 by timing and vessel size; no substitution or rebooking-priority rights are published by the SCA (unchanged). Read on rerouting options, as the methodology provides where no booking system exists, the month widened them: four Gemini services are moving from the Cape routing to Suez with one passage recorded at tracker grade and four vessels in passage (observations below). The overland bypass behind the Red Sea export terminal, shut on 10 September and suspended from mid-month, is loading again at tracker grade from 27–28 September — still below the state the previous edition recorded, so it is not read as a widening.

  Evidence label: Press-reported — Reported in press; not published by the authority.

• Service reversions (Gemini return)Third-party estimate

  The four Gemini services moving from the Cape to Suez — Maersk/Hapag-Lloyd codes AE5/NE4, AE11/SE2, AE12/SE1 and ME2/IEX, joining AE15 and AE19 already on the routing (carrier notices of 14 September; Hapag-Lloyd's SE1 is Maersk's AE12, not a fifth service) — are observed on the water. Maastricht Maersk (AE5, eastbound) made a southbound Suez passage on 28 September (anchorage departure 01:54 UTC; canal 03:21–12:22 UTC; tracker-only, recorded at observed — the first passage of this return, no authority record). Cornelia Maersk (ME2, westbound) was inside the canal northbound at 05:46 UTC on 1 October, exit not yet observed. Antonia Maersk (AE11) and Marchen Maersk (AE5) were inside the Red Sea north of Bab el-Mandeb at 06:35 and 05:04 UTC on 1 October with declared Suez arrivals of 2 October — northbound Bab el-Mandeb passages completed by position, crossing times unobserved. Maersk Halifax (AE11, eastbound, departed Vado Ligure 28 September 18:09 UTC) declares Suez for 2 October. AE15's first Salalah calls are tracker-confirmed — Madison Maersk 26–27 September, Morten Maersk 27–28 September (Morten's second Suez passage, 22 September, remains tracker-only). The first SE1/AE12 sailing is named: Umm Qarn westbound ex-Tanjung Pelepas 27 October and eastbound ex-Algeciras 5 December (Hapag-Lloyd and Maersk customer advisories, 28 September) — scheduled, not observed; her last AIS fix (about 27 September, stale) is consistent with the schedule and supports no position statement. Services are not vessels: four services, six named vessels. OOCL Portugal's SCA-confirmed southbound passage of 16 September is closed at Bab el-Mandeb by reappearance in the South China Sea on 1 October (02:15 UTC); the crossing date is unobserved. Declarations are not passages, and each is recorded as such.

  Evidence label: Third-party estimate — An attributed external estimate.

• Red Sea bypass (East-West pipeline / Yanbu)Third-party estimate

  Crude loadings at Yanbu — suspended from mid-September after drone strikes on pumping stations shut the East-West pipeline on 10 September (Saudi energy ministry via wires) — were observed resumed on 27–28 September at tracker grade: satellite imagery of 27 September showed crude loading at Yanbu and Al Muajjiz (TankerTrackers.com, via Reuters), and two VLCCs and two Suezmaxes were loading about 6.2 million barrels at Yanbu on 28 September with all seven crude berths occupied the day before (Windward citing Vortexa, via The Maritime Executive). Reported separately and never combined: loadings about 2 million barrels a day since the week of 21 September (two trade sources via Reuters, 29 September); pipeline throughput about 2 million barrels a day (industry source) to 2.65 million (Kpler), with Kpler expecting 3–4 million within days and the full rate (about 5.5 million) in roughly a month; Aramco issued customers its October Yanbu loading schedule on 28 September (refining source via Reuters). Throughput is not loadings and loadings are not exports. No authority or operator confirmation exists — JMIC Update 100 (27 September, 15:00 UTC) recorded regular loadings as not independently confirmed — so the loading milestone is recorded as observed at tracker grade, not verified.

  Evidence label: Third-party estimate — An attributed external estimate.

Outlook

In brief

≈4% of reference traffic (3.6/day in September); access via Iranian permits or US-facilitated passage; 18 of 30 September days below 80% of the August level; UKMTO's strait warnings ran 133-26 → 146-26, with the first fatality in the tracked sequence on 23 September.

