This post is a guide for those wishing to understand how to enhance their performance on a bike by improving their aerodynamics. I have included some of the science and many of the insights we have gained from wind tunnel testing of cyclists. There is no need to buy expensive kit, indeed you can go 10 minutes quicker in your next 100 mile sportive just by changing what you wear.

The data for this post was collected as part of Michael Acuna Lesmes’ third year engineering individual project at the University of Southampton.

Putting aerodynamics into context

As well as the aerodynamic resistance, a cyclist’s power also needs to overcome the rolling resistance of the tyres. Although I won’t consider it in detail, it’s worth having a look at how rolling resistance and aerodynamic resistance compare. The good thing about rolling resistance is that reducing it is simply an equipment choice (weight loss is important too – see Nutrition off the Bike: Feast or Famine? Carb or Cow?), and there are data published to help us choose, e.g. Using data for the best and worst performing tyres from this site we can plot an upper and lower bound for the power required to overcome the rolling resistance alongside the aerodynamic drag for a standard road bike drop handlebar position.

Comparison of power to overcome the rolling resistance of ‘fast’ tyres (PR(lower)) and ‘slow tyres (PR(upper)), and the aerodynamic resistance of a comfortable drop handlebar position (PA). The total power using average tyres is also shown (PA+R) as is the extra power required due to a drivetrain of low (93%) and high (98%) efficiency  (Ptotal(low) and Ptotal(high)).

Clearly it’s worth choosing good tyres but, at higher speeds, aerodynamics is by far the most important consideration. Finally the efficiency of the drivetrain needs to be taken into account. We need to divide the power by between 0.93 and 0.98 depending on how well the drivetrain is running (run-in chain, not ‘crossing’ the chain, etc.). In summary, at 30 mph, rolling resistance accounts for ~33-67 W of power required, drivetrain efficiency for ~12-45 W and aerodynamics for ~543 W. The rest of this post will look at how much variation there is (and therefore possibilities for reduction) in this aerodynamic component of the required power.

How do aerodynamics affect speed and power?

Aerodynamic drag depends on some key quantities, which can be expressed in the equation:


where ρ is the air density, V the velocity of the air (i.e. the speed of the cyclist plus any head wind), CD is the drag coefficient, and S is area (the frontal area of the cyclist and bike). Sometimes, quite understandably, is used to denote area. In aerospace engineering we use S, so I’m sticking with that.

The dynamic pressure ½ρV(and so the drag) rises along with V2, making speed our biggest enemy. So, let’s rearrange things and move V to the left hand side of the equation to better understand what we need to do to go faster:


(apologies for the poor equation formatting). Many riders now use power meters and so power is a more familiar quantity to work with than drag (it’s power output that you feel in your legs too). The power, P, required to overcome the aerodynamic drag is simply the drag multiplied by the velocity, so we can write:




To increase our speed we need to make P bigger by doing more training, but this post is about increasing speed by making the bottom of the fraction smaller.

Weather dependent

Unfortunately we can’t do much about the air density, ρ, except choose to ride on a day when it is hot and the atmospheric pressure is low (density is proportional to pressure and inversely proportional to temperature). Riding at a record UK low of 926 mb (31st January 1902, Aberdeen) compared to the record high of 1054 mb (26th January 1884, Ochtertyre) would result in an increase in V of approximately 4.5%, assuming all other factors remained constant. Moving from freezing conditions to 30°C increases V by approximately 3.5%.

Of course such favourable conditions would be the same for everyone, except for an individual record attempt, which is why there was so much talk about air pressure in the build up to Sir Bradley Wiggins’s hour record (which he rode at an unfavourable 1030 mb).

Something we can change

The frontal area, S, and drag coefficient, CD, are particular to the rider and their equipment. That is, we can control and optimise these quantities.

It is easy to get confused about the effect of CD and S on the aerodynamic drag, and so the required power. CD is a property of a body’s shape and surface finish. A large truck may have a lower CD than a convertible sports car; it is long and thin and has various drag reducing fairings, whereas the sports car has a windscreen sticking up into the airflow, fancy spoilers, etc. However, when the much bigger S is taken into account, the actual drag of the truck is higher.

For a cyclist, S and CD are inherently linked. For example, crouching over the handlebars will change both S and CD. Some equipment changes may be made that affect only S or CD but, in general, we should consider both quantities as one variable we are trying to reduce: CDS. What we actually feel on the bike is not CDS but, for a given speed, the power needed to stay at this speed. Given a constant air density and speed, we can report changes in CDS as power:


Fine-tuning reductions in CDS through changes in position is very rider-dependent. The variety of time-trialling positions adopted by the top riders is not just down to some riders not having the opportunity (or not taking the opportunity) to use a wind tunnel. Riders’ body shape and style leads to, e.g. flat, wide forearms being better for some while the angled, close-elbow ‘praying mantis’ position being better for others. Although I will consider the benefits of optimising position in a future post, for now we will concentrate on more profound changes in position, clothing and equipment, the results of which will be applicable to most riders.


We calculated the power for a range of positions, clothing and equipment using the R J Mitchell Wind Tunnel at the University of Southampton. This is the facility where UK Sport and Scott Drawer’s ‘Secret Squirrel Club’, with the University, made all those marginal gains for UK cyclists in the build up to the Beijing and London Olympics.

