Preprints
https://doi.org/10.5194/gmd-2021-125
https://doi.org/10.5194/gmd-2021-125

Submitted as: model description paper 21 May 2021

Submitted as: model description paper | 21 May 2021

Review status: this preprint is currently under review for the journal GMD.

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

Camilo Fernando Rodríguez-Genó1 and Léster Alfonso2 Camilo Fernando Rodríguez-Genó and Léster Alfonso
  • 1Atmospheric Sciences Centre, National Autonomous University of Mexico, Mexico City, 04510, Mexico
  • 2Autonomous University of Mexico City, Mexico City, 09790, Mexico

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 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.

Camilo Fernando Rodríguez-Genó and Léster Alfonso

Status: open (until 16 Jul 2021)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse

Camilo Fernando Rodríguez-Genó and Léster Alfonso

Model code and software

Program for the solution of the stochastic coalescence equation: One-dimensional cloud microphysics Andreas Bott https://www2.meteo.uni-bonn.de/forschung/gruppen/tgwww/people/abott/fortran/coad1d.f

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

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

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

Camilo Fernando Rodríguez-Genó and Léster Alfonso

Viewed

Total article views: 168 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
130 35 3 168 3 1
  • HTML: 130
  • PDF: 35
  • XML: 3
  • Total: 168
  • BibTeX: 3
  • EndNote: 1
Views and downloads (calculated since 21 May 2021)
Cumulative views and downloads (calculated since 21 May 2021)

Viewed (geographical distribution)

Total article views: 157 (including HTML, PDF, and XML) Thereof 157 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 15 Jun 2021
Download
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 very well the evolution of actual droplet size distributions.