Your DollarShadow
7/05/07 9/28/10 12/3 1/8/11  Research notes on How to Understand the True Scale of Your Energy Use

Five True Energy Budget Stories: 
1) What AVG means 2) Scale of Hidden Energy  3) PV area  for AVG$ 4) Suburban Home 5) Consumer Products 6) Global PV Footprint 7 IEA Data & Trend

for example - Volume of CO2 per $ GDP :  0.46 kgCO2/$  *  2.1m3/kgCO2  => 0.97 m3CO2/$ of GDP
so if you make $50,000 a year that's 50,000 cubic meters of CO2 produced for you as a world average, the volume of ~20 three story townhouses...
Pollution is often removed by hiding it, with outsourcing or reducing visible part, quieting the messenger so its silent, but still deadly.


1. What AVG means:

The world average energy use per dollar of GDP is the amount of purchased energy resourced from around the world to deliver an average $'s value of goods or services.   Because it's so much larger than the visible energy uses of commerce it is a much more accurate measure than even very carefully tracing visible energy uses (f).

To measure individual economic impacts on the earth $1 1 average share of the world economy, its total energy use and all its impacts, since every $ needs the services of the whole economy (i) (ii)... Which turns out to be surprisingly accurate for energy use (f)

AVG = 8000btu/$ (or 2.4kWh/$) (IEA 2006 world TPES/GDP-PPP) = the usual accumulative total energy demand of money   

i) Average values are true for individual cases "on average".. and most products require such highly divers energy uses to deliver them it's also rare for the energy required to deliver the $1 of value to be far from average (f).
ii) I finally found people interested and able to study  the sources and equations to confirm these figures from Table 1 below, in a discussion on Azimuth with John Baez.  There was one "small" error found, now corrected, made in revising the whole calculation for this update of the old work.

2.  The scales of visible and hidden energy use:

Why our estimates of our energy use are usually Way off!

Systems Compared Financial Budget Visible Energy for Fuel Use Energy Demand  for Economic Services Total Energy Demand Units The error in believing your visible energy use
Wind Farm (per kW size) (iii) $213 0.10 0.45 0.58 kW 532%
Family in town w/car (iv) $50,000 120.95 105.48 226.43 mW 187%
Family "off grid" w/car (v) $52,000 101.26 109.70 210.96 mW 208%
Family "off grid" walking (vi) $52,000 5.13 109.70 114.83 mW 2237%

Each dollar of a family budget has an energy impact by requesting services from people who will consume energy in delivering the service, making the purchase of services an energy burden and source of impacts on the economy and environment.

iii) The SEA study of a wind farm showed that the direct energy for producing wind energy was only about 18% of the business energy demand for economic services.
iv) A modest income family in a 2000sf house for might pay $2500 for heat and electric and $1800/yr for gas  in the car for a short commute and running around.
v)  The same family, generating its own energy by Wind or PV might be able to generate their own heat & electric, a little less efficiently than commercial energy
vi) The same family, walking to work and store, generating their own energy and just using a taxi now and then.

Your $'s as an area of land, for collecting the energy

If the economy got its energy from The Sun, every dollar would... cast a shadow.  
comparing the energy used by the economy to the areas of land needed to collect the energy, per $ of GDP


3. Calculation of PV Area

- "average" solar radiation availability picked from radiation map below,  note: mismatch between population areas and solar radiation availability -

Table 1.  
Solar energy available for PV (a)                      meters^2                                                       feet^2
(See Figure A)       kWh/day 5000 kWh/m^2.dy = 465 kWh/sf.dy
per year 1825 kWh/m^2.yr = 169 kWh/sf.yr
BTU/year 6229 kbtu/m^2.yr = 579 kbtu/sf.yr
High Performance PV          
* reduce for efficiency 0.19     110 kbtu/sf.yr
** reduce for coverage 0.5     55.0 kbtu/sf.yr
*** reduce for EROI cost     45.8 kbtu/sf.yr
At avg. btu/$ GDP (b or Table 3)          
world GDP (PPP) 2008 63865 Billion $      
world energy use 2008 512285 Quad btu's      
to make $1, at world avg intensity 8021 btu/$ *      
Energy content of oil 138000 btu/gal      
energy equivalent of oil per $ 6.5 oz/$      
Area needed per $          

to make 1$ per year

0.016 m^2   .175 sf **

to make 1$ per day

5.94 m^2   63.9 sf**

* A nominal peak efficiency of 19% is now achievable for solar panels. It might be 50% in the future, but at high cost.
** A rough estimate that solar panels would cover 50% of the land area in a solar farm
*** The EROI (energy return on energy invested) of 6:1 is a rough estimate using SEA (d)  for the "usual" net energy produced by high performance PV solar farms in the US economy.    It seem likely to be optimal rather than average, but would take study.   The value would be quite sensitive to the costs of the land, operations and equipment, as well as the average radiation from the sun at a particular location.  The value of 6:1 means that 1/6 of the energy gain is lost in the effort to produce energy.

Figure A. Annual average daily solar radiation available in the US overlaid with 1990 US population density (a, e)

One reason to think of our impacts on the earth as proportion to the money we earn and spend is that every dollar is one equal share of the total product of the whole world economy.   All economic impacts are energy impacts, and all the things that cause economic impacts are purchased.  For energy use the "average energy per dollar" also turns out to be a fairly accurate estimate of the real total use of energy and production of CO2 pollution, seemingly +/-15% mostly.    What you can be sure of is it will be fairly close most of the time, and will be a far more truthful measure of the scale of impacts for  any purchase choice than zero.

