Interpretation of ASCAT Radar Scatterometer Observations Over Land: A Case Study Over Southwestern France

Round 1

Reviewer 1 Report

Manuscript: ‘Interpretation of radar scatterometer observations over land: a case study over southwestern France’

The authors investigated the impact of leaf area index (LAI) and surface soil moisture (SSM) on satellite-derived radar backscatter (σ°) observations over southwestern France. Unlike similar works, observations from the Advanced Scatterometer (ASCAT) are compared to simulated σ° values produced by the Water Cloud Model (WCM) coupled to the Interactions between Soil, Biosphere, and Atmosphere (ISBA) land surface model. The LAI and SSM variables used by the WCM are derived from satellite observations and from ISBA simulations, respectively. The results from statistical evaluation metrics revealed that the WCM could be used as an observation operator for the assimilation of ASCAT σ° observations into the ISBA LSM provided that large model errors over karstic areas and over wheat croplands at springtime are accounted for. The study addresses a quite interesting topic, in term of operational applications focused on the detection of land‐cover changes. Furthermore, this is a relevant topic lies within the scope of the MDPI remote sensing journal. The article is well organized and neatly written with the appropriate scientific content. However, I found a minor problem that, in my opinion, must be addressed before the publication.

********************************

Title: it fits perfectly the paper content. 

Abstract: it is quite adjusted to the paper content.

Introduction: this section provides sufficient background and includes relevant references about the application of ASCAT over land to monitor variables via backscatter observations. Objectives and the novelty of the study are also clearly stated.

Materials and Methods: the description of the study area and datasets are clearly stated. For the sake of clarity, I think that this section could be significantly improved if the authors add a flow chart with the different methods described in text, highlighting inputs, applied analysis/procedure, and outputs so that readers could understand this section easier. 

Line 95, Figure1: authors could add color to subpanels c to h.

Results: the results have been presented clearly.

Line 274: fix ‘loamy (they are classified as Andosols in the French soil classification system [39]).’

Discussion: it has been presented clearly.  

Conclusions: they’re clear and concise and are supported by the results.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

This study investigates the ability of a Water Cloud Model (WCM) to simulate ASCAT σ° observations with SSM simulated by ISBA and satellite-derived LAI product as inputs over southwestern France. It presents the response of the observed and simulated σ° to changes in LAI/vegetation cover and SSM across seasons and land cover types.

Comments

The study is well structured with a clear presentation of results. However, the accuracy of the LAI product used in the paper needs to be examined over the study area and especially in the Landes forest region. On L362-363: “Before the storm, the LAI of the pine forest presents a marked annual cycle and ranges from 1.2 m²m^-2 at wintertime to 4.5 m²m^-2 at summertime”. This is in contrast to small seasonal variations of LAI based on field measurements in evergreen needleleaf forests in other studies, such as in Heiskanen et al [2012]. I wonder whether this is due to the use of the coarse resolution LAI data (0.25x0.25deg) in the study? So that the LAI is from mixed land cover types instead of pure pine forest. I suggest the authors examine the LAI data at its original 1km resolution over the Landes forest area, and see if the seasonality of the LAI for pine forest is reasonable relative to field measurements. If it does, then they can focus their analyses on relatively pure pixels of pine forest at 0.25deg resolution.

Heiskanena, J., M. Rautiainena , P. Stenberga , M. Mõttusb , V-H Vesantob , L. Korhonenc , T. Majasalmi, Seasonal variation in MODIS LAI for a boreal forest area in Finland, Remote Sensing of Environment, 2012, https://doi.org/10.1016/j.rse.2012.08.001.

The writing of the paper can be improved. For example, in several places the paper describes that the observed σ° values are compared to simulated (in the abstract), and the observed are lower than the simulated (e.g. L379). Given the observations are used as the reference, I suggest the authors re-write such sentences.

 

L305, change “that” to “than”

L322-323, note the observed σ° is used as reference, I suggest change the sentence to something like the simulated σ° values at springtime are systematically higher than the observed…

L334-344, I wonder if it would be more informative and/or straightforward to simply plot time series of Sigma0 and LAI for the Landes forest to show the changes before and after the storm? I understand this is to follow Teuling et al [13], however, that study was to present the evidence of cloud cover enhancement over the forested region relative to non-forested agriculture areas. What are the advantages of using the difference here?

In addition, can you provide some explanation/causes to the observed σ° change before and after the storm? Is it due to changes in land cover? Are there field evidence/pictures?

