Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Paper content is a review on developed methods for soil moisture estimation with the use of surfial soil moisture by RS-data. Advantages and disadvantages for 4 categories of methods are described and discussed. First, the importance of RZSM with hybrid and assimilation methods is well described. Often statistical or physical-mechanistic approaches exist.

With empirical methods (2.1) often statistical inversion methods with surface parameters and spectral indices are used. Here we miss description of spatial accuracy and evaluation for different soil depths? In 2.2 semi-empirical methods as ExpF or SMAR-model are described, problem of adjustment T with Topt is named - the limitations for these methods for higher spatial discretization must be more considered. Soil-physical and or -hydraulical parameter are needed to use SMAR-model - what is the advantage of SMAR in relation to physically based hydrological catchment models with their soil mositure tool? (discuss this in discussion chapter).

In 2.3 use of data assimilation methods are described. EnKF is a most popular algorithm, but resampling methods can be problematic. Here no description of problems for spatial extrapolation possibilities and accuration for soil depths exist. In 2.3.2 physical model use is well described, where higher spatial resolution data from TIR and SAR-data can be used e.g. in SVAT-model. Here we miss discussion of studies, which have used SMAPL4 or SVAT or Hydrus-1D with TIR, SAR or LST data - how data confusion is done and describe short with lit. references these problems more in detail named in rows 463-468.

With table 1 in 2.4 a good overview (20 ML´s) on ML-methods is given; mostly RF and ANN are used with best accuracy of RF. Authors point out that different input variables give good results for RZSM estimation and extrapolation to unobserved regions is difficult - give an research example for this. - In 3.1 comparison of RZSM-methods is done with good overview with table 2. Advantages and disadvantages with lit. references are good discussed; also in 3.2 a good summary shows limitation of RS-based inversion methods for RZSM.  - Paper shows that different methods in relation to available ground data (e.g. SSM, land-soil parameter) can give good results - semi-empirical methods seems to have best accuracy.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Thanks to the authors for this comprehensive summary of the state of the art of the techniques to infer sub-surface soil moisture. The paper is well written and easy to read, with numerous citations. I think it will be very useful for the readers.

I have just a few minor comments, and the a few more on the English Language, that I list in the next box.

- Line 16. I would not say "Remote sensing technology...", but "Remote sensing techniques". Technologies refers to the instruments, techniques encompass the processing algorithms etc.

- Line 30: I would not say "state variable" but "essential climate variable" instead.

- Line 49. Please briefly mention how optical data can be used "to estimate soil moisture". For a novel reader it may be difficult, but actually the first soil moisture retrieval algorithms used VNIR and TIR data...

- Figure 1 does not covey any informaton that is not already given in the paragraph above.

- Eqn. (1) is useless: it does not convey any information, and it is already said in the text below. If the authors want to keep it, the "f" in line 117 should be in italics, as written in the equation.

- Lines 140-148: In several places it appears R:XX or R^2:XX. I would use either R or R^2, but I won't be mixing them. Also, I do not know if the authors meant ":" or "=".

- Line 153: This equation should not be inline, it deserves to be separated, and the SCA variable properly defined. Also, the pixel size must be explicited, because it will affect the calculation of angle beta. Also, iin the equation it is not clear it looks like an upper case i ("I"), but I wonder if it is a "ln" (i.e. natural logarithm). Please clarify.

- Line 162: Add reference to MFD algorithm.

- Line 2.1.2. Cross-correlation analysis.

In this section I ould suggest to include the work from Herbert et al. 2020 [https://www.mdpi.com/2072-4292/12/16/2614] in which a "Dynamic Time Warp" (DTW) technique to assess the temporal evolution of the subsurface soil moisture at different depths, including SMOS surface soil moisture estimates, in situ soil moisture data from the REMEDHUS network, rain rate etc.

- Section 2.1.3. Eqns. 2-4 are not clear. Use of \Delta and \Delta_p is confusing. It seems they should be the same. In iii) it is stated "the relationship between \Delta and theta_s is ..." and the eqn. (3) uses \delta_p (in italics).

Later in eqn. (4) \theta_CDF is the predicted surface SM, but \theta_s is the surface Soil Moisture (from which source?), and \delta_p is "the difference between SSM and RZSM". Can you please revise, and clarify?

- Lines 320 and 321. Are brackets used to indicate the units? If yes, then it is confusing, because brackets are also used in eqn. (8) and it is not clear to me what for (index of Z?). Please clarify.

- Lines 321 and 322: please write the "1" as subscripts, as they appear in eqn. (8).

Comments on the Quality of English Language

Please find below a few instances where I would writte it differently:

- Line 15: "methods to measure soil moisture..."

- Line 16: "Remote sensing technologies..."

- Line 20: "This paper analyzes and summarizes..."

- Line 24: "development of RZSM estimation methods methods is made"

- Line 30: "crucial Essential Climate Variable (ECV) that plays"

- Line 39: delete "and so on."

- Line 46: "observation techniques to monitor soil moisture"

- Line 49: "visible and infrared bands"

- Line 53: "is largely unaffected"

- Line 100: "The principle to estimate RZSM"

- Line 117: "RSindex (in italics) denotes a remote sensing index"

- Line 135: "compared to"

- Line 355: "methods to estimate"

- Line 425 : "Land Surface Models (LSMs) are widely"

- Line 534: "... are fully descussed."

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf