The MOP, equation 5-18b calls for p02, which the MOP describes as the "Percent passing the 0.02 mm sieve". I assume that this is the percent passing for the subgrade.
The percent passing parameter in the equation IS taken from the subgrade layer inputs. However, there is a typo in the IRI equation in the MOP. The IRI model actually uses the percent passing of the #4 sieve and not the p02 sieve of the first subgrade layer in the pavement structure.
Initially, the IRI model used the p02 sieve in the original NCHRP 1-37A project but was adjusted in subsequent work and the documentation for the MOP wasn’t updated. I wasn’t able to find a specific reference documenting when this change was implemented into the software but has been included since at least 2015 and likely before that.
The corrected equations for the site factor are shown below.
$$IRI=InitialIRI + C1\times TotalRutting + C2\times TotalCracking + C3\times TotalTransverse + C4\times SiteFactor$$
$$SiteFactor=(Frost + Swell)\times Age ^{1.5}$$
$$Frost=Ln[(Precipitation+1)\times(FreezingIndex+1)\times P_4 ]$$
$$Swell=Ln[(Precipitation+1)\times(PlasticityIndex+1)\times P_{200} ]$$
SiteFactor = Site factor
Age = Pavement age (years)
FreezingIndex = Mean annual Freezing index, °F-days.
PlasticityIndex = Plasticity Index
Precipitation=Mean annual precipitation(in.)
P4 = Percent subgrade material passing No. 4 sieve
P200 = Percent subgrade material passing No. 200 sieve