Forecasting Snowmelt Season Temperatures in the Mountainous Area of Northern Xinjiang of China

Round 1

Reviewer 1 Report (New Reviewer)

1. In line 158 on page 4, maybe the distribution of grid points could be better explained for the benefit of the reader.

2. Maybe the authors can add summary of how this research might benefit the common people or independent meteorological/commercial organizations.

I think the overall quality of the English language is pretty good and maybe a brief proof-read might be sufficient.

Author Response

Comments and Suggestions for Authors

1. In line 158 on page 4, maybe the distribution of grid points could be better explained for the benefit of the reader.

Reply: Thank you, you are very careful.

The number of grid points and the spatial resolution of distribution in the study area of this paper have been described in lines 116-118 on page 3.

2. Maybe the authors can add summary of how this research might benefit the common people or independent meteorological/commercial organizations.

Reply: Thank you for your suggestion. The following contents have been added in the summary.

The common people need more accurate temperature forecast value,so the algorithm in this paper can provide technical reference for independent meteorological/commercial organizations.

Comments on the Quality of English Language

I think the overall quality of the English language is pretty good and maybe a brief proof-read might be sufficient.

Reply: The article has asked professional editing services to help proofread, and the author has proofread twice after this minor revision.

Reviewer 2 Report (New Reviewer)

The article is interesting and fits the profile of Water. The melting of the snow cover is crucial to the amount of water runoff from a catchment - especially in mountainous areas with steep slopes and low soil retention capacities. I recommend the article for publication after taking into account the following comments.

The article deals with forecasting air temperature during the snow melt period. Nevertheless, other factors cannot be ignored. The amount of sunshine, the structure of the snow, the surface coverage of the snow cover (and consequently the amount of albedo). The text should be expanded to include these issues, even if they are not the purpose of the study. It should be indicated, based on existing knowledge, what influence they have on the rate of melting of the snow cover.

The paper does not have a Discussion chapter. The paper refers to 40 publications of which 38 are in the Introduction chapter. The results obtained should be set against the background of existing knowledge in this area.

Author Response

The article deals with forecasting air temperature during the snow melt period. Nevertheless, other factors cannot be ignored. The amount of sunshine, the structure of the snow, the surface coverage of the snow cover (and consequently the amount of albedo). The text should be expanded to include these issues, even if they are not the purpose of the study. It should be indicated, based on existing knowledge, what influence they have on the rate of melting of the snow cover.

Reply: Your suggestion is particularly good, the article added a discussion section.

4.3. Discussion

This paper discusses the problem of temperature forecast during snowmelt period.  However, there are many factors affecting snowmelt,such as evapotranspiration[41],the amount of sunshine,the structure of the snow, the surface coverage of the snow cover[42] and precipitation[43-44] etc,which have an effect on snowmelt,in that more aspects should be considered in the later stage of this study.

The paper does not have a Discussion chapter. The paper refers to 40 publications of which 38 are in the Introduction chapter. The results obtained should be set against the background of existing knowledge in this area.

Reply: The article added a discussion section.At the same time, related literature been added.

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

The manuscript is about applying two correction algorithms to forecast data of Central Met. Observatory (SCMOC) of China by using the Chinese Land Data Assimilation System (CLDAS) data as real data. Dynamic unitary linear regression and average filtering revision schemes were used as the algorithms. The algorithm performances were tested for the mountainous areas having altitudes above 1300 m in northern Xinjiang in China. The authors used the corrected temperature forecasts to forecast daily maximum temperature, temperature rise range, snowmelt temperature and daily snowmelt duration. They used several indices to evaluate the forecasted temperatures. The forecasted temperature is an effective parameter to model the snowmelt. As the authors indicated as the novelty of their study, dependable temperature forecast may help in improving the snowmelt prediction in terms of flood time, flood peak discharge and volume. However, I doubt the novelty of the manuscript.

 I am afraid the scientific message of the paper is not so clear. In addition to that it is hard to follow the text. There are so many abbreviations and performance indices where I could not understand easily. The reason of choosing only two algorithms for the correction is not explained well. Considering the assimilated temperature product as real data may have uncertainty. How do the authors consider this uncertainty in their temperature forecast corrections? What is the reason to forecast 10 days but not correcting the forecasts as the new observations exist? Why do the authors not consider the filtering techniques using Kalman theory? Evaluating the performance of the forecasts in terms of performance indices for the number of grids do not make sense that much. It would be good to see the effect on estimating the snowmelt amount. I do recommend the authors to include more methods to improve the correction of forecasts and include snowmelt modelling to highlight the performance of the corrected temperature forecasts.  

Line 50: The temperature had become an important element of snow.  What does it mean? Temperature is always an important parameter for snowmelt.

