Learning Prognostic Variables for AI Convective Parameterizations via Symbolic Distillation
Researchers learn symbolic prognostic variables so AI convective parameterizations retain memory and improve climate statistics.
Hybrid AI-physics climate models usually use diagnostic parameterizations that depend only on the current coarse state. The authors compress past information with an autoencoder, then replace it with symbolic equations—a forced multivariate linear ODE—that evolve latent memory variables alongside the resolved atmosphere. Tested online on Lorenz-96 and offline on high-resolution surface precipitation, the approach recovers most of the autoencoder’s gains and improves climate statistics, temporal structure, and the diurnal cycle of tropical land precipitation versus memory-free baselines.
- Autoencoder compresses past states into a low-dimensional latent memory
- Symbolic distillation replaces the encoder with a forced linear ODE
- Evaluated online on Lorenz-96 and offline on precipitation
- Improves climate statistics and the tropical land diurnal cycle
Full article205 words · extracted from arxiv.org · click to collapse
Hybrid AI-physics climate modeling aims to improve coarse (~100km-resolution) Earth system models by learning to parameterize subgrid processes from high-fidelity data. However, this so far mostly involves local-in-time, diagnostic parameterizations, in which the subgrid state depends only on the current coarse state with no memory of previous states, which is unrealistic for processes such as convection that have intrinsic persistence. To address this, we enhance local-in-time parameterizations by learning prognostic variables that compactly carry important, additional past information where no explicit sub-grid information is available. First we compress past information into a low-dimensional latent space using an autoencoder, which then informs a neural network trained to parameterize targeted subgrid-scale processes. We then replace the autoencoder with symbolic equations that govern the time evolution of the latent variables, yielding additional prognostic memory variables that can be integrated alongside the resolved atmospheric state. We evaluate this approach on two systems: the Lorenz-96 model (online) and surface precipitation from high-resolution atmospheric simulations (offline). A forced multivariate linear ordinary differential equation recovers most of the added value achieved by the autoencoder-based approach in both experiments. Benchmarked against diagnostic parameterizations without memory, our memory-informed approach improves climate statistics and temporal structure, including a realistic diurnal cycle of tropical land precipitation.
Text extracted automatically; images, tables and formatting may be missing. Original: https://arxiv.org/abs/2609.24882