未明确命名

本研究构建了一个新的气象场降尺度基准和数据集，用于从低分辨率气象场中获取任意散点站尺度的气象状态。数据集主要包含ERA5再分析数据作为气象场数据，以及Himawari-8卫星的L1级网格数据和Weather2K数据集的站点观测数据作为观测数据。该数据集的目标是通过整合多源、多尺度、多模态的观测数据，实现对低分辨率气象场的降尺度，以达到任意散点站尺度的连续气象场建模。应用领域主要集中在提高天气预报的精度和分辨率，特别是在风速和表面压力等变量的预测上。

This study develops a novel benchmark dataset for meteorological field downscaling, which is designed to derive meteorological states at arbitrary scatter station scales from low-resolution meteorological fields. This dataset primarily includes ERA5 reanalysis data as meteorological field data, as well as L1-level grid data from the Himawari-8 satellite and in-situ station observation data from the Weather2K dataset as observational data. The objective of this dataset is to integrate multi-source, multi-scale, and multi-modal observational data to accomplish downscaling of low-resolution meteorological fields, thereby enabling continuous meteorological field modeling at arbitrary scatter station scales. Its application scenarios mainly focus on enhancing the accuracy and resolution of weather forecasts, particularly for the prediction of variables such as wind speed and surface pressure.