Round 1

Reviewer 1 Report

This paper describes a new automated method based on time-dependent curve fitting by using TIR to detect lake ice phenology parameters. It depends on a combined daily MODIS Terra and Aqua product to fill gaps caused by cloud cover. The unfrozen water fraction is calculated after initial application of thresholds and logistic curve fitting.  The sensitivty study gives insights into the impact in regard to different LSWT temperatures and its variabiloity for lake ice phenology events. 

This paper is a good illustration of the need for less time pressure in science. The idea and concept of the automated method to extract lake ice phenology events is well developed, but it lacks a thorough validation and discussion. Overall, the writing needs improvement to fulfill a good scientific writing style especially the usage of appropriate scientific terms. 

Main comments:

• Well developed and applied concept of lake ice phenology detection using TIR
• Lacks details information to frame results and validation section
• Misses an independent validation using high-resolution satellite sensors or insitu measurements
• Needs thorough reworking of the discussion section

Detailed comments:

Abstract:

• ‚as they provide four observations per day with a high spatial resolution‚ - Use ‚from 250 to 1 km spatial resolution.‘ As this is not a high spatial resolution. 
• ‚by a quasi-universal time function‘ Please replace the term with a more appropriate one (ff in the text)
• ‚showed good agreement with various satellite data sources.‘ Please quantify ‚good agreement‘ e.g. the accuracy, in addition mention the validation products used
• ‚revealed trends to a later freeze-up period and a faster freeze-up rate, as well as trends to longer break-up periods for the lakes in the Tibetan Plateau and shorter break-up peri-ods for the lakes in northeastern lowland China’ Quantify the observed trends and their significance niveau as well as the time period where these trends relate to.

Introduction:

• ‚have also enjoyed popularity as they provide higher spatial resolutions (250 m, 500 m, 1 km) than that of AVHRR data‘ Change sentence and wording ‚enjoyed popularity‘. There are multiple reasons why MODIS products are often used. Furthermore, there is no advantage if you use the 1km MOD11A1/MYD11A1.
• Mention for example in addition higher resolution satellites such as VIIRS , Sentinel-1, Sentinel-2
•  ‚lake ice phenology ,’ Space error.
• ‚However, the spatial and temporal resolutions of optical sensing data are strongly limited by cloud obstruction, especially for some regions with strong convection, such as Tibetan plateau, where the higher cloud rates during the day compare with the night, with a result of diurnal convection [28].‘ Reformulate sentence.
• ‚MODIS daily tem-perature products have a unique advantage than other products that could provide two observations during the day and two observations during the night., which could increase the amount of available data.‘ Punctation marks error.
• ‚Currently, various methods have been developed to reduce the cloud problem and increase data availability, such as multiple sensor data or tem-poral-spatial combinations [29-31].‘ Change to ‚reduce and fill data gaps due to cloud coverage, ‚such as multi sensor data products … ‚ In addition, the cited papers are rather old, ‚currently‘ implies newer published approaches which are also available.
• ‚utilized binary thresholds‘ Please reformulate ‚utilized thresholds for binary classification‘. 
• ‚time series of averaging near-infrared reflectance and surface temperature values over all lake pixels [18]‘ Not correct, [18] used only near-infrared reflectance values. The combined averaging of near-infrared and surface temperature was the contribution (development) of [20].
• ‚The efficiency of a threshold method may vary among different climatic regimes or different remote sensing data sources‘ Please differentiate between fixed thresholds (applied for multiple lakes) and thresholds derived from the lakes itself. The later is not dependent on a specific climate regimes.
•  ‚Based on the idea of the time-dependent parameterization of the freezing/thawing process, we propose a new automated approach to extract lake ice phenology from MODIS daily temperature products based on the time-dependent pa-rameterization of the freezing/thawing process.‘ Check sentence.
• ‚In our study, the approach is validated using data from three seasonally ice-covered lakes‘ Check Grammar and see discussion validation section.
• ‚in the ice phenology are evaluated‚ Replace ‚‘evaluated‘ with ‚analysed‘.

Study Area:

• Table 1: Add sources of lake characteristic information listed in the table. 

Data:

• ‚The MODIS mounted‚ Replace with ‚the ‚MODIS sensor mounted‘
• ‚at approximately 10:30 and 22:30 local time from Terra and at approximately 1:30 and 13:30 local time from Aqua‘ Please confirm, is this really the local time, not the euatorial crossing time?
•  ‚The MODIS daily snow product and MODIS daily reflectance data (MOD09GQ, ver-sion 6) were used to validate the change pattern‘ See section 4.2
• Indicate the time period covered by the MOD09GQ as well as MOD10A1 data used in the text (not only in Table A1, first reference for Table A1 is in section 4.2).
• Why did you prefer to use MOD09GQ and MOD10A1 and not  MYD09GQ and MYD10A1 or both as combined product? I would see a clear advantage of using combined Terra and Aqua products for this comparison.
• Regarding the MOD10A1: Which NDSI threshold was used for the snow/no snow classification? This is important to know.
• Add here a short paragraph describing the AMSR-E/2 data product used for the comparison.
• Add air temperature data used.

Method:

• Using four MODIS temperature products per dayis certainly a benefit to reduce missing information caused by clouds. However, variations in the composite (MODIS_merge) due to the diurnal temperature cycle of lake and even more day-night differences have certainly an impact on the detection of lake ice phenology events. These impacts are either mentioned in the method section nor in the discussion section, but are important to mention and discuss to explain uncertainties in the methods and likely deviations in the results. A merged product containing day and night temperature values the pixels will contain a large range of LSWT values which reduces the accuracy of the final extracted lake ich phenology dates.

Method 3.2.

• a), b) How were the thresholds 0.5°C and -0.5°C respectively defined? Its not clear until the sensitivity study (which also suggest slightly different thresholds. Please clarify and describe in the text.
• How do the authors deal with pixels having LSWT temperatures > -0.5 and < 0.5?
• ‚Afterwards, the adjacent temporal filter ‚ Please clarify what a adjacent temporal filter is.
• ‚Afterwards, the adjacent temporal filter was applied to recover the information from pixels on the preceding and following days to deduce the water/ice status from cloud-covered pixels. ‚ Please define here the number of preceding and following days i.e. the time window used for the gap filling.
• I don’t understand why your data contain at this point ‚remaining cloud pixels’ which need to be Flagge. Please explain. 
• ‚temporal window lengths of 5 and 7 days‘ Temporal window length: Is this +/- 5 and 7 days? Or 5 days for preceding anf 7 days for following days? Please clarify. 5 and 7 days seem quite long to me since lake ice phenology freeze-up or break-up phases are fast changing. How did you define the number of days?
• Did you do any weightening applied for the gap filling? 
• ‚contained abnormal unfrozen water cover fraction values‘  Wording ‚abnormal‘, please repplace (also ff).
• ‚preceding the date when the mean air temperature in autumn dropped below 0°C ‚ Missing data source and e.g. spatial and temporal resolution for air temperature used. Add to the data section.
• ‚The latter date was‘ What is lauter referreing too? Please clarify.
• ‚Second, the unfrozen water cover fraction was removed if the percentage of “invalid” pixels in an image was larger than 80%.‘ Clarify removed. Does it mean you obtained a data gap on that specific day?
• ‚was employed to evaluate lake ice phenology’ Word choice ‚evaluate‘. Please change.
•  Figure 3. Please explain ‚original‘

Method 3.3

• Please discuss how the curve fitting accounts for multiple break ups and freeze ups in the respective phases.
• ‚where ? is the days of the hydrological year, ? is the unfrozen water cover fraction value at time ?; ? is the maximum value of the unfrozen water cover fraction, which is equal to 1 in our study‘ Please change order accordingly to formula.
• Figure 4: Low quality of figure (blurred). Please replace.

Results: 

Results 4.1.

• Note, very extensive description, which could be shortened.
• ‚Overall, after cloud removal by daily combination, it has been shown obvious increases in the number of valid pixels in MOD11_Merge images and the number of images that represented “valid” pixels reaching more than 80% of the whole lake area during the 2002-2016 period.‘ Check sentence Grammar.
• ‚negative cloud effects‘  Wording. Find a appropriate replacement.

Results 4.2.

• This is not an independent validation if it is the same sensor.  Thus I would highly suggest to call it comparison. What is needed is a independent validation from a higher resolution sensor (such as Landsat, Sentinel-2 or insitu measurements).
• As indicated earlier: These comparison misses details such as the time period where a comparison has been conducted. This accounts for both MOD09GQ as well as MOD10A1. Its first mentioned in 4.3 which is by far too late.
• ‚in agreement with previous studies [15, 24, 26].‘ Please add the date Ranges / accuracy of other studies to put you results in context.
• Figure A3 and A4. Please add label for the MOD11_Merge Temperature product. In addition a label for clouds would be helpful for the readers.

Results 4.3

• Trend calculation misses significance niveaus (see also 4.4)
• ‚Overall, the interannual variations in ice phenology showed good agreement be-tween our study and existing studies, except for inconsistent trends in the freeze-up dates‘ Please Elaborate what you mean with ‚inconsistent trends‘. 

Results 4.4.

• Consistent use of FUS, FUE, BUS. BUE
• ‚When validated on three different types of lakes, the proposed lake ice phenology extraction approach has proven its robustness when avoiding lake-specific thresholds.‘ Im not convinced about the robustes of the method since it lacks a rigoros validation as mentioned earlier.
• Table 5: Please reformalte table header. It is hard to read.

Discussion

• The discussion section needs to be reworked to clearly discuss the advantage and disadvantages of the method, its impacts (e.g. clouds, composite effects of day and night LSWT values, etc). 
• A thorough discussion of the to adress the impacts and its effects on the results of the proposed method is missing.
•  Example: Cloud/ice discrimination and/or Cloud masking. Discuss uncertainties in cloud masking, quality during night time.
• Discuss possible application (e.g. needed requirements such as lake size) to apply you method to other lakes.

Discussion 5.1.

• Starts with a 1/3 page Literatur review. These basics are important, but should be part of the introduction or shortened.
• The sensitivity study is of high interest. A big benefit and added value for this study as well as the scientific community would be to use the sensitivity study (of the discussion section) to derive uncertainties for the final lake ice phenology product.

Discussion 5.2.

• Nice and important section. However, please explain how you obtained those results. It only says in the Figure caption that you performed a correlation.
• Missing data source of Weather stations used. 

