I’ve never seen anything like this: an academic study that assumes that a privately held power company will continuously violate state and federal environmental laws.
That’s just one of several surprising flaws in a recent paper by researchers from Duke University and the University of Calgary. The study purports to show that, under certain assumptions, exporting coal to Korea will reduce the amount of CO2 emitted per megawatt-hour of power produced globally. Yet the study is based on premises so absurd that they render the authors’ conclusions meaningless.
The study, titled “Environmental Implications of United States Coal Exports: A Comparative Life Cycle Assessment of Future Power System Scenarios”, compares two alternative scenarios for coal consumption and exports in the Pacific Northwest.
- A “business as usual” scenario, in which two aging coal-fired power plants in Washington and Oregon would be retrofitted to comply with state and federal air pollution laws, and then continue to burn coal from the Powder River Basin (PRB) long into the future.
- An “export” scenario, in which the two Northwest coal-fired plants would be promptly shut down and replaced by natural gas-fired power plants, while the PRB coal that would have been burned in the Northwest instead would be shipped to South Korea.
But designing the scenarios this way is literally absurd, for at least 3 separate reasons.
- The authors’ “business as usual” scenario is flat-out illegal.
The study describes its baseline scenario—under which retrofitted coal plants in Centralia and Boardman continue to burn coal indefinitely—as “business as usual.”
Yet this “business as usual” scenario violates both state and federal law. There is no room here for interpretation or nuance; no ifs, ands, or buts. According to SB 5769, signed by Washington’s governor in 2011, one of Centralia’s two generating units must emit no more greenhouse gases per megawatt-hour than a modern natural gas-fired power plant by 2020. By 2025, the second boiler must follow suit. Soon after the law was passed, the state revised its federal Clean Air Act State Implementation Plan (SIP) to reflect the change in emissions. The EPA approved the SIP, so the Centralia closure plan now has the force of federal law.
The case is similar for Boardman: Oregon’s Department of Environmental Quality, as well as federal regulators, have approved Portland General Electric’s (PGE) decision to close the Boardman coal plant by 2020. It’s literally preposterous to assume that illegally retrofitting Centralia and Boardman to burn coal indefinitely represents “business as usual.”
- The “business as usual” scenario makes absolutely no economic sense.
PGE has stated time and again that retrofitting Boardman, and continuing to operate the plant past 2020, would be uneconomic. As The Oregonian reported, the utility lobbied to close the plant early:
Based on its analysis of carbon and natural gas prices…PGE maintains that a 2020 shutdown would be the low-cost, least-risk plan for utility ratepayers and shareholders.
The Oregonian later described cost savings from early closure of the plant:
It allows PGE to install $60 million to $90 million of pollution controls instead of the $500 million that would have been required for PGE to operate the plant through at least 2040. Given that cost differential, PGE concluded that closing the plant early was less expensive.
So not only is the “business as usual” scenario illegal, it is actually a crazy business proposition that was explicitly repudiated by Boardman’s owner.
- There is no logical connection whatsoever between the two scenarios.
Out of the huge range of possibilities the authors could have studied, the authors happened to choose two in which the exact same amount of coal, from the exact same mines, will be shipped to the Northwest. Yet there’s literally no reason, either logical or economic, that this core assumption should have guided both scenarios. In addition, there’s literally no reason to link the closure of Boardman and Centralia with coal exports. None. They’re just not linked.
Find this article interesting? Support more research like this with a gift!
So, the authors settled on an illegal and uneconomic “business as usual” scenario, and designed a second scenario that arbitrarily linked coal exports with the closure of the coal plants at Centralia and Boardman. Those core assumptions—illegal, uneconomic, and illogical as they are—rig the entire study in favor of exports.
What benefits, again?
What’s more, the purported “benefits” of coal exports largely vanish under scrutiny. For example, the study found that low-energy PRB coal would actually produce slightly more CO2 per megawatt-hour than the typical mix of coal that is currently burned in Korea. It also found that shipping coal from Northwest ports to Korea could boost emissions from ocean-going vessels. So when the authors looked at the impact of shipping a ton of coal overseas, they found that coal exports significantly boost emissions:
If instead of estimating emissions per MWh we had chosen to look at emissions per tonne of PRB coal combusted, the MPP export scenario would not be advantageous…In this case, combusting a tonne of PRB coal in South Korea creates 29%−36% more CO2e emissions than burning the same tonne in the U.S…As the U.S. substitutes coal for shale gas in its electricity generation, the best choice to reduce air emissions would of course be to leave the unused U.S. coal in the ground.