What changed since the last edition

Direction by dimension (6 applicable) — Methodology §4 · what each measure means

Not applicable here, so not counted: nominal capacity and reservable capacity.

• Effective capacity — deteriorating (worse): September mean 3.6/day over 30 days vs 4.7 for August — a 23% fall, outside the 5% band; 18 of 30 days below 80% of the August level, longest run 6 (PortWatch, full-month vintage).
• Waiting time — stable: No authority publishes a waiting measure and no comparable published measure changed; the press-reported lengthening of Gulf of Oman ship-to-ship operations (~10 days from 5–7) is a workaround queue outside the strait.
• Cost of secured passage — deteriorating (worse): Freight records — Baltic TD3C MEG–China VLCC TCE US$1.21M/day (17 Sep), LSEG US$1.27M/day (21 Sep); Maersk's US$1,000/box Hormuz fee; war-risk 6–9% (Reuters 24 Sep) beside the 8 Sep 7.5–12.5% — no series.
• Allocation rules — deteriorating (worse): OFAC designated the BitBank payment rail for the Iranian transit fee on 17 Sep (Treasury sb0632) — an announced enforcement change on the non-facilitated path; permit regime, NCAGS routing and MARAD 2026-011 unchanged.
• Access uniformity — deteriorating (worse): PGSA Non-Compliance List 56 (2 Sep) → 80 on the PGSA's page (1 Oct), now extended to any vessel cooperating with a listed one via STS or transshipment — criteria widened; 26 of 43 UKMTO strikes on the Omani route.
• Recovery flexibility — deteriorating (worse): Crude workarounds degraded in-month: East-West pipeline shut 10–22 Sep; Yanbu loadings observed 27–28 Sep (satellite/tracker; trade sources ~2 mb/d from the week of 21 Sep; Aramco silent); Gulf of Oman STS at its limits.

Key figures and evidence

Each figure shows its evidence label; open it for the full text.

• TransitsDerived

  3.6/day for September (IMF PortWatch, full month, finalised 8 October; daily counts 1–8) vs 4.7 for August, 102 at the June-2023 reference and 78.3 in February 2026, the last full month before the crisis — ≈4% of reference and ≈5% of February; 18 of 30 days sat below 80% of the August level (longest run 6 days). AIS counts are a floor: UKMTO's VRA Overview Report No. 8 (25 September) puts US-stated facilitated movements on the southern route near 30 vessels a day against about 6 AIS-derived transits a day over 17–23 September — an approximate 5:1 ratio printed with UKMTO's caveat that it cannot independently verify facilitated figures; the two populations are never summed here. High-frequency nowcasts are separate populations and vintage-dependent: Reuters' preliminary ship-tracking count printed 9 for Thursday 24 September (8 out, 1 in; 10-day average near 18 — Reuters wire of 25 September, which names Kpler only for the Bab el-Mandeb count), 5 on Saturday 26 September and none on Sunday 27 September (first prints, preliminary; via Gulf News), while Windward — a different tracker whose counts include dark detections — put transits at 16 on 28 September and 17 on 29 September (6 in, 11 out, 7 of them dark — Windward's own daily brief, data as of 29 September; the two totals also carried by The Maritime Executive). No Reuters daily print was found for Monday 28 to Wednesday 30 September, so the Windward figures are not read as a continuation of the Reuters series. None of these series is spliced with another or with the PortWatch line.

  Evidence label: Derived — Computed by WRI from primary figures by a stated method.

• Estimated commercial flow (context)Third-party estimate

  Kpler's preliminary September accounting as re-issued by Reuters (29 September) puts crude exports from the key Middle East producers at 16.3 million b/d against 19.5 million in February, with about 9.7 million b/d 'via the Strait of Hormuz' on a tally that includes Gulf of Oman ship-to-ship transfers, Fujairah and Oman terminal shipments and Red Sea exports; the wire's first issue (28 September) printed 12.8 and 7.4 million on a narrower cut — two scopes, not a revision; AIS-off movements excluded. A separate seven-day cut (Kpler via CNBC, 30 September) puts crude transiting the strait at 13.5 million b/d to 28 September, matching Kpler's pre-war baseline, and refined products at 677,000 b/d against 3.6 million b/d before the war; Kpler also estimates that about 40% of Gulf crude now leaves by pipeline without crossing the strait (17% before the war). The windows and populations differ and are not combined; reconstructed flow estimates do not change the observed-access reading and are never combined with the PortWatch series.

