An analysis of the effect of decreasing the size resolution in two techniques to solve the stochastic collection equation for growing cloud drops

A comparison is made between the methods for solving the stochastic collection equation of Berry and Reinhardt, using 72 categories, and the method of moments as improved by Tzivion et al., using only 36 classes. The computations, carried out for different kernels and several initial mean drop radii, showed that the numerical acceleration inherent to the Bleckʹs method is also present in the approach of Tzivion et al., but in a weaker form. This acceleration depends on the type of kernel and, for cloud conditions allowing coalescence development, it remains between acceptable limits and decreases with time. When the method of moments is run for 72 categories, the numerical acceleration practically disappears. Although, in terms of mass storage, the requirements of both methods are the same, in terms of computational speed, this approach has advantages over Berry and Reinhardtʹs method for use in cloud models with detailed description of microphysical processes.