Given these findings, the only way to find a climate “benefit” from coal exports is by assuming that US mining is inevitable and that exports would somehow be linked with the closure of coal plants at Boardman and Centralia—a linkage that, as we’ve seen, doesn’t even pass the laugh test.
The Verdict: Unnecessary Errors
There are a number of other factual and logical flaws in the study. But they pale in comparison to the deep and troubling problems with the core assumptions that inform their scenarios and guide their analysis.
What’s so disappointing is that the authors easily could have avoided these errors. With just a few minutes on Google, or a brief conversation with literally anyone who’s familiar with the legal or economic issues of coal power in the Northwest, the authors could have designed a “business as usual” scenario that wasn’t preposterous. But instead of choosing caution, accuracy, and relevance, the authors chose a flashy headline and meaningless results. It’s a disappointment; the issue is important, and it deserves more careful treatment.
Do you suppose they were paid to do a study as to how to continue to burn coal? So they did, and may have been aware of their poor arguments.
Hard to know. A note at the end of the study says that “The authors declare no competing financial interest.” And they acknowledge financial support from the Center for Climate and Energy Decision Making funded by the National Science Foundation.
Perhaps the most charitable view is that this is the sort of thing that happens when engineers get out of their depth. 🙂
There is no reason to think they were paid by the coal industry. The abstract of the article indicates that “If instead PRB coal were to replace natural gas or nuclear generation in South Korea, green house gas emissions per unit of electricity generated would increase”. And the discussion section says “as the US substitutes coal for shale gas in its electricity generation, the best choice to reduce air emissions would of course be to leave the unused U.S. coal in the ground”.
I believe there is linkage between the closure of american coal fired generation systems and increase efforts to export coal. The widespread closures and stalled permitting of coal generation has stimulated the desire to export.
Asian countries are building new generation and will burn coal, either US coal or coal sourced elsewhere.
When we ban exports, we shoot ourselves in the foot economically and accomplish nothing environmentally. This is not good policy.
Devoting more effort toward influencing fuel source choices and pollution controls in Asia would be much more productive.
I believe you are mostly wrong.
The decline in domestic coal consumption may have increased the level of *desperation* that US coal executives feel about exports. But the decision to seek out exports was NOT spurred by low US demand. It was spurred by high Asian prices.
For a while — 2010 through early 2013 — exporting PRB coal was extremely lucrative, because Pacific Rim prices were high. The state of the US market was mostly irrelevant: because international prices were high and international sales were profitable, US producers would have sought out exports no matter what was happening domestically.
Regardless, there is absolutely reason to believe that exports are CAUSING the closure of Centralia or Boardman’s coal plants. That, in essence, what the study presumes: namely, that you can attribute the emissions reductions from converting Boardman and Centralia to gas to the decision to export the same coal overseas.
Finally, one of the few ways we have to influence fuels source choices in Asia is not to further flood the seaborne coal market with more coal. Hawking cheap coal, while gently admonishing China not to burn it, would do nothing.
Preventing exports of coal will only hurt the US. It does nothing to better the environment unless all of the worlds substantial coal reserves are depleated.That wont happen for some time. If all the effort being wasted on stopping boats and trains that carry coal were instead directed to educating the folks burning coal, perhaps some reduction in coal consumption could be attained.
The US coal plants are closing because of policy uncertainty. Utilities are concerned that the cost to meet environmental regulation is or will be higher than cleaner sources of energy.
A good tirade to be sure, but I would argue that it is you, Clark, who are approaching this study on several faulty premises. And as a result, you miss the point entirely.
Firstly, and most importantly, this is an academic paper, not a white paper. It is clear to this reader that the study explored two hypothetical scenarios aimed at answering the question of how bad coal exports are in comparison to the burning of coal domestically (I paste verbatim from the study itself: “In this hypothetical Baseline scenario, Boardman would need to be retrofitted with air emissions controls…” — Did you even read it?). The study is not aimed at answering the question: “What set of scenarios are most economical, most legally sound, and most beneficial to the Northwest?” It is a research study presenting a methodology for examining the impacts of coal exports, nothing more. Hypothetical, repeatable for other such export scenarios, and meant to explore macro-type emissions questions. Morrow Pacific is not active, not looking likely to happen, and we don’t even know where coal it might ship will end up. But I don’t see you attacking the study’s uncertainty in this regard. I recommend removing your policy wonk hat and reading the paper again.
