Why is physical optics an approximate, not a precise technique?

Question

Consider the problem of known electromagnetic source in free space near a perfect electric conductor (PEC) object. Finding the total electromagnetic field in this situation can be done through solving Maxwell's equations numerically. At high frequencies, this becomes very computationally expensive in terms of time and memory. Approximate techniques like Physical Optics (PO) are then used.

PO in the above situation is based on finding the electric currents on the surface of the PEC object, getting the exact scattered field due to these currents, and adding it to the original fields radiated by the source.

From reading of various resources and references, it is always mentioned that PO has two inherently introduced approximations:

1- ignoring the diffraction effect, which is complemented afterwards through methods like PTD (physical theory of diffraction).

2- the PO current on the PEC surface is itself approximate.

And here comes my question: the justification of point 2 above. Why is the PO current considered approximate, although it is calculated from a true boundary condition at the PEC surface?

Answer 1

The method called physical optics as originated by Macdonald ignores the induced surface currents in the shadow region and calculate the induced surface currents only on the directly illuminated side on the surface of the scatterer, the surface current on the shadow side is assumed to be zero. Despite this obvious non-physicality the method is very successful in various reflector antenna problems. To some extent the approximation involved in the original PO method can be improved by assuming not a perfectly conducting but a more general impedance surface. See 1 and references therein.

[1] Umul: Physical optics theory for the diffraction of waves by impedance surfaces, JOSA Vol. 28, No. 2 / February 2011