What is the cavity standing wave dielectric loss of the earth?
Down below 20KHz? Rule of thumb: the system-Q for the Earth's lowest resonance is commonly quoted as 8-10. So, any ~11Hz pulse will pass around the Earth over 8 times times before decreasing to 66%. (Imagine passing an AC pulse ten times across an entire a continent-wide power grid, with losses only 33%.)
The Corums (the Tesla-phile engineers behind the Viziv tower) published an engineering analysis in the early 1990s that showed that Tesla's system would still work, even with a Q that low. Their main point: when the system is in operation, the Earth's losses, (the low Q,) represents a constant loss independent of throughput, basically a resistor parallel to the power supply. For a large system, the loss would be ?tens? of megawatts. But if we keep the operating voltage always the same, then these Earth-losses remain constant, no matter how much wattage is being drawn by the distant receivers. Just make sure to sell gigawatt-rates of energy to customers, if your grid losses are constant megawatts.
The losses are analogous to corona-leakage from a power grid the size of a continent. Even with the conventional 60Hz power grid, if nearly everyone stopped using it, the efficiency would plummet to near zero, with megawatts of net corona-loss relative to such reduced throughput wattage. (Also remember, the Tesla system isn't made of thin wires. It's planes. The analog to "cable resistance" apparently is insignificant, when compared to that parallel leakage in the air between the high voltage and ground. But I think this fact is taken from theory only, and has never been verified empirically. Better make sure your receiver acts like a lossless sky-coupling capacitor, not like a giant neon lamp!)
But in the mean time, there was a 1990s Sutton/Spaniol paper from NASA where they built an unusual exotic antenna, and measured Q-values in Shumann-lines with values of many hundreds. Not ten! The greater VLF research community has never accepted this. These results are still maverick/fringe, and like the Corum papers, banished to Tesla conference publications.
Sutton/Spaniol explained the problem: conventional Schumann resonance Q-measurements require long integration times, and unexpectedly, the Earth's sharp resonant lines are not stable in frequency, but wander around over a scale of many seconds. If your measurement-window is on a scale of minutes (in order to pull the natural signal out of the extreme thunderstorm VLF static,) then a wandering-yet-high-Q resonance will be measured as an artifact: a broader low-Q resonance. And also, with minutes of integration time it becomes impossible to even detect such wandering behavior. Sutton/Spaniol's weird neg-resistance neg-inductor antenna-driver was actively altering the space impedance, essentially amplifying the signal as fields, just outside the antenna-coil. An artificial superconductor loop. They obtained relatively huge signals, performed FFT on them, and saw extremely narrow Shumann spikes going way up into the KHz. But it was just two NASA guys with weird results, versus an entire radio-research community, where the two guys say it's been done wrong for decades. I bet their paper is PDF somewhere, with their graphs of VLF spectra above 100Hz, so-called "Black Hole Antenna." (Ripe for hobbyist replications! But probably not if you live in normal em-noisy neighborhoods.)
Note that if Earth's system-Q really was more like 800 rather than 8, then any VLF pulse actually passes around the planet many hundreds of times before significant attenuation. (If Sutton/Spaniol aren't fools, then that would mean that Tesla's claim of insignificant losses might have been genuine after all. Yet that part wasn't even necessary to his success.)
FYI, the Q of a microwave oven chamber is more like 10,000 - 50,000. If the box was made of parallel front-surf mirrors w/99.99% reflection, you could see way deep into the "infinite tunnel," tens of thousands of distant copies before the "tunnel" looked too dark. In a box, you'd see miles of 3D space, with a billion copies of any emitter, all shining upon you. So, a Tesla-style worldwide power-broadcast system may rival the efficiency of a wired grid thousands of miles wide, but its no match for the incredible efficiency of your kitchen's nook-ular dinner-cooker box! Now we just gotta get rid of that hot magnetron. Replace it with nice cold mosfet blocks. (We had one of those in our lab last year: a 300watt plasma lamp, with a pair of 2.5GHz, 150watt "microwave oven" mosfet drivers.)