中国1千米分辨率逐日全天候地表土壤水分数据集（2003-2024）

(1、2025年5月19日对V2.0版本进行了最新更新，本次更新将数据集覆盖时段延伸至2024年年末。
2、2023年5月数据更新提示：本数据集的V2.0版本目前已经更新至2022年底，同时填补了2011年10月至2012年6月的空白时段，V2.0版本整体估算结果与V1.0原始版本相同，已下载的V1.0版本数据亦可放心使用，详情请参阅附件"2023年5月数据更新说明.pdf"。)
地表土壤水分（SSM）是了解地球表面水文过程的关键参数。长期以来，被动微波（PM）技术一直是在卫星遥感尺度上估算SSM的主要选择，而另一方面，PM观测的粗分辨率（通常>10 km）阻碍了其在更细尺度上的应用。虽然已经提出了定量研究，以缩小基于卫星PM的SSM的规模，但很少有产品可供公众使用，以满足1km分辨率和全天候条件下每日重访周期的要求。因此，在本研究中，我们在中国开发了一种具有所有这些特征的SSM产品。该产品是通过在36 km处对基于AMSR-E和AMSR-2的SSM进行降尺度生成的，涵盖了2003-2019年间两台辐射计的所有在轨时间。MODIS光学反射率数据和在多云条件下填补空白的每日热红外地表温度（LST）是降尺度模型的主要数据输入，以实现SSM降尺度结果的“全天候”质量。4月至9月期间，这一开发的SSM产品的每日图像在全国范围内实现了准完全覆盖。在其他月份，与最初的每日PM观测值相比，开发产品的全国覆盖率也大大提高。我们根据2000多个专业气象和土壤水分观测站的现场土壤水分测量结果对该产品进行了评估，发现该产品的精度在晴空到多云的所有天气条件下都是稳定的，无偏RMSE的站平均值在0.053 cm3/cm3到0.056 cm3/cm3之间。此外，评估结果还表明，开发的产品在1km分辨率下明显优于广为人知的SMAP Sentinel（主被动微波）组合SSM产品。这表明，我们开发的产品在改善未来水文过程、农业、水资源和环境管理相关调查方面可能带来的潜在重要效益。

1. The latest update to Version 2.0 was carried out on May 19, 2025, which extended the coverage period of this dataset to the end of 2024.
2. Note on the May 2023 data update: The V2.0 version of this dataset has been updated to cover up to the end of 2022, and the gap period from October 2011 to June 2012 has been filled. The overall estimation results of the V2.0 version are identical to those of the original V1.0 version, and users who have downloaded the V1.0 data can use it with confidence. For details, please refer to the attachment "2023年5月数据更新说明.pdf".

Surface Soil Moisture (SSM) is a critical parameter for understanding Earth's surface hydrological processes. For a long time, passive microwave (PM) technology has been the primary option for estimating SSM at satellite remote sensing scales. However, the coarse resolution (typically >10 km) of PM observations has hindered its application at finer spatial scales. Although quantitative studies have been proposed to downscale satellite PM-based SSM, few publicly available products meet the requirements of 1 km resolution, all-weather conditions, and daily revisit cycle. Therefore, in this study, we developed an SSM product with all these characteristics for China.

This product was generated by downscaling AMSR-E and AMSR-2 based SSM at 36 km resolution, covering all in-orbit operating periods of the two radiometers from 2003 to 2019. MODIS optical reflectance data and daily thermal infrared Land Surface Temperature (LST) that fills gaps under cloudy conditions serve as the main data inputs for the downscaling model, to achieve the "all-weather" quality of the SSM downscaling results. From April to September, the daily images of this developed SSM product achieved quasi-complete national coverage. In other months, the national coverage rate of the developed product has also been substantially improved compared to the original daily PM observations.

We evaluated this product using in-situ soil moisture measurements from over 2,000 professional meteorological and soil moisture observation stations, and found that its accuracy remains stable across all weather conditions ranging from clear skies to cloudy periods, with the station-averaged unbiased RMSE falling between 0.053 cm³/cm³ and 0.056 cm³/cm³. Additionally, the evaluation results demonstrate that the developed product significantly outperforms the well-known SMAP Sentinel (active-passive microwave combined) SSM product at 1 km resolution. This indicates that the developed product may bring potentially significant benefits for advancing future investigations related to hydrological processes, agriculture, water resources, and environmental management.

AI搜集汇总

背景与挑战

背景概述

该数据集是中国地区2003年至2024年的逐日地表土壤水分数据，具有1千米高空间分辨率和全天候覆盖特点，通过降尺度被动微波数据并结合MODIS光学反射率与热红外地表温度生成，有效填补了多云条件下的数据空白。数据精度稳定，经实地验证无偏RMSE在0.053-0.056 cm³/cm³之间，优于主流SMAP Sentinel产品，适用于水文、农业和环境管理等领域的研究。