Peak Oil and The Energy Utilization Chain (EUC)

This is a guest post, related to net energy, by Doug Reynolds. Dr. Reynolds is an associate professor of oil and energy economics at the University of Alaska Fairbanks, and author of "Scarcity and Growth Considering Oil and Energy", and "Alaska and North Slope Natural Gas".

I met Doug at the recent ASPO/USA conference in Boston. We both were intersted in Charlie Hall's working paper on "The Minimal EROI Required for Society" and struck up a conversation on net energy, and the peaking of Russian oil. I invited him to do some guest posts on his areas of expertise of economics/net energy and Peak Oil/Soviet Union. Below the fold is Doug's guest post on the "Energy Utilization Chain", followed by my comments.

The Maglev Train.
It has often been cited how economics does not incorporate or look at physics and biology when it comes to sustainability issues. However it can also be said that physicists and biologists have failed to recognize economics when it comes to studying Peak oil or how to adapt to Peak oil. One of the more interesting dimensions in the discussions of net energy analysis is the Energy Utilization Chain (EUC). Instead of looking only at the net energy from say the oil well to the car wheel (well to wheel), it is important to also look at the energy service itself. You can have a wheel on a car, but you can also take a train or develop other forms of transportation that can use high net energy sources. In economics, this is called substitution and the degree of that substitution is called the elasticity of substitution.

However, the problem from the economics side is that economists are quick to say that substitution is possible, which gives the appearance that the economists--the Julian Simon's Ultimate Resource crowd--have won the day. But they then fail to consider physics and the entropy law. For example if the net energy for an alcohol fueled car is low, then just use a coal fired steam locomotive train instead or use nuclear power to run electric trains. But if that is your substitute you need to ask, how many railroad tracks and electric corridors are you going to need to build, to replace all the automobiles you have? Such infrastructure would take a long time to build, but more than that it would also require a lot of energy. Thus the net energy of the EUC from in-situ energy source all the way to the energy service is important.

One reason oil is so valuable is because it is in the physical state of being a liquid as opposed to a gas or solid. Solar energy is in the form of an energy field, i.e. a field state, which is the lowest state. The state of the energy resource--the energy state grade--explains an additional value of each energy resource. Coal isn't as valuable as oil or natural gas because it has a lower solid state grade which is why you often pay a premium for oil or natural gas over coal. What is particularly great about a liquid energy resource is that you can take a single drop of that resource, burn it, and release the exhaust all within a split second. That has made the internal combustion engine possible which has made Large Independent Mobil Machinery (LIMMs) possible. The internal combustion engine--as opposed to a coal fired steam engine, which is an external combustion engine--has a great power to weight ratio making LIMMs possible. Coal or nuclear power cannot do that. This is why the oil EUC gives the economy such fantastic service.

Nevertheless, the switch to lower state grade energy resources implies more energy use in order to make heavier, clunkier coal steam engines, as well as electric transport systems, fit into the economy. Also electric transport systems will have a number of power losses along the way because power lines often lose a lot of energy due to heating the lines. So the net energy concept needs to expand to look at the entire EUC from original in-situ resource to the energy service that is being provided. Along the way there will be energy inputs needed to simply build new infrastructure.

Coal fired steam locomotive

Maglev train

For example what is the energy necessary to build two airports, one in Los Angeles and one San Francisco, compared to building, say, a high speed magnetic levitation (Maglev) rail all the way between the cities. Once the energy for building the rail is accounted for, and even some sort of energy rate of return on that initial energy use to compensate the investment, I suspect airports actually have a much lower net energy use. Even if plenty of nuclear power is available, and with out ancillary problems such as storage and nuclear weapons proliferation, you still need a liquid fuel to run all the LIMM construction machinery and LIMM construction vehicles that will build the Maglev line.

Trans-alaska pipeline

Alaksa pipeline pathway

The trans-Alaska pipeline cost $6 billion in 1975 dollars to construct, a cost overrun of six times, and undoubtedly all that construction used a lot of liquid energy. Luckily it was built to transport pure energy, upwards of 2 million barrels of oil a day. Building a similarly difficult maglev would only transport a few hundred people a day at most. Although without a maglev, and no planes being possible without liquid fuels, then steam trains and some higher cost electric trains would be left. That implies a major reduction in our standard of living, which will manifest itself as a GDP decline--a recession or a great depression. It would be similar to the Soviet Unions economic collapse. We would still need to construct a lot of new infrastructure just to accommodate steam locomotives or electric transport systems. Therefore replacing the oil EUC will create a lot of energy problems.

Other economic issues besides this must also be included into concepts of the Hubbert curve. Many such economic concepts can be read in my book, Scarcity and Growth Considering Oil and Energy: An alternative neo-classical view.


Reynolds, Douglas B., (2002) Scarcity and Growth Considering Oil and Energy: An Alternative Neo-Classical View, sole author academic monograph, The Edwin Mellen Press, 240 pages.

_________. (2000) "Energy Utilization Chain: Determining Viable Oil Alternative Technology," Energy Sources, Volume 22, Number 3, April, pp. 215-226.

_________. (1999) "Modeling OPEC Behavior: Theories of Risk Aversion for Oil Producer Decisions," Energy Policy, Volume 27, pp. 901-912.

I have 3 main insights/takeaways from Professor Reynolds post.

  1. Using wide energy boundaries is important
  2. Societal demand impacts net energy
  3. The timing of (net)energy flows matters.

(I will put them in separate comments to organize any discussion).

First the issue of boundaries. As Professor Reynolds points out, an energy technology might have a higher net energy than its fossil fuel counterpart, but when the widest boundaries of inputs are considered, it may not.

There are few standards (so far) in comparing alternative energy technologies on EROIs - some researchers, (like Patzek and Pimental), use more inclusive (in my opinion more correct) boundaries on what they consider as energy inputs (e.g. not only the energy needed to RUN the tractor to harvest the corn but also the energy needed to MAKE the tractor)

I am going to write a post on this soon, but essentially, oil is used in every layer of our societies transport system.  It is therefore not enough to measure the energy needed to create a wind turbine, or nuclear plant;  we have to include all the ancillary energy used in the highways, delivery of food for employees, delivery of parts, and the delivery of their parts, etc. This type of analysis reduces the `headline energy' and shows just how globally dependent on oil we have become.

Also, though it's hard to do, when comparing energy choices, we should include an energy cost for environmental externalities. If two energy technologies both have 3:1 EROIs but one has double the greenhouse gas emissions, clearly we would choose the other. The harder choice will be if the dirtier technology has a higher EROI (Coal-to-liquids comes to mind)-unless we have protocols for this in place ahead of time, the dirtier, higher energy options will always win out (which means lose out, in the long term)

About that EROEI boundary...

I think that since every layer of our society is dependent on a way or in an other on fossil fuels, I think that assessing actual EROEI would be kind of taken accounted for by calculating the actual cost of producing that energy.

Say for ethanol for example, cost calculated before subsides and tax break should amount to something more than what a comparable amount of oil (or gasoline) would cost.

If the cost is 3,50$ and the gasoline is actually 2,50$, it would imply that it is needing 1 $ more of gasoline input that is lost in the process.

So in order to get the same amount of ethanol, you need more energy than what is given by it.

Well, that need to be studied more in detail but Jeff Vail came with something near I think.

What do you think about it?


This is the crux of the issue. In theory you are corect. However in reality, using dollars to generate decisions gives you a constant moving target, based on the markets sloth at recognizing scarcity. Your decisions would change monthly in an energy crisis. Net energy analysis attempts to jump ahead of this by acknowledging that energy  is important (implicitly, energy is more central to our needs, whereas the market is good at pricing our wants), and pricing things in energy terms so that accurate planning can be made in advance. I suspect that as we begin to run faster on our natural gas treadmill, the value of pricing ethanol in energy terms instead of $ will become apparent.

In a perfectly functioning market that has no subsidies, that values externalities, and that doesnt exhibit very steep discount rates (valuing the present more than it should), pricing in dollars would equal pricing in energy, but we are a long way from that.

sorry I missed you in Portland...(next summer!)


