I’ve been writing a lot recently about the historic transition from coal to gas-fired power generation ongoing in the United States (see here, here, and here). Water is almost always front and center in these stories (see here and here), as the large amounts of water consumed in hydraulic fracturing, or “fracking,” to unlock natural gas trapped in shale formations has brought the water-energy nexus to the fore. 

So I felt a bit embarrassed recently when, in a friendly email debate about the environmental merits and demerits of shale gas, I was asked whether or not I was concerned about the millions of gallons of water consumed to frack each of the thousands of shale gas wells now dotting America. I realized I couldn’t really answer that question!

In reporting on the shale gas boom, I’ve been guilty of writing about the “millions of gallons per well” or the “billions of gallons of water consumed annually,” and leaving it at that. That sure sounds like a lot of water. Look at all those “millions” and “billions,” right?! But is that really a lot of water? Compared to what? Where’s the context for me or our readers to interpret what that means?

I frequently chide people writing about energy to try to find a way of translating their facts and figures into terms people can actually understand, since nobody knows what the heck a kilowatt-hour or a ton of oil equivalent or quadrillion BTUs really is anyway. The same goes for water, so I'd better follow my own advice!

So in this edition of Friday Energy Facts, we’ll try to tackle this question: how much water does fracking for shale gas really consume?

 

Part 1: A Nationwide Estimate 

To get a nationwide picture, here's a back of the envelope estimate that should give us a good idea of the orders of magnitude involved… 

There were 27,000 new gas wells completed in the US in 2011, according to the US Energy Information Administration (EIA). Shale gas wells represented virtually all of the increase in gas production from 2010 to 2011, so let's just assume for simplicity that these were all shale wells that were hydraulically fractured (rather than any conventional wells). So we’ll assume approximately 27,000 shale gas wells drilled in 2011, which was the peak year to date in the U.S. for shale gas well completions (shale exploration slowed in 2012 as gas prices plummeted).

Water consumption for hydraulic fracturing or fracking of shale gas wells (source: FracFocus.org)Then for simplicity again, let's assume each well consumes 5 million gallons of water per well on average for the fracturing and completion of the well. The graphic at right, which shows the distribution of water consumption per frack job reported via the voluntary industry reporting database FracFocus.org (raw data available here). As you can see, the average here appears to be around 2 million gallons per frack job, but (a) some wells are fracked multiple times and (b) there may be some systemic downward bias inherent in the voluntary nature of the FracFocus.org data set. So 5 million gallons per looks like a good solid average estimate.

Given those assumptions, all shale gas wells completed in 2011 across the United States consumed on the order of 135 billion gallons of water. 

That sure sounds like a lot! But let's put it into context because I have no idea what 135 billion gallons of water means, do you? 

All freshwater water withdrawals (surface and groundwater) totaled about 127,750 billion gallons in 2005 (see the infographic below from Lawrence Livermore National Labs, noting the different units [MGal/day] used there). Lets just assume 2005 is a fair baseline for total annual water withdrawals in the United States (remember we’re just looking for an order of magnitude estimate). Given that baseline, that means water consumed to frack all U.S. shale gas wells in 2011 represents on the order of 0.1 percent of total U.S. freshwater withdrawals. 

Water Consumption and Withdrawals in the United States by Use

So far though we’ve been talking about withdrawals of fresh water, or water removed from surface water bodies and groundwater. That differs from actual fresh water consumption, since a lot of that water, such as water used for power plant cooling, is discharged back into waterways without any contamination or treated in wastewater treatment plants and discharged. Since most of the water used in fracking for shale is actually consumed – about 80 percent of the fracking water remains stuck in the shale deposit, while about 20 percent flows back up the well as contaminated wastewater, which is typically disposed of in deep injection wells. So looking at consumption is probably a more relevant/fair comparison. 

Using the same infographic for 2005 as our baseline, total U.S. annual freshwater actually consumed (e.g. either evaporated or contaminated and stored) came to 43,800 billion gallons. With that as our baseline, total estimated 2011 water consumption for all shale wells completed that year represents about 0.3 percent of total U.S. freshwater consumption.

That's all a rough order of magnitude estimate, but it indicates that shale wells are not as significant a consumer of freshwater as I would have suspected prior to running these numbers. By far the biggest sources of water consumption in the United States remains agriculture, which consumes on the order of 32,850 billion gallons of water annually, or more than 243 times more water than fracking for shale gas. 