To be fair, you are probably confusing the study’s conclusions (which, if you actually read them, offer several caveats and alternatives) with the cherry-picked media coverage it has garnered. One only needs to read the abstract (which describes emissions increasing under certain conditions!) to realize its conclusions are very different than the headlines suggest. This is called “spin.” I know you must be familiar with this concept.
Also, the study does not forecast emissions “long into the future,” as you claim. Rather, it takes a snapshot in time of two hypothetical scenarios. i.e. “What would happen if tomorrow we exported coal and burned natural gas instead of coal?” Your arguments with regard to the legally-mandated closure of Boardman and Centralia in later years have precisely zero meaning in a hypothetical macro research study, and less than zero meaning in this case, since the study does not include a timeline of emissions.
You also criticize the study for not having a causal link between coal exports and the closure of Boardman and Centralia. Am I missing something, or do events have to be causally linked for someone to study their co-existence? This is certainly your strangest argument.
Lastly, what you call “scrutiny” is explained at length in the paper– you just chose to offer a snippet of information to prove your point– not unlike the work of internet media crafters who slice and dice scientific research into news bites every day.
Funny thing is, DavidK, when started reading the article itself I was actually expecting to find more reason to criticize the coverage of the study than the study itself. I figured that the coverage of the study had trumpeted a few factoids out of context. And yes, I think that some of the coverage missed many of the important nuances and caveats of the study.
But as I read and re-read the study, and then dove into the supplementary material, I became more and more appalled at the assumptions that guided the analysis and that informed the study’s scenarios.
So, about the issue whether all elements of alternative scenarios causal link: no, there doesn’t always have to be a causal link between all elements of an alternative. But in this case, the “benefits” of coal exports can be entirely attributed to the conversion of Boardman and Centralia from coal to gas, which has nothing to do with exports. Calling it an “export” scenario, and comingling the closure of coal plants with the export of coal, produces conclusions that are somewhere between misleading and irrelevant.
Consider, for example, two scenarios that really do compare exports with no exports: one in which Boardman and Centralia are retrofitted and no coal exports take place, and a scenario in which Boardman and Centralia are retrofitted and 8 million tons of additional coal are exported to Korea. In this case, it’s truly an “export” vs. “baseline” scenario, since the only thing that changes is the volume of exports.
In that case, emissions from the retrofits and the combustion of coal domestically drop out, and you’re just left with the effects of exports. It’s easy enough to estimate, at least in broad strokes, what that analysis would show.
The study itself says that:
“it is clear that if the MPP scenario materializes, PRB coal will be the dirtiest fossil fuel available to South Korean power generators in every emissions category except sulfur oxides. These higher emissions are mostly a result of the emissions associated with the transportation phases, but also a consequence of the increased NOx, PM, and CO2 emissions of PRB relative to the higher rank bituminous coal that constitutes 92% of coal currently used in South Korea.” [Emphasis added.]
The supplementary information also gives information on the relative performance of Wyoming PRB vs. the current mix of coal burned in South Korean power plants. And they find that Wyoming PRB coal emits about 9% more CO2 per kWh than the standard mix of coal currently burned in Korea (2.28 tonnes per kWh for Wyoming PRB, compared with 2.09 lbs per kWh for the current bituminous mix). This is from a table on page S9 of the supplementary material, but note that the table is actually mislabeled “lbs/MWh” rather than “lbs/kWh” for CO2 emissions.
So the study itself says that shipping Wyoming PRB coal to South Korea produces higher emissions, both total and per MWh produced, than the existing mix of coal burned in South Korean coal plants. Once you take the closure of Centralia and Boardman out of the picture, exports are clearly worse, by every metric the study uses.
To me, this simple hypothetical shows shows that the study’s so-called “export” scenario only shows lower emissions per megawatt hour because of the closure of Boardman and Centralia’s coal plants.
I tried to come up with an analogy, but couldn’t come up with a good one. But I’ll try this: imagine 2 scenarios, a “baseline” in which ALL coal power plants in the country are retrofitted and continue to burn coal for 20 years, and another in which all coal power plants in the country are converted to natural gas and 8 million tons of coal is shipped to Korea. In that case, the “export” scenario would show huge benefits both for total emissions and to emissions per MWh — but only because of the closure of US coal plants! The situation is similar here, but just on a much smaller scale: a purported result about “exports” really just shows that converting US coal plants to natural gas reduces emissions.
