Your calculation is correct (well, I assume it is, it looks like it should be) but your reasoning is still wrong. If the correct answer (horizontal and vertical are independent, as green1 has said) seems unintuitive, think about angles. You might be travelling at 22mph when you hit the ground but you hit it at an angle and slide - it's a glancing blow. The faster you are travelling horizontally or the slower vertically, the more glancing it is - until, in the limit, at 0mph vertically you don't hit it at all and merely skim it forever - in that case you can see quite clearly that the horizontal component makes no difference at all to impact speed because there is no impact when you're travelling parallel with teh ground. It's the vertical component (strictly, the component perpendicular to the ground, if there are hills, humps, or kerbs involved) that's the relevant bitI think that Green1 is probably right, although he didn’t really explain how he reached his conclusion. Others may have also reached this conclusion but I couldn’t be bothered to read all the posts It is all about vectors as someone has already stated. Imagine a right angled triangle with the forward speed of 20 mph represented by the long side and , say, 10 mph the vertical speed of your head, by the sort side. The hypotenuse represents the speed that your head will hit the tarmac, which is the square root of the sum of the squares of 20 and 10. ie 400 + 100. Square root of 500 is about 22.3. I might be totally wrong as it is a long time since I failed to pay attention at school, but seems logical to me. However, if you pay attention to your riding, rather than trying to solve questions like this when you are out cycling, then you probably will never find out whether I am right or not.
(Friction, now, that's a separate issue. If the helmet rubs against the ground and stops very quickly, or even worse catches on something instead of sliding, that's when you want to start worrying about horizontal speed)