Challenges and Evolution of Water Level Monitoring towards a Comprehensive, World-Scale Coverage with Remote Sensing

Round 1

Reviewer 1 Report

Dear authors,

Water level monitoring is an essential task in 21 century. Are your research refer also to other indicators like Equivalent of water height (EWH) if yes, then it would be nice to compare some values with the COST-G action or EGSIEM project.
The manuscript clearly describes the experiment. Nevertheless, they are the following minor issues:
line 46 several web platforms - list please those platforms with reference
Line 236 basckatter ? or maybe backscatter or backwater, please correct. Please check the vocabulary in the whole draft.
line 233 - (troposphere, dry troposphere )
line 287 equation 2 did you invite this formula? If not, please add reference
line 302 figure 4 - at the scale of the figure for the top and bottom figures. What is mean the colors of the dots on the river mean?
Table 3, Table 4, Table 5 - In some tables, the values are bolded; why?
In other tables, none of the values are bolded. Please unify this, and explain why some values are bolded, for example, in the caption.

In figure 6 the labels of the time-series graph are too small, and on the right side, the map should have the units as well as the north direction (it is easy to add in such software as qgis or arcgis).

line 378 - 379 Climate science often studies large-scale Spatio-temporal phenomena depicted by 378
long data records - please paraphrase this sentence; maybe climate scientists often study. Science can not study.

Figure 7 -  I am unsure whether this figure is necessary for this article. Such a figure would be good in some tutorial or appendix, not in the scientifical article. What is the message of this figure?


Figure 8, line 410 -411 - Such figures should have a scale bar in a visible place. The black labels on the dark green background are hard to read.

Is this approach work well also in the smaller river basins in western Europe, if yes please add them to analyze(one as a representative example). 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

This paper provide a nice overview of the challenges in altimetry based water level monitoring and provide an interesting approach for increasing the global coverage of space based water levels globally using new automated pipelines for multi-mission water processing.    

 

Good points:

·    ·       Automatized pipeline to extract virtual stations (teroVIR) and to process water levels (teroWAT). Well represented on figures 2 and 3.

·   ·        Increases the number of monitored water bodies and gets more out of altimetry missions (mainly Sentinel-3, figure 5) in developing countries and remote areas. Nice visualization on figure 4

·    ·       Nice scoring option (teroSCO) which considers the number of points in the time series as the Peason correlation coefficient and the RMSE tend to be higher for very low number of points

·   ·        Interesting introduction displaying the different altimetry databases, some of their processing methods and their altimeters

·  ·         A step towards a combination of the different databases (Jason, Saral, Sentinel-3)

·   ·        Possibility to use both L1 and L2 data (though there is not much detail on their implementation of this teroWAT)

 

Suggestions and questions:

·         Simple summary mentions a solution for water bodies monitoring - but not what type of monitoring - maybe worth adding Water Levels monitoring

·         Line 71, maybe include a column on table 1 showing the processing process instead of the useless column T_agg which is “NRT” for all the datasets

·         Line 130, maybe mention that laser data (e.g., IceSat-2) is out of scope? Sentinel-6?

·         Line 162, are these lines buffered to create an AOI? How big is the buffer and would that include nearby water bodies?

·         Line 182, does that operate for each tile independently or would that also deal with a lake and the boundary between two GSWE tiles?

·         Line 190, not sure to understand

·         Line 202, I think there is an issue with your first formula, the unit would be m^4 not m^2. Also define al and ac as the resolutions

·         Line 240, does it not thereby remove extreme events?

·         Line 2550, why SRTM and not a newer DEM like the Copernicus?