As for the question of time scale, the authors use the 20-year projected life of the MPP project, not a snapshot in time. Feel free to check the supplementary material for confirmation. It’s standard for life cycle analysis to look over the life cycle of a project or process. That’s actually the point of LCA — snapshots in time are meaningless. And for the record, I have no problem with their use of LCA for this purpose, and their model and general approach to evaluating impacts seems defensible enough. It’s the design of the scenarios themselves that renders their conclusions both misleading and meaningless.
Clark, it sounds like the authors were interested in something different than you’re interested in. It sounds like you want to know how dirty it is to export coal (based on the scenario you lay out in which the baseline in both cases is a retrofitted Boardman/Centralia). In that case, you have a basic control/treatment-type experiment, where you can subtract out the baseline and figure out the net emissions effects from exporting coal. That’s fine and simple, and is both derived and reported in the paper!
What the study design tells me is that the authors wanted to follow the same amount of coal from the PRB to Boardman/Centralia and from the PRB to South Korea, to see not only what the life cycle emissions difference is (this would get us to your experimental design), but to take the boundary out a step further and consider what fuels are displaced/replaced in the US and South Korea. Because, I think what you’re failing to realize is that life cycle assessment in this case doesn’t make sense unless you’re replacing energy/capacity that you’ve removed with other energy/capacity.
And there’s really no point in pursuing the line of research you lay out (by itself), because OF COURSE it will be dirtier to burn coal domestically AND export it. And so, to me, this seems like the main reason it’s important to investigate what happens when replacing domestic coal capacity with natural gas capacity. But further, it makes sense because the trend in the US is replacement of coal with gas. The authors state this in their paper. And they also provide a number of (5?) fuel displacement scenarios in South Korea. To me, this is a more complete life cycle assessment than you propose.
And I understand that you have an issue with the accounting of emissions– whether they result from coal-to-gas replacement or South Korean coal plant efficiency— but I have now seen you pull a couple pieces from the study that explain this exactly. And there’s this from the Discussion (!) section: “On an emissions per MWh basis, burning PRB coal in South Korea results in about 10% more CO2e than burning it in the United States. But export of PRB coal actually becomes favorable when widening the boundary of the study to include the rest of South Korea’s energy mix and coal-to-gas fuel replacement in the U.S.” It doesn’t get much clearer than that, Clark.
This begs the question: If we can pull this type of clear information from the paper itself, how is it misleading? Again, I think you’re confusing the media headlines (i.e. coal exports can reduce emissions 21%) with the actual title of the study, which says nothing about coal exports reducing emissions, and simply says it’s going to compare some power system scenarios related to coal exports using LCA. It then poses the question of whether it would be better (purely from an emissions perspective!) for the Northwest to export coal and burn natural gas in its place, or not to bother and just to burn the coal domestically. I, for one, think this is a very interesting hypothetical question to ask, with great value in informing the way we think about energy export projects.
And in regards to the timeline issue, you are not correct—what they did was to use a 20-year lifecycle of the Morrow Pacific project to calculate its total cost for purposes of using the EIOLCA model, but for direct emissions from power plants they used simple emissions/MWh data from various sources. Look at the sources. There is no emissions timeline, only an operational timeline for the export project in order to use the EIOLCA model. It states all this very clearly in the Methods section.
Yes, I think the authors thought did what you said:
“the authors wanted to follow the same amount of coal from the PRB to Boardman/Centralia and from the PRB to South Korea, to see not only what the life cycle emissions difference is (this would get us to your experimental design), but to take the boundary out a step further and consider what fuels are displaced/replaced in the US and South Korea.”
That’s the EXACT problem: they took the boundary out a step, but got it wrong. They assumed that shipping coal to Korea, as you put it, “displaces” coal at Boardman and Centralia. That’s not correct. Because of both legal and economic realities, exporting 8M tons of coal to South Korea does absolutely nothing to the fuel mix at Centralia or Boardman. Those plants are still going to gas.
I’ll check into the timing issues, though.
