I agree that there are other water cycles involved that should be treated independently. I guess, from my view, this seems like a fairly straightforward evaluat

I agree that there are other water cycles involved that should be treated independently. I guess, from my view, this seems like a fairly straightforward evaluation (under ideal conditions): if the ship (or space station, or colony, or whatever) has N gallons of water per astronaut, astronauts require 1 gallon of water per day (according to the article), and 98% of water used by astronauts can be recovered, then the rundown time would be d days, where d satisfies: 0.98^d == 1/N Past that point, you have less than 1 gallon of water per astronaut, which entails rationing and sub-optimal performance etc. Rearranging the equation we can solve for d as: d = -Log2(N)/Log2(0.98) = 34.31 Log2(N) If you bring 2 gallons of potable water per astronaut, for example, after ~34 days you will be down to slightly more than 1 gallon, and after 35 days you will be in suboptimal conditions as you will have less than 1 gallon of potable water remaining from this system alone. In order for d to be 100 you would need a bit less than 8 gallons of potable water per astronaut at mission outset under this model. Edit: I suppose the much briefer way of putting this is that you can't count time until the water runs out; you have to count time until the water is insufficient to meet requirements (in this case, 1gal/day/astronaut).