Will the Keystone XL Pipeline Significantly Increase Carbon Emissions? The Numbers Behind the KXL Debate [UPDATED]

Will the Keystone XL pipeline lead to a significant increase in global carbon emissions? That’s the key question at the center of the debate over the controversial tar sands pipeline. 

The State Department released an updated final environmental impact statement (EIS) for Keystone XL Friday that buoyed the hopes of pipeline advocates and opponents alike and gave Secretary of State John Kerry and President Barack Obama considerable leeway to make their final decision on the merits of the project.

The environmental report’s headline conclusion was that approval or denial of the Keystone XL pipeline would not have a “significant impact” on the rate of Canadian tar sands extraction. That’s because, under what the report’s authors consider the most likely assumptions, oil from the tar sands will find its way to market with or without the Keystone project. 

“Approval or denial of any one crude oil transport project, including the proposed [Keystone XL pipeline], remains unlikely to significantly impact the rate of extraction in the oil sands, or the continued demand for heavy crude oil at refineries in the U.S.,” the report says, reiterating the central findings of the an earlier draft EIS released in March 2013.

Pipeline advocates, including the American Petroleum Institute, Republican members of Congress, and TransCanada, the owners of the $5.4 billion project, all seized on that headline finding to call for immediate approval of the long-delayed pipeline. 

“The case for Keystone XL, in our view, both pre and post this report, is as strong as ever,” Russ Girling, chief executive of TransCanada told reporters Friday.

But a deeper look at the report finds that the Keystone pipeline would not be without impact: depending on assumptions, State’s final EIS concludes that, if built, the 830,000 barrel per day pipeline could cause an extra 1.3 million to 27.4 million metric tons of carbon dioxide (CO₂) emissions each year. That’s the equivalent of putting an extra 250,000 to 5.5 million cars on the road, estimates the Washington Post’s Brad Plumer. That’s more than enough carbon to give climate advocates ammunition in the debate as well.

“The report concluded that in a scenario where we take climate change seriously and regulate climate pollution, this pipeline will indeed have a ‘significant impact’ on climate change,” said a statement from Bill McKibben, founder of 350.org, the environmental organization that has spearheaded opposition to the pipeline. “So now we’ll find out if that’s the world Barack Obama and John Kerry want. This report gives President Obama everything he needs in order to block this project.”

President Obama declared in June that he would sign off on Keystone XL only if the pipeline “does not significantly exacerbate the climate problem.” 

The final EIS released Friday gives the president enough ammunition to go either way on the project. According to White House spokesman Matt Lehrich, the EIS “includes a range of estimates of the project’s climate impacts, and that information will now need to be closely evaluated by Secretary Kerry and other relevant agency heads in the weeks ahead.” 

The State Department will now have 30 days to collect public comments on the EIS and 90 days for other executive agencies, including the Environmental Protection Agency and Departments of Energy and Defense, to weigh in. 

As with so much of the debate surrounding the controversial pipeline, what you believe — and what President Obama ultimately decides — about the Keystone XL pipeline depends on what assumptions you make.

To get at the numbers behind the Keystone XL debate, we are republishing the detailed analysis below, originally published after release of State’s draft EIS in March 2013. This analysis unveils the key issues and assumptions behind the divergent views on the pipeline’s climate impact.

Which scenario do you believe is most likely? Should President Obama and Secretary Kerry reject or approve the pipeline? How do you read the conclusions of the final EIS? Weigh in with your comments…

-Updated, February 3, 2014-

Keystone XL, the proposed oil pipeline connecting Canada’s Alberta tar sands region to the refineries and markets of the American Gulf Coast, has become the front lines in the battle between climate campaigners and the fossil fuel industry.

The pipeline would carry up to 830,000 barrels per day of heavy tar sands oil, enough to supply more than 4 percent of U.S. oil demand. With that much oil, money, and carbon at stake, it is no surprise the pipeline has become a key political flashpoint on both sides of the U.S.-Canada border.

On one side of the debate, climate activists charge that the pipeline would carry some of the world’s dirtiest and environmentally destructive crude and result in up to 181 million additional metric tons of carbon dioxide (CO₂) annually. That’s as much carbon as the annual tailpipe emissions of 37.7 million cars, according to Oil Change International, an environmental NGO. That’s more than the total number of cars registered in California, Oregon, Washington, Florida, Michigan and New York combined.

