Parameterization of the collision–coalescence process using series of basis functions: COLNETv1.0.0 model development using a machine learning approach

• Published in: Geoscientific Model Development Vol. 15; no. 2; pp. 493 - 507
• Main Authors: Rodríguez Genó, Camilo Fernando, Alfonso, Léster
• Format: Journal Article
• Language: English
• Published: Katlenburg-Lindau Copernicus GmbH 21.01.2022; Copernicus Publications
• Subjects: Accuracy; Algorithms; Analysis; Artificial neural networks; Basis functions; Cloud droplets; Cloud physics; Cloud water; Clouds; Coalescence; Coalescing; Distribution; Distribution (Probability theory); Distribution functions; Distribution moments; Dynamics; Learning algorithms; Machine learning; Mathematical models; Methods; Neural networks; Numerical analysis; Parameterization; Parameters; Rain; Rain water; Rain-water (Water-supply); Raindrops
• ISSN: 1991-9603; 1991-959X; 1991-962X; 1991-9603; 1991-962X
• DOI: 10.5194/gmd-15-493-2022

A parameterization for the collision–coalescence process is presented based on the methodology of basis functions. The whole drop spectrum is depicted as a linear combination of two lognormal distribution functions, leaving no parameters fixed. This basis function parameterization avoids the classification of drops in artificial categories such as cloud water (cloud droplets) or rainwater (raindrops). The total moment tendencies are predicted 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 and validated using randomly generated data over a wide range of parameters employed by the parameterization. An analysis of the predicted total moment errors was performed, aimed to establish the accuracy of the parameterization at reproducing the integrated distribution moments representative of physical variables. The applied machine learning approach shows a good accuracy level when compared to the output of an explicit collision–coalescence model.