When calculating EROI in this way, any bias in the dollar value of energy divides out. One might have to take an average over a period of time to account for variations in market forces of different energy types, but this is easy to do. Similarly, all energy produced with fossil fuels is discounted similarly and this bias exists both in the numerator and the denominator in an EROI calculation.

Using a more rigorous approach (following energy expenditures all the way up the chain), you still cannot account for externalities (nuclear waste buried for centuries, global warming effects, missing mountaintops in WV, quite a mess in Alberta, etc. etc.). What is the energy value of those? What is the energy cost to society of disrupting the food supply?

Also, improvements in technology improve EROI--it is not a static number. Pimentel has been often criticised (and then his arguments dismissed) for using supposedly obsolete numbers.

I suggested something similar further down
Actually not a response to you, but I was just wondering if I could get more of the contents of Mr. Reneynold's book "Scarcity and Growth Considering Oil and Energy" without having to buy the book. Cost over $100 new - an awful lot for a book that I might read two chapters of. It would really surprise me if the library here in Germany would have it, and the one-page abstract linked above was less than a teaser..

Cheers, Dom

Im not sure - you could email him at the address listed on his home page in intro
Second, societal demand impacts EROI

The energy input into an energy harvesting process is subtracted from the gross energy to get the net energy available.  But societies a)choices and b) infrastructure also impact EROI.  For example, if everyone was vegetarians, there would be no demand for dry distiller grains for cattle. These DDGs get an energy credit that increases the EROI of ethanol production by allocating away some of the inputs.If there was no demand for DDGs, the net energy of corn ethanol would decline.

The choice to drive cars reduces the net energy we receive from cellulosic fermentation (if we demanded electricity instead of liquid fuel, we would lose less of the energy in the lignin or bagasse).  We continually face the tradeoff between energy quantity (in Btus) and energy quality (in whatever energy services society demands).

Our infrastructure also impacts EROI. For example, a wind entrepreneur has a 5:1 EROI on small scale wind turbines, which he imports parts for, assembles, then distributes. He currently uses railcars to ship his components - if the rail industry grows too slowly to accommodate all rail demand as energy prices increase, he will be forced to use less efficient means, perhaps shipping his turbine parts in semi-trucks, or some such.  This will increase not only his dollar budget, but his energy budget, putting downward pressure on the 'net energy' from the wind turbines.

Essentially, Energy Return on Investment(EROI) is a snapshot in time of a technology combined with an energy source combined with an infrastructure. Oil is pervasive in impacting the EROIs of alternatives. The main point is that we can impact EROI not only by technology, but by demand choices.

If there was no demand for DDGs, the net energy of corn ethanol would decline

This hasn't gotten much attention, but it is in fact happening. The price of DDGs is falling steadily even as the price of corn rises. This is adding to ethanol costs, and is only projected to get worse. I think we will get to the point that excess DDGs are burned for fuel.

Aha!  About a month or so ago I brought up some questions regarding the legitimacy of taking a credit for the energy input of the DDGS byproduct (or co-product, depending on one's point of view).

I think one of the points I was trying to make was that if the DDGS did not exist, cattle would be fed something else, something that would have its own energy content. So, the energy credit (if there really is one) should really be the energy content of the most likely alternative feed that the DDGS displaces, not the DDGS itself. This may or may not be the correct way of looking at it, but I've had a gut feel from the very start that something was a little dodgy about the DDGS credit.

I know very little about agriculture, but is it not also true that cattle can only tolerate a certain percentage of DDGS in their feed mix without suffering adverse effects?  If so, then it would not be too hard to picture the market for DDGS getting saturated pretty quickly as more and more ethanol plants come on line.

 You are probably right that at some point a large fraction of the DDGS will be burned just to recover some heat value from an otherwise next to worthless material.

Very true that cattle only tolerate a limited abmount of DDG, but 30% of their total feed is very reasonable.

The other big change coming is that companies will start extracting more of the corn oil from the DDG product (already happening at a few plants).  This is positive in several ways.  From what I understand, less energy is than required to dry the DDG, we pick up another high-value product stream (corn oil -- think biodiesel) and the resulting DDG product can be used for the pork and poultry industries, which so far have been unable to use much DDG at all.

Although there will be short term market issues due to the extremely rapid growth of the ethanol business, we will run out of land to grow corn long, long before we run out of a way to utilize the by-product.

I saw on a recent ag report that nationwide, cattle are not finishing out as well this past year.  Do you have any theories about this?  They mentioned drought conditions being one, and many people are jumping to the conclusion that its the DDG being fed.  What are your thoughts on this?
That is a very intresting side note.  Getting cattle to "finish well" means getting internal fat (marbling, what makes a steak juicy), without external fat, the stuff around the edges.

There is raging debate in the industry about how much feeding DDG is responsible for what has been a significant decline in cattle grading over the last couple of years.

From our personal experience, I think DDG is a significant culprit, and we are investigating how to address the issue.

So DDG makes for a leaner healthier steer???

Very interesting.

It sounds as if this nation's steers are all on the Adkin's diet.  Now there's a big market--using DDGs for a human diet product.  
Ideally, the drying step would be eliminated completely.  This can be done if the fermentation is integrated with or near to a cattle-feeding operation.
Absolutly not true on DDG falling steadily.  

We feed cattle, so use a lot of DDG's.  Earlier this summer we were contracted DDG from our local ethanol plant for $88 per ton, corn was about $2.00 a bushel at that time.  Corn is now worth $3.50 and DDG has shot up as well.  Trying to buy some now would cost about $130 per ton.  About a 50% price rise.  National market reports show this same thing happening across the country.

As long as corn to ethanol is heavily subsidized, DDG's will seldom be burned for fuel as they are too valuable as a feed ingredient.  

In the big picture, what will probably happen over the next few years is that corn planting will rise substantially, soybean planting will decline.  Historically, the livestock industry has gotten it's protien from soybeans (meal) -- much of that demand will now be filled by DDG.

I wonder what the GHG impact of a switch from soy to DDG would be. I know that cattle can't fully digest DDGs and the resultant methane is not insignificant.

Here is some Christmas cattle trivia: The MROI (Milk return on Energy invested) is about 5:1

Absolutly not true on DDG falling steadily.

Absolutely is. I have seen several news reports on this. Here is a blurb from Ethanol Producer magazine:

As the start of September came around, the DDGS market was still reeling from the dog days of summer. August prices, which began the month in the $60s FOB Midwest, moved to the mid-$50s in central Iowa as new plants started up without DDGS railcars in place.

I also have a reference around here from November that I will try to track down. Finally, I have seen an analysis that predicts that the DDGs market will completely saturate pretty soon as ethanol production continues to ramp up.

A couple more blurbs on this. From the National Corn Grower's Association:

DDGS prices will continue to decline as expanding ethanol production expands available supplies.

From a USDA report:

The value of byproduct credits declined from 30 cents per gallon in 2003 and 2004 to about 22 cents per gallon in 2005.

That's from a PDF. You can see the exerpts here:

Scroll down to comments by "Randall Parker."

DDGS prices can be found here, current thru Oct. from USDA ($75/ton wholesale, Lawrenceburg IN)


Thanks for that. I have searched and searched for just that kind of month by month data on DDGs prices. DDGs prices have fallen steadily since April from $95 down to $75 per ton.
You are very welcome. I worked on the design of a large corn ethanol plant back in 1983 and even then the only thing that made it economically feasible was the subsidy. I think it is still in operation someplace in IL.

My opinion is that MTBE substitution is the current driver of ethanol production in the US, not use as an alternative fuel. The economics are just not there.

Regards and thanks for your frequent and useful comments.

