Lets look at a model. You are making the claim, so you provide the cited numbers to fill it in. To begin with - lets assume all joules are equally polluting. T

Lets look at a model. You are making the claim, so you provide the cited numbers to fill it in. To begin with - lets assume all joules are equally polluting. This isn't true, but we'll get to that. Jp = the number of joules needed to move a person a distance. Je = the number of joules needed to deliver Jp to the point of use. Jt = the total number of joules for a mode of transport Simply: Jt = Jp + Je Therefore a first approximation is to compare Jt(bike) with Jt(car) However there are other factors that must be accounted for in this equation: Jm - the total number of joules needed to manufacture the transport (full supply chain), and deliver it to first point of use. Jl - the number of joules needed by a person anyway - farming can't just be eliminated, people require food to remain people. (Jl stands for joule's life) Jl' - the number of joules needed to transport the minimal "keep people living" joules to point of use. It should be noted, that Je and Jl' are related in the bike case, so Je in the bike case will only be the extra energy required to deliver additional food, not the energy required if the Je for a bike were delivered in an entirely separate manner. With Je for a car - the energy source, energy delivery etc are an entirely different infrastructure. So factoring those in we get: Jt = Jp + Je + Jm + Jl + Jl' Of course - at the beginning I mentioned not all joules are created equal. So we need some sort of pollution index. I don't know a simple polution index formula, but assuming it exists, we will call it P. The output will be a normalized value that can be simply added to understand total polution. So: P(Jt)= P(Jp) + P(Je) + P(Jm) + P(Jl) + P(Jl') I want to note about the pollution function - the pollution of P(Jp) is essentially 0 in the bike case, because it is just metabolism. P(Jl) will most always be close 0, because it is metabolism again. The polution from delivery is captured in other variables. This model could be further refined into pollution per travelled km, rather than the pollution per joule used. It could be better refined by including the Jm costs as an amortization of efficiency over the total number of travelled km. Etc. But for now, the model as stands is a good way of determining the claims first order validity.