On the other side, a draft environmental impact statement (EIS) commissioned by the U.S. State Department claims that completion of the pipeline is “unlikely to significantly impact the rate of extraction in the oil sands” and would result in as little as 70,000 metric tons to as much as 5.3 million tons of additional CO₂, depending on whether alternative pipelines would be built to replace a foiled Keystone. Even if environmentalists managed to block all new pipelines out of Alberta, State’s EIS concluded that it would reduce tar sands production by just 2 to 4 percent by 2030. (The State Department must ultimately bless the international pipeline project as “in the national interest” before construction can commence, placing the agency in the crosshairs of environmental activists, who have blasted State’s draft EIS.)

Wait! Did you catch that?

Estimates of the climate impact of the Keystone XL project differ by up to four orders of magnitude from the State Department’s low-end assessment of as little as 70,000 tons to Oil Change International’s charge that Keystone will drive up emissions by 181 million tons annually.

So which is it? Is Keystone XL “ground zero” in the effort to stave off climate disaster, or is it a sideshow with negligible impact on future emissions?

Where you come down on that question hinges centrally on vastly different answers to the question: what happens if Keystone XL isn’t built? Will oil flow through alternative pipelines or rail lines as State’s EIS believes? Or without Keystone, will transport bottlenecks effectively “shut in” tar sands oil, as environmental activists hope?

Beyond this central question, one must also consider how global oil markets respond to new flows of tar sands oil and what types of crude Keystone’s oil will ultimately displace?

In this post, I’ll unpack the assumptions and the economics of the Keystone debate, describing several possible scenarios that cross the range of claims in this debate — and then I’ll leave it to you to decide which set of assumptions you find most believable. Please add your views to the comments below.

Strap in for a tour through the nest of issues behind the Keystone XL debate…

1. Background

We will begin with a little background on tar sands oil and the economics of Canadian oil. If you’re familiar with these topics, skip right ahead to the discussion of scenarios below…

How Dirty is Tar Sands Oil?

First, it’s important to understand that tar sands oil isn’t your conventional crude that flows out of the ground through a normal well. Tar sands oil is derived from “bitumen,” an extremely dense and viscous form of petroleum that is actually trapped in layers of loose sand or sandstone. Raw bitumen is a sticky, nearly solid, tar-like substance at room temperature – hence the name “tar sands oil.”

To get at the bitumen, shallow bitumen deposits are strip-mined out of the ground while deeper deposits are recovered “in situ” by injecting a stream of hot gas into an underground bitumen reservoir, cooking the heavy oil and making it fluid enough to pump out of the ground. In situ production is less environmentally damaging than surface strip mining, but more energy intensive. Currently, mining accounts for just over half of Canadian tar sands production, although with 80 percent of recoverable bitumen reserves located too deep for surface mining, in situ production is expected to grow in the future.

Once out of the ground, the tar-like raw bitumen is too sticky to be transported via pipeline at all. Bitumen is therefore either upgraded into a synthetic crude oil (aka “syncrude”) or diluted with other liquid hydrocarbons into a mixture called “diluted bitumen” or “dilbit.” Since Keystone will carry dilbit, we’ll focus on that form in this post.

Once the dilbit is transported to a refinery, the heavy oil has to be “upgraded” into a lighter hydrocarbon before being refined into products like gasoline and diesel fuel. That process requires still more energy and releases more carbon.

The result of all this: in its journey from the oil fields to your gas tank and on to combustion in your vehicle’s engine, gasoline derived from tar sands produces 10 to 30 percent more greenhouse gas emissions than the average gallon of U.S. gasoline, according to a March 2013 Congressional Research Service (CRS) survey of the available literature.

As the graphic above illustrates, the “upstream” part of the tar sands “well-to-wheels” pathway related to production and transport of tar sands oil is about 50 to 150 percent worse than many other crudes.

That said, about 80 percent of the emissions associated with burning a gallon of gasoline are due to the combustion of the gasoline itself. And tar sands gasoline contains just as much carbon once it gets to your tank as a normal gallon of gas.

