the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 09 Aug 2023
Verifying triple and single Doppler lidar wind measurements with sonic anemometer data based on a new filter strategy for virtual tower measurements
Abstract. In this study, we compare the wind measurements of a virtual tower triple Doppler lidar setup to those of a sonic anemometer located at a height of 90 m above ground on an instrumented tower and with those of two single Doppler lidars to evaluate the effect of the horizontal homogeneity assumption used for single Doppler lidar applications on the measurement accuracy. The triple lidar setup was operated in a 90 m stare and a step/stare mode at six heights between 90 and 500 m above ground, while the single lidars were operated in a continuous scan Velocity-Azimuth-Display (VAD) mode where one of them had a zenith angle of 54.7 ° and the other one of 28.0 °. The instruments were set up at the boundary-layer field site of the German Meteorological Service (DWD) in July and August of 2020 during the FESST@MOL (Field Experiment on sub-mesoscale spatio-temporal variability at the Meteorological Observatory Lindenberg) 2020 campaign. Overall, we found good agreement of the lidar methods for the whole study period for different averaging times and scan modes compared to the sonic anemometer. For the step/stare mode wind speed measurements, the comparability between the triple lidar and the sonic anemometer was 0.47 m s-1 at an average time of 30 minutes with a bias value of -0.34 m s-1. For wind speed measured by one single lidar setup for the same period with an averaging time of 30 minutes, we found a comparability of 0.32 m s-1 at an averaging time of 30 minutes and a bias value of 0.07 m s-1 and values of 0.47 m s-1 and -0.34 m s-1 for the other one, respectively. We also compared the wind velocity measurements of the single and triple lidars at different heights and we found a decreasing agreement between them with increasing measurement height up to 495 m above ground for the single lidar systems. We found, that the single Doppler lidar with the increased zenith angle produced a poorer agreement with the triple Doppler lidar setup than the one with the lower zenith angle, especially at higher altitudes. At a height of 495 m above ground and with an averaging time of 30 minutes the comparability and bias for the larger zenith angle were 0.71 m s-1 and -0.50 m s-1, respectively, compared to values of 0.57 m s-1 and -0.28 m s-1 for the smaller zenith angle. Our results confirm that a single Doppler lidar provides reliable wind speed and direction data over heterogeneous but basically flat terrain in different scan configurations. For the virtual tower scanning strategies, we developed a new filtering approach based on a Median Absolute Deviation (MAD) filter combined with a relatively relaxed filtering criterion for the signal-to-noise-ratio output by the instrument.
- Preprint
(2304 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2023-1704', Anonymous Referee #1, 20 Sep 2023
The authors conducted an observational validation and statistical comparison study between various Doppler lidar scanning strategies within the planetary boundary layer during a northern Germany field campaign in Summer 2020. Notably, triple and single Doppler lidar setups are compared against a tower-mounted sonic anemometer to determine the relative accuracy of each type of measurements at different heights, while a new filtering method is introduced for the virtual tower scanning strategy performed by the triple lidar setup. The rigorous statistical analyses provide a detailed understanding of the lidar measurements conducted and the efficacy of the filtering technique.
Nevertheless, the manuscript has grammatical errors throughout that should be addressed before the next submission. And probably more importantly, the uniqueness and new contributions for the field of lidar measurements and the usefulness of the new filtering method should be written more succinctly and emphasized more often. With these overall changes in mind, I return the paper to the authors for major revisions.
Specific Comments
Introduction: I think that the uniqueness of your triple DL setup and VT scan mode should be placed better in the context of past studies. How is your setup different? How are your objectives different than past work? This will allow the reader to better understand the motivation for the DL setup and scanning methods, and provide a basis for the reader to decipher how the Results shown later are useful to the community.
Starting with Line 98: I think the sentences starting with “Our objective…” should be a separate paragraph since these sentences clearly outline your paper’s objectives, and having them be a separate paragraph will draw more attention to them. Perhaps consider rephrasing some of these sentences as ‘Research Questions’ that are numbered and formulating your discussion in the Results section around the Research Questions.