Conclusions

• ‚We developed a new automatic algorithm to characterize lake ice phenology using MODIS daily temperature products from 2002 to 2016 by calculating time series of unfro-zen water cover fractions for freeze-up/break-up periods and then fitted by a quasi-lo-gistic function capturing the temporal pattern of the transitional water/ice phase and ex-tracting the lake ice phenology characteristics.‘ Check Sentence length.
• ‚allowing us to trace the ice phenology on a more detailed level than simple binary ice-on/ice-off events‘ Please Elaborate.
• Summarize main advantages and disadvantages of your developed method.
• Add quantification of your results (i.e. trends) and its accuracy / comparison.

Author Response

Reviewer#1

Comments and Suggestions for Authors

This paper describes a new automated method based on time-dependent curve fitting by using TIR to detect lake ice phenology parameters. It depends on a combined daily MODIS Terra and Aqua product to fill gaps caused by cloud cover. The unfrozen water fraction is calculated after initial application of thresholds and logistic curve fitting.  The sensitivity study gives insights into the impact in regard to different LSWT temperatures and its variability for lake ice phenology events. 

This paper is a good illustration of the need for less time pressure in science. The idea and concept of the automated method to extract lake ice phenology events is well developed, but it lacks a thorough validation and discussion. Overall, the writing needs improvement to fulfill a good scientific writing style especially the usage of appropriate scientific terms. 

Main comments:

Well developed and applied concept of lake ice phenology detection using TIR;

Lacks details information to frame results and validation section;

Misses an independent validation using high-resolution satellite sensors or insitu measurements;

Needs thorough reworking of the discussion section;

Response: we thank the reviewer for the suggestion. Following the reviewer’s suggestion, we have added the comparison between our results and Landsat images with a high spatial resolution of 30 m in section 4.3. Also, the advantages and limitation of the methods have been added in section 5.3 in discussion.

 

More details are provided in the specific comments below.

Detailed comments:

Abstract:

1 ‚as they provide four observations per day with a high spatial resolution‚ - Use ‚from 250 to 1 km spatial resolution.‘ As this is not a high spatial resolution. 

Response: Thanks for the suggestion. The old sentance has been changed to “The algorithm is based on Moderate Resolution Imaging Spectroradiometer (MODIS) daily temperature products, which have unique potential for monitoring lake ice cover by providing four observations per day at 1 km spatial resolution” (lines 22-25)

 

2.by a quasi-universal time function ‘ Please replace the term with a more appropriate one (ff in the text)

Response: it has been changed to “by a parameterized time function”

 

3. ‚showed good agreement with various satellite data sources.‘ Please quantify ‚good agreement‘ e.g. the accuracy, in addition mention the validation products used.

Response: the sentence has been changed to “The temporal pattern of water/ice transition phase was performed by unfrozen water cover fraction extracted from MODIS daily temperature data, and validated by MODIS snow products with a good agreement with R2 of above 0.8. The annual variation of extracted ice phenology dates showed good consistency with MODIS reflectance and AMSR-E/2 products” (lines 27-31)

 

4. ‚revealed trends to a later freeze-up period and a faster freeze-up rate, as well as trends to longer break-up periods for the lakes in the Tibetan Plateau and shorter break-up periods for the lakes in northeastern lowland China’ Quantify the observed trends and their significance niveau as well as the time period where these trends relate to.

Response: the interannual variation of ice phenology among the lakes are added as “The results showed that Lake Qinghai had the most obvious trend to delayed freeze up start date (FUS) (6.31 days/10 yr) among the lakes in Tibetan plateau, and Lake Hulun the obvious tendency to earlier break up start and end dates (BUS and BUE) with the change rate of -3.73 days/10 yr and -5.02 days/10 yr, respectively”. (lines 33-36)

 

Introduction:

1.‚have also enjoyed popularity as they provide higher spatial resolutions (250 m, 500 m, 1 km) than that of AVHRR data‘ Change sentence and wording ‚enjoyed popularity‘. There are multiple reasons why MODIS products are often used. Furthermore, there is no advantage if you use the 1km MOD11A1/MYD11A1.

Mention for example in addition higher resolution satellites such as VIIRS , Sentinel-1, Sentinel-2

Response: the sentences has been changed to “Visible spectrum sensors have also been widely used for monitoring lake ice dynamics, e.g., the Advanced Very High Resolution Radiometer (AVHRR, 1.1 km spatial resolution), the Moderate Resolution Imaging Spectroradiometer (MODIS), with different spatial resolutions (250 m, 500 m, 1 km), Visible Infrared Imaging Radiometer Suite (VIIRS), Sentinel-1 and Sentinel-2 with a higher resolution.” (lines 60-62)

 

2 ‚lake ice phenology ,’ Space error.

Response: it has been removed.

 

3 ‚However, the spatial and temporal resolutions of optical sensing data are strongly limited by cloud obstruction, especially for some regions with strong convection, such as Tibetan plateau, where the higher cloud rates during the day compare with the night, with a result of diurnal convection [28].‘ Reformulate sentence.

Response: it has been changed to “However, the optical sensing data are strongly limited by cloud obstruction, especially in regions with strong convection in the lower atmosphere, like the Tibetan plateau [25].” (lines 66-67)

 

4‚MODIS daily temperature products have a unique advantage than other products that could provide two observations during the day and two observations during the night., which could increase the amount of available data.‘ Punctation marks error.

Response: it has been removed.

 

5 ‚Currently, various methods have been developed to reduce the cloud problem and increase data availability, such as multiple sensor data or temporal-spatial combinations [29-31].‘ Change to ‚reduce and fill data gaps due to cloud coverage, such as multi sensor data products … ‚ In addition, the cited papers are rather old, currently‘ implies newer published approaches which are also available.

Response: the sentences has been changed and the new papers has been cited as listed below:

Hall, D. K., Riggs, G. A., DiGirolamo, N. E., and Román, M. O.: Evaluation of MODIS and VIIRS cloud-gap-filled snow-cover products for production of an Earth science data record, Hydrol. Earth Syst. Sci., 23, 5227–5241, https://doi.org/10.5194/hess-23-5227-2019, 2019

Li, X., Jing, Y., Shen, H., and Zhang, L.: The recent developments in cloud removal approaches of MODIS snow cover product, Hydrol. Earth Syst. Sci., 23, 2401–2416, https://doi.org/10.5194/hess-23-2401-2019, 2019.

 

6.‚utilized binary thresholds‘ Please reformulate ‚utilized thresholds for binary classification‘. 

Response: it has been changed.

 

7.‚time series of averaging near-infrared reflectance and surface temperature values over all lake pixels [18]‘ Not correct, [18] used only near-infrared reflectance values. The combined averaging of near-infrared and surface temperature was the contribution (development) of [20].

Response: the sentence has been changes to “...such as the threshold obtained from near-infrared reflectance [34], or combined averaging near-infrared and surface temperature values [35, 36.” (lines 79-82 )

 

8‚The efficiency of a threshold method may vary among different climatic regimes or different remote sensing data sources‘ Please differentiate between fixed thresholds (applied for multiple lakes) and thresholds derived from the lakes itself. The later is not dependent on specific climate regimes.

Response: as we referred to the efficiency of fixed threshold here, the sentence has been change to “The efficiency of those fixed threshold method may vary among different climatic regimes or different remote sensing data sources”

 

9‚Based on the idea of the time-dependent parameterization of the freezing/thawing process, we propose a new automated approach to extract lake ice phenology from MODIS daily temperature products based on the time-dependent parameterization of the freezing/thawing process. ‘Check sentence.

Response: it has been changed as “In this study, we propose a new automated extraction approach using MODIS daily temperature products based on the idea of the time-dependent parameterization of the freezing/thawing process.” (lines 88-90)

 

10.In our study, the approach is validated using data from three seasonally ice-covered lakes‘  Check Grammar and see discussion validation section.

Response: it has been changed as “In our study, the approach is developed from three seasonally ice-covered lakes”. The modification of validation will be responded in corresponding comments.

 

11.in the ice phenology are evaluated‚ Replace ‚‘evaluated‘ with ‚analysed‘.

Response: it has been changed to “The spatiotemporal changes observed in the ice phenology are analyzed.”

 

Study Area:

Table 1: Add sources of lake characteristic information listed in the table. 

Response: the sources of lake characteristic information has been cited in Table 1 caption.

 

Data:

1‚The MODIS mounted‚ Replace with ‚the ‚MODIS sensor mounted‘

Response: it has been added.

 

2‚at approximately 10:30 and 22:30 local time from Terra and at approximately 1:30 and 13:30 local time from Aqua‘ Please confirm, is this really the local time, not the euatorial crossing time?

Response: it has been changed to “The daily temperature data are available four times per day: at approximately 10:30 LT(local time) and 22:30 LT Equatorial crossing from Terra and at approximately 1:30 LT and 13:30 LT Equatorial crossing from Aqua.” (lines 138-140)

 

3.Indicate the time period covered by the MOD09GQ as well as MOD10A1 data used in the text (not only in Table A1, first reference for Table A1 is in section 4.2).

Response: the time period has been added, “The MODIS daily snow product (MOD10A1, version 6) from 2002 to 2016 was used to validate the change pattern of the lake ice cover from the MODIS daily temperature product.” (lines 143-145)

And “The cloud-free images from MODIS Daily Reflectance products and Landsat images from 2002 to 2016 were selected to validate the lake water and ice cover information.” (lines 158-160)

 

4.Why did you prefer to use MOD09GQ and MOD10A1 and not  MYD09GQ and MYD10A1 or both as combined product? I would see a clear advantage of using combined Terra and Aqua products for this comparison.

Response: in our study, we only employed MOD10A1 from version 6, as the new version of snow cover product added several new features in Aqua, “In this study, we used the snow cover data from Terra only, since the cloud/snow discrimination errors in Aqua cloud-mask algorithm compared with Terra [30, 47]. Also, the snow cover data from Aqua in version 6 used band 6 to replace band 7 in the NDSI calculation which is different from the previous version. More details about the MODIS snow products can be found in [30, 46, 47]”. (lines 153-157)

 

5. Regarding the MOD10A1: Which NDSI threshold was used for the snow/no snow classification? This is important to know.

Response: this information has been added in line 166-168: “...are derived from SNOWMAP algorithm that automatically uses the normalized difference snow index (NDSI) and decision strategies to identify snow cover [33, 46]. The pixel is labeled as snow when NDSI ≥0.4, as recommended by Riggs, G. A., et al. [47].” (lines 147-150)

 

6. Add here a short paragraph describing the AMSR-E/2 data product used for the comparison.

Response: The data source and study period of AMSR-E/2 and MODIS reflectance products and other publishing datasets were added in Table 4 in section 4.3, where extracted ice phenology dates in our study compared with other datasets introduced.

