A while back I tried to figure out whether we can keep running things the way we're running them: do we need to transition off of fossil fuels, how fast, and would we be able to provide today's level of energy supply using alternatives? I didn't include many of the details in that post, so I'd like to consider a few additional questions and look at this topic in greater depth.
I finally got around to re-doing these calculations (so I could present them in a talk last week), and wanted to go through the numbers.
How fast do we need to transition?
In his excellent talk on climate change and energy, Saul Griffith analyzes how fast climate change might require a transition off of fossil fuels. While there is good reason to think that 350 ppm of CO2 should be our global target, we passed that point in the late 1980s and have shown no signs of turning back. Ultimately the goal is to avoid exceeding 2C of warming, as it marks the rough threshold at which many climate feedbacks are likely to kick in, including permafrost melt and loss of the Amazon. Griffith observes that 400 ppm should be a target since it gives us a reasonable shot of staying below 2C, but we're only 2-3 years away from 400 ppm (we'll be at about 393 ppm by the end of the year) so it seems that is unlikely to happen. Griffith does his calculations assuming a target of 450 ppm, which by various projections only gives us a 1/3 to 1/2 chance of staying below 2C of warming. (As a point of reference, the "agreement" at the Copenhagen climate conference would have taken us to 770 ppm, which equates to roughly 5C of warming.)
Given a target of 450 ppm, Griffith calculates we need to transition in about 20 years. I think about it as follows: we're increasing CO2 by about 2.5 ppm / year at the moment (a little less when the economy isn't growing quickly - this is a whole other subject I'd like to explore). Any real-world deployment of alternatives would necessarily face some ramp-up and would be producing much more towards the end of the transition, and thus we'd probably keep fossil fuel generation going until near the end of the transition period. As a result, there'd be little drop-off in emissions until near the end, and so we should aim to transition over a period of about 20 years to avoid overshooting 450 ppm. As for transitioning faster, we can look to the Hirsch report, which argued that a 20 year energy crash program is about as fast as it can be done. So 20 years it is.
I had done some quick calculations last time around, but I wanted to delve into more detail on industrial production numbers. Specifically, what does the alternative energy industry say? I figure that each industry is its own best advocate, so it's likely that their numbers will be on the optimistic end of the spectrum.
First, how much would we need to build to provide 15 TW globally using mostly alternatives in 20 years? Griffith calculates that we'd need to build 2 TW each of Solar PV, Solar Thermal, Wind, and Geothermal, 3 TW of Nuclear, and 0.5 TW of Algae fuel (we'd also keep some existing fossil fuel and other production capacity). While I could quibble with the particulars, it's a fairly balanced profile and is a reasonable starting point for a calculation.
Let's start with Solar PV, with this blurb that says we'll be at 28 GW / year of (nameplate) production by 2012. Let's round that up to 30 GW / year and use a capacity factor of 15% (considering the 200 W / m2 that's available in most temperate zones). That yields about 4.5 GW / year of production. Let's allow for a steady 20% yearly increase in production over the next two decades. Combined production over 20 years will thus produce about 745 GW of PV capacity.
I'm going to assume that Solar Thermal can easily meet its 2 TW production slice, since there's not much to it (mirrors, motors, pumps, generators).
How about Wind? The Global Wind Energy Council expects 2500 GW of wind nameplate capacity by the 2030s. Using a standard 30% capacity factor gives us 750 GW of Wind capacity in 20 years.
Geothermal is a bit complicated, as there really aren't that many easy places to tap geothermal energy. The International Geothermal Association expects a 9 GW nameplate capacity increase over 5 years. Assuming the same growth trend, this yields 94.5 GW of capacity. Including a 65% capacity factor, this yields about 61 GW of new geothermal capacity in 20 years.
Biofuels are also complicated. While there are numerous current-day biofuels, including ethanol from corn and sugarcane, Griffith rightly considers better options such as algae-based oil. Biofuels digest estimates a production capacity of 1.6 billion gallons / year by 2014. That's about half of one day's global production of crude oil, per year. Given the 15% production growth rate, this yields 26 billion gallons / year of capacity in 20 years, which is about 1.7 million barrels of oil per day (about 2% of today's oil production). Assuming the same energy density as gasoline, this yields about 105 GW of capacity in 20 years.
Finally, nuclear. There are no good sources for expected production capacity, so to be optimistic (side note: I don't actually think we should build any more nuclear, but that's another issue) let's use the peak rate of construction ever achieved, 30 GW / year of nameplate capacity construction (MacKay looks at a 60 year construction horizon, which is far too long). This would yield about 600 GW of capacity in 20 years, assuming no loss thanks to nuclear's high capacity factor.
In total this yields about 4.2 TW of new capacity to add to an existing 2 TW of fossil fuels, 1 TW of nuclear, and 0.5 of hydro, yielding 7.7 TW - about half of the target. That is, even assuming optimistic rates of production of alternative energy sources, we'd be about 50% short of our energy target in 20 years.
A few scenarios
Doing the analysis above reminded me that a wholesale transition to alternatives seems unlikely to deliver energy at current levels of consumption/production. I'd like to briefly consider a few possible trajectories / scenarios here, which I'll explore in more depth another time.