Solar and lunar semidiurnal wind and temperature oscillations from the surface to 400 km are simulated for average solar activity conditions utilizing the numerical tidal model described in part 1. Hough mode decomposition of the solstitial solar semidiurnal tide excited by H2O and O3 insolation absorption below 80 km indicates that it is characterized by a predominance of (2, 2) below 60 km; about equal contributions of (2, 2), (2, 3), and (2, 4) between 70 and 90 km; mostly (2, 4) with some (2, 5) and (2, 2) between 90 and 120 km; and a predominance of (2, 2) above 140 km with secondary contributions from (2, 3) and (2, 5). The solstitial lunar semidiurnal tide is characterized by a predominance of (2, 2) below 70 km; (2, 4) with secondary contributions from (2, 2), (2, 3), and (2, 5) between 80 and 110 km; and by (2, 2) with a fairly strong secondary contribution from (2, 3) above 120 km. Similar descriptions for both the solar and lunar oscillations apply at equinox, except that the asymmetric (2, 3) and (2, 5) modes are absent. Relative to the solar component, lunar temperature and wind amplitudes are sufficiently large (~20%) in the lower and upper thermosphere to account for some of the observed variability previously attributed to the solar tide. Amplitudes of the (2, 4) and (2, 5) modes above 100 km are significantly affected by mesospheric eddy diffusion coefficients in excess of 2¿106 cm2 s-1. The total solar semidiurnal exospheric temperature oscillation for average solar conditions ranges from ~10--20 ¿K poleward of 40¿N to ~40--60 ¿K equatorward of 30¿N and has its origin in three sources of excitation all of comparable importance: (1) EUV solar radiation absorption in the lower thermosphere (100--200 km), (2) in situ momentum coupling due to the interaction of diurnal winds and diurnally-varying ion drag, and (3) tides propagating upwards from below 100 km. In addition, the calculations are compared with rocket, meteor radar, partial reflection, Thomson scatter, and satellite data to evaluate the consistency of the model with observations.