# Machine Learning Emulation of 3D Cloud Radiative Effects

arXiv: https://arxiv.org/abs/2103.11919 · HTML: https://mlsanity.com/paper/2103.11919/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 5 of its 55 references that are in the map.
- First posted on arXiv 22 Mar 2021.

## Builds on (most cited first, 5 of 5)

- [TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems](https://mlsanity.com/paper/1603.04467/) (2016)
- [Deep learning to represent sub-grid processes in climate models](https://mlsanity.com/paper/1806.04731/) (2018)
- [Applying machine learning to improve simulations of a chaotic dynamical system using empirical error correction](https://mlsanity.com/paper/1904.10904/) (2019)
- [Copula-based synthetic data augmentation for machine-learning emulators](https://mlsanity.com/paper/2012.09037/) (2020)
- [Predicting atmospheric optical properties for radiative transfer computations using neural networks](https://mlsanity.com/paper/2005.02265/) (2020)

## Built on by (most cited first, 5 of 5)

- [Forecasting Global Weather with Graph Neural Networks](https://mlsanity.com/paper/2202.07575/) (2022)
- [A posteriori learning for quasi-geostrophic turbulence parametrization](https://mlsanity.com/paper/2204.03911/) (2022)
- [ClimART: A Benchmark Dataset for Emulating Atmospheric Radiative Transfer in Weather and Climate Models](https://mlsanity.com/paper/2111.14671/) (2021)
- [Calibrated Physics-Informed Uncertainty Quantification](https://mlsanity.com/paper/2502.04406/) (2025)
- [Machine Learning Emulation of Urban Land Surface Processes](https://mlsanity.com/paper/2112.11429/) (2021)

## Abstract (arXiv)

The treatment of cloud structure in numerical weather and climate models is often greatly simplified to make them computationally affordable. Here we propose to correct the European Centre for Medium-Range Weather Forecasts 1D radiation scheme ecRad for 3D cloud effects using computationally cheap neural networks. 3D cloud effects are learned as the difference between ecRad's fast 1D Tripleclouds solver that neglects them and its 3D SPARTACUS (SPeedy Algorithm for Radiative TrAnsfer through CloUd Sides) solver that includes them but is about five times more computationally expensive. With typical errors between 20 % and 30 % of the 3D signal, neural networks improve Tripleclouds' accuracy for about 1 % increase in runtime. Thus, rather than emulating the whole of SPARTACUS, we keep Tripleclouds unchanged for cloud-free parts of the atmosphere and 3D-correct it elsewhere. The focus on the comparably small 3D correction instead of the entire signal allows us to improve predictions significantly if we assume a similar signal-to-noise ratio for both.