  Evidence label: Third-party estimate — An attributed external estimate.

• Nominal / reservable capacityPrimary-published

  Not applicable — no design capacity is published for a strait and no booking mechanism exists; Iran's PGSA permit is a case-by-case administrative gate, not a reservable slot.

  Evidence label: Primary-published — The authority's own published figure.

• WaitingPress-reported

  No formal queue and no authority publishes waiting data; tracker-estimated counts of vessels holding position vary with the population counted and are not comparable across providers. The workaround queue outside the strait lengthened: a Gulf of Oman ship-to-ship operation now takes nearly 10 days against five to seven previously, with transfer capacity described as at its limits (Vortexa and trade sources via Reuters, 25 September) — a measure of the shuttle system, not a strait waiting time; the Cape diversion adds 10–14 days per leg.

  Evidence label: Press-reported — Reported in press; not published by the authority.

• Cost of secured passagePress-reported

  Carrier tariff: Maersk's Middle East operational update 48 (17 September) published the first carrier price for secured Gulf passage — emergency freight of US$1,800 / 3,000 / 3,800 per 20ft / 40ft / reefer container across Gulf destinations plus US$1,000 per container for any Hormuz transit, stated to cover insurance premiums and crew-risk compensation and replacing the prior landbridge charges (carrier-published). Freight: the Baltic Exchange's TD3C Middle East Gulf–China VLCC assessment reached a round-voyage time-charter equivalent of US$1,212,503/day on 17 September (exchange-published; Worldscale 1,140), and LSEG data via Reuters put the Middle East–China VLCC daily rate at a record US$1.27 million on 21 September (press-reported); the Gulf of Oman–China route assessment the September edition carried (US$358,201/day on 8 September, Lloyd's List) is a different route and is not compared; Kpler estimates a freight premium near US$10 a barrel for loading inside the Gulf against the Gulf of Oman (third-party estimate). War-risk cover: Reuters (24 September; four industry sources declining to be named) quotes premiums for Hormuz transits at 6–9% of a vessel's value, noting that no public data exists for agreed rates, which can differ from quotes — a separate press quotation from a different set of sources, printed beside, not against, the 8 September indications of 7.5–12.5% of hull value per transit (≈0.25% pre-conflict) carried by the September edition; no movement is read between the two. Iranian transit fees remain press-reported at US$150k–2M; paying them is OFAC-sanctionable and voids Lloyd's war-risk cover, and since 17 September the fee's payment rail (BitBank) is itself designated.

  Evidence label: Press-reported — Reported in press; not published by the authority.

• Allocation rulesPrimary-published

  Access runs through Iran's Persian Gulf Strait Authority (PGSA) case-by-case permits (no reservation product) and parallel US-facilitated passage on the southern, Omani-side route coordinated through US Naval Forces Central Command's Naval Cooperation and Guidance for Shipping (NCAGS). The permit regime and the coalition routing guidance are textually unchanged, and MARAD advisory 2026-011 (effective 9 September 2026, expires 8 March 2027) remains the US framework advisory, listed Active on MARAD's page at 1 October (2026-006, the Red Sea advisory, now sits in the Cancelled table after 2026-013 superseded it). The enforcement envelope tightened: OFAC designated the BitBank payment rail used for the Iranian transit fee on 17 September (Treasury release sb0632), completing a three-tier designation — the PGSA (27 May), the Hormuz Safe Marine Services Authority (29 July; a distinct entity) and now the payment channel — with the payer rule (OFAC FAQ 1249) unchanged; no further Hormuz-related OFAC action followed between 24 September and 1 October. The Financial Times has reported, from US naval coordination centre emails to maritime advisers, that US air-defence protection for tankers narrowed to two daily time slots around the start of September after night attacks increased (press-reported; recorded by the September edition on 15 September). Enforcement and facilitation series, each in its own population: CENTCOM's cumulative count of commercial vessels redirected under its blockade of Iranian ports ran 109 (20 September), 115 (23 September), 122 (25 September) and 125 (30 September) on its own posts — a belligerent-party statement, read as an increment, never a rate; JMIC Update 100 (27 September) records 89 US-facilitated strait transits over 24–26 September, about 30 a day; UKMTO's Overview Report No. 8 puts US-stated facilitated movements since 13 July at 669 outbound and 674 inbound. The three are never merged.