 

7. Add air temperature data used.

Response: the information has been added in lines 187-190 : “The air temperature is collected from 2002 to 2016 from the China Meteorological Administration (referred to as CMA) dataset which has been checked by thorough quality control [48].” (lines 165-167)

 

Method:

Using four MODIS temperature products per day is certainly a benefit to reduce missing information caused by clouds. However, variations in the composite (MODIS_merge) due to the diurnal temperature cycle of lake and even more day-night differences have certainly an impact on the detection of lake ice phenology events. These impacts are either mentioned in the method section nor in the discussion section, but are important to mention and discuss to explain uncertainties in the methods and likely deviations in the results. A merged product containing day and night temperature values the pixels will contain a large range of LSWT values which reduces the accuracy of the final extracted lake ich phenology dates.

Response: thank for the reviewer’s suggestion. We have added the discussion about the limitations caused by the diurnal temperature difference during the combination of Terra and Aqua data in section 5.3 in discussion, “Additional uncertainties are caused by the effect of diurnal temperature variations when combined Terra and Aqua data. The difference between day and night temperatures becomes especially complex during freeze-up and break-up periods, since MODIS temperature referred to ice temperature rather than water surface temperature when the lake is frozen. Hence, additional temperature validation may be required during the transition phase between ice and open water.” Lines(509-513)

Method 3.2.

1. a), b) How were the thresholds 0.5°C and -0.5°C respectively defined? Its not clear until the sensitivity study (which also suggest slightly different thresholds. Please clarify and describe in the text.

Response: the information has been reorganized “The water/ice status in MOD11_Merge image was classified by a temperature threshold initially on a pixel-by-pixel basis. The temperature thresholds were ±0.5 °C here and their sensitivity analysis was discussed in section 5.1.”. (lines 202-205).

The reason for the thresholds 0.5°C and -0.5°C was detailed explained in Discussion 5.1, “The temperature classification applied for the water/ice status should consider the lake surface situation in the freeze-up pattern and break-up pattern, respectively. The freezing temperature can be influenced by several factors. Water salinity can decrease the freezing temperature and delay ice formation [52]. Moreover, the ice pixel temperature can be influenced by the bubble contents, impurities, and fracture patterns of ice surfaces [53]. The melting temperature should consider the ice surface condition. Ice break-up starts after the ice/snow surface is warmed up to the melting point [54]. Melting snow, snow-ice and black ice have different reflectance values [55, 56]. What’s more, the day and night melt-freeze events produce meltwater on the ice surface before the ice cover break-up completely. To take into account these uncertainties, and physically basis of ice cover, we choose the classified pixel as “ice” when its temperature was lower than -0.5°C and classified pixel as “water” when its temperature was higher than 0.5°C.” (lines 445-456)

 

2. How do the authors deal with pixels having LSWT temperatures > -0.5 and < 0.5?

Response: Those pixels would be regard as “Pixel_{ice-water-mixed}”, as this temperature ranges still need to further investigate the state of pixel. Considering this, those pixels will not involve into the unfrozen water cover fraction calculation.

 

3‚Afterwards, the adjacent temporal filter ‚ Please clarify what an adjacent temporal filter is.

Response: the definition of adjacent temporal in our study is referred to “Afterward, the adjacent temporal filter was applied to deduce the continuous stable ice-covered or ice-free pixel from cloud-covered pixels via the same pixel in the previous and subsequent days without reducing and spatial and temporal resolution.” (lines 216-218)

 

4. I don’t understand why your data contain at this point ‚remaining cloud pixels’ which need to be Flagge. Please explain. 

Response: “remaining cloud pixels” referred that the pixels were possibly still polluted when the cloud cover severely exist. We further used adjacent temporal filter to deduce the pixels status from the same pixels in previous and subsequent days.

 

5. And ‚Afterwards, the adjacent temporal filter was applied to recover the information from pixels on the preceding and following days to deduce the water/ice status from cloud-covered pixels. ‚ Please define here the number of preceding and following days i.e. the time window used for the gap filling.

6 ‚temporal window lengths of 5 and 7 days‘ Temporal window length: Is this +/- 5 and 7 days? Or 5 days for preceding anf 7 days for following days? Please clarify. 5 and 7 days seem quite long to me since lake ice phenology freeze-up or break-up phases are fast changing. How did you define the number of days?

Response: as comment #5 and # 6 state the similar issues, we responded as followed: we use same temporal window length for previous day and following day, but we use two different temporal window lengths, 5 days and 7 days, respectively. As adjacent temporal filter was to deduce the status pixel, the relative long temporal window length was reasonable. We further precise the paragraph about adjacent temporal filter, “As the adjacent temporal filter was mainly used for stable status pixel deduction in our study, the two temporal windows of 5 days and 7 days were chosen, respectively.”( lines 219-220)

 

7. Did you do any weightening applied for the gap filling? 

Response: the gap filling is involved into the curve fitting.

 

8. ‚contained abnormal unfrozen water cover fraction values‘  Wording ‚abnormal‘, please repplace (also ff).

Response: the “abnormal unfrozen water cover fraction values” has been changes to “the outliers in unfrozen water cover fraction”.

 

9. ‚preceding the date when the mean air temperature in autumn dropped below 0°C ‚ Missing data source and e.g. spatial and temporal resolution for air temperature used. Add to the data section.

Response: the data source has been added in section 2.2.2 “The air temperature is collected from 2002 to 2016 from the China Meteorological Administration (referred to as CMA) dataset which has been checked by thorough quality control [48]” (lines 165-167).

 

10‚The latter date was‘ What is latter referreing too? Please clarify.

Response: To be precisely understood, this paragraph has been merge into the previous one, “The outliers in unfrozen water cover fraction were mainly caused by the high percentage of cloud cover, as shown in Figure 3. The outliers were eliminated by the following steps. First, the unfrozen water cover fraction was set to 1 for dates before the mean air temperatures in autumn dropped below 0°C. The mean air temperature values were preliminarily smoothed by a 31-day running average to remove synoptic variability [51]. Second, the unfrozen water cover fraction was removed if the percentage of “invalid” pixels in an image was larger than 80%. After the outlier removal, the temporal pattern of the unfrozen water cover fraction (increase or decrease) was applied to describe the water/ice transition process and then extract lake ice phenology.” (Lines 221-229)

 

11‚Second, the unfrozen water cover fraction was removed if the percentage of “invalid” pixels in an image was larger than 80%.‘ Clarify removed. Does it mean you obtained a data gap on that specific day?

Response: the percentage of “invalid” pixels in an image was larger than 80% would be regarded as the outlier that need to remove. Those “invalid” pixels were mainly caused by severe cloud pollution, and the gap caused would not temporal filled.

 

12‚was employed to evaluate lake ice phenology’ Word choice ‚evaluate ‘. Please change.

Response: the sentence has been changed as “After the outlier removal, the temporal pattern of the unfrozen water cover fraction (increase or decrease) was applied to describe the water/ice transition process and then extract lake ice phenology.”(lines 227-229)

 

13 Figure 3. Please explain ‚original ‘

Response: the explanation has been added in figure 3 caption “The blue dot line represents the unfrozen water cover fraction without outlier removal, and the red dot line represents the unfrozen water cover fraction after outliner removal.”

 

Method 3.3

1.Please discuss how the curve fitting accounts for multiple break ups and freeze ups in the respective phases.

Response: the information about the multiple break ups and freeze ups has been added “The logistic function was applied by fitting and  separately to the freeze-up and break-up periods in each hydrologic year.” (line 252)

 

2.‚where ? is the days of the hydrological year, ? is the unfrozen water cover fraction value at time ?; ? is the maximum value of the unfrozen water cover fraction, which is equal to 1 in our study‘ Please change order accordingly to formula.

Response: It has been changed

 

3.Figure 4: Low quality of figure (blurred). Please replace.

Response: it has been replaced.

 

Results: 

Results 4.1.

1. Note, very extensive description, which could be shortened.

Response: the note was intended to explain how to obtain valid pixel fraction and valid day fraction in Table 2. This part has been precisely added into the context, which can be found in lines 227-280: “The valid pixel fraction calculated by the number of valid pixels in the lake mask, and the valid day fraction calculated by the number of the day that image was covered by 80% valid pixel, were jointly analyzed to evaluate the reduction in cloud contamination (Table 2)..”

 

2‚Overall, after cloud removal by daily combination, it has been shown obvious increases in the number of valid pixels in MOD11_Merge images and the number of images that represented “valid” pixels reaching more than 80% of the whole lake area during the 2002-2016 period. ‘ Check sentence Grammar.

Response: the sentence has been changed to “Overall, after cloud removal by daily combination, it has been shown obvious increases in the valid pixel fraction and the valid day fraction in MOD11_Merge images.”

 

3.‚negative cloud effects’Wording. Find a appropriate replacement.

Response: the sentences has been change to “the proposed approach effectively increases data availability”

 

Results 4.2.

1.This is not an independent validation if it is the same sensor.  Thus I would highly suggest to call it comparison. What is needed is a independent validation from a higher resolution sensor (such as Landsat, Sentinel-2 or insitu measurements).

As indicated earlier: These comparison misses details such as the time period where a comparison has been conducted. This accounts for both MOD09GQ as well as MOD10A1. Its first mentioned in 4.3 which is by far too late.

Response: Thanks for the reviewer’s suggestion. The information of data used for the comparison has been added in section 2.2.2, including time periods, the algorithm used in MODIS snow products, and also the basic information of MOD09GQ and Landsat data.

The Landsat images with a spatial resolution of 30 m were added to provide an independent validation. The detail comparison has been described in section 4.2, which could find in line 308-317, “The visual inspections of lake ice under clear-day conditions from the MODIS daily reflectance products and Landsat images were further used as a reference. The consistency of unfrozen water cover fraction in the phases of break-up process was found among MODIS reflectance products, MODIS snow products, and MOD11_Merge temperature images, as shown in Figure A2-A4. This consistency was notably seen in the developed and final phases of ice cover than in the initial phase, especially in smaller Lake Ngoring (Figure A4). Thin ice When Compared with Landsat images with a high resolution of 30 m, thin ice or fast ice at the lakeshores was hardly detectable from the MODIS products, either from snow products and temperature products, as shown in Figure A5-A7. One of the reasons is the misclassification of thin ice due to thin cloud cover [23].”

 

2. Figure A3 and A4. Please add label for the MOD11_Merge Temperature product. In addition a label for clouds would be helpful for the readers.

Responses: it has been added.

 

Results 4.3

1.Trend calculation misses significance niveaus (see also 4.4)

Response: the related sentences has been added in lines 369-371: “Among the nine lakes, only Lake Xinkai had a significant break-up trend. The short length of data record (15 years) may be one of reasons that few lakes have significance trends.”

 

2.Overall, the interannual variations in ice phenology showed good agreement be-tween our study and existing studies, except for inconsistent trends in the freeze-up dates‘ Please Elaborate what you mean with ‚inconsistent trends‘. 

Response: the “inconsistent trend” were replaced by the sentences “.But the trends of freeze-up dates showed inconsistency. For example, the trend of FUS in Lake Qinghai was -4.09 days/10 yr while it was 4.00 days/10 yr from Cai, Y., et al. [15] and 6.31 days/10 yr in our study” (lines 339-341)

 

Results 4.4.

1.Consistent use of FUS, FUE, BUS. BUE

Response: the items has been changed.

 

2‚When validated on three different types of lakes, the proposed lake ice phenology extraction approach has proven its robustness when avoiding lake-specific thresholds.‘ Im not convinced about the robustes of the method since it lacks a rigoros validation as mentioned earlier.

Response: the independent validation from the higher spatial resolution images has been added. This part has been found detailed in section 4.2 and the previous response.

 

3.Table 5: Please reformate table header. It is hard to read.

Response: it has been changed.

 

Discussion

The discussion section needs to be reworked to clearly discuss the advantage and disadvantages of the method, its impacts (e.g. clouds, composite effects of day and night LSWT values, etc). 

A thorough discussion of the to adress the impacts and its effects on the results of the proposed method is missing.

 Example: Cloud/ice discrimination and/or Cloud masking. Discuss uncertainties in cloud masking, quality during night time.

Discuss possible application (e.g. needed requirements such as lake size) to apply you method to other lakes.

Response: we thank reviewer for suggestions. The advantages and limitation of the algorithm has been discussed in section 5.3: “ The proposed algorithm of ice phenology extraction possesses several distinct advantages on the wide-scale application on seasonal ice-covered lakes. An important qualitative advance is to scale the transition period between the open water and the ice-covered stage by a parameterized function of time (Eq. 3). The scaling-based method has proven to be robust to outliers and applied to lakes with various morphometry. Also, the characterization of the seasonal ice phenology can be extended beyond the binary ice-on/ice-off events by the direct assessment of the “partial ice cover” duration. The latter characteristic is crucial for understanding the lake response to the ice cover formation, especially in medium-to-large lakes. During the ice formation in autumn, the length of the transitional period determines the rate of cooling across the entire water column: A faster formation of the complete ice cover allows more heat to be stored in the near-bottom part of the lake, affecting the water-ice heat fluxes during the entire ice-covered period and the ice thickness [62, 67, 68]. In turn, the length of the transition period to the ice-free state in spring causes long-lasting effects on the vertical thermal stratification in the following summer and affects thereby the crucial lake water quality characteristics, such as the dissolved oxygen content [62, 68]. Both autumn and spring transitional periods are the results of interaction between the air-lake heat transport, the available solar radiation, and the heat storage in the lake water column. Therefore, they serve as sensitive indicators of the regional response to global climate change. When applied to long-term remote sensing observations, the proposed approach would allow tracing the climatically driven effects on non-linear interactions between various components of the surface heat balance during autumn and spring. The proposed approach also builds on the strengths of the MODIS daily temperature products. The Terra and Aqua daily combination from four times observation with a time step of 4 to 7 hours demonstrates the effectiveness in reducing the cloud contamination while retaining the essential spatial information. The automatic temperature-based algorithm also shown the ability to correctly extract ice phenology on different lake types concerning potential influences caused by salinity, morphometry, and geographic location.

The limitations of the proposed algorithm are intrinsic for remote sensing products. While the data availability in MOD11_Merge images has been improved by Terra and Aqua daily combination, the data gaps remain when severe cloud contaminations exist. Additional uncertainties are caused by the effect of diurnal temperature variations when combined Terra and Aqua data. The difference between day and night temperatures becomes especially complex during freeze-up and break-up periods, since MODIS temperature referred to ice temperature rather than water surface temperature when the lake is frozen. Hence, additional temperature validation may be required during the transition phase between ice and open water. Finally, although the spatial resolution of 1 km from MODIS temperature products is relatively high compared to other remote sensing products, like AMSR-E, it is still coarse for phenology studies on lakes with horizontal dimensions of less than several kilometers, especially located at high elevations or in areas covered by persistent snow cover. A progress in phenology observations on small lakes may be envisaged with obtaining sub-kilometer spatial resolution while retaining sub-daily temporal resolution of data. “ (lines 474-516)

”

Discussion 5.1.

1. Starts with a 1/3 page Literature review. These basics are important, but should be part of the introduction or shortened.

Response: the literature review has been shortened and also removed to the end of this section, which can be found in lines 442-453 “The temperature classification applied for the water/ice status should consider the lake surface situation in the freeze-up pattern and break-up pattern, respectively. The freezing temperature can be influenced by several factors. Water salinity can decrease the freezing temperature and delay ice formation [57]. Moreover, the ice pixel temperature can be influenced by the bubble contents, impurities, and fracture patterns of ice surfaces [58]. The melting temperature should consider the ice surface condition. Ice break-up starts after the ice/snow surface is warmed up to the melting point [59]. Melting snow, snow-ice and black ice have different reflectance values [60, 61]. What’s more, the day and night melt-freeze events produce meltwater on the ice surface before the ice cover break-up completely.”

 

The sensitivity study is of high interest. A big benefit and added value for this study as well as the scientific community would be to use the sensitivity study (of the discussion section) to derive uncertainties for the final lake ice phenology product.

Discussion 5.2.

Nice and important section. However, please explain how you obtained those results. It only says in the Figure caption that you performed a correlation. And Missing data source of Weather stations used. 

Response: the related information is added in lines 390-394: “the correlations between ice phenology time series extracted in our study and possible influencing factors including lake specific factors and major climatic factors were analyzed. The climatic variables, including air temperature, wind speed, solar radiation, and snow depth, were obtained from 2002 to 2016 from the China Meteorological Adminstration dataset based on the nearest weather station.”

 

Conclusions

‚We developed a new automatic algorithm to characterize lake ice phenology using MODIS daily temperature products from 2002 to 2016 by calculating time series of unfrozen water cover fractions for freeze-up/break-up periods and then fitted by a quasi-logistic function capturing the temporal pattern of the transitional water/ice phase and ex-tracting the lake ice phenology characteristics.‘ Check Sentence length.

Response: It has been shortened to “we developed a new automatic method to characterize lake ice phenology using MODIS daily temperature products from 2002 to 2016 by parameterizing the temporal pattern of water/ice transition process”

 

2.‚allowing us to trace the ice phenology on a more detailed level than simple binary ice-on/ice-off events‘ Please Elaborate.

Response: it has been changed to “allowing us to trace the ice phenology on a more detailed level than traditional binary classification.”

 

3. Summarize main advantages and disadvantages of your developed method. Add quantification of your results (i.e. trends) and its accuracy / comparison.

Response: the quantification of the results was written in lines 522-527: “The results showed that Lake Qinghai had the most obvious trend to delayed freeze up start date (FUS) (6.31 days/10 yr) among the lakes of Tibetan plateau, and Lake Hulun the obvious tendency to earlier break up start and end dates (BUS and BUE) with the change rate of -3.73 days/10 yr and -5.02 days/10 yr, respectively”

Further, we descripted the advantages and implication of the algorithm in line 527-537: ”The proposed method has two major advantages, making it robust and applicable: (i) it avoids any lake-specific or empirical thresholds, allowing the effective treatment of data gaps caused by atmospheric conditions, and (ii) it utilizes a time-dependent parameterized function to characterize the water/ice transition process, allowing us to trace the ice phenology on a detailed level than traditional binary classification. The algorithm could provide the data support for the correct interpretation of the global-scale climate change effects on lake ice phenology: apart from the general shortening of the ice-cover period, the interplay between atmospheric warming and seasonally available shortwave solar radiation may produce significant changes in the duration of transitional—partially ice-covered—periods with important consequences for air-lake interactions and internal lake mixing.”

 

Author Response File: Author Response.docx

Reviewer 2 Report

The authors have developed an automated system for identifiying lake ice phenology that utilizes MODIS imagery.  The process is well described and generates good results. Specific comments, mainly questions are embedded in the attached PDF.  There is need for more specific reporting on typcial climate variable values  during the freeze up and break up period early in the paper.  Is there a salinity trend in any of these lakes that is impacting ice formation or loss?  It would be of value in section 2.1 to briefly describe the geographic progression of freezeup or breakup in each lake.  Does it begin on the western shore, or in the shallowest sections etc. 

For future reviews please make sure to have line numbers in your template, this makes reviewing and responding to reviews much easier.

Author Response

Reviewer#2

Comments and Suggestions for Authors

The authors have developed an automated system for identifying lake ice phenology that utilizes MODIS imagery.  The process is well described and generates good results. Specific comments, mainly questions are embedded in the attached PDF.  There is need for more specific reporting on typical climate variable values during the freeze up and break up period early in the paper.  Is there a salinity trend in any of these lakes that is impacting ice formation or loss?  It would be of value in section 2.1 to briefly describe the geographic progression of freeze up or breakup in each lake.  Does it begin on the western shore, or in the shallowest sections etc. For future reviews please make sure to have line numbers in your template, this makes reviewing and responding to reviews much easier.

Response: Thanks for the reviewer’s suggestion. Firstly, the line number has been added. Secondly, the uncertainty caused by salinity was considered when we analyst sensitivity of classification of water/ice pixel, which could be found in section 5.2. Finally, the geographic progression of freeze-up and break-up process have been added, which could be found in section 2.1 in lines 111-120:  “Lake Qinghai is a large brackish dimictic lake located on the northeastern Tibetan Plateau. It is the China’s largest lake inland lake and is subject to a typical semiarid continental climate characterized by warm summers, cold winters, and more precipitation in summer than in other seasons. The lake starts to freeze part from east coast to northeastern and northwestern parts, and the melting part starts from northeastern and northwestern coast and gradually toward the center [19]. Lake Ngoring is the highest (~4200 m asl) freshwater lake in China and is located in the source region of the Yellow River, surrounded by hills covered with alpine meadows. The south and east shoreline area freeze and melt firstly. Lake Hulun is a shallow monomictic lake located in the eastern part of the Mongolian Plateau in the northern part of China and is affected by semiarid continental and monsoon climates. The lake freezes from the complex shoreline area of lake, and melting starts from northwest coast and gradually toward the east coast [42].”

 

Author Response File: Author Response.docx

Reviewer 3 Report

Dear Editor,

In this study, Zhang et al. propose a novel algorithm that employs data from the Moderate Resolution Imaging Spectroradiometer (MODIS) to quantify the ice coverage over lakes located in mountainous regions and harsh environments where in-situ measurements are not sufficient to achieve this goal.

The presented algorithm employs daily temperatures estimated by MODIS provided at a high temporal (four observations per day) and spatial resolution (500 meters).

Thanks to high resolution and observations’ repeat cycle, the presented algorithm returns a time series in which the effect of noise introduced by cloud coverage is minimized, allowing the complete characterization of the seasonal variability of ice coverage (lake ice phenology).

In the second part of the study, the authors employ the presented approach to study the variability of lake ice phenology and its relation with several climatic variables at nine locations within the Chinese territory.

In my opinion, the study is of technical interest and fits within the scope of Remote Sensing.  However, I can’t recommend this manuscript for publication, at least in the present form.  I will provide my main observations below:

• Considering the technical goal of this manuscript, the authors should describe the proposed algorithm more clearly and in further detail.
• The authors should also improve the English used in this manuscript that, in my opinion, is not suitable for publication in an international journal.

 

I will be happy to recommend this manuscript for publication once the authors have improved the manuscript and addressed my comments.

Author Response

Reviewer#3

Dear Editor,

In this study, Zhang et al. propose a novel algorithm that employs data from the Moderate Resolution Imaging Spectroradiometer (MODIS) to quantify the ice coverage over lakes located in mountainous regions and harsh environments where in-situ measurements are not sufficient to achieve this goal.

The presented algorithm employs daily temperatures estimated by MODIS provided at a high temporal (four observations per day) and spatial resolution (500 meters).Thanks to high resolution and observations’ repeat cycle, the presented algorithm returns a time series in which the effect of noise introduced by cloud coverage is minimized, allowing the complete characterization of the seasonal variability of ice coverage (lake ice phenology). In the second part of the study, the authors employ the presented approach to study the variability of lake ice phenology and its relation with several climatic variables at nine locations within the Chinese territory.

In my opinion, the study is of technical interest and fits within the scope of Remote Sensing.  However, I can’t recommend this manuscript for publication, at least in the present form.  I will provide my main observations below: Considering the technical goal of this manuscript, the authors should describe the proposed algorithm more clearly and in further detail. The authors should also improve the English used in this manuscript that, in my opinion, is not suitable for publication in an international journal.

I will be happy to recommend this manuscript for publication once the authors have improved the manuscript and addressed my comments.

 

Response: Thanks for the reviewer’s suggestion. This paper is to develop a robust and applicable method to extract lake ice phenology automatically based on MODIS daily temperature products. Three seasonally ice-covered lakes with different characteristics in different climate regions were selected to test the applicability of method during the period of 2002- 2016. The annual variation of extracted ice phenology dates showed good consistency with independent data sources. The proposed method can be suitable for estimating and monitoring ice phenology on different types of lakes over large scales and has a strong potential to provide valuable information on the responses of ice processes to climate change.

In the revised version, there were three main aspects we improved and added. Firstly, we reorganized the result section. we added the validation of our result by the Landsat images with a high spatial resolution of 30 m, as a independent data source from MODIS products The unfrozen water cover fraction was firstly compared with MODIS snow products with a spatial resolution of 500m, and further checked by Landsat images with a spatial resolution of 30 m and MODIS reflectance product with a spatial resolution of 500m. The extracted ice phenology dates from this algorithm were further compared with other lake ice datasets, such as AMSE-R/2, that annual variation of ice phenology dates showed the good consistency with our results. Secondly, we reframed the discussion section. The advantages and limitation of the algorithm were thoroughly addressed in section 5.3. Thirdly, the detailed information has been added in data and method sections, including the algorithm to extract ice from MODIS snow products, the adjacent temporal filter. we also have carefully checked the grammar and improved the writing with the help of language check.

It is our belief that the manuscript is substantially improved after making the suggestion edits. We will appreciate your consideration.

 

Round 2

Reviewer 1 Report

Dear Authors,

I appreciate the work and time investment to improve the manuscript. The manuscript improved overall by far. 

A small list of comments is added, which requires minor revision. 

However, I still have two major issues.

- For me the neglection of the diurnal LSWT cycle (combing day and night time data) is still a large concern, which is now at least very briefly mentioned but the impact is not thoroughly discussed. The impact due to the combined day and night temperature e.g. in terms of temperature differences for the specific lakes should have been retrieved and discussed.

- In addition, the publication still misses a thorough validation with higher resolution sensors (apart from MODIS products). The visual checks using 6 Landsat images are poor, but I recognize your efforts in the limited time. For me, the method is still lacking a robust, independent validation. Moreover, since you applied your proposed method to the MODIS Snow Products for your 'validation'. This lacks independency. I highly recommend that you rethink the term ‚validation’ and replace it with a more appropriate such as ‚comparison‘ or ‚evaluation‘. Furthermore I suggest that you write in a clear and way more careful word choice about the comparison/evaluation and its results. 

 

Abstract

- ‚Previous satellite-based lake ice detec- tion algorithms used mainly empirical or lake-specific thresholds.' Delete in abstract since nowadays several methods exist to derive lake ice phenology dates without empirical or lake-specific thresholds.

- Please mention the Landsat validation.

- Please add the significance niveau to the described results.

 

1. Intro

- ‚The latter method was found to effectively improve data availability by approximately 10%-20% [31-33].‘ Please be more precise, it is not a single method, e.g. ‚Approaches using temporal-spatial combinations were found…‘

- ‚or combined averaging near-infrared and surface temperature values [35, 36].‘  Remove ‚averaging’ since ‚combined averaging’ is missleading.

- ‚An alternative approach is to use a spatially resolved threshold distribution, e.g., 5% and 95% of a lake area.‘ This ist not the right citation source, it has been earlier implemented by e.g. [34, 36]. 

- Again: ‚The efficiency of those fixed threshold methods may vary among different climatic regimes or different remote sensing data sources.‘  Not all of the listed methods/references used ‚fixed thresholds‘. Rewrite, e.g.: ‚While fixed threshold methods may be limited to different climatic regimes, dynamic threshold approaches, as listet above, are better suited across differenct climatic regimes and remote sensing data sources’.

- ‚pattern of water /ice transition‘ Space error. 

 

2.1 Study area

- ‚Nine large annual ice-covered lakes are chosen for validation and application of the method distributed across northern China (Figure 1 and Table 1)‘ Please write more precise e.g., ‚Nine large annual ice-covered lakes were chosen to develop, evaluate and  and test the proposed approach, distributed …‘

- Please add reference for the description of Lake Ngoring. 

- The basic information of the selected lakes[43]. Space error.

 

2.2.1 

- ‚(https://modis.gsfc.nasa.gov/data/dataprod/mod11.php).‘ Check font size, see also Links in 2.2 and ff.

- ‚at approximately 10:30 LT (local time) and 22:30 LT Equatorial crossing from Terra and at approximately 1:30 LT and 13:30 LT Equatorial crossing from Aqua.‘ Please check ‚(local time)‘, while the others are Equatorial Crossing Time? Add ‚Time‘ after ‚Equatorial Crossing‘.

 

2.2.2

- ‚~93% for distinguishing land cover from snow [46].‘ Replace with ‚snow-free land from snow cover.‘

- ‚In this study, we used the snow cover data from Terra only, since the cloud/snow discrimination errors in Aqua cloud-mask algorithm compared with Terra [30, 47].‘ Check sentence.

- ‚The Landsat 5,7 and 8' Add comma ‚5,7, and‘.

- ‚resolution of 16 days, and available from‘ Replace with ‚‘and are available’.

- Mention the time period and number of scenes of Landsat data.

 

4 Results

4.1

- Correct Table 2 layout

 

4.2

- see overall comment: For me, the method is still lacking a robust, independent validation. I highly recommend that you rethink the term ‚validation’ and replace it with a more appropriate such as ‚comparison‘ or ‚evaluation‘. Furthermore I suggest that you write in a clear and way more careful word choice about the comparison/evaluation and its results.

- 'The performance of the unfrozen water cover fraction obtained from MOD11_Merge daily images was validated against MODIS daily snow products, MODIS daily reflectance products, and Landsat images.' See comment to 'validation' above and change accordingly.

- ‚The comparison patterns in break-up period among the three lakes were shown a higher consistency compared with freeze-up period,‘ Check sentence grammar.

- ‚This may be due to the thin ice or initial ice information was not easily detectable, especially in large lakes [14, 15].‘ Check sentence Grammar.

- ‚When Compared with Landsat images‘ Please correct capital letter.

- Table 4.3 Please write Break-up Season, reformat layout

- Figure 5: At this scale, differences are almost impossible to see. Could you add as add. example a single year only?

 

4.3.

- Correct references (Qiu, Y., et al. [55]), (Cai, Y., et al. [15]), Cai, Y., et al. [15] 

 

4.4.

- Mention the significance niveau/level of the trends directly when writing e.g. 6.31 day/10 yr. The authors need to write the significance level for the derived trends as e.g. in 5.1

- ‚The short length of data record (15 years) may be one of reasons that few lakes have significance trends.”‘ Replace with: ‚—why the other lakes show no significant trends.‘ One is not few.

- Figure caption 7 ‚ice cover.Table 6.‘ Add space. 

 

 5.1.

‚The growth and decay of lake ice cover are governed by thermodynamic interactions between atmospheric-water-ice interfaces and are affected by climatic and lake-specific factors. The major climatic factors controlling the heat exchange with the atmosphere are air temperature, solar radiation, precipitation, and wind speed. Temperature changes in water and air are regarded as the most prominent factors affecting lake ice cover [5, 8]. The amount of solar radiation and snow accumulation significantly affect the ice cover break-up process [10]. Lake-specific factors, such as lake morphometry (depth, area, and volume) and location (longitude, latitude, and altitude), determine heat storage in lakes and additionally affect lake ice formation [57].‘ Please move to the introduction part. It does not belong to the discussion. 