What is your opinion on the possibility of increasing efficiencies in ethanol and biodiesel production?  As noted earlier, some plants are beginning to extract more oil from DDGs.  One company looking at this is GS CleanTech  There appears to be about a 5% gain in productivity from extracting more oil from DDGs.  Other technologies this company is promoting are some form of gasification of DDGs and biodiesel reactors linked to ethanol plants CO2 emissions.  Personally I think corn and soybean are odd and possibly dangerous, especially corn which I like to eat fairly often, starting points for a biofuels program, but I do wonder, even with these starting points how much room is their for incremental improvement as the industry matures?
I was going to write something technical in nature, but then I noticed the co. you mention is publicly traded (penny stock) so I will pass.

Nevertheless, I firmly believe that biodiesel, ethanol, CTL, GTL, tar sands and in general all the "whatever-to-liquids" programs are driven by the quest to keep the whole post-petroleum era looking and acting as much as possible like the Oil Era. There is too much capital invested in the whole liquid fuels regime for it to change overnight. The internal combustion engine drives our planetary economy and society.

Yet the quest is ultimately destined to fail, for reasons analyzed extremely well in this site (EROEI + GHG). Oh, they may very well serve as stopgap measures as we (if we) transition to a purely Electric Era, but not much more beyond that. The sooner we realize that the Sun is the only source of permanent energy, the better our future.



There are some much more serious firms looking at bio-diesel from the ethanol by-product.  The biggest "dry-mill" ethanol plant operator for one:

Can't disagree on your observation regarding the sun, but is it possible that biomass done right provides a nice "battery" in that equation?

all the "whatever-to-liquids" programs are driven by the quest to keep the whole post-petroleum era looking and acting as much as possible like the Oil Era. There is too much capital invested in the whole liquid fuels regime for it to change overnight.
Indeed, but we already have enough capital invested in the electric regime to allow a shift to begin.  The average age of a modern car is about 8.5 years; if it only takes that long to turn half the vehicle fleet into high-range PHEV's, we'll be on top of the problem.

The key is to get started now.

Guys, your data has serious lag time.  Corn market started going up about the first of October, so prices earlier in the year on DDG were reflective of very low priced corn (and soybean meal, the other commodity it can replace).  

Look at this weekly report issues by USDA -- DDG currently $110 to $130 per ton..  mnreports/SJ_GR225.txt  

Not sure where everything will shake out, but DDG is certianly not cheaper.  There will be serious mis-location issues over the next couple of years, made worse with high cost of transportation, but the feed market is pretty efficent.  DDG will be priced at its feed value.  Main value of DDG is it's protien, and the US/north American/global protien market to date is very large.

Oh, and really appreciate the cow picture.. drives home the point that under skyrocketing energy costs, we will all eat less meat!! and who knows, even though I feed cattle, we may be more healthy as a nation by cutting meat consumption :(

Correct link to weekly DDG price:  
Thks vm, I was wondering why DDG's were lower while corn was zooming.

Thks again.

"even though I feed cattle, we may be more healthy as a nation by cutting meat consumption"

You are a scholar and a gentleman!

I'm not sure soybean planting will decline much since the demand for soybeans is also growing with increasing #s of biodiesel producing plants.  (And with high wheat prices, wheat is also competing for acreage.)
Soybeans make an extremely poor bio-diesel crop.  A soybean is only 20% oil, compared to 40% for crops like canola and sunflowers.  And even those bio-diesel yields are pretty pathetic in the whole scheme of things.

Soybeans are 75% soybean meal (which itself is about 48% protien), so with the coming flood of DDG's, we will be awash in protien and I really believe that soybean acerage will drop significantly under current policy.

One interesting side note, in our part of the world (northeast Kansas), we can grow about as much bio-diesel an acre from a wheat-doublecrop sunflower program as a single crop of soybeans, and have a full "food" crop to harvest that year as well.

However, in reality, soybeans are being used to produce biodiesel in spite of other seeds containing the higher levels of oil...

Bunge North America, the North American operating arm of Bunge Limited (NYSE: BG), announced that it is expanding the crush capacity of its soybean processing plant in Council Bluffs, Iowa, by more than 11 million bushels per year. When the expansion is complete by harvest 2008, the facility will have an annual crush capacity of nearly 77 million bushels, the largest in the United States.

One interesting side note, in our part of the world (northeast Kansas), we can grow about as much bio-diesel an acre from a wheat-doublecrop sunflower program as a single crop of soybeans, and have a full "food" crop to harvest that year as well

Wow, now just get yourself and your neighbors to run your tractors on straight sunflower oil and at least we'll know we won't starve.  I'm only half joking.

I think we will get to the point that excess DDGs are burned for fuel.

That would be a waste.

So the dry distillers' grains is a pre-emergent herbicide, in other words it stops weed seeds from sprouting. As soon as they sprout, they die. That's because what I've done by putting this stuff in the soil is I've fed seed-eating fungi and bacteria. There's a population explosion, when the weed seeds sprout, that exploding biology eats the little roots of the weed seeds. It's really a tricky little system, really fun. So the farmer doesn't need to buy Round-Up anymore, because he's got it built in, and he doesn't need to buy the GMO corn.

That may be, but I have a hard time taking Blume seriously after some of his previous claims.
I don't know the accuracy of this, but a few farmers in Nebr. that I have talked to said they new people that were spreading DDG's on their corn fields, and that was last Spring.
You don't have to find him 'serious' just because he feels making booze for fuel is a good plan.

Examine what others are doing in the field, look into the patent claims.   Look at the comments of farmers who are using the method.  

Taking the DDG and making a biochar to make a terra preta would also be a fine plan.   But burning for energy?   Using the waste for betting the food supply change is a FAR better plan.

What keeps these fungi and whatnot from eating the root hairs of the crop seeds?
If you plant the food crop seeds during the time the fungi blooms, you loose the crop.    Directly seeded crop would be planted once the fungal bloom is on the downswing.  The act of trnching for the seeds will kill off the fungus, as should seed treatment with Magnesium (I know of no studies about seed coating with magnesium however.)

If you use transplants, they can survive some fungal activity, in fact to the point were black mold can grow right up the plant, if you go overboard on the spent-grain fungal activity.

Third, the timing of energy flows and fixed vs marginal EROI

The other point Professor Reynolds alludes to is the concept of fixed vs marginal EROI.  Given dueling constraints of time and limited liquid fuel, the timing of energy inputs vs energy outputs in the future should have bearing on a technologies acceptance. As energy prices go up, high EROI alternative technologies that are BUILT (hydro, wind, etc) will go up in value immensely as their inputs occured in a cheap oil environment, however those PROPOSED (on paper) will find their high EROI (on paper) unattainable, as inefficiencies in the transportation system will eat away at the high original net energy (at the snapshot in time it was analyzed, when oil was abundant).

In effect, a fixed energy system throwing off 4-1 EROI must be differentiated from a 'future' system that might throw off 10:1.  An `energy discount rate' or some such, could be used to make comparisons of energy technologies with disparate timelines more commensurate.

Good organization of the analysis.

I've posted before on the idea that a usefule, albeit more complex, approach to EROI studies would be on the lines of a 'black box' (BB) from which we draw a certain type of energy needed for some use (eg transportation energy). Within the BB there might be a number of different types of energy extraction/conversion processes going on to achieve the overall output. Any one of the processes, by itself, might be EROI negative but, by contributing to an overall positive EROI for the whole BB process, would be well worthwhile. This actually relates more to #1 about wide boundaries.

There is definitely a tendency among the number-crunchers to focus on each tree and not see the forest.

At some point like now, a research grant needs to be provided to develop a model that would include each tree.
well said! And I am happy to announce that there are several things happening on this front following the net energy presentations at ASPO - apparently a meeting in January at RPI in NY to further work on Global Energy Systems Modeling.
Third, the timing of energy flows and fixed vs marginal EROI

Nate - looking forward to your article. I sense that a workable Net Energy Analysis would help resolve some of the debates that constantly come up on TOD.

With regard to the timing of energy flows it would seem appropriate to model them on the basis of lifecycle costs. For example, major producers of airline gas turbines are willing to sell engines at cost as the purchase price is 30% of the full lifecycle cost for the turbine. Once you have an installed base you can make your profit on sales of spares and support contracts. You basically have an annuity income.

