The "Green" Energy Transition

Below are a number of "back of the envelope" calculations I made in 2021 which indicate how daunting is the task of eliminating CO2 emitting energy sources. The numbers are grim. Filling the atmosphere with CO2 isn't going well, although it does make for a comfortable life for now. The "Green" alternatives might provide half the per capita net energy consumption, but that's for a relatively sparsely populated country such as the USA or Russia. For the UK, Japan, India, and many other places, the prospects are dismal.

I gathered my numbers mostly from public US government web sites including the Department of Energy (DoE), the Orwellian titled Energy Information Administration (EIA), and the National Renewable Energy Laboratory (NREL).


To power the US at its current net energy consumption using "renewable" energy would require devoting over 200,000 square miles to a mix of wind and solar farms. That's 1/15 of the continental US, or an area the size of Kansas, Nebraska, and South Dakota combined. Realistically, we'll have to get by on much less than today, which wouldn't be too painful.

What's to Be Replaced

The following graphic gives the energy flows for the USA in 2019. (One can read the numbers by zooming the image.)

The energy units are quads, which is one quadrillion BTUs. Since most of the alternative energy sources, such as wind and solar, produce electricity directly, it will be useful to have the conversion to electric energy, and 1 quad is 0.3 PW-h or 300 TW-h or 300,000 GW-h.

In their graphic, Livermore broke the energy flows down to four categories of consumption: Residential, Commercial, Industrial, and Transportation. The numbers in the pink boxes show the energy which passes into (and out of) these sectors. These each put out two forms of energy, Rejected Energy (which is waste) and Energy Services, which is the net useful energy. Note that the net energy consumption of 33 quads is one third of the primary energy sources (shown the left side of the graph).

Primary energy consumption was 100 quads, a staggering number. Two-thirds of this becomes waste heat, but Green sources don't create as much waste heat, so you can get by with less from them. I estimate the US can match it's current net energy consumption with "only" 50 quads of electricity. Even this reduced requirement will be very difficult to match with "Green" sources, so the only way to make ends meet is to greatly cut net consumption, probably by two-thirds.

How much electricity is needed

First, estimate what electrical energy will be needed to deliver the Energy Services (net consumption) for each of the four categories.


Livermore shows about 12 quads (11.9) was used for residential. Of that, 5 was for electricity (appliances, lighting, internet, etc), and the remaining 7 quads would mostly have been for heating (air and water). For heating, a heat pump can be used to deliver about 10 units of heat energy for each unit of electric energy, so we'd need maybe 1 quad of electricity for heating, giving about 6 quads total for the current housing stock.


This is offices and stores and the like and has a similar energy profile as residential, so it ends up about 5 quads of electricity.


Of the 26.4 quads used by industry, only 3.3 was electric and almost 21 was fossil fuels. The remaining 2.5 was biomass which is more or less "green" already.

The fossil fuel consumption will be mostly process and space heating, and I didn't try to come up with solid numbers as to what it takes to replace each form of it with electricity. Livermore places useful energy as nearly 13 quads for industry. About 3 quads were input as electricity, most (but not all) of which would be useful. That implies about 10 quads of useful heat.

To make the 10 quads thermal needed, space heating could be heat pumps (10 to 1 heat production), but some processes will require higher temperatures and use resistance heating (1 to 1). Heavy heating processes such as steel and cement will require combustion, either hydrogen or hydrocarbons, and both types can be made with electricity (a process possibly enhanced with waste heat from nuclear plants). Combustion fuel production is currently not very efficient, and I'll use 60% as an average. (That 60% is likely optimistic, with hydrogen better but hydrocarbon worse.) For now, I resorted to a scientific wild ass guess (SWAG) of the amounts of each form of heating required and came up with an overall conversion of 75% from electricity to heat for industry. This low number reflects losses in the production and distribution of combustion fuel overwhelming the gains from heat pumps.

I thus came up with 13 quads (10 / 0.75) plus the 3 quads already used by industry as electricity, giving about 16 quads total electric for industry.


One quarter of primary energy consumption currently goes towards transportation. This will be cars, trucks, trains, and airplanes, and the latter three for the most part won't work well off batteries, so gaseous or liquid fuels will be required, creating energy losses in their production.

