Speed through the water matters not one whit.
It does to what SD is trying to determine.
Speed over the bottom matters with regards to when you're going to get somewhere. But it doesn't matter to the boat or the engines. For any given rpm they work just as hard going downstream as they do going upstream if current is the only factor. Wind has a different effect.
SD is trying to determine his optimum engine setting for efficiency through the water, which means coming up with the best combination of rpm, fuel consumption, and boat speed. Speed over the ground is useless for calculating this because it will vary, and perhaps vary widely with the current. But the boat doesn't see this variation--- it just knows what it has to do to move through the water. Which way or how fast the whole body of water it's moving in is going is irrelevant to the boat.
If SD can run the tests at slack current or use timed runs over a measured distance in both directions in a current and then take the average that will work, of course. Or as psneeld said, if the current is slow enough the GPS measurement of speed over the bottom will be close enough for SD's calculations.
But where he boats, I'm guessing currents can perhaps get up into the double digits in speed. They can around here and he's much farther north with bigger tide changes. Slack current sometimes only lasts 15 minutes or so, which is usually not enough time to make the different runs that are needed to get the results he's after. So under these conditions, speed through the water is what you want to use because then it doesn't matter what the current is doing as far as gathering the data SD needs to gather.
Once he's come up wth the optimum power setting for his boat, it's up to him if he wants to exceed this going up-current so he can get where he wants to go by x-time, or back off the power going down-current to burn even less fuel. But he needs a baseline, and for that he needs to know what the boat is doing going through the water, not over the bottom.