 

L349, change “on” to “one”

L362, according to ESA CCI Land Cover map (https://maps.elie.ucl.ac.be/CCI/viewer/) and field measurements in Heiskanen et al [2012], evergreen coniferous forest dominates the Landes forest region, which has small seasonal variations in LAI. Are there field measurements of LAI in Landes forest area? Winter LAI of 1.2 appears to be too small for evergreen coniferous trees. Is this due to the use of coarse resolution LAI data? See above.

L349-370, it would be helpful and perhaps makes the paper more interesting to link the changes in LAI, Sigma0, and WCM parameters to changes in land cover available from the annual CCI land cover maps (see link above).

 

L371, Discussion. The same WCM has been used in previous studies to assimilate ASCAT σ° into land surface models (i.e. [7]). Please provide some comparisons in terms of methods/results between this study and previous studies. For example, what are the ranges of the four WCM parameters in previous studies? Do they encounter similar issues over karstic areas and wheat croplands at springtime?

L379, suggest change the sentence to the simulated values are higher than the observed

L403-409, can you provide the basis for the relation between observed σ° and LAI/SSM? I believe that it can be found in the literatures, which would be helpful to better understand the impact of LAI/SSM on the simulated/observed σ°.

Please make sure the links to the references work, I can’t open some of them, such as 19-21.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Dear Authors,

 

Thank you for choosing Remote Sensing for your work. I found it very interesting as it deals with not very well investigated topics of assimilation of radar backscater data from a scatterometer. Although the research design is generally appropriate I was looking for your field observations which verify that the studied objects (different ecosystems) are properly monitored. Although the data you use to investigate and test your hypothesis is of good quality, i.e. Copernicus Land Monitoring Service LAI and the ASCAT data, the low resolution and model uncertainties (for LAI estimation) input a bias in the very beginning. My question is how do you deal with that and how you can overcome the error in LAI estimates. LAI tend to saturate above a value of 4 and should be taken with caution especially when modelled. My proposal to you is to scale down your study to a higher resolution and after you get meaningful results to upscale. However, the choice of the approach is yours but if you aim at working on global scale with the models it will still be needed further investigations to overcome (if it is possible at all) the limitations which I already mentioned. Thank you in advance for taking into consideration my remarks.

 

Kind regards,

Reviewer

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

This study explored the use of WCM with LAI and SSM as input variables to simulate ASCAT σ° observations in contrasting land cover conditions.

Comments/Suggestions for Authors:

TITLE - the title does not reflect what is done in the paper.  The title is too broad, "Interpretation of radar scatterometer observations over land".

Abstract:

Line 13: LAI of what vegetation? Line 19: What are these four parameters? What is the importance of these four parameters? Line 21-22: ASCAT can detect rapid change in vegetation... What is this rapid change? Describe what this is and what caused it. Line 24: What do you mean by marked seasonality? What are these seasons (for people from countries that does not have the same season with your study area). Line 27 just repeated lines 19-20. The abstract did not include the importance of this study and what is new in this research, its contribution to science. What is really the focus of this paper?

Introduction:

1. Inrtoduction is too short did not provide essential information what this paper is about. It is poorly structured. Was not able to establish why this paper is important and what is new in this paper.

2. Line 49: What is this observation operator?

3. Line 49-51: Too many acronyms. readers will have a hard time remembering what the acronyms stand for.

4. Line 59-60: How do the models used by Vreugdenhil and Lievens differ?

5. Line 75-76: This can be deleted.

Materials

I suggest materials be merged with Methods (Materials and Methods). Line 89-92: You mentioned crop rotation systems, what other crops are planted in place of corn in the south and wheat in the east? Lines 110-119 - This can be shortened. Line 130-133: Confusing sentences. Please rephrase. GEOV2 LAI is derive d from SPOT VGT from 199 to 2014 and from PROBA-V from 2014 - present.

Methods:

Line 142-154: This should be included in the Introduction. Line 196-198: how does your research differ from the work of Al-Yaari? Line 233: Result analysis can be replaced with Statistical analysis.

Results:

Line 286: what are these numbers - 209327 and 204520?

The Title, Abstract and Introduction is disconnected from what is written in the method section, Results, Discussion and Conclusion.

Overall, the paper is poorly written.

 

 

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The authors have satisfactorily addressed all my comments. I only have one minor comment: I’d suggest to incorporate Section 4.5 (Are satellite-derived LAI values reliable?) into Section 2.4 (LAI observations). In addition, note that collection 6 MODIS LAI product has improved over collection 5 https://www.mdpi.com/2072-4292/8/6/460.

Author Response

Please see the attachment

Author Response File: Author Response.pdf