With regard to the NEA between 2 airports and a single rail-line between the same two points, I find it hard to believe that the airports will be more economical. Airlines are limited in freight capacity, the rail line is dual use.

Hopefully your article would provide an introduction to NEA methodology that would assist all of us in a better understanding of trade offs.

Cheers, and best seasonal wishes to all of TOD!
igh speed rail (TGV, ICE, bullet trains (shinshanken ?)) are NOT dual use, but pax only.  details like banking and grade prevent higher density cargo. although at least one mail train runs on TGV.  Maglev is pax only.

See my post at the bottom.

SBB (Swiss rail) plans to run 200 & 250 kph pax trains with freight.  This seems to be the upper limit for flat straight rail lines.

Best Hopes,


To clarify, is it being said here that the eroei of a given energy conversion system will decline as the eroei of its inputs declines?  

I'm not entirely comfortable with the use of value in this post.  Value to whom.  What about the problem of depreciation of the asset?  At any point in time, is not value determined by the net energy yet to be realized from any two assets?

Thank-you so much for your work in bringing Dr. Reynolds to these pages and for your own continued emphasis on a matter central to the peak oil event.  I wish I had more time at the moment to pursue this thread. Alas, once again, I have left my winter solstice duties to the last moment.  

For your enjoyment:

"There are some reasons why I stress here...the irrevocability of the entropic process.  One reason interests the economist in particular.  If the entropic process were not irrevocable, i.e., if the energy of a piece of coal or of uranium could be used over and over again ad infinitum, scarcity would hardly exist in man's life.  Up to a certain level, even an increase in population would not create scarcity: mankind would simply have to use the existing stocks more frequently.  Another reason is of more general interest.  It concerns one of man's weaknesses, namely, our reluctance to recognize our limitations in relation to space, to time, and to matter and energy. It is because of this weakness that, even though no one would go so far as to maintain that it is possible to heat the boiler with some ashes, the idea that we may defeat the Entropy Law by bootlegging low entropy with the aid of some ingenious device has its periodical fits of fashion.  Alternatively, man is prone to believe that there must exist some form of energy with a self-perpetuating power."  

Nicholas Georgescu-Roegen, The Entropy Law and the Economic Process, Harvard University Press, 1971, p. 6.

Gosh, I really did enjoy that!

As an example, I have run into political leaders waiting for free energy devices to keep the highway system going.

Tribal thinking also prevails.  Many people appear immune to credible information if it calls into question the decision made by them and their associates.  Institutional norms and policies become defended vigorously by people who have participated and benefited from those institutions--no matter what information is provided that contradicts those norms.  

I would like more discussion on TOD of the possible evolutionary basis for this kind of behavior.  It is something engineers, scientists, empiricists, etc. need to be aware of.  

Folks, also consider this a reminder to positively rate these articles (using the icons under the tags in the story title) at reddit, digg, and if you are so inclined. Also, don't forget to submit them to your favorite link farms, such as metafilter, stumbleupon, slashdot, fark, boingboing, furl, or any of the others. These posts are a lot of work, and the authors appreciate your helping them get more readers for their work however you can.

Cheers and Happy Holidays from The Oil Drum!

Prof. Goose,

Did someone put up a post at some point explaining how these websites (reddit, digg, etc.) work?  I've tried rating these articles many times in the past, but I can't figure out if I'm doing it correctly or if it's having an effect.  For instance, I went to the reddit page for this article having logged in at reddit.  I got a note that you had posted a link to the article, with a pair of arrows on the side of the note.  Then reddit gave me a blank comment box.  I clicked the top arrow, which changed color, but otherwise nothing seemed to happen.  I keep expecting to see that the article has increased a notch in ratings, but there never seems to be any indication of that.  Are comments required?

I went to digg and it seemed to have no record of the article yet.  I'm never sure in this case whether I did something wrong or if I'm really the first person to submit a comment.  On another article a few days ago, it seemed to create a separate article comment chain from one you had started.  Does that increase or decrease the ranking of the article?  How does digg decide which category the article goes into?  

I'm sure these websites are valuable somehow (or you wouldn't make such an effort to remind us to use them), but I can't make heads or tails of their documentation.  Could someone who understands these sites write some kind of a primer for TODers who want to help?

Once again, someone else points out the energy cost associated with the change in technologies AND that energy has to come from someplace.  I can take all the raw materials for a 1 square meter solar panel (high efficiency design) and allow the sun to shine on it for 1500 to 2500 days.  But no solar panel spontaneously appears at the end of that time.  

The quality of the energy source makes a huge difference.  Nuclear powered aircraft just aren't in the cards.  

Technically, it's 1500 to 2500 minutes to replace the energy for a solar panel. 1500 to 2500 hours is to replace the energy for the entire silicon, wafer fab, and panel assembly plants manufacturing chain.
It works out the same if you use the cheaper but less efficient thin film solar panels.
Also maglev unlike internal combustion is an angry and unforgiving technology, see the recent accident in Germany.
The tragic maglev crash in Germany was the result of a train colliding with another piece of equipment on the track i.e. "human error", not, I think its fair to say due to some inherent aspect of the propulsion system of the train. I think an IC or conventioanl electric train is subject to the same sort of risks.

Crash story available here:,,1878980,00.html

See the 'angry and unforgiven technology' of me parking a vat of asphalt (a maintenance vehicle) infront of your garage door when you back out tomorrow morning.
Yes, I always back out at 350km/h.
They settled on a super-expensive maglev variant specifically in order that no train collision could ever occur: this design decision (and the track involved), coupled with the fact that a demonstration route must be scenic, necessitated hours of cleanup of debris every night.  By vehicles that weren't propelled by the linear motor.

When someone screwed up the scheduling of those vehicles, obviously things turned bad.  Just as an airport runway repair crew who screws up their schedules can make things turn bad.

That doesn't mean that the technology is any more 'angry and unforgiving' than air travel or HSR.  That's a prejudicial way to describe a deadly human error.

I agree in principle that to be correct a net energy analysis should have wide boundaries so as to capture as much of the energy content of a particular product of activity.

However, I maintain that if those boundaries of analysis get too broad, then one can get into all sorts of trouble. This is because at the outer edges of the boundaries one has multiple uses of the energy-consuming entity, and therefore one has to get into allocating energy consumption among the various uses.  But to do this, one has to make all sorts of assumptions, value judgements, and just plain wild-ass guesses.

Take the tractor that helps make the corn that makes the ethanol that makes the car go. A certain amount of energy went into making that tractor. But the farmer doesn't just use the tractor to plant corn, but possibly to also plant other crops and to do a variety of other farm chores. Going further back up the chain, the factory that makes the tractor might make a whole line of other farm machinery, so one must allocate there. Then you have roads, which obviously have multiple uses. It gets messy real fast.

This leads me to the view that once the boundary gets past a certain width, the net energy analysis might be getting less rather than more accurate. Thus, I think there is something to be said about stopping once you get to a point several steps back up the energy chain. Otherwise, thing tend to degenerate into big allocation game, in which two equally legimate analyses can arrive at markedly different numbers.

I mean where do you stop .... when you're at the point of including the energy content of the ham and eggs the farmer ate for breakfast the day he goes out to plow his fields?

Good point. I dont have an easy answer to that. Wide boundaries are essential but the widest boundaries are prohibitive. Use of systems analysis by decisionmakers will be a step in the right direction. Using energy as a decision criteria in addition to dollars would be another
While it might seem ridiculously simplistic, I'm beginning to believe that a better way to determine EROEI is to calculate the total cost (in today's dollars) of producing the product (corn, windmill, PV, etc.), convert that to energy (at today's  dollars for the product), and then divide that into the total energy produced in use. Subsidies on the product (such as for ethanol or PV) would be added to the production cost, while the sales value of other uses (distiller grains, e.g.) would be subtracted from it--both prior to the currency-energy conversion.

As an example, if the production cost for ethanol that would have the energy utility of a gallon of gasoline is equal to or exceeds that of a gallon of gas retail, then it has an EROEI less than 1 and it also (obviously) makes no economic sense to proceed. It still means allocating the cost of the tractor (to the farmer) between its various uses, but one doesn't have to figure out the energy costs for manufacturing everything the farmer needs.

