Let's see if I can clear this up.
Let's take a look at something a little easier than air:
On a flat surface, dashboard for example, place an object. At low acceleration, the object doesn't slide. Why? Because the force of friction is greater than the force required to accelerate the object at the car's acceleration.
Let's say the car is accelerating at 1 meter per second each second. (increasing the velocity by 1 m/s every s, or abbreviated by the physics folks as 1 m/s^2
Now, we use the coefficient of static friction for the object against the dashboard (which would probably have to be determined experimentally)
Why do we want the static friction? Because we don't want the object sliding on the dashboard (this is the kinetic coefficient of friction which will generally be lower)
Force of friction = coef * Normal force
F(f) = mu*F(n)
(on a level surface, normal force is equal in magnitude to the weight.)
So, the car is accelerating at 1 m/s^2
the acceleration of the car is caused by a force in the direction of the motion of the car. This happens where the tire meets the road. The tire pushes backward against the road as it tries to spin, and <equal and opposite forces> the road pushes back against the tire. Anyway, the car is accelerating at 1 m/s^2.
For the object to stay on the same spot on the dashboard, it needs to have the same acceleration as the car.
since, F=ma, and a = 1.0 m/s^2, then it needs a force equal to it's mass times 1.0 m/s^2
This force is provided by the friction between the object and dashboard.
Let's say it has a mass of 10 kg, and a coefficient of static friction of .30 (so I don't have to use so many symbols)
Then, it needs a force of m*a = 10kg *1m/s^2 = 10 kgm/s^2 or 10 newtons.
The force of static friction that can be provided is F(f) = .30 * weight.
the weight, from the acceleration of gravity, is 98.1 Newtons (F=mg)
.30 * 98.1 = 29.4 Newtons. So, up to 29.4 newtons of force can be provided by the static friction. Only 10 is necessary for the acceleration, therefore, the object doesn't slide on the dashboard.
**Note: if you double the mass of the object, you double the amount of force necessary for acceleration. However, you double the normal force as well, therefore you double the frictional force. It all balances out.
Now, if you attempt to accelerate the car at 3m/s^2 (roughly increase your speed by 6 mph every second, or 0 to 60 in 10 seconds (actually, more like 9.1 seconds))
Available force of friction is the same. Using F=ma, with the 10kg object sitting on the dashboard, it needs a force of 30 newtons now to accelerate. Unfortunately, static friction provides only 29.4 Newtons. Therefore, the object starts to slide.
Now, once it starts sliding, this is kinetic friction. The coefficient of kinetic friction will be lower, let's say around .20
So, friction is providing a force of .20*98.1Newtons = 19.6 Newtons.
F=ma
F/m = a
19.6N/10kg = 1.96 m/s^2
The object will accelerate at 1.96 m/s^2.
However, the car is accelerating at 3.00 m/s^2. Relative to the car, the object is moving backwards, at an acceleration just over 1 m/s^2.
So, if you're leaning backward in your seat, and accelerate the car at 3m/s^2. Expect a whole lot of coffee to be landing in your lap.
Now, back to the fly...
The force for acceleration of the fly will be provided by air resistance. Simply because without motion, there is no air resistance, the fly is going to end up moving toward the back of the car. Air resistance is proportional to (often used) velocity squared. If v=0, air resistance = 0.
However, the air itself is also going to be accelerating toward the back of the car (relative to the car) Because of inertia, the air doesn't want to accelerate. This results in higher density of air at the rear of the car and lower density of air at the front of the car. An interesting experiment to verify this: Hang a balloon on a string from the ceiling of the car, and have a helium balloon floating at about the same height. As the car accelerates forward, the regular air balloon will go backwards (as will a fly clinging to its surface, and a fly that's hovering in the air. However, as the car accelerates forward, the helium balloon will actually move forward in the car. I've got a video of this on my computer that I downloaded somewhere on the net, but don't have the time to search for it now (where it's located on the net.) It's a pretty common demonstration in college physics.