Quantification and Correction of Wave-Induced Turbulence Intensity Bias for a Floating LIDAR System
Round 1
Reviewer 1 Report
Overview
The authors present a method to increase the precision of the estimation of turbulence intensity using floating wind lidars. The method is based on the coupling of the 10-minute mean wind statistics with high frequency observations of the motion of the buoy. The presented methodology has the advantage that doesn't require the storage of high frequency wind lidar data and their synchronisation with motion data, which in relation to similar research studies, leads to a simplified way of compensating the impact of the motion of the buoy. The method and the results presented in this manuscript are interesting and relevant to the wind lidar research in wind energy applications. Please find below the comments and suggestions of my review, which I hope that will help improving the comprehension of the study and the interpretation of the presented results.
General Comments
- “LIDAR” is a general term that describes a remote sensing technique that it is not exclusively measure wind. Therefore, I suggest that the authors explicitly refer to “wind lidar(s)” in the introduction instead of just “LIDAR”
- The authors use the term “relative velocity” to refer to the radial wind speed measurements. This is not a widely used term. I suggest replacing it with a more appropriate term as “radial wind speed” or “radial velocity of the wind”. Otherwise, I suggest explaining what is meant with the term “relative velocity”. First time that it is used in the text in on line 59.
- The authors refer to a 10-minute period as “10min” and as “10-minute” please use the same term throughout the document. Also, please have always a space between numeric values and their corresponding units
- The authors have entitled section 3.1 “A model for the estimating the wave-induced variance in the LIDAR measurements”. However, there is not really a model presented here, just a common definitions of wind speed statistics. I suggest renaming this section with a more descriptive title.
- In the manuscript is included a section where a theoretical study of the impact that the motion of the buoy has on the mean and the turbulence intensity estimation. Especially the sub-sections 4.2.1.1 and 4.2.1.2 are presenting simplified studies of how the motion of the buoy impacts the mean radial wind speed. I do not understand how is this relevant to the study of the TI bias. I recommend that the authors should consider moving part or all of this section to the appendix and state their main conclusions from this study in the main manuscript.
- The abstract contains a detailed summary of the method used, but it does not include any information about the results and the conclusions of this study. I suggest that this should be added in the abstract.
- The authors state in the lines 296-298, they state the method presented in this manuscript is based on simplified hypothesis in relation to other already published studies. I think that it is going to be very helpful for the understanding of the value of this method, to discuss in the results section how does this method perform in relation to other already published studies. This does not have to be an extensive review, but a short comment on how does the presented improvement in the estimations of TI compare to other models/methods.
Specific Comments
Line 29 Do the authors mean that the wind power sector has been growing for the past 30 years or is a verb missing after the “has been”. I suggest simplifying and clarifying this sentence.
Lines 31-33: Please add a reference here.
Lines 34-36: The way that the sentence is structured it is understood that wind fluctuation decrease the structural fatigue. Please clarify this sentence.
Line 41: This is true only for the case of wind lidars. I suggest adding the word “wind” in between “profiler LIDAR” and replace the “a vertical wind speed” with “the vertical wind speed”. Furthermore, since the lidar is used as an acronym and denoted with capital letters, I suggest presenting the full description of the acronym the first time is mentioned.
Line 49: it is written that “a large part of the planet is still unknown”, please specify that you mean in terms of available wind resources or wind conditions.
Line 52: It would be helpful to the reader if you can also mention what are the 6 degrees of wave-induced motion.
Line 53: Please add a reference in the sentence ending with “impacted by sea motion”.
Line 55: “on pulsed LIDAR technology “. I suggest re-write as: “on the pulsed LIDAR technology”.
Line 57: Can you please check that this is the correct reference?
Line 60: “… between the two devices or on a storage …” the word “or” should be replaced by “and”. It is necessary to store the high frequency data in order to perform the analysis presented in Kelberlau et al. 2020.
Lines 59-61: The authors state that a method to correct the measurements of the wind vector using a wind lidar was introduced by Kelberlau et al. 2020. However, they omit mentioning a lot of other studies that they have presented something similar methods. I recommend to either include more references here or mention that a detailed review of previous studies is presented in Kelberlau et al. 2020. Please note that I am not neither the author or one of the co-authors of the Kelberlau et al. 2020 article.
Line 75 – 77: Please add a reference here
Line 78 – 79: This statement is not correct, coherent wind lidars are not comparing the frequency of the backscattered radiation to the transmitted one in order to estimate the Doppler frequency. This is implemented in the direct-detection wind lidar (see for example sections 7.4.1 and 7.4.2 in Henderson, Sammy W., et al. "Wind Lidar." Laser Remote Sensing. CRC Press, 2005. 487-740.