The graphic from Livermore has a little over 26 quads of fossil fuels being used in transportation. The EIA has numbers for the relative amounts of transportation fuels used:

	gasoline     56 %
	distillate   24     diesel, etc
	jet           9
	natural gas   4
	biofuel       5	    alcohol and oils 

Cars and short haul trains and trucks can (in principle) be done with batteries or overhead wires, so I assume all the gasoline and natural gas used in transportation can be replaced with battery power, or about 60% of the total. Livermore gives 6 quads of useful work in transportation, and 60% of that is about 3.6 quads.

One 2020 SAE paper gave 70% as a favorable estimate for the overall efficiency converting electricity at a generating plant to motion in a battery electric vehicle. This would include line, charging, discharging, and motor losses plus cabin heating and cooling draws as well as cold weather effects on the batteries, etc. Applying the 70% to 3.6 quads, one gets about 5 quads electric needed.

The heavy haulers—on highway trucks, locomotives through rural areas, and medium and long distance airplanes—will have to stay with gaseous or liquid fuels, which I've for now put at a 60% conversion efficiency. This will represent the distillate and jet fuel categories from the EIA which is about 33% of the fossil fuel consumed, so 26 * 0.33 or about 8 quads, which at a 60% efficiency of production is around 14 quads electric energy.

For transportation, the total comes to about 19 quads.


Adding up the 4 categories, it's 46 quads of electric delivered, but there will be some overhead.

The EIA estimates 5% of electricity is lost through transmission (high voltage) and distribution (low voltage). I expect this to rise as the length of transmission lines increases to carry green electricity from the rural areas where it's created to major cities, especially those along the coasts.

There's the issue of storage. Solar and wind are variable and seasonal. Having many long transmission lines might balance the load around the country (while creating more line losses), but battery back-up is being created to level out the daily peaks. Batteries are about 85% efficient, so there will be daily losses from that.

The US imports more in manufactured goods than it exports. Other industrial countries are too densely populated to have enough collecting area for Green energy to support their industry (more below), so the tally for Industrial use of green energy would have to increase to keep the level of goods here the same.

Overall, I just rounded up the 46 quads to an even 50 quads needed from electric generating plants but realize it should probably be bumped up further. However, just modest improvement in energy efficiency should bring it back to 50 quads.

Scaling up Green energy sources

Whence these 50 quads? There are only four proven, major sources of Green energy: Wind, Photovoltaic (PV), Hydro, and Nuclear. These deliver primarily electricity, although nuclear creates much thermal waste which might prove useful for chemical engineering. Switching to the more common energy units for electric energy, those 50 quads are 15 PW-h (15 PetaWatt-hours).

For comparison, the amount from each "green" source as actually produced in 2019 is found in the EIA Monthly Review:

	nuclear		 809
	hydro		 288
	biomass		  58
	geothermal	  15
	PV		  72
	wind		 296
	  total		1538  TW-h 

One finds the electric generation and distribution in the U.S. was only 1.5 PWh or 1/10 of what will be needed. (From all electric sources, the grid carried 4.1 PWh or about one quarter of would be needed. As an update, the Review shows less than 1.7 PWh from "green" sources in 2022.)

To get an idea of the magnitude of the problem, I next considered each of these four sources as if each alone were to provide the 15 PWh. In practice it would be a mix, of course.


In 2019, wind produced only 0.3 PWh. We'd need to increase production by 50 times. What would that take in terms of land?

A 2023 article in the Wall Street Journal mentioned the average capacity factory for a wind turbine on land ("on shore") in the US is running only 30% of the nameplate rating. In 2019, the average wind turbine was rated at 2.5 MW, but today 15 MW are on the drawing boards and 10 MW are in service (but currently have a high failure rate). I'll go with 5 MW since those are in service and it's currently not possible to move the blades of the largest turbines over land by either truck or rail.

So, 15 PW-h from 5 MW turbines running at an average output of 30% of capacity implies about 1,100,000 turbines across the country. That's one big turbine every three square miles across the entire continental US.

Unfortunately, much of the country isn't the best for wind power, and the existing wind turbines will have been placed in the favorable locations, so that 30% capacity factor may not be easy to maintain as wind deployment increases. On the other hand, the larger turbines need higher masts and this places them in a slightly higher average wind.