Tables 2, 3, 4, 5: I think it would be more beneficial to show the “Number of Observed Periods” instead as a percentage of the total number of possible observed periods in order to better emphasize the amount of data removed when applying a certain filter.
Given that you have 7 Tables in this manuscript, I wondered if it could be possible to change one or multiple to a figure? Perhaps a box and whisker plot or a bar plot with error bars would be possible? I feel that it would perhaps be easier to digest the comparisons if less are in tabular format.
Lines 497-499: I would change the end of this sentence to indicate that ‘the VT wind measurements perform equally well to wind measurements from the VAD DLs’ since I think you want to show that using the VT method is a reputable method, correct?
Lines 516-518: Sentences like this make it more difficult to understand the objectives of the paper. Are you trying to show that the VT method is better and therefore should be used more widely in the community? Are you trying to indicate that a single DL can be used instead of a triple DL setup? If your answer is no to these questions, then this sentence would seem to undercut your hypotheses.
Line 553: Change “proofing” to “proving” if you would still like to keep this word. However, I would additionally suggest being careful from using this more absolute language. Using words like “indicating” or “reinforcing” I think tends to be safer since as scientists it is quite difficult for us to definitively show that a certain method is wholly justified.
Lines 568-569: These two sentences seem to undermine the new filtering method you have outlined in this study. If the advanced filters do not improve the resulting comparison with the sonic anemometer data, would that not suggest that your filtering method should be altered? Or that the older filtering methods are still adequate to use?
Technical Comments
I would suggest indenting each paragraph or placing a blank line between paragraphs to improve readability throughout.
Lines 23 and 25: “average time” and “averaging time” are both used in the same sentence position. However, for consistency, only one of these should be chosen.
Lines 28-29: “We found, that…” should be changed to “We found that”.
Lines 43-46: “…can contribute to further increase the quality of the weather forecast…” could be written more succinctly. Perhaps “…can contribute to further increases in weather forecast quality…”.
Line 53: Change “To determine wind speed and direction from a single DL a set of different scanning methods are well-established…” to something like “There are a set of well-established scanning methods to determine wind speed and direction from a single DL”.
Line 57: I think it is generally not good practice to start a sentence with “that”. I would suggest changing to something like “Alternatively, the 3-D winds can be obtained using different scanning methods like…”.
Line 70: Change “especially, when” to “especially when”.
Line 84: Change “respective DL they found” to “respective DL, they found”.
Line 117: “number” can be removed to be more concise.
Line 118: Change “study,” to “study:”.
Line 153: Change “…for a comparison with the sonic anemometer” to “for comparison with the sonic anemometer”.
Line 161: Change “Both these scan modes” to “Both of these scan modes”.
Line 181: I would suggest identifying (1) as “Equation (1)” when used in the text. Also, if possible, I think centering the equations horizontally would be more correct formatting to use.
Line 202: Change “In Figure 3” to “In Figure 3,”.
Figure 3: Change “Overview over the meteorological conditions” to “Overview of the meteorological conditions”.
Line 216: Change “from the north” to “from north”.
Line 220: Change from “measurements” to “measurements,”.
Line 223: Change “assure” to “ensure”.
Line 231: Change “(see Table 2)” to “(see Table 2),”.
Line 232: Change “…filter which…” to “…filter, which…”.
Line 233: Change “…30 % that…” to “…30%, that…”.
Line 244: Change “30 minutes averaging interval” to “30-minute averaging interval”.
Lines 267-268: Change “When applying the MAD filter to constantly moving scanning strategies an increase” to “When applying the MAD filter to
constantly moving scanning strategies, an increase”.
Lines 268-269: Change “If the actual radial velocity is around 0 ms-1 the filter” to “If the actual radial velocity is around 0 ms-1, the filter”.
Table 2: Change “Person’s” to “Pearson’s”.
Line 282: Change “all the filters” to “all of the filters”.
Line 293: Change “as in the study of” to “as in”.
Lines 295-296 (similarly for Lines 297 and 300): I believe that standard figure citations should be “Figure 5a” and “Figure 5b” in this sentence.
Line 300: Change “To” to “to”.
Figure 6, right y-axis label: Change “m s-1” to “(m s-1)”.
