Evaluation of global and CCIR models for estimation of rain rate statistics

The prediction of the statistics of attenuation by rain on a radio wave propagation path requires first the prediction of rain rate statistics for a point on the path and then the conditional prediction of the path-averaged attenuation, given the rain rate statistics. In this paper, two models with worldwide applicability, the Global and CCIR models for effecting the first step of the prediction process, are evaluated by comparing model predictions with rain rate measurements. The use of single-year, measured rain rate statistics as a model for the prediction of the statistics for subsequent years is also evaluated. Surface rain rate distributions for one or more years of observatins at 52 locations were employed for the model evaluations. On the basis of the available data the Global model performed best. The location-to-location variations in rain rate at specified probability levels and in probability at specified rain rate levels were found to be well described by lognormal distributions. No significant differences were found between year-to-year and location-to-location variability estimates relative to the two rain rate prediction models. The data show that the climate models are equivalent to inference from limited-duration records and suggest that prediction variability can be reduced by subdivision of the rain climate zones into smaller regions.