The authors attempted to evaluate the predictive performance of eight satellite data-derived root-zone soil moisture data. The authors provided a quantitative statistical analysis of each RZSM product based on average values of in-situ, and remote sensing estimates for the Huai River Basin. The authors concluded that the GLDAS CLSM RZMS products outperform other RZSM products. While the information contained in the manuscript is quantitative, I ended up questioning the potential contribution of this paper to the readers. I have tried to address why I think 

- Major comments:

1. Although the study provides a significant amount of comparative analysis between remote sensing-derived RZSM products as well as against observational data, the mechanistic understanding and explanation of each result are widely missing across the manuscript. Therefore, the author's rationale about the causes of differences was often too obvious or uncertain.
2. Also, the authors argued that GLDAS CLSM-derived RZSM outperforms other RZSM estimates. However, rather than trying to explain the different predictive performances of each RZSM product (against in-situ observations) in relation to soil properties, land cover/use, and vegetation in the study catchment, the authors just used the average of 58 in-situ data as well as satellite products, resulting in 'all-lumped' single time series for each dataset. Thus, it is not convincible to say a certain remote sensing-derived estimates outperform others since the performance differences can be revealed differently depending on soil properties, vegetation, land cover/use, etc.
3. It is also not indicated how each satellite-based soil moisture (at multiple depths) and RZSM 'with different spatiotemporal resolutions were aggregated (again, spatially and temporally) to come up with the sets of time series that require consistent temporal scales between them. The method used for spatial aggregation of the gridded-RZSM also needs to be manifested (i.e., methods).
4. As this is site-specific, it sounds even less convincing that CLSM-derived soil moisture products outperform, and thus it gets more confusing what the authors want to argue from the RZSM products comparison. 
5. There is significant inconsistency (due to the randomness in estimating RZSM from the remote sensing data) between RZSM estimation methods. For example, the authors tried to estimate RZSM using a depth-weighted method, but equation 1 used for in-situ RZSM is different from equation 2, which was used for RZSM estimation from satellite-derived modeled soil moisture. 

-Specific comments:

line 39-40: is this sentence needed?

line 42: duplicate definition of RZSM?

line 99-100: by this sentence, do you intend not to include any process-based explanation for the soil moisture products? What about attempting to explain the performance differences found among the RZSM products (as this is essentially modeled data) in relation to model structure? Why does CLSM outperform other land models in terms of RZSM products? 

Chapter 2.4: the information on the spatial and temporal resolution of each data needs to be revisited and clearly indicated.

Chapter 3.2: why did you estimate satellite-derived RZSM different from in-situ RZSM? Why equation 1 and 2 are different? How convincing are the RZSM comparisons based on equation 1 and 2?

line 293: instead of averaging all in-situ stations, can you think of disaggregating the study basin (and stations) using any available information such as surface soil properties, orography (e.g., slope, and elevation), land cover, and/or vegetation? That will help the readers get more generalizable information and references.

line 306-309: This needs to be rephrased. It is hard to understand what is meant.

line 311: can you explain why SMOS L4 showed less rapid changes and smoother trends?

line 321: can you explain why they did a better job in the wet season compared to the dry season?

line 360: can you explain why individual satellite-based RZSM products showed different probabilistic distributions? Some are log-normal and the others are normal. Can you add more explanation on this matter?

line 375: how does this ground-based observation of precipitation (840 mm/year) represent the average precipitation of the basin area? You also compared gridded-precipitation with this in-situ precipitation observation (line 430). Can you clarify how solid the comparison of this in-situ precipitation with gridded precipitation is?

line 432: do you think MERRA-2 and GLDAS-CLSM would outperform other satellite-derived RZSM in other basins (or area) as well? What if you perform a continental-scale study, will you still think there will be a certain winner? If not, how can you limit the scale of this sort of comparison study to be meaningful and convincing?  

line 436-438: the sentences need to be re-structured to clarify the argument.

line 453: in-situ RZSM observation does not capture irrigation effect? Can you explain how the irrigation water supply does not impact the soil moisture content?

line 485-489: can you add more information on how the soil properties could end up in certain ranges of soil moisture values?

See supplement