The local impact is much greater though if the water is pumped out of the ground and then shipped out of state. All of your other examples end up back in the local water table. It's silly to argue that 30 million gallons of water being shipped out of range of the local reservoir is inconsequential -- especially in a state with constant droughts and rapidly depleting reservoirs.

(I've helped manage large aquaculture projects dealing with much larger quantities of water, so I'm not reacting to what seems like a big number. It's striking how differently the EPA treats the diversion of ground water vs. surface water)