EROI wouldnt work this way. The reason to use net energy analysis is to give a clearer picture of scarcity and value based on natural laws. Introducing the metric of dollars to NEA, in a market designed for just in time inventories, would be akin to being half-pregnant.
Disagree. All the energy inputs for manufacturing a PV cell, for example, are eventually reflected in the cost to buy it, which is determined by the cost to manufacture and market it.

I'm still talking NEA, but just a different way to estimate the complex chain of inputs. The variances due to market forces are probably no more than those encountered when trying to pin down all the energy inputs needed to fabricate something.

All the energy inputs for manufacturing a PV cell, for example, are eventually reflected in the cost to buy it

The key word here is "eventually".  It is a certainty that we have less oil remaining in the ground than we did a year ago, because we used around 30 billion barrels of it, yet prices are lower.  Is long term availability of oil more or less scarce than it was a year ago?  The reason to use NEA over dollars is to bypass the timing/uncertainty of the market getting it right.

Consider an exogenous event dramatically increasing energy prices - like the shutting down of trade through the Strait of Hormuz, or natural gas depletion/cold weather causing shortages. The dollar cost of ethanol would skyrocket, yet the EROI would be little changed.

Awareness of peak oil is happening gradually, but time is our most precious resource when it comes to adaptation and mitigation. Making decisions in energy terms instead of dollar terms, buys us time.

Consider an exogenous event dramatically increasing energy prices - like the shutting down of trade through the Strait of Hormuz, or natural gas depletion/cold weather causing shortages. The dollar cost of ethanol would skyrocket, yet the EROI would be little changed.
True, but the dollar cost of the the oil and natural gas equivalent would skyrocket as well--the dollar is only an instantaneous itermediary in calculating EROI.

It's not a perfect approach, but I think the difficulties inherent in trying to stick with energy all the way up the chain are such that nobody will ever be convinced--certainly not economists, and they seem to have the ears of too many of the PTB.

JB - that is kind of the point. Economists have been instrumental on a (relatively) empty planet in efficiently allocating resources. On a full planet, their role will give way to those embracing more biophysical approaches.

That is why I was keen to have Doug Reynolds contribute a post - he is a neo-classically trained economist, that is thinking in these terms. There need to be more like him that straddle both paradigms

Costs are a human artifact that vary over time (and often irrationally) unlike physical units such as BTUs and joules.  For example, you are starving and I have an extra loaf of bread that I normally would sell for a dollar.  However, my costs for wheat are going up so I raise my price to a dollar and a half.  There has been no change in energy inputs to make the loaf, yet, under a cost-based EROEI, one would have to assume that energy is involved in the price change.  I say stick to physical units.


This is one of the classic problems with EROEI as a universal decisionmaking tool.  I've sat in on plenty of online debates about this, and I don't have an answer either.  I will, however, attempt to float a few ideas in that direction.  Please do tell me what you think.

One question should be asked up front, and that is, what are you planning to use the EROEI results for?  If you're trying to come up with a absolute figure that is totally inclusive, and that can be compared to independently-developed absolute EROEI figures for other technologies, you've created a very, very hard -- maybe impossible -- problem for yourself.  A universally-applicable EROEI number requires infinitely wide boundary conditions, and that's just not feasible, as far as I can tell.

So, if credible estimates of absolute EROEI are not possible, what does that leave us with?  Alot, really, because we can still develop comparative EROEI figures that contrast two or more technologies, processes or systems.

If you're going to develop useful comparative EROEIs, then probably the most important thing is to make sure that the boundary conditions are drawn consistently for all the subjects of your analysis.  That is, if you're going to include the embodied energy of the farmer's breakfast, you'd better also include the righand's breakfast as well.  So, starting out, you set the boundaries of your detailed analysis based on what is feasible for all of the technologies under examination.

Beyond these boundaries, it's less important to know detailed energy consumption information.  But it's still important to have a sense of relative second/third/fourth-order energy requirements, because a major discrepancy could invalidate your results.  I would suggest an order-of-magnitude analysis of discrepancies.

For example, you don't want to have to analyze in detail the energy used in commuting by workers involved in, say, etahnol production vs. gas production.  But it's probably worth looking at roughly how many people have to travel how far in each case.  If the numbers are roughly equivalent, then you're core analysis is probably still valid.  If, on the other hand, you find that there are ten times as many people involved in ethanol production as there are involved in gasoline production, then you may need to take a closer look at those figures, or at least note that discrepancy as a caveat to your core analysis.

I would also suggest that the discrepancy thresholds should become more permissive as you get further from the core processes being analyzed, due to diversity of use (joule's point about farm equipment being used for multiple tasks) if nothing else.  If you really want to consider the energy used to support the factory workers that make the farm equipment used to harvest the corn or pump the oil, you could probably estimate those numbers, compare them, and then ignore those figures unless there is a discrepancy of 10-100 times between the two processes.

My feeling about EROEI is that it is a useful, valuable metric, but that it is only helpful in a relative fashion.  In this way, it's like the calculation of personal ecological footprints that has become so popular.  I don't really think that these calculations can reasonably determine whether my own footprint is 5 acres, 7, or 10.  And I'm not sure that number would be meaningful, at any rate (how many biologically productive acres does it take to make my Powerbook?).  But I am very comfortable saying that someone who has a 30 acre footprint is doing alot more ecological damage than my 10 acre footprint.  Maybe not three times as much, specifically, but certainly more.

One question should be asked up front, and that is, what are you planning to use the EROEI results for?

Good points. And I agree - you want wide enough boundaries to catch the majority of the things that make technology A different technology B - to go all the way leads to near infinite regress.

And to plan, as Matthew Simmons has often suggested, based on an accounting of a global energy balance sheet, using EROI as one tool of many in a systems approach would be the end objective. What do we have? What do we need? How best to get there?

I'm not sure you do agree with me.  You said you want wide enough boundaries to catch the majority of the things that make technology A different technology B - to go all the way leads to near infinite regress, which seems to me to put us back in the same problem we started with, which is trying to develop a metric that is too complicated to ever get right.

Matt Simmons' proposal (which I have heard elsewhere; I'm not clear it originated with him) also seems to be focused on absolute EROEI, rather than relative/comparative EROEI, in that it wants to ecompass the entire system.  That would be nice, no doubt, but I don't think its feasible.

There's also the fact that EROEI considers only energy, and there are many other factors that determine what is sustainable and what is preferable.  I don't think that limitation makes it useless, but I do think it limits its applicability as a decisionmaking tool.

It seems to me that (realistically computable) EROEI is a useful tool within a relatively narrow context: it lets you compare the energy balance of different technologies.  I don't think it's the be-all, end-all metric.

on that we agree - if you email me, I'll send you the paper I referred to above.
I don't think EROEI is helpful at all.

I have three processes.

  1. A black box. You pour in one gallon of gas, it pours out two.

  2. A similar box, you pour in one gallon of gas, and 10,000 gallons of high quality farm soil, it produces five gallons of gas.

  3. A similar box, you pour in one gallon of gas, and a 5 year old child, and it gives you 10,000 gallons of gas.

Which of these is the best? I prefer #1, even though it has by far the lowest EROEI. Why is that? What does EROEI really say about a process, if anything at all?

#3 has an EROEI of 10,000, but can we power the country with it? We could certainly power anything we wanted with #1 for as long as we want. #2 might work, but would cause serious problems in the long term.

EROEI is a limited metric, obviously, since it considers only energy.  I don't think that means it is without merit.  Although see my response to Nate, above.
I recently co-wrote a paper on EROI and Multicriteria Analysis addressing this exact topic - once it is published (pending for several months now) I will write a post here on it.

Essentially you are right, we care about energy, but we care about water and soil, etc as well. What the limiting factor of a system is not always known.  It seems like liquid fuels will be ours, but who knows?

In general, if the output of the system is liquid fuel, then it's limiting factor had better not be liquid fuels. Indeed, you don't see any such system in existence, so all is well with the world.

