D of earth = 12756km, and assuming a sphere (my physics prof would be proud), I'm computing that 730km covers a ~6.6 degree arc. Also assuming a 100m underground tunnel, the distance would be 21m shorter than its above-ground measurement. Someone needs to recheck my math, though.
According Arethuza's comment below, at least one of the labs is 1400 meters below ground. Also you need to calculate the straight-line distance between the two points, not the distance over the surface. Although for calculating the difference that's probably a good enough approximation.
So if both stations are 1400 deep and your calculation is correct, the difference should be 14x21=300 meters. Which is a lot more than the 15m effect they found, which suggests they thought about this.
They claim the geodesic measurements are done to within 0.6m, and the majority of that uncertainty is introduced by the need to triangulate down the tunnel to the experiment.
