PaveKeep Sign in
Knowledge base

PCI and IRI: what each one measures, and why you need both

Two numbers describe the same road and disagree with each other, and both are right. Knowing which one answers which question is the difference between a budget submission that stands up and one that comes back for more evidence.

A council officer reporting road condition will meet two acronyms almost immediately. PCI is a 0–100 score built from what a surveyor or a camera can see. IRI is a measurement of how rough the ride is. They are not two versions of the same thing, and a road can score well on one and badly on the other.

What IRI measures

Roughness is the deviation of a road surface from its intended longitudinal profile. Austroads describes it as the parameter most widely used to rate condition at a network level, because of its effect on vehicle dynamics and therefore on vehicle operating costs, driver comfort and the dynamic load the pavement takes (Austroads, Guide to Pavement Technology Part 5, 2025).

It is not measured by feel. A survey vehicle carrying an inertial laser profilometer records the longitudinal profile of both wheelpaths, and the response of a hypothetical vehicle — a quarter-car simulation travelling at 80 km/h — is computed against that profile. The index that falls out is expressed in metres of simulated suspension movement per kilometre travelled: m/km. Current methods capture surface irregularities with wavelengths between 0.5 m and 50 m.

The method is deliberately international. It was defined in 1986 through the World Bank's International Road Roughness Experiment and published as World Bank Technical Papers 45 and 46, precisely so that two agencies with different equipment could produce comparable numbers. ASTM E1926 specifies the computation from a measured profile.

For Australian practice there is one number to report. Austroads has determined that at network level roughness is always reported as the International Roughness Index of the measurement lane — Lane IRIqc, usually just IRI (m/km) — under Austroads test method AG:AM/T001-16. It is the average of the two single-wheelpath values, and results are normally processed at 100 m intervals and reported to no more than two decimal places.

Austroads is unambiguous that this is now the only reporting unit. Its Guide to Asset Management Part 5B: Roughness states that the former unit — NAASRA Roughness Meter counts — "is no longer considered appropriate and all roughness results should now be reported exclusively in terms of IRI".

If your historic data is in NAASRA counts

Many councils hold years of roughness data in NAASRA Roughness Meter counts per kilometre, from the response-type devices that preceded profilometers. That data is not lost. Austroads gives the correlation:

NAASRA counts/km = 26.49 × (lane-IRI) − 1.27

Austroads' own worked example of an intervention level shows the two side by side: a maintenance response of surface regulation for lengths with roughness above IRI 4.2 m/km (110 NRM counts/km).

Two things follow. The old series can be placed on the same axis as the new one. And the conversion is a correlation, not an identity — so a step in a trend at the point where a council changed method is a change of instrument, not a change in the road. If your condition history does not record which method produced each survey, you cannot tell those two apart. That is worth fixing before it matters.

What PCI measures

PCI is a visual-distress index. A surveyor or an automated system records what kinds of distress are present, how severe they are and how much of the surface they cover, and those combine into a 0–100 score where 100 is a new surface. The international methodology is ASTM D6433.

Australian local-government practice adds something useful. The LGAM Pavement Condition Assessment Manual — derived from Bundaberg Regional Council's own manual — describes sealed-road condition with two indices rather than one:

Both include roughness, which is why the two families are related rather than independent. Each 10 m of road is scored, and the segment score is the average. The manual publishes the weightings, which differ for urban, rural and short urban (under 150 m) segments — in a rural segment roughness carries far more of the PCI than it does in a short urban one.

That difference matters more than it looks. A PCI is not a universal quantity. It is the output of a weighting somebody chose, and two councils using different weightings produce numbers that should not be compared without saying so.

Why a council needs both

They answer different questions.

A road with fine cracking and no deformation yet can ride perfectly well and be two years from needing far more than a reseal. A road with no visible distress can ride badly because of settlement under it. Running one index alone means one of those two roads is invisible to you.

What this means for a report

Three habits make condition figures survive review:

  1. State which index, which method, and which year. "PCI 62" is not a fact until it says whose weighting produced it.
  2. State coverage beside every average. An average over the roads that were surveyed is not an average over the network — unless you say which, it reads as the second.
  3. Record the method with the survey. Vendor, equipment, rating method. The NAASRA correlation above is exactly why.

How PaveKeep handles this

PaveKeep scores each surface type by the method that fits how it fails, and every figure it produces can be traced back to the survey it came from, the setting that shaped it and the arithmetic that produced it. Condition over time shows how the network has actually changed between surveys rather than a single snapshot.

What it does not yet do: survey records do not carry the vendor, equipment or rating method, so a step caused by a change of method is visible but not explained. That is recorded as an open item rather than left for a council to discover.

PaveKeep is built for small and regional councils and is not yet in use at any council.

Sources