AEGIS: Differentiable Mars Climate Model with Neural Closures

2026-10-02

AI Summary

This study addresses the limitations of existing Mars climate models, which rely on legacy code lacking gradient support and thereby hinder parameter calibration and machine learning-based optimization. To overcome these constraints, this work introduces differentiable programming into Martian atmospheric simulation for the first time, constructing a modular differentiable climate model that couples the Dinosaur dynamical core with a neural closure interface. This architecture enables end-to-end physical calibration and neural network training, transcending the capabilities of traditional general circulation models (GCMs). Experimental results demonstrate that the proposed model successfully reproduces decadal-scale seasonal CO2 cycles while accurately capturing surface temperature structures and pressure distributions. Furthermore, the accuracy and computational efficiency of the derived gradients are rigorously validated, establishing a promising foundation for physics-informed machine learning in planetary climate modeling.

Abstract

General circulation models (GCMs) are the primary tool for simulating planetary atmospheres. They play a vital role in understanding Mars's atmosphere, as forecasting its unique weather is mission-critical for operations such as entry, descent, and landing. Mars poses unusual challenges for these models, as observations are sparse compared to Earth. In addition, a thin \co{} atmosphere alongside a radiatively active dust cycle creates a volatile atmosphere with large diurnal temperature swings and no true terrestrial analog for validation. Existing Mars GCMs, including the LMD PCM, the NASA Ames Mars GCM, and PlanetWRF, are mature and physically detailed but are implemented in legacy Fortran with finite-difference or finite-volume solvers, and they do not expose gradients for calibration or machine learning. Here we present AEGIS, a modular differentiable Mars climate model that couples Mars's unique atmospheric physics to the Dinosaur dynamical core, with interfaces for neural closures. We showcase stable ten-Mars-year simulations that reproduce the seasonal \co{} cycle while conserving the total \co{} inventory, capture realistic large-scale surface-temperature structure, and produce surface pressure that follows Mars Orbiter Laser Altimeter (MOLA) topography. Gradients through coupled trajectories agree with finite differences and support physical calibration and neural training. We compare with conventional GCMs, highlighting the framework's computational efficiency and differentiability.

Problem

• Mars climate model
• General circulation models
• Differentiability
• Neural closures
• Model calibration

Innovation

• Differentiable climate model
• Neural closures
• Mars general circulation model
• Dynamical core coupling
• Gradient-based calibration