General comments:

This paper describes a ground surface temperature (GST) monitoring network established in a specific region of the Qinghai-Tibet Plateau. Temperature sensors were deployed across areas of varying surface characteristics to monitor changes in GST under different landcover conditions. The collected monitoring data is abundant and of reasonably high quality. The authors have conducted a thorough analysis of the acquired data, providing readers with a more in-depth understanding about the freeze-thaw state during that period. Overall, the English writing in this paper is clear and coherent, and the obtained data can serve as valuable input for modeling or validation of surface processes. However, there are still some issues that the authors should consider. I would be highly appreciated if the authors could address them.

Specific comments:

1. What is the difference between the ground surface temperature (GST) mentioned in the paper and the land surface temperature (LST) commonly referred to in the remote sensing field, as well as soil temperature? Additionally, the description “topsoil temperature” in the data website provided by the authors raises questions about the physical meaning of the variables discussed in the paper. It is recommended that the authors either standardize their terminology or provide additional explanations within the text to ensure a clearer representation.
2. In the Introduction section, the authors mentioned that some scholars have already deployed GST monitoring networks in the northeastern part of the Qinghai-Tibet Plateau (e.g., Luo et al., 2020; Serban et al., 2023). What distinguishes the observational data in this study from those previous efforts? Perhaps the authors placed their monitoring network in mountainous regions? However, it seems that the data analysis by the authors did not include a specific analysis of mountainous characteristics. Despite some sections discussing elevation, the more unique features of mountainous regions such as three-dimensional structure and illumination conditions were not addressed.
3. The title of the paper mentions a “multiscale observation network…” but typically, multiscale implies different sensor observation fields (e.g., ground stations, drones, satellites). However, in this study, all sensors used for observations are ground-based and have the same observation field, with differences only in their placement. Additionally, it cannot be claimed that the sensors observed data at “local scale”, “landscape scale”, and “regional scale” because the instruments still provide sparse point observations and do not comprehensively cover an area. In summary, I am concerned about the validity and accuracy of the description “multiscale observation” in the paper.
4. The authors mentioned that some sensors were malfunctioning. What is the current status of these sensors? Are they now operational, or are they still not functioning correctly? Is there a possibility of acquiring more comprehensive observational data in the future?
5. Page 6, line 154. The authors mentioned a data collection interval of 3 hours for ground observations. Does this mean that data is recorded once every 3 hours, or is it recorded multiple times and then averaged using a specific algorithm? I suggest providing a brief explanation in the paper for clarity.
6. Page 9, line 206. How were the 165 “biased” data points mentioned in the paper determined? Were they identified through manual inspection or using a specific criterion (e.g., three times the standard deviation screening)?
7. Page 14, line 266. Why is it that a 14-meter distance can observe larger GST differences for the same type of landcover type?
8. Page 17, line 290. Although the authors have provided some explanations regarding the relationship between MAGST and elevation, it might be more intuitive to include a graphical representation of the MAGST and elevation relationship.
9. Page 19, lines 341-348. While it is understandable that the authors compare the results of FDD calculations with previous satellite-based calculations, is it meaningful to compare the results with very distant regions like Antarctica or other islands (especially when the timeframes are not consistent)?
10. Page 20, lines 374-376. While the authors mention that GST monitoring can provide a better assessment of the presence or absence of permafrost, they also note the high spatial variability of permafrost thaw. In my view, for an accurate determination of permafrost status, even when using GST as an indicator, a highly dense sensor network would be necessary, which does not seem to be currently feasible. Therefore, the authors need to further explain why they chose GST monitoring for assessing permafrost status over other methods such as borehole measurements (considering factors like cost, convenience, data uncertainty, etc.).
11. Page 22, line 425. While the authors mention the potential significance of this dataset for improving modeling methods, the entire paper analyzes the relationship between GST and freeze-thaw without specifying the advantages of higher spatial resolution GST monitoring data for model improvement (compared to using satellite data). Considering that large-scale snow and ice state analysis typically relies on satellite observations, is there a genuine necessity for such dense sensor deployment?

Technical corrections:

1. Page 2, line 46. The term “permafrost areal extents” is also a component of “model accuracies”, so there is no need to repeat it.
2. In Figure 1, there is an issue with the legend labels. “locale” should be “local”. Additionally, please confirm whether “Qingshui’he” should be “Qingshuihe”.
3. Page 9, line 194. Are the mentioned four failed sensors included among the previously mentioned 11 sensors, or are they an additional set of four sensors?
4. Page 10. In the title of Figure 3, there is no need to repeatedly provide the full term of “GST”.
5. Page 12. I suggest adding a legend to Figure 5.
6. Page 17, line 318. “… TDD of 320 m and 180 ℃ day”, remove “m”.
7. Page 18, line 328. When the authors mention “… most of the sites”, I suggest giving the exact percent.

