The challenge of life-cycle analysis in a world of rapid innovation
There was a big stink this week when a published study, led by University of Virginia civil engineering professor Andres Clarens, concluded that producing biofuels from algae isn’t as climate-friendly as many people believe, at least when compared to getting biofuels from switchgrass, canola, and — Huh? — even corn. The results, according to an abstract of the study, “indicate that these conventional crops have a lower environmental impact than algae in energy use, greenhouse gas emissions, and water regardless of cultivation location.” Why? Because of the need to supply more nutrients — i.e. fertilizer — to algae to stimulate growth, and fertilizer is energy-intensive to produce.
The problem with this conclusion? Clarens based the life-cycle analysis on data that was mostly 10 years old. For example, some current algae cultivation practices, particularly those based on wastewater or sea water, tackle the fertilizer issue head on. So the age of the data is an important bit of information that should have been made very clear in the study — even the abstract. Ten years in the world of technology, particular cleantech, is a long time. I mean, the big R&D push around algae-based fuels only began three or four years ago, and 10 years ago the “cleantech” sector didn’t exist in name. Ten years ago the world was still wrapping its head around Y2K, George W. Bush was just getting into office, Google was still a start-up years from going public, and the TV show CSI (the original one) had its world premiere. In other words, you can expect data about algae cultivation to be, well, rather useless as a reflection of current practices.
This isn’t to blame Clarens. As he told the New York Times’ Green Inc., the most current data out there is simply unavailable to academia. It’s proprietary. “I’d be happy to model it if somebody produces it,” he said. This, of course, is a general problem with a lot of studies looking into lifecycle analyses. Researchers can only go with the data they can get, and perhaps this explains a lot of the earlier controversy around ethanol from corn. It’s still something we want to move away from, but certainly not as bad as guys like David Pimental of Cornell University like to paint it. I’d argue these studies should do two things: make a greater effort of emphasizing data limitations; and make a clear distinction between technologies/processes already deployed and those in pre-commercial phase.
Look at it this way: Can you imagine a study coming out in 2010 comparing different Internet search engines, but basing it on data available in 2000? Now, the Internet isn’t cleantech, but in certain areas there’s no reason to believe that the pace of innovation is any different.
Life-cycle analysis is hugely important work, but if it can’t keep up with innovation then it can become dated before it’s even published.