0.3% of water used for irrigation would be totally fine.
The heat of formation of water is 13 MJ/kg. If we can convert hydrogen to power at 50% HHV efficiency, this means 1000 TWh of storage would require hydrogen from about 1/4 of a cubic kilometer of water. This is small compared to the water used in the US for irrigation (about 100 cubic kilometers per year). (It would be somewhat higher than this due to the need to reject waste heat from the bottoming cycle of the CC plant, and at the electrolysers and hydrogen compressors.)
A key point, though, is that nuclear uses water too, for cooling. For every MJ of power from a reactor, 2 MJ of waste heat is produced, and this heat goes out the cooling towers (there are dry cooling solutions, but they make nuclear even more expensive, less efficient, and less competitive). The heat of evaporation of water is 2.26 MJ/m^3, so 1000 TWh of power from nuclear would require evaporating 1.6 cubic kilometers of cooling water.
But it's worse than that. Hydrogen only has to handle the last slice of power demand (that 500 TWh is about 1/9th the US annual power consumption; your 3 weeks out of 52 would be even less), but nuclear would have to produce all of it, or at least a major chunk (and the reactors have to operate at high capacity factor or the cost of their power becomes even more ludicrous). So the water use of nuclear would actually be an order of magnitude higher than that.
The water argument is an argument for wind and solar (and hydrogen), not an argument for nuclear.
The heat of formation of water is 13 MJ/kg. If we can convert hydrogen to power at 50% HHV efficiency, this means 1000 TWh of storage would require hydrogen from about 1/4 of a cubic kilometer of water. This is small compared to the water used in the US for irrigation (about 100 cubic kilometers per year). (It would be somewhat higher than this due to the need to reject waste heat from the bottoming cycle of the CC plant, and at the electrolysers and hydrogen compressors.)
A key point, though, is that nuclear uses water too, for cooling. For every MJ of power from a reactor, 2 MJ of waste heat is produced, and this heat goes out the cooling towers (there are dry cooling solutions, but they make nuclear even more expensive, less efficient, and less competitive). The heat of evaporation of water is 2.26 MJ/m^3, so 1000 TWh of power from nuclear would require evaporating 1.6 cubic kilometers of cooling water.
But it's worse than that. Hydrogen only has to handle the last slice of power demand (that 500 TWh is about 1/9th the US annual power consumption; your 3 weeks out of 52 would be even less), but nuclear would have to produce all of it, or at least a major chunk (and the reactors have to operate at high capacity factor or the cost of their power becomes even more ludicrous). So the water use of nuclear would actually be an order of magnitude higher than that.
The water argument is an argument for wind and solar (and hydrogen), not an argument for nuclear.