In summary, the manufacture and transport of that one kilogram bottle of Fiji water consumed 26.88 kilograms of water (7.1 gallons) .849 Kilograms of fossil fuel (one litre or .26 gal) and emitted 562 grams of Greenhouse Gases (1.2 pounds).
UPDATE: Due to the incredible response to this post, Pablo obtained more information and has recalculated, and it is not as bad as previously stated but still staggeringly bad, 6.74 times as much water as is in the bottle.
Nearly seven times as much water used to make it than you actually drink. Staggering is an understatement. Click here for original source at ::Triple Pundit
or read portions of it below.
Here is a portion of the original:
I once heard Julia "Butterfly" Hill (everyone's favorite tree-sitting sweetheart) say that it pollutes several times more water to make the plastic bottle than it actually holds. We might as well put that myth to the test while we're at it. Where do we begin? Well, I doubt that Fiji has a booming plastics industry so they probably get the bottles in the form of "Blanks" from China, which are then expanded to their final size and shaped by a process called "stretch blow molding." The total mass of the empty 1 liter bottle is probably around 0.025kg (25g) and it is made from PET (Polyethylene terephthalate) Plastics of this type use around 6.45kg of oil per kg, 294.2kg of water per kg, and result in 3.723kg of greenhouse gas emissions per kg. So, with a quick check (200kg/kg x 0.025kg = 5kg of water) we find that Butterfly is indeed correct. Based on my calculations a bottle that holds 1 liter requires 5 liters of water in its manufacturing process (this includes power plant cooling water).
Let's take a look at the transportation aspect to see what the total ecological impact of an imported bottle of water might be. A container vessel uses 9g of fuel per tkm (that's metric tons carried x distance traveled), 80g of water per tkm, and releases 17g of GHGs per tkm. The distance from China to Fiji is 8,000km, which gives us exactly 0.25tkm ( (0.025kg / 1t/1000kg) x 8,000km = 1.0tkm). So, 2.3g of fossil fuels, 20g of water, and 4.3g of GHGs per bottle delivered to Fiji from China.
Now let's look at the trip to the US. The distance from Fiji to San Francisco is 8,700km. But this time the bottles will be full, so they will have a mass of 1.025kg each. This gives us a much larger value of 9.8tkm ( (1.025kg / 1t/1000kg) x 8,700km = 8.9tkm) which I will round up to 9tkm. So, 81g of fossil fuels, 720g of water, and 153g of GHGs per bottle delivered to the US from Fiji.
Since the fossil fuels end up being accounted for in the GHG emissions I'll ignore those values for now. The total amount of water used to produce and deliver one bottle of imported water is 6.74kg (5kg + 20g + 1kg + 720g)! And the amount of GHGs released amount to 250g (93g + 4.3g + 153g), or 0.25kg, or 0.00025 tons.