the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 07 Mar 2025
China Coastal GNSS Network: Advancing Precipitable Water Vapor Monitoring and Applications in Climate Analysis
Abstract. The Global Navigation Satellite System (GNSS) offers precise, continuous monitoring of atmospheric water vapor, essential for weather forecasting and climate research. This study presents a high-accuracy precipitable water vapor (PWV) dataset from 55 GNSS stations along China’s coast (2009–2019). PWV retrievals utilized weighted mean temperature (Tm) and zenith hydrostatic delay (ZHD) derived from fifth-generation European ReAnalysis (ERA5) products. After rigorous quality control, the dataset achieved an average completeness rate of 70 %. Validation against ERA5 PWV products showed strong agreement (mean bias: 0.80 mm; RMS error: 2.52 mm), while comparisons with radiosonde profiles yielded a mean bias of 0.90 mm and an RMS error of 3.01 mm, confirming its accuracy and reliability. Spatial analysis revealed PWV values ranging from 0 to 88.57 mm, with minima decreasing with increasing latitude and concentrated around the Yangtze River estuary. Temporal patterns exhibited prominent annual and semi-annual cycles, particularly in higher latitudes. PWV showed a strong positive correlation with sea surface temperature (SST; r = 0.76), with a 1 K SST increase leading to a 2.4 mm (7 %) PWV rise. This dataset supports high-precision applications, including PWV validation, extreme weather prediction, and climate trend analysis. The processed ZTD and PWV datasets from 55 CGN stations are accessible at https://zenodo.org/records/14639032.
- Preprint
(22608 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 14 May 2025)
-
RC1: 'Comment on essd-2025-24', David Adams, 02 Apr 2025
reply
Review of "China Coastal GNSS Network: Advancing Precipitable Water Vapor Monitoring and Applications in Climate Analysis" by Zhilu Wu et al. April 2025
David K. Adams - dave.k.adams@gmail.com
General Commnents.
I think, in general, this article is very straightforward and should be accepted after some minor corrections. There are a few issues I raise in the paper and I have made lots of small corrections to the English. One thing I would recommend, is to include some yearly timeseries distributions of a couple of sites of PWV from different stations, as in Figure 12, but just for one year so that the reader can have a better idea of the accuracy and general climate in terms of humidity. Also, I thought your limits of PWV maximum and minimum values are a bit extreme (see below)
Specific Comments.
Line 25. You should probably be a bit more precise here instead of just referring to "water vapor". With respect to water vapor as a variable, what is typically most valuable for modeling, weather prediction, global climate studies is its vertical distribution and the total column water vapor or "precipitable water vapor".
Line 32. There have also been numerous field campaigns around the world employing GNSS meteorology, you should mention some from different regions of the world. I will let you choose and make no specific recommendation.
Line 47. Write "Recent research has utilized GNSS ..."
Line 48. What do you mean " project proposing water vapor products from ..." ? This idea is unclear.
Line 74. Write " ...providing reference positions for coastal research..."
Line 87 Write " At the outset, only observations from the GPS and GLONASS satellite constellations were available."
Line 87 Write " In recent years, with the advancement of the Galileo..."
Line 99 Write" ...University (Shi et al., 2008; Liu and Ge, 2003) using the static precise point position ..."
Line 102 Write " and an elevation-dependent weighting function was applied."
Line 110 Write "...ZTD consists of a hydrostatic part..."
Line 111 Write " ...(pressure and temperature) provided by Global Pressure and Temperature...,"
Line 114 This idea is a bit unclear. What do you mean by " Batch least-squares estimator" ?
Line 127 c. Validation of GNSS ZTDs based on ERA5 products
You do not have any local surface meteorological stations collocated or near the GNSS antennas?
You could use these surface met. stations for the surface pressure and to derive Tm with a simple model and then calcuate PWV. This would be good to compare against ERA5 since these ERA5 data are very smoothed in some respect (~ 25km x 25km grid
Line 164 Your PWV limiting values for outliers are very strange (0.72 mm and 86.21mm)
It is not physically possible to have PWV values near 0 nor near 90mm. Even under typhoon/hurricane conditions, the maximum PWV should be near 80mm at the highest. And PWV can never be near 0mm in these region under any conditions.
