(Part of the No Sweat Solutions series.)

Before discussing water savings, we need to define what we mean by “use.” The EPA refers to withdrawal and consumption. Withdrawal is the amount taken from surface water and the water table. Consumption refers to the amount chemically combined with something (so that it is no longer fresh water) or evaporated. Water discarded instead of consumed is referred to as “returns,” because it is supposedly reusable. This does not even approximate the impact of water use.

One example the EPA gives is power plant cooling. The water is withdrawn and used to cool the plant. A little evaporates, and the rest returned (still more or less clean) to the source. This overlooks a certain amount of impact (fish killed during withdrawals, aquatic plant, fungal, and microbial growth encouraged by the change in water temperature), but is basically correct. However, they apply the same logic to water used for irrigation. With very few exceptions, irrigation water “returns” are loaded with fertilizer salts, growth hormones, microbes, and often pesticides and herbicides as well. Even runoff from organic farms usually contains salts from the manure and composts used.

So the proper way to count water is consumption plus polluted returns — in most cases, all withdrawals.

Grist thanks its sponsors. Become one.

The table below translates standard EPA figures into real consumption numbers[1] for the U.S.:

EPA Classification Withdrawals (%) Consumption (%) Withdrawals (millions of Gallons) Consumption (millions of gallons) Consumption + polluted returns = use (millions of gallons) % total use (excludes clean returns)
Irrigation 40% 81% 137,000 76,200 137,000 66.38%
Thermoelectric cooling 39% 4% 131,000 3,500 3,500 1.70%
Industrial + mining 8.2% 5% 27,800 4,500 27,800 13.47%
Domestic 7.5% 6% 25,300 5,900 25,300 12.26%
Commercial 2.4% 1% 8,300 900 8,300 4.02%
public uses and losses (clean returns) 1.6% 5,500 0 0 0.00%
Livestock 1.3% 3% 4,500 3,000 4,500 2.18%

Note that by any classification, the single largest use of water in the U.S. is irrigation — around two thirds of total water consumed or polluted. Internationally, the figures are different: rainy climates use less for irrigation, dry more. At any rate, for all categories we have no-hair-shirt ways to greatly reduce water use, with the exception of public use and direct consumption by livestock.


My previous suggestions on food sustainability would save water by default.

Reader support helps sustain our work. Donate today to keep our climate news free. All donations DOUBLED!

Grist thanks its sponsors. Become one.

To the extent that grazed cattle are substituted for lot cattle (fed irrigated grain and soybeans) you reduce water consumption by about 75 percent, basically to the water the animals drink directly.

Similarly a biodiverse, no-till, low-input approach to growing row crops would save between 30 percent and 50 percent of irrigation water. Soil with good structure holds water better than poor soil; less runoff means more available to plants. Further, good soil structure allows plants to access more of what water is there. So the water goes even further. This also allows better utilization of dissolved nutrients, and if you absolutely need small amounts, to herbicides and pesticides. That means what runoff you get is less polluted; remember, polluting water less is another way to use less. If conventional farming is like hooking soil on heroin, and organic farming is supporting the soil entirely without artificial help, low-input no-till biodiverse farming is like giving the soil an aspirin when needed and letting it have an occasional beer with dinner.

Another huge saving can come from simply growing crops in vaguely appropriate places. I’m not talking about growing native plants only, or even perfectly climate-appropriate imports only. With our current population, we are way past the point where that is possible. But certainly we can avoid certain extremes — like growing cotton in the Arizona Desert. (Cotton is one of the world’s most water-intensive plants in net output ratio to water consumed.)

The combination of rotational grazing, no-till, and avoiding extremes of climate inappropriateness can cut water per unit of output by well over half, when reduced water pollution is taken into consideration.

Convert all less efficient irrigation to low-energy, precise-application micro-sprinkler, drip irrigation, subsurface irrigation, and other ultra-efficient irrigation methods[2]. This can save an additional 33 percent in the U.S., which already uses such means more than most nations — meaning that worldwide, the untapped potential for high-efficiency irrigation is even higher than in the U.S. There is even a kind of bucket-based drip irrigation, which has low capital costs but high labor costs — and thus is economic in poor nations where capital is expensive and labor cheap. (In the long run, we hope this won’t stay the same.)

Lastly, because low-input biodiverse no-till requires drainage in any case, and produces much cleaner runoff than other forms of agriculture, we can capture and recycle that runoff[3], saving another 27 percent. That means we can save around 75 percent of the total water required by no-till agriculture per capita.


In residences, toilet flushing and showering alone can be responsible for about 70 percent of water use. Dual flush toilets like the Caroma, which use different amount of water depending on whether dealing with solids or liquid, can cut flushing use by half or more. Advanced compressed-air toilets, like the Microphor can cut this in half again, but at a cost that takes about seven years to pay for itself at 5.5 percent interest in a typical household. Low-flow showerheads such as the Bricor B100Max can cut bathing costs in half at a price from $30-$75.

Washing machines and dishwashers are also major water users in homes that have them. There are many high-efficiency washers out there these days; the LG WM1814c is an example of a comparatively inexpensive one; the LG Tromm Steam Washer of a pricier one with fancier features. Similarly, you can find a long list of dishwashers that get much better results than the U.S. EnergyStar standard from the U.S. D.O.E.

Sink aerators can cut the flow in bathroom sinks by 60 percent, at cost of $4-$15. This can be combined with a hands-free faucet conversions kits for kitchen and bathrooms sinks at around $50 each. (Kick or knee-pedal conversion are more expensive manual alternatives in the $130-$600 ranges.)

Commercial buildings can gain similar savings. Because commercial bathrooms are used more intensively than homes, they get payback from things like one quart compressed air driven standard toilets, and waterless urinals.

