Posted by & filed under CSAG Blog, Frontpage.

So it’s the carbon-footprint season… First Smaug the dragon, then CSAGer, the climate scientist. Time for another monster – the cyclist. Why? I cycle to work. And nothing makes one’s day better than an opportunity to feel smug about his or her contribution to resolving one of the most daring challenges that society faces today. And by comparison to the users of oil-fuelled machines, we cyclists seem to have a reason to feel so. And unfortunately, the CSAG’s carbon budget does not give us (the cyclists) the opportunity to show it.

So what is the carbon footprint of the daily commute as executed on the bicycle saddle? There is some confusion about that. One extreme is to assume that cyclists do not emit anything. This approach is present in e.g UCT’s carbon footprint calculations, which notes that 16% of staff and students commute by walking and cycling, and sets their carbon emissions at 0. But that’s obviously, as shown below, a simplification.

On the other side of the spectrum is the radical position which considers that because humans metabolise and respire, they are a source of CO2, no matter whether they cycle, drive, walk or sleep. Since carbon emissions are calculated in the context of climate change mitigation, to achieve a truly deep reduction of emissions we all, cyclists, drivers and walkers, would have to cease to be. According to that standpoint, because the demand of ridding the planet of all life is absurd, there is no point in even calculating how much who emits and why.

The truth, as it usually does, lies somewhere in between. There are a number of factors contributing to CO2 emissions of a bicycle commute, some more, some less important. I’ll try to detail as many of them as possible within the space of this blog. Of course, calculations like this have been done before, for example by MIT students, by the European Cyclists’ Federation and within a number of blogs. They differ in assumptions and details, and because I did not find them particularly consistent, I decided to do my own version (although I have borrowed heavily from them).
Boundary conditions
Before we start – some boundary conditions:

  • What we’re interested in is carbon emissions occurring during a cycling commute or in relation to it.
  • In this we implicitly assume that there is no alternative to commuting from home to office or school, although there is a choice of how to commute or move about.
  • Different means of transport will have different factors contributing to their total emissions. What we are interested in are those that are specific to cycling – different than these occurring during alternative ways of commuting.
  • To assess the level of moral and environmental superiority cyclists can achieve, we will compare their emissions with these generated by commuting drivers. Why drivers and not public transport commuters? The majority of UCT staff and students (42%)[1] commute by private cars. If you are commuting by Jammie Shuttle, or other public transport – and want to feel better – write your own blog! Same applies to motorcyclists;-)


Direct emissions from fuel
Bicycles are powered by cyclists. Cyclists are powered by food. To move their muscles, they burn carbohydrates, fats and proteins as well as organic acids, polyols, and ethanol in the process of aerobic or anaerobic celular respiration. The simple way to obtain direct carbon emissions from bicycle fuel is to look at carbon dioxide content in the exhaust gases. Cyclists breathe out more CO2 than drivers sitting comfortably behind the wheel. This is obvious even to a proverbial American politician – one recently noted that “the act of riding a bike results in greater emissions of carbon dioxide from the rider”, and decided to tax cyclists for that. It did not work, though.

So how much is this anyway? A person riding bicycle slowly (say 15 km/h, which I consider equivalent to doing light work) breathes 0.08 m3 of CO2 per hour. A hard riding cyclist (say 25 km/h, which is like doing hard work, at least for me) would breathe 0.38 m3 of CO2 per hour. These translate into 10.5 and 24.9 g CO2 per km of commute. For comparison, a comfortably sitting driver moving with a speed of 50 km/h would breathe out 0.8 g CO2 per km travelled (unless they get involved in a shouting match with a cyclist, that is). So the cyclist emits quite much more than the driver. Right? Right. But from mitigation point of view this carbon does not matter because it originates from food we eat, and we eat plant and animal products. We do not eat coal or crude oil. To provide fuel for the bicycle commute, we do not introduce new carbon into the atmosphere. Rather, we simply recycle what is already there.

For comparison, cars powered by internal combustion engine burn hydrocarbon fuel. “Tail-pipe” emissions from that vary widely depending on vehicle type and size, engine type, fuel type and quality, road conditions, driving distance, weather conditions and driving style. As a result we cannot, unfortunately, be too precise here. In general, one can consider direct fuel emissions from a personal petrol or diesel fuelled car to fall within the 140-260 gCO2/km[2, 3]


Indirect emissions from fuel
Apart from the direct, “tailpipe” emissions from fuel, one should consider fuel’s indirect carbon footprint.
In case of cycling, production and distribution of food has its own carbon footprint resulting from the use of fertilisers, transport, distribution, packaging etc. We can calculate this footprint knowing how much energy does cyclist require over that needed by the driver, and knowing the carbon footprint of food required to provide that extra energy.

