The H Balmer &agr; nightglow is investigated by using Monte Carlo models of asymmetric geocoronal atomic hydrogen distributions as input to a radiative transfer model of solar Lyman &bgr; radiation in the thermosphere and exposhere. The radiative transfer model includes all orders of scattering, temperature variation with altitude and solar zenith angle, and anisotropic velocity distributions. The influences of multiple scattering of Lyman &bgr; radiation and observing geometry on the H Balmer &agr; intensity and effective temperature are evaluated in detail. Morning and evening hydrogen distributions for minimum, medium, and maximum solar activity are used in calculations of nightglow emission rates and line profiles. For each of the hydrogen models of H Balmer &agr; intensity and effective temperature are displayed as a function of solar depression angle, observation zenith angle, and azimuth relative to the sun, to make possible detailed comparison to observations. Except for rather unique observing conditions, multiple scattering effects cannot be ignored for the determination of either intensity or temperature.

Observed effective temperatures which are significantly less than the exobase temperature may be due either to low-altitude thermospheric hydrogen or to high-altitude expospheric hydrogen and depend on the observing conditions and the level of solar activity. Previous (selected) observations of the H Balmer &agr; nightglow and effective temperatures are compared to models which most closely represent the solar conditions of the observations. Results of the theoretical investigation of H Balmer &agr; emissions and of model-data comparisons indicate that predicted morning-evening concentration variations yield small intensity variations for solar depression angles greater than 20¿. The variations with solar depression angle of observed intensities and temperatures both show discrepancies in comparison with model results. A substantial disagreement occurs in absolute intensity: the long-standing result that the observations are a factor of 2 higher than model results is reaffirmed with recent observational data and improved estimates of the solar H Lyman &bgr; flux. Observing schemes are discussed which lead to first approximations to hydrogen content independent of solar flux and to the high-altitude atomic hydrogen distribution. ¿ American Geophysical Union