Why it's fairly accurate most of the time is because energy is the universal resource, and traded around the world at a world price, as discussed a bit further below.   The best present research on this method of analysis is a draft paper Defining a standard measure of energy use for businesses, using a wind farm and its energy return on investment (EROI) as an example.   By switching from adding up the receipts businesses collect from purchasing energy, to treating the prices of things as a receipt for average energy use in all the work that went into what was purchased, four times the energy uses being employed are found.   The main reason is  ... *because we pay for them* ... and the price we pay is an accumulation of all the prices reflecting the work of all the people and machines along the way that brought the product or service to us.   That's the key.   It will change how energy accounting is done entirely when it is appreciated that economic footprints are far easier and more accurately measured with money than trying to follow almost untraceable threads connecting diverse causes and effects, as well as the significance of finding a 500%+/- typical error in our primary measure of sustainability....   9/28/10

4. Suburban Home Example:


Budget for an "average" suburban home with a 1500sf footprint and family income of $75k/yr,


Energy budget

CO2 budget

step 1 Energy/GDP => $1 share ≈ 8000btu/$ (or 2.4kWh/$ ) CO2/GDP ≈ 16oz/$ (or .47kg/$)
step 2 Using Solar Panels to produce a $1 share  ≈ .175 sf for a year (table 1).  needing to burry purchased charcoal to  sequester the carbon costing $.16/kg (iii)
step 3 needs 13,125 sf area for PV panels, produces =32.3ton CO2 in global fuel uses
step 4 needing a solar farm ~9 times the home's roof area,
to be truly grid neutral
spending $11.3k/yr on sequestration
to be truly carbon neutral

iii) Using Charcoal to Fix the price of carbon emissions assumes no continuing land use cost to protect the buried carbon
iv) "truly grid neutral" and "truly carbon neutral" are a quick way to say that the family provides as much energy and carbon sequestration for the economy as the family takes from the economy as a whole, "on average".  "Normal" economic growth of 3.1% per year would increase the PV area and cost of CO2 sequestration ~1.8% per year, slower than the growth rate due to normal efficiency improvement.

5. Consumer product Example:

Money spent on a soda or glass of wine, like a receipt for a share of world energy and carbon costs. As a measure of energy use it's quite accurate, on average, and makes spending or earning a dollar responsible for one equal share of all the impacts of energy use too. You can appreciate the meaning of "average share" using it to measure how much CO2 (your carbon footprint) is created with normal expenses. Try to compare the benefits and the costs with these two examples. (b Table 1)

So... a 16oz soda for $1.50 would put ~24oz of CO2 into the atmosphere, as average spending  from conventional energy sources, and use  the equivalent of 10 oz of oil.   To use solar energy it would take the use of a 100sf of a solar farm for a day to generate the same energy with PV.  

And... a 6oz glass of wine costing $6 would produce ~96oz (6lb) of CO2, 16 times the weight of the drink as average spending for conventional energy sources, using the energy of 39 oz of oil,  or... require the use of a 384sf of a solar farm for the day to get the same energy from PV

6. Global footprint Example:

The global "dollar shadow" for using PV to supply the present world energy demand, estimated in %'s of the World Agricultural Land Area

I  did the estimate using the PV values from Table 1. and agricultural land area figures from (c), and IEA economic statistics (b) as shown in the table above.     That gives 1$ of economic product needing 5.9 sq meters of sunshine for a day, or .016 sq meters per yr, to deliver the world average amount of energy per $.    It's rough, because the solar radiation varies so widely from place to place, so to be accurate it would need to be averaged over all the land area where people might use PV.    I then use the historic energy use "growth constant" (b) of the past 40 years,1.89%/yr  which results in a 37 year doubling period for energy use.    World GDP $46.9 trillion in 2000(b) that would have cast a dollar shadow of 15,490,000 sq km,  with world agricultural land area 48,033,854 sq km as 40% of the total land mass of 120,000,000 sq km.

Table 2.

World dollar shadow as a % of present total agricultural land and the world landmass

% of Agricultural Land in 2000 



as % of landmass in 2000


% in 37 yr  



% in 37 yr  


% in 74 yr  



% in 74 yr  


% in 111 yr 



% in 111 yr 



Does anyone see a problem with this?? 
..or with needing to locate the solar farms near the people?? ..or with the cost of renting the land for it, ..or what's wrong with the figures??


7. Global Trend Data and Relationships:

 IEA World economic data - Key indicators GDP Energy & CO2
partial 2008 data for Table 3. and Figure 1.,  for 2010 IEA source data see reference (b)

Table 3.








