Improved models for long-term prediction of tropospheric scintillation on slant paths

The prediction models for tropospheric scintillation on Earth-satellite paths from Karasawa, Yamada, and Allnutt (1988) and the ITU-R are compared with measurement results from satellite links in Europe, the United States, and Japan at frequencies from 7 to 30 GHz and elevation angles of 3 to 33/spl deg/. The existing prediction models relate the long-term average scintillation intensity to the wet term of refractivity at ground level. The comparison shows that the seasonal variation of scintillation intensity is well predicted by this relation, but for the annual average some additional meteorological information is needed. A much better agreement with measurement results is found when a parameter representing the average water content of heavy clouds is incorporated. This confirms the assumption that scintillation is, at least partly, associated with turbulence inside clouds. The asymmetry between the distributions of signal fade and enhancement can also be explained by turbulence inside clouds. The asymmetry depends on the intensity of the scintillation, which is consistent with the theory assuming a thin layer of cloudy turbulence. A new model based on this theory predicts the distributions of signal fade and enhancement significantly better.