One problem will be the spacing of the turbines. For efficiency, it's recommended they be spaced at least 7 rotor diameters apart and preferrably twice that. (I don't think I've yet seen a turbine field spaced to even the lower recommendation.) The 5 MW turbines have rotors of about 130 m. For the preferred spacing (which would be about 1 mile), the area to be covered with turbines is fantastic, over a million square miles or 1/3 of the continental US. At the lower recommended spacing, it's only 350,000 square miles of wind farms, or just the states of ND, SD, NE, KS, plus half of OK converted to wind farms with 3 big turbines every square mile.

Some government studies are useful for comparison:

My estimate of the land use required is reasonable—so long as you consider reasonable turning five entire states into wind farms. I can't see the people living in those states putting up with that. Indeed, there are already protests with only a tiny fraction of that number of turbines in place. Further, the best wind areas are right down the Central Flyway for migratory birds; the carnage would be remarkable.

Servicing such a fleet would be labor intensive. The blades on wind turbines have a life of about 25 years, so about 200 turbines will be off-line daily to have the blades replaced, ground up, and disposed. However, this should be no more effort than trains hauling coal around the country daily.

The need for storage is high even for wind. EIA numbers show total wind power across the U.S. varies 4 to 1 over a span of weeks, and this is on top of a seasonal cycle. That seasonal and weekly variation in wind across the nation is not all bad news. One would have to size the turbine fleet for the low end, which implies there can at times be upwards of triple the electricity left over. This can be used to make hydrogen and hydrocarbon fuels for industry and transportation, which is why in my calculations above I've assumed those are available.

While finding space for a million turbines would not be easy, making them is feasible. Turbines cost about a million dollars per MW plus land and connection fees, so the 5 MW jobs would cost about 7 million installed. Assuming a century life span (the blades will have to be replaced about 4 times in that span), that's 10,000 per year at 7 million apiece or 70 billion per year, a small fraction of the annual automobile sales in the US each year. This of course neglects the nettlesome problem of storage for such an intermittent source of energy.

Off Shore Wind

Off shore, wind turbines might attain a higher capacity factor, and there are proposals for 15 MW turbines floating beyond the continental shelf, where there are no people living to complain about their sea vistas being ruined. Unfortunately, it hasn't been proven such giants (they will reach about 80 stories tall) can withstand a full force Atlantic hurricane.

As I read it, only a few floating turbines have been made, and the existing fleet of fixed bottom, 10 MW off shore turbines is suffering a high failure rate. A 2021 NREL paper estimates the floating turbines will cost about 4 times as much as those on land (the fixed bottom jobs are more than twice the cost of those on land). We'd still need about 350,000 of them. The effects on avian and oceanic wildlife are unknown, but concerns have already been raised. I can imagine a whale or shark cruising along in the dark being wounded running into the cables that anchor such structures, or sea birds being slapped down by turbine blades.


PV isn't quite as dismal as wind, at least with regards to land requirements. An NREL paper from 2013 reported the average total land-use requirements of a solar farm are 3.6 acres/GWh/yr. For 15 PWh, that's 86,000 square miles, or roughly the area of all of Kansas.

Solar cells wear out, so this would have to be renewed every generation or accept an even larger area blanketed with cells. It would also have to be supported with literally a mountain range volume of batteries, but those might be placed directly under the cells. That is, those mountains of batteries would just be spread out over an area the size of Kansas.

Perhaps such an area is not impossible to maintain. One report claims 18,000 square miles of roadway in the US, so the area of PV required is about 5 times what we maintain in roads. Covering roofs with PV cells won't cut it. One 2018 report reckons only 1.4 PWh of rooftop potential in the US, or about 1/10 of what would be needed. Over time, buildings would be replaced with those with better roof orientations and so would improve upon that. On the other hand, as people move to more densely populated cities to save energy, rooftop potential would fall.


This source, while useful, is a dead end, for most of the potential sources in the U.S. are already tapped out, so to speak. The EIA reports 0.3 TWh, or only about 2% of what's desired.


Nuclear is great in that it's reliable and no energy storage is required. Plants prefer to run flat out, so in the low power times (at night) the excess could be used to make combustion fuels. The EIA says only 0.8 TWh were produced in 2019, so to meet all the demand, the generating capacity would have to increase about 18 times the current level. While not unfeasible from the standpoint of making the facilities, there simply isn't enough fuel to go around with the current fuel cycle. Beginning in the 1970s, the US foolishly chose to ignore the development of the breeder and recycling techniques needed to make nuclear viable any time soon.

