the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Doppler power spectrum processing methods for a vertical sensing 94 GHz millimeter-wavelength cloud radar
Jie Wang
Zhenhua Chen
Junxiang Ge
Abstract. A set of Doppler power spectrum processing methods for TJ-II, a vertical sensing 94 GHz millimeter-wave cloud radar(MMCR), is proposed to distinguish clouds and enhance the data quality. The noise level is estimated by a modified segment method with 2-D segments. A two-step cloud signal mask method is proposed to distinguish clouds and noise. A Gaussian filter with adaptive standard variance is used to improve detection performance at the boundary of clouds and noise. Square signal blocks were constructed to test our method. Velocity dealiasing is carried out combined with a pre-dealiasing processing and the dual Pulse Repetition Frequency (PRF) technique to address a particular phenomenon called half-folded in MMCRs, which can not be fixed by post-processing at the base datum stage. Some observations of TJ-II are used to demonstrate our method. A comparison between our method as pre-processing and the method as post-processing proposed by Key Laboratory for Semi-Arid Climate Change of the Ministry of Education and College of Atmospheric Sciences, Lanzhou University, was carried out using one-day observation. It was found that our method shows some advantages in cloud base detection.
Hai Lin et al.
Status: final response (author comments only)
-
RC1: 'Comment on amt-2022-302', Anonymous Referee #1, 01 Jan 2023
Review of the manuscript “Doppler power spectrum processing methods for a vertical sensing 94 GHz millimeter-wavelength cloud radar” by Hai Lin et al.
The manuscript presents a processing chain of a W-band radar. The processing chain includes several stages: (a) noise level estimation, (b) signal detection including spatial and temporal filtering for improved sensitivity, (c) dealiasing based on dual pulse repetition frequency. The described routines are applied to measurements from the vertically pointed W-band radar developed by the authors.
The manuscript has a number of weaknesses. Here are the major ones:
- The lengthy introduction section covers irrelevant details of general cloud research but contains absolutely nothing about the actual problem addressed in the manuscript. The introduction must include a general overview of problems in the data processing, literature review on this topic, and explain which of these problems are solved in the study. In the current status, the motivation of the manuscript is not defined.
- The manuscript describes already existing methods slightly fine-tuned for the presented radar system. In my opinion, there is a severe lack of novelty in the manuscript. (1) In the section 3 the authors apply the segment method developed decades ago, the only difference is that the method is applied not just to a spectrum in a single range bin but to several range bins. It is hard to consider this as a novel approach. Also, it is not clear why the authors try to solve the noise estimation topic at all, since with modern computers there is absolutely no problem in applying the Hildebrand-Sekhon algorithm in real time. (2) In the section 4, the authors suggest using a time-space filter to improve sensitivity. Such filters, however, have been routinely applied for decades (e.g. Clothiaux et al. 1995 JAOT, Marchand et al. 2008 JAOT). The performance of the authors new filter is very close to the Gaussian one. For a single presented case study (which is hard to consider as statistically significant), FAR and MDR are nearly the same (just a few % difference). (3) In the section 5 the aliasing problem is solved by applying the dual pulse repetition frequency technique which has also been used for decades. Authors themselves give a long list of references. The spectral dealiasing is also not a new thing (e.g. Kuehler et al. 2017, Maahn and Kollias 2012).
Taking into account the two above-mentioned items, I, unfortunately, cannot recommend the manuscript for publication and suggest a rejection.
Citation: https://doi.org/10.5194/amt-2022-302-RC1 - AC1: 'Reply on RC1', Hai Lin, 07 Jan 2023
-
RC2: 'Comment on amt-2022-302', Anonymous Referee #2, 08 Mar 2023
This manuscript presents a step by step procedure to process Doppler spectrum data collected by a millimeter wavelength weather radar. The presentation is clear and detailed but it is unclear what is new in the proposed technique.
In the first step of the processing for example the noise floor is estimated using the standard segment method, but extended to multiple range gates. Seems like an insufficient contribution to prior publications. Also, addressing such a mature topic should include a comparison of other techniques based on the quality of results. The well known Hildebrand and Sekhon (1974) (H-S) noise estimation technique for example is mentioned to require significant calculations. This may have been a deterent for real-time processing in 1974, but since then computers have significantly improved so a comparison using real and simulated data of the proposed 2-D segment method would be interesting. If the H-S method is still computationally too demanding, then let's show that.
It is also not clearly stated if any new technique is proposed in the second step of identifying the cloud signals. It seems that this processing sequence follows the two references provided, yet it is presented with great detail like this was a novel method. Similarly the velocity dealiasing using dual PRF is not new. The presentation needs to make a clearer distinction between what's new and when an existing technique is applied - when the reference is sufficient instead of a lot of detail.
Overall this is nice work, but it is more like a documentation of the radar than a journal publication advancing the state of science.
Citation: https://doi.org/10.5194/amt-2022-302-RC2 - AC2: 'Reply on RC2', Hai Lin, 21 Mar 2023
Hai Lin et al.
Hai Lin et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 241 | 71 | 17 | 329 | 7 | 7 |
- HTML: 241
- PDF: 71
- XML: 17
- Total: 329
- BibTeX: 7
- EndNote: 7
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1