% change


Mt CO2 Sectoral method




















TPES   (QuadBtu)










TPES  (QuadkWh)










GDP   (billion 2000 US$)










GDP PPP   (billion 2000 US$)




























































Population   (millions)










kg CO2 / GDP 2000 US$










kg CO2 / GDP PPP  2000 US$)










Ton CO2 per capita)










Mton CO2/QBtu











 x - original data converted to Btu & kWh units
* energy/$ values (economic intensity) declining over time at 1.3%, ~25% in 30 years
** ratios of GDP adjusted and unadjusted for "purchasing power parity" PPP to remove monetary exchange rate discrepancies, for the US unadjusted GDP$ is only ~5% higher than its PPP adjusted value. 
*** For climate change the key observation is that steadily improving producer efficiency completely overcomes consumer efficiency, to continually multiply energy use, it also does NOT change the share of fossil fuel used to produce energy at all, as reflected in the near constant CO2/Btu ratio... Info on CO2 from Wikipedia

Figure B (f, h)


a - National Renewable Energy Laboratory,  Historical average solar radiation at the ground:(picking a "typical" average mid latitude location)
b -
IEA World Highlights:  & 2010 historical data: 
c - Agricultural land use figures
d - Method for measuring whole system energy use and EROI: System Energy Assessment (SEA)

e - ThoughtForm population map
f -  see System Energy Assessment (SEA) measuring EROI for Whole Business Systems, Henshaw 2011 extending a 2009 talk
h - understanding why the economy behaves as a whole The curious use of Stimulus for Constraint, Henshaw 2011

Phil Henshaw   id at synapse9-com 

The following is for record, from my pre 2009 approach to the subject. 
There's good content, notes and references, differing only slightly from above, and some useful complex reasoning too.  

ed 7/05/07, 7/07/07, 7/08/07, 8/10/07, 8/11/07, 9/13/07, 10/29/07, 11/29/07, 12/4/07, 1/3/08, 1/20/08, 3/17/08, 5/24/08, 5/3/09
Understanding a dollar's shadow on the earth

The Science How to Use It CO2 Inventory resources References - Discussion -
Correction & Edit notes -
9/05/07 10/29/07 1/20/08

Your Energy Footprints & Physical Shadow on the Earth

Measuring the whole impact of your Choices

tatement of the principle:

What's most untraceable is what people independently choose to do with their money...But you know it's in the total!

The economy requires energy to deliver the goods and services we buy using diverse supporting services in your community and around the world.    Average spending is then going to be responsible for an average share of it (see details below).    That's the principle.    An average share of the total energy the world uses is  ~8000btu/$ (1995$) according to the US DOE.   Due to inflation and improved efficiency that’s ~6000btu/$ in 2008.   To help convey how much energy that is, think of the solar energy that could be made into electricity in an hour with a high performance PV panel measuring 100ft on a side.    That's around the same 6000btu.   This estimate is based on a 40deg north latitude location, averaged over a 24 hr, 360 day period, with normal weather, with 18% collection efficiency as if for expected future high performance solar cells (1)(4)(corr3)(corr4)

The ethical and moral choice is fascinating.   By paying for products we also choose to directly request and pay for the whole diverse web of things that went into them.   We consume the product of those contributions to what we buy, and see we’re physically responsible for it.   We have not traditionally thought of being ethically or morally responsible for it, though.   Usually we feel ethically and morally responsible only for our own personal acts, seeing other people's acts as their own responsibility.   Now it turns out that effects our choices and we want to have an effect on how our choices affect our world.  Unless we know the whole impact, though, we can’t make effective choices about it.    When we have better information we can not only make better choices, but have them be effective.   It's also an opportunity to extend our own ethical and moral responsibility much further into the whole system of the world, if we choose.

This graph shows an overlay of two of the figures from reference (1), showing the world trends in GDP and energy use Intensity.   The differences between the developed economies (OECD, red lines) and the rest of the world’s economies (blue lines).  That the two follow similar curves, and that money flows fan out extensively is part of why all spending is assumed to be average unless otherwise determined.

The critical question is whether treating all spending as having average energy content unless shown otherwise can be broken into several parts.

1.     Why is directly measuring the total contributing impacts difficult?

2.     Does the uses of money always distribute very widely through the normal uses people make of it?  

3.     Can you adjust the implied average for known measurable impacts?

4.     Are there hidden high or low impacts embodied in some choices that might introduce errors?

5.     Is there any other way to estimate the error in counting only the easily visible impacts?

6.     Will better statistics, less dependent on theory, become available as people use this approach?

Each of these could be an essay, but first I’ll try to answer each simply, and then treat the remainder of this attachment as further discussion.

1.     It is quite hard enough to find out what a project’s direct energy uses will be, especially during design when decisions are being made.   It is also not actually possible to add up the things no one keeps records of, and that includes the majority of the spending.   The majority of spending for things goes to the people who assist in delivering them, all the way down through the supply chain.

2.     Yes.  If you just think of all the very many ways you distribute the money you receive for the work you do, and then of the ways those people distribute the part they get from you begin to see.   The product you help make costs $100 and the business passes parts of that on to each of 1000 diverse kinds of contributions to delivering it, including yours.   If each of them does what you do, spend their share on 1000 diversely different kinds of things, the one product choice is responsible for enabling 1000 times as many other choices at each step.   In three steps that’s a billion choices, in four a trillion.  It’s likely to equally support all the different kinds of uses people do in proportion.   There are also some other issues that touch on, but do not alter, this conclusion.