Line 94-98 not clear. Rewording is needed.

Line 200-201 not clear. 

Author Response

Thank you for your careful and patient guidance. Based on your suggestions, the article has been revised as follows. Please check if it is suitable.

The manuscript is about applying two correction algorithms to forecast data of Central Met. Observatory (SCMOC) of China by using the Chinese Land Data Assimilation System (CLDAS) data as real data. Dynamic unitary linear regression and average filtering revision schemes were used as the algorithms. The algorithm performances were tested for the mountainous areas having altitudes above 1300 m in northern Xinjiang in China. The authors used the corrected temperature forecasts to forecast daily maximum temperature, temperature rise range, snowmelt temperature and daily snowmelt duration. They used several indices to evaluate the forecasted temperatures. The forecasted temperature is an effective parameter to model the snowmelt. As the authors indicated as the novelty of their study, dependable temperature forecast may help in improving the snowmelt prediction in terms of flood time, flood peak discharge and volume. However, I doubt the novelty of the manuscript.

Reply:

In this paper, two correction algorithms and four high temperature prediction algorithms are involved.

Among the two correction algorithms, dynamic linear regression algorithm is an improved algorithm based on regression algorithm which is innovative by the author.This algorithm has a good effect when applied in ground stations forecast.According to the conclusion of this paper, it is not suitable in mountainous areas.

The average filter correction algorithm is an algorithm that the author borrowed from himself achievements, which is quoted in the paper. From the conclusion of this paper, it is very suitable in mountain areas.

In this paper, four high temperature prediction algorithms make the temperature rise process forecast more accurate. The author consulted a number of experts in related fields to discuss and determine the relevant algorithms, completed the implementation of this algorithm and verified the forecast products in this paper. These four high temperature algorithms are original.

 I am afraid the scientific message of the paper is not so clear. In addition to that it is hard to follow the text. There are so many abbreviations and performance indices where I could not understand easily. The reason of choosing only two algorithms for the correction is not explained well. Considering the assimilated temperature product as real data may have uncertainty. How do the authors consider this uncertainty in their temperature forecast corrections? What is the reason to forecast 10 days but not correcting the forecasts as the new observations exist? Why do the authors not consider the filtering techniques using Kalman theory? Evaluating the performance of the forecasts in terms of performance indices for the number of grids do not make sense that much. It would be good to see the effect on estimating the snowmelt amount. I do recommend the authors to include more methods to improve the correction of forecasts and include snowmelt modelling to highlight the performance of the corrected temperature forecasts.  

Reply:

There are three temperature forecast products,which including original product (represented by 'XJ') was obtained from the temperature guidance forecast product of the China Meteorological Administration (SCMOC),the revised product (represented by 'HG') that the original forecast product has been modified by the dynamic linear regression algorithm and the revised product (represented by 'PJ') that the original forecast product has been modified by the average filter correction algorithm.The temperature forecast products future 240h in 3 h intervals value with a spatial resolution of 0.05°×0.05°,and reported starting from 08 AM Beijing time every day,during March to May 2021.

Real data was simulated value,which obtained from Chinese Land Data Assimilation System (CLDAS) hourly data with a spatial resolution of 0.05°×0.05°.The advantage of the real product was the combination of multi-source fusion data,such as ground observation data,ECMWF numerical product,GFS numerical product,FY2 Precipitation product,FY2 satellite full disk nominal map and DEM data.

The dynamic linear regression algorithm mainly dynamically updates the coefficients in the equation. The average filtering algorithm mainly dynamically improves the error. The different improvement points of the two algorithms were all based on Kalman filter correction theory.In order to predict the temperature rise process more accurately,this paper designs 4 innovative temperature forecast methods that including daily maximum temperature,daily temperature rise range,snowmelt temperature and daily snowmelt duration.Subsequently,the algorithm can be applied to test other areas and combined with the advantages of the algorithms to design a comprehensive algorithm.

Line 50: The temperature had become an important element of snow.  What does it mean? Temperature is always an important parameter for snowmelt.

Reply:Thank you, the statement has been modified.

The temperature had an important element of snowmelt.

Line 94-98 not clear. Rewording is needed.

Reply:In this study,two aspects were considered to design the correction algorithms. First,the mountainous terrain is steep and the temperature changes dramatically with the increase of altitude. Designed dynamic linear regression algorithm, which the latest real value is used to dynamically update the slope of the regression equation of temperature prediction to forecast the temperature change. Secondly,the amount of grid data and the system forecast error in mountainous area are large.Designed the average filter correction algorithm, which used the latest real value to dynamically update the forecast error of different forecast times to further forecast the temperature in mountainous area.

Line 200-201 not clear. 

