In the coming years, there will be more usage of the millimetre/sub-millimetre frequencies due to congestion of the lower frequencies. At these frequencies, precipitation greatly affects the quality of service, by attenuating signals, hence the need for a thorough study and understanding in order to design mitigation techniques for improved signal quality. Previous studies modelled rainfall data using mostly unimodal statistical distributions, which may not fit multimodality encountered in the data. This paper looks at the prediction of the number of modes, given rain rates and wind speeds by looking at the occurrence of multimodality in rainfall data captured at Chilbolton Observatory, southern England from 2003 to 2009. From the drop size distributions, it develops a novel model based on the Gaussian mixture model. This enables the multimodal distributions observed by various researchers to be modelled. It provides expressions for calculating model parameters as a function of rain rate, R (mm/h) and wind speed, W (m/s). The model parameters include number of modes N m , standard deviation σ 1 – σ m of each mode along with corresponding means, μ 1– μ m . The study concludes that multimodality exists, and the average number of modes tends to increase with increasing wind speeds and rain rates.