3.     Yes. The normal accountable energy to be factored in is the electric and gas bill and the gallons of gas and things.  The simple rule of thumb I use to get the right scale of adjustment for these hard measures of fuels is to just add their btu equivalent to the total.   You’d think, perhaps, of factoring them in, adding their btu equivalent while subtracting their cost from total.   The odd thing is that the money you pay for gas doesn’t go to nature for the flammable liquid, it all goes to other people, who use it to consume things throughout the economy, and incur average impacts from that.   

4.     Not that I know of.

5.     What will happen is that businesses will see the need to reduce the energy content of their supply stream, and pass on their locally lowered impact intensity to the consumer, so they can sell things at a higher price for lower impacts.   They will need their suppliers to do that, and pass those savings on to them.   It will result in their whole supply chains passing detailed energy intensity information along, making the end user choices ever more effective.  


Those energy uses that are spread throughout the economies is what this measure captures.   In the end, most of the untraceable energy uses come from the money you give to people.  It causes the current energy impact models to miss 90% or more of the real energy costs of what we do (14,15).  This interpretation  was looked at by the life-cycle impact economist Wayne Trusty (author of the Athena life-cycle impact tool) and found to at least be theoretically correct.   Measuring your energy use as a share of the whole provides a true statistical measure that lets you see the difference between what is and is not accountable.  It gives you  a) the real scale of your energy choices and b) a guide to locating where they're hidden and why they're growing.   For most US home owners the $shadow height of a collector to support their lifestyle would be the width of the home and over a mile high (sim 11).   That's more than large!    It's also not in our control, and so a little home efficiency won't touch it.   There's one thing you can do.  Take the material small steps that lead to a new future.   The only feasible way to compensate for such large excesses is to contribute to our finding new ways to think and act in the future.  We need a different way to measure luxury than in terms of multiplying money and energy use.

[Double counting note: This way of calculating energy & CO2 impacts works because it counts the whole cascade of contributions that occur as a consequence of spending.   It measures the whole effect of choices.   That also means you should take care to not double count contributions.  If the cost of your salary is counted as part of the costs of your company's products, the two should not be added.  There's dual responsibility in that both you and the people why pay you are responsible for the energy consumed by the money you spend.   Economists are careful to not double count what they include in GDP.   You would use their same method to count whole environmental impacts using this tool so they can be added without overlap.]

The scientific idea: The 'embodied energy' (or 'energy intensity') of any product or service is the sum total of all the energy uses needed to provide it.  The problem of adding that all up is that in an economy most of them are unaccountable.   Driving a car both burns energy in the engine as well as in making and maintaining the car.  Keeping insurance for it, supporting the gas station as a business and the consumption of the people at the refinery are also all in there.  Your choices are responsible for energy consumption of very many kinds throughout the entire network of people that take part in bringing you what you purchase.   The money you give them supports both the energy consuming work they do at their jobs and also the energy consumption of their entire lifestyles.  It's not possible to count up since you can't ask them what they do with your money and they wouldn't be able to give you useful answers anyway.  It's actually prohibitively difficult to trace, and so while while not at all invisible, it also remains completely 'unaccountable'.   

The real problem is that the unaccountable part is so much larger.  The distribution has what is called a 'fat tail' in the sense that most of the embodied energy for products is located in the tiny contributions scattered beyond your ability to identify them.  These sources remain 'hidden' because the information gathering task is too difficult.    Using the average value for all spending to estimate the energy diffusely consumed throughout the system is a great shortcut, particularly for getting you to look at the difference between what you can and can't account for.   More work will find more exceptions, but it seems quite likely to be very accurate for most spending, simply because of how widely people distribute money, from one source to many many destinations.

Understanding that a $1 apple purchased in New York, has a hidden energy cost equal to a $1 share of the energy used by the whole economy that New York is part of, takes some thinking.    You need to add up all the little bits of energy use in the world that are required to bring that apple to you where you are in New York.  That includes supporting the farm and all the activities of the farmer and his family, all the goods and services resourced from all over the world to support the work of farming and the consumption that the farmer's whole family relies on the money for.  As you count it up it becomes clear that it's the whole economy that is delivering that apple as a $1 product.   There's also the important insight that most kinds of products are essential companion products of others.  Part of buying an apple is the service of the whole city's systems in bringing it to you, made possible by your having clothes and a place to work, the whole environment of interacting parts that make the exchange possible.   It's good general reason to accept that any part of the whole economic system should be credited with it's share of the whole system's impacts.

Still you might say, it just doesn't look like money and energy are the same thing, so how can they be equal?   There are two things that have been hiding their direct connection from us.   It's not coincidental that energy is nature's universal resource for making things physically happen and money is our universal resource for making things.   We only pay people, and consider nature's resources as free, but what we are actually getting when we buy things is packaged energy.   Yes, we only see the 'package' in a sense, but the whole process of making things consumes energy and so that becomes 'embodied' in the product.   The most specific reason seems to be that when we choose to give money to people we select those who deliver products for the least energy.  That ties maximizing efficiency and a necessary amount of market determined energy content in every step of product delivery.  Globally that makes price a direct measure of energy use for every process.  That all the economies give energy about the same economic value and price and are improving their efficiency at the about the same rate everywhere, then, fairly assures that the relationships between money and energy will be uniform in the world and steady.  

