Improving the Accuracy of TanDEM-X Digital Elevation Model Using Least Squares Collocation Method

Round 1

Reviewer 1 Report

Dear authors, please address the following findings.

1. I do not see a list of existing studies addressing a similar issue and probably, it's better to use a separate sub-section documenting similar studies and their pitfalls.  then the next section could be the contributions sub-section where authors can height lights their contributions in this regard.

2. Figure 1. Please provide satellite imagery for Pannel (d) instead of land use and that would make more generalized sense of the study site.

3. Datasets, I would recommend using a reference DEM from ALS systems as well for ground truthing the whole thing. probably that can be used to quantify the overall accuracy of ICESat-2 as well.

4. Line 123_ 129, within the main body text, Fig.2 can be referenced here as you are explaining the working principle of the SAR system.

5. ICESat-2 is another major data source that has been used in this study, but the explanation is not too details. There are no references related to ICESat-2 data and its explanation. I recommend taking some material from this paper to highlight some aspects of ICESat-2 data as the authors used it for DEM assessments and its relevance to this study.

Remote Sensing | Free Full-Text | NASA ICESat-2: Space-Borne LiDAR for Geological Education and Field Mapping of Aeolian Sand Dune Environments (mdpi.com)

3.2. Error analysis,

please explain these terms briefly systematic error and random errors, better to depict them with some drawings as it is useful information to understand the problem and then see how you corrected these errors.

Figure 3. Please make a third panel by differencing the a and b to see the differences otherwise it is not visible.

Incorporate an ALS-DEM in the study as a standard benchmark to check the quality of generated DEM as ICESat-2 cannot provide a complete DEM.

The rest of the article seems fine to me.

 

Good luck

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The subject matter of the manuscript "TanDEM-X Digital Elevation Model Correction Using Least Squares Collocation Theory" is very interesting. The work is well structured and the technical language adopted is quite appropriate and scientifically valid. Topic is covered in an exhaustive way, making use of a large and appropriate bibliography. Moreover, the study provides interesting insights for future developments. The paper can be considered as worthy of scientific publication. However, the following minor improvements have to be made: 

 

ABSTRACT: 

 1. According to the template, the maximum number of words is 200 (The authors have written 288 words in this single paragraph). 

2.  Write BPNN in full “Back Propagation in Neural Networks” from the outset; thereafter only the acronym may be used.

3. Leave a space between the numerical value and the unit of measurement.

 

INTRODUCTION:

1. Pay attention to repeated words in the text. Remove 'currently' in this sentence:

(“However, the currently commonly used TanDEM-X DEM is currently limited to publicly…”).

 

SECTION 2:

1. In the figures the text written in the axes and the table of contents should be enlarged (font similar to that used in the text) (i.e Figure 1 a,b,c,d). 

2. Leave space between paragraphs (i.e between 2.2.1 and 2.2.2).

3. Write the full acronym NSIDC DAAC.

 

SECTION 3: 

1. In Figure 2 the authors have left out S1 and S2 to indicate the position of the two SAR satellites. 

2. In the figure 4 the value number written in the axes should be enlarged; they are too small and slightly blurred. 

 

SECTION 5: 

1. Why did you decide to group the two formulas on ME and RMSE with a single number (15)? It is usually preferable to number each formula with different numbers and mention them in the text. 

2. Fix punctuation, i.e. remove the dot before the reference [35]; “…of TanDEM-X DEM precision assessment. [35]. However, the grid points exhibiting”. 

3. The Figure 8a is not mentioned in the text; moreover it is preferable to modify graphs 8a and 8b (i.e. the format of the numbers should be enlarged). Do it for all figures where this problem occurs. 

4. Figure 9 is not mentioned in the text; the authors probably wrote Figure 6 instead of Figure 9. 

5. Table 1: the format is not respected regarding the upper and lower edges.

Comments for author File: Comments.pdf

The English language used in the paper is sufficiently acceptable but should be ameliorated. In some parts the manuscript can be simplified for ease of reading. Here below authors can find some suggestions for improvements prior to publication:

- The Abstract contains the item, the aim and the results but its first part should be adjusted to be more linear by showing a clearer order for the item and aim;

- Pay attention to the appropriate insertion of articles (the, a, an); sometimes they are lacking (i.e On the other hand, SRTM DEM and TanDEM-X DEM generated based on  Interferometric Synthetic Aperture Radar (InSAR) technology are more prone to larger  terrain inclinations' influence); 

- Some sentences can be better formulated and repetition of vocabulary should be avoided (see for example "However, the currently commonly used TanDEM-X DEM is currently limited to publicly available 90m resolution data released by the DLR [17], which directly impacts the potential application scope of TanDEM-X DEM data".); 

 

- Authors should check verb tenses as some errors can be found (i.e This approach utilizing the variance of external auxiliary DEMs as a reference for identifying abnormal points in lifting rail DEMs and obtaining the final DEM through weighted fusion of coherence coefficient of interference pairs and vertical baseline height);

- It is preferable not to use expressions like "we" and "us". Please, change the sentences with them (i.e. In Section 4, we construct a least squares collocation-based error correction model for TanDEM-X DEM.);

- For the description of the results obtained reported in Section 5, authors should use the simple past tense rather than the simple present. This is valid also for the Conclusion section. 

The authors can examine the following sentences listed below for correction:

 

Abstract

- The TanDEM-X DEM generated by InDAR technology is frequengly utilized as fundamental data for geological analyses, 

- The findings indicate that the application of this method has resulted in a significant reduction 

-...... revealing that the accuracy of TanDEM- X DEM corrected using

the algorithm proposed in this paper has been enhanced by 6.5%, 7.6%, and 18.1% respectively.  

