New Zealand imports 100% of its refined fuel. The Marsden Point refinery (Northland) was decommissioned and converted to an import-only terminal in April 2022. All inbound tankers carry refined product (petrol, diesel, jet fuel) — no crude oil is imported.

New Zealand’s refined fuel supply chain depends heavily on two refining hubs that are, in turn, heavily dependent on Middle Eastern crude oil transiting the Strait of Hormuz: South Korea (~48% of NZ imports), Singapore (~33%), Australia (~12%), and other sources (~7%). A sustained closure of the Strait would propagate through to NZ fuel supply with a lag determined by transit times and refinery crude buffer stocks.

2.1 Opening stock levels

All stock data is sourced from the MBIE Weekly Oil Stock Update published 12 March 2026, reporting data as at 8 March 2026 (Day +8 from crisis onset on 28 February).

2.2 Minimum Stock Obligation (MSO) thresholds

New Zealand maintains Minimum Stock Obligations under IEA stockholding commitments.

2.3 Daily national fuel consumption

NZ daily fuel consumption of approximately 22 million litres per day (ML/d) across all refined products. Used for the DWT-to-supply-days conversion (§3.4) but does not directly drive the stock model, which operates in MBIE’s “days of demand” unit.

2.4 Import source shares

Treated as constant. In practice, shares fluctuate quarter to quarter, but 48/33/12/7 represents a reasonable central estimate of the structural dependency.

2.5 Transit times

Standard laden voyage times for medium-range (MR) and long-range (LR) product tankers on established routes.

2.6 Refinery crude oil dependencies

Each refining hub’s dependency on Middle Eastern crude determines the “crude crash” — the point at which refineries exhaust their ME crude buffer and must fall back on alternative crudes.

2.7 Refinery crude buffer stocks

Days each hub can continue operating at current throughput before ME crude stocks are exhausted.

2.8 Confirmed vessel data

As at 27 March 2026, 15 active refined product tankers tracked. Silver Philippa discharged at Timaru (Day 13–18) and departed NZ — supply captured in MBIE Day 15 baseline. Hafnia Falcon discharged Marsden Pt + Tauranga, coastal: Wellington 28 Mar (slipped from 26 Mar), Bluff 01 Apr. Pacific Crystal discharged (Lyttelton + Bluff + Dunedin), multi-port complete. Torm Herdis DEPARTED Marsden Pt 25 Mar. Torm Diana DOCKED Tauranga 25 Mar 14:46 (after 7d at anchor), ETD 28 Mar 15:00 (slipped from 27 Mar), next Wellington. Diamond Express en route Nelson from TGA (departed 25 Mar), returns TGA 01 Apr via Wgtn. Magnolia Express at Port Otago 28 Mar (from Timaru), departs 29 Mar for Nelson. Multi-port: Lyttelton→Timaru→Pt Otago→Nelson. Oak Express arrived Wellington 27 Mar from Port Taranaki, shifts berths 28 Mar, next TGA 03 Apr. STI Magic at Wellington (arrived 25 Mar), Napier 29–30 Mar, TGA 02 Apr (slipped from 01 Apr). CS Fujairah AIS ETA slipped to 3 Apr (from 2 Apr), stale AIS 7d (mid-ocean). Front Pollux passing New Caledonia, AIS ETA 30 Mar 04:00. Four additional vessels confirmed via VesselFinder AIS and NZ Tanker Watch: Redwood Mariner (ETA 8 Apr, Tauranga, IMO 9902861), Chang Hang Kai Tuo (ETA 10 Apr, Marsden Pt), Oriental Aquamarine (ETA 11 Apr, Tauranga — dest changed from Lyttelton 27 Mar), Hafnia Expedite (ETA 8 Apr, Marsden Pt, IMO 9735593, departed 23 Mar). Departure gap: 3 days (Day 27 minus Day 24). No new source port departures detected.

Total confirmed supply: 35.4 days across 15 vessels, arriving Day 19–42 (19 Mar–11 Apr 2026). Silver Philippa supply (~1.5d) already captured in MBIE Day 15 onshore baseline. CS Fujairah departed Day 17 and AMASYA departed Day 20 — both captured in MBIE Day 22 on-water baseline. Redwood Mariner departed Day 22, also captured. Redwood Mariner, Chang Hang Kai Tuo, Oriental Aquamarine, and Hafnia Expedite confirmed via NZ Tanker Watch and VesselFinder AIS 22–24 Mar — not yet on standard port schedules.
Excluded: White Pearl (bitumen), Awanuia (small bunker tanker, ~3–5k DWT — immaterial), Stolt Hagi (chemicals).

Each assumption is categorised by its nature and the degree to which it is supported by evidence.

3.1 No crude oil is imported into New Zealand Confirmed

Since the closure of the Marsden Point refinery in April 2022, NZ has imported no crude oil. All inbound tankers carry refined product. This applies to all vessel classes — including larger tankers (Torm Herdis at 115k DWT, Front Pollux at 110k DWT) which carry refined product on NZ-bound voyages.

