Monitoring Soil Moisture Drought over Northern High Latitudes from Space

Round 1

Reviewer 1 Report

The authors used raw brightness temperature (TB) observations from SMOS for drought monitoring. The idea is novel but looks questionable to me.

Fairly speaking, the proposed Standardized Brightness Temperature Index is able to capture brightness temperature anomalies. However, the TB anomalies do not have to be caused by drought. As we know, the satellite observed TB is governed by land properties such as vegetation, soil, and open water. For wetter-than-normal conditions, we may see decreased Tb due to expanded water bodies and wetter soil which may decrease TB; or increased TB due to denser vegetation; and vice versa for the drier-than-normal conditions. Therefore, it is hard to draw any conclusion directly from the TB dynamics. For example, the false drought detection in August 2018 (Fig. 4-5) may be caused by recovery of vegetation.     

The approach is worth exploring; however, unless the above concerns are reasonably addressed, the paper is not suitable for publication.

Author Response

Dear Reviewer#1

We would like to thank you for reviewing our manuscript. For detailed response on your comments, please see the attachment. 

Author Response File: Author Response.doc

Reviewer 2 Report

Overall comments:
In this study, the authors use satellite-based brightness temperature data and land surface models to develop and validate a new drought index called STBI. The STBI is validated against drought indices from the LDAS-Monde, two satellite-based SSI and an SPI. My main concern relates to the data quality that the authors used in this study. For example, the model-based index was used as reference values to validate STBI. However, there are no evidence or supporting works that these indices can reliably serve as reference values. Even, the basic hydrological variables from model and precipitation data have not been validated over these study areas. Also, the other problems with this study are their method based on Tb data, which does not have any theoretical meaning. And spatial resolution mismatch between different data sources is another problem. Also, the masking process of Tb makes no senses. Which category of drought indexes does STBI belong to (meteorological drought or agricultural drought)? Before applying a new index in extreme environmental conditions, it is recommended to apply a new index to an extreme environment after verification in a typical environment condition.

I enjoyed this paper and look forward to seeing it published. There are a couple of major problems that need to be addressed before publication.

Further comments are listed below.

Specific comments:

Abstract

The current abstract is too ambiguous and lack of details. The abstract should include the type a summary that identifies the purpose, problem, methods, results, and conclusion of this research project. Also, this abstract does not provide any specific results of your research. For example, what is “two satellite-derived SSM (which satellites?)”, what is “the before mentioned drought”? What is the actual part of the “superior part” of STBI? How exactly “the STBI can capture the 2018 drought onset, severity and extent”? Which precipitation data did you use? Satellite or assimilation data? What additional information STBI? Also, if SMOS has a problem with retrieving SSM which is based on Tb value, how Tb data can provide more information regarding surface-related conditions?

L8: derived

L14: SSM retrieval is hard to make.

L18: (WMO) shows

L27: dry surface conditions

L29: more references are needed

L37: Nordic regions are getting wetter, but why droughts might still occur in the future?

L42: surface soil moisture (SSM)

L48: Not only VWC but also snow cover will strongly affect the Tb value. How can you control this fact? Also, if Tb value is considered for VWC, why don’t we just use VWC data or VOD data instead of Tb value? What are the advantages of using Tb instead of these data sets?

L54: The reason for having a large spatial and temporal gap is not a problem of satellite observation systems. Instead, it is very natural results because the soil is actually frozen.

L56: How drought is defined when the ground is frozen?

L73: This means, you have a different number of data sets for SM and Tb. You should report the number of data used and discuss possible biased that caused by this difference.

L80: Why TbV showed different results with TbH?

L83: Under dry conditions, microwave emissions are originated from deeper soil layer (more than 5 cm) than wet conditions.

L86: This does not make sense. SM in deep soil layer provides plants with sufficient soil water to maintain a higher photosynthetic during the dry season.

L100: In this study, you masked out SM data because of its low quality of data. However, for Tb data you did not mask out even the quality of data can be low. To make this study fairer, the SM data that have low quality should not have been masked out.

L105: What ESA CCI SM data is used in this study? It also includes SMOS SM data which is not independent of SMOS data you used.

L113: More detail information regarding P data is required.

L113: The method of the moment might be appropriate since you do not have many monthly data. In addition, it seems that the author tested their model’s adequacy each pixel. Moreover, why the authors chose Spapiro-Wilk test?

L146: similar to 2.5, the model’s accuracy test should be conducted for the beta probability distribution. If the SMOS SSM data only retrieved a few days per month, the data monthly data must be highly skewed. Why the authors chose MLE instead of MOM?

L151: The reason for selecting SPI-1 is unclear.

L154: Why empirical CDF was considered?

L158: So, the author calculated drought index from P values using 69 years data, while SSI for only nine years?

L162: With frozen soils, dry or wet conditions cannot be defined.

L167: The map of p-values should be included (what is the results that when you set the significance level at 0.01 level?).

L168: This is a duplicate statement that already mentioned in the method part.

L172-175: Do not need to explain all this. Or if it must be, it should have been mentioned in the methodology part.

L180: Figure 1b do not have any meaning—it only shows the spatial distribution of Tb_H which is not related to your fitting results.

L184: I do not get why you need the SSI_ESA.

L191: Why D0? Detail explanation of D0 is also required.

L199: Why LDAS-based index was considered as accurate index?

L202, 203: It does not mean that LDAS data is superior to other source of data sets over this area. Please add more previous research.

L216: Does LDAS includes SMOS?

Figure 3: Set the range of the color bar from 0 to 1.

Table 1: How may data per pixels were used to calculate R values?

L254: How Tb-based index can be related to the dry or wet condition of the surface? What is the exact theoretical reason?

L295: references needed

L301: Again, what exactly the Tb-based index provide more information regarding the dryness of surface conditions over frozen soil?

L330: Tb values can also provide an opposite trend of surface dryness conditions.

L343: To fully investigate the advantages of using STBI, various drought indexes over non-frozen areas should be considered.

L348: SMAP, Tb_H, LDAS-Monde, etc (no dot need to make acronyms again).

Author Response

Dear Reviewer#2

We would like to thank you for reviewing our manuscript. For detailed response to your comments, please see the attachment. 

Author Response File: Author Response.doc

Reviewer 3 Report

please find attached pdf with my review

Comments for author File: Comments.pdf

Author Response

Dear Reviewer#3 

We would like to thank you for reviewing our manuscript. For detailed response on your comments, please see the attachment. 

Author Response File: Author Response.doc

Round 2

Reviewer 1 Report

The revision looks good to me.