 

5.2.

- ‚The melting temperature should consider the ice surface condition.‘ Repetition to paragraph sentence one, please delete. 

- Figure 10. The caption missed the description for the different panels (e.g. please repeat K_b etc).

 

5.3.

‚The automatic temperature-based algorithm also shown the ability‘ Sentence grammar.

6

- add significance niveaus/levels to trends

Author Response

Reviewer #1

Comment: Dear Authors,

I appreciate the work and time investment to improve the manuscript. The manuscript improved overall by far. A small list of comments is added, which requires minor revision. However, I still have two major issues.

1. For me the neglection of the diurnal LSWT cycle (combing day and night time data) is still a large concern, which is now at least very briefly mentioned but the impact is not thoroughly discussed. The impact due to the combined day and night temperature e.g. in terms of temperature differences for the specific lakes should have been retrieved and discussed.

Response: we sincerely thank the reviewer’s suggestion. the additional uncertainties would be caused by the day and night temperature combination from MODIS data. The validation of this part would be the main topic in future work, before the application of this approached in a large scale. currently, several sentences have been added into the discussion, as seen in “Firstly, the thin ice cover was formed when lake surface temperature approaches the freezing temperature, but it may be repeatedly destroyed during windy periods with prolonged fluctuation of the surface temperature around 0°C [63, 69]. Therefore, the thin ice is difficult to detect by the temperature-based methods. The results also showed the thin ice layer was not easier detected by the proposed methods. Secondly, the severe cloud contaminations caused inevitable data gaps, although the data availability have been improved by Terra and Aqua daily combination in our study. Also, MODIS-derived LSWT was revealed a negative bias compared with in situ observations, especially during the daytime. The negative bias was possibly caused by several reasons, such as the undetected cloud cover [46, 51], cool skin and warm layer effects [70-72]. Hence, additional uncertainties are caused by the effect of diurnal temperature variations when combined Terra and Aqua data. Furthermore, the differences between day and night temperature become especially complex during freeze-up and break-up periods, since MODIS-derived temperature referred to the ice surface temperature rather than water surface temperature when the lake is frozen. Hence, additional temperature validation is required to classify the water/ice status during the transition phases between ice and open water in the range between -1 °C and 1 °C.” (lines 509-523).

 

Comment: 2 In addition, the publication still misses a thorough validation with higher resolution sensors (apart from MODIS products). The visual checks using 6 Landsat images are poor, but I recognize your efforts in the limited time. For me, the method is still lacking a robust, independent validation. Moreover, since you applied your proposed method to the MODIS Snow Products for your 'validation'. This lacks independency. I highly recommend that you rethink the term ‚validation’ and replace it with a more appropriate such as ‚comparison‘ or ‚evaluation‘. Furthermore, I suggest that you write in a clear and way more careful word choice about the comparison/evaluation and its results. 

Response: thanks for the reviewer’s suggestion. As suggested, the title of section 4.2 has been changed to “Comparison of the unfrozen water cover fraction with other datasets”, and the corresponding context has been modified by the word “evaluated” or “compared”.

 

Comment: Abstract 1. Previous satellite-based lake ice detection algorithms used mainly empirical or lake-specific thresholds.' Delete in abstract since nowadays several methods exist to derive lake ice phenology dates without empirical or lake-specific thresholds.

Response: it has been deleted.

Comment: 2. Please mention the Landsat validation.

Response: the sentences have been changed to “The temporal pattern of water/ice transition phase was performed by unfrozen water cover fraction extracted from MODIS daily temperature data, and was compared with MODIS snow and reflectance products and Landsat images.” (lines 25-27)

Comment: 3. Please add the significance niveau to the described results.

Response: the statistical significance of the trend for 9 lakes was illustrated in Table 6 and section 4.3. and the described results has been modified as “The strongest tendency to later freeze up start date was revealed in Lake Qinghai (6.31 days/10 yr) among the lakes in Tibetan plateau, and the break up start and end dates shifted to be earlier rapidly in Lake Hulun (-3.73 days/10 yr; -5.02 days/10 yr).”

 

Comment: Introduction

1. The latter method was found to effectively improve data availability by approximately 10%-20% [31-33].‘ Please be more precise, it is not a single method, e.g. ‚Approaches using temporal-spatial combinations were found…‘

Response: it has been change to “Approaches using temporal-spatial combinations was found to effectively improve data availability by approximately 10%-20% [31-33]”.(lines 69-70)

Comment: 2. or combined averaging near-infrared and surface temperature values [35, 36]. ‘  Remove ‚averaging’ since ‚combined averaging’ is missleading.

Response: it has been removed.

Comment: 3. An alternative approach is to use a spatially resolved threshold distribution, e.g., 5% and 95% of a lake area.‘ This is not the right citation source, it has been earlier implemented by e.g. [34, 36]. 

Response: the citation has been checked and the sentence has been changed to “An alternative approach is to use a spatially resolved threshold distribution, e.g., 5% and 95% of a lake area [22, 23].” (lines 81-84)

Comment: 4. Again: ‚The efficiency of those fixed threshold methods may vary among different climatic regimes or different remote sensing data sources.‘  Not all of the listed methods/references used ‚fixed thresholds‘. Rewrite, e.g.: ‚While fixed threshold methods may be limited to different climatic regimes, dynamic threshold approaches, as listed above, are better suited across different climatic regimes and remote sensing data sources’.

Response: thank you for the suggestion. The sentences have been changed as it suggested.

Comment: 5. pattern of water /ice transition‘ Space error. 

 Response: it has been removed.

Comment:

2.1 Study area

1. Nine large annual ice-covered lakes are chosen for validation and application of the method distributed across northern China (Figure 1 and Table 1)‘ Please write more precise e.g., ‚Nine large annual ice-covered lakes were chosen to develop, evaluate and  and test the proposed approach, distributed …‘

Response: The sentence has been changed to “Nine large seasonally ice-covered lakes with different characteristics, latitudes, altitudes, areas, salinities, and climatic conditions distributed across northern China are chosen to develop, evaluate, and test the proposed approach (Figure 1 and Table 1)” (lines 102-105)

Comment: 2. Please add reference for the description of Lake Ngoring. 

Response: the citation has been update “The south and east shorelines freeze and melt firstly [42]”.

Comment: 3. The basic information of the selected lakes[43]. Space error.

  Response: it has been removed.

 

Comment: 2.2.1 MODIS Daily temperature data

1. (https://modis.gsfc.nasa.gov/data/dataprod/mod11.php).‘ Check font size, see also Links in 2.2 and ff.

 Response: it has been changed.

Comment: 2. at approximately 10:30 LT (local time) and 22:30 LT Equatorial crossing from Terra and at approximately 1:30 LT and 13:30 LT Equatorial crossing from Aqua.‘ Please check ‚(local time)‘, while the others are Equatorial Crossing Time? Add ‚Time‘ after ‚Equatorial Crossing‘.

  Response: it has been added.

 

Comment: 2.2.2 Other MODIS data and Landsat data

1. ~93% for distinguishing land cover from snow [46].‘ Replace with ‚snow-free land from snow cover.‘

 Response: the sentences has been rewritten as “The assessment of MODIS snow cover products have an overall accuracy of ~93% under clear-sky conditions [47]. ”

 

Comment: 2. In this study, we used the snow cover data from Terra only, since the cloud/snow discrimination errors in Aqua cloud-mask algorithm compared with Terra [30, 47].‘ Check sentence.

Response: the sentence has been updated to “ In this study, we used the snow cover data from Terra only, since the cloud/snow discrimination has errors in Aqua cloud-mask algorithm [30, 48]. ”

Comment: 3. The Landsat 5,7 and 8' Add comma ‚5,7, and‘.

Response: it has been added.

Comment: 4. resolution of 16 days, and available from‘ Replace with ‚‘and are available’.

Response: it has been added.

Comment: 5. Mention the time period and number of scenes of Landsat data.

Response: the related information has been added as “Due to the 16-day temporal resolution and the influence of cloud cover, the availability of Landsat images is limited and cloud-free images were preferably selected. In our study, 4 images from Landsat under clear sky condition were chosen in each lake during freeze-up and break-up period, respectively” (lines 164-167)

 

Comment: 4 Results 4.1 Cloud contamination removal by Terra and Aqua combination

- Correct Table 2 layout

 Response: it has been modified.

 

Comment: 4.2 Comparison of the unfrozen water cover fraction with other datasets

1. see overall comment: For me, the method is still lacking a robust, independent validation. I highly recommend that you rethink the term ‚validation’ and replace it with a more appropriate such as ‚comparison ‘or ‚evaluation‘. Furthermore, I suggest that you write in a clear and way more careful word choice about the comparison/evaluation and its results.

Response: we thank the reviewer’s detailed suggestion. As suggested, the title of section 4.2 has been changed to “Comparison of the unfrozen water cover fraction with other datasets”, and the related term has been changed to “comparison” or “evaluation”.

 

Comment: 2. The performance of the unfrozen water cover fraction obtained from MOD11_Merge daily images was validated against MODIS daily snow products, MODIS daily reflectance products, and Landsat images.' See comment to 'validation' above and change accordingly.

Response: the sentence has been changed to “The performance of the unfrozen water cover fraction obtained from MOD11_Merge daily images was compared with that from MODIS daily snow products, MODIS daily reflectance products, and Landsat images.”

Comment: 3. The comparison patterns in break-up period among the three lakes were shown a higher consistency compared with freeze-up period, ‘Check sentence grammar.

Response: the sentence has been changed to “The comparison patterns of the break-up period were shown a higher consistency than the comparisons pattern of the freeze-up period, with the R² values of 0.912, 0.889, and 0.948, respectively.”

Comment: 4. This may be due to the thin ice or initial ice information was not easily detectable, especially in large lakes [14, 15].‘ Check sentence Grammar.

Response: the sentence has been changed to “This may be due to the thin ice or initial ice information that was not easily detectable, especially in large lakes [14, 15].”

Comment: 5. When Compared with Landsat images‘ Please correct capital letter.

Response: it has been corrected.

Comment: 6. Table 4.3 Please write Break-up Season, reformat layout

Response: it has been modified.

Comment: 7. Figure 5: At this scale, differences are almost impossible to see. Could you add as add. example a single year only?

