Parameterization of the collision-coalescence process using series of basis functions: COLNETv1.0.0 model development using a machine learning approach (original) (raw)

Geoscientific Model Development Discussions, 2021

Abstract

Abstract. A parameterization for the collision-coalescence process is presented, based on the methodology of basis functions. The whole drop spectra is depicted as a linear combination of two lognormal distribution functions, in which all distribution parameters are formulated by means of six distribution moments included in a system of equations, thus eliminating the need of fixing any parameters. This basis functions parameterization avoids the classification of drops in artificial categories such as cloud water (cloud droplets) or rain water (raindrops). The total moment tendencies are calculated using a machine learning approach, in which one deep neural network was trained for each of the total moment orders involved. The neural networks were trained using randomly generated data following a uniform distribution, over a wide range of parameters employed by the parameterization. An analysis of the predicted total moment errors was performed, aimed to stablish the accuracy of the...

Léster Alfonso hasn't uploaded this paper.

Let Léster know you want this paper to be uploaded.

Ask for this paper to be uploaded.