According to Compendium of Physical Activities, resting requires approximately 1 kcal per hour per kilogram of body mass. The closest I could find to driving a car (rather than being driven in a car) is “sitting – arts and crafts, light effort”, perhaps equivalent to shifting gears and pressing accelerator, requires 1.5 kcal/kg/h. For comparison, “bicycling, 16 km/h, leisure, to work or for pleasure” requires metabolic rate of 4.0 kcal/kg/hour. Riding hard, “22-25 km/h, racing or leisure, fast, vigorous effort” requires 10 kcal/kg/h. These translate into extra 2.5-8.5 kcal/kg/h of energy needed by a cyclist over that needed by a driver. One should note at this moment, that from the point of view of this story it does not matter if our cyclist eats more or less on a particular day, whether they burn they accumulated body fat, or sugar and proteins they ate for breakfast. Conservation of energy rules. The extra energy the cyclist needs has to come from somewhere. Since the only way human body can obtain energy in the long term is from food, so cyclists have to use more of their food for commuting than the drivers do.

How does the extra food intake translate into carbon emissions depends on what one eats, because various foods have various carbon footprint. Data from South Africa are difficult to come by, but in US it looks like this:

This of course lends itself to an argument whether a carbon footprint of an aggressively cycling “beefarian” is comparable with that of a vegan driving a small car. Such a question would be valid for an individual, but if we want to scale up the analysis to the population of cyclists and drivers (or more general non-cyclists), the real question is:


do cyclists eat differently than the non-cyclists do?


I have not managed to find any data on that. But from semi-rigorous lunch-time survey, it seems that the eating habits of cyclists and non-cyclists may vary between nationalities. South African (n=4), Polish (n=1) and British (n=1) cyclists living and working in South Africa eat relatively similarly – a diet dominated by cereals and pulses, occasionally incorporating meat. The diet of South African non-cyclists is not significantly different. However, French (n=1) and British (n=1) non-cyclists more frequently than not indulge in greasy fast food of suspect provenance and unknown composition. Surveys continue to obtain a representative sample of diet of American and Australian non-cyclists who recently joined the group, but at this stage, we will assume that there is hardly any difference between the eating habits of cyclists and non-cyclists.

Under the above assumption, we can consider carbon emission associated with the “average” diet from the figure above applicable to both cyclists and non-cyclists (although perhaps we should reduce the total calories intake – average Americans are not epitome of slim). Carbon emissions associated with providing 3600 kcal per person per day in the form of average american diet amount to 2.5 t CO2 per year, or 1.75 g CO2 per kcal. Note the the american diet has only 66% “food intake efficiency”, i.e. a person takes in only 66 % of calories they actually bought to eat. In EU, the values are a little different. An average european needs 3600 kcal/day, and the provisioning of the “average” food needed to satisfy their energy demand generates 1.83 t CO2/year[3]. This is equivalent to 1.44 g CO2/kcal. Although the source does not explicitly state it, it seems that the EU numbers too account for “food intake efficiency”, and rightly so. If we use EU numbers, the commuting by bicycle generates 3.6-12.2 CO2/km from “fuel”, depending whether one rides soft or hard.

For cars, the indirect carbon emissions from fuel are often termed “well-to-tank” emissions. These result from energy expenditure and emissions occurring during the process of crude oil extraction, processing and transportint products all the way to a filling station. The “well-to-tank” emissions vary from country to country depending on source of oil. In US the “well-to-tank” emissions are estimated at 24-46 gCO2/km[2].
The vehicle
But fuel is not all. One has to consider carbon footprint associated with manufacturing and maintenance of the bicycle. Under assumption that the average commuter bicycle weighs 19.9kg, and is composed of 14.6 kg aluminium, 3.7 kg steel and 1.6 kg rubber and that the bicycle will last 8 years and cover a distance of 2400 km each year, bicycle production and maintenance generates approximately 5 grams CO2 5/km[2]. That’s one heavy bike, though. Similar calculations for US, although with slightly different assumptions (15 year lifetime, 2000 miles/year, US manufacturing emissions, unspecified weight of the bike), give 5.6 g CO2/km travelled[3]. Of course one can argue here about lighter bikes, and country-to-country differences in carbon efficiency of energy and raw materials production, as well as bicycle manufacturing process. But the point is, 5 gCO2 per km travelled resulting from bike manufacturing is roughly 11-19 % of that of a car (in US: 30 gCO2/km for a sedan, and 36 gCO2/km for a SUV. In EU – 42 gCO2/km for a sedan). The difference between bike’s and car’s per km emissions is in fact rather low – one would expect that manufacturing a bike is equivalent to manufacturing a tiny fraction of a car, say, a car door, and the difference would be larger. However, the larger lifetime mileage obtained from a car as compared to that from a bicycle offsets the per km emissions.

