US Mitigation Math

cross posted from Prometheus, the Science Policy Blog

The mathematics of United States carbon dioxide emissions are not actually that complicated. The figure below from the U.S. Energy Information Agency shows that the 5,991 million metric tonnes (MMt) of carbon dioxide emitted by the U.S. came from 3 sources: coal, natural gas, and petroleum (see three inputs in the upper left of the graph).

Each of these fossil fuels, plus renewables and nuclear power make up the total energy consumption in the United States. Energy consumption is measured using a unit call a "quad" which means a quadrillion BTUs (British Thermal Units). In 2007 the United States used 101.4 quads of energy (data). This amount of energy can be broken down by source as follows.

The 15.2 quads of energy from nuclear and renewable sources resulted in negligible carbon dioxide emissions. The amount of carbon dioxide emitted due to each quad of fossil fuel energy depends upon the source, as their carbon intensities differ. For the analysis that follows I use the following values, distilled from the EIA information provided here in .xls.

Coal = 94 MMt Carbon Dioxide per Quad
Natural Gas = 53 MMt Carbon Dioxide per Quad
Petroleum = 65 MMt Carbon Dioxide per Quad

Thus, to calculate total U.S. carbon dioxide emissions simply requires multiplying quads of energy by carbon dioxide per quad and summing across the three fuels. This simple math results in the following:

(94 * 22.8 [Coal]) + (53 * 23.6 [Natural Gas]) + (65 * 39.8 [Petroleum]) = 5,981 MMt carbon dioxide

This total compares quite well with the total of 5,991 MMt carbon dioxide reported for 2007 by EIA (see figure above). We can use this information to ask some straightforward questions about how an emissions reduction target of 14% below 2005 levels (5,095 MMt carbon dioxide) might be reached by 2020.

We can do a bit of hypothetical "stress testing" of these numbers, by asking, in theory, what sort of actions might lead to reaching the emissions reductions target. Before we do this, we do need to make a guess as to 2020 US energy consumption. The EIA projects that energy consumption will grow at a rate of 0.5% per year (calculated from information here). Because GDP growth is expected to be higher than this rate, it already builds in an assumption of gains in energy efficiency. But let's use the EIA estimate, which suggests that US energy consumption in 2020 will be 108.6 quads, of which 21 quads will come from renewables plus nuclear energy, representing a growth of about 40% on top of 2007 values. This leaves 87.2 quads to be produced by fossil fuels.

Here are a few examples of the effects of different hypothetical strategies:

1) What would happen if all coal consumption were to be replaced with natural gas?

Answer: In 2020 total emissions would be 5,110 MMt carbon dioxide, very close to the 2020 target.

2) By how much would renewables plus nuclear have to displace coal to reach the target?

Answer: The target could be reached if coal consumption were reduced by about 42%, and the displaced 9.2 quads of energy were replaced by renewables plus nuclear, implying more than doubling of renewable plus nuclear energy supply, to comprise 30% of all energy consumption.

If renewables alone (i.e., non-nuclear) are to carry the weight of displacing coal, then they would have to increase their role in consumption by a factor of 4.7 over 2007 values. If growth in renewable energy supply is restricted to solar and wind only, then these sources would have to increase their role in consumption by a factor of 80 (that is, e-i-g-h-t-y). The reason for this big difference is that biomass and geothermal provided about 6.4 quads of energy in 2007, whereas wind and solar only 0.4 quads. The Obama Administration's goal of doubling wind, solar, and biofuels production within 3 years may indeed be a worthwhile policy, but it is not consistent with a goal of displacing sufficient coal to reach the 14% 2020 target using wind and solar (and while biofuels have their own complexities as a policy issue, they are not really a substitute for coal in any case).

3) By how much would energy consumption have to be reduced to meet the target assuming no changes in the energy consumption mix?

Answer: Energy consumption would have to be about 85.5 quads in 2020, about equal to 1992 values when the US economy was 35% smaller than in 2007.

Some Comments on the Stress Tests

First, number (1) above is really not desirable if the goal of mitigation policy is ultimately a reduction in emission of 80% or more. The reason for this is that while natural gas is less carbon intensive than goal, it is still carbon intensive. Locking in a large natural gas infrastructure is not compatible with large emissions reductions. Consider that in the hypothetical case that all US fossil fuel needs were to be met by natural gas, then 2007 carbon dioxide emission would have been 5,375 MMt, less than observed in 2007, but not consistent with any low stabilization target.

Second, number (2) is theoretically promising but practically daunting. The following is worth repeating -- for wind and solar to displace enough coal to reach the 14% target by 2020 would require that it increase by a factor of 80 in absolute terms from 2007 production. President Obama's policy of a tripling in wind and solar energy supply in the next three years would leave a need for another increase by a factor of about 25 over the next 8 years if wind and solar are to displace sufficent coal to meet the target.