·         Line 273, why do define MAD and not ubRMSE? You do not use it in your teroSCO

·         The subplots of figure 6 could be made bigger

·         Figure 7 is not used and illegible

·         Line 400, yes it should not be hard to combine VS corresponding to a same lake

 

 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report

1
The manuscript under review provides an overview of altimetry-derived (lake) water level monitoring and introduces a new methodology, namely teroVIR & teroWAT,
to exploit altimetric raw data in the form of L1 and L2 products to effectively retrack surface height variations over non-ocean water bodies. I would like to
congratulate the authors on a very well-written manuscript, which effectively describes the current status and the proposed state of the art. I especially
appreciate the introductory part in lines 64-81, where in a very coherent and simple way the authors state the current limitations one faces when using a
method designed for open ocean in areas close to the coastline and over land areas. There are only a few points that need to be addressed, which do not remove
anything from the overall high-merit of the paper. After the below mentioned points are considered the paper can be published, hence the recommendation is for
a minor review.
Line 103: An L2 data....
Line 105: please rephrase "operationalised". One option is to replace it with "...data based algorithm [28] has been developed and..."
Line 111: Do you mean statistical metrics?
Line 113: "...previously uncovered" or better "previously unsurveyed"
Line 127: It is true that Cryosat-2 is currently in a drifting geodetic mode, hence not on an exact repeat mission. But, it offers high spatial resolution,
so could it be used for the verification and homogenization of the other missions when in-sync at some area? For instance a dual-satellite crossover may well
serve this purpose (validation) and referencing all satellites to a common frame, hence each satellite's absolute altimetric bias would not pose a problem.
Of course, in many cases Jason-2 serves this purpose depending on the agency that provides the altimetric records.
Line 144: ....by crossing the altimeter's....
Line 152: What do you mean exacty by "cleaning of geometries"? Some more info is needed.
Line 180: ...the output collection...
Line 183: ...components, which actually...
Line 202: Please define pi, al(tr), ac(tr). They are given in Table 6 but they should de be defined in their first occurrence.
2
Line 211-212: Rephrase as: The altimetry L2 products provide directly only the satellite ranges, whereas the L1 ones consist...
Line 216: The necessary inputs
Line 218: It is then possible...
Line 234: What about the inverse barometer correction? Also, what do you mean by "computed"? Doesn't the algorithm use the already provided geophysical corrections?
Line 238: Provide a reference for the Interquartile Range Rule
Line 247: What do you mean by "too noisy"? How much is too noisy and how much is not? Please quantify it or provide some reference.
Line 251: SRTM may be deemed as "old" and provides in some cases inferior quality to the Copernicus DEM and the one by JAXA. Could this create any problems in
identifying blunders?
Line 258-260: This address part of the comment on 127, but some more details are needed.
Line 264-265: ...platform adapters are only needed
Line 281: The unbiased root mean square error (ubRMSE) and the correlation coefficient (R) should be devfined here.
Line 302: Before Figure 4, please give some introductory text and not immediately just the Figure.
Line 306: Table 2 shows that....
Line 323: ...for the SRL, JS 1-2-3 and S3A-3B missions.
Line 357: This title does not sound correct. You actually mean "Adapting to user needs" or something similar. Please rephrase.
I think that it would add to the value of the paper to include results on the application of the teroVIR and teroWAT algorithms. For instance, some examples on the
identification by teroVIR of virtual stations and the identification of false ones would help realize on its applicability. Also, for teroWAT it would be nice to
include:
a) where L2 altimetry data can be downloaded from? Is is compatible with data from Copernicus, AVISO, NASA, RADS, etc.?
b) examples L2 of blunders identified over water bodies
3
c) L1 noisy waveform fitlering output
d) examples of retracked L1 data that do not agree with the SRTM DEM
e) exaples of results on the mutli-mission comparison and bias identification. To which mission are the data referred to? If they are referred to the overall mean,
then this is some arbitray value that changes if the datasets are changed (either by including/excluding missions or for different time spans)
A final question refers to whether the algorithm(s) will be developed into a new service avaialbe for the scientific community, or whether it will be included in one
of the available services.

Comments for author File: Comments.pdf

Author Response

Please see the attachment

Author Response File: Author Response.docx