Yards and landscaping can follow many of the principles outlined under row crops — choosing plants suitable to the environment, biodiversity, soil preservation.

Rainwater capture can provide a high percentage of domestic and commercial needs, both at the individual building level (PDF) and the neighborhood scale. Treatment is comparable to that required for well water, though often rain water is higher quality. Separation of blackwater (toilet water and possible garbage disposal water) and greywater (everything else) allows treatment of greywater for reuse. Sewage can be treated on the neighborhood level with living machines, which will have a much easier time processing pure blackwater, with none of the toxins often found in greywater added.

Industrial Water Use

There is a similarly huge potential for water savings at the industrial level: a good example is the computer chip industry, where more efficient filters[4], reduction in output waste combined with recycling[5], and slowing the speed of rinse processes[6] can reduce water consumption by 80 percent or more. There are hundreds, perhaps thousands of techniques that can reduce water use in industry. To name a few: counter-current washing uses waste water from the cleanest process as rinse for the next cleanest, and so on. Some industries can reuse the same water five times. Quite often, dry processes can be substituted for wet ones, such as pigging or super-critical carbon dioxide. On the lower tech, cleaning can start with dry or damp cloths (applied by automated processes, not manually) with rinse steps following. Similarly, spray processes can substitute for bath processes. Other tricks include shaping and sizing vessels in baths to minimize the ratio of rinse water to stuff being rinsed. Mini-reactors can isolate steps requiring ultra-clean or ultra-pure environments from those that can use normal atmosphere and rinse water.

Overall, we can cut water input per unit of production by around three quarters. However, this is not quite enough to make water use sustainable. Population growth would reduce savings to about half. Providing clean drinking and washing water and waste treatment to everyone on the planet instead of letting people die miserably due to lack of it would increase water consumption a bit. (Not a lot; domestic use is not where most water goes.)

Now, if it were not for climate chaos, that would be the end of it; but in the face of it we will probably need to expand irrigation in certain parts of Africa (and as we have seen Australia). It is hard to put a hard number on how much of the savings this will lose. But my intuition is that combining efficiency increases with new demand will give us a 30 percent technically and economically feasible overall reduction in draws on water tables and snow melt. This is unsustainable in at least as many areas as it sustainable, but it takes a bite out of part of the problem. Also, since water is used more efficiently in this scenario, because we squeeze more GDP out of each acre-foot of water, it also allows us to pay more per acre-foot.

And that lets us fill the gap with desalinization. There are well known techniques for desalinating seawater: reverse osmosis driven by electricity and more old-fashioned low-pressure distillation techniques driven by waste heat from various processes. In a renewable scenario, I’d suggest the electricity be produced by wind plants at a time when electrical demand is low (the wind equivalent of off-peak) and that the waste heat be from solar thermal electric plants in the desert. Even with that, desalinated water will cost about twice water from conventional sources (unless major breakthroughs in the technology occur). But if we are using that water four times as efficiently, paying double for it won’t be a great hardship. And if we have already reduced absolute demand by 30 percent, then such desalinated water will be providing around another 30 percent, cutting draws on groundwater and snowmelt in half.


[1] United States Environmental Protection Agency, How We Use Water In These United States. 18/March 2003, United States Environmental Protection Agency, 06/Jul/2005.

[2] Micro-irrigation system (drip + sprinkler) about 5.7% of total irrigated acreage.

Various gravity forms (at 50 percent) are about 43.9 percent of total irrigated acreage.

Other sprinklers irrigate about 51.2 percent of total irrigated acreage.

United States Department of Agriculture National Agricultural Statistics Department, 2003 Farm & Ranch Irrigation Survey (2002 Census of Agriculture| Volume 3, Special Studies, Part 1)(PDF). Nov 2004. United States Department of Agriculture National Agricultural Statistics Department, 28/Oct/2005. p8.

Table 4. Land Irrigated by Method of Water Distribution: 2003 and 1998.

Micro irrigation systems average around 82.5% irrigation efficiency.

Gravity irrigation systems average around 50% irrigation efficiency.

Other sprinkler average around 70% irrigation efficiency.

Michael D. Dukes, Types and Efficiency of Florida Irrigation Systems(PDF), (Note: Data used was from national sources). Dec 2002. University of Florida – Agricultural and Biological Engineering Dept, 28/Oct/2005. p8.

So applying the efficiency numbers from the second source to the acreage in the first, we can calculate that current average irrigation efficiency is around 62 percent. If that average efficiency was upgrade to micro-irrigation levels we would reduce water use for irrigation nationally by an average of one third, internationally by substantially more.

[3] I. Broner, Irrigation: Tailwater Recovery for Surface Irrigation. Crop Series, 4.709. 1998. Colorado State University Cooperative Extension, 17/Sep/2005.

[4] Pacific Northwest Pollution Prevention Resource Center, Topical Reports, “Energy and Water Efficiency for Semiconductor Manufacturing,” Pollution Prevention (P2) Pays – N.C. Division of Pollution Prevention and Environmental Assistance, Feb 2000, Pacific Northwest Pollution Prevention Resource Center, 17/Sep/2005.

[5] Hidetoshi Wakamatsu, Akira Mayuzumi, and Norio Tanaka, Effective Utilization Technology for Ultra Purewater, Chemical Liquids and Waste Materials on Semiconductor Manufacturing Plant (PDF), OKI Technical Review 68, no. 188: Special Edition on the Environment Dec 2001, Oki Industry Co. Ltd – Environment Division, 23/May/2004. pp23 – 27.

[6] Stanford University News Service, Can Computer Chip Makers Reduce Environmental Impact? 5/Jun 1996, Stanford University News Service, 4/Jun/2004.