Because energy is such a universal commodity, and flows to wherever it is most needed, it turns out that money has almost the same energy intensity everywhere.   The measures show that the economies employ fuels at about the same btu/$ efficiency rate in every economy, rich and poor, and that the trends of change in all economies follow the same norms.   This amazing evidence was gathered by the US Dept. of Energy in a 2004 study (1) and are further verified by the updated 2007 EU IEA data (6.1)  That all the economies behave as a whole in how they use fuel is another way to say why individual shares of GDP are a good direct measure of individual shares of global energy use.    Why the economies have consistent matching global behavior, treating energy as a universally interchangeable part with a universal matching $ value is harder to explain.  It takes an exploration of complex natural systems, from multiple points of view.   Perhaps the best shortcut way to explain is just that all of nature treats energy that way too.    Energy is the universal interchangeable resource of all systems.

Another way to understand it has to do with the 'liquidity' of energy and money and the economic principle of the 'flat earth'(5)(6).  Every money event and every energy event have ripple effects that spread throughout the economic system.   Some settle out quickly and some more slowly, but they all tend to seek a single common level, like ripples in a pool.  You can see this in the energy intensity curves for individual economies (1). Even though the individual economies are all are heading in the same direction they each do so in a different way.    You can also see it in the way economies sell whole market baskets of products, not individual ones.   No one product is either useful or producible without an extended network of 'companion products'.  For the energy intensity of spending to vary from place to place would require change in the whole network of companion products to constitute a local 'product space'  (16).   Logic suggests that choosing to buy products from product communities (product spaces) with unusually low energy intensity, for example, would only be possible from within them.   They probably exist and your lifestyle or community might develop a system learning path toward becoming part of them, but doing that is not a readily available choice for most people.  The economies are very thoroughly integrated on price alone.   You can think your economic world is local, but so much of the true network of dollar flows is probably global.    

It would certainly be nice to be know about low-impact product communities, how they might develop and what it takes to encourage them.  That's an area of research that's wide open it seems.   The obvious one is the loose idea of choosing to live and work in a 'green world', with everyone in the 'network living simply.   It's physically possible for that to work and for such networks to become stable evolving and self-sufficient.    There are also lots of hidden flaws in the idea that help explain why the way people normally think of doing it usually fails.   Product networks that separate you from the larger economy tend to wither is the main one.    There's also a lot of the 'you can't get there from here' problem.   Product spaces have natural whole system learning paths that enable or restrict their development and it would be good if more study was put into understand them better.   It seems very likely that reducing the energy intensity of spending while retaining high quality services requires  whole system change. 

That whole system change in energy efficiency is realistic and happening naturally anyway is evident in the DOE data curve above, showing a typical decay trend, and in the  35 year IEA world energy intensity data (6.1) which shows more detail.   A close look at the detailed world energy intensity curve shows a 'stairstep' shape to the curve, which seems likely to indicate alternating periods of 'retooling' and 'using' emerging kinds of systemic efficiencies.   Studying the subject would probably lead to a better understanding of whole system efficiencies develop.

How you might use it The first principle of systems learning is to just start with small accumulative steps.    Make sure you 'look around', find a few things to put 'in the box', including ideas for looking further, and add up the total.   Basing your choices on 100% rather than less than 10% of your direct energy impacts on the earth is good.  It's not an answer, it's a better question.    Better information about our whole impacts will make our choices, of all kinds, more effective.   It helps correct the flaw of economies that the natural systems on which we entirely depend are universally assigned the value of $0.   That's a very curious error, and not easy to correct.  

If you wanted to assign a value to nature, what would it be and who would you pay, anyway?   It just does not fit the model.   The economies do not recognize the value of their environments in much the same way as a formula is completely self-contained and can't change itself in response to changing behaviors of the world around it.   One thing we can do, now that we see that $=energy quite directly, is understand that defining 'good' in terms of the rate of multiplying $'s is inherently bad.   It's missing all the relevant connected values.   We need other definitions of good, ones that take nature into account. We need things like learning to use rigorous whole system measures to give the word "sustainability" more substantive meaning as an end product.

Using global impact measures is essentially just another way to guide project design to produce a better product, like the rigorous credit point checklists used in LEED, the energy rating estimator used in Green Globes, or my 4Dsustainability learning and evaluation model, or other methods.  The difference with accounting for the whole impacts of things is that you can then measure whether your adaptations have increased or decreased your impacts on the earth.   The point systems like LEED don't actually give you that information, though.   Consequently most traditional or green projects still produce large and increasing environmental impacts.   LEED just measures quality, and doesn't measure quantity except as a qualitative ratio, so there's not 'total'.   We've learned how to increase our impacts on the earth more efficiently... but that's not good enough.

The steps for using the $Shadow measure on a building project begins with comparing  the total energy use implied for average btu/$ spending with the particular energy uses you can measure.  That means multiplying an initial total project costs by 8000btu/$ and comparing it to the fuels and other things the project would consume.   Then you'd try to explain the difference and adjust your estimate up or down accordingly.    It can be done with either complete analytical rigor or just rounded up or down based on judgment.   You'd want to do this in a way that is simple at first and lets you come back to refine it.   Then you'd do the same thing for a baseline reference project costs.   That might be the prior use of the site or a prior service being replaced, something to compare in a meaningful way to give you the change in the earth, before and after doing your project.   Two ways of doing this for a sample project are in my resources (11, 14)

The next step is to choose one or two more additional total project  impact measures, such as using the $Shadow method for CO2 as well (fairly easy) or the Energy Star project energy estimator (fairly easy) and the EF global environmental footprint method (a little more work).  Then you'd have a picture of before and after for energy, for CO2 and for renewable eco-systems services.   You might also want to add onto that using the greenhouse gas inventory method of which is likely to become a reporting requirement for all businesses and the Athena life-cycle assessment tool for a complete impact picture (a lot more work).  

