I think several things are being confused here.
'Long range' is a very broad definition, certainly above around 30MHz (it varies considerably) the effects of ionospheric reflection are minimal and radio waves become 'line of sight', that limits them to the distance to the horizon, allowing for antenna height.
At higher frequencies, up to a few GHz, signals travel with relatively little attenuation from the atmosphere.
As the frequency gets higher, generally from around 10GHz and above, the effects of chemical absorbsion start to be more noticable and specific frequencies (according to the type of medium) have higher attenuation or reflectivity.
The speed of data that can be carried on a radio signal is related to the Nyquist theorem (although VMSK is theoretically possible) so for more data a higher carrier frequency is needed.
For imaging, the wavelength is of prime importance. If I draw an analogy, trying to measure tiny distances with a 1M rule is far more difficult than measuring them with a 1nM one. Generally, you want your measuring units to be much smaller than the item you are measuring. Beamwidth has little to do with it except in circumstance where (over long distances) it is easier to focus a signal with shorter wavelength. For example, the beam width attainable from a parabolic reflector is proprtional to it's area relative to wavelength.
Brian.