地球系统模型参数优化的物理-机器学习多保真度策略：准地转概念验证
A Physics--ML Multi-Fidelity Strategy for Earth System Model Parameter Optimization: A QG Proof-of-Concept
浏览论文内容
中文总结 AI 辅助
提出一种结合格林函数优化与代理模型的多保真度框架，用于地球系统模型参数优化，在准地转模型中实现65%左右的改进，显著提升样本效率。
中文摘要 AI 辅助
地球系统模型依赖于可调的子网格尺度参数化方案，但优化这些参数的计算成本很高，尤其是当非线性相互作用需要大量模拟时。我们提出了一种混合的物理-机器学习多保真度框架，该框架将格林函数优化（GFO）与高斯过程或神经网络代理优化相结合。使用准地转湍流模型，GFO首先在归一化坐标中对参数敏感性进行排序，并选择缩减的活跃子集。然后，非线性代理使用廉价的30天模拟来探索该子集，之后再用180天的模拟对有希望的候选方案进行细化。在七种策略和35个成员的集合中，GFO-MultiGP和GFO-MultiNN分别实现了64.6%和65.2%的平均改进，并在3,060和2,520模拟天后达到实际饱和。相应的独立GP和NN实现了61.1%和42.0%的改进，并需要7,740和6,660模拟天。这些结果表明，所测试的混合流程具有端到端的样本效率优势。由于筛选、降维、初始化和保真度调度同时发生变化，它们各自的贡献未被单独分离。
英文摘要
Earth System Models rely on tunable subgrid-scale parameterizations, but optimizing these parameters is computationally expensive, particularly when nonlinear interactions require many simulations. We present a hybrid Physics-ML multi-fidelity framework that combines Green's Function Optimization (GFO) with Gaussian Process or Neural Network surrogate optimization. Using a quasi-geostrophic turbulence model, GFO first ranks parameter sensitivities in normalized coordinates and selects a reduced active subset. Nonlinear surrogates then explore this subset using inexpensive 30-day simulations before refining promising candidates with 180-day simulations. Across seven strategies and a 35-member ensemble, GFO-MultiGP and GFO-MultiNN achieved mean improvements of 64.6 percent and 65.2 percent, respectively, while reaching practical saturation after 3,060 and 2,520 simulation-days. The corresponding standalone GP and NN achieved 61.1 percent and 42.0 percent improvements and required 7,740 and 6,660 simulation-days. These results demonstrate an end-to-end sample-efficiency advantage for the tested hybrid pipelines. Because screening, dimensionality reduction, initialization, and fidelity scheduling change simultaneously, their individual contributions are not isolated.
发表机构
- Global Modeling and Assimilation Office, NASA Goddard Space Flight Center（全球建模与同化办公室，NASA戈达德太空飞行中心）
- Department of Earth, Environmental and Atmospheric Sciences, Western Kentucky University（西肯塔基大学地球、环境与大气科学系）
机构由 AI 辅助整理，请以论文原文为准。
