A plain-language guide to EMC shielding effectiveness: how absorption loss and reflection loss combine (SE = A + R + B), why near-field sources behave differently from far-field plane waves, and why real enclosures are usually limited by apertures rather than the solid barrier.
A metal barrier blocks an electromagnetic field two ways at once, and they work for completely different reasons. Absorption happens because the field induces eddy currents as it penetrates the metal, and those currents dissipate its energy as heat — the deeper the barrier, relative to the material's skin depth, the more gets absorbed. Reflection happens purely from an impedance mismatch at the surface: a good conductor looks like a near-short-circuit to an electromagnetic wave, so most of the wave simply bounces off before it ever gets inside. Schelkunoff's classical treatment adds the two together in decibels, plus a small correction for energy that reflects back and forth inside a barrier too thin to absorb it fully: SE = A + R + B.
Absorption loss is where shielding theory and skin-effect theory are the same physics — A(dB) = 8.686·(t/δ), where δ is exactly the classical skin depth (see this site's Skin Depth guide). Every skin depth of thickness costs about 8.7dB, so absorption rises quickly with frequency (skin depth shrinks as f increases) and with thickness. Copper and aluminum, despite being non-magnetic, are excellent absorbers at RF frequencies simply because they're such good conductors that their skin depth is tiny.
Reflection loss compares the wave's own impedance against the shield's — the bigger the mismatch, the more bounces off. Far from a source, any field settles into a plane wave with a fixed impedance of 377Ω (free space). But close to a source — inside roughly a sixth of a wavelength — the field impedance depends on what kind of source is radiating it. A high-voltage, low-current source (like an unterminated trace or antenna) looks like an electric dipole, with a high near-field impedance that's easy to reflect. A low-voltage, high-current source (like a current loop, busbar, or cable carrying ripple current) looks like a magnetic dipole, with a low near-field impedance that reflects poorly — which is exactly why magnetic fields are the hard case for shielding at short range, and why absorption (not reflection) tends to dominate the design for low-frequency magnetic sources.