Argument from incredulity

A few days back one of my favorite writers, David Roberts at Grist posted an article titled We can solve climate change, but it won’t be cheap or easy. This was based on an article in WIREs Climate Change, A critical review of global decarbonization scenarios: what do they tell us about feasibility? by a group of authors including Jesse Jenkins, a young energy analyst I've also been following for a while. Andy Revkin at dot Earth picked up on Roberts' post as a positive sign, while Joe Romm at Climate Progress had a rather pointed critique.

I don't want to get into the details of the report or most of these arguments here; there was just one point that I thought was very odd when highlighted in Roberts' post, and exemplified by the following figure from the report:

Roberts quotes the authors as follows on this:

These studies also envision a normalized build-out of generating capacity in the range of 5-23 GW/year/$T of GDP, or 1.4-15 times faster than historical experience. These unprecedented rates are a consequence of both the relatively low-capacity factors of wind and solar as well as increased demand due to the assumed widespread electrification of the economy.

What struck me was this measure of "build-out of generating capacity" in GW/year/$T of GDP. The graph suggests past experience on this measure should be a guide to the future, and therefore the Jacobson, WWF, etc. scenarios are unlikely to be feasible. But why did the authors pick this measure?

"Generating capacity" may be a metric favored by the utility industry but as the authors point out, for the same quantity of electric power delivered, renewable sources almost always require significantly more installed capacity, because the sun only shines and the wind only blows for part of the day. "Generating capacity" as a metric forces renewable sources to look like a much bigger deal, comparatively. If you want 1 kW of steady power, you're going to need to install maybe 5 kW of solar or wind peak generation; everybody understands that. There's really nothing particularly special about "peak" power levels other than that attached electrical equipment (inverters, batteries) needs to be sized for it. What matters most for the purposes of deployment is (A) average power (product of peak capacity and capacity factor) and (B) the cost per kW of that average supply.

Old utility power sources (nuclear, coal) typically run at 60-90+% capacity factors, so their average power supply is very close to their peak generating power. For all intents and purposes, then, the above figure would have been much more appropriately scaled according to average power supply rather than peak generating capacity - let's rewrite the above paragraph and imagine the red double-headed arrow in the above figure rescaled by about a factor of 4 (the ratio between typical existing generators and renewable capacity factors):

These studies also envision a normalized build-out of average generating power in the range of 1.25-5.75 GW/year/$T of GDP, or 0.35-3.75 times faster than historical experience. The high end of these rates is a consequence of an assumed widespread electrification of the economy; aside from that the range is well within historical precedent.

There, much better. It's amazing how presenting data different ways can change one's perspective. Unfortunately far too much of the argument about renewables that I've found seems to boil down over the years to this sort of "argument from incredulity". The numbers really aren't that bad when you think about it carefully. And the solar PV cost decline (the "learning curve") has been amazing and shows little sign of slowing down.