Yep, and as Brangdon points out, inverse square laws apply. I'd figure the best non-nuke emp might make it to the horizon if popped at ground level, maybe 12 miles if they really pump the local field up and aren't in close proximity to any big conductors (including underground). The gauss level is what we would measure I reckon just to get a good basis. Figure it'd have to be about 100 times more voltage than your standard lightning bolt, so we're probably looking at a 10 to the 7th or 8th power voltage level with a high spiking square wave type leading edge. That'd generate enough harmonics to beat the slew rate on most solid state junctions within that range.

In contrast, it isn't all about the amount of voltage, but the type. Like I was saying before, when you get the frequency component up to ionizing levels, it can compound the effects. Propogational loss is diminished in this case (the inverse square law doesn't work quite the same, but still is a factor). Popping a nuke at 40 miles up with a 10 to the 11th or 12th order (we're talking a really good hydrogen bomb thermo-nuclear detonation, probably 15 megaton or larger) voltage level would probably effect at least a 200 mile radius on the ground, likely a whole lot more if the pulse finds it's way into the grid and surges down the line a lot further. The more x-ray and gamma ray generation, the more effective it will be at slamming all those transistors into oblivion (greater penetration of obstructions).

BTW, humans aren't immune to the emp. I believe that at 5,000 rem the central nervous system blows, and our switch gets abruptly flipped off. I can't remember what the conversion is from eV to rem, but it is not a small number.