Estimating Surface Soil Heat Flux in Permafrost Regions Using Remote Sensing-Based Models on the Northern Qinghai-Tibetan Plateau under Clear-Sky Conditions

Round 1

Reviewer 1 Report

The authors have improved both the presentation and the contents of the manuscript, and this work is highly appreciated.

The minor issues and concerns are the structure and the clarity of some explanations in the text. The limitations of this study should be explicitly discussed in detail in the introductory part of the text. The explanation concerning the negative G0 values should be placed earlier in the text, and should be much more detailed.

Some other minor comments are found in the attached file.

Comments for author File: Comments.pdf

Author Response

L51 Delete or displace.

Response:

Thanks for your comments. The sentence has been deleted in the revised manuscript.

Author Response File: Author Response.docx

Reviewer 2 Report

Dear authors, 

overall I believe this manuscript to be valuable and I think it should be published. I have several comments that I believe should absolutely be addressed before. I have listed them below:

Major Changes:

In the beginning it is not very clear what the purpose of your study is. Are you using in situ data to validate results obtained by using satellite observations or the other way around? The purpose becomes clearer later, but you have to revise your introduction and abstract in a way that clearly defines the purpose of your work. 

Line 62 (and other places in the manuscript): The authors integrate many satellite data sets. Also they mention soil moisture data would be useful but can not be directly obtained from optical satellite data. What about radar (scatterometer or SAR) soil moisture (e.g. ASCAT)? Is it not suitable for this study, why was it not included? Include or justify!

Line 109: Why was this site chosen? Justify the choice clearly. 

Figure 1: what is the background map. Give source clearly!

Line 151: vegetation reflectance? You always talk about NDVI in other parts of the manuscript. Specify. 

Methods section: This section should be overall made more clear. The way it is structured now is quite hard to read. 

Equation 8 + 9: These equations are not needed.

Results section: overall not very clear. Must be improved. For example row 255 - 257.

Figure 14: please put the legend somewhere else, for example below all three parts of the figure

Minor Changes: 

There are many equations. Are they really all necessary? please revise. 

Why are some parts of the text red?

Line 47-48: revise sentence to clarify

Line 94-98: this would fit better in the beginning of the introduction to set up the paper and make it more interesting for readers.

Table 1: Some lines are to close together.

Equation 13 + 14: to small

Figure 5 + 6: the x-achis, does it need to go until -100? there is not data there. 

Author Response

There are many equations. Are they really all necessary? please revise. 

Response:

Thanks for your comments. Formulas 8, 9 and 14 in the original manuscript have been deleted from the revised manuscript.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

In my opinion, the paper was significantly improved. The authors addressed my comments sufficiently. I appreciate the work especially in clarifying the text and improving the tables/figures.

I urge the editors to accept this manuscript, I recommend only minor revisions (see below).

Throughout the manuscript there are discrepancies in the abbreviations. For example Maimpr vs M_aimpr 

The conclusion section could be more convincing. 

For example I would suggest writing "....in permafrost regions and leads to a deeper understanding these areas."

And: "....multi-site observations would improve the understanding of energy flow processes..."

As for the conclusion, the abstract could also use some "strengthening".

Don't use abbreviations without explaining them first (for example "QTP")

Change "We validated several existing remote sensing-based models to estimate G0 by analyzing the in-situ measurement data." to "We validated several existing remote sensing-based models to estimate G0 by analyzing in-situ measurement data."

Change "When the error of the remote sensing measured  land surface temperature is less than 1 K and the surface albedo measured is less than 0.02, then the accuracy of estimates based on remote sensing data for G0 will be less than 5%."  to "When the error of the remotely sensed land surface temperature is less than 1 K and the surface albedo measured is less than 0.02, the accuracy of estimates based on remote sensing data for G0 will be less than 5%."

Change "This study enhances our understanding of the impacts of climate changes on the ground thermal regime of permafrost and the land surface processes between atmosphere and ground surface in cold regions." to "This study enhances our understanding of the impacts of climate change on the ground thermal regime of permafrost and the land surface processes between atmosphere and ground surface in cold regions."

The image quality of the graphical abstract could be improved. I seems distorted and low resolution. 

