We study the evolution of spectra of acoustic waves that are generated in the convection zone and propagate upward into the photosphere, where we compare the simulated acoustic spectra with the spectrum observed in an Fe I line. Although there is no pronounced 3 min component in the spectra generated in the convection zone, there are dominant 3 min features in the theoretical spectra, in agreement with the observed spectrum. We interpret the occurrence of the 3 min features as the response of the solar atmosphere to the acoustic waves which shifts high frequency wave energy to low frequencies. We also find qualitative agreement for the acoustic power between the wave simulations and the observations.

We study the response of the solar atmosphere to excitations by large amplitude acoustic waves with radiation damping now included. Monochromatic adiabatic waves, due to unbalanced heating, generate continuously rising chromospheric temperature plateaus in which the low frequency resonances quickly die out. All non-adiabatic calculations lead to stable mean chromospheric temperature distributions determined by shock dissipation and radiative cooling. For non-adiabatic monochromatic wave excitation, a critical frequency f_cr ~ 1/25 Hz is confirmed, which separates domains of different resonance behaviour. For waves of frequency f < f_cr, the resonances decay, while for waves of f > f_cr persistent resonance oscillations occur, which are perpetuated by shock merging. Excitation with acoustic frequency spectra produces distinct dynamical mean chromosphere models where the detailed temperature distributions depend on the shape of the assumed spectra. The stochasticity of the spectra and the ongoing shock merging lead to a persistent resonance behaviour of the atmosphere. The acoustic spectra show a distinct shape evolution with height such that at great height a pure 3 min band becomes increasingly dominant. With our Eulerian code we did not find appreciable mass flows at the top boundary.