  Evidence label: Primary-published — The authority's own published figure.

• Access uniformityPrimary-published

  Identity-conditional (both): access varies by flag, affiliation, ownership and destination — Iran-linked and PGSA-permitted vessels advantaged on the northern, Iranian-side route; US/allied-linked tonnage and vessels without crisis-priced cover disadvantaged. The identity gate widened again: the PGSA's Non-Compliance List, read on the PGSA's own page on 1 October (06:30 UTC; the page prints no date), carries 80 vessels, against 56 on 2 September and the 57 recorded by the September edition on 20 September; the press ladder in between ran 77 (Seatrade Maritime, earlier in September) and 81 (The Maritime Executive, 25 September; Seatrade's 28 September table), and the list is not monotonic — against Seatrade's 28 September table one vessel is gone and one is new. The page states that any vessel cooperating with a listed vessel 'via STS, transshipment, etc.' will itself be listed and that cargo owners must check the list before chartering — an extension of the published criteria to counterparties (a party's instrument; its enforcement assertions are claims). The PGSA's 26 September post warns that charterers forcing vessels onto 'illegal routes' expose them to listing (claim-class; no post after 26 September, and the list changes were not announced). The incident pattern is consistent with differentiation, the weakest evidence class: UKMTO's Overview Report No. 8 (25 September) counts 43 projectile-strike incidents since 6 July, 26 of them on the southern Omani route and 25 with engine rooms struck; the count includes incidents without warning numbers and is never equated with the numbered sequence. The IRGC Navy's claim (27 September) to have targeted 19 vessels on 25–26 September for sailing outside Iran-authorised lanes has no corresponding UKMTO entry and JMIC Update 100 records no confirmed attack for 24–27 September — a claim, uncorroborated; absence of a report is not refutation given the time-late pattern.

  Evidence label: Primary-published — The authority's own published figure.

• Recovery flexibilityPress-reported

  No booking system exists, so recovery means rerouting or a workaround, read cargo class by cargo class. Crude: the East-West pipeline to Yanbu, shut on 10 September after drone strikes on its pumping stations (Saudi energy ministry via wires; Reuters dates the shutdown 11 September), was reported restarted at a reduced rate on 22 September, and loadings at Yanbu were observed again on 27–28 September by satellite and tracker (TankerTrackers.com imagery of 27 September; Vortexa via Windward on 28 September) — observed-loading grade, with no Aramco confirmation. Trade sources put crude loadings near 2 million b/d since the week of 21 September (Reuters, 29 September); pipeline throughput is estimated at about 2 million b/d (industry source) to 2.65 million (Kpler) against a pre-attack rate Kpler puts near 5.5 million, with full return estimated at about another month (Reuters, 29 September); pipeline throughput, terminal loadings and exports are different quantities and are not combined. Gulf of Oman ship-to-ship transfer, the other crude workaround, is described as at its limits (trade sources via Reuters, 25 September), and Aramco has sold more than 60 million barrels for transfer off Sohar for September–October loading (trade sources; sales are not deliveries). LNG: QatarEnergy's chief executive said on the record on 20 September that Qatar has decided not to build bypass pipelines and will only use the strait — no alternative exists for Gulf LNG. Containers: Maersk's update 48 restored a priced Hormuz-transit service to Gulf destinations in place of its landbridge. For cargo already outside the Gulf, Cape of Good Hope rerouting remains the practical option.

  Evidence label: Press-reported — Reported in press; not published by the authority.