Line 166 " In addition, a station-specific outlier detection method was employed."
Line 167 " the median PWV value was calculated within a 15-day moving window centered on the specific day." Why do you employ such a long moving window? It is typically hourly or daily change in PWV that is of interest.
Line 182 I would not call ERA5 Numerical Weather model data. It is reanalysis data which include both observations and model output.
Line 240 " Additionally, the quality of the RS profiles may contribute to the larger biases observed."
Another thing to consider is that the RS can be biased if they rise through cloud/rainy conditions leading to higher PWV values than the GNSS PWV which has a large cone of observation ( ~20km diameter) can may contain clear skies in addition to the cloudy/rainy skies. These "saturated" soundings can be easily identified visually,
Line 325 Write "In addition, the spatiotemporal characteristics of coastal PWV in China were analyzed" And again, 0mm PWV values are not possible, there should always be a couple of mm of PWV even in very cold, dry weather in this region.
Citation: https://doi.org/10.5194/essd-2025-24-RC1 -
AC1: 'Reply on RC1', Zhilu Wu, 24 Apr 2025
reply
Dear Reviewer:
Thank you for the constructive and encouraging comments regarding our manuscript. We have enclosed a carefully revised manuscript according to the comments and suggestions provided, and provide an item-by-item response to all comments in the accompanying rebuttal document.
Best
Zhilu Wu
-
AC1: 'Reply on RC1', Zhilu Wu, 24 Apr 2025
reply
-
CC1: 'Comment on essd-2025-24', Zhangfeng Ma, 23 Apr 2025
reply
This paper presents a decade-long dataset of GNSS-derived Precipitable Water Vapor (PWV) from the China Coastal region, offering a significant contribution to research on climate change and weather forecasting. I find the manuscript to be well-written, with rigorous processing steps for generating the GNSS product. I recommend minor revisions prior to acceptance, as I have only a few minor concerns outlined below:
Line 56: Wang et al. (2017) also utilized the CGN station to generate a PWV product. Could you specify the differences in the PWV generation algorithms between your study and theirs?
Line 97: The GNSS observations have a sampling rate of 30 seconds, yet the final GNSS PWV product is provided at an hourly resolution. Could you explain the rationale for this choice?
Line 115: The statement, "The zenith wet delay (ZWD) was processed as an unknown piecewise constant parameter with an interval of 60 mins. The tropospheric horizontal gradient was processed as an unknown piecewise constant parameter with an interval of 720 mins," could be clearer. Please revise this sentence to improve readability and precision.
Line 250: I observed that data completeness is relatively lower at some stations. Could you provide an explanation for this?
Line 270: The maximum PWV values reported appear quite large. Have you compared these with findings from other studies or datasets for validation?
Line 306: Please refine the incomplete expression, "As shown in Fig. 15(a), a 1 K increase in SST results in a PWV increase ranging from …" Please refine the sentence.
Citation: https://doi.org/10.5194/essd-2025-24-CC1 -
AC2: 'Reply on CC1', Zhilu Wu, 28 Apr 2025
reply
Thank you for the constructive and encouraging comments regarding our manuscript. We have enclosed a carefully revised manuscript according to the comments and suggestions provided, and provide an item-by-item response to all comments.
Line 56: Wang et al. (2017) also utilized the CGN station to generate a PWV product. Could you specify the differences in the PWV generation algorithms between your study and theirs?
Response: Thank you so much for your suggestion. The contribution done by Wang et al. (2017) only used three years observation with much less stations. Moreover, the algorithm for PWV retrieval in the contribution is not state-of-art as this one, specific the algorithm for the ZHD and Tm, which is the key parameter for PWV retrieval. Besides, a more strict quality control is conducted
Line 250: I observed that data completeness is relatively lower at some stations. Could you provide an explanation for this?
Response: Thank you for highlighting the issue of data completeness. The stations with the largest gaps are located on small, unstaffed islands along the coast. Because these sites can be reached only by boat, any receiver outage caused by power loss, salt-spray corrosion, or storm damage may persist for weeks to months before a maintenance team can safely access the equipment. These logistical constraints, rather than deficiencies in the processing chain, are the primary reason for the reduced data availability noted at those stations
Line 115: The statement, "The zenith wet delay (ZWD) was processed as an unknown piecewise constant parameter with an interval of 60 mins. The tropospheric horizontal gradient was processed as an unknown piecewise constant parameter with an interval of 720 mins," could be clearer. Please revise this sentence to improve readability and precision.