The one thing you're always certain is not the limiting factor is the output, and that's the only thing EROEI measures, so what is the point of a number that only measures the one thing known apriori to be irrlevant?

EROEI is that it is a useful, valuable metric, but that it is only helpful in a relative fashion.

EROEI would be usefully supplemented by a measure of sustainability. If you have a low EROEI process where all the externalities are accounted for as inputs to another viable set of processes that "process set" should be of greater overall value then a unique independent process with high EROEI and commensurate high externalities.

The example provided earlier of dual cropping a food crop and fuel crop resulting in DDG that is both a soil amendment and and an input to a livestock program should be of greater "value" than a set of liquids fuel processes that result in externalities that threaten the long term survival of 50% of present species.

I do not know if there is yet a term for this. Inter-process efficiencies are what are being described. Both "efficiency" and "sustainability" may have competing meanings. This is really an entropic criteria; call it "EC" and solve for max EC. The externality problem in classical economics results in a solution for "bounded EC" and this results in a whole host of intractable problems.

GreenEngineer -

That was truly excellent!  Best clarification I've seen yet on this whole EROEI question.

Your comments re EROEI are very close to my own thinking on the subject, but are far better developed.

I think you really hit the nail on the head by asking what these EROEI numbers are really going to be used for. And I think we both agree that their purpose is not to track energy content all the way back to some infinitely broad boundary, but rather to enable a reasonably good (albeit imperfect) comparision of two different energy-producing schemes.

EROEI might be somewhat analogous to EPA fuel mileage. While it's generally a poor indicator of the actual mileage you will get, it is a very good indicator of whether one car will get better mileage than another car.

(In the meantime, I'm still wrestling over the question of who consumes more calories for breakfast: an Iowa farmer or an offshore oil rigger.)

This is not likely to become a serious problem in practice. In general, the farther you go, the smaller the contribution. The fuel for the tractor may contribut 100 units, the energy to make the tractor itself may be 10 units (or maybe it's only 5, if the tractor has multiple uses), the factory may be 1 unit (maybe 0.5 if the factory makes multiple things), etc... Just round it up to 120 and you've included the farmer's clothes and his breakfast too, and the number is about the same (i.e. 105.5 if you really wanted to split hairs and had excellent knowledge).

For something like a power plant, it's really is irrelevant. They generally have 60% thermodynamic losses, and about another 1% or so goes into mining the fuel, transporting it, building the structures themselves, etc... It doesn't really matter if that's 1% (for 39% efficiency) or 2% (for 38% efficiency). It probably wouldn't even matter if it was 10%, under most circumstances.

Something like ethanol has severe problems without considering the majority of these things. In terms of acres needed per gallon, it will never be close to producing as much as we need, so who cares if it takes the farmer 1 gallon to plow a field or 20? Even if it took 0, there isn't enough farmland, which is the real problem with ethanol, not  how many gallons of gas it takes to make a farmer's cup of morning coffee.

I'm not a great fan of economists for....well anything really. I think they are a pseudo science of wet-fingers-in-the-air type and run on guessimates that tend to make things worse rather than better. However if you can add something useful to the debate then great.

However I'd have to point up one particular statement, that 'energy fields' are somehow the lowest state. Take the innovation engineering methodology TRIZ and energy fields and their use is the 'highest' state. Playing around with burning and turning is considered to be less developed; less elegant.

Energy density and engineering sophistication are two very different things. Don't confuse one with the other.

Using diesel today to build railway lines for tomorrow makes me happier...
HROI (Happiness Return on Energy Invested) is something we should all try to increase....;)
Let us all have that as our top New Year's Resolution for 2007. Now down to specifics: How are we to maximize Happiness Return on Energy Invested?
  1. More GAS (Great Aerobic Sex).
  2. More bicycling.
  3. More walking.
  4. Much more sailing.

Any other suggestions?
I would list paying work and less girlfriend distance for more hugs before GAS. And swicth out sailing and walking for building something, otherwise it looks ok. :-)
How about restoring a wooden sailboat?

By the way, I am planning to move to a neighborhood where almost everything is within walking distance--which is also much closer to some of my children than I now am.

More time with family!

I am drunk as a skunk and want to make a bold statement:

I want one of these running on clean electricity running all the way from Aberdeen to the Alps by 2012±3years

And Don - I'm glad you're back!

..or do I? Having now read the article and looked at all the pictures twice.
The obligatory comment.

EROEI is a useless metric, please, please, PLEASE stop pretending it has use. You want a metric, use something like barrels of oil depleted from field per person/mile, or something of actual relevance.

If an output is also an input, the factor it out to figure out some meaningful concept of efficiency. "It takes us 1 ton of coal to mine 100 tons of coal...." Irrelevant, entirely irrelevant. However.... "After subtracting any resources needed to extract the coal, this coal seam will produce a net gain of X tons of coal." Completely relevant, and actually useful. Nobody cares a flying fig how many tons of coal it takes to produce a ton of coal, they just care how many net tons can be produced from a deposit, and nothing more. If it takes 2 tons of coal to produce 1 ton, then the net tons of coal produceable from this deposit is 0. Simple, easy, sensible.

Now, on to some of the more interesting myths...

  1. It takes energy to produce energy. True, to a degree, but pretty much every major means of producing energy takes a pittance to produce compared to what it produces. Do the calcs to see how long it takes a solar panel to pay back its energy investment, or a windmil, or a nuke plant, in all cases it's a matter of months or less. The energy required to make one of these things (which will last a decade or more) is a rounding error, don't spend so much time harping on something of little or no relevance. This should be handled by considering net energy of the producer, which in any normal case will be almost exactly equal to gross energy produced by said provider. It doesn't matter how the energy swirls around (how much power does a coal plant use to pump around water? It doesn't matter, just figure out how many MW it makes per ton of coal and be done with it, water pumping gets lumped in with thermodynamic losses, and the cost of flushing the mens room toilet).

  2. But it takes oil to produce (insert thing here) that is needed to make (insert energy source here). True, kindof. If (as noted in #1) total energy required to produce the producer is a pittance, then even converting from one form to another (electricity -> hydrogen -> methanol -> whatever) is then 2 or 3 pittances, and hence, not of any particular relevance. Just subtract correctly at the end. If a coal power plant requires 5 barrels of oil, and oil can be made for x tons of coal per barrel, or y watts of electricity per barrel, just subtract one or the other from the inputs or the outputs, respectively.

So, to recap. When considering how much power a power plant (lets say a wind turbine here) produces, just take its output, subtract the energy cost to make it, being careful to convert from electricity to coal/oil/naturalgas with an appropriate conversion ratio, and there you are. That's how much it will produce in its lifetime, so if you need X times that much, then you neeed X turbines. Simple, easy, relevant. This is especially nice, because the actual resource in question (in this case wind) is now directly correlated to the output (electricity) produced. It matters not a whit whether the losses are due to friction in the turbine, or the energy needed to mine the metal to make the blades. Nobody cares, and it has no possible real world relevance.

Also, a word on conversion factors. In most cases, there is a plausible route for conversion. From electricity to gas or coal, it's easy to see that adding the electricity to the grid will spare X units of the relevant fuel from being burned, so there's your conversion factor. For the conversion to oil, either consider what would displace the required quantity of oil (oil heat, for instance), or consider something like conversion to hydrogen that would make more gas from fewer barrels of oil. Here you can go as hog wild as you like, it won't matter. Consider the ratio to be 1-1, 10-1, whatever, probably makes little if any difference.

That's my rant. In short, have some meaningful concept of efficiency (units of output / units of fuel), as it then becomes useful.

I'll end with the example of ethanol. Lets say its EROEI is 2 to 1. If I had a magic box with a 2-1 EROEI, we would have no troubles at all. Pour in 1 gallon of gas, get 2 gallons out, that's fantastic, but ethanol sucks, why? Well, the thing of relevance isn't EROEI at all, but rather pounds of feedstock per unit of output (or acres of land, whatever). That's the problem.

