Existing ML-based atmospheric models are not suitable for climate prediction,
which requires long-term stability and physical consistency. We present ACE
(AI2 Climate Emulator), a 200M-parameter, autoregressive machine learning
emulator of an existing comprehensive 100-km resolution global atmospheric
model. The formulation of ACE allows evaluation of physical laws such as the
conservation of mass and moisture. The emulator is stable for 10 years, nearly
conserves column moisture without explicit constraints and faithfully
reproduces the reference model's climate, outperforming a challenging baseline
on over 80% of tracked variables. ACE requires nearly 100x less wall clock time
and is 100x more energy efficient than the reference model using typically
available resources.

现有的基于机器学习的大气模型不适用于气候预测，我们提出了 ACE（AI2 Climate Emulator），一个 200M 参数的自回归机器学习模拟器，用于现有的包含 100km 分辨率全球大气模型。ACE 的构建允许评估质量和水分守恒等物理规律。该模拟器在 10 年内保持稳定，几乎在没有明确限制的情况下保持了纵向水分的守恒，并忠实地再现了参考模型的气候，相对于一个具有挑战性的基准模型，在超过 80% 的跟踪变量上表现更佳。ACE 所需的墙钟时间几乎比参考模型少 100 倍，并且使用常见资源效率更高 100 倍。