IRI vs. PCI: what each number is actually telling you

If you've sat through a resurfacing budget meeting, you've probably watched someone point at a map and say "this segment's a 65" without specifying which 65. That ambiguity causes real arguments, because IRI and PCI are not the same measurement and they don't always move together.
IRI measures how the road rides, not how it looks
The International Roughness Index comes from a profilometer run, usually a vehicle with laser sensors reading the longitudinal profile of the wheel path at highway speed. The output is a single number, typically reported in meters per kilometer or inches per mile, that feeds a quarter-car simulation model. A low IRI means a smooth ride. A high IRI means the car bounces, and above a certain threshold it starts costing fuel economy and suspension wear for everyone who drives it.
Here's the part that trips people up: IRI has nothing to do with cracking. A segment can be laced with alligator cracking and still post a decent IRI if the cracks haven't opened up into potholes or caused rutting yet. The profilometer doesn't see distress. It sees vertical movement.
PCI measures what's wrong with the surface
The Pavement Condition Index, as defined in ASTM D6433, is a 0-100 score built from a manual distress survey: a rater walks or drives the sample units, logs every crack type, severity, and extent, plus rutting, raveling, patching, and the rest of the distress catalog, then runs it through deduct curves to land on a single index. A PCI of 85 means low-severity, low-extent distress. A PCI of 40 means something is actively failing, whether or not a driver feels it at 55 mph.
That's the key difference people miss in handoff meetings: PCI is a condition assessment of the pavement structure and surface. IRI is a ride-quality measurement. One is about what's happening to the asphalt. The other is about what's happening to the vehicle on top of it.
Where the two numbers disagree
A few situations come up often enough that any shop running both metrics has seen them:
Fresh overlay on a failing base. IRI reads smooth right after paving, but the PCI of the layer underneath, or the reflective cracking that shows up within a season or two, tells you the fix didn't address the structural problem. The smooth ride buys a year, maybe two, before the cracking telegraphs back through.
Early-stage alligator cracking with no rutting. This hits PCI hard well before it touches IRI. Classic load-related fatigue cracking in the wheel path can rack up distress points long before the surface deforms enough to register as rough.
Rough ride with a clean surface. Some corridors ride badly because of base settlement, frost heave, or utility cut patches that are structurally fine but geometrically uneven, without the kind of cracking or raveling that drags a PCI score down.
Each metric answers a different question, and a program that only tracks one is missing half the picture. A roughness vs. distress rating comparison on the same segment, done honestly, usually explains a surprising amount of why last year's "worst road" complaint calls didn't match the "worst road" PCI list.
Why this matters for the annual list
Most agencies still build the PCI side of this from windshield surveys: a crew drives the network, logs distress by eye, and that pass happens once every few years because it's slow and the rating is only as consistent as whoever's behind the wheel that week. IRI collection via profilometer van is usually tighter and more frequent, which is part of why the two datasets drift apart in age and in what they're measuring.
Road Surface Condition covers the distress half of that gap. It reads the whole network from VHR imagery once a year and classifies where cracking, ravelling, and patching are concentrated by segment, giving budget season a distress picture that doesn't depend on whatever a crew happened to drive past that season. Run it alongside your IRI dataset and both halves of the condition picture stay current instead of one aging ahead of the other. If your resurfacing priority list is leaning on a PCI pass that's older than your last IRI run, that's worth closing before the next budget cycle starts.