Third, with respect to number (3), while there is a lot of potential to exploit in increasing energy efficiency, to reach the 14% would require a reduction of US energy use by about 2 quads per year for the next decade. Assuming that policy makers and citizens want economic growth to resume, this is a Herculean task. If you factor in that the EIA estimates to 2020 already include a good bit of efficiency gain in the BAU scenario, the task could be even larger if these assumed gains do not occur or if economic growth happens at a faster rate than assumed.

In reality, of course, none of these "stress tests" would be applied alone; there would be a combination of all three approaches discussed above. However, I challenge readers to present a scenario combining decarbonization of the energy supply and efficiency gain that has a realistic chance of succeeding in meeting a 14% emissions reduction (below 2005) by 2020. I am not saying that it can't be done, but I am saying that I don't see how it can be done. The comments are open, have at it.

Setting an emissions target and timetable, allocating emissions permits, and then saying that the magic of the market will efficiently take care of the task is exactly the answer I'd expect if one doesn't have an answer. Markets can't make the impossible possible, and when they are used in such a manner, often have undesirable results.

Comments

---

This post is confusing on the topic of primary energy use vs. usable energy (electrons on the grid).

The AEO forecast quads are all for primary energy -- this is the raw energy in the fuel itself. (1 kg of solid coal has a certain amount of "primary energy".) Given a OECD average coal power plant efficiency of 37%, the 22.75 quads of coal burned in 2007 generated only 8.4 quads of electricity. (This is before transmission line losses, etc, but those are the same for all technologies, except efficiency.) Natural gas plants are a bit more efficient, at 45%. The transportation sector uses primary energy directly (liquid fuel into your car), so that part is "100% efficient" compared with electricity. (The inefficiency is counted in a different place.) Since not all petroleum is used in cars, let's assume it's 95% efficient when counting "usable energy". AEO Renewables are calculated directly from the energy in the lines, not from "primary energy" of sunlight falling on a given square meter, so they are effectively 100% efficient for this comparison. Nuclear is counted the same way (I think). The upshot of this is that instead of working with 102 quads of primary energy, we ought to be thinking about 72.6 quads of "used" energy in 2007. The most recent AEO numbers (downloaded from http://www.eia.doe.gov/oiaf/forecasting.html and then adjusted in Excel)) show this "used" number growing to 75.2 quads in 2020.

I'm a big efficiency fan, but let's assume that number is un-changeable. Can we shuffle the proportion around between sources to add up to 75.2, while still cutting greenhouse gas emissions by 14% before 2007? Of course. Start by cutting coal in half. If the average coal plant has a lifetime of 30 years, this should be happening anyway over the next 15 or so years, as long as we don't build new plants. 4.5 quads of used coal electricity comes from 12.2 quads of primary coal energy. Expand natural gas a bit: 13.1 quads of used natural gas power comes from 29.1 quads of primary energy (AEO forecasts 24 quads of NG). Leave petroleum where the AEO says, at 39 quads of primary energy. The AEO forecasts renewables growing from 6.3 to 9.4 quads. If this is accelerated to 11.5 quads, and nuclear stays where it is, we have 75.2 quads of "usable" energy on the grids/roads, using only 100.7 quads of primary energy, and emitting 86.3% of the carbon emitted in 2007.

Electric efficiency enters the computation the same way renewables do (and in fact better because it avoids the few% transmission line losses). So, if you don't think we can grow renewables at a big boost over the AEO rate, just think about how to save 2.1 quads of end-user energy (per year) without hurting quality of life. Given that this is just 3% or so of total use, it's not hard to imagine. (Refrigerators, for example, can easily be 20% more efficient today -- look at all the Energy Star models available --and that's not built into the AEO model.)

What happens if we set the "used" energy in 2020 equal to the "used" energy in 2007? If we just need a "useable" energy of 72.6 quads, we can even leave renewables where the AEO predicts them to be in 2020, and just do a straight swap of coal for natural gas and efficiency. Accelerated renewable R&D beyond the AEO forecast can push the natural gas number down, too.

It's fun to play with numbers, but this post creates a false problem by ignoring the fact that the AEO numbers quoted are in _primary energy_ terms. I haven't addressed any policy question here of how one would do this, but any option which results in enough increased renewables, efficiency, and natural gas at the expense of coal (relative to the AEO baseline) ought to work.

(comment also posted at Prometheus)

By Asa Hopkins on 2009 03 16

---

Peak Oil and Peak Natural Gas make for greater dependence on renewables, nuclear and coal. As it becomes more expensive to mine, refine and transport fossil fuels we have to make a serious committment to embrace sustainable sources of energy for future generations.

By M Vandewark on 2009 03 12

---

What about using natural gas and nuclear to replace coal? Yes natural gas is more carbon intensive, but 200 GW of gas turbines were built in about 10 years. It is probably the fastest carbon reducer. Nuclear plants in Asia can be built in about 5 years a piece. I know everyone prefers renewables, but gas and nuclear have the scale (and do not need energy storage) required for such a large replacement program.

By R Margolis on 2009 03 12