Evaluation of GEOS-Simulated L-Band Microwave Brightness Temperature Using Aquarius Observations over Non-Frozen Land across North America

Round 1

Reviewer 1 Report

1. The spatial resolution of Aquarius is not consistent with SMOS and GOES RTM-based TB, which will affect the evaluation results of GOES RTM-based TB. The TB evaluation results in this paper is obtained for what spatial scale (SMOS, GOES RTM or Aquarius)?
2. The GOES RTM-based TB is calibrated from the SMOS TB. Thus, the different between the GEOS RTM-based TB and the Aquarius TB is inherited from the difference between the SMOS TB and the Aquarius TB. In other words, part of the difference of GOES RTM-based TB and Aquarius TB is due to the difference between the SMOS TB and the Aquarius TB, and the other part is due to the uncertainty of RTM. Can we analyze the influence of these two factors on the brightness temperature difference?
3. For Figure 7, after calibration the brightness temperature in the north is missing a lot, please give specific reasons. Can we calibrate the GOES RTM-based TB based on the SMOS TB?
4. In the Results and Discussion, the error dependence on Soil Hydraulic Parameters (wilting point and porosity), Vegetation Type and irrigation area was investigated. Can this paper analyze the contribution rate of three factors to brightness temperature error? Or pixel by pixel contribution rate?
5. For the irrigated areas, I thought that the GOES RTM-based TB cannot capture the change of Surface soil moisture, and mainly reflect large-scale climate and vegetation changes. The satellite TB can reflect the soil moisture state of the earth's surface, so it should have a lower brightness temperature. Therefore, the TB bias (RTM TB - Aquarius TB) should be positive, which is inconsistent with the analysis results in this paper. Although the explanation of surface roughness is given in this paper, I hope the author can make a more adequate explanation in this respect.

Author Response

Thank you for providing valuable comments. Please see the attached pdf file for point-by-point response to the reviewer's comment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The paper describes the use of a radiative transfer model, that is calibrated using SMOS Tb, to model Aquarius Tb. The results showed a lower performance over high porosity/high wilting point and higher performance over forests.

The paper is well written and it was a pleasure to read.   I only have minor comments which I think might make the paper slightly easier to understand.   First, It is very interesting to see the good performance over forests, as this is usually a more ‘problematic’ land cover type at least for microwave based soil moisture retrievals. As now grass and cropland are within one class and the lesser performance of this class is attributed to irrigation, it would be very interesting to see the land cover split for grassland, crop land irrigated and non-irrigated.    Please describe better why you use SMOS for calibration to model Aquarius Tb. At the moment this is not completely clear when reading the first time what the purpose is of this. In addition, do I understand correctly that the input data is soil moisture, soil temperature and vwc from NASA CLSM? Was this then used for calibration with SMOS and then again to predict Aquarius Tb? I could be wrong, but this is how I interpreted it. If correct, I suggest to make this more clear in the text.   The comparison of calibrated and uncalibrated bias, rmse and urmse is maybe not the best, as the uncalibrated areas are areas where any RTM has many issues due to lakes, frozen soils and snow (masking is not always perfect). Hence it is to be expected that the statistics will be worse over these areas and it is hard to contribute this to calibration only. Would it be possible to select an area where calibration is possible, and also run RTM without the calibration and with the LUT and mean values? This would make for a more fair comparison and a stronger argument.    Is there any effect of oversampling Aquarius? It would be interesting to see how the calibrated parameters vary within one Aquarius pixel. I guess ca. 4-9 SMOS pixels go into 1 Aquarius pixel?   What was the cutoff for number of valid observations? Line 251   Some minor textual points: Sometimes the unit is directly behind the value, other times it isn’t. Especially when discussing two values, I noticed sometimes the unit is only behind the second value, other times at both. Eg line 258 compared to 263.   Line 206 and 207, in the PDF I had the Roman caps are not working.   In the text you refer to a b c d e f in figure 6 but the sub numbering is not in the figure.   Line 338 needLEleaf

Author Response

Thank you for providing valuable comments. Please see the attached pdf file addressing point-by-point response to the reviewer's comment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Review of manuscript “Evaluation of GEOS L-Band Microwave Brightness Temperature using Aquarius Observations over Non-Frozen Land across North America” submitted to Remote Sensing

 

The manuscript presents a comparison of a L-Band microwave Brightness temperature simulated from an RTM implemented in the Goddard Earth Observing System, calibrated from SMOS observations, with Aquarius observed TB in non-frozen area in north America. The study evaluate the difference looking at variables such as soil type, vegetation cover and Aquarius passes.

 

The manuscript is clear and well written. However, the scientific contribution of the study seems weak. The results section is comprehensive, but at the end, it is not clear what important information it brings to the microwave community. Maybe a comprehensive discussion on the outcomes of the research could palliate for the lack. In addition, the biases look high to me. At least, it would be important to quantify the impact of such errors on eventual soil moisture retrievals. So the manuscript would need major revision before publication in remote sensing.

 

1. Title: “GEOS L-Band Microwave Brightness Temperature » is confusing. Maybe GOES simulated L-Band Tb… or something like that?

 

2. Line 29-30 : “that in turn affects the 30 measured Tb ». This part of the sentence is a repetition of the last sentence.

 

3. Line 30 : dielectric constant of water « at L-Band »

 

4. Line 48 : carries “3 L-band radiometers”

 

5. Line 95: “canopy temperature that is assumed to 95 equal Ts ». I am surprise that GEOS does not simulate canopy temperature. What impact could it have on the simulation?

 

6. Line 109: remove “it is evident”

 

7. Would be useful to mentioned in the figure label why northern region are so poorly calibrated.

 

8. Line 239 : It would have been possible to interpolate the Tb from the multi-angular SMOS observations.

 

9. It is difficult to distinguish if the higher biases in Northern Canada are related to the lakes and everything, or it could be related to the fact that the model was not calibrated in these regions. Something to discuss?

Minor

 

1. Line 113: Section 33.2?

 

2. Line 301: 55.3

Author Response

Thank you for providing comments. Please see the attached pdf file addessing the point-by-point response to the reviewer's comment.

Author Response File: Author Response.pdf