 Response: as suggested, the comparison in single year (e.g. 2011) has been added, seen in Figure 6.

 

Comment: 4.3. Comparison of derived lake ice phenology with other lake ice datasets

- Correct references (Qiu, Y., et al. [55]), (Cai, Y., et al. [15]), Cai, Y., et al. [15] 

 Response: it has been corrected.

 

Comment: 4.4. Interannual variability in lake Ice phenology

1. Mention the significance niveau/level of the trends directly when writing e.g. 6.31 day/10 yr. The authors need to write the significance level for the derived trends as e.g. in 5.1

Response: the statistical significance has been added in Table 6.

Comment: 2. The short length of data record (15 years) may be one of reasons that few lakes have significance trends.”‘ Replace with: ‚—why the other lakes show no significant trends.‘ One is not few.

Response: The sentence has been changed to “The short length of data record (15 years) may be one of reasons that trends have no significance.”

Comment: 3.  Figure caption 7 ‚ice cover.Table 6.‘ Add space. 

 Response: it has been corrected.

 

 Comment: 5.1. Factors influencing lake ice phenology

‚The growth and decay of lake ice cover are governed by thermodynamic interactions between atmospheric-water-ice interfaces and are affected by climatic and lake-specific factors. The major climatic factors controlling the heat exchange with the atmosphere are air temperature, solar radiation, precipitation, and wind speed. Temperature changes in water and air are regarded as the most prominent factors affecting lake ice cover [5, 8]. The amount of solar radiation and snow accumulation significantly affect the ice cover break-up process [10]. Lake-specific factors, such as lake morphometry (depth, area, and volume) and location (longitude, latitude, and altitude), determine heat storage in lakes and additionally affect lake ice formation [57].‘ Please move to the introduction part. It does not belong to the discussion. 

 Response: the context has been shortened to “The growth and decay of lake ice cover are governed by thermodynamic interactions between atmospheric-water-ice interfaces and are affected by climatic and lake-specific factors. The major climatic factors controlling the heat exchange with the atmosphere are air temperature, solar radiation, precipitation, and wind speed [5, 8, 10]. Lake-specific factors, such as lake morphometry (depth, area, and volume) and location (longitude, latitude, and altitude), determine heat storage in lakes and additionally affect lake ice formation [58]”, as it briefly illustrated how lake-specific and climatic factors contributed to the freeze-up and break-up patterns

 

Comment: 5.2. Sensitivity analysis of classification for water/ice status pixels

1. The melting temperature should consider the ice surface condition. ‘ Repetition to paragraph sentence one, please delete. 

Response: it has been corrected.

Comment: 2. Figure 10. The caption missed the description for the different panels (e.g. please repeat K_b etc).

Response: it has been added as “The first row presents the time series of  among the three lakes (a-c), and the second row presents the time series of  (d-f), and the third row presents the time series of  (g-i), and the fourth row presents the time series of  (g-j).”.

 

Comment: 5.3. The advantages and limitations of the method

1. The automatic temperature-based algorithm also shown the ability‘ Sentence grammar.

Response: it has been changed to “The automatic temperature-based algorithm was also shown the ability”.

 

Comment: 6. Conclusions

- add significance niveaus/levels to trends

Response: as the short period of trends, the statistical significance was not found in many lakes. Hence, we prefer to illustrate the significance level of the trend briefly in table 6 and section 4.4.

Author Response File: Author Response.docx

Reviewer 3 Report

Dear Editor, 

Thank you for sending the revised version of the manuscript. I would like, first of all, to thank the authors for their work to improve the study by following my comments and suggestions. The new version of the article provides a more exhaustive presentation of the proposed approach and the obtained results. From this perspective, the manuscript has been definitely improved. Remains, however, the need to improve the level of the English used in the text that, in my opinion, is still not suitable for a publication in an international journal like Remote Sensing. Aside from this major criticism, I think that, after extensive editing of the English used in the text, the manuscript could be considered for publication. 

I report, below, some further suggested editing that could improve the manuscript.

Figure 1: By adding the geographic coordinates of the considered lakes inside the relative sub-figure, the authors could make this figure more clearly.

Table1: Please, provide more details in the table's caption.

Figure 2: add more details to the figure caption.

Formula 1:  separate this formula from the main text.

Table 3: What do you mean by "Statistical Information"?

Figure 6: This figure could be improved by adding a grid and using a better color selection.

Table 4: The authors could consider moving this table to the appendix section.

Figures 6 and 7: The figures could be improved by changing the color selection and different line styles and adding a grid.

Figure 9: Probably reorganizing this figure from a 2 by 3 grid to a 3 by to grid could make the figure more readable.

Author Response

Reviewer #3

Comment: Dear Editor, 

Thank you for sending the revised version of the manuscript. I would like, first of all, to thank the authors for their work to improve the study by following my comments and suggestions. The new version of the article provides a more exhaustive presentation of the proposed approach and the obtained results. From this perspective, the manuscript has been definitely improved. Remains, however, the need to improve the level of the English used in the text that, in my opinion, is still not suitable for a publication in an international journal like Remote Sensing. Aside from this major criticism, I think that, after extensive editing of the English used in the text, the manuscript could be considered for publication. 

Response: we sincerely thank for the reviewer’s suggestions. The writing has been extensively improved by carefully choosing the appropriate words and phrases and double checking grammar in revised version. Following the reviewer’s concrete suggestions and comments, several aspects have been improved in the revised paper. Firstly, the paper title is modified as “An automatic method to detect lake ice phenology using MODIS daily temperature imagery”. Secondly, the abstract has been reconstructed and rewritten in a precise way. Thirdly, the results and the figures have been modified. And the detailed discussion has been added about the impact of daytime and nighttime combination which could be found in Section 5.3. Fourthly, the paper has been reorganized and readjusted in a clear way to make it more logical and readable.

 

Comment: I report, below, some further suggested editing that could improve the manuscript.

Figure 1: By adding the geographic coordinates of the considered lakes inside the relative sub-figure, the authors could make this figure more clearly.

Response: the geographic coordination has been added in Figure 1.

 

Comment : Table1: Please, provide more details in the table's caption.

Response: the caption has been changed to “The areas, depths, locations, altitude, and salinity for the selected lakes”

 

Comment : Figure 2: add more details to the figure caption.

Response: the caption has been changed to “Flowchart of the automated detection of lake ice phenology from MODIS daily temperature products.”

 

Comment : Formula 1:  separate this formula from the main text.

Response: it has been changed.

 

Comment : Table 3: What do you mean by "Statistical Information"?

Response: the caption has been changed to “Assessment information including Bias, SD, RMSE, and R² for the comparison of unfrozen water cover fraction between the MOD11_Merge and MODIS snow product data.”

 

Comment : Figure 6: This figure could be improved by adding a grid and using a better color selection.

Response: the color has been re-selected; the grid has been added, and it has changed to be Figure 7 in revised version.

 

Comment : Table 4: The authors could consider moving this table to the appendix section.

Response: Table 4 added to provide the time period and data sources for the comparison of lake ice phenology, as Reviewer#1 suggested.

 

Comment : Figures 6 and 7: The figures could be improved by changing the color selection and different line styles and adding a grid.

Response: thanks for the suggestion. The grid has been added in Figures with new different color selections and line styles.

 

Comment : Figure 9: Probably reorganizing this figure from a 2 by 3 grid to a 3 by to grid could make the figure more readable.

Response: as the first column is to represent the unfrozen water cover fraction during freeze-up period, and the second column is to present during the break-up period, we consider it could be easily readable by two different periods in 2 by 3 grid than 3 by 2 grid.

Author Response File: Author Response.docx

Round 3

Reviewer 1 Report

Dear authors,

your changes and answers to the second revision round improved the manuscript again a lot. I appreciate the changes especially regarding my main concerns. I would have wished a more comprehensive discussion of these points, but I recognise the short time. I’m looking forward to learn about your future research regarding e.g. the impact of temperature cycle on the combined product.

Overall, the review process is a great sucess, which improved the manuscript by far and therefore I congratulate you to your improvements. 

Please make sure, you don't miss editing the following points:

Abstract

• 'In this study, a new automated method was developed to extract lake ice phenology. In this study, a new automated method was developed to extract lake ice phenology parameters by capturing the temporal pattern of the tran- sitional water/ice phase by a parameterized time function.' Please Remove double sentence.

 

2.2.1 MODIS Daily temperature data

2. at approximately 10:30 LT (local time) and 22:30 LT Equatorial crossing from Terra and at approximately 1:30 LT and 13:30 LT Equatorial crossing from Aqua.‘ Please check ‚(local time)‘, while the others are Equatorial Crossing Time?

 

2.2.2. Other MODIS data and Landsat data

'The MODIS daily snow product (MOD10A1, version 6) with a spatial resolution of 500m and daily temporal resolution from 2002 to 2016 is used to validate the' Please change 'validate' to compare or evaluate.

5.3. The advantages and limitations of the method

'Hence, additional uncertainties are caused by the effect of diurnal temperature variations when combined Terra and Aqua data.'  Check sentence grammar.

Author Response

Reviewers' Comments to the Authors:

Reviewer #1

Dear authors,

Your changes and answers to the second revision round improved the manuscript again a lot. I appreciate the changes especially regarding my main concerns. I would have wished a more comprehensive discussion of these points, but I recognize the short time. I’m looking forward to learn about your future research regarding e.g. the impact of temperature cycle on the combined product.

Overall, the review process is a great success, which improved the manuscript by far and therefore I congratulate you to your improvements. 

Please make sure, you don't miss editing the following points:

Comment: Abstract: 'In this study, a new automated method was developed to extract lake ice phenology. In this study, a new automated method was developed to extract lake ice phenology parameters by capturing the temporal pattern of the transitional water/ice phase by a parameterized time function.' Please Remove double sentence.

Response: It has been removed.

 

Comment: 2.2.1 MODIS Daily temperature data: at approximately 10:30 LT (local time) and 22:30 LT Equatorial crossing from Terra and at approximately 1:30 LT and 13:30 LT Equatorial crossing from Aqua.‘ Please check ‚(local time)‘, while the others are Equatorial Crossing Time?

Response: It has been changed to “The daily temperature data are available four times per day: at approximately 10:30 and 22:30 from Terra and at approximately 1:30 and 13:30 from Aqua (local time)”

 

Comment: 2.2.2. Other MODIS data and Landsat data: 'The MODIS daily snow product (MOD10A1, version 6) with a spatial resolution of 500m and daily temporal resolution from 2002 to 2016 is used to validate the' Please change 'validate' to compare or evaluate.

