The Long-term, High-accuracy and Seamless Soil Moisture (LHS-SM) dataset over the Qinghai-Tibet Plateau: part 2 (2011-2020)

Soil moisture (SM) is a vital variable in the water-energy cycle and characterizing its spatiotemporal dynamics is crucial for understanding the impacts of climate change. Although substantial efforts have been devoted to derive SM data at fine scale, there is still a research gap in obtaining the long-term, high-accuracy and high-resolution SM data over the Qinghai-Tibet Plateau (QTP) due to its complex topography. Therefore, this study generated the long-term, high-accuracy and seamless soil moisture (LHS-SM) dataset over the QTP during 2001-2020 using a two-step downscaling method. First the daily SM data from the Climate Change Initiative program of the European Space Agency (ESA CCI) was downscaled to 1km utilizing five machine learning approaches. Then a dynamic data merging method that considers the spatiotemporal nonstationary error was applied to derive the final LHS-SM data. Results indicated that LHS-SM data exhibited satisfying accuracy (mean R = 0.55, ubRMSE = 0.049 m³/m³) and certain improvement to the ESA CCI SM data both at station and network scales. The dataset can be used for various regional hydrology, meteorology, ecological analysis and modeling.土壤湿度（Soil moisture, SM）是水-能量循环中的关键变量，解析其时空动态特征，对于理解气候变化的影响至关重要。尽管学界已开展大量工作以获取精细尺度的SM数据，但受复杂地形制约，青藏高原（Qinghai-Tibet Plateau, QTP）区域仍存在长时序、高精度、高分辨率SM数据集的研究空白。为此，本研究采用两步降尺度方法，构建了2001-2020年青藏高原区域的长时序、高精度、无缝隙土壤湿度（Long-term high-accuracy and seamless soil moisture, LHS-SM）数据集。首先，本研究采用五种机器学习方法，将欧洲空间局气候变化倡议项目（European Space Agency Climate Change Initiative, ESA CCI）的逐日SM数据降尺度至1公里分辨率。随后，本研究引入考虑时空非平稳误差的动态数据融合方法，生成最终的LHS-SM数据集。验证结果表明，LHS-SM数据集精度表现优异（平均相关系数R=0.55，无偏根均方误差ubRMSE=0.049 m³/m³），且在站点与网络尺度上均较ESA CCI SM原始数据集有所提升。该数据集可广泛应用于区域水文、气象、生态相关的分析与模拟研究。