meteorological conditions would be more uniform. The similarity in the curve shape indicates that the fluctuations in the 10 min averages are related to atmospheric effects and are not due to random instrument fluctuations.As before, the standard deviations in refractivity obtained from the meteorological instruments are about a factor of four larger than those derived from the receiver fluctuations measured. This is attributed to the path-averaging experienced by the propagating microwave signals.v. RESULTS AND DISCUSSION OF COMPARISONS OF PREDICTED AND
MEASURED SIGNAL STRENGTH FLUCTUATIONSMeteorological measurements at eight heights plus the surface made at one location near the RF signal propagation path provided the information required to calculate the refractive index profile and its time variations. The 10 min averaged data sampled on the hour provided the refractivity profile needed by the RF signal prediction model. Calculations were made to estimate the expected signal intensity changes from hour-to-hour at the ten receivers. The predicted changes were compared with the hour-to-hour changes obtained from the averaged 10 min data at each receiver.There was also significant interest in evaluating the ground plane effects on the signal distribution at the receivers. The long term averaged receiver pattern did not correspond to the theoretical model using a ground reflection coefficient of -1. The peak in the measured signal distribution occurred at about 0.8 of the predicted height. The lowered peak could be the result of variations in the grazing angle over the Fresnel zone and the associated variations in reflection phase. The absence of sharp nulls could result from additional multipath signals and the fact that short term refractive changes would tend to cause rapid and extreme changes at the expected null points. The time averaging process used would result in null filling.To establish a reference profile for the prediction model, the reflection coefficient was varied with grazing angle to obtain a best fit to the measured profile "unperturbed" by atmospheric inhomogenities. Variations about this reference profile were then used for comparison of the predicted and measured effects of atmospheric inhomogenities. This is shown in Fig. 6 for eight different levels. The shape of the curves compare best during the evening hours when there is usually less mixing. In addition, the predicted changes are generally a factor of two or more larger than the measured changes. This latter effect can be attributed to the fact that the model predictions are based on meteorological measurements made at one point, but the propagating signal experiences the average meteorological conditions existing over the propagation path.It should be noted that the measured intensity changes at levels 2 and 10 are almost identical in shape and that the measured intensity changes at levels 2 and 6 are mirror images. This indicates that the dominant effect of the refractive change is to shift and stretch or compress the theoretica...