GIS Tour de Force on Electric Plant Siting Delivers Bad News

The Department of Energy (DoE) site at Oak Ridge National Laboratory (ORNL) (if I read right, funded by the Electric Power Research Institute (EPRI)) has undertaken a massive brute force Geographic Information System (GIS) analysis of the lower 48 states of the United States of America (USA), regarding siting of electric power plants.

In another small sign of journalistic progress, a recent New York Times article pleasantly surprises by pointing directly to the report. (Admittedly, the source is just a New York Times “blog”.)

The methodology was to divide the entire territory of the continental US into seven hundred million sites (100 meters square), and check each of them (using a computer, obviously) against some objective criteria of plant siting. The Times reports

The study, carried out by the Energy Department’s Oak Ridge National Laboratory, found locations for 515 gigawatts’ worth of new nuclear plants — nearly five times what exists now — based on considerations like the availability of cooling water and relatively low population density. There is also space for 168 gigawatts of “advanced coal” plants. Should the plants be designed to sequester the carbon dioxide they produce rather than emit it, however, the study did not factor in how far the carbon would have to be piped.

But potential locations for solar thermal plants, which use the sun’s heat to make steam and then electricity, are far more limited; if the plants are cooled by water, there is space for only about 18 gigawatts, the study said. If the plants are cooled by air, which reduces their efficiency, there would be space for 60 gigawatts, the authors found.

This would be more useful if calibrated, so I dug a bit. Current peak demand is about 800 GW. The 60 GW of solar thermal can’t do what we need even if it’s off by an order of magnitude. At least if the Times summary is correct and the Oak Ridge study is correct this is very bad news.

I would imagine that if these issues are at the margin in America they would be insurmountable in Europe, though. I think that when push comes to shove siting issues can be overcome. The report (which after all does come from coal country) foresees a major role for coal with CCS. But even so, it’s hard to see how decarbonization can be achieved with these constraints.

It’s important to note, though, that while DoE reports are presumably less prejudiced than, say, think tank reports, they are not actually peer reviewed. I have a suspicion of “inevitable research” here. Not that this approach isn’t useful, but perhaps the implementation shouldn’t be taken entirely at face value.

Still, I’m getting a picture that sheer siting issues will be a problem, and that nuclear and CCS are BOTH going to be needed.

The research did not include wind siting. But the siting problems with wind, and its poor match to demand cycles, are well-known. The latter can be ameliorated with a combination smart grid/electric vehicle fleet strategy, or with new compressed air energy storage, whose siting issues are discussed in the report.

The report is, just the same, of some considerable entertainment value to a map geek.


  1. These numbers for solar thermal sound too low. Earlier studies by the NREL found 285 GW potential for Arizona alone and about 1,000 GW for the southwest states.

    Closed loop Heller type cooling systems don't use very much water. There was no mention of them?

    One study done for the Western Governors Association found over 300 GW near existing power lines.

  2. It's rare to see detailed attempts at these macro constraint problems: people either seem to ignore them or, at the other extreme, assume energy armageddon. Here's an older attempt: Nathan Lewis' presentation, using (among other stuff) info from: Hoffert et al. ‘Energy Implications of Future Stabilization of Atmospheric CO2 Content’. Nature 395, no. 6705 (October 29, 1998): 881–884.

    Lewis makes a stab at the total land area required assuming just solar, without considering any geographical constraints (he picks a US state he's not too fussed about...!). Different techs have different constraints, but they all require them.

    Relatedly, Brian Walker reviews Reinventing Fire: I've asked over there whether it contains any thoughts on these top-level constraints.

    So there's one future - the most likely one, unless we can exactly substitute our current demand - where civilisation needs to pass through an energy bottleneck. Whether that's eighty, fifty or ten percent of current demand, the political problem is ricidulously hard - or at least seems so at the moment. We're left with the problem of needing to leave carbon fuels in the ground at the same time as reducing overall energy use. What politician is going to (successfully) campaign on that kind of platform?

    I've asked over at Hot Topic (the link above): does the author of Reinventing Fire address any of those issues, or it it assumed that a smooth substitution is going to take place - a bit like a magician pulling the tablecloth off without disturbing the diners.

  3. The survey is of limited practical use for decision making, because it can't handle PV and wind was excluded. See pages 95 and 139.

