No, spacing the elements horizontally will narrow your azimuth beam (horizontal direction) IN ADDITION TO the elevation beam (aiming at the horizon). For ground antennas and vehicles at a distant airport you need BOTH. In any event, without the proper test equipment it is very difficult to get the maximum or proper performance from an array such as yours, the losses in spacing and phasing may offset much of the advantage of the multiple elements, but as I said before the additional two elements and a few more feet in height were probably not going to make much difference here, well tuned or otherwise, so you are not missing much.
The best thing to do now is use it for a while and look for common areas where the performance is below what is needed, then take your time and figure out if anything can realistically be done about it. For example, rarely can you go from a signal not being heard at all to solid copy, a jump of about 12db, which is sixteen times the signal strength, so even if your array is only running at half efficiency and you get it tuned perfectly you'd still need another eight times the signal strength to pull that walkie-talkie out of the mud.
The same goes for elevation, but it's even more depressing... each doubling of tower height gets you about 50% more range. Right now the radio horizon for your antenna is about 14km, so if you double the height to 30m that only gets you to about 20km, which is close to your requirement and should get base station antennas pretty solid, but still a bit less than needed for solid ground mobile traffic.
If it were me in your situation I would probably opt for elevation over complex array, a really good, wide band coaxial dipole antenna up as high as I can get it on a guyed mast extension and hardline (Heliax) transmission line.
The attached photo shows two arrays that between them are able to pick up all the ground and airborne traffic at an airport 14km distant, and the elevation of the arrays is just a few feet above yours, the 4-element beam being at about 16m and the coax dipole a few meters higher. The gain of the coax dipole over a normal ground plane is 2-3db and the gain of the beam is about 9db, both are very wide band and exhibit that gain across the aviation band. The other 2-element beam to the right is aimed at a different coverage area (out to sea).
So you see, for me to get the same performance from an airport 22km away I would need to raise those arrays to more than double their current height, like 40m, which would be pretty tough for all the antennas, but relatively easy for that omni on the top. By the way, there is little on the ground that the beam picks up that the omni cannot pick up as well, when we surveyed the location we did it with the omni only, but the beam was added to make solid copy of the planes at their gates, cars and hand-held radios on the ground. That's what an additional 6db will get you... things that are already heard being heard better.