3.2 No vessels discharged cargo between 8 and 17 March Confirmed (Silver Philippa discharging at Timaru from 17 Mar)

AIS tracking confirms no product tanker arrivals or discharges at NZ ports during this 7-day period. Onshore stock declined by exactly 7 days of demand during the roll-forward period, with no offsetting arrivals.

3.3 Singapore throughput reduced to 60%, effective Day 8 Confirmed

Singapore Refining Corporation (SRC) confirmed a throughput reduction. The model uses 60% of normal capacity from Day 8 onward as the default (adjustable by user).

3.4 DWT-to-supply conversion: 50,000 DWT ≈ 2.5 days Derived

Based on daily national consumption of ~22 ML/d and standard petroleum product density (0.82–0.85 t/m³). A 50,000 DWT tanker at typical utilisation yields ~55 ML — approximately 2.5 days of NZ national consumption. The conversion is linear across vessel sizes.

3.5 Departure day = ETA minus 17 days Estimate

The model infers each vessel’s departure by subtracting the 17-day transit time from the AIS-reported ETA. Used to determine whether a vessel departed after the MBIE data date (Day 8).

3.6 Product allocation proportional to MBIE on-water split Assumption

Known vessel departures are aggregate — we don’t know per-product cargo composition. The model allocates using each product’s share of total on-water stock:

3.7 Demand is constant at 1 day per day Assumption

Demand continues at the normal rate unless the user applies a reduction. NZ fuel demand has mild seasonality, and demand destruction or rationing would likely occur before physical stockouts.

3.8 Throughput ramp between cutback levels: 20 days Assumption

When a refinery transitions from its initial cutback level to a deeper cut, the model applies a linear 20-day ramp rather than a step change. Reflects operational reality that refineries adjust run rates gradually.

3.9 S. Korea throughput estimated at 50%, effective Day 10 Estimate

Unlike Singapore (confirmed by SRC), the South Korean refinery response is estimated from industry reporting. The model defaults to 50% throughput from Day 10.

3.10 Australia surge capped at 2× normal share Assumption

Australia can increase supply to NZ in response to reduced Korean/Singaporean throughput. Surge is capped at 2× Australia’s normal share (12% → max 24%). Applied with a 4-day lag matching transit time. Default is zero (no surge assumed).

3.11 “Other” sources maintain 100% throughput Assumption

The 7% sourced from countries other than S. Korea, Singapore, and Australia continues at full rate. These sources are diversified and generally not ME-dependent. The small share limits the impact even if this assumption is wrong.

3.12 S. Korean export ban recovery caps at 80% Assumption

If a S. Korean export ban is enacted and subsequently lifted, Korean export volumes recover to only 80% of pre-crisis levels — reflecting likely domestic stockpiling priority.

3.13 Supply restoration lag: 45 days Estimate

After Hormuz reopens, crude takes 18–20 days to reach the refinery, then 45 days for throughput to ramp back to 100%. Total recovery from reopening to normal NZ supply: approximately 62–82 days.

3.14 Ships on the water carry full cargoes High confidence

The “full water toggle” (default: on) treats all vessels loaded before today as carrying 100% cargo, regardless of throughput settings. Throughput reductions only affect future loadings — ships already at sea carry what they were loaded with.

3.15 No unknown vessels will depart for NZ Conservative

The model includes 15 active confirmed vessels as of 27 March (Hafnia Falcon, Pacific Crystal, Magnolia Express now discharged/discharging in NZ; Silver Philippa discharged at Timaru, supply in MBIE baseline; CS Fujairah and AMASYA confirmed via MarineTraffic AIS direct search; Diamond Express and STI Magic now arrived; Redwood Mariner, Chang Hang Kai Tuo, Oriental Aquamarine, and Hafnia Expedite confirmed via VesselFinder AIS and NZ Tanker Watch). After the confirmed arrival window ends (Day 42 — Oriental Aquamarine), it relies entirely on the smoothed supply model.

3.16 Onshore depletion is tracked independently of MSO Design choice

The model tracks two distinct risk metrics:

MSO breach — total stock (onshore + on-water) falls below the regulatory minimum. A compliance and policy trigger.

Onshore depletion — physical stock at terminals approaches zero. Fuel on ships in the Pacific cannot be dispensed at petrol stations.

Any MSO breach or onshore depletion within the 90-day forecast triggers at least a Level 2 alert. This addresses scenarios where total stock remains above MSO floors while onshore stock runs critically low.

The model includes three presets that populate input parameters with internally consistent assumptions. All parameters remain adjustable after selection.

MBIE publishes fuel stock levels every Wednesday as “days of cover” for each product. Understanding their methodology is critical for interpreting changes between releases.

Calculation

Days of cover = physical stock (ML) ÷ average daily demand (ML/day)

The denominator is fixed: “the average daily demand for the 12 months ending 4 months before the reporting period begins.” For the current reporting period this gives:

What “on-water” means

MBIE's on-water figure counts “vessels that have already departed” as at the data date. A vessel still loading at a refinery port is NOT in the on-water count. Once departed, it enters the count and stays there until it arrives in NZ and discharges.

Comparison with our model

Our model was previously using 22 ML/day total daily consumption. This has been updated to 23.6 ML/day to match MBIE's denominator exactly. The model's “days of cover” are now directly comparable to MBIE figures.