Answers For DeLong About The SuperFreaks, Part Three: Solar Power And Warming Debts

This is part three of a three-part series. Read parts one and two here.

Blogging economist J. Bradford DeLong has read the “global cooling” chapter of SuperFreakonomics and has asked six wonkish questions about climate science and policy. DeLong’s final two questions were about the lifecycle costs of deploying solar power. In SuperFreakonomics, Dubner and Levitt cite billionaire mosquito-laser inventor Nathan Myhrvold’s argument that solar power is not actually a “good thing” when it comes to tackling global warming.

Solar panels, Myhrvold argues, create both an “albedo debt” and a “warming debt.” If a “black” solar panel is placed on a light-colored surface, even as it generates electricity it will increase air temperatures. Furthermore, the construction of large-scale solar plants generates global warming pollution, which Myhrvold claims would counteract the benefit of replacing coal-burning plants. He makes the radical claim that the “warming debt” from solar plant construction would make “emissions and global warming worse every year until we’re done building out the solar plants, which could take 30 to 50 years.” DeLong, not surprisingly, finds these claims a bit dubious:

5: “The problem with solar cells is that they are black… designed to absorb light from the sun…. But only about 12 percent gets turned into electricity, and the rest… contributes to global warming.” Surely the heat energy reradiated from a solar panel is a small fraction of the heat trapped by all the carbon dioxide that would be produced by the coal-fired plants that would otherwise generate the electricity, isn’t it?

6: “The energy consumed by building the thousands of new solar plants necessary to replace coal-burning and other power plants would create a huge long-term ‘warming debt’.” I had thought that practically none of the power plants that we will use in 2050 are now in operation, and that building them — whether for open-carbon cycle, closed-carbon cycle, or non-carbon — will cost about the same amount of energy, and thus that there is no significant extra power-plant construction debt from going green in our new power-plant construction over the next forty years as long as it is done gradually. Am I wrong?

Myhrvold has defended his arguments, saying that when he said “black,” he didn’t mean black, just, well, rather dark. Although Dubner and Levitt radically misrepresented Ken Caldeira’s opinions in their chapter, they were spot on with Myhrvold, who blogged:

If we go hell-bent for leather in building solar plants for the next 50 years or so, it is entirely possible that we won’t see much small benefit for 30 to 50 years.

This is nonsense. Take a simplified model of the world that starts with 100 percent high-emission coal plants emitting 10,000 MMT of carbon dioxide a year and no zero-emission solar plants. Let’s assume that the construction of each solar plant has a three-year “warming debt” and that the use of each plant has a two-year “albedo debt,” in line with Myhrvold’s estimates. We’ll also assume slow growth in total energy demand (an assumption which does not affect the results of this thought experiment). If all the coal plants are replaced over a forty-year period (by 2050), the world starts seeing the benefit in only twenty years (by 2030):

Chart on left graphs annual emissions against number of solar plants, with total replacement by 2050. Chart on right shows cumulative emissions difference over time, showing cut of two-thirds by 2100. Click charts to enlarge.


What if instead of taking forty years to replace every coal plant, the world follows Al Gore’s advice and goes completely to clean electricity in only ten years? Surely, if it takes five years for a solar plant to pay off its warming debt, this crash effort must have serious negative consequences, right?

In fact, although such an effort creates a brief spike in cumulative emissions, benefits rapidly accumulate:

Chart on left graphs annual emissions against number of solar plants, with total replacement by 2020. Chart on right shows cumulative emissions difference over time, showing cut of 84% by 2100. Click charts to enlarge.

Cumulative emissions accelerate briefly, increasing by one-third over the reference scenario, but then plummet with respect to the reference, such that gains are seen by 2023, years faster than the slower transition. By 2100 total emissions are 84% lower than they would have been, a major improvement over a slower transition.


This simple modeling ignores other real-world factors that make Myhrvold’s assertions even less plausible. Legislation such as the Waxman-Markey American Clean Energy and Security Act would promote both an increase in clean energy and in efficiency — delivering emissions reductions both through cleaner supply and reduced demand. In short, the SuperFreak boogeyman of a great “warming debt” caused by a rapid transition to clean energy is highly misplaced.


Ken Caldeira, the climate scientist whose research is misrepresented in SuperFreakonomics, continues to pummel the book’s approach to climate science:

To talk about global cooling at the end of the hottest decade the planet has experienced in many thousands of years is ridiculous.