• Security incidents (record)Primary-published

  UKMTO's numbered warnings in the strait ran from 133-26 (10 September) to 146-26 by month-end: thirteen strait warnings, twelve classed Attack and one Suspicious Activity (136-26 in the number sequence is a Gulf of Aden entry). Since 142-26 (23 September): 143-26 — a time-late, third-party report of a vessel struck by a suspected unknown projectile on the evening of 28 September, fire extinguished, vessel underway, crew safe, which UKMTO classes as Suspicious Activity, not Attack (report 29 September 05:30 UTC); then three Attack warnings issued 30 September for 29 September, report time 'TBC', source 'Verified' — 144-26, a crude oil tanker struck on the port side; 145-26, a tanker struck while transiting inbound; 146-26, a tanker struck. UKMTO names no vessel, casualty, damage or time for the three; 146-26 as re-issued on 30 September (10:35 UTC) describes 'a Tanker' with no sub-type — an earlier issue of the same warning, since withdrawn, described 'an LNG tanker', which is why press reports carry 'LNG'; neither sub-type is carried as current here; no UKMTO entry was created between 23 September (11:58 UTC) and 29 September (10:28 UTC), and JMIC Update 100 reports no confirmed attack for 24–27 September. The IMO's list stands at 88 confirmed incidents and 24 seafarer fatalities as at 30 September — entries through 29 September, now including the Cape Dao (23 September, one seafarer fatality) and three further strait entries, Al Funtas (28 September), Sinbad and Mersin Prosperity (29 September), each 'Damaged. No pollution.' — the IMO's own identifications and count, not matched to UKMTO's warning numbers and never summed with them. UKMTO's Overview Report No. 8 tiles for the Strait of Hormuz, Arabian Gulf and Gulf of Oman since 28 February: 87 reports of damage to vessels, 14 near-miss reports, 5 confirmed total losses, 106 events in all, and a combined 'injury / fatalities' tile of 33 that is not a fatality count. Overview Report No. 9 (due 2 October) and JMIC Update 101 were not posted at the data cut.

  Evidence label: Primary-published — The authority's own published figure.

Outlook

Completed passages and milestones

Per-gateway ledger — Methodology Annex C

Completed passages whose completion is stated by an authority record, a corroborated arrival or reporting of the delivery are labelled single completions; a passage read from tracker logs alone sits at observed; milestones follow the ladder announced → scheduled → observed → verified. Date is the event date; Checked is the date WRI checked the entry against its cited source (verified_date in the CSV). Full ledger: milestone ledger (CSV) · provenance.

Revisions between vintages

High-frequency source revisions — Methodology Annex B

Of the 41 observation-days on the two daily high-frequency series this edition cites (Kpler counts as carried by Reuters and the named outlets; observation days 26 August to 27 September, vintages 27 August to 28 September; 69 prints in all in the published table), 19 were printed at more than one vintage (14 Strait of Hormuz, 5 Bab el-Mandeb) and every one of them changed: 18 revised upward and 1 downward (Bab el-Mandeb, 1 September, 18 to 17 — a single vessel). The largest revision was the Strait of Hormuz count for 26 August, 8 to 17 (+9, more than doubled); the largest proportional revision was 15 September, 4 to 12 (tripled); six observation-days at least doubled. On low-count days a first print is closer to noise than to a floor, and the 10-day average revises with its days. The later-September prints cited on the Strait of Hormuz page (24 to 29 September) are first prints — floors, not readings; the last observation-day in the table printed at two vintages, Wednesday 23 September, moved 10 to 14 between its 24 and 25 September prints. IMF PortWatch's day-level series was captured at two vintages this cycle — the release vintage (to 27 September) and the full-September vintage published on 6 October: of the 406 observation-days in both (seven gateways, 1 August–27 September), none was revised, and 28–30 September were added; the September transit figures were finalised on that vintage by the dated data refresh of 8 October.

Observation date and vintage date are recorded separately and never compared across vintages; first prints on these series are floors until a later vintage confirms them. Full observation-by-vintage table: vintage log (CSV) · provenance.

Transparency scorecard

Who publishes what