If there was a system that had a terrible EROEI (1.0001) but used some entirely useless material (I don't know, sea water or something) it would be a godsend, but something with an excellend EROEI (10,000:1) that is fueled by 5 year old children isn't really useful. EROEI has absolutely (positively, without a doubt, I can't stress this enough) no relationship with the utility of a process. None. None.

One more time, NONE.

The obligatory response.


First of all let me point out our common ground.  

After subtracting any resources needed to extract the coal, this coal seam will produce a net gain of X tons of coal." Completely relevant, and actually useful.

This, by definition my slaphappy friend, is net energy analysis. EROI is a tool of net energy analysis - the comparison of different energy harvesting technologies.

Do the calcs to see how long it takes a solar panel to pay back its energy investment, or a windmil, or a nuke plant, in all cases it's a matter of months or less.

This is plain wrong. Would you accept National Renewable Energy Labs (narrow boundary) estimate of 3-4 years for solar panels?  For energy payback time to be a 'pittance', on a product with a 20-30 year life, the EROI would have to be over 100:1

the example of ethanol. Lets say its EROEI is 2 to 1. If I had a magic box with a 2-1 EROEI, we would have no troubles at all. Pour in 1 gallon of gas, get 2 gallons out, that's fantastic, but ethanol sucks, why? Well, the thing of relevance isn't EROEI at all, but rather pounds of feedstock per unit of output (or acres of land, whatever). That's the problem.

We partially agree here. The other factores needed to scale ethanol are a big part of the problem. But society has been built around an energy dense infrastructure requiring high EROI fuels - we cannot replace oil with something 2:1 unless we divert resources from other productive areas of society.

EROEI has absolutely (positively, without a doubt, I can't stress this enough) no relationship with the utility of a process.

Well does your paycheck and the withholding tax rate on it suggest anything about how well you do your job?

I sense our disconnect is in the following area. Peak Oil is a declining net energy story. If we have a finite amount of an energy feedstock, we should care how it gets used. Best use is to invest it into future energy harvesting technologies. I would say you might agree with this.  EROI is just a tool to do the comparison, quality corrected, of such choices.

Well said.
Since you brought up the econonomics, here's your analogy for EROEI. I have an investment, put in $10 and some other assorted things, and the investment will pay you $50. So, is it a good investment? It seems like it's a good investment if and only if the other assorted things cost less than $40, an more importantly, if you can actually get them. The EROEI on this process would be 5:1, but that is the least relevant thing you can say about it.

Also, if the investment had a horizon of 50 years until payback, then maybe that also has a dramatic effect, but it is also completely uncaptured by EROEI. The only thing EROEI captures is the part of the equation that factors out (the $10 before is just subtracted from the $50 after), which is the only thing that is not relevant.

Sorry man, talk all you like about liquid fuels infrastructure, but you have a fundamental problem. You're basing your analysis on a number that is entirely useless. What if the process above created $100 in output, but required $60 in input (and the same other stuff)? Nothing has changed, it's still only good if the other stuff costs less than $40, but now the EROEI is 10:6, wildly different. In fact, any EROEI > 1 can be turned into any other EROEI > 1 (even infinite) by just factoring correctly. This is a useful number how?

You would then say things about "Well, you just consider the cost of the other stuff...." Sure, but eventually you get to the point where the "other stuff" doesn't have a well defined cost. How expensive is an ex-president, not really clear, is it? Same thing with EROEI. What's the energy cost of a ton of topsoil? How about a million acres of habitat? There is no straightforward conversion here, but this is what is really being consumed by the process, and it's completely uncaptured by EROEI.

Also, some more direct responses...

"Well does your paycheck and the withholding tax rate on it suggest anything about how well you do your job?"

Maybe, very vaguely, but it says nothing about what job I do. If you want to build road it might take you $1,000,000 in wages, so you could just hire several of me, right? Not so fast, what makes you think I know anything about building roads? You say a road takes X slaphappies to build, but this isn't really a relevant thing to say, as the number is probably far higher. I'm not as good at building roads as I am at other things, so your analysis is fatally flawed (probably by a factor of 100 or more) already.

"This, by definition my slaphappy friend, is net energy analysis. EROI is a tool of net energy analysis - the comparison of different energy harvesting technologies."

No, not true. You see, EROEI has this whole "energy invested" part, whereas net energy analysis factors out all the invested energy. If your product is energy, all invested energy is subtracted from the outputs, so the EROEI of this sort of analysis would always be infinite (anything nonzero divided by zero). This is the root of the problem. If one of your outputs is also an input, then by just passing through input to the output (or the other way) you can make this EROEI number whatever you want, so how can it possibly capture anything meaningful about the process? An ethanol plant can have its EROEI go arbitrarily close to 1, or arbitrarily high just by adding a pipe here and there to take things from one tank and put them in anohter, but it has NOTHING to do with the process itself.

If you run all your farm equipment on ethanol, then its EROEI is infinite, will that not be good enough? What is the  actual problem? It has nothing to do with EROEI, so what is it, and why do you throw around a meaningless number?

"Would you accept National Renewable Energy Labs (narrow boundary) estimate of 3-4 years for solar panels?"

Well, this NREL estimate isn't really good to use in that way.

If you look at the citations, you'll see that they're all quite old, except for one from 2004 that deals with peripheral issues - they range in age from 6 to 18 years! That’s the stoneage, as far as PV is concerned.

Why haven’t they updated it? Well, first, they included “recent” and “anticipated” figures intended to partially address this problem.

2nd, they never anticipated that anyone would care about EROEI after it was shown that it was roughly 10:1 or better. It just seemed to answer the EROEI question sufficiently. I’d be surprised if anyone has done any later studies, because why waste your time if the question is answered?

There is another problem with convetional EROEI analysis: it doesn’t account for quality of energy. In this case, the process heat used to melt silicon would only generate 1/3 as much electricity if used for that purpose, so the EROEI is understated by a factor of 3. Why doesn’t the NREL try to address this? Because, again, it didn’t seem necessary.

The fact is that PV EROEI has continued to improve every year, by perhaps 20% per year, and the improvement is accelerating. You may be aware that polysilicon supplies have been a limiting factor for PV lately. Well, they’re by far the most energy-intensive part of PV manufacturing, and in order to deal with the shortage manufacturers have been aggressively reducing the thickness of the silicon wafers. 300 microns to 200, 200 to 150, and so on. Silicon supplies have grown by perhaps 10% per year, while cell production has grow by 40% - you can see the effect this has on silicon and energy content.

Finally, thin film energy content is plunging. For instance, Nanosolar is building a 400MW fab plant in San Diego which will have an electrical supply of 4MW. That suggests that the manufacturing energy contribution (usually by far the single largest contributor) will be paid back in about 3 weeks ((4MW/400MW)/18% capacity factor x 52 weeks per year).

In closing, I would suggest that if you want to use this as a reference, you should definitely use the “anticipated” figures - as we are definitely in the time frame for which they were intended, and that would be true to the intent of the authors. I think that they are now too high, but I can’t provide you with a reference, and I think the NREL “anticipated” figures should be sufficient.

Does that make sense to you?

Rocket fuel almost certainly has low EROEI << ethanol. For example which some would say is an essential fuel to make our military-industrial system work.

The transition to lower EROEI suggests we should be investing a decent percentage  of remaining fossil fuel energy towards lower yielding sources in the future; examples are smelting of silicon and metalworking of wind turbine towers. I believe the current FF energy 'set aside' is too low and needs something like a WW2 commitment. In my opinion the market can't foresee the looming urgency of the problem and we will find ourselves with massive underinvestment post peak.

Just a couple errors/omissions/misleading items in the story:

  • Steam trains are thoroughly obsolete.  If we ran out of petroleum to run trains, the easiest thing to do would be to convert the diesels to burn powdered coal.  This is not clean, but it's been done and it works.

  • Electrified trains could probably be run on coal more cleanly (excepting CO2 emissions) than engines using on-board diesels.  IGCC can run cleaner than any engine on a vehicle.

