Consider something like the Morse potential
, which would be a decent approximation for the potential energy surface of a diatomic molecule like iodine. As you should hopefully know, the vibronic energy levels get closer together at higher energies because there is a dissociation point (dissociation energy), i.e., an energy above which the two iodine nuclei are no longer bonded. Above the dissociation energy, the molecule is no longer in a bound state. That is, the system is no longer confined by a potential well, particle energies/observables are no longer effectively quantized, etc. Like the states of a particle-in-a-box as the box dimension approaching infinity: the states become infinitely close together.
Intensity of peaks decreasing - no, this is not directly related to the energy level spacing. Rather, it is related to the Franck-Condon overlap integral
between the ground and excited vibronic state wavefunctions.
(As a final note, the various factors that influence peak intensities and etc change if one is talking about vibronic spectra or pure vibrational spectra.)