Near field, the impedance can be very different (at DC, the impedance of a static magnetic field is zero and the impedance of a static electric field is infinite, but it also doesn't propagate). Veeeery roughly, the ratio of impedances is the ratio of attenuation (power in versus power radiated), or the Q*, or ratio of dimensions to the wavelength, or stuff like that.
*If you can actually calculate or measure the radiation resistance independently from other losses, this is actually a less rough, and more useful, matter.
Analogously, the impedance of a section of transmission line (whether a wide pad, a thin trace, or a component lead, or...) varies from characteristic impedance below the cutoff frequency (say under 1/4 wave or so) to inductive or capacitive a proportional amount away. The only reason we approximate e.g. traces as inductive (for low system impedances) or capacitive (for high system impedances) is because the line impedance is sufficiently different not to mind the precise difference; but if you're looking at it over a very wide bandwidth, even the transmission line stuff becomes relevant, and it's important to remember its origin.
Tim