There are quite a few possibilities above and beyond this, such as electric engines powered by batteries of some type.  If the refillable zinc-air cells come to market, they'd be perfect.  They would allow the rail system to run without wires and get their power from anything between wind farms (elecrolytic regeneration of zinc metal from oxide) and charcoal (chemical regeneration).
Nice post, Dr. Reynolds.  Thanks for the overview of the amount of energy used in our larger systems and the difficulties in evaluating EROEI.  This can also occur in subtle ways with small, everyday decisions.  3 years ago Barbara and I bought a new, top-of-the-line, energy saving refrigerator.  The reasoning was clear:  Energy savings alone would pay for the new fridge as the old one was over 10 years old and an energy hog.  Well, it didn't work out that way.  After 3 years we had the main electronic power board fail in the refrigerator.  To repair the machine cost us $69 for a service call, $80 for labor to replace the part, and $125 for the part, for a total of $274.  The repair wiped out the dollar energy savings we expected with our new machine.  The old fridge didn't suffer from such complexities and frailties.  It will be a lot harder for us to really reduce our energy use than we think, if my fridge experience is any guide.  To be truly successful the new refrigerator would need to be as reliable as the old one, as well as being energy efficient.  
A couple of comments.

Maglev is an energy disaster; high capital & high consumption.  The Concorde of railroads.

I do not support high speed rail (see TGV in France, ICE in Germany, bullet trains in Japan) in the US.  Seperate ROWs for pax & freight are excessive and demand, time and US capital investment is unlikely to be there. (The money proposed for HSR between San Francisco/Sacramento-Los Angeles/San Diego could build out about every Urban Rail proposal in California (hundreds of miles in LA Basin), and it would move zero freight).

I have proposed a mix of 110 mph top speed (100 mph avg) pax trains combined with 100 mph top speed (90 mph avg) medium and light density freight (electrified).  Pax service would be focused on regional service (<250 miles pre-Peak Oil, perhaps <500 miles post-Peak Oil) and not cross-country.

Link major cities within 250 miles, like the NorthEast Corridor.  A good example would be DC-Richmond-Charlotte-(spur to Atlanta)-Jacksonville-Orlando (spur to Tampa)-Ft. Lauderdale-Miami.  Pax service between city pairs, freight longer distance than most pax.

Not nationwide, not universal, no (or rare) stops in small towns slowing down service for minimal pax.  Express high value rail freight (currently carried by truck or air) would "pay the freight" with pax service as an extra.

I took a tour of the new Greenbush commuter rail line under construction south of Boston after the ASPO conference.  The energy required to build it would be comparable to a two lane road of equal length.  Very little maintenance for at least 50 years and perhaps 100 years (switches and a few other pieces will not last that long IMO).  Basically a century long investment.  High EROEI :-)

Improvements over older commuter lines are cracked granite ballast, concrete ties, heavier rail, high speed switches.  All but the switches add significantly to service life, and EROEI.

Best Hopes,


A couple of comments:

On the general issue of systems engineering, or systems analysis, or EROEI, my background is in computers and I have the scars of the hard systems analysis approach to prove it.

I agree with Joule, in that assumptions, value judgements, wild ass guesses, and just the cumulative effect often give the the wiegtings of numbers attached, very little meaning. What is more many of the assumptions may be disputed or not made explicit. For example it turns out that many of the wind turbines deployed in England (not Scotland) donot give the 30% utiliasation claimed but only 20%. We have encountered these sort of issues with wind in the past (cost of storage, cost of connection (Prof Cutler Cleveland

Which is why I tend towards the soft systems approach, as developed by Peter Checkland at the University of Lancaster.
"A more practical approach is to use SSM to generate HARD questions which can then be dealt with by the, more traditional, HARD methodologies."

i.e. Use soft systems (metamethodology) to make the world view (Weltanschauung) and other assumptions explicit.

Now on to trains (or not trains in the case of maglev).

Wasn't Concorde beautiful engineering ?

I support maglev in the UK, see uk ultraspeeds site (I have no financial or other interest in UK ultraspeed).

The reasons are complex, and not just peak oil related.
If you want a peak oil related solution, Alan Storkey
(An economist) suggests puting coaches on the national moterway network. (Sort of trains on Motorways, as the current rail network is oversubscribed).

1. We already have in the UK what Alan is proposing for the US.

1.1 A Passeneger service at approx 125 mph.

1.2 In the UK the PAX (at 125Mph), and regional rail (including commuting) absorb capacity on both regional and national strategic routes. Far from freight 'paying the frieght', it is reduced to low value bulk commodities like coal, with higher value goods going by road or air (including the mail!).

1.3 In fact there is talk of a dedicated frieght North-South route.

2. Travel is time dependant. 311mph is more attractive than 125mph and even the 186mph of the TGV.

2.1 People will not leave thier cars or planes for solutions that do not give a significant advantage. Speed matters.

2.2 Other large countries (by population) have invested in High Speed Rail (PAX only), the UK has not.

2.3 Moving PAX off conventional rail frees it for freight.

2.4 Maglev competes with air travel (and cars) for (light, high value) freight and people.

3. Maglev competes with Air Travel, but is more like rail.

3.1 See the fact book

3.2 Maglev does not have capacity measured in hundreds of seats (or less) like aircraft but capacity of 1000+ per train (1,200 for UK Ultraspeed).

This also applies to Rail in general including Light Rail

3.3 In the UK travel time is comparable to air travel.

4. It is about regional/economic development.

4.1 You will find references to the Northen way. If you look at it from a transport point of view, the North of England, and Scotland are a long way from the centre of Europe (which is moving East with new members). It only takes 2hrs 35 min to Paris from London via TGV.

4.2 It is only by using maglev you can get a similar journey time to Glasgow.

4.3. Maglev for the Northen way (including Scotlands central belt) connects cities, to make a logical conurbation equiverlent to London in size.

5. It's expensive !

5.1 Compared to road, not unless you are prepared to go the coach route (and that is utilising existing investment).

5.2 To high speed or even mid-speed rail.
Not compared to trying to increase capacity on the existing network

"On a like-for-like basis each new seat-kilometre of transport capacity created by 500km/h Ultraspeed will cost between 83 and 97 pence. Upgrading the West Coast Main Line railway to only 200km/h cost between £8 and £32 per seat-kilometre of capacity created.[1]"

5.3 It requires less land than road or rail.
"A maglev line requires up to 8 times less land and around half the vehicle fleet compared to high speed rail. Its modular track design makes construction costs more predictable"

5.4 The environment stupid.
"In terms of land take, noise and emissions en route, maglev creates less all-round environmental impact that any other transport system, including rail. If powered by electricity generated at today’s power stations, Ultraspeed will emit around five to eight times less carbon dioxide than air travel and three to four times less than cars for the same transport capacity. If powered by the cleaner low-carbon electricity we are increasingly likely to see in the future, Ultraspeed can operate with much lower, or even zero, emissions. (And all of this whilst consuming up to 45 times less land than motorways.)"

6. No it's your Weltanschauung stupid!

Someone said it was a liquid fuels crisis (it is), we also have an short term energy crisis. Maglev does not use liquid fuel, maglev has a much higher carrying capacity (1200 per train, train every few minutes) than road or aircraft. Maglev meets the need of PAX better than the existing network while freeing up capacity for lower speed rail to do what it does best.

Good luck getting people on to coaches unless they have no choice.

And what are economists useful for (tinder ?).

Too little time, but Maglev is the Concorde of rail transport. Energy hog supreme compared to 125 mph or even 300 kph high speed rail. Expensive to build, sound issues require wide ROW (useful for some types of farming, so not a total waste).

I found the Concorde to be quite ugly engineering. A horrid waste of resources for no benefit.

BTW, they are not going to dig another Chunnel for maglev, and the Swiss are not going to drill another 57 km tunnel for maglev either.

Your last comment showed a complete disregard for cost-benefit analysis, and that is a pre-condition for supporting mag-lev. Society will simply not join your support for maglev "no metter the cost".

best Hopes for Realistic Planning,