The power we measure is that required to overcome the drag of the air pushing the cyclist and bike backwards. Note that it does not include the power required to turn the wheels, so we cannot fully assess wheel aerodynamic performance with the setup used to obtain the data for this post.

The test subject is me: height 183 cm, weight 67 kg. The default ‘road’ bike is my Specialized S-Works Roubaix XL with Shimano Ultegra Di-2, Fulcrum Racing 5LG CX wheels with 25 mm Michelin Lithion 2 tyres and a 750 ml bottle in a down-tube cage.

The test bike. It’s not supposed to be a super aero machine – just a bike.

Day 1

I cycle the short distance to the tunnel (with a rucksack full of kit options that arrived from Zepnat the day before). The rig that holds the bike to the force balance is all ready and my bike is quickly installed. The rig is like a very narrow set of rollers, but the bike is held front and rear by its wheel axles like a turbo trainer. Unlike rollers, the front roller is driven by a delicate cable, which means brakes are contraband – a rule reinforced by the large ‘NO BRAKES’ message being projected on the floor in front of me as I mount the bike.

As the tunnel starts, the noise and rush of air is a little intimidating and escalates rather rapidly, resulting in a slight panic that perhaps something’s gone wrong, the whole thing’s out of control and I’m about to be blended. Soon the feeling is one of slight smugness that I’m in a wind tunnel going really fast. Michael (the student getting his degree with the data we’re collecting) gradually ramps up the speed from 20 to 47 miles per hour, taking 30 seconds of force data at each speed, for a range of everyday cycling positions: on the hoods, drops, standing with hands on the hoods, and a standing sprinting position with hands on the drops. The idea is to test `normal’ positions, so on the hoods I’m in a relaxed position you might adopt on a club run, rather than crouched over the bars, getting as low as possible.

The tunnel speed is projected onto the floor in front of me so that I can try and match my wheel speed to it. At the fastest setting of 47 miles per hour (‘Chris Hoy speed’ the tunnel manager calls it), it’s hard to stay relaxed; I’m hanging on to the bars to stop myself being dragged off the bike.

As well as the four positions, we methodically work our way through a number of configurations representing potential kit choices for races/sportives. These configurations are detailed in table 1.

A number of tests were performed for each configuration at each speed. Minor changes in position, sensitivity of the measurement equipment, etc. mean that no two tests produce the same result. The results in the table are average values and are only reported if the probability that the measured difference between configurations was due to experimental errors is less than 5%, i.e. we are 95% confident that the result is due to the change we made to the configuration.

Table 1: day 1 test data (% difference are relative to baseline drops positions and only reported if statistically significant).

To a certain extent the numbers speak for themselves, but I will summarise the findings below and look at how we can interpret them.

Position matters

The largest changes we see in the drag data relate to changes in position. With a 10% increase in power required in the sprint position, you need to think twice before getting out of the saddle. This is relative to a ‘comfortable’ position on the drops, so compared to a tucked, finishing straight position, the penalty for standing up would be greater. One would normally only be en danceur on a slow climb when drag is less of a consideration. However, battling in to a stiff headwind, the 36% increase in drag will certainly make a difference.

Draggy bottle?

As well as assessing different positions, we wanted to look at the effect of some simple clothing and accessory choices. The first question we had was: does carrying a bottle make a difference? Working against gravity uphill it surely does, but as far as aerodynamics go, the data wasn’t convincing; while we often measured a lower drag with no bottle and cage, the results are not significant.

Helmet cover

The British Cycling team like to cover the vents on their helmets, and more and more pros are doing the same. So, does it make any difference? We tried out the Lazer Helium Aeroshell and it’s a resounding yes, with a 3% reduction in power.

Do you really need to wear a skinsuit?

Although yet to be de rigueur at café stops, skinsuits are ubiquitous in time trials and becoming more and more popular in road races. Is wearing one for a sportive worth the inevitable jibes? That may depend on how beach-ready your body is, but the power savings are considerable at 8% in the standard drop handlebar position.

So how much faster can I go?

Let’s say you’re riding with your hands on your hoods at 16 mph. Referring to table 1, your CDS=0.38 (actually that’s my CDS, but for this relative analysis my results will be indicative of gains you can make). From the figure below, reading off 16 mph on the horizontal axes, we see at the reference CDS=0.38 (the horizontal dashed line) the colour scale indicates 16mph too. Now, let’s say you decided to ride out of the saddle on the hoods. From table 1 we use CDS=0.48. At the same reference speed, moving up to 0.48 on the vertical axis, we see the colour scale indicates ~14.5 mph. This is the speed you would drop to riding in this position at the same power. If you decided to ride on the drops in a skinsuit, at CDS=0.32, we see your speed would increase to over 17 mph. Not bad!

Potential speed gains due to changes in CDS (for a constant power).

In the next figure you can see the implications on the time taken for a 100 mile ride. At the on-the-hoods speed of 16 mph, this is going to take you 6 1/4 hours. For the above change from CDS=0.38 to CDS=0.32, the time saving over 100 miles is around half and hour. Time to take in a café stop! But would you go into a café wearing a skinsuit?

Time to ride 100 miles due to changes in CDS (for a constant power).

In the next post I’ll look at some more clothing, positions and equipment (including a time-trial bike). You might also be interested in Aerodynamics: the Sagan position.

One thought on “All you need to know about bicycle aerodynamics: part I

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s