Prognostic Validation of a Neural Network Unified Physics Parameterization

TLDR

Weather and climate models approximate diabatic and sub‑grid processes with human‑designed parameterizations that are efficient but oversimplified, yet high‑resolution simulations now enable machine‑learning approaches. This letter trains a neural‑network parameterization on a near‑global 4‑km aqua‑planet simulation (NG‑Aqua). The network predicts heat and moisture source terms for 160‑km grid boxes and achieves numerical stability by minimizing multi‑step prediction error. Prognostic single‑column tests show the scheme reproduces NG‑Aqua fluctuations and equilibrium more accurately than the Community Atmosphere Model.

Abstract

Abstract Weather and climate models approximate diabatic and sub‐grid‐scale processes in terms of grid‐scale variables using parameterizations. Current parameterizations are designed by humans based on physical understanding, observations, and process modeling. As a result, they are numerically efficient and interpretable, but potentially oversimplified. However, the advent of global high‐resolution simulations and observations enables a more robust approach based on machine learning. In this letter, a neural network‐based parameterization is trained using a near‐global aqua‐planet simulation with a 4‐km resolution (NG‐Aqua). The neural network predicts the apparent sources of heat and moisture averaged onto (160 km) 2 grid boxes. A numerically stable scheme is obtained by minimizing the prediction error over multiple time steps rather than single one. In prognostic single‐column model tests, this scheme matches both the fluctuations and equilibrium of NG‐Aqua simulation better than the Community Atmosphere Model does.