The authors provided a valuable dataset of GST observations at various spatial scales in the Headwater Area of the Yellow River (HAYR). GST datasets were collected at 39 sites between 2019 and 2020. The authors showed how the measurements could be used for permafrost research.

General Comments

(1) Overall picture

While the authors provide a very detailed comparison of GST at different scales, this study generally lacks an overall picture. An easy way to do this would be to examine the lapse rate of MAGST. There should be a new figure with the x-axis representing elevation and the y-axis representing MAGST. You could even use different colors to represent vegetation cover.

(2) Permafrost borehole temperature datasets

A borehole temperature measurement from Luo et al., 2018 was used to determine whether permafrost was present. As an additional dataset, I suggest authors make the borehole temperature measurements public open.

(3) Review of GST measurements

In the CMA monitoring network, GST has been measured since the 1950s on the QTP and even the entire country. In spite of this, the measurement algorithm is inconsistent, making direct use of the dataset problematic (see Cui et al., 2020, Cao et al., 2023). Therefore, the datasets here are valuable. It would be helpful if you reviewed the measurement algorithms and clarified your significance.

Specific Comments

L37: …approximate or about 55%..

L39: Cao et al., 2019 PPP reported the permafrost zonation index map based on a statistical model and various measurements. Please consider citing here.

L44: Cao et al., 2018, JGR-Atmospheres reported the permafrost changes over the Northeastern QTP.

L56: Cao et al., 2020 TC (Table 1) reported how the MAGST combined with thermal offset can be used as an indicator for permafrost presence/absence. 

L104: Please clrify the landcover and microtopography information here.

L135: “…for some sites…”, please give the number of sites which have similar landcover.

L170: change larger to greater

L173: The principle behind SO and TO is the effects of vegetation cover, and soil properties (soil organic content, soil moisture). Please clarify here.

L193: change “delete” to remove

L231: "Differences larger than 2.5 ºC were observed mainly at the sites at elevations above 4600 m a. s. l., regardless of the landcover types in the plots." why?

Fig.1: Please add the specific distance for each scale in the legend.

References

Cao, B., Zhang, T., Peng, X., Mu, C., Wang, Q., Zheng, L., Wang, K., & Zhong, X. (2018). Thermal Characteristics and Recent Changes of Permafrost in the Upper Reaches of the Heihe River Basin, Western China. Journal of Geophysical Research: Atmospheres, 123(15), 7935–7949. https://doi.org/10.1029/2018JD028442

Cao, B., Zhang, T., Wu, Q., Sheng, Y., Zhao, L., & Zou, D. (2019). Permafrost zonation index map and statistics over the Qinghai-Tibet Plateau based on field evidence. Permafrost and Periglacial Processes, 30(3), 178–194. https://doi.org/10.1002/ppp.2006

Cao, B., Zhang, T., Wu, Q., Sheng, Y., Zhao, L., & Zou, D. (2019). Brief communication: Evaluation and inter-comparisons of Qinghai–Tibet Plateau permafrost maps based on a new inventory of field evidence. The Cryosphere, 13(2), 511–519. https://doi.org/10.5194/tc-13-511-2019

Cao, B., Wang, S., Hao, J., Sun, W., & Zhang, K. (2023). Inconsistency and correction of manually observed ground surface temperatures over snow-covered regions. Agricultural and Forest Meteorology, 338(November 2022), 109518. https://doi.org/10.1016/j.agrformet.2023.109518

Cui, Y., Xu, W., Zhou, Z., Zhao, C., Ding, Y., Ao, X., & Zhou, X. (2020). Bias Analysis and Correction of Ground Surface Temperature Observations across China. Journal of Meteorological Research, 34(6), 1324–1334. https://doi.org/10.1007/s13351-020-0031-9

Luo, D., Jin, H., Jin, X., He, R., Li, X., Muskett, R. R., Marchenko, S. S., & Romanovsky, V. E. (2018). Elevation-dependent thermal regime and dynamics of frozen ground in the Bayan Har Mountains, northeastern Qinghai-Tibet Plateau, southwest China. Permafrost and Periglacial Processes, 29(4), 257–270. https://doi.org/10.1002/ppp.1988