Response: Thank you so much for your suggestion. Amended.
Line 270: The maximum PWV values reported appear quite large. Have you compared these with findings from other studies or datasets for validation?
Response: Thank you so much for your suggestion. Our data shows PWV reaching up to 86.21 mm, and we note that values exceeding 90 mm are not unprecedented in extreme weather scenarios. Studies such as Gao et al. (2024) and Zhao et al. (2018) have recorded PWV measurements surpassing 90 mm during typhoon events, leveraging high temporal resolution data (5-minute intervals). These observations suggest that under intense atmospheric conditions, PWV can indeed exceed the 80 mm threshold you mentioned. To enhance clarity and provide further support for these findings, we have updated the manuscript to include citations to Yuan et al. (2023), Gao et al. (2024), and Zhao et al. (2018). These references should offer additional context for the observed PWV variability.
Zhao Q, Yao Y, Yao W. GPS-based PWV for precipitation forecasting and its application to a typhoon event[J]. Journal of Atmospheric and Solar-Terrestrial Physics, 2018, 167: 124-133.
Gao Y, Wang X. Analysis of the Response Relationship Between PWV and Meteorological Parameters Using Combined GNSS and ERA5 Data: A Case Study of Typhoon Lekima[J]. Atmosphere, 2024, 15(10): 1249.
Yuan P, Blewitt G, Kreemer C, et al. An enhanced integrated water vapour dataset from more than 10 000 global ground-based GPS stations in 2020[J]. Earth System Science Data, 2023, 15(2): 723-743.
Line 306: Please refine the incomplete expression, "As shown in Fig. 15(a), a 1 K increase in SST results in a PWV increase ranging from …" Please refine the sentence.
Response: Thank you so much for your suggestion. Amended.
Citation: https://doi.org/10.5194/essd-2025-24-AC2
-
AC2: 'Reply on CC1', Zhilu Wu, 28 Apr 2025
reply
-
RC2: 'Comment on essd-2025-24', Anonymous Referee #2, 01 May 2025
reply
The manuscript „China Coastal GNSS Network …” submitted by Wu et al. described PWV determination and results for a network of GNSS stations located along the Chinese coastlines. The paper is well written and supported by figures and numbers. Nevertheless, I have some general questions and suggest modifications to improve the draft before submission.
Major questions:
- The authors processed GNSS data from 55 stations; however, the PVW results are available, but the data and metadata are not. It would be beneficial to get access to the GNSS data and to know whether there are plans to release this important dataset. In any case, I recommend updating the station table in the dataset with more accurate locations (currently roughly at the 10km level), start/end dates, and equipment for the stations.
- The title contains “climate,” but the authors processed only the years 2009-2019. At least 20-30 years of data are required to derive climate trends. I expect the authors to extend the dataset to 2024 (at least 15 years) or provide a profound explanation if this is impossible.
- I suggest adding more information to the PWV records in the dataset. Why not provide the inputs taken from ERA4 (Eqn. 3 and 4)? Furthermore, I miss the uncertainty information. This is a minor detail, but I wonder about the different spacing of the PWV values in the dataset (+/—3s). Is there a reason for this?
Minor comments:
- Is there a particular reason for using the ESA GNSS products?
- Is there a public source for the radiosonde data? Same for the sea surface temperatures.
- Provide a consistent description for the GNSS processing with suitable references. A reference to ESA products is missing. The reference for antenna information is outdated. Add the length of the troposphere interval (1h) [l115].
- I somehow miss the climate aspect of the discussion - the trend estimation against SST covers, to my understanding, seasonal variations.
Citation: https://doi.org/10.5194/essd-2025-24-RC2
Data sets
China Coastal GNSS Network: Advancing Precipitable Water Vapor Monitoring and Applications in Climate Analysis Zhilu Wu and Bofeng Li https://zenodo.org/records/14639032
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 320 | 50 | 12 | 382 | 12 | 10 |
- HTML: 320
- PDF: 50
- XML: 12
- Total: 382
- BibTeX: 12
- EndNote: 10
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1