Scientific American‘s grand plan to provide a bit over a third of U.S. energy from solar sources provides insight into what it would cost to phase out all or most U.S. greenhouse emissions. Bottom line: a lot less than current military spending.

The total cost of the SciAm plan: $420 billion over the course of that 40 years, or slightly over ten billion dollars per year — less than current fossil fuel subsidies, less than the new subsidies “clean coal” would require.

The authors suggest phasing out fossil-fuel powered electricity over the course of forty years, using a solar dominated electricity grid. They suggest Compressed Air Electricity Storage (CAES) and thermal storage to compensate for the intermittent nature of solar electricity, and High Voltage Direct Current (HVDC) transmission lines to move solar electricity from where it is generated to where it is needed.

However, we can’t wait 40 years, and we especially can’t wait 40 years for a 35% reduction in emissions. So suppose we tripled the investment, and spent over the course of 20 years. That would be about $1.26 trillion, or $63 billion a year over twenty years — a rounding error in the Pentagon budget.

Unfortunately, it is not that simple. The “Grand Plan” saves a lot of money via slow implementation, giving the technology time to develop. Implementing it more quickly, with less mature technology, would cost more, probably requiring more solar thermal and less photovoltaic power (unless PV prices drop a lot faster than SciAm projects). So we can double to ~$2.5 trillion, or $126 billion per year. This is still a fraction of our military budget.

One more step raises costs further. In the SciAm projections, around two thirds of the electricity passes through CAES storage. CAES is a hybrid system, burning natural gas to use the compressed air more efficiently. Emissions from this natural gas are around 30% per kWh compared to our present grid. That implies total emissions of 20% per kWh compared to today, or an 80% reduction.

Global warming is just that — global. If we, the most intense greenhouse polluters, cut our emissions by only 80%, that does not leave room for development by the Global South, which produces one fifth or fewer GHG per person than we do. Reasonably, we need to reduce by 95% or better, to leave some possibility of fossil fuel use for others.

So, no more than 30% of our kWh should pass through CAES storage. SciAm suggests that the next least expensive storage method costs about double CAES. At most that increases total costs of the proposed system by a third. Extrapolating from SciAm’s own figures, that suggests a total cost of ~$3.4 trillion over the course of twenty years, or less than $170 billion a year to completely eliminate fossil fuels over the course of 20 years.

However, SciAm actually made a number of expensive technology choices. Making other choices could lower these costs further.

First, the grand plan contains almost no demand reduction, other than what would occur as a side effect of less fossil fuel use, plus electrification of transport. However, we know there is a huge potential for inexpensive efficiency improvements. At the low end, the McKinsey group calculates (PDF) that we could reduce by around 11% in absolute terms at a savings. My own book suggests that efficiency improvements alone could reduce emissions by more than 60%, at a cost lower than the current cost of coal electricity.

Secondly, we can use non-electrical solar to displace a lot of generation. Even in existing buildings, low temperature solar thermal could replace a lot of electricity used for space, hot water heating, and other low temperature purposes. (SciAm does already suggest a smart grid, including low temperature storage of climate control energy in buildings.) In new buildings passive solar is the next cheapest choice after efficiency.

Third, under-using wind costs the SciAm plan a lot. I understand why the authors concentrated on solar. Solar potential in the U.S. is hundreds, perhaps thousands of times current U.S. consumption. Commercial wind potential, in contrast, may be as little as ten times projected U.S. consumption in 2100. But wind electricity is also currently much cheaper than solar electricity. In addition, studies have shown that wind, when produced at multiple sites a great distance apart, has potential to provide a certain amount of very reliable power, even before storage is used.

Still another saving is that the SciAm plan included overbuilding to handle an extreme case where volcanic activity greatly reduce available solar energy for a year or two. A grid that was more evenly divided between sun and wind would still need overbuilding for both the volcano case and for wind draughts. But since the two are unlikely to occur simultaneously over as long a period as either one by itself, the overbuilding would not need to be as large. In addition, while wind by itself needs less storage than sun by itself, a grid that combines approximately even amounts of both will need even less storage. In fact, one very important part of the designing a combined solar/wind grid is to figure out the ratio of wind to sun generation capabilities to get maximum reliability. (For example, the Midwest and Great Plains are probably a lot steadier source of power than California.)

This assumes pretty complete electrification of transport and industry, and that backup for solar in climate control comes from renewable electricity, not fossil fuels. It also assumes we replace at least feedstock for chemicals with biomass, and perhaps tiny amounts of biofuel to run freight trains, short run freight trucks, and perhaps backup engines on plug-in hybrids as well. I don’t think we can do completely without liquid fuel, but we can probably reduce liquid fuel consumption to four or five quads.

If natural gas inputs into the electrical grid are already producing all the emissions we can afford, then we have to get liquid fuel from net zero emission sources. Alternatively, maybe this is an argument for spending another 70 billion or so annually on storage so that our electrical grid is truly fossil fuel free, and then using 3-4 quads of fossil fuels (mostly natural gas) for transport and industry.

So conservatively, the cost of eliminating 95% of fossil fuel use over the course of 20 years would be $170 billion annually if we can get small amounts of genuinely sustainable, net-zero-emissions biomass. Alternatively, for 240 billion annually we can do it with no biofuels. So we can completely phase out fossil fuels for around a third of current military spending.

(And zero biofuels is absurd. If nothing else we want to tap methane from waste, if for no other reason than to convert that methane into CO2 that will produce far less warming. Similarly, after we close all coal mines, we want to tap the non-biomass methane they emit into the atmosphere. Sustainable biomass potential may be small, but it is not zero.)

This assumes no breakthroughs in renewable production or storage — which is also absurd.

This only tackles fossil fuels, not agriculture and forestry. Though these are extremely important, I also suspect that converting them to sustainability would cost an order of magnitude less than eliminating fossil fuels. And in the course of increasing energy efficiency, we would probably make a start by reducing paper use, substituting agricultural waste for a large part of wood use, and supporting more energy efficiency in agriculture. I further suspect that half or more of the cost subsidizing more sustainability could be paid by converting existing subsidies into less perverse incentives.