A daily high-resolution surface net radiation dataset in China (2000-2019)

Surface net radiation (Rn) characterizes the energy available at the Earth's surface and is essential for studying atmospheric, water, and carbon cycles. While some global-scale Rn products are available, they often suffer from issues such as data gaps, low resolution, and large uncertainties. Therefore, we developed a high-resolution (0.05°×0.05°) daily Rn dataset (named CHiRAD) from 2000 to 2019 in China using routine meteorological variables from more than 2400 stations and remotely sensed albedo. To ensure the reliability of the dataset, we tested a series of net shortwave and longwave algorithms using ground-based measurements and then employed the optimal combination of algorithms to generate this dataset. The dataset was validated against Rn observations from 43 flux towers across China. However, one may expect to use Rn data beyond this period in practice. To address this requirement, we also used AVHRR albedo to force the RI-PE algorithm and generated a long-series (1982–2020) Rn dataset (named Ext_CHiRAD). The only difference between the Ext_CHiRAD and CHiRAD, except for the length of the data series, is the source of albedo data. This dataset may not be as accurate as CHiRAD, but it has the advantage of a longer time span (1982–2020). This advantage makes it an ideal source of Rn data for hydrologic and environmental models to simulate long-term changes in target variables.地表净辐射（Surface net radiation, Rn）表征地球表面可获取的能量，是研究大气循环、水文循环与碳循环的关键要素。目前虽已有部分全球尺度的Rn产品问世，但这类产品普遍存在数据缺失、分辨率偏低以及不确定性较大等问题。为此，本研究基于全国2400余个常规气象站点的观测变量与遥感反演反照率（albedo），构建了2000年至2019年中国区域高分辨率（0.05°×0.05°）逐日地表净辐射数据集，命名为CHiRAD。为保障数据集的可靠性，研究团队先利用地面实测数据对一系列短波净辐射与长波净辐射算法开展对比测试，随后选取最优算法组合完成该数据集的生成。本数据集通过全国43座通量塔的Rn观测数据进行了精度验证。不过实际应用中，用户往往需要使用该时间范围以外的Rn数据。为满足这一需求，研究团队还利用AVHRR反照率驱动RI-PE算法，生成了时间跨度为1982–2020年的长时序地表净辐射数据集，命名为Ext_CHiRAD。除数据序列长度外，Ext_CHiRAD与CHiRAD的唯一差异在于反照率数据源的不同。尽管Ext_CHiRAD的精度或不及CHiRAD，但其具备更长的时间覆盖范围（1982–2020年）这一核心优势。该优势使其成为水文与环境模型模拟目标变量长期变化的理想Rn数据源。