A Maximum Entropy Modelling of the Rain Drop Size Distribution

This paper presents a maximum entropy approach to Rain Drop Size Distribution (RDSD) modelling. It is shown that this approach allows (1) to use a physically consistent rationale to select a particular probability density function (pdf) (2) to provide an alternative method for parameter estimation based on expectations of the population instead of sample moments and (3) to develop a progressive method of modelling by updating the pdf as new empirical information becomes available. The method is illustrated with both synthetic and real RDSD data, the latest coming from a laser disdrometer network specifically designed to measure the spatial variability of the RDSD.