Abstract
A stochastic regression model is used in modeling rain rate. Under some conditions on the model parameters, it is shown that rain rate is asymptotically lognormal. An application of the model to the GATE (global atmospheric research program, Atlantic tropical experiment) data shows a remarkable agreement between the assumed and estimated model parameters for rain rate averaged over sufficiently large area and a sampling interval of 15 min.