There's a real value of keeping things simple, and as rigorously complete as you can.  That's partly because you want to be able to adjust them over and over, and if it's too complicated it's not going to be useful.   It may be the most important lesson of all to recognize, in designing our new way of design we've been doing the opposite.   We've 'innocently' been stepping into a job of micro-managing whole environmental systems and our increasingly complex interactions with them.  Yes, that is partly forced on us by everyone being caught off guard by our massive interference in the earth's natural systems.   We also err in thinking that taking on ever more complicated problems is solvable.   It's not.   It means taking on ever steeper learning curves.   Hm...   Learning how nature does things of exceeding complexity, very simply, is on this learning path, but a ways along, so the immediate next step is to add a couple more rows to your project's whole impact measure chart (14,15).

The next things to look at are your adjustments.   With any estimating method you need to make adjustments to balance what's accountable and what's not, combining direct measures with contingencies for what's unmeasurable.  Cost estimating always does that, and impact estimating needs to do that too.  For any method you use if the 'contingency' is not there, you need to plug one in.   For the $shadow method the entire estimate is for 'unaccountable' energy costs, and you need to adjust it for the accountable 'non-average' impacts you can see.   The simple, and nearly correct way to do that is simply adjust the project $shadow by adding all the direct fuel uses as direct impacts.   That means if you can count a certain number of btu equivalents for electricity uses, just add it to the total, since the $Shadow measure large estimates the consumption of the people behind the product.  Other materials might have more or less than the average.   Concrete, for example uses direct energy six times the btu/$ of average spending, or around 48,000btu/$.   These kinds of rules of thumb are needed for a wide variety of materials and systems.  The only one I can think of that might significantly reduce the embodied energy of spending would is when the main value of the product is as art, and most of the 'value added' is in the artist's signature.

The next step really shows you why this needs to be kept fairly simple.   It's to compare your target scenario and compensations.   The target scenario could be to raise project quality in a way that makes reducing project impacts to somewhere around 1950 levels for the same functions.  That's very roughly the global warming objective.   One could pick a target in lots of other ways too, as Architecture2030 does, for example.    Then you compare that with your proposed project totals and that tells you how much you need to compensate for somehow.     Designing your compensations is the true creative challenge.  Make real, honest estimates of the beneficial impacts for project choices that would have direct community or environmental benefits or long range effects on the future.    A publisher, for example, might devote a portion of their staff hours to maintaining a community resource website for any target community they think might benefit from having help communicating, local or international.   The diversity of ways to effect the future is immense, of course.    The big one, and the toughest and most important, is helping people figure how we can stop having growing impacts on the earth as a whole world.   That's not just a lifestyle change, but true whole community learning, and the first small accumulating material steps are of extraordinary real value.

Any particular project would have different results, but for one five story project the $Shadow estimate was that it would take high performance PV panels 125 times the size of the building footprint to supply the energy for its combined operating and amortized development costs.  That's a multiple of 22 times the estimated footprint of the reference prior site use, small scale brownstones.   One compensation goal considered was to materially contribute, to the degree the project missed the target, to reversing energy use growth in the building's stakeholder community toward achieving the IPCC 80% CO2 reduction target.    Asking how to do that stimulated two main ideas, and a new path of learning sustainability in the stakeholder community.   One was that we could reduce the building size by finding a collaborator in the neighborhood to share some functions, and share the expensive centerpiece of the design in this case, so that expensive piece could have multiple uses.  That would greatly reduce both the footprint and the compensation target.   Then we also began looking in an open ended way at who the stakeholders in the project really were, and how their interests could be combined to create other value for free.   

One idea was to pitch in on the city sustainability plan, initially considering storm water retention to help prevent polluting runoff and restore the ecology, but also deciding to go well beyond the intent in use the LEED education point.  That would serve as one of the project's 'bright green spots' and a good research and experimentation opportunity.    Another idea was to influence the future by the project becoming a neighborhood center for helping people with their energy and CO2 inventories.   We also considered compensations in relation to 1) their lasting accumulative direct effect,  2) their value as important symbols, and  3) carefully examining them and avoiding those possibly having reverse effects.   Of course the plan included to make efficient buildings and measure progress with other measures like LEED.  The fact what can have the most effect is whole system learning, and accountable impacts like efficiency measures address may control only 10% of the problem, does change the picture.   In many cases it's hard to imagine how a building could effect the future, particularly enough to reverse it's own excess impacts.    Learning curves always start slow, though, with small steps.   What's important is the accumulation of steps and the quality of the learning.   It's what a finite, fragile, and truly beautiful blue ball in space seems to need from us.   



