This could be really useful on any frequency low enough to have a fairly stable phase delay between all the stations involved.
for example, if everyone was able to receive wwvb, (or nist corrected GPS) they could open loop record a slice of spectrum that included wwvb and the desired signal, (or record the spectrum of interest with a GPS timestamp).
The next step could take as long as needed, as it is recorded and no longer happening in real time. take an FFT of all the recordings, pick an arbitrary recording as the "master" then adjust the timing alignment until the WWVB signals in all of the recordings are in phase. apply a timing differential based on each stations distance from the transmitter you are trying to receive, and the tiny signals of the desired signals should add when all FFTs are averaged.
as far as implementation, there are probably a thousand ways, but if one is willing to record raw spectrum, and process it later, rather than in real time, this becomes much easier.
if anyone has any additions, subtractions, or criticisms for this idea, I would love to hear them