power and decision-making

not what you might think; more a disquisition on the value of complete data in decision support.

example 1: whenever we consider strategies that individuals can implement for reducing carbon emissions, the first thought seems to be to throw on a bunch of solar panels. this particularly because, in california, of tax rebates and solar installation subsidies. the reality is that the first thing to do is increase energy efficiency in the home rather than change the means of energy generation. household energy consumption could be reduced by between 15% and 25% -- an immediate savings that is significantly out of proportion to the cost of an efficiency audit and the usually inexpensive fixes that they recommend (a heater blanket, auto-off power plugs, properly sealed windows and doors, etc).

example 2, and similarly: biofuels (particularly corn ethanol) have become fashionable as environmentally-friendly substitutes for miscellaneous petrochemicals. only recently has systematic research been published on the degree to which biofuels reduce net greenhouse gas emissions -- and the indications appear to be that biofuels may actually harm rather than help. why is this? several reasons, all associated with absence of full-cost accounting when evaluating biofuels:

• land-use patterns changed dramatically when biofuels became popular. two articles (here's one, and the other) in the feb 8 issue of science argue that increased demand for biofuels changes land-use in ways that were difficult to predict and which liberate greenhouse gases. (a gross oversimplification of one of the cited causes: converting US corn production to ethanol production causes, among other things, land in the tropics to be deforested for other grain and seed production to compensate. no good.)
  
• nitrogen-based fertilizers used in producing biofuel increase the amount of nitrogen dioxide, also a greenhouse gas, liberated into the atmosphere. (the reference is from crutzen's recent paper in ACPD.)
  
• moreover, the amount of energy required to produce ethanol from raw biomass only infrequently reduces total energy use. corn ethanol is an instructive example. to grow and produce 1 calorie of energy from corn grown in the US generally requires 10 calories of petroleum-based energy (the variation in the input-output ratio varies, but is never even close to 1:1). part of this is due to the general inefficiency of the corn plant as an ethanol producer. most of the biomass of a corn plant -- the stalky, leafy part -- is not amenable to ethanol conversion. we spend a lot of energy growing an annual plant, about 85% of the biomass of which cannot be used for ethanol. new processes for converting what is currently waste cellulose to so-called cellulosic ethanol may change this, but are still only in initial pilot phases.

whether implicitly or explicitly, we make decisions about complex systems using models that have far fewer degrees of complexity. sometimes that just doesn't cut it.