One aim of the Propagation en Air Clair et Météorologie 2 (PACEM2) experiment was to obtain a model for clear air propagation applicable to space diversity reception. The statistical results derived from nearly 4 years of continuous recording show an asymmetry between both diversity channels, the higher antenna being significantly more affected by multipath effects than the lower one. A possible effect of the antenna heights on multipath propagation occurrence is included in the latest version of the International Radio Consultative Committee (CCIR) prediction formula, which takes the inclination angle of the path into consideration, but the difference experimentally observed between both channels is much greater than the prediction derived from this CCIR formula. The object of this paper is to give a physical interpretation of the effect of antenna height on multipath occurrence leading to a quantitative agreement with the data. The physical model we use is built on available statistics of radio meteorological data. For each particular meteorological situation the shape and size of the region of interference between multiple propagation paths are then obtained by means of ray tracing. The results of these computations lead to the determination of the relative probability of occurrence of multipath propagation as a function of path length and antenna heights. A rather good agreement with the data is obtained when these parameters are given the values of the experimental link.