Abstract
A model for numerical prediction of the 1000-mb surface is developed which includes a term expressing the interchange of sensible heat between the air and the underlying surface as well as the effect of terrain-induced vertical motion. In spite of the crudeness of the non-adiabatic representation, the model shows a definite improvement over a similar adiabatic model when the two are compared in a series of prognoses. Moreover, when monthly mean isotherms may be used to represent the temperature of the underlying surface, the non-adiabatic term may be combined with the orographic term and the earth's vorticity so that there is no work added to the prognostic routine.