References - top

World economic statistics from US DOE, Source of the 8000btu per $ global constant, DOE charts for global energy productivity in a report on the global carbon trends:

1- website:  document:
2- World Highlights:

Conversions used from btu to square feet of earth at 40 deg N. lat. 15% energy extraction efficiency:
3- wind energy manual
4- bioenergy calc sketches

Basic references for World energy economics:

    - A great compilation from longer and better IEA and Angus Maddison Historical data showing the whole history of growth and
our present stair steps of improving efficiency in converting energy to wealth.
My (not updated yet) discussion of the DOE data's meaning for the long term sustainability of growth and my slides of the DOE figures to help in seeing the relation as a flowing change :

HDS Excel (rough templates for organizing these complicated conversions)

HDS Sustainable Design Resources
11- - comparing site use before & after
4Dsustainability Design Process model & Wiki

HDS TotalBalance  with CO2Inventory
Spreadsheet for projects showing
- Adjusted whole system impacts, prior and proposed, with future & compensation targets, for multiple measures
14 - - summary page PDF
15 - - model speradsheet

Other Research
16 - Product Space, a network model of the learning paths of natural economic product communities
17 - World energy & economic efficiency curves Current Trends - 2000 year World GDP, current GDP, energy & efficiency, OECD 1000 yr  projection

Physics Principles
20 - Principle of conservation of energy, that energy is transferred or transformed, not created or destroyed
21 - Principle of thermodynamics, corollary of (20), energy change  = energy transfer - 'work'
22 - Principle of limits - entropy, also known as principle of waste & decay, all energy transfer takes 'work',
23 - Principle of continuity, derived corollary of (20) & speed of light providing bounds of organizational development
24 - Principle of diminishing returns, the limit of perfecting all directions of progress, corollary of (22) & (23), Jevon's law

[Double counting note: This way of calculating energy & CO2 impacts works because it counts the whole cascade of contributions that occur as a consequence of spending.   It measures the whole effect of choices.   That also means you should take care to not double count contributions.  If the cost of your salary is counted as part of the costs of your company's products, the two should not be added.  There's dual responsibility in that both you and the people why pay you are responsible for the energy consumed by the money you spend.   Economists are careful to not double count what they include in GDP.   You would use their same method to count whole environmental impacts using this tool so they can be added without overlap.]

[Source data note:
- For CO2 inventory, the same DOE data(1) provides .57Metric Tons per $1000 (1995$), (or 12oz/$1) for average CO2 content spending.   The interpretation is similar and the averages still valid, but since energy sources vary in how much CO2 they produce, and CO2 is not a priced commodity (yet) CO2 content per $ will vary more, and so more adjustment of average embodied CO2 for non-average content would be needed for accuracy.
The DOE figures(1) are only for energy purchased as fuels, and omit the direct solar energy used by the economy.  That raises the broad question of  unaccounted 'natural system services' that are even more 'hidden' from the economic statistical measures than distributed purchased fuel uses that a $shadow measures
- The EU world statistics from the IEA and historical world statistics from Angus Maddison provide a more detailed picture of the whole evolution of energy consumption and our present stair steps of improving efficiency in converting energy to wealth (6.1) ]


Correction & edit notes -  top

First POSTED TO THE A.I.A. Committee On The Environment Forum - 7/05/07
Archive at  

1-  The DOE figures for total energy used by the economies do not include any solar contribution, only purchases of fuels.   That includes some renewables like hydro power as purchased electricity, but ignores the solar energy in growing corn.    This means that the actual embodied energy per dollar is higher, and somewhat harder to calculate.   Some figures are kept by the International Energy Agency in Paris does (, but these have not been investigated.   The more interesting question is whether including the solar contribution would change the decay curve shape of the historic btu/% curve that seems to say that the economies have long been approaching an asymptotic limit for the economic values of people.    It’s unlikely that that would change, but the question open.     9/5/07

2-  After some delay.. I finally got around to proofing the conversion from btu's to shadow area, and found a factor of 10 error (arg!) but still trust the more basic principle, that whatever error you make if it's the same one over and over the results are still comparable.   There remain some application issues, with my main update being to realize that the measure is most accurate and valid for the energy uses a dollar is responsible for that are hardest to trace.  That's very cool.    10/29/07

3-  I got a longer and better data set than the US DOE data, from DOE researcher who knew the EU IEA 2007 data sources.  That's displayed in (6.1).  It seems to confirm the basic 8000btu/$ metric, but has a different apparent growth rate for the energy intensity curve than the DOE 2004 data .  That may be due to counting different kinds of energy, I'm not sure.   I had been saying the DOE world GDP trend was +3.5% and the decay of energy intensity -1.8%, projecting the curves visually.    The IEA data shows world GDP growing at +3.02% and the decay of energy intensity at -1.23%, both much slower rates; doubling GDP every 24 years and halving EI every 60 years.   Until a more complete analysis and online tools are developed the original rules of thumb are nearly accurate and would only confuse things to change.    1/20/08