Reply:To forecast of temperature rise range of the date, the maximum temperature forecast of the date after and the day before the date is required,so 'Hou' represents the daily maximum temperature of the following day,and 'Qian' represents the daily maximum temperature of the previous day.

Reviewer 2 Report

The manuscript entitled “Forecasting snowmelt season temperatures in the mountainous area of Northern Xinjiang of China.” by Zhang et al.

Overall, this study and its results are interesting. However, scientifically, many aspects need to be considered. In addition, the whole manuscript should be rearranged according to the journal format, including the reference list. I recommend a major revision before this paper can be accepted for publication. My comments are as follows,

General comments:

1. Line 58-58, “ Therefore, SCMOC, a grid forecasting product (0.05°×0.05°, forecasting the following 10 days in 3 h intervals)....”. The forecast data has a very high spatial resolution, which is very good compared to the temporal resolution, which is 3 hours. What is the accuracy of the 10-day forecast? Authors need to write more about the model configuration with proper scientific reference and the accuracy of the 10-day forecast. 

2. Line 76-80, “ “Li et al. established horizontal ……. ….temperature forecast products of intelligent grid forecast (SCMOC) (Liu et al., 2020).” These sentences are not clear; rewrite the whole section.

3. Multiple inconsistencies exist between the words, semicolons, commas, and space. Check the whole manuscript carefully and according to the journal format. 

4. In addition, the author used XJ, HG, and PJ in multiple places but nowhere defined these works' full form or meaning. Example in line 135 “Dynamic unitary linear regression (HG) revision scheme.” Why HG? Need to explain clearly.

5. Line 123 does not clarify the meaning of “Real data.” Is it observation, derived product, or simulated value? Write a few more words with proper references.

6. Line 123-124, “Chinese Land Data Assimilation System (CLDAS)” It would be good to write more sentences about the CLDAS system with proper references.

7. Line 124-125, “The advantage of the real product was the combination of multi-source fusion data, such as ground and satellite data.” Specify the list of all ground and satellite data. What type of satellite data was used? Need to specify with details.

8. Line 129 “the results showed that the non-independence test of CLDAS was 0.972.” non-independent test of CLDAS. The line is not clear; rewrite the line.

9. Section 2.2.1 Dynamic unitary linear regression (HG) revision scheme

Section 2.1.2. Average filtering (PJ) revision scheme

Section 2.2.1. Forecast of daily maximum temperature

Section 2.2.3. Forecast of snowmelt temperature and daily snowmelt duration

These sections require appropriate scientific references, how authors derive these equations and the conditions. Further, if the method is applicable in another region, I can see a few references cited by the author's paper in non-English language.

10.  All the references need to be rearranged according to the journal format.

11. And my final question is, what specific improvements should the authors consider regarding the methodology? Can the results be compared with other methods? What further controls should be considered?

Comments for author File: Comments.pdf

Minor editing of English language required

Author Response

Thank you for your careful and patient guidance. Based on your suggestions, the article has been revised as follows. Please check if it is suitable.

1. Line 58-58, “ Therefore, SCMOC, a grid forecasting product (0.05°×0.05°, forecasting the following 10 days in 3 h intervals)....”. The forecast data has a very high spatial resolution, which is very good compared to the temporal resolution, which is 3 hours.What is the accuracy of the 10-day forecast? Authors need to write more about the model configuration with proper scientific reference and the accuracy of the 10-day forecast.

Reply: This forecast product has average temperature forecast accuracy of less than 50% in mountainous areas.Therefore, it is necessary to design a correction algorithm to improve the accuracy of the original forecast product in mountainous areas.

The data is stable and services have been delivered.

2. Line 76-80, “ “Li et al. established horizontal ……. ….temperature forecast products of intelligent grid forecast (SCMOC) (Liu et al., 2020).” These sentences are not clear;rewrite the whole section.

Reply:Li et al. established three forecast models including horizontal, longitudinal, and horizontal and longitudinal integrated forecast models (Li et al., 2020) and Liu et al. used three algorithms including the wavelet analysis, sliding training algorithm, and optimal fusion algorithm (Liu et al., 2020) were to correct the temperature forecast products (SCMOC).

3.Multiple inconsistencies exist between the words, semicolons, commas, and space. Check the whole manuscript carefully and according to the journal format.

Reply:Thank you for your reminding. I have reviewed the manuscript again, but I don't know if it is suitable. If the sentence format is still not suitable, could you give me another chance to ask professional editing service?

4. In addition, the author used XJ, HG, and PJ in multiple places but nowhere defined these works' full form or meaning. Example in line 135 “Dynamic unitary linear regression (HG) revision scheme.” Why HG? Need to explain clearly.

