Error Estimation of Pathfinder Version 5.3 Level-3C SST Using Extended Triple Collocation Analysis
Round 1
Reviewer 1 Report
I add my report as a pdf file because I have used some formulas.
Comments for author File: Comments.pdf
Author Response
Please see the attachment for detailed reply.
Author Response File: Author Response.docx
Reviewer 2 Report
To record primary geophysical parameters is a prerequisite for monitoring their long-term variability and several platforms are providing public with global time series data with high quality. And the conventional way to validate these data is by comparing with an in situ data set. However, this prevalent bivariate way cannot avoid inherent uncertainties in both testing and reference data sets. Therefore, triple collocation error analysis has been brought into attention to overcome this disadvantage.
Saha et al adopted this method and analyzed sea surface temperate data from Pathfinder v5.3, an global twice-daily data set covering almost data for four decades. They conducted triple collocation analysis with other independent data sets (the independency was confirmed by random error analysis.). The study confirmed the low RMSE of the two tested data sets, which make sense with regards to the high unbiased Signal to Noise ratio.
Before any editorial recommendation would be made, I would love to raise a few of my comments here.
1 Though the authors gave numerous space to introduce the background of this triple collocation method, the result and discussion part are unexpectedly weak. Literally only one section (section 4.3) is on the result from triple collocation analysis per se. Thus I suggest these authors give more discussion on the TC analysis (primarily RMSE result here). Some discussion could be the comparison with other TC analysis on SST data, such as the cited Xu 2010 and 2016.
2 Conclusion section is also a kinda "big head dull" which means the background and the findings are imbalanced. In essence, this section should be superbly concise. Only would primary findings and perspectives for future studies be summarized.
3 Any coding document related to this research ought to be uploaded to some public access database for further methodology examination.
Author Response
Please see the attachment for detailed reply.
Author Response File: Author Response.docx
Reviewer 3 Report
Evaluation by "Extended Triple Collocation Analysis" to guarantee the quality of satellite SST is an important study. There are no significant issues with the method or results, but consider modifying the following:
1) Section 3.1 describes the application of Pathfinder (atmospheric and ocean modeling, coral-reef bleaching, understanding the variability of fisheries yields and analysis of extreme climate events). It is generally not conceivable that such coarse SST data will be useful for fisheries and coral reef studies, and references to these applications should be written.
2) Similarly, in "3.1", describe "GHRSST quality flag (qf = 5)" in more detail. Please explain the structure of the flags, which will be helpful to readers who evaluate other satellite SSTs.
3) How should I consider the difference between buoy point data and satellite spatial resolution (4 km) data in "3.2"?
4) In “5. Conclusion”, you should mention in more detail the limitations (problems) of your proposed method and whether it can (or should) be used for other satellite SST evaluations.
That's all
Author Response
Please see the attachment for detailed reply.
Author Response File: Author Response.docx