People are finally realizing that we are disrupting the climactic conditions that allowed our species to evolve from hunter gatherers to sophisticated agricultural societies. As a result, there is increasing interest in regulating carbon dioxide. CO2 emissions are quite complex and indirect in some cases, but automotive transport should be one of the easier places to measure carbon emission, use, and efficiency. For example, many newer cars come with dashboard display units that tell you what your fuel efficiency is. Or so we are led to believe.
A year and a half ago, I blogged on the fuel efficiency of my old motorcycle, which was comparable to Jim and Callan’s Priuses. This autumn, I started tracking the fuel efficiency of our family car, a small turbodiesel European wagon. I did this the same way I did for the bike- dividing the fuel purchased by the distance driven to get the liters per 100 kilometers- the standard efficiency measure for metric driving.
As I was doing this, I started to notice that the numbers I was calculating did not agree with the dashboard display. While the fuel use method is known to be imprecise (it is hard to fill the tank to exactly the same level every time), it should be accurate over the long term. So I started recording both numbers. And after half a dozen refills, I have the following data:
Figure 1. Blue line is calculated efficiency, Pink line is values reported on dashboard computer.
Figure 2. Same data, converted into miles per gallon for the sake of unscientific Americans
With an average of 800 km between refills, the average fuel difference is about 5.5 liters per refill, or about 11%.
I checked the computer distance traveled against the odometer- the numbers are the same. The six refills came from four different service stations operated by two different companies (Shell and BP). No obvious difference. At this point, my working hypothesis is that the computer underestimates its fuel use, but I’m open to suggestions for how to test this hypothesis. In the mean time, if anyone else out there want's to check a different car model, it would be interesting to see what y'all come up with.