Reply:There are three temperature forecast products,which including original product (represented by 'XJ') was obtained from the temperature guidance forecast product of the China Meteorological Administration (SCMOC),the revised product (represented by 'HG') that the original forecast product has been modified by the dynamic linear regression algorithm and the revised product (represented by 'PJ') that the original forecast product has been modified by the average filter correction algorithm.The temperature forecast products future 240h in 3 h intervals value with a spatial resolution of 0.05°× 0.05°,and reported starting from 08 AM Beijing time every day,during March to May 2021.

Line 123 does not clarify the meaning of “Real data.” Is it observation, derived product, or simulated value? Write a few more words with proper references

Reply: Real data was simulated value,which obtained from Chinese Land Data Assimilation System (CLDAS) hourly data with a spatial resolution of 0.05° × 0.05°.

Line 123-124, “Chinese Land Data Assimilation System (CLDAS)” It would be good to write more sentences about the CLDAS system with proper references.

Reply:The advantage of the real product was the combination of multi-source fusion data, such as ground observation data,ECMWF numerical product,GFS numerical product,FY2 Precipitation product,FY2 satellite full disk nominal map and DEM data.

http://data.cma.cn/data/detail/dataCode/NAFP_CLDAS2.0_NRT.html

Line 124-125, “The advantage of the real product was the combination of multi-source fusion data, such as ground and satellite data.” Specify the list of all ground and satellite data. What type of satellite data was used? Need to specify with details.

Reply:The advantage of the real product was the combination of multi-source fusion data, such as ground observation data,ECMWF numerical product,GFS numerical product,FY2 Precipitation product,FY2 satellite full disk nominal map and DEM data.

Line 129 “the results showed that the non-independence test of CLDAS was 0.972.” non-independent test of CLDAS. The line is not clear; rewrite the line

Reply:After Liu Ying et al. evaluating the temperature of CLDAS-V2.0 in Northwest China including this study area, the test results of CLDAS showed that the non-independence test was 0.972, the independence test was 0.950, the mean deviation was -0.271 ℃, the RMSE was 2.406 ℃ and the mean absolute error was 1.588 ℃ (Liu et al., 2021).

Section 2.2.1 Dynamic unitary linear regression (HG) revision scheme Section 2.1.2. Average filtering (PJ) revision scheme Section 2.2.1. Forecast of daily maximum temperature Section 2.2.3. Forecast of snowmelt temperature and daily snowmelt duration These sections require appropriate scientific references, how authors derive these equations and the conditions. Further, if the method is applicable in another region, I can see a few references cited by the author's paper in non-English language.

Reply: In this paper, two correction algorithms and four high temperature prediction algorithms are involved.

Among the two correction algorithms, dynamic linear regression algorithm is an improved algorithm based on regression algorithm which is innovative by the author.    This algorithm has a good effect when applied in ground stations forecast.    According to the conclusion of this paper, it is not suitable in mountainous areas.

The average filter correction algorithm is an algorithm that the author borrowed from himself achievements, which is quoted in the paper. From the conclusion of this paper, it is very suitable in mountain areas.

In this paper, four high temperature prediction algorithms make the temperature rise process forecast more accurate. The author consulted a number of experts in related fields to discuss and determine the relevant algorithms, completed the implementation of this algorithm and verified the forecast products in this paper. These four high temperature algorithms are original.

All the references need to be rearranged according to the journal format.

Reply:Thank you, references and citations have been modified according to the journal format.. Please check if it is suitable.

11. And my final question is, what specific improvements should the authors consider regarding the methodology?  Can the results be compared with other methods?  What further controls should be considered

Reply:Thank you, already added in summary:

The dynamic linear regression algorithm mainly dynamically updates the coefficients in the equation. The average filtering algorithm mainly dynamically improves the error. The different improvement points of the two algorithms were all based on Kalman filter correction theory.In order to predict the temperature rise process more accurately,this paper designs 4 innovative temperature forecast methods that including daily maximum temperature,daily temperature rise range,snowmelt temperature and daily snowmelt duration.Subsequently,the algorithm can be applied to test other areas and combined with the advantages of the algorithms to design a comprehensive algorithm.

Round 2

Reviewer 2 Report

I am not satisfied with the revised version of the manuscript, and most of my previous questions (Q. 1, 3, 4-11) still need to be clarified. 

In addition, in the replay of Q11. 

The authors mentioned, without any reference, “Kalman filter correction theory.” What is this theory? I know that the Kalman Filter theory is used for data assimilation. 

Also, I do not find any scientific knowledge in this paper in its present form. The revised version needs to be more evident than the previous version. Any lines will be confusing to all readers. 

Note that the link provided by the authors needs to be fixed for me. 

What is FY2, what is “DEM data”?

Authors must check carefully and rewrite the entire manuscript with a professional editing service.

Extensive editing of the English language is required.