Line 80: ”we focus on pulsed” please rewrite as ”we focus on the pulsed”
Line 83: The author state that the WindCube v2 uses an optical scanner to emit laser pulses to each of the five fixed directions. Can the authors describe how does this optical scanner work in order to emit radiation both vertically and with an elevation angle of 28 degrees?
Line 84: ”… measured from the north of the LIDAR … ” This statement can be confusing. Please specify that is a the “north” here is just a reference axis that can have an offset from the magnetic North (as accurately stated in line 102)
Line 88: Please add the symbol used to refer to the LOS relative velocity
Line 91: Please specify here that the “u”, “v” and “w” components of the wind vector are specified in the coordinate system used in the lidar and they are not referred to an absolute coordinate system.
Line 114 – 115: A weighting function is describing the filtering effect that the probe volume has on the measured radial wind speed. Please clarify what is meant with this sentence and also add a reference.
Line 115-117: Is the mean radial wind speed calculated at the end of each scan or at the end of each line-of-sight measurement? Please add a reference regarding the signal processing of the WindCube measurements. The reference [15] could be used here.
Line 119: Please provide a description for the acronym DBS
Lines 123-126: Please explain why the component of the line-of-sight unit vector has a negative sign in equation (6) and subsequently why the radial wind speed is then expressed as a sum of positive terms.
Lines 137 – 138: The reference added here concerns an inter-comparison between 3 staring pulse lidars and a sonic anemometer it is not relevant to the estimation of the instrumental error of modern commercial lidar systems.
Line 140: The motion of the buoy has an effect on the measurements of the lidar, it does not alter them.
Line 141: Please explain what is meant with the term “measurement contamination”.
Line 143: I suggest replacing the verb “appreciate” with “quantify”
Line 151: The rotation matrix presented in equation 9 is not the same as the one presented in equation (6) of the reference [19]. Please specify what types of rotations have been applied to result to equation (9)
Line 153: “The rotation is applied …” please specify what does this rotation represent.
Line 157: The authors state that the lidar and the IMU have been aligned and thus no offset is expected between the two instruments. Can the author specify how was this alignment performed between the two devices?
Line 178-179: Why do the authors use the word “Therefore” and how is the 10-minute mean wind speed simulated? Do the authors mean “simulate” or “examine”? Also, if it is not a specific 10-minute mean period then the “during the 10-minute measurement period” should be changed to “during a 10-minute measurement period”.
Line 182: The equation 12 is inconsistent with equation 1. One presents a transpose matrix.
Line 187: Equation 13: Please specify what is “ti” and what is the “Vhor”
Line 190: Please add references regarding the approach used here for the estimation of the variance.
Line 191: Please explain more this sentence. It is not clear to which model this sentence refers to (an equation number could be added for example).
Lines 205 - 208: Can the authors explain how the wind shear decrease the influence of the cone geometry variation on the measured velocity? What is the “k” factor presented in the reference [16] and does this mean that the authors recommend to not consider the wind shear?
Line 213 - 222: Please specify what does the “DLOS” and “PLOS” stand for. “DLOS” is defined in line 223 but it is first used in line 213.
Line 245: Equation number is missing. Also, please specify that the components of this vector represent the velocity of the lidar along the three axes of the coordinate system used here.
Line 258: Is it correct that the terms of the translational velocity of the lidar should be divided by the sine of the elevation angle of the line-of-sight in the expressions of the “u” and the “v” component?
Line 285: Which is the “slight difference” that the authors refer to in this sentence?
Lines 296 – 298: The word “is” is probably missing after the part “However, our method …”
Line 316: I suggest that the symbols of the roll, pitch and yaw angles should replace the words “roll, pitch and yaw”
Line 321: I suggest adding a reference regarding the K-means clustering.
Line 327: How close was the closest ERA5 point?
Line 346: It is stated that wave data were sampled every 30 minutes, therefore how are the “last 30 seconds” extracted?
Lines 348-349: Can the authors clarify which is the quantity that the low and high values refer to?
Lines 356 – 358: The authors present two cases of “rough” and “calm” sea states. They show that the amplitude of the tilt angle is higher in the “rough” sea in relation to the “calm” sea. For these cases they find the same period (“about 4 to 5 seconds). However in lines 348 and 349 they state “as wave height and period are highly correlated …”. Can the author explain why in the examples presented here such a correlation is not observed? Also, how was the duration of the period quantified?
Lines 375 – 376: The authors state that: “Several phenomena overlap, making it very difficult to predict”. First, they should specify that they mean that it is difficult to predict the impact of the motion of the buoy to the measured radial wind speed. Also, the term “several phenomena overlap” is too abstract in a scientific content. I suggest to either list those parameters that they have in impact or add a reference where those parameters are described.