Response: It has been changed to “compare”.

 

Comment: 5.3. The advantages and limitations of the method: 'Hence, additional uncertainties are caused by the effect of diurnal temperature variations when combined Terra and Aqua data.'  Check sentence grammar.

Response: It has been changed to “Furthermore, Terra and Aqua daily combination could cause the additional uncertainties due to ignoring the diurnal variation of lake surface temperature.”

 

Submission Date

14 May 2021

Date of this review

29 Jun 2021 15:42:26

Author Response File: Author Response.docx

Reviewer 3 Report

Dear Editor,

Thank you for sending the new version of the manuscript. I would like to thank the authors for addressing my comments. In my opinion, the manuscript has been significantly improved compared to the previous version. 

As I reported in the previous review of the article, I don’t have significant concerns about the scientific and technical value of the proposed study. Still, I think the English used in the text, even if improved compared to the old version, should be further revised before the manuscript publication on remote sensing.

I will report several suggestions below:

• Page 2: what do you mean by “optical sensing data”? Maybe using an expression like “data from optical sensors” would be more appropriate.
• Page 2: I would change “different climate zone” to simply “different climate”. 
• Page 3: what do you mean by “freeze part” and “Melting part”?
• Figure 1: I would show the actual lake’s location with a marker on the MainMap. Some of the considered lakes are small and not visible on the side of the relative label.
• Table 1 - Caption: change the “selected lakes” to the “considered lakes”.
• Page 4 - “The assessment of MODIS snow cover products have an overall accuracy of ~93% under clear-sky conditions.” - Rephrase into something like: “The assessment shows that MODIS snow cover products ...”
• Page 4 - what do you mean by “ lake water and ice cover information”?
• Page 5 - “The pixels were marked as “valid” …. Quality flags were “cloud effects” or others” -  Please rephrase this sentence. It is still confusing. The same suggestion applies to points a,b, and c reported below.
• Page 6 - “sensitivity analysis was discussed in section 5.1.” - You are using the past, but the reader has not gotten to section 5.1 yet.
• Page 6 - What do you mean by “the adjacent temporal filter”??
• Figure 4 - Caption: Please, clarify what you mean by “The solid circles represent the midpoints of the transition dates”
• Page 9 - “Comparison of the unfrozen water cover fraction with other datasets” - You could change this title to simply “Comparison with other datasets”.
• Page 9 - “The unfrozen water cover .... especially in large lakes [14, 15].” I would rephrase this period. It is still not clear. 
• Page 9 - “further used to evaluate.” - to evaluate what?
• Page 9 - “The consistency of unfrozen water cover fraction ...” - Rephrase.
• Page 9 - what do you mean by? - “One of the reasons is the misclassification of thin ice due to thin cloud cover.”
• Pages 11 -13: Eva though their meaning can be understood from the context, the acronyms FUS, FUE, BUE, etc., should be clearly defined in the text.
• Figure 6 - Caption - “taking the example of the year 2011.” - here saying something like “taking the year 2011 as an example”.
• Table 3 - What do you mean by “ Assessment information”?
• Figure 7 - change  “existing studies” into something like previous studies or other studies.

 

I hope the authors will find these suggestions helpful.

Author Response

Review #3

Dear Editor,

Thank you for sending the new version of the manuscript. I would like to thank the authors for addressing my comments. In my opinion, the manuscript has been significantly improved compared to the previous version. 

As I reported in the previous review of the article, I don’t have significant concerns about the scientific and technical value of the proposed study. Still, I think the English used in the text, even if improved compared to the old version, should be further revised before the manuscript publication on remote sensing.

I will report several suggestions below:

Comment: Page 2: what do you mean by “optical sensing data”? Maybe using an expression like “data from optical sensors” would be more appropriate.

Response: thanks for the reviewer’s suggestion. It has been corrected.

 

Comment: Page 2: I would change “different climate zone” to simply “different climate”. 

Response: It has been changed.

 

Comment: Page 3: what do you mean by “freeze part” and “Melting part”?

Response: It has been deleted.

 

Comment: Figure 1: I would show the actual lake’s location with a marker on the MainMap. Some of the considered lakes are small and not visible on the side of the relative label.

Response: It has been added.

 

Comment: Table 1 - Caption: change the “selected lakes” to the “considered lakes”.

Response: It has been changed.

 

Comment: Page 4 - “The assessment of MODIS snow cover products have an overall accuracy of ~93% under clear-sky conditions.” - Rephrase into something like: “The assessment shows that MODIS snow cover products ...”

Response: It has been changed to “The assessment showed that MODIS snow cover products have an overall accuracy of ~93% under clear-sky conditions [47]”.

 

Comment: Page 4 - what do you mean by “ lake water and ice cover information”?

Response: It has been modified to “The cloud-free images from the MODIS Daily Reflectance product and Landsat images from 2002 to 2016 were also selected to evaluate the lake water/ ice cover information from MODIS temperature products.”

 

Comment: Page 5 - “The pixels were marked as “valid” …. Quality flags were “cloud effects” or others” -  Please rephrase this sentence. It is still confusing. The same suggestion applies to points a,b, and c reported below.

Response: the sentence has been changed to “ When the pixel quality flags were “good quality”，“average land surface temperature error ≤ 0.01”, “average emissivity error ≤ 0.01”, or “average emissivity error ≤ 0.02”, the corresponding pixels were identified as “valid pixels”. When the pixel quality flags were “cloud effects” or others, the corresponding pixels were identified as “invalid pixels”.”

And the combination process has been rephrased as followed:

• If same pixel marked as “valid pixel” is found in more than two images in a given day, the value of corresponding pixel in MOD11_Merge was calculated by the average temperature from the “valid pixels” and also flagged as “valid”;
• If same pixel marked as “valid pixel” was found in only one image in a given day, the corresponding pixel in MOD11_Merge was from “valid pixel” and also flagged as “valid”;
• If same pixel marked as “valid pixel” was not found in any images in a given day, the corresponding pixel was marked as “invalid”.

 

Comment: Page 6 - “sensitivity analysis was discussed in section 5.1.” - You are using the past, but the reader has not gotten to section 5.1 yet.

Response: It has been changed to the present.

 

Comment: Page 6 - What do you mean by “the adjacent temporal filter”?

Response: it has been modified to “Afterward, the adjacent temporal filter was applied to deduce the continuous stable ice-covered or ice-free pixel from cloud-covered pixels is to use the same pixel in the previous and subsequent days without reducing and spatial and temporal resolution to deduce the continuous stable ice-covered or ice-free pixel from cloud-covered pixels. As the adjacent temporal filter was mainly used for stable status pixel deduction in our study, the two temporal windows of ±5 days and ±7 days were chosen, respectively.”

 

Comment: Figure 4 - Caption: Please, clarify what you mean by “The solid circles represent the midpoints of the transition dates”

Response: It has been changed to “Schematic representation of the logistic function. The solid circles represent the midpoints of the freeze-up and break-up period. The shaded areas are the freeze-up and break-up periods.”

 

Comment: Page 9 - “Comparison of the unfrozen water cover fraction with other datasets” - You could change this title to simply “Comparison with other datasets”.

Response: As the section 4.2 and section 4.3 present the comparison results with other datasets in different aspects, we use the detailed title of “Comparison of the unfrozen water cover fraction with other datasets” in section 4.2 to distinguish with section 4.3.

 

Comment: Page 9 - “The unfrozen water cover .... especially in large lakes [14, 15].” I would rephrase this period. It is still not clear. 

Response: It has been modified to “The unfrozen water cover fraction calculation and the outlier removal method (Section 3.2) were applied to the MODIS daily snow cover products. As shown in Figure 5, the unfrozen water cover fraction obtained from MOD11_Merge was overall consistent with that of the MODIS snow cover product, with the values of R² larger than 0.70 during the freeze-up pattern and 0.89 during the break-up pattern (Table 3). Compared with the freeze-up patterns, the break-up patterns in MOD11_Merge and MODIS snow cover product showed a higher agreement, with the R² values of 0.912, 0.889, and 0.948, respectively. The freeze-up pattern in MOD11_Merge and MODIS snow cover in Lake Qinghai has a lower R², Bias, and RMSE than other two lakes, with the value of 0.72, 0.22, and -0.55, respectively. This may be due to the thin ice or initial ice information which was not easily detectable, especially in large lakes [14, 15].”

 

Comment: Page 9 - “further used to evaluate.” - to evaluate what?

Response: It has been changed to “The visual inspections of lake ice under clear-day conditions from the MODIS daily reflectance products and Landsat images were further used to evaluate the water/ice cover information”

 

Comment: Page 9 - “The consistency of unfrozen water cover fraction ...” - Rephrase.

Response: It has been changed to “The unfrozen water cover fraction in the phases of break-up process was consistent with MODIS reflectance products, MODIS snow products, and MOD11_Merge temperature images”

 

Comment: Page 9 - what do you mean by? - “One of the reasons is the misclassification of thin ice due to thin cloud cover.”

Response: It has been modified to “The misclassification of thin ice is possibly caused by the lower temperature due to undetected cloud cover from MODIS temperature products.”

 

Comment: Pages 11 -13: Eva though their meaning can be understood from the context, the acronyms FUS, FUE, BUE, etc., should be clearly defined in the text.

Response: the definition has been added in section 3.3, “the specific ice phenology characteristics, freeze-up start and end dates (FUS and FUE), break-up start and end dates (BUS and BUE), frozen ice cover duration (days between freeze-up start date and break-up end date, FIC), and complete ice cover duration (days between freeze-up end date and break-up start date, CID))”

 

Comment: Figure 6 - Caption - “taking the example of the year 2011.” - here saying something like “taking the year 2011 as an example”.

Response: It has been changed.

 

Comment: Table 3 - What do you mean by “ Assessment information”?

Response: The caption has been modified to “The Bias, SD, RMSE, and R² for the comparison of unfrozen water cover fraction between the MOD11_Merge and MODIS snow product data.”

 

Comment: Figure 7 - change  “existing studies” into something like previous studies or other studies.

Response: It has been changed to “Comparison of lake ice phenology dates between our study (the right y-axis) and other studies”

I hope the authors will find these suggestions helpful.

 

Submission Date

14 May 2021

Date of this review

28 Jun 2021 11:50:59

 

Author Response File: Author Response.docx