Iron Ore Sampling at Port Hedland: Who Picks the Points Picks the Number
Quote from chief_editor on May 18, 2026, 3:30 pmIron ore quality is determined by sampling. The sampling protocol determines the sample. Buyers who don't specify the protocol accept the seller's version.
Port Hedland, Western Australia. The world's largest bulk export terminal by tonnage. In any given week, vessels are loading iron ore for Chinese, Japanese, South Korean, and European steel mills. The quality of what goes on each vessel — the iron content, the moisture, the silica and alumina levels — determines the price, the yield in the blast furnace, and whether the receiving mill's quality control department accepts or rejects the parcel.
The quality is determined by sampling. And sampling, at Port Hedland as everywhere else in bulk commodity trade, is a technical process governed by standards that most cargo buyers have never read.
ISO 3082 covers the sampling and sample preparation of iron ores. It specifies the minimum number of increments required for a lot of a given size, the equipment standards for mechanical samplers, and the procedures for sample division and preparation. A 170,000-tonne Capesize cargo loading at Port Hedland requires a minimum number of increments specified in the standard that depends on the precision required and the variability of the material. The standard permits calculation of minimum increment numbers — the point being that higher variability materials require more sampling to achieve the same confidence level.
The Precision Stated in the Certificate Depends on Assumptions Nobody Disclosed
When a certificate of analysis arrives stating that an iron ore cargo is 61.2% Fe content, the number looks precise. It is a measurement result from a laboratory analysis. What the certificate typically does not state is the confidence interval around that measurement — the range within which the true mean of the cargo, if you could test every particle, would fall with a specified probability.
Industry estimates for precision in iron ore sampling suggest that even under ISO 3082-compliant sampling, the 95% confidence interval for Fe content on a large bulk cargo can be plus or minus 0.3 to 0.5 percentage points. This means a certificate showing 61.2% Fe could, with statistical validity, represent a cargo whose true mean Fe content is anywhere from 60.7% to 61.7%. On a high-grade iron ore priced at around $120 per dry metric tonne, a 0.5 percentage point difference in Fe content translates to a price difference of roughly $3 to $5 per tonne — material on a 170,000-tonne parcel.
This imprecision is not a defect in the measurement system. It is inherent to the sampling of heterogeneous bulk materials, and the ISO standard is designed to manage it within acceptable bounds. The problem arises when buyers treat the certificate number as exact when the contractual specification allows no tolerance — when a contract says "minimum 61.0% Fe" and the cargo tests 60.8%, the buyer invokes a quality claim, and the argument that the difference is within the sampling uncertainty becomes an arbitration question rather than a technical fact.
Major iron ore producers — Rio Tinto, BHP, Fortescue — have contracts with large steel mills that include detailed quality tolerance provisions precisely because both sides understand the measurement limitations. Smaller trades, where the buyer is a trading company or smaller mill and the seller is an intermediate trader or smaller producer, frequently have contracts that do not account for sampling uncertainty, creating disputes that the technical reality does not justify.
Who Controls the Sampling Infrastructure Controls the Number
At Port Hedland, the mechanical sampling infrastructure at the major terminals is owned and operated by the terminal operators — in most cases, entities closely affiliated with the major miners. The sampling is conducted according to the operators' established procedures, calibrated against their equipment. Independent verification of sampler performance requires access that third parties cannot always obtain freely.
This is not an allegation of systematic manipulation. It is a structural observation: the entity that controls the sampling infrastructure has practical influence over the sampling process, regardless of their intentions. Buyers who want independent verification of loading quality need to either appoint a third-party inspector with access to verify the sampling procedure or negotiate for split samples that the buyer's laboratory can independently analyze.
The gap between what a quality certificate states and what an independently conducted sampling exercise would reveal on the same cargo is, in most trades, small and commercially immaterial. In some trades, particularly when the cargo is on the boundary of specification, that gap determines whether there is a valid quality claim or not.
Iron ore quality is determined by sampling. The sampling protocol determines the sample. Buyers who don't specify the protocol accept the seller's version.
Port Hedland, Western Australia. The world's largest bulk export terminal by tonnage. In any given week, vessels are loading iron ore for Chinese, Japanese, South Korean, and European steel mills. The quality of what goes on each vessel — the iron content, the moisture, the silica and alumina levels — determines the price, the yield in the blast furnace, and whether the receiving mill's quality control department accepts or rejects the parcel.
The quality is determined by sampling. And sampling, at Port Hedland as everywhere else in bulk commodity trade, is a technical process governed by standards that most cargo buyers have never read.
ISO 3082 covers the sampling and sample preparation of iron ores. It specifies the minimum number of increments required for a lot of a given size, the equipment standards for mechanical samplers, and the procedures for sample division and preparation. A 170,000-tonne Capesize cargo loading at Port Hedland requires a minimum number of increments specified in the standard that depends on the precision required and the variability of the material. The standard permits calculation of minimum increment numbers — the point being that higher variability materials require more sampling to achieve the same confidence level.
The Precision Stated in the Certificate Depends on Assumptions Nobody Disclosed
When a certificate of analysis arrives stating that an iron ore cargo is 61.2% Fe content, the number looks precise. It is a measurement result from a laboratory analysis. What the certificate typically does not state is the confidence interval around that measurement — the range within which the true mean of the cargo, if you could test every particle, would fall with a specified probability.
Industry estimates for precision in iron ore sampling suggest that even under ISO 3082-compliant sampling, the 95% confidence interval for Fe content on a large bulk cargo can be plus or minus 0.3 to 0.5 percentage points. This means a certificate showing 61.2% Fe could, with statistical validity, represent a cargo whose true mean Fe content is anywhere from 60.7% to 61.7%. On a high-grade iron ore priced at around $120 per dry metric tonne, a 0.5 percentage point difference in Fe content translates to a price difference of roughly $3 to $5 per tonne — material on a 170,000-tonne parcel.
This imprecision is not a defect in the measurement system. It is inherent to the sampling of heterogeneous bulk materials, and the ISO standard is designed to manage it within acceptable bounds. The problem arises when buyers treat the certificate number as exact when the contractual specification allows no tolerance — when a contract says "minimum 61.0% Fe" and the cargo tests 60.8%, the buyer invokes a quality claim, and the argument that the difference is within the sampling uncertainty becomes an arbitration question rather than a technical fact.
Major iron ore producers — Rio Tinto, BHP, Fortescue — have contracts with large steel mills that include detailed quality tolerance provisions precisely because both sides understand the measurement limitations. Smaller trades, where the buyer is a trading company or smaller mill and the seller is an intermediate trader or smaller producer, frequently have contracts that do not account for sampling uncertainty, creating disputes that the technical reality does not justify.
Who Controls the Sampling Infrastructure Controls the Number
At Port Hedland, the mechanical sampling infrastructure at the major terminals is owned and operated by the terminal operators — in most cases, entities closely affiliated with the major miners. The sampling is conducted according to the operators' established procedures, calibrated against their equipment. Independent verification of sampler performance requires access that third parties cannot always obtain freely.
This is not an allegation of systematic manipulation. It is a structural observation: the entity that controls the sampling infrastructure has practical influence over the sampling process, regardless of their intentions. Buyers who want independent verification of loading quality need to either appoint a third-party inspector with access to verify the sampling procedure or negotiate for split samples that the buyer's laboratory can independently analyze.
The gap between what a quality certificate states and what an independently conducted sampling exercise would reveal on the same cargo is, in most trades, small and commercially immaterial. In some trades, particularly when the cargo is on the boundary of specification, that gap determines whether there is a valid quality claim or not.