If the soil is generating a voltage, it will also be generating a current. If this current is comparable to the meter’s probe current in resistance mode, you’ll get significantly wrong resistance readings. Essentially it’s like trying to measure a resistor whilst it’s in a circuit and with the circuit switched on.

The solution is to measure the dynamic resistance: apply two different test currents and measure the voltage drop for each. The resistance is the difference between the two voltages divided by the difference in the two currents. The Mooshimeter would be fine for measuring the voltage drop, but you’ll need an external current source.