This paper reviews experimental and theoretical research on cross polarization that occurs during clear-air conditions on terrestrial microwave links. After consideration of the various clear-air mechanisms that have been suggested, the conclusion is drawn that multipath mechanisms interacting with the cross-polarized patterns of the antennas are the most important. Theoretical models that explain these mechanisms and semi-empirical models that characterize the cross-polarized signal statistics are then discussed. Particular attention is given to the semi-empirical models since they can be used most readily for statistical predictions in design applications. Some new insight into the similarities and differences between models developed in France, Japan, and the United States is provided. Finally, application of the semi-empirical models in the design of digital systems is briefly considered, and the relative importance of cross polarization during clear-air and precipitation conditions is discussed. Paper number 80S 1794. 0048-6604/81/0910-1794501.00 631 632 R. L. OLSEN o o o o o o CROSS POLARIZATION DURING CLEAR AIR 633 ß o 634 R.L. OLSEN
CLEAR-AIR CROSS-POLARIZATION MECHANISMSThe first cross-polarization measurements during clear-air conditions were those of Morita [ 1971] carried out in 1961. Since that time the results of a number of experiments have been published. In all these experiments, significant reductions in the measured XPD or XPI were observed. Table 1 lists the various experiments reported and provides the measured values of XPD not exceeded for 0.1% and 0.01% of the time where these statistics are available. In combination with a small number of theoretical analyses, the results of these experiments shed considerable light on the main clear-air cross-polarization mechanisms involved. Virtually all observations of severe deterioration of XPD during clear-weather conditions have been associated with the deep fading of the co-polarized signal that occurs during multipath conditions. A typical record of the co-polar and cross-polar signals in such conditions is illustrated in Figure 1 (R. S. Butler, private communication, 1980). In this example, the cross-polar signal is quite well correlated with the co-polar signal, except during the time interval of the deep co-polar fade. The highly correlated parts of the cross-polar signal records observed in the example of Figure 1 and in all experiments have been explained in terms of the various imperfections in the antennas and channel-separation networks which allow a constant fraction of the co-polar signal power to be coupled to the cross-polar channel (see section 3.2). The uncorrelated parts that occur during deep 0-CROSS POLARIZATION DURING CLEAR AIR 635