    What do you mean by But the siting problems with wind, and its poor match to demand cycles, are well-known.
    In Europe we're managing to keep the lights on using wind.

    You should also consider the effect of improving efficiency to cut consumption.

  4. My current (admittedly half-baked) understanding is that solar thermal blows photovoltaics out of the water at scale.

    Otherwise, this strikes me as more plausible. The amount of raw solar energy hitting undeveloped public lands in the west is surely enormous.

    I'm not finding a number for efficiency of conversion of solar thermal energy.

    A sunny area the size of Arizona getting 10% efficiency at 300 W 24-hour-average would yield, hmm
    30 w * 295 e 9 m^2 = 8850 GW. Granted that is a lot of land, but that is a lot of energy as well.

  5. I would contend that distributed rooftop photovoltaic combined with utility scale CST for grid scale with a smart grid to manage the energy flows is a large part of the "solution," assuming that solving the energy predicament is all we need (which, of course, is very far from the truth).

    If I may be so bold as to suggest, you might consider asking Dr. Tom Murphy, a physics professor at UCSD and host of Do the Math to contribute to this site.

  6. Rob, +1 to a Do The Math guestpost or repost. Hadn't come across it before, cheers. The current top article has a nice list of his present take on the 'chief problems' and (only a quick glance) the energy matrix is an interesting idea / must have taken a while...!

    He seems to argue for both a steady state future as a necessity (true ultimately, but in the near future? And have we really understand the difference between growth and development and how that shows up in our stats?) and for peak oil as an inevitable, within-the-next-decade outcome. Some of the economics of the problem get forgotten; the Grauniad has highlighted a report today that, if the figures are right, the economics are vital: high prices = resource / land grab boom. If you assume peak oil is inevitable, that means being unprepared when high energy prices feed cheaper carbon fuels into the system - undermining alternative sources, unless we have some other way of keeping carbon fuels in the ground...

  7. Of course ORNL is a national lab and so operates under the direction of a political appontee (who this time is quite good IMO). I assure you theree isn't a coal mine for a long way in any direction and the history of the lab is that was where the unranium separaation was done in WWII; it used to be more of a nuclear center but all the national labs have rather branced out.

    The lack of sufficient sites for expanding nuclear power plants (NPPs) more than not quite 5 times is disturbing. It is worse in that every western river of any size is included. The example I know something about is the Columbia and Lower Snake rivers. Those are salmonid migratory routes and the fish have enough troubles without adding the extra burden of heating the water. So I'm dubious of the ORNL result which doesn't consider fisheries matters.

    Next, neither NPPs nor solar thermal does not have to require water for cooling; at efficiency expense it suffices to use air cooling to remove the reject heat at the bottom of the Rankine cycle. In addition probably it suffices to clean the mirrors for solar thermal using mostly air.

    Therefore the study is of limited usefulness for detailed planning but it does make plain that other studies going forward will have to consider the inefficiency caused by air cooling.

  8. I forgot about evaporative cooling. In the west other things are for drinking and water is for fighting over. Substantive withdrawals are not a good plan either, but some may well not cause much harm. On the other hand, the predictions are that much of the west is going to dry up. So once again considering air cooling appears advisable.

  9. Wouldn't (thermal) nuclear power plants presumably be replacing (thermal) coal / natural gas plants, and therefore require similar amounts of water? Or do nuclear plants require more water per watt?

    Also, our standards for nuclear plant siting might be rather subjective. I'm not sure if we have a very good objective handle on the risks of nuclear power. And newer plant designs are significantly decreasing those risks.

  10. NPPs require more water per watt than the newer coal burners (because those have higher thermal efficiency). For older ones its about the same.

    I don't know how to compare NPPs to combined cycle gas units.

    Of course NRC has to approve the siting of every proposed NPP with regards to safety. [Others approve regarding other siting concerns.] So possibly the existing sites for coal burners could be used for NPPs. However, we'll need quite a bit more electricity to displace transportation fuels as those become overly expensive.

    Professor Cohen's "Understanding Risks" is available online. In that essay one discovers that current NPPs are about as risky as eating peanut butter. The new Gen III+ designs are several orders of magnitude safer, something like 100--1000 according to NRC's standard probabilistic risk assessment.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.