The manuscript presents a soil moisture dataset for the contiguous United States with fine temporal and spatial resolution. The authors disaggregate the fine soil moisture data from a daily level to a 3-hour level. I also note that the spatial coverage of soil moisture data is complete. It seems to make the dataset more useful.

I have a concern. In section 4.1, the authors used the drought in four states as a case study to display the decline in soil moisture. However, since drought is always a slow process, it seems the data have potential to characterize changes in drought using the daily soil moisture data. I suggest that the authors supply fast-forming disasters as a case study further to amplify the importance of hourly soil moisture data.

There are also some minor suggestions as follows:

1). The units of RMSE, Bias, and ubRMSE in Tables 2-5 should be provided.

2). Section 2.2: The VIPSTF model includes two different versions. Which version was used in the manuscript? Please explain it.

3). Line 265: What is the size of the soil moisture dataset? What is the data format? I suggest the authors to provide more detailed information about the STF_SSM dataset.

4). Figure 3: Is the scene in the first column an average of all the intraday scenes? Please clarify the specific time for each scene.

5). Line 475: The potential for hourly soil moisture data applications needs to be further emphasized. That is the main driver of fine soil moisture dataset development.

6). The references listed also provide 1-km soil moisture data or downscaling methods, which may be helpful to your work. In addition, what the difference is between the proposed and listed methods.

https://doi.org/10.1016/j.jag.2023.103572

https://doi.org/10.1016/j.rse.2022.113334

https://doi.org/10.1016/j.rse.2024.114579

This manuscript produces a 3-hour, 1-km soil moisture (SM) dataset generated using a spatiotemporal fusion approach that integrates the 3-hour, 9-km SMAP L4 soil moisture product with the 1-day, 1-km Crop-CASMA soil moisture data. The resulting dataset is evaluated against in-situ SM observations. Overall, the manuscript is well-structured, and the generated SM dataset holds significant potential for the scientific community. However, several aspects require clarification and further discussion:

1. Validation Approach: The methodology leverages the higher-accuracy SMAP L4 product to capture temporal variations while using the lower-accuracy but higher-resolution Crop-CASMA data to retain spatial details. Consequently, the accuracy of the fused product should theoretically be higher than that of Crop-CASMA but lower than SMAP L4. 
2. Comparative Analysis: The authors compare their product only with SMAP L4 and Crop-CASMA but do not benchmark it against other 1-km resolution datasets or even higher-resolution (30-m) products (DOI: 10.1038/s41597-021-01050-2). Including such comparisons, or at least discussing them, would provide a more comprehensive evaluation of the dataset's performance.
3. Temporal Variability Discussion: It is recommended that the authors expand their discussion on the temporal variations of the generated SM dataset within a single day, in addition to the analysis presented in Figures 6 and 7. This would help highlight the advantages of the product in capturing sudden SM changes compared to daily-scale products.
4. Figure 3: The date and time of the SM data should be explicitly stated in the figure title for clarity.
5. Figure 8 / Table 4: It is recommended that the authors provide the number of validation sites corresponding to each land cover type, either in Figure 8 or Table 4, to enhance transparency in the validation process.
6. Figure 9: If feasible, the authors are encouraged to analyze and present the relationship between RMSE, slope, and altitude, as this could provide additional insights into the dataset’s accuracy under varying topographic conditions.

I see that the authors have addressed several major concerns raised by the other reviewers. The paper is already in good shape and suitable for the journal, with only a few minor comments from my side related to data operations:

The availability of an open-access soil moisture dataset with higher spatial and temporal resolution is desirable for both the Ag sector and the Earth system science community. My main concern is the long-term sustainability of the dataset. The current dataset covers the historical period from 2015 to 2023. Will data for the most recent years be made available in the future? The manuscript lacks a discussion on the sustainability and future availability of the dataset, which is crucial for readers who intend to reuse or build upon this work.

It is good to know the paper used soil moisture data products on Crop-CASMA as inputs and benchmark data. While Crop-CASMA is designed primarily to support USDA NASS operations and is produced operationally during the growing season, it is important to discuss how the new dataset can be used in more real-world applications (like near-real-time crop condition monitoring) beyond historical analysis. Meanwhile, I recommend including a data management plan and ensuring code availability in the manuscript, especially given its submission to a top data journal.