Enhancing Sea Surface Height Retrieval with Triple Features Using Support Vector Regression
Round 1
Reviewer 1 Report
This paper investigates the feasibility of using features obtained from SNR arc to predict sea level height. The accuracy of sea level height retrieval is improved based on the proposed model. This work may be an interesting topic for potential readers. However, more details and the implications of this work should be further clarified so that it can be published in Remote Sensing
(1) Could you please provide detailed information about the methodology employed for sample selection in the experiment and describe the process of assigning those selected samples? (2) In the context of environments with large tidal variations, why do the authors consider the conventional method to be limited? In addition, what do the authors think GNSS techniques introduce errors that affect the accuracy of the results?
(3) What is the rationale behind selecting frequency, amplitude, and phase as the feature parameters extracted from GNSS-R for sea level height retrieval? Why are these parameters considered important? (4) In the case of traditional models, I am curious to know if the authors implemented certain corrections and other related quality controls. Could you provide information on how these factors were addressed in the study? (5) Could you please elaborate on the advantages that the authors attribute to the proposed method compared to traditional methods? What makes the proposed method stand out in terms of its benefits and potential improvements over existing approaches?
Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
Reviewer 2 Report
The manuscript title must be corrected. "... Sea Surface Wave Height ..." - waves are always on the sea surface . The manuscript was said about the sea surface height in the strong tide area . "Wave" is it a tide or wind waves?
Figure 1 - "Sea Level" to replace with "sea surface height", since the tides heights are present in the study.
There is no Figure and information, where is the highlighted area with which the information is obtained? At what distance from the stations is this area?
Maybe such errors arose due to a large distance between the Illuminated area, the antenna and a tidal gauges?
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 3 Report
The manuscript shows an application of a machine learning technique (Support Vector Regression) for sea level estimation in two specific regions. The results are robust, however some aspects of the manuscript need to be improved for publication. Below are some comments and suggestions.
Major:
1) Shows all hyparameters used in defining the SVR model. Including target information. Note that this is not clear in sections 2.2 and 3.2 (flowchart). All research needs to be reproducible.
2) It is unclear whether any preprocessing of the data for training was done. If so, make this clear in the manuscript. In this case, think about other users who want to reproduce your results, with the information given in the manuscript is this possible?
3) Apparently there is an imbalance in sea level data (more positive than negative) for SC02, which I did not observe in BRST. This seems to affect the results shown in 3.3.1, note that there is a skill difference for positive and negative sea level values. If this data imbalance really exists, can the model be influenced by this distribution? If so, it is necessary to investigate this effect and bring new results (trained balanced SVR) to the manuscript. Please insert BIAS between metrics. Also evaluate the metrics for two classes: positive and negative sea level values.
Minor:
1) Missing punctuation at the end of some sentences. Please check this in the manuscript.
2) define lambda in eq. 3.
3) In 2.2 "an empirical value set by hand", which value? (remenber, reproducibility in scientific research is important)
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Round 2
Reviewer 3 Report
Dear Authors,
Thanks for your effort to make your manuscript clearer.
Best wishes