[Update: As another point of comparison, golf courses in the United States consume about 0.5 percent of all freswhater used in the country, according to the Professional Golf Association. Tip of the hat to Robert Wilson for pointing this out on Twitter]

Those figures are for 2011 levels of shale gas exploration and may certainly grow over time if shale gas expands in the U.S. Or it may fall if efforts to conserve and reduce freshwater usage continue, as I suspect they will incrementally over time. 

Efforts are now underway to commercialize recycling technology to reclaim the portion of the water that flows back to the surface as flowback water. That could reduce net freshwater consumption by about 20 percent (the rest of the water generally remains in the wells and can’t be recycled). Several companies are also working on waterless fracking methods that use propane-based substances or other materials that then flow back and are recaptured as a gas at the surface and reused. I'm not very familiar with this technology (some of readers could chime in here in the comments), but it's another possible way to reduce water usage or even eliminate it.

Summary: All shale gas wells drilled and completed in the United States in 2011 consumed on the order of 135 billion gallons of water, equivalent to about 0.3 percent of total U.S. freshwater consumption.

 

Part 2: Comparing Shale Gas to Other Fuels

Here’s one other way to look at this: how much water does shale gas consume per unit of energy produced, and how does this compare to other energy sources? Coal mining and processing consumes a lot of water, for example, so how does shale compare to other energy sources?

Luckily, a 2010 paper by Erik Mielke, Laura Diaz Anadon, and Vankatesh Narayanamurti of Harvard’s Belfer Center has data on exactly that question (no need for back of the envelope here!). Here’s what they had to say:

Water Consumption For Extraction and Processing by FuelThe recent shale gas transformation of the U.S. natural gas industry has also focused attention on the water-energy nexus, although the water consumption for the production of shale gas appears to be lower (0.6 to 1.8 gal/MMBtu) than that for other fossil fuels (1 to 8 gal/MMBtu for coal mining and washing, and 1 to 62 gal/MMBtu for U.S. onshore oil production). The increased role of shale gas in the U.S. energy sector could result in reduced water consumption (Chart ES-1). The water used for releasing the gas (hydraulic fracturing), however, has to be carefully managed at a local level. Concerns about potential contamination of freshwater supplies with hydrofracking fluids also need to be addressed. Natural gas-fired combined cycle power plants (CCGT) also have some of the lowest consumption of water per unit of electricity generated, helped by the relatively high thermal efficiency of CCGT plants (Chart ES-2). 

A couple points here: not only does shale gas extraction consume less water per unit of energy provided as coal or oil, combined cycle gas-fired power plants currently offer the most efficient way to turn fossil fuels into electricity. A pulverized coal-fired power plant will consume about 30-50 percent more fuel than an efficient combined cycle gas plant to produce an equal amount of electricity.

As a result, if shale gas displaces coal in the electric power sector, as has been occurring in recent years, then total water consumption per unit of electricity provided will actually decrease – by a lot. Assuming values from the middle of the ranges reported by the Belfer Center paper, and assuming a coal plant consumes 30 percent more fuel than a combined cycle gas plant, water use per kWh could fall by on the order of 80 percent.

Summary: Shale gas consumes about 0.6-1.8 gallons of water per million BTUs of energy produced. If shale gas is used to generate electricity at a combined cycle gas plant and displace coal-fired power, the quantity of water consumed per unit of electricity generated could fall by on the order of 80 percent.


Part 3: What About Texas? Shale Gas in Arid Regions 

The Belfer paper’s point is well taken though that water consumption is a local concern really. There’s plenty of water in my native Pacific Northwest for example, but what about down in Texas or up in North Dakota where shale plays are being actively developed?

Here we’ll turn to one last paper, a 2012 article in Environmental Science and Technology by by Jean-Philippe Nicot and Bridget Scanlon of University of Texas at Austin.  Their paper notes that as of 2011, total annual water consumption for fracking in the Barnett Shale, the largest play in Texas, is equal to about 9% of the annual water consumption of the city of Dallas, for comparison. They also report that total water use for all shale gas wells in Texas amounted to below 1% of all freshwater withdrawals in the state, although again, "local impacts vary with water availability and competing demand." 