Line 328: Change “sonic anemometers” to “sonic anemometer”.
Line 405: Change “time of the day” to “diurnal cycle”.
Lines 406-407: This sentence is probably not necessary. I suggest simply mentioning “sonic anemometer” in the previous sentence.
Lines 411 and 524: Should “sinus-wave-like” be “sine-wave-like”? I understand that sinus is likely derived from “sinusoidal”, but I believe that the most typical usage is “sine” instead of “sinus”.
Lines 414-415: I do not think that this sentence is really necessary. I would expect that a reader would be able to understand that the x-axis between two plots that show different variables may have a different extent.
Line 438: Change “comparisons this” to “comparisons, this”.
Line 445: Change “stability we” to “stability, we”.
Line 452: Change “We had to choose relatively relaxed filter criteria” to “We had to choose relatively relaxed filtering criteria”.
Lines 459-460: I think that “as an example day” can be deleted.
Line 471: ‘since only one instrument must be operated’ can be deleted for succinctness.
Lines 471-474: ‘Moreover, the setup of the instruments presents another error source when using multiple DLs compared to a single instrument, since the north orientation has to be carefully verified for all three systems. Hence, single DLs are certainly easier to handle and more cost-efficient than using a three DL setup’ can be deleted. In my opinion, this information should be fairly intuitive to the reader.
Line 484: Change ‘…VT mode another measurement regime’ to ‘…VT mode, another measurement regime…’.
Lines 513: Reword this part “based on the with height increasing spatial averaging area for the VAD measurements” to be grammatically correct and add a comma at the end of this phrase.
Lines 513-514: Reword “the VT measurements present the actual state of the wind field” to be grammatically correct.
Lines 560-561: Adding “they also give reliable mean wind values as well” onto the end of this sentence seems to make it a run-on sentence. I would suggest removing this phrase.
Lines 583-584: I would suggest making this portion more succinct – “other RHI-related scan modes should be chosen for the triple DL or a single DL should be used to perform a VAD scan or similar pattern”.
Line 592: Change “further insides” to “further insights”.
Line 593: Change “multi” to “multiple”.
Lines 594-595: Remove “6 Code availability” from the end of this paragraph.
Citation: https://doi.org/10.5194/egusphere-2023-1704-RC1 -
CC1: 'Comment on egusphere-2023-1704', Vincent Baltazart, 22 Feb 2024
The authors conducted a field campaign comparing various Doppler lidar scanning strategies for measuring winds (speed, direction) within the planetary boundary layer.
The paper highlights the competence and the scientific soundness of the authors in this field.
However, I suggest to select a paper title which better reflects the main idea of the work and/or the new contribution in the field.The word "Verifying" may be definitely replaced by "comparing" at least.
According to my understanding of the paper content, I may suggest the following title:
Assessing the single Doppler lidar method for reliable/accurate wind speed measurement within the planetary boundary layer.
A new title may help the authors to give a common thread to the paper, and may inspire some modifications to the paper structure accordingly.As for the first reviewer, I then suggest major revisions.Most typesetting errors have been already listed by the first reviewer.
Citation: https://doi.org/10.5194/egusphere-2023-1704-CC1 -
RC2: 'Comment on Verifying triple and single Doppler lidar wind measurements with sonic anemometer data based on a new filter strategy for virtual tower measurements', Anonymous Referee #2, 07 Mar 2024
The authors present data from a field campaign in which triple-Doppler lidar measurements wind measurements were compared with those from a sonic anemometer on a 90-m tower and standard velocity-azimuth display (VAD) wind profile measurements from two different lidars operating in separate modes. Robust statistical analysis is presented, particularly of the triple-Doppler virtual tower measurements, showing the quality of the data. Additionally, the authors present a new data filtering methodology.
While the results appear to be scientifically sound, there are concerns over the uniqueness and significance of the study. There have been numerous past similar studies evaluating wind measurements from single and multi-Doppler lidars that the authors briefly review in the introduction. However, the authors need to do a better job of integrating their findings with respect to the literature and previous findings throughout the entire paper. Additionally, the authors should state up front at the end of the introduction what is unique about their analysis compared to the previous studies, as it appears very similar. As such, I recommend this paper be reconsidered after major revision in which the above general and following specific comments, building upon other reviewers’ concerns, are addressed.
