I have tried soil moisture sensors before, but not for an extended period of time (in my garden). You seem to be saying that a DC current will cause corrosion over a time, and effect the reading? But AC will not? Reversing the polarity reduceses the corrosion?
If that's the case, could we use two output pins to provide ground, and +5v to the voltage divider. Set them in the forward direction for 1/100 of a second, take our reading from the center tap. Switch the two outputs to provide a reverse voltage for a calculated time, then wait for 100 seconds, and do it again ? Now the probe is not seeing DC, it is seeing AC. Just a thought.