Creation and Verification of a High-Resolution Multi-Parameter Surface Meteorological Assimilation Dataset for the Tibetan Plateau for 2010–2020 Available Online
, Xian Zhu 3, Maoshan Li 1, Mei Chen 1, Shaobo Zhang 1, Xianyu Yang 1, Zhiyuan Zheng 4,5,6, Yikun Qin 7, Yu Zhang 1 and Shihua Lv 1
Round 1
Reviewer 1 Report
The revised manuscript “Creation and verification of a multi-parameter surface meteorological dataset for the Tibetan Plateau for 2010-2020, available online” provides us a high-resolution atmospheric dataset over the Qinghai-Tibetan Plateau (QTP) with many additional data in field experiments are assimilated, which is important for atmospheric analysis and research on the complex terrain over the QTP. The use of 3D-Var method in WRF model is a straightforward way and relevant to produce the high-resolution reanalysis data to show us reliable description of the atmospheric state, which is a difficulty for existing reanalysis data.
The authors have made a great improvement according to the comments given by the reviewers. Some issues, however, still exist as displayed.
1. Just as the answered by the authors to reviewer 2 (point 2), the diurnal cycle of the QTP-HRAD data show obvious hourly bias such as at Garz, Nayingchi and Pulan. Well accordance to the observation is shown at other 6 stations, i.e. Qamdo, Hotan, Nagqu, Zoige, Tuotuohe and Yushu. What is the reason? Does this result relate to the land surface classification? The availability of the hourly data should be noted in the manuscript. Alternatively, suggestion to readers can be given (in lines 373-380) for using the data, for example daily average is recommended.
2. Comma in the title can be removed
3. Line 98, replace “great” with “large”
4. Line 403, what is the orientation of the station?
5. Fig.6, low correlation coefficient (CC) of the 10-m wind speed shows the large error of distribution pattern of horizontal wind, contrasting to the high CC of 2-m temperature and all surface quantities. Does it mean the error growth with height? Additional discuss needed.
Author Response
- Just as the answered by the authors to reviewer 2 (point 2), the diurnal cycle of the QTP-HRAD data show obvious hourly bias such as at Garz, Nayingchi and Pulan. Well accordance to the observation is shown at other 6 stations, i.e. Qamdo, Hotan, Nagqu, Zoige, Tuotuohe and Yushu. What is the reason? Does this result relate to the land surface classification? The availability of the hourly data should be noted in the manuscript. Alternatively, suggestion to readers can be given (in lines 373-380) for using the data, for example daily average is recommended.
Answer: The discrepancies between the simulated and observed hourly values at Garz, Nayingchi , and Pulan stations may be attributed to their locations in regions with significant altitude variations, resulting in inconsistencies between the model grid points and observation points in terms of altitude. The daily observation and simulation values at these three stations also exhibit higher errors compared to other stations located in flat areas. As suggestion in the manuscript, a pressure-altitude correction should be applied when comparing the assimilation data with the station data. Please see lines 374-379 and 418-421 for further details.
- Comma in the title can be removed
Answer: it has been removed.
- Line 98, replace “great” with “large”
Answer: it has been corrected.
- Line 403, what is the orientation of the station?
Answer: The stations we used were national observation stations established by the China Meteorological Administration, located in the Qinghai-Tibet Plateau region. The Hotan, Nagqu, and Zoige stations are located in the northwest, central-southern, and eastern parts of the plateau, respectively, and are standard meteorological observation stations oriented north-south. The geographical locations of the ten meteorological stations can be found in Figure 2.
- 6, low correlation coefficient (CC) of the 10-m wind speed shows the large error of distribution pattern of horizontal wind, contrasting to the high CC of 2-m temperature and all surface quantities. Does it mean the error growth with height? Additional discuss needed.
Answer: It does not mean the error growth with height. In general, the accuracy of numerical weather prediction models for simulating near-surface wind speed and direction is lower than that for temperature, humidity, pressure, and other variables. This is because the simulation of wind speed requires consideration of many parameterization schemes, such as turbulence parameterization schemes and boundary layer parameterization schemes, and inaccurate estimation of parameters such as surface roughness within these schemes can cause errors. Typically, surface wind speed is considered to be the wind speed at 10 meters above the surface rather than 2 meters, representing the lowest boundary layer where free flow of air occurs.
We have discussed the reasons for the overestimation of wind speed simulation in our manuscript. Please see lines 484-492 for details.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Reviewer's comments:
Manuscript ID: remotesensing-2345559-peer-review-v2：“Creation and verification of a multi-parameter surface meteorological dataset for the Tibetan Plateau for 2010-2020, available online”
Review Summary:
I have read the above paper carefully and found that the presented work is very important and interesting but before considering it for publication, some modifications are needed to refine this manuscript for more meaningful outcomes. Therefore, I suggest that major modifications have to be done before considering it for publication. According to my review, some MAJOR modifications are needed by the author(s).
Please find below my specific comments:
Review comments:
1. Please modify the title, which clearly reflects all the phases of your work including the purpose of conducting this work, methodology/modeling, and the final output. Here only the goal and the final output are highlighted in the title.
2. Please explain the acronym used for the first time.
3. What do 2 m temperature and 10 m wind speed mean here? Please specify it to remove the confusion. Also, make clear the distinction between land surface temperature and air temperature.