  Introduction

-Consequently, an increasing number of global digital elevations with varying scales, resolutions, and accuracies are being publicly released.

- This solution utilizes a backpropagation algorithm for neural network training and minimizes the error between the target layer and output layer by employing mean square error as the cost function

- The Least Squares Collocation (LSC) in Generalized Measurement Adjustment can simultaneously processing the random signals and propensity parameters of observation objects 

- This method was originally developed for gravity anomaly analysis [23], and has since been widely utilized in determining quasi geoid 

 

Section 2

- Furthermore, the study area displays considerable slope variability as depicted in Figure 1c's terrain slope analysis. 

- In this study, we utilized the ATL08 dataset which is a version 5 of ATLAS/ICESat-2 L3A Land and Vegetation Height product. 

 

Section 3

- For the TanDEM-X DEM that has been generated, errors primarily consist of systematic and random components.

- To mitigate the impact of terrain on unwrapping, an external DEM is commonly employed and differential phase unwrapping technique is utilized to recover residual terrain information 

- After correcting for the orbital and terrain phase errors that affect phase unwrapping in TanDEM-X DEM, it was determined that the system error is primarily influenced by terrain features such as slope, aspect, and elevation. 

- To investigate the relationship beween elevation and TanDEM-X DEM error, we divided the error into four groups

- After conducting the aforementioned analysis, it has been determined that the error present in TanDEM-X DEM is primarily associated with elevation and terrain slope.

  - Consequently, a systematic error model for TanDEM-X DEM has been established as follows

 

Section 4

- The former is those that coincide with ICESat-2 points, while the latter do not. 

- Only by ensuring that the estimated field is a stationary random field and removing any trend, can the covariance function accurately reflect the 

 - Simultaneously, by subtracting the value of the system error trend surface from ∆? at the coincidence point, one can obtain an error value that removes the trend system error at said point this is known as initial random error value (signal) at said coincidence point.

 

Section 5

- Following correction by the LSC-TXC algorithm presented in this paper, these data gaps were filled 

- However, the correction effect of random errors is not statistically significant in the vicinity of 2000m along the profile line 

- The RMSE for water areas has increased post-correction (i.e., decreased accuracy), which is attributed to low error correlation in these regions and persistent gaps in TanDEM-X DEM coverage. 

 

Section 6

- This paper proposes a least squares configuration-based algorithm (LSC-TXC) for correcting both systematic and random errors in TanDEM-X DEM, resulting in improved accuracy

- Additionally, we conducted a comparative analysis of the correction outcomes with three commonly used algorithms 

- Nevertheless, it does not delve deeper into exploring the correlation between terrain characteristics, land use types and random errors.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

 I appreciate the opportunity to review the article titled TanDEM-X Digital Elevation Model Correction Using Least 2 Squares Collocation Theory. After careful reading, I'd say the article has number of flaws which need to be fixed. Having said that, I have recommendations that I believe the authors will find useful and that, if implemented, would make this research suitable for publication in the journal.

 

(1)   Title of the article is somewhat unscientific hence need to be modified. In my perception it is some sort of book chapter not like a full scientific article

(2)   The given abstract requires significant change in terms of motivation and implication because it is not acceptable at its current stage. An abstract should begin with a general statement about the topic. The beginning statement is not up to the mark as per the journal standard, please consider rewrite the statement. Methods and possible outcomes are not expressed academically. The study's implication is unclear. At the end of the abstract, kindly include a statement outlining the study's implications.

(3)   Introduction is written poorly without proper background and discussion of previous related studies and most importantly the research gap which helped to formulate this particular research. I do not see the idea of the article has been established properly. Without scientific referencing and detailed analysis of those approaches, the conclusion and argument of the article do not sound appropriate.  I recommend the author to look at the latest scientific publication before jumping to any conclusion, as a lot might have been changed in the decade-long period. In other way I would say in the Introduction, it is necessary to reconstruct the storyline of this manuscript. The introduction is the opening part of the scientific story to attract an audience and suggest the direction of your research. For this reason, you need to identify the problem that drives the research and introduce the key characters. If then, using the key characters, it is required to intertwine the scientific story concisely, systematically, and logically. Please rearrange the keywords so that the background of the study sounds appropriate. Please also include advantage of TanDEM-X Digital Elevation Model over other digital elevation model.

(4)   Which region of Spain has been chosen as the study location is unclear to me.  Give a justification for choosing this area for the study. Please replace the study area map with a world map and designate the study area so that people from all around the world may find it. I can find no references to figure d in this area.

(5)   Entire methodology section is poor and vague which make the research unclear in terms of methodological design. It is hard to understand the direction of this research. It is necessary to reorganize the hierarchy of Materials and Methods such, overview of the method, Data Collection, and Data Analysis, validation etc. Please include a brief flowchart to help readers who are unfamiliar with this topic comprehend the study methodology. Although the flowchart is given in the later section but it has not been fit in relevant section.

(6)   Please provide separate the result and discussion sections which might be beneficial for the reader. I do not see rigorous discussion regarding the method, outcomes, errors and validation in comparision to the other relevant studies. Please provide clear and concise discussion about the limitation and future direction of the present analysis and also scientific significance in geospatial science.

(7)   Should be revised. Supplementary document is highly desirable otherwise reader would not be benefited. Please shift the statement on future direction to the discussion.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Thank you for the revisions. The authors have addressed all the questions and comments in the revised version; therefore, I recommend it for publication in its current form.

Good luck

Reviewer 3 Report

im ok with the corrections