The 43-year global 9km remotely sensed soil moisture product is estimated by the fusion of two kinds of microwave soil moisture products using the spatial temporal fusion model (STFM). One product is the Climate Change Initiative (CCI) 0.25° passive soil moisture product in version 6.1. European Space Agency (ESA) integrates multi-source passive microwave observation data from 1978 to 2020 for CCI 0.25° passive soil moisture estimation. Another product is the Soil moisture Active and Passive (SMAP) 9km soil moisture product in version 3. The SMAP 9km data is less than three months (from April 13–July 7) as the failure of SMAP radar. The soil moisture STFM takes the known CCI 0.25°data and SMAP 9km data at the same date as the reference, and then to fuse other date CCI soil moisture for the unknown 9km soil moisture estimation at the date. The estimated 9km soil moisture covers from 1978 to 2020 in global scale, one image per day, 15402 in total and the data volume up to 13.6 G. The estimated long time series 9km soil moisture will play an important role in the researches and applications at regional scale.

1. The 9km soil moisture was estiamted by dwonscaling of the European Space Agency Climate Change Initiative (CCI) data with the assistance of the Soil Moisture Active and Passvie (SMAP) data,2. The unit of the 9km soil moisture is in volumetric (m3/m3).3. The data is in integer format and the valid range is from 1 to 10000, null or invalid values are represented by 0. To obatin the true soil moisture, the data should be multiplied by 0.0001.4. This data inherits the spatial coverage of CCI data. Therefore, there is a certain null value region in the spatial distribution of the data.5. Based on the theoretical basis of spatiotemporal fusion model,the data can be taken as the SMAP-like soil moisture as the similar spatial distribution.6. By evaluation against in-situ data from ISMN(International Soil Moisture Network), it shows that the accuary of the estimated 9km soil moisture is comparable with the accucary of CCI and has slightly better temporal correlation and unbiased root mean square error.