Why Canada’s Oil Industry Loves Pipelines

Now for some economics…

The West Texas Intermediate (WTI) price index reflects the typical value of high quality crude oil (or in industry-speak, “light sweet crude”) delivered to the main North American trading hub at Cushing, Oklahoma. WTI represents the underlying commodity of the New York Mercantile Exchange’s oil futures contracts. Since 2011, WTI crude has hovered around $95 per barrel of oil.

Since diluted bitumen is a “heavy oil” of lower quality than WTI crude, it is typically priced based on a different index, known as Western Canada Select (WCS). From 2005 to 2012, WCS oil traded at a discount of $10 to $20 per barrel relative to WTI, reflecting both the lower quality of the oil and the cost of shipping the oil from Canada to the North American hub reflected in the WTI index.

That means that if WTI crude is worth $95 per barrel, WCS oil – aka tar sands oil – is worth only $75 to $85 per barrel.

In the first three months of 2013, however, the WTI-WCS spread jumped up to about $32 per barrel, driving down the value of tar sands oil to about $62 per barrel.

Understanding what’s going on here is key to understanding why the Canadian oil industry is so keen on building Keystone.

Alberta’s oil fields are landlocked. Getting that oil to market and on to consumers in the U.S. or overseas therefore requires somehow shipping it overland.

Pipelines are the cheapest way to ship oil overland, at about $8.00 to $9.50 per barrel, according to State Department estimates.

In recent years, with both tar sands production in Western Canada and shale oil in the Bakken formation of North Dakota and Montana on the rise, the pipelines used to move oil through the Midwestern United States and on to refineries and export terminals in the Gulf Coast are now at capacity.

Without any free space in existing pipelines, oil producers in the Bakken and Canada’s tar sands fields have turned to rail to ship their oil to market, and that raises costs.

The State Department’s evaluation of Keystone XL estimates that shipping oil by rail to the Gulf Coast costs about $15.50 per barrel, although Washington Post reporter Brad Plumer quotes oil producers who say shipping by rail costs as much as $30 per barrel.

Bottom Line: Shipping oil by rail is roughly $6 to $22 per barrel more expensive than shipping via pipeline. And that increase in transport costs is what accounts for the declining value of tar sands oil relative to the WTI crude oil price index. At a capacity of 830,000 barrels per day, even if Keystone saves just $6 per barrel in transport costs relative to rail, that’s more than $1.8 billion in increased profits for tar sands producers each year.

2. What Happens if Keystone Isn’t Built? Three Scenarios at the Heart of the Keystone Debate

Now we’ve got our background. But as I indicated in the intro, the true climate impact of Keystone XL hinges on what you assume will happen if the pipeline is not built.

There are three main scenarios to consider… 

Scenario 1: If Not Keystone, Another Pipeline

First up: if climate campaigners succeed in blocking Keystone XL, the industry may end up constructing an alternative pipeline to get their oil to market. According to Inside Climate News, the oil industry has announced the intention to build 10,000 miles of pipelines at a cost of $40 billion over the next five years in an effort to bring Canada’s tar sands crude to global markets.

Graphic Source: InsideClimateNews.org, see full article here for more.

The main alternative to Keystone XL is the Enbridge Northern Gateway project, which would bring oil west to the port of Kitimat in British Columbia. From there, it would be loaded in oil tankers for shipment to consumers in California or Asia.

With a shorter distance from Alberta to deepwater ports along the British Columbia coast, State’s EIS estimates that export costs to Asia are actually comparable in cost to transporting oil to the Gulf Coast via Keystone.  Shipping oil by tanker is remarkably cheap, costing just $2 to $3 per barrel from British Columbia to Asia, for example, and less to California. 

However, the Northern Gateway project just suffered a major setback, with the provincial government of British Columbia expressing official opposition to the proposed pipeline last week, saying the project fails to address the province’s environmental concerns. The Canadian federal government ultimately has authority over the pipeline and says the project’s review process is still “ongoing.”

Alternative routes to market via the east coast would be longer and slightly more costly. That could depress the amount of oil produced at the margins in Canada, though not by much – perhaps up to $2 per barrel.

In this scenario then, building Keystone XL would avoid sending tar sands oil to market via longer alternative routes, increasing the value of tar sands oil by up to $2 per barrel, an approximately 3 percent increase over current WCS prices.

As the graphic below illustrates, the resulting impact on tar sands production depends on how responsive producers are to this price change, or what economists call “the price elasticity of supply,” or the ratio of the percentage change in supply for any given percentage change in price.