2 more comments:
First, I’m actually VERY interested in the question of what happens when you expand the analytic boundaries of coal exports one step, or several steps. I could go on at length about many possible legitimate ways they could have thought about that “next step” — some of which I think could *reduce* the calculated climate impacts of coal exports compared with their assumptions, some of which would increase them.
But under NO circumstances would those expanded boundaries include the closure of Boardman and Centralia. Boardman was slated for closure before Morrow Pacific was announced, before ANY of the northwest coal terminals were being publicly discussed. Centralia too was slated for closure before Morrow Pacific was announced. The study boundaries for the “export” scenario strike me as arbitrary and indefensible.
Second, you ask how a study is misleading if it includes this statement: “export of PRB coal actually becomes favorable when widening the boundary of the study to include the rest of South Korea’s energy mix and coal-to-gas fuel replacement in the U.S.”
I find it inherently misleading to describe coal-to-gas fuel replacement at Boardman and Centralia as part of “export of PRB coal.” Words have meanings, and they’ve stretched the meaning of “export of PRB coal” beyond all recognition.
It struck me that this kind of language would never pass muster in other areas of research. “The pain relief benefits of using aspirin for a headache disappear when widening the boundary to include dropping a bowling ball on a patient’s foot.” Abstract: “The aspirin use scenario results in 21 percent increase in patient discomfort…the effects of aspirin use can vary, depending on the boundary of the analysis.” That sort of thing would be absurd in medical research.
Sure, all LCA studies have to draw boundaries. But in this case the author’s boundaries are indefensible: they devised a “baseline” scenario that makes no economic sense and that violates the law, and an “export” scenario that includes two major coal-to-gas projects that are completely unrelated to exports.
Rapidly reducing coal burning worldwide is critical to reduce greenhouse gases and other pollution. Mining less US coal and not increasing exports are obviously part of this. Its not that complicated except to people who want to profit from coal exports.
Don’t know why I couldn’t reply to your latest comment, Clark, but probably just as well as I’m spending way too much time trying to explain this. For what it’s worth, I was inclined to share my perspective because I felt your criticism of the study was unfairly harsh and a bit short-sighted in my opinion. You may be happy to know this is my last post:
There are no definitively wrong assumptions in a hypothetical study. Period. It’s hypothetical. It doesn’t exist in reality. Again, you seem to be approaching this research assuming the media headlines derived from it are perfectly representative of the results it reports. This study explores a hypothetical scenario to try to put context to the life cycle emissions of coal exports. It does so by (explicitly) coupling export and the prevailing national trend in coal-to-gas replacement, and comparing this to what it would look like if Boardman and Centralia were brought in line with EPA standards (so as to avoid closing in their longer-term theoretical operational lives) to burn the same amount of coal themselves.
It’s not a likely, or as you point out, even possible scenario in the long-term since B/C are being shuttered by 2025. That’s not the point. The study (unlike the headlines) is not recommending any course of action to the Northwest (if anything, as it points out in the Discussion, the coal should be left in the ground and other environmental impacts of export should be studied at length). To me, it seems simply to be putting some context and scale to the emissions impacts of coal exports in relation to how we burn coal at home– and, I will add, it’s the only study I’ve seen that does this. From the text, is seems the authors chose Morrow Pacific/Northwest because LCA is complicated and it’s a relatively simple case study (coal travels roughly the same route to B/C as it would during the first leg of Morrow Pacific, B & C are the only coal plants in the Northwest, and together they burn roughly the same amount of coal as would be exported). In other words, B/C are treated as example US coal plants (the authors discuss the replicability of this type of study for other energy export proposals, and look at the title! It says nothing about the Northwest or Morrow Pacific! It says “US exports”).
I think one of your sticking points is not understanding that the hypothetical scenario in the study is not stretched over time. It’s a snapshot (in emissions per MWh) of what the world would look like tomorrow under the two hypothetical scenarios– for comparison purposes only, as the title of the study implies (read the title again). Your perspective would be more warranted if such an emissions timeline existed (since Boardman and Centralia are being shut down completely after 2025). But even so, at the time this paper was published, Morrow Pacific was planning a 2015 start date. That’s 10 years before B&C are both completely offline. That’s 10 years of burning coal at B/C while also exporting. That’s why I’m glad they compared coal burned in South Korea with coal burned here. It adds context and insight; it DOES NOT recommend a strategy to be used in the Northwest. I think people need to be asking these climate questions, even if they do not reflect what is likely to play out in reality.
Thanks for the good debate.