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

Rodríguez Genó, Camilo Fernando; Alfonso, Léster

doi:https://doi.org/10.5194/gmd-15-493-2022

Articles | Volume 15, issue 2

https://doi.org/10.5194/gmd-15-493-2022

© Author(s) 2022. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/gmd-15-493-2022

© Author(s) 2022. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 15, issue 2

Model description paper

|

21 Jan 2022

Model description paper |

| 21 Jan 2022

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

Camilo Fernando Rodríguez Genó and Léster Alfonso

Model code and software

COLNETv1.0.0 Camilo Fernando Rodríguez-Genó and Lester Alfonso https://doi.org/10.5281/zenodo.4740061

COLNETv1.0.0 neural network training scripts Camilo Fernando Rodríguez-Genó and Lester Alfonso https://doi.org/10.5281/zenodo.4740129

COLNETv1.0.0 graphics generation scripts Camilo Fernando Rodríguez-Genó and Lester Alfonso https://doi.org/10.5281/zenodo.4740184

Program for the solution of the stochastic coalescence equation: One-dimensional cloud microphysics Andreas Bott https://doi.org/10.5281/zenodo.5660185

CP2000CCv1.0.0 WDM6 Parameterization Camilo Fernando Rodríguez-Genó and Lester Alfonso https://doi.org/10.5281/zenodo.5196706

Short summary

The representation of the collision–coalescence process in models of different scales has been a great source of uncertainty for many years. The aim of this paper is to show that machine learning techniques can be a useful tool in order to incorporate this process by emulating the explicit treatment of microphysics. Our results show that the machine learning parameterization mimics the evolution of actual droplet size distributions very well.