The infrastructure
Now, we are getting into a rather shaky ground. The next batch of emissions is associated with infrastructure – paths, roads, parking lots, both their creation and maintenance. This is where calculations become rather tricky. One can, of course commute without, or with minimal road infrastructure, both by bike and by an off-road vehicle. But in a typical urban setting one rarely does. So all the carbon emissions associated with creating that infrastructure should be included into the total carbon footprint of a commute. One can commute by bicycle without special bicycle paths, simply sharing roads with cars, and without special bicycle parking spaces, keeping bikes on corridors and in office rooms. But that’s some sort of carbon emissions free-ride. This is OK for an individual or a couple of them. Scaling up the numbers would require some dedicated construction work with all the emissions thereof. Under US conditions, the carbon costs of road network creation and maintenance turn out to be 37-44 gCO2/km for a car[3], the spread attributed to the range of weight of the vehicle, and thus damage to the road caused. For dedicated bicycle lines, either build on purpose, or appropriated from existing roads, these turn out to be 11 gCO2 per km travelled[3].


One last detail
Before we go to final conclusions, we have to note that the table above presents results for a vehicle, not for a commuter. It is not likely, although possible (believe me, I’ve lived in the source of all cycling evil – the Netherlands), to commute with two people to a bike (we would have to change the calculations slightly then). But is is likely (and perhaps increasingly so with the higher and higher fuel prices) that there is more commuters to an average car. However, numbers are not that optimistic. In US, 75% commuting cars have only one passenger. In EU – average occupancy is 1.16 commuters per private vehicle with average occupancy rate for all private car trips of 1.57[2]. For South Africa, average occupancy rate for private cars in Johannesburg is 1.4 [4]. Let’s be generous and consider that UCT staff and students commute with 1.4 person per car.
and some exotic factors…
There is a number of factors that could potentially be included into the calculations, but ultimately they would disappear within the sea of uncertainty. Consider for example these:
– car’s commute route is usually approx. 5% longer that that of a bicycle. Implications to the emissions calculated per km are obvious.
– cyclists may or may not need to shower after the ride, and may or may not need to wash their clothes more frequently, depending on the level of tolerance of their workmates. Showering and washing have their carbon footprints.
– there are health implications to riding the bicycle. Regular cycling brings a life expectancy gain of 3-14 months against a loss of 0.8-40 days due to exposure to fine dust particles and 5-9 days due to traffic accidents (that’s when you commute for most of your life). One could look at emissions associated with health care, as well as these associated with the 14 more years of being a resources and energy consuming retiree.
– an interesting effect may occur in relation to differences in time necessary to cover the distance of a commute. In a typical urban setting, cycling is faster than driving up to a certain critical commute distance. That distance is approximately 6-12 km. One could wonder what happens with the extra time the driver (or the cyclist) get. There will be consequences if the one with spare time engages in carbon-emitting activities (such as, for example, any activity leading to the increase of GDP).

to summarize…

Carbon dioxide emissions during commute, in gCO2 per km travelled per person

source by bicycle by car*
“tail-pipe” emissions: 0** 100-185
“well-to-tank” emissions: 3.6-12.2 17-33***
manufacturing and maintenance emissions: 5 21-30
infrastructure-related emissions: 11 26-31
total 19.6-28.2 165-280

* per person, with average occupancy rate of 1.4 person per vehicle
**CO2 respiration only recycles CO2
*** increased driver’s metabolism related to the excitement of driving and the need to shift gears, steer and remain alert is ignored


That’s basically it. An average commuting cyclist have carbon footprint roughly equivalent to 10 % of that of an average commuter by a personal car. But does the difference matter? Let’s consider a person living 10 km from UCT, who cycles (slowly) to office every day, for 200 days a year. He/she would create 80 kg of carbon dioxide emissions. Average commuter riding a small car would generate 660 kg of CO2. Considering that CSAG’s overall carbon footprint is 10400 kg per capita per year, cycling to work instead of driving can offset 5.5% of that total. CSAG’s carbon budget used a figure of 200 kg of CO2 for carbon emissions of a one-way Cape Town-Johannessburg flight. A year of cycling instead of driving offsets emissions of almost three of such flights.


That’s not much. But not little either. Every percent matters. What is also important is this:


1. Letete et al. 2010. University of Cape Town Carbon Footprint.
2. European Cyclists’s Federation. 2011. Cycle more to cool down the planet.
3. Dave.S. 2010. Life Cycle Assessment of Transportation Options for Commuters. MIT, Massachusets.
4. Merven et al. 2012. Quantifying the energy needs of the transport sector for South Africa: A bottom‐up model. Energy Research Centre, University of Cape Town.

Figure/image credits:

Comments are closed.