Traditionally, most oil producers are pretty “inelastic,” meaning they do not respond much to changes in the value of oil. However, for a few producers with production costs close to the current value of oil – these producers are said to be “on the margin” – even small price changes could be material.

Thus, I present a range of three possible price elasticities of supply in each of these scenarios:

1. A very low supply elasticity of 0.05, consistent with estimates of average global oil supply elasticity and appropriate if price changes are small enough that they are unlikely to effect whether or not tar sands extraction projects are profitable.
2. A supply elasticity of 0.2, consistent with a case where a few tar sands producers are “on the margin” and would not invest in production unless prices are high enough, while most other producers are not responsive to price changes.
3. A supply elasticity 1.0, indicating that tar sands producers are perfectly responsive to changes in the value of tar sands oil – every 1 percent change in price leads to a 1 percent change in production. This would be consistent with a case where many tar sands projects are on the margin, and every increase in the value of oil brings new projects into profitability.

In addition, to determine the final impact of an increase in tar sands production on CO₂ emissions, we must make an estimate as to how global oil markets respond to this new increase in supply.

Traditionally, markets would respond to a supply shock like this by re-equilibrating at a new level of demand. That is, an increase in supply from tar sands producers would lower global oil prices, all else equal. That would spur both an increase in demand and a reduction in production by other oil suppliers. The new equilibrium will depend on the elasticities of both global oil supply and demand. With the elasticity of oil supply and demand both well below 1.0, the result will be that net oil consumption increases by less than the increase in supply coming from tar sands oil – some portion of the tar sands oil will be incremental to global oil consumption, while the remainder will simply displace other oil from expensive producers on the margin elsewhere in the world.

I present two such case in each of these scenarios, one where 50 percent of tar sands production causes an incremental increase in global oil consumption (corresponding to an estimated global oil supply elasticity of 0.025) and another where 80 percent of tar sands production is incremental (corresponding to a global oil supply elasticity of 0.1, more consistent with the highest price elasticities observed from OPEC producers in recent years). In both cases, the long-run elasticity of demand is assumed to be -0.1 (about the middle of this range of estimates for 10 nations). 

Alternatively, one might consider global oil markets to be supply constrained – that is, global oil consumers want to consume as much oil as producers can produce. Constraints in oil supplies are holding back consumption, in which case any new output from tar sands oil will lead to an equivalent increase in consumption. In this case, tar sands output is 100 percent incremental.

If the elasticity of tar sands supply is very inelastic, building Keystone XL may increase tar sands production by just 2,800 barrels per day (bbl/day). The ultimate climate impacts would be negligible, resulting in less than 1 million metric tons of additional CO₂ emissions annually. This is probably the most likely outcome given the relatively small change in the value of tar sands oil associated with this scenario. 

If tar sands producers are somehow more responsive to this price change, the result may be an increase up to 9 million metric tons, depending on how global oil markets respond to an increase in tar sands output. That would be the equivalent of the annual CO₂ emissions of up to 3 average sized coal-fired power plants.

  
(Click to enlarge)

Bottom line: If Keystone XL isn’t built but tar sands oil finds its way to market via other pipeline routes, the climate impacts of blocking Keystone would be fairly small. Building Keystone would increase the value of tar sands oil by only up to $2 per barrel, a 3 percent increase. This is the scenario behind the State Department’s (very) low end estimate that blocking Keystone would save just 70,000 metric tons of CO2 annually. Assuming such a small price increase has little impact on tar sands production, the climate impacts would be negligible. If that small increase in value somehow makes the difference between profitability and not for some tar sands projects, Keystone’s climate impact may be somewhat larger, on the scale of the annual CO₂ emissions of one to three coal-fired power plants.

Scenario 2: Rail Saves the Tar Sands

Under this scenario, environmentalists in the U.S. and Canada are successful in blocking all (or nearly all) new pipeline proposals. In this case, building Keystone XL would relieve existing bottlenecks in pipeline capacity and allow tar sands producers to avoid shipping their diluted bitumen via more costly rail lines. The value of tar sands oil would rise relative to WTI by about $6 to $22 per barrel, an increase of 10 to 36 percent relative to current WCS prices.