In other words, even in drought-stricken Texas, presumably the most arid of all regions home to shale gas development, shale wells currently represent a fairly small use of water, relative to other activities—cities and agriculture being the largest sources of true consumption and power plants being another large source of withdrawals, although water used for power plant cooling is subsequently discharged back into waterways. 

The paper projects that water use in the Texas plays will grow three-fold by 2020 if shale production expands. So that would start to become a more significant consumer of water, equal to a medium-sized city perhaps. But still less than 3% of all Texas water withdrawals. The paper also notes that usage may shift to brackish water to reduce freshwater needs in the future.

Summary: All shale gas wells drilled and completed in Texas in 2011 amounted to less than 1 percent of all water withdrawals in the state of Texas. That figure could grow roughly three-fold by 2020 as shale production rises, although other developments could reduce the amount of freshwater consumed per well.


Part 4: Location, Location, Location - Water Consumption is a Local Issue

[Note this is an update with additional info prompted by the excellent comments by Bob Meinetz and Matthew Lewis below]

As mentioned previously though, like politics, all water consumption is really a local issue. So let's drill down from the state level even further.

Dimmit County TexasThe New York Times reports here on concerns about water consumption due to fracking in arid Dimmit County, Texas (location seen at right) and neighboring counties, all part of the Eagle Ford shale gas play. According to the Times. "in 2011, nearly a quarter of the water used in Dimmit County went toward fracking ... that the figure will rise to about a third in a few years."

Groundwater withdrawals from the local Carrizo-Wilcox Aquifer used for fracking exceeds by one-third the amount of water recharged in the aquifer b annual precipitation and other factors, according to the Times

Clearly, depending on the source of water used for fracking, the local availability of water, and the competing uses in the region, water demands for fracking can be far from trivial.

As Matthew Lewis writes in the comments below, "in a state like Texas, where demand for water typically exceeds supply, an increase in demand of 1% is not trivial as it instantly throws the proposed use into conflict with other users."

The flip side is also true in regions where water is more plentiful. 

In the Marcellus shale gas region encompassing parts of Pennsylvania, Ohio, and New York, rainfall and surface water is far more abundant than in Texas's gas producing regions. The graphics below compare average monthly precipitation totals in Pittsburgh, PA in the Marcellus region and Carrizo Springs, TX in the Eagle Ford region. As illustrated below, monthly average rainfall in the driest months in Pittsburgh is on par with the wettest months in Carrizo Springs.

Monthly average precipitation in Pittsburgh, PA

Monthly average precipitation in Pittsburgh, PA

 
Monthly average precipitation in Carrizo Springs, TX
Monthly average precipitation in Carrizo Springs, TX
 
In contrast with the situation in drought-prone Texas, Will West, a pro-gas blogger at GoMarcellusShale.com (and apparently a resident of the Marcellus region), offers a view on what water consumption for fracking in the Marcellus region looks like:

Let's take an actual example. A drilling unit near me covers 351 acres, with two drilling pads. ... One pad has 1 horizontal bore, the other has 3. Let's assume ... 3 million gallons [of water will be used to frack] each of the 4 legs. To produce the required 12 million gallons would require rainfall of 1.26 inches on 351 acres. Even in the driest month, October, with an average monthly rainfall in Pittsburgh of 2.25 inches, this means that the water required for fracking all legs would be replaced in just 17 days of rain on just this unit.

Of course, the drillers are not going to use rainfall as a water source. They'll probably take water from the creek. I couldn't find the flow rate of our creek but for one of similar size I found 80 cubic feet/second.  The total 12 million gallons could be taken from the creek in less than 6 hours, if it were taken all at once.  In other words, every day the creek carries 4 times as much water  as needed to frack all 4 legs.

Summary: Like politics, water consumption is a local issue. Fracking presents a major source of water consumption in arid locales like Dimmit County, Texas in the Eagle Ford shale region, where fracking represents on the order of one-quarter of the entire county's water consumption. In contrast, in the more rainy Marcellus shale region of Pennsylvania, Ohio, and New York, the water needs for an entire fracking operation represent about 17 days of average local rainfall in even the driest months of the year.


Conclusions


So what do you think? Is freshwater consumption a show-stopper for shale gas? Where does this weigh in the range of environmental pros and cons involved in shale gas production? Please drop a comment below…