Improvement of a Neural Network Convection Scheme by Including Triggering and Evaluation in Present and Future Climates
Abstract
In this study, we improve a neural network (NN) parameterization of deep convection in the global atmosphere model ARP-GEM. To take into account the sporadic nature of convection, we develop a NN parameterization that includes a triggering mechanism that can detect whether deep convection is active or not within a grid-cell. This new data-driven parameterization outperforms the existing NN parameterization in present climate when replacing the original deep convection scheme of ARP-GEM. Online simulations with the NN parameterization run without stability issues. Then, this NN parameterization is evaluated online in a warmer climate. We confirm that using relative humidity instead of the specific total humidity as input for the NN (trained with present data) improves the performance and generalization in warmer climate. Finally, we perform the training of the NN parameterization with data from a warmer climate and this configuration get similar results when used in simulations in present or warmer climates.
Related
- Climatology and trends of extreme precipitation in France: evaluation of an explicit-convection regional climate model
- Including the triple isotopic composition of dissolved oxygen in the ocean into the iLOVECLIM model (version 1.1.7): development and evaluation
- Evaluation and improvement of CAMS-derived CCN number concentrations using in-situ measurements
- Matching Wind-Solar Portfolios with Electricity Consumption Across Climates: A Physical Framework to Minimize Residual Demand
- The diversity of planetary climates in the Solar System as a basis for reflection and work on the study of exoplanet climates
BibTeX
@article{germain2026improvement,
title = {{Improvement of a Neural Network Convection Scheme by Including Triggering and Evaluation in Present and Future Climates}},
author = {Hugo Germain and Blanka Balogh and Olivier Geoffroy and David Saint-Martin},
journal = {Artificial Intelligence for the Earth Systems},
year = {2026},
month = oct,
doi = {10.1175/aies-d-25-0108.1},
url = {https://meteofrance.hal.science/meteo-05840839v1},
note = {CC BY},
}
RIS
TY - JOUR
TI - Improvement of a Neural Network Convection Scheme by Including Triggering and Evaluation in Present and Future Climates
AU - Hugo Germain
AU - Blanka Balogh
AU - Olivier Geoffroy
AU - David Saint-Martin
PY - 2026
DA - 2026/10/27
JO - Artificial Intelligence for the Earth Systems
DO - 10.1175/aies-d-25-0108.1
UR - https://meteofrance.hal.science/meteo-05840839v1
AB - In this study, we improve a neural network (NN) parameterization of deep convection in the global atmosphere model ARP-GEM. To take into account the sporadic nature of convection, we develop a NN parameterization that includes a triggering mechanism that can detect whether deep convection is active or not within a grid-cell. This new data-driven parameterization outperforms the existing NN parameterization in present climate when replacing the original deep convection scheme of ARP-GEM. Online simulations with the NN parameterization run without stability issues. Then, this NN parameterization is evaluated online in a warmer climate. We confirm that using relative humidity instead of the specific total humidity as input for the NN (trained with present data) improves the performance and generalization in warmer climate. Finally, we perform the training of the NN parameterization with data from a warmer climate and this configuration get similar results when used in simulations in present or warmer climates.
SN - 2769-7525
ER -