4. Does the final output of the metrological dataset at 5 km x 5 km enough keeping in view the high spatial variability of precipitation over the geographical area? Various studies have been previously conducted over Tibetan Plateau for downscaling the precipitation dataset at 1 km. Why those datasets have not been used for generating the meteorological dataset at 1 km x 1 km?
5. I have not seen any comparison for the taken meteorological variables with already available online datasets for the accuracy evaluation.
Author Response
- Please modify the title, which clearly reflects all the phases of your work including the purpose of conducting this work, methodology/modeling, and the final output. Here only the goal and the final output are highlighted in the title.
Answer: Thank you for your comments. The title has been changed to “Creation and verification of a high-resolution multi-parameter surface meteorological assimilation dataset for the Tibetan Plateau for 2010-2020 available online”.
- Please explain the acronym used for the first time.
Answer: Thank you for your comments. We have carefully reviewed the entire text and provided the complete spellings for all acronyms.
- What do 2 m temperature and 10 m wind speed mean here? Please specify it to remove the confusion. Also, make clear the distinction between land surface temperature and air temperature.
Answer: The temperature at 2 meters and the wind speed at 10 meters both represent near-surface temperature and wind speed. However, as defined by the World Meteorological Organization (WMO), near-surface wind speed refers to the wind speed at a height of 10 meters above the ground. Therefore, meteorological stations measure near-surface wind speed at a height of 10 meters, and the wind speed output in models is also at this height. On the other hand, near-surface meteorological elements such as air temperature, pressure, and humidity are measured at a height of 2 meters above the ground. Land surface temperature is referred to as "surface temperature" in the text to distinguish it from "2m air temperature".
- Does the final output of the metrological dataset at 5 km x 5 km enough keeping in view the high spatial variability of precipitation over the geographical area? Various studies have been previously conducted over Tibetan Plateau for downscaling the precipitation dataset at 1 km. Why those datasets have not been used for generating the meteorological dataset at 1 km x 1 km?
Answer: Thank you for your comment. The 5 km x 5 km meteorological assimilation dataset is already very detailed for the Qinghai-Tibet Plateau, and it can reflect the spatial variation of precipitation. However, to generate a 1x1 km model dataset for the Qinghai-Tibet Plateau, it would require an astonishing 3500x1500 spatial grid points, which would entail a very large amount of computation and time. Some previous studies have released 1 km precipitation data or surface temperature data based on satellite data and ground observation stations, but these can only generate single-element data and have lower temporal resolutions (daily or monthly data). Using regional numerical models, we can generate multiple meteorological element data at the same time, and the temporal resolution can reach 1 hour, which was previously not achievable in similar studies.
- I have not seen any comparison for the taken meteorological variables with already available online datasets for the accuracy evaluation.
Answer: As the purpose of this article is to introduce the method and usage of this dataset to readers, a comparison and error analysis was conducted with the data from the National Surface Observation Stations of China Meteorological Administration, and a preliminary comparison was made with ERA-5 data. However, a thorough comparison and validation with other online published datasets would require a significant amount of work, which will be a focus of our future research. After the release of this dataset online, other researchers can also verify and compare it with other similar datasets and provide feedback to us for further improvement. This has been explained in lines 695-699 of the manuscript.
Author Response File: Author Response.pdf
Reviewer 3 Report
The paper present technical work to build a meteorological analysis of weather parameters over the Tibetan plateau. The significance of the work and the importance of the product relate to the importance of the Tibetan plateau for the formation of climate at the global scale. It is the biggest mountain massif in the world, center to the biggest land mass – the Eurasian. It influences circulation patterns and drives thermal contrast at the global scale. The authors illustrate well this importance in the introduction.
The lower space resolution of the existing global reanalysis products limits their applicability over the Tibetan region. Therefore the author decided to produce higher resolution analysis by assimilating available data from newer automatic stations – both ground and air-sounding. They used existing tools like the 3DVAR of the WRF model.
They verified their finer analysis against their own data and showed correlations. The conclusion is that the WRF tool simulates well the weather over the Tibetan plateau with higher resolution sufficiently well.
The 10-year analysis is freely available and can be used for weather and climate simulations as well as to feed hydrological models with weather data.
Author Response
The paper present technical work to build a meteorological analysis of weather parameters over the Tibetan plateau. The significance of the work and the importance of the product relate to the importance of the Tibetan plateau for the formation of climate at the global scale. It is the biggest mountain massif in the world, center to the biggest land mass – the Eurasian. It influences circulation patterns and drives thermal contrast at the global scale. The authors illustrate well this importance in the introduction.
The lower space resolution of the existing global reanalysis products limits their applicability over the Tibetan region. Therefore the author decided to produce higher resolution analysis by assimilating available data from newer automatic stations – both ground and air-sounding. They used existing tools like the 3DVAR of the WRF model.
They verified their finer analysis against their own data and showed correlations. The conclusion is that the WRF tool simulates well the weather over the Tibetan plateau with higher resolution sufficiently well.
The 10-year analysis is freely available and can be used for weather and climate simulations as well as to feed hydrological models with weather data.
Answer: Thank you very much for your comments.
Author Response File: Author Response.pdf
