The industrial meters I have used have used 1000 HZ square wave or 50 HZ sinewave. The wave form is not important neither is its frequency, although it should not be too high else capacitive losses because a problem.
Have you worked out the likely range of the ground conductance? with this value and the correction factor of your probes, you then can work out the likely range of resistance your are dealing with.
I have only dealt with the conductance of de-ionised water which can be a rather good insulator, 2µ Mhos was our spec. (~500 KΩ) limit. drinking water was 600µ Mhos. So from these figures I would say that dry sand was better the 10 MΩ , and wet soil, higher then 1K.
If I was doing this job, I would use a common or garden DVM to measure the volt drop between your probes, and feed them, from say 20V low frequency sinewave, via a series of resistors. providing the voltage source is constant, relative conductivities can be made quickly, and if super accurate measurement are required a bit of maths can compensate for the change of current.
The problem is a high voltage sine wave oscillator (as I guess you want it to be battery powered), The actual waveform is not critical sine /square.. but is would be useful if its level was stabilized. it is extremely likely that you will need some sort of transformer to get from battery voltages to some thing over 20V AC .
lets see what the combined brains of this forum can suggest
Frank