提出物理一致性评估框架，检验机器学习天气模型是否符合自然规律。
PhysMetrics.Weather: An Evaluation Framework for Physical Consistency in ML Weather Models
- 设计三类物理指标：守恒性、谱特性、动力学行为，评估模型合理性。
- 通过定量分析发现现有模型在能量守恒等关键物理量上存在偏差。
- 适合开发物理引导的气象模型或需高可靠性预测的研究者使用。
机器学习天气预测（MLWP）模型在计算成本远低于传统物理方法的情况下，实现了令人瞩目的预报性能。然而，这些模型主要依赖数据驱动，且仅用像素级误差指标（如均方根误差）评估，无法保证其预报结果符合已知物理定律。本文提出 PhysMetrics.Weather，一个评估 MLWP 模型物理真实性的框架，涵盖三类指标：守恒性、谱特性和动力学行为。该框架通过量化物理一致性，指导物理信息架构的设计，并帮助判断 MLWP 模型是否适用于实际业务场景。相关代码已开源，地址为 https://github.com/Emmakast/PhysMetrics.Weather。
原文摘要 · Abstract (English)
Machine learning weather prediction (MLWP) models have achieved impressive forecasting performance at a small fraction of the computational costs required for traditional physics-based methods. However, they are primarily (1) data-driven and (2) evaluated using pixel-wide error metrics (e.g., RMSE), so there are no guarantees that their forecasts are consistent with known physical laws. We introduce PhysMetrics$.$Weather, an evaluation framework that assesses the physical realism of MLWP models across three types of metrics: conservation, spectral, and dynamical. By quantifying physical realism, this tool guides the development of physics-informed architectures and helps evaluate whether MLWP models are reliable for operational use. Our framework is available on Github at https://github.com/Emmakast/PhysMetrics.Weather.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。