general comments

The topic is highly relevant and supports the projects to improve climatological data quality. This is important to understand studies on trends in climate associated to climate change. Attention is given to the movement of observing sites due to urbanisation in connection to harmonisation of datasets. Less attention is given to global activities, under the governance of WMO, to improve harmisation of climate datasets world wide although these activities are recomended in the conclusions. The examples provided give a good idea of the issues with harmonisation in practice, in particular for an immense area like China.

specific comments

Although to topic is clear and plenty references are provided, a refrence to the WMO World Climate Programme (WCP) is missing, in particular to the World Climate Data and Monitoring Programme (WCDMP) (see https://community.wmo.int/en/world-climate-programme-wcp). Within this program a number of guides are published which are relevant to learn from, like WMO/TD- No. 1186; WCDMP- No. 53 (Guidelines on climate metadata and homogenization). Also the constraint to have appropriate metadata in order to classify the quality and ability to process data is not regarded. In the introduction (2nd all.) surface friction is mentioned, but in the paper no atention is given to the parameter terrain roughness as defined and explained in WMO-No. 8, Vol. I (Wind Chapter). Trends in roughness are an essential analyses critria when evaluation wind datasets. The quality of wind data is associated with siting criteria, required functional specifications of wind sensors and their maintenace policy. These topics are not refered to,but relavant because trends in these items will affect data homogenity.

specific items, line by line

[22] reference to the WMO World Climate Programme is missing

[24] (~10m), to be 10 m [value and unit to be separated by a space]; 

[35] homogenization to become data homogenization

[77] refer also to WMO World Climate Programme is missing, including policies on data homogenization

[77] missing reference to WMO, 2003: WMO/TD - No. 1186; WCDMP - No. 53, Guidelines on Climate Metadata and Homogenization, https://library.wmo.int/doc_num.php?explnum_id=10751

[90] Chosen is for the Thiessen Polygon approach (Thiessen, 1991), but explanation for this choice is not provided.

[115] 1.85km and 1.85*cosφ to become 1.85 km and 1.85 × cos φ [the symbol "×" should be used for multiplication, not "*"]

[146] for "the frequency of zeros", "zeros" should be defined (not trivial): Stans "zero" for no data or for no wind.

[151] "accuaracy" stands for a subjective expression.  For a quantitative use of accuracy the word "uncertainty" shall be used, not "accuracy".

[157] {see [151]}

[159] missing: information on maintenance, functional specifications of wind sensor (including changes) and proven traceability to SI (calibration, see also WMO-No. 8). Also the required siting specifications are missing (including changes), see WMO-No. 8, Vol. I.

[183] missing: roughness data and their trend. Not determined? Are necessary to understand further analyses of data (see WMO-No. 8, Vol. I, Wind Chapter)

[320] sentence ends with a comma, not a dot.

[323] "lacked"; preference for "missing".

[328] Refrence to WMO Worls Climate Programme, in particular the World Climate Data and Monitoring Programme stimulating the proposed actions.

[476] Not a "WMO Technical Report", but a WMO Guide; to become WMO-No. 1203. Refrence to be replaced by principal source, https://library.wmo.int/doc_num.php?explnum_id=4166

[479] Not a "WMO Technical Report, but a WMO Guide; to become WMO-No. 1245.

General comments:

The meteorological data continuity is crucial for climate and climate change areas. The paper focuses on the discontinuity of the wind speed datasets across China and quantifies the possible impacts of anemograph changes, data logging methods, and site relocations on long-term wind speed trends. The authors found that the use of advanced anemographs can increase the observation frequency and improve the instrument sensitivity, thereby enhancing data accuracy. Additionally, the paper checks station relocations by examining the altitude, latitude, and longitude information of all observation stations, identifying approximately 90 stations in China that are affected by relocation and exhibit an excessive increase in wind speed trends. Based on the discontinuity analysis, problematic stations are recognized and excluded. The wind speed data series is reconstructed and a new wind speed trend is also given in this paper.

The topic of the paper is suitable for Atmospheric Measurement Techniques. The authors’ careful and comprehensive work in examining the wind data series across China is appreciated, and I also agree with their call for the improvement of data quality based on user feedback.

This paper requires some minor revisions before publication.

Specific comments:

1. Line 101: To maintain consistency in Table S1, the "(3)" should be added before the third criterion.
2. Line 169: What do the “EL” and “EC” stand for?
3. Line 284: More detailed information about the “Thiessen Polygon” method should be given.
4. Line 285-287: Large weights of the wind speed were given in Northwestern China and the Tibetan Plateau, but these regions are also complex terrains compared to Eastern China. How does the author consider the problem of station representation?
5. Line 295-297: The phenomenon of the increasing trend in wind speed across China in recent decades after using the Thiessen Polygon should be further explained.
6. It is suggested that the author DO MORE WORK TO CLARIFY the wind speed trends in different regions of China, at least over eastern China.