Resolving Low Cloud Feedbacks Globally with HR-MMF: Agreement with LES but Stronger Shortwave Effects

This study uses the Multiscale Modeling Framework (MMF) of the Energy Exascale Earth System Model (E3SM-MMF) to explore low cloud feedbacks. E3SM-MMF embeds a two-dimensional Cloud-Resolving Model (CRM) based on the System for Atmospheric Modeling (SAM) within each column of a host Global Climate Model (GCM). The CRM employs a one-moment microphysics scheme, Smagorinsky turbulence closure, and the Rapid Radiative Transfer Model for GCM Applications-Parallel (RRTMGP) for radiative transfer, optimized for GPU performance. CRM columns are grouped for radiative calculations to reduce computational costs without affecting solution quality. The simulations use an optimized E3SM-MMF variant with enhancements such as horizontal hyperviscosity, cloud droplet sedimentation, refined autoconversion thresholds for liquid (qcw0qcw0 = 5 × 10−4) and ice (qci0qci0 = 5 × 10−5), and non-local CRM variance transport. The CRM grid has 200 m horizontal resolution (Δx) with 256 columns (51.2 km domain), while the GCM uses a ``ne30pg2'' horizontal grid (~1.5∘ resolution) and a 125-level vertical grid (L125), with Δz = 20 m resolution between 500–1,800 m to resolve marine stratocumulus inversions. A 2 s CRM time step ensures numerical stability. Two 5-year simulations were conducted: a control (0 K) experiment with default NOAA SST and sea ice data, and a warmed (4 K) experiment with SST uniformly increased by 4 K. Both simulations used identical configurations, ensuring reasonable low cloud amounts in the control climate. The latest E3SM version is available at \url{https://github.com/E3SM-Project/E3SM.git}.本研究依托能源百亿亿次地球系统模型（Energy Exascale Earth System Model，E3SM）的多尺度建模框架（Multiscale Modeling Framework，MMF，即E3SM-MMF），开展低云反馈机制的探索研究。E3SM-MMF在宿主全球气候模型（Global Climate Model，GCM）的每一列网格中，嵌入了基于大气建模系统（System for Atmospheric Modeling，SAM）的二维云解析模型（Cloud-Resolving Model，CRM）。该云解析模型采用单矩微物理方案、斯马戈林斯基（Smagorinsky）湍流闭合方案，以及面向全球气候模型应用的快速辐射传输并行模型（Rapid Radiative Transfer Model for GCM Applications-Parallel，RRTMGP）进行辐射传输计算，并针对GPU性能进行了优化。为降低计算成本且不影响求解质量，研究中将云解析模型的网格列进行分组以开展辐射计算。 本次模拟采用经过优化的E3SM-MMF变体，其改进内容包括水平超粘性方案、云滴沉降过程、液态水（qcw0qcw0 = 5 × 10−4）与冰相（qci0qci0 = 5 × 10−5）的精细化自动转换阈值，以及非局地云解析模型方差输运方案。云解析模型的水平分辨率为200米（Δx），共包含256个网格列（模拟区域跨度51.2千米）；宿主全球气候模型则采用“ne30pg2”水平网格（分辨率约1.5°）与125层垂直网格（L125），在500–1800米高度区间内采用20米的垂直分辨率（Δz），以精准解析海洋层云逆温结构。模拟采用2秒的云解析模型时间步长以保障数值稳定性。 本研究共开展两组为期5年的模拟试验：一组为对照试验（0 K），采用默认的美国国家海洋和大气管理局（National Oceanic and Atmospheric Administration，NOAA）海表温度与海冰数据；另一组为增温试验（4 K），将海表温度均匀提升4 K。两组模拟的配置完全一致，确保对照气候下的低云量处于合理范围。E3SM的最新版本可通过以下链接获取：https://github.com/E3SM-Project/E3SM.git。