Author Response

Throughout the manuscript there are discrepancies in the abbreviations. For example Maimpr vs M_aimpr 

Response:

Thanks for your comments. This is our mistake. In the revised manuscript, the expression forms of the improved G₀ parameterization scheme are all Ma_impr, and the wrong expression M_aimpr is all modified to Ma_impr. See L399-400 of the revised manuscript for details：

Table 8. Average sensitivity of the G₀ parameterization scheme Ma_impr to land surface temperature T_s and land surface albedo α

Author Response File: Author Response.docx

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

The revised manuscript is well-organized and has better readability as compared to the previous version. However, there are several concerns I would like to raise here, some of them - repeatedly.

Firstly, the English language still lacks quality; I had corrected some major errors, but not all of them.

Secondly, at the first stage the G0 values are obtained using two different methods, plates and TDEC. These two methods are uncomparable in their quality, as discussed in the Introduction, but are nevertheless used further as a single dataset with uniform quality. No information is given on the performance of the two methods, e.g., at the same weather station (a control experiment). 

Thirdly, the origin and quality of the initial data are unclear in some important cases. E.g., in the Table 3, is the albedo measured or modeled  value? What is the source of the albedo data? Was it observed, or calculated using equation 11? Why the land surface emissivity is taken as 0.95 in the parametrization scheme tests (Line 176-177), but is calculated from narrowband emissivities (equation 13) on the G0 estimation step? Why not use consistent and uniform approach to all data throughout the verification, validation and estimation processes?

And lastly, my major scientific concern is the use of modeled data in the improvement of another model. This is scientifically incorrect. As is claerly stated in the manuscript, only one station had direct observations on the surface temperature Ts (SPAM) - Lines 171-173. In fact, for all other stations Ts is a modeled, synthetic value, and not a true data. Nonetheless, the flowchart (Figure 2) does not mention Ts calculation as a step in G0 optimization. This is ambiguous and misleading.

The purpose of the G0 parametrization is to estimate the G0 values from remote sensing. In this case, you need observed G0 values (famous "ground truth") in order to validate your parametrizations. Instead, you use yet another model to calculate G0 (equation 7), which uses (a) a G0/Rn ratio also obtained from remote sensing, and (b) an Rn value, which was only observed at two weather stations, NPAM and SPAM (see Table 1). Remote sensing data can not be validated by the remote sensing data.

Brief, the manuscript can not be accepted, since in its present form it is ambiguous and misleading in its most important part, concerning observed and remotely sensed data. The authors have to stay consistent in the description of the study setup, its methodology, they should not trick the audience by avoiding using - and showing - the ground truth.

Some comments to the manuscript are given in the comments in an attached file. The corrections in the file are not exhaustive, and they repeat - in part - my last review.

Comments for author File: Comments.pdf

Author Response

Firstly, the English language still lacks quality; I had corrected some major errors, but not all of them.

Response:

Thanks for your comments. We have polished the language of the manuscript by Letpub.

Author Response File: Author Response.docx

Reviewer 2 Report

see attached comments

Comments for author File: Comments.pdf

Author Response

L540 the ratio rs?

Response:

The ratio r_s is the ratio of soil heat flux to net radiation for bare soil. In order to maintain consistency with the original (Su, 2002), r_s were changed to Γ_s in the revised manuscript (L532-539)：

The ratio Γ_s of G₀ to R_n in SEBS is 0.315 in bare soil [33]. Because the underlying surface of ZNH is bare soil (Fig. A2) and the vegetation of AYK is very sparse, the two sites are treated as bare soil. In the study, Γ_s on the northern Qinghai-Tibetan Plateau can be obtained by linear fitting of the data from these two stations and takes the value of 0.25 (Fig.A3). And the scheme Moran_adj can also prove this point. In the case of bare soil, the NDVI value is 0 and it is substituted into Moran_adj (Eq.8A) to obtain The ratio Γ_s of G₀ to R_n (0.237), which is quite different from 0.315, but close to 0.25. It can be seen that Γ_s in the scheme SEBS (0.315) is overestimated on the plateau. Replace 0.315 with 0.25 to get SEBS_adj

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

In this revision round, I would raise an issue of the incapability of the model to reproduce negative G0 values, that is clearly present in the Figures 5, 6, 13 and B1. It is worth noting, importantly, that both Ma and Ma-impr schemes fail to regress below zero (Figure 13), but Ma-impr even less than Ma. It is worth verifying then, how do other parametrization schemes reproduce negative values! Maybe they do better from this point, and other parametrization should be treated as optimal. The manuscript should provide a discussion on this topic.