That rising value could trigger an increase in tar sands production, relative to the case where Keystone XL is not built. How much of an increase depends again on the elasticity of supply for tar sands producers, and that depends on the marginal costs of producing oil from the tar sands. 

Actual production costs are a closely guarded industry secret. But State Department’s assessment of Keystone estimates that after accounting for the discount between WTI and tar sands (WCS), it takes WTI prices in the range of $51-61 per barrel for new in situ production and $66-76 per barrel for new mining projects. Mining projects that upgrade the tar sands oil to synthetic crude require a WTI price of $86-96 to break even. The graphic below displays these “break even” ranges, denominated in WTI prices.

If rail transport adds $6 to $22 per barrel to the cost of transporting tar sands to market, those break-even prices would rise by a corresponding amount.

With WTI oil now trading around $95, you can see that, as State’s EIS concludes, “particularly in the shorter term, the most expensive oil sands projects … are economically challenged” by this increase in transport costs. While most in situ projects could still progress at these prices, some mining projects would be on the margin, and synthetic crude upgrader projects would be unprofitable. 

State stresses that “in the shorter term” part however, because most industry and government forecasts project steadily rising global oil prices, with WTI prices steadily above $100 per barrel by the 2016-2018 time frame. At those prices, the marginal cost of rail shipment won’t affect the economics of too many projects.

To represent this long-term uncertainty, I’ve once again modeled the impacts of higher rail shipment costs under a range of different elasticities of supply. 

(Click to Enlarge)

As the graphic above illustrates, if shipping oil via the Keystone pipeline instead of rail doesn’t impact the break-even price of most projects – e.g. tar sands producers have a low elasticity of supply of 0.05 – the project’s construction may result in 8,300 to 30,500 barrels per day of increased tar sands oil production.  The result would be an increase of only 1 to 5 million metric tons of CO₂ annually, or about the an emissions of a single coal-fired power plant. This could be the case if global oil prices remain high, keeping the value of tar sands

If instead, shipping via cheaper pipeline makes the difference between profitability or not for a set of tar sands producers and the elasticity of supply is about 0.2, building the pipeline could increase bitumen production by up to 122,000 barrels per day, a 7 percent increase over current Canadian tar sands production. That in turn would result in 3 to 21 million tons of additional CO₂ annually, or the equivalent output of a handful of new coal plants.

Finally, if tar sands producers are perfectly elastic (elasticity of 1.0), they might increase output by up to 610,000 barrels per day, a more than one-third increase over current production levels. That would drive up CO₂ emissions by anywhere from 14 million tons to 103 million tons annually, equivalent to the emissions of up to 29 coal-fired power plants or about 2 percent of total U.S. energy-related CO₂ emissions.

As you can see, even in the case of perfect elasticity, tar sands production increases in this scenario by less than the total capacity of the Keystone XL pipeline. The elasticity of tar sands supply would have to be very large, on the order of 1.4 to 4.6 before the price difference from rail versus pipeline shipment could cause tar sands production to increase by the full capacity of the Keystone pipeline, or 830,000 barrels per day.

Note that this scenario is the one most consistent with the core scenarios in the State Department’s EIS, which assumes that rail lines could expand to ship oil to market if Keystone is blocked. State assumes that tar sands production will fall by 40,000 to 170,000 barrels per day under this scenario, consistent with a supply elasticity of about 0.2 to 1.0 if rail costs just $6.00 more per barrel than pipeline, or 0.07 to 0.3 if rail costs $22 more per barrel. 

Bottom line: Under a scenario where climate campaigners blocked all proposed pipelines out of the tar sands region and rail shipments expand to carry the oil instead, building Keystone XL could increase the value of tar sands oil by $6 to $22 per barrel, or up to 36 percent relative to current WCS prices. If global oil prices are fairly low, that price increase may make the difference between profitability and not for a number of tar sands projects. In this case, blocking Keystone may prevent as much as 103 million tons of CO₂ annually, or up to 2 percent of total U.S. energy-related CO₂ emissions. In contrast, if global oil prices rise, building Keystone or not may not make much of a difference to the economics of tar sands oil. In this case, blocking Keystone could prevent as little as 1 to 5 million tons of CO₂ annually, about the output of a single coal-fired power plant.

Scenario 3: Tar Sands Hit a Bottleneck

The final scenario corresponds to a world in which there simply isn’t enough transport capacity to ship oil out of the tar sands region to market. All pipelines are blocked, and rail lines can’t expand fast enough to carry the tar sands oil instead.