Fusion remains utterly unproven.

What it implies

We've painted ourselves into a corner. Getting out will not be easy.

The idea of turning several states into wind farms strikes me as infeasible. PV is denser, but relies heavily on batteries, which also doesn't seem very feasible. At least with wind, you get some output all day, all year. A mix of the two would let PV fill in some of the gaps in wind. Nuclear fission can provide some of the baseline load as well as useful process heat.

Overall, I reckon the US would want to use no more than one half the net energy consumption per person and probably no more than one third. That doesn't mean we'd be in poverty, just that we'd have a far less luxurious life. In the above calculations, the big consumers are industry and heavy transportation. Having a house 1/2 the size, and 1/3 the clothing, and not using food to put blubber on ourselves, and driving half the number of cars (tiny little cars at that) would reduce the consumption from transportation and industry greatly.

Ultimately, an economy needs to provide shelter, food, medical care, transportation, and some entertainment. Cutting to 1/2 the shelter would ripple through the economy reducing energy consumption proportionaly, and there are further gains to be had in energy efficiency of the US housing stock. Entertainment by internet is far less energy intensive than trips to Disneyland or vacations across the seas. Energy intensive foods like meat would have to be reduced, but we eat too much of that to be healthy anyway. Production of medical supplies and equipment you wouldn't want to change much, though. Overall, it would be like turning the clock back to about 1920, but with far more effective medical care and with the diversions of the internet. My grandparents never complained about life being unbearable circa 1920. It's obviously viable, just not as pleasant.

To further put things in perspective, compare the US today to other countries. One chart in the Wikipedia says the US uses about 79 MWh of energy per person per year. Japan is 40, France is 36, Germany is 41, and the UK is 30. These are all industrial economies getting along on roughly half the consumption the US does. In part, this is the result of living in small dwellings in densely packed cities, thereby reducing heating and transportation costs. Another part is refusing to carry their share of the military defense burden.

"Green" energy is, roughly, produced by the area available to collect energy from wind and solar, and so population density determines the per capita energy availability. The population density of the EU is triple that of the US, and so people there will have to get by on 1/3 of the Green energy. Let's say the US can manage to produce 40% of what it uses today, so 32 MWh per person. The EU would then have to get by on about 11 MWh, so they are looking at about a 2/3 reduction in standard of living. With a population density of 4 times that of the US, China is in a similar predicament.

The industrial powerhouse Japan has nearly 10 times the population density; the UK is 8 times, Germany (if treated separately from the EU) is nearly 7 times, and India is a miserable 12. With "Green" energy, these countries will be in poverty. Actually, I imagine they have little choice but to keep using fossil fuels until their population density drops.

You must be wrong

Possibly I am. These are only very rough estimates, but I did look for other estimates.

The Electric Power Research Institute in it's 2021 paper "Powering Decarbonization: Strategies for Net-Zero CO2 Emissions" has a figure (the text adds no details) that shows current energy use is about 87 quads and that the electrical energy replacement will be about 42 quads. This is compared to my 50 quads which was based on Livermore's report of 100 quads used currently. I've seen the 100 quads estimate in other papers, and I don't know from where EPRI took their baseline.

The 2015 study by Jacobson and 3 others, all from Stanford and UC-Berkeley, titled "Low-cost solution to the grid reliability problem with 100% penetration of intermittent wind, water, and solar for all purposes" did a far more detailed study of the problem than I have. Their analysis included looking at detailed wind and solar influx maps for the entire country. Their solution assumed some energy and capture technologies I didn't consider, such as geothermal and solar-thermal. They rely heavily on hydrogen production whereas I assume hydrocarbon production for some uses. They also rely on thermal storage to try to alleviate the problem of batteries. Their results of a computer simulation of 6 years of energy flows across the contiguous US was an average annual production of 14.5 PWh of electricty, nearly all from wind and PV (plus another 1 PWh thermal from solar), as compared to my rough estimate of 15 PWh.

So there.

Closing Comments

The ability of wind turbines across the continent to deliver some electricity all the time seemed a huge benefit over PV. This would reduce the problem of storage while allowing the land below to remain productive for agriculture. That it takes more than five times the area deployed I had not expected. The Jacobson study found that half wind and half PV was a good mix for the US, and that ratio was what I used to come up with the 200,000 square miles estimate in the Summary section.

11/2023 - 02/2024