Specific Comments
- Figure 2: It would be extremely helpful to add a scale to the figure in the lower left or right to show distance.
- Line 135: I appreciate the methods used here to assess the azimuth angle with respect to true north as a robust way to determine the heading. Was a similar approach scanning up/down hard targets used as well to determine the pitch/roll of the lidar to assure it was level so that the elevation angle is known precisely (particularly over the virtual tower)?
- Line 203: The data used here covers nearly two months; it would be helpful to show the precipitation accumulation or rate as well. Furthermore, was data excluded for time periods when there were precipitation? During these time periods, the data may appear ok but the DL is really measuring the hydrometeor instead of aerosol movement, which is quite different than the air speed (particularly the vertical fall rate). Strongly suggest removing any periods of precipitation from subsequent analysis.
- Line 215: ‘… western side of the lattice tower and pointing towards 190 deg’. Is this correct? If the sonic were on the western side, I’d expect it to be around 270 deg, as 190 would be on the southern side if I’m interpreting this correctly. Should it be the ‘southern side’?
- Line 249: What is the cause of the erroneous radial velocity values around 0 m/s? Is it known, such as due to hard target hits?
- Line 271: Given the various stages of the data quality editing and filtering (e.g., SNR threshold, dip statistic, MAD, different time windows), it would be helpful to add a block diagram showing the flow of the different steps as a new figure.
- Line 304: Could you add a brief description here of why the radial velocity values from the DL in Figure 5 (both panels) appear to have vertical lines of a large number of points at around 1, 3, and 5 m/s? This looks unrealistic, but maybe it’s some artifact of underlying shear in the time period presented.
- Line 313: Could you succinctly explain why the VAD coverage is so poor at night? This is unexpected given the higher coverage from the VT, thus aerosols are present at higher altitudes and there are no clouds attenuating the signal. Is the poor coverage an artifact of the different filtering for the VAD measurements compared to the step/stare, due to the smaller number of pulses/rays, or something else? In general, throughout the entire paper, I feel like the discussion of the VAD results is lacking and insufficient consideration is given to fundamental differences in the lidar scanning parameters (zenith angle, pulse accumulation, focus length, etc.) with regard to explaining the results in the entire study.
- Figures 8/9: Recommend combining these two figures together with different coloring for what is currently in each figure to differentiate the two. For example, overlay the data of Figure 9 on current Figure 8, changing the colors of the data to red or some other color to differentiate the two. This will make it easier to visually see the differences.
- Lines 412-415 and Figure 10: Suggest removing the figure and discussion as it does not add any value to what is presented. It is already clear that there is more disagreement between VT and VAD 1 at higher altitudes from previous data in the table, and data in Figure 10 and can be largely gleaned from what’s in Figures 8/9.
- Lines 461: Are these higher radial velocities associated possibly with second trip echoes or possibly contamination of clouds or other artifacts in the background noise (see Manninen et al 2016)?
- Line 479: Here, should specifically call out increasing the number of pulses/ray for additional integrations to increase the range of the DLs operating in VAD.
- Line 589: The current dataset from this experiment would actually be conducive to analyzing a double-DL setup as well. Simply remove the vertically pointing lidar from the analysis and it could be easily done.
- Line 595: The words ‘6. Code availability’ should be removed here as the section is below it.
References
Manninen, A.J., O'Connor, E.J., Vakkari, V. and Petäjä, T., 2016. A generalised background correction algorithm for a Halo Doppler lidar and its application to data from Finland. Atmospheric Measurement Techniques, 9(2), pp.817-827.
Citation: https://doi.org/10.5194/egusphere-2023-1704-RC2 - AC1: 'Comment on egusphere-2023-1704', Kevin Wolz, 12 Apr 2024
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 245 | 129 | 32 | 406 | 23 | 24 |
- HTML: 245
- PDF: 129
- XML: 32
- Total: 406
- BibTeX: 23
- EndNote: 24
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1