Coarse-grained Near-global Aqua-planet Simulation with Computed Dynamical Tendencies

This dataset includes the coarse-grained 3D state of the near-global CRM simulations (NG-Aqua). The simulation is run at a 4km resolution using the System for Atmospheric Modeling (SAM) A dataset derived from the same simulation is included at the 10.5281/zenodo.1226370. This current posting supplements this dataset with the dynamical tendencies for total water and liquid-ice potential temperature, respectively given by FQT and FSLI. These are computed by initializing SAM run at a 160km resolution with the coarse-grained fields from NG-Aqua; evolving the state forward for 10 30 second time steps; saving the output; and finally computing the difference with the initial condition. This netCDF dataset is split into several part files for more robust uploading/downloading. To download this data, download each "part" file, and combine them with the "cat" linux command: cat noBlur.nc.part?? > noBlur.nc If using this with the uwnet code repository, you should then move this file to "data/processed/training/noBlur.nc", creating that folder if necessary.本数据集涵盖近全球云分辨模式（Cloud Resolving Model, CRM）模拟（NG-Aqua）的粗粒度三维状态场。本次模拟采用大气模拟系统（System for Atmospheric Modeling, SAM），空间分辨率为4千米。基于同一模拟衍生的另一数据集已上传至DOI：10.5281/zenodo.1226370。本数据集为前述数据集补充了总水量与液态冰位温的动力倾向，该倾向分别由FQT与FSLI表征。该动力倾向的计算流程如下：以NG-Aqua的粗粒度场作为初始场，在160千米分辨率下启动SAM模拟，将模拟状态向前积分10个时长为30秒的时间步，保存模拟输出结果，最后计算输出与初始状态的差值。本netCDF数据集被拆分为多个分卷文件，以提升上传与下载的稳定性。如需下载该数据，请下载全部分卷文件，并通过Linux系统的cat命令合并：cat noBlur.nc.part?? > noBlur.nc。若将该数据集与uwnet代码库配合使用，请将合并后的文件移动至"data/processed/training/noBlur.nc"路径，如需可提前创建对应目录。