Notes & discussion -  top

1-Whether you call the $1 energy constant 8000 btu or 1,000 sf-hours (1995 constant $), it reflects the ability of energy to satisfy the values that people actually have.   Discounting efficiency gains as small compared to growth, energy for increased $'s will come from fossil fuels, nuclear, geo-thermal or the sun.   The principle flexibility in getting energy from the sun is competing land uses.  For every 8000btu of fossil fuel you want to offset with renewables you need to set aside an additional .1sf of high return land for energy production.  At first it seems easy, and then it gets hard, because converting land to energy production comes out of land already in use for something.    This draws very clearly the limits of the earth, and brings into question our curious worldwide devotion to multiplying money, as if it could come out of a magic hole in the ground.   Apparently we think that because much of  it actually did!   ;-)    Why people thought doubling the size of economies about every 20 years forever would be a reasonable plan for the future, requiring ever expanding resources or products and services using vanishingly small amounts of them, is a mystery.   The idea that our 200 year experience of magic energy from the ground in a pristine world would automatically be repeated over and over, well, looks like wishful thinking with support coming largely from our lack of evidence.  1/20/08

2- deleted

3- We've been waiting for 'the big crunch' when the growth economy hits the limits of the earth.  Clearly, prices and technologies have been responding to the future of energy and environment problems for many years, but also in several ways accelerating toward while ignoring the approach of the very firm limits of the earth and it's ecological capacities entirely.     I could be off, but think we're likely to look back and see that June 2007 marked the beginning of the real collision between our unconditional growth system and the finite earth.   It was that interesting little milk/corn price jump we had.    The bump in prices came directly from the technology 'surge'  to replace fossil fuels with methanol, and the 'unexpected' discovery that there was no free land to supply it.    In the complex system of feedbacks that kind of discovery sets off ripples of the economy's natural steering system.    I think it's likely to be the first of a serious of multiplying rather than damping swerves in response.   Exactly how is hard to judge, but the reactions of a growth economy to rigid energy growth limits is quite inevitably going to intensify because of the $/btu constant, fixed land mass and major interests in limiting our impacts on the earth.   Whether I picked the start of it correctly or not, we probably don't need to worry about whether this whole event will be noticed!    1/20/08

4- Doesn't there have to be some hidden 'give' in the picture.   We don't know the limits of lots of things, but no one is going to change the size of the earth or principles like the conservation of energy(20) or the principle of diminishing returns(23, 24).   Both are highly trusted principles of physics that we are still discovering the real meaning of.   What people have always valued and desired is not likely to change abruptly either.  We're just not going to buy products that have vanishing amounts of resource content.  We're also not going to change the properties of the systems we rely on, nor turn technologies that are running out of improvements to make into ones with limitless horizons again.  These are all 'discovered' facts not 'made up' ones.   The main one that prevents continuation of the kind of growth economies have relied on is the diminishing returns of investing in the earth, evident in our finding resources in smaller and harder to get at amounts.  It is also clearly evident in the stable and steadily slowing improvement in the whole system energy efficiency curves.   Energy efficiency improvements are now coming at less than half the rate expanding energy uses(17).    After the big reserves of easy to use resources are used what's left are the smaller reserves of ones that are harder to use, and it takes more investment.  That's diminishing returns, when it takes more physical investment to get the same physical return.

When miles of landscape are turned into housing sprawl, it does not appear on the news and we don't see the global sweep of eliminating the opportunities that were there before, but the opportunity the empty space once represented is gone.    After a sweeping conversion of land to buildings and highways you can't just rearrange it for higher density and grow corn again!    We don't see that, but instead hear stories like the announcement of a possible 40% efficiency solar cell, seeming to suggest solar power efficiency could multiply by a factor of 10 soon and create a whole new source of energy.   The curves of how the whole system is producing energy efficiencies show only a long term potential for 1.2% a year improvement, not jump by factors of 10.  At 1.2% a year a factor of 10 improvement would take almost 200 years.   That apparent contradiction is confusing.   

What the long range curves show is the collective behavior of the whole system, including all the millions of brilliant inventors doing their very best to improve the performance of everything that is sold by every means possible.   That's a primary way for business to remain competitive.   Individual anecdotes simply are not 'vetted' for the truth of their appearances that way, shaken down from many many perspectives at once, and tested for how they'll integrate with everything else.   The global trends showing whole system maturation are simply direct indications of it.  They're not theories, and they're not projections.   They're measures.    1/20/08

5- cut

6- It's important when working with measures that they be accurate.  What's even more important is that they be consistent.   Whether measures are correctly calibrated and so reproducible by a different means, if they're at least consistent means that your comparisons are still valid.   However measure the ratio of $'s of economic product to energy use,  the usefulness of the measure is providing an easily understood basis of comparison.  That's the reason for using the number 8000(btu) or 1000(sf) or 12(ozCO2) for a 1995$ of resources.  They're close enough to greatly improve the estimated scale of your impacts, and if you use them consistently you'll have comparisons from one use to the next that are completely valid.    1/20/08

7- The real reason for interpreting energy use as a physical shadow on the earth is not just the equivalence of fossil fuel energy and solar energy.   It has to do with the global warming choice.  If we increase total economic energy use or not, we need to move a lot of fossil fuel uses, up to 80%, to a new source of supply.   The choice is between 1) accepting the effect of long term pollutants produced by fossil and nuclear energy, 2) using such large amounts of energy from the sun that it has major impacts on other things already using it like a physical shadow, or 3) or changing the whole idea of what wealth is.   Anyone can easily understand drawing a 100 foot square on a hillside and needing to have exclusive use of it for 1 hour to earn each and every $1 they spend (or 0.1 sf/yr), and that makes it easy to calculate and imagine the physical scale of impacts of anything measurable in dollars.     1/20/08



pfh Explorations