In this case, blocking Keystone XL has the real chance to “shut in” tar sands production. In other words, if you believe there simply isn’t enough rail capacity to ship oil to market and you believe that climate campaigners can kill all proposed pipelines, then blocking Keystone is the first stage in truly constraining the growth of the Canadian tar sands. This scenario is what environmental campaigners have in mind when they target Keystone XL. 

In this case, it’s entirely fair to assume blocking Keystone means keeping at least 830,000 barrels per day of oil out of the global market. That would mean blocking Keystone XL could prevent 70 to 140 million tons of CO₂ emissions annually — equivalent to the output of 20 to 40 typical coal-fired power plants, or as much as 2.6 percent of U.S. energy-related CO₂ emissions.

(As an aside, I must note that even in this scenario, I can’t quite reproduce the emissions figure in the April 2013 Oil Change International report frequently cited by 350.org and other opponents of the pipeline. That report claims the pipeline will emit “at least 181 million metric tons of carbon dioxide equivalent (CO₂e) each year.” Their estimate assumes 100 percent of the oil carried in the Keystone XL pipeline is new tar sands production and 100 percent of that oil leads to incremental global consumption. Furthermore, their estimate relies on a well-to-wheels CO₂ estimate for tar sands oil of 598 kg or 1,318.4 pounds per barrel. That is 29 percent greater than the middle-range estimate from CRS’s review of the available academic literature and 16 percent greater than the highest estimate CRS could find in the literature. Oil Change International claims that prior research has not fully accounted for emissions associated with petroluem coak, a coal-like by-product of upgrading tar sands oil at refineries and a substitute for coal in power plants.) [Update: Lorne Stockman, author of hte Oil Change International report on Keystone provides additional explanation of their methods in the comments below.]

(Click to Enlarge)

So can rail lines really scale up to ultimately handle a couple million barrels of new tar sands oil shipments per year?

In many ways, the Keystone debate hinges on this question.

State’s environmental and market review of the pipeline concretely answers, yes. The U.S. Environmental Protection Agency has challenged the State Department over this assumption, as have many environmentalists.

State’s case in a nutshell is that rail shipments out of the Bakken region in North Dakota and Montana are similarly constrained by a lack of available pipeline capacity and have successfully turned to rail to ship oil to market. If Canadian producers can expand rail capacity as quickly as the Bakken region has in recent years, it would reach 800,000 barrels of capacity – aka about the same as Keystone – by the end of 2015.

But can rail freight keep pace with the Canadian oil industry’s plans to more than double current tar sands production to 3.2 million barrels per day by 2020 and further to as high as 5 million barrels by 2030?

Maybe. Maybe not. Expanding rail freight capacity at that rate may run into real bottlenecks. On the other hand, the entire coal industry ships its product to market via rail, barge and other means (for obvious reasons, it’s impossible to ship coal via pipeline!). If tar sands oil is a valuable global commodity, there’s a strong incentive for it to find a way to market. [Update: Well what do you know?! Coal can be transported by pipeline, as Robert Joshi informs me on Twitter.]

[Update: Geoff Styles adds this bit of context in the comments below]

“At a conservative 600 barrels per tank car load (limited by weight for heavy crude, vs. a full 750 bbl for lighter oil) the entire volume of Keystone would equate to 500,000 tank car loads per year.  That sounds monumental until you notice that US rail car loads of coal declined by more than that volume from 2011 to 2012.  It would roughly double 2012 rail shipments of “petroleum”, which includes LPG and other products as well as crude oil.”

(For more on whether rail can save the tar sands, see these excellent articles by Brad Plumer, Elana Schor, and Steven Mufson and Juliet Eilperin)

Bottom line: if rail freight capacity can’t expand fast enough to make up for the loss of Keystone XL, blocking the pipeline project could keep at least 830,000 barrels per day out of the global market. That could have large climate impacts, preventing 70 to 140 million tons of CO₂ emissions annually — equivalent to the output of 20 to 40 typical coal-fired power plants, or as much as 2.6 percent of U.S. energy-related CO₂ emissions.

So what do you think? Please comment below about which scenario you find most plausible, and why you either support or oppose the Keystone XL pipeline…