Garvin Heath and his colleagues at the National Renewable Energy Laboratory last month published analyses of the indirect greenhouse gas emissions from electricity generation technologies. This work was also summarized in a set of articles that appeared in a special issue of the Journal of Industrial Ecology (which also includes some additional articles analyzing other energy technologies). The results are the product of what’s called “life cycle analysis”, where emissions from all parts of the lifecycle of an energy technology, including exploration, production, transportation, use, and decommissioning are carefully tallied. Such analysis allows consistent comparisons of emissions from the whole system, not just one phase in the life cycle.
The key findings are summarized in two figures, which I reproduce below. The first, which was published in the IPCC Special Report on Renewables, is more comprehensive, and it includes estimates as published, without any adjustments. The second figure summarizes results that have been harmonized to make the comparison as consistent as possible.
Figure 1: Comparison of as-published lifecycle greenhouse gas emission estimates for electricity generation technologies. The impacts of the land use change are excluded from this analysis
Figure 2: Comparison of as-published and harmonized lifecycle greenhouse gas emission estimates for selected electricity generation technologies
What I conclude from both figures is that all non-fossil generation technologies have modest indirect emissions on average compared to fossil fuels, with most of the non-combustion renewables (like wind, concentrating solar, ocean, and hydro) having the lowest uncertainty about their emissions. The big range for biomass comes from different assumptions about whether the fuel source is harvested sustainably or not, while the range for nuclear power depends on the fuel enrichment process used (some are more electricity intensive than others) and on the source of the electricity for that process. I think the range for photovoltaics is because of the many different technologies used to generate electricity from sunlight, but I’m not certain, so I’ll ask Garvin to enlighten us and report back.
The other important point about these indirect emissions is (as Saul Griffith points out) that even though they are small for most non-fossil resources, we are under tight time constraints (and a tight carbon budget) for keeping the earth from warming more than 2 Celsius degrees from preindustrial times. Whatever infrastructure we build to meet that challenge will eat up some of the carbon budget, even if we use the source with the lowest indirect emissions, so we really only get one shot at creating a stable climate. That is an example of path dependence in its purest form, and keeping in mind the importance of life-cycle emissions can help us think more clearly about how to create a more hopeful (and cooler) future.
The journal has made the articles freely downloadable, as a public service.
The Journal of Industrial Ecology is an international peer-reviewed bimonthly journal owned by Yale University and published by Wiley-Blackwell. It is the official journal of the International Society for Industrial Ecology.
Addendum, June 26, 2012: Garvin replied to my email with this explanation about the ranges for solar photovoltaics and biomass:
As for the range from PV, it is due to both technology and manufacturing process variation but also solar resource (some estimates considered low solar resource sites) and to some extent the GHG intensity of the source energy mix.
As for biomass, one key variation is whether the feedstock is residue/waste (implying no burdens from feedstock production) or a crop one grows intentionally (which carries the burdens of the production). But the bioenergy system is very complex, so there are a host of other differences – climate, feedstock selection, irrigation requirements, transportation requirements, fertilizer usage, tillage regime, processing requirements (especially drying), etc. NREL has developed a version of our popular System Advisor Model for biopower which allows the user to specify their “system” and returns an estimate of the life cycle GHG emissions: https://sam.nrel.gov/ to download and find documentation.