In a numerical climate model, fine-scale processes are represented by parameterizations. A new method to develop parameterizations consists in using Artificial Intelligence (AI) techniques, by learning the representation of these processes from data coming from high resolution simulations. But a number of problems remain to be solved before the learned parameterizations can be used in a climate model. The objective of this thesis is to study some problems preventing the use of AI parameterizations in a climate model: numerical stability and online performance. These problems are studied through two studies on toy models. The results obtained are then placed in perspective, through the learning of a parameterization implemented in a climate ...
Abstract The fidelity of climate projections is often undermined by biases in climate models due to ...
Parameterization and parameter tuning are central aspects of climate modeling, and there is widespre...
The climate system can be regarded as a dynamic nonlinear system. Thus, traditional linear statistic...
In a numerical climate model, fine-scale processes are represented by parameterizations. A new metho...
Climate predictions and weather forecasting strongly rely on simulations of the Earth’s oceans and a...
An essential challenge for the climate science community is to provide trustful information about th...
International audience• We apply uncertainty quantification to single-column model/large-eddy simula...
A promising approach to improve climate-model simulations is to replace traditional subgrid paramete...
Weather and climate prediction is dominated by high dimensionality, interactions on many different s...
Development of neural network (NN) emulations for fast calculations of physical processes in numeric...
Climate modeling is a key tool for understanding the climate system and for making projections of it...
Climate is a complex system resulting from various interactions between its different components and...
Dynamical weather and climate prediction models underpin many studies of the Earth system and hold t...
When modeling complex phenomena in nature and in technological systems, one is often faced with the...
The representation of nonlinear subgrid processes, especially clouds, has been a major source of unc...
Abstract The fidelity of climate projections is often undermined by biases in climate models due to ...
Parameterization and parameter tuning are central aspects of climate modeling, and there is widespre...
The climate system can be regarded as a dynamic nonlinear system. Thus, traditional linear statistic...
In a numerical climate model, fine-scale processes are represented by parameterizations. A new metho...
Climate predictions and weather forecasting strongly rely on simulations of the Earth’s oceans and a...
An essential challenge for the climate science community is to provide trustful information about th...
International audience• We apply uncertainty quantification to single-column model/large-eddy simula...
A promising approach to improve climate-model simulations is to replace traditional subgrid paramete...
Weather and climate prediction is dominated by high dimensionality, interactions on many different s...
Development of neural network (NN) emulations for fast calculations of physical processes in numeric...
Climate modeling is a key tool for understanding the climate system and for making projections of it...
Climate is a complex system resulting from various interactions between its different components and...
Dynamical weather and climate prediction models underpin many studies of the Earth system and hold t...
When modeling complex phenomena in nature and in technological systems, one is often faced with the...
The representation of nonlinear subgrid processes, especially clouds, has been a major source of unc...
Abstract The fidelity of climate projections is often undermined by biases in climate models due to ...
Parameterization and parameter tuning are central aspects of climate modeling, and there is widespre...
The climate system can be regarded as a dynamic nonlinear system. Thus, traditional linear statistic...