Lines 387: Please define the variable P.
Lines 424 – 426: Can they authors explain what is the difference between the V1 and V2 and u and v in equation (7)? Also, it is not it is clear what is actually presented in Figure 5b. How is the time-averaged value of the radial wind speed calculated?
Line 442: How is the variation of the phi and theta angles defined here?
Lines 528 – 530: It is not clear what it is meant here. The way that the model used in this study is constructed the mean horizontal wind is constant over 10 minutes. Therefore, the only impact on the radial wind speed estimation is the motion of the buoy during the 0.8 – 1.0 seconds that it takes to perform one measurement. I am not sure why the authors state that the averaging effect is added to the subsampling effect. Can you please explain this part a bit better?
Line 535: Please add the number of the equation used to estimate the “relative velocity” in Figure 9.
Lines 545-637: The authors present an expression of three wind vectors in the coordinate system of a pulsed wind lidar in equation 21. They use this equation in this section to reconstruct the horizontal wind speed. However, they don’t explain what is the time scale that they use for the estimation of each horizontal wind speed. As far as I know, WindCube is performing on estimate of the horizontal wind speed and direction after the completion of the measurement of each individual line-of-sight. Resulting in having approximately one measurement per second. Could it be that the authors perform an estimation of the horizontal wind every 4 seconds and thus the estimated value of the “resonance period”.
Lines 630-637: How is the overestimation of the TI defined?
Line 650: Please add a reference
Line 654: Can the authors explain a bit more why does the lidar underestimate TI at low TI values. They mention that this is due to the “difference in technology” and to the “low measurement altitude”. It would be very helpful to the reader to elaborate more on this.
Lines 656- 659. What do the black lines depict in Figures 12 (a) and (b)? Are they the graphical representation of the linear regression? Especially in the Figure12b, according to the linear regression the linear offset is estimated to be equal to 0.029 but visually it looks that it is almost 0.5 (estimated very roughly by finding where does the black line meet the y-axis). This comment applies also for the results presented in Figure 15.
Line 674. How is the turbulence intensity measurement error defined? The values presented in Figure 13a are quite low in relation to the scattering observed in Figure12b.
Lines 728 – 731: Can the authors explain more how do they filter the data?
Author Response
Dear Reviewer,
Please find the answer to your comments and questions in the attached file.
Regards,
Thibault Désert
Author Response File: 
Author Response.docx
Reviewer 2 Report
Review of manuscript “Floating LIDAR System: Error quantification and correction of wave-induced turbulence intensity bias” by Désert et al. / Comments
The manuscript is generally very well elaborated and presents the developed methodology in great detail. I have collected a few comments, which may help to improve the manuscript thought and would like to ask the authors to consider them thoroughly.
Comments in order of their appearance in manuscript:
[ll 2-3] Please revise the title – I think, for instance, “Error quantification… for a floating LIDAR system” would be more suitable.
[l 14] Generally, I think, you should put some more focus on the floating platform defining the ”floating WindCube”. As you also point out in your conclusions the type of platform may have a large effect on the motion impacts, so you are actually rather studying floating platform affects than wave effects on the lidar measurements. For the abstract and introduction, it would help to mention the type of platform of your case study and a bit of generalisation.
[ll 34-35] I don’t agree with your statement “It induces fluctuations in the wind’s potential energy”. Wind potential is studied on larger scales, and the same applies to its fluctuations. TI is most relevant for the consideration of turbine loading (so, your reference fits) or potentially the modelling of wind farm effects (so, this may have a link to the wind potential at least offshore).
[l 39] Only referring to “LIDAR technology” here is very specific – I would suggest to narrow this down to e.g. Doppler wind LIDAR or also mention the wind energy application in this context.
[l 43] Please specify what you mean by a “modern profiler LIDAR”.
[l 45] I would not say that “atmospheric turbulence measurement” by LIDAR “is not yet validated”. Please be a bit more specific here referring to the available overview studies and also up-to-date industry work on this topic.
[l 47] I believe these are not just “uncertainties” but also systematic biases to some assumed reference.
[ll 55-57] I don’t agree with the sentence “The turbulence intensity…” What does “0.2” refer to – a factor or bias? Please also check the reference – I don’t think it includes considerations on floating lidars.
[whole section 1.] As already pointed out above, I think the issue about the (floating) platform should also raised here. At least when mentioning the case study at the end of the sections, you should also referred to the (overall) floating lidar system, with which normally the lidar together with the platform is meant.
[Figure 1] why a capital first letter for Surge, Sway and Heave but small letters for roll, pitch and yaw?
[l 110 and l 82] Please revise the section title – how exactly do you differentiate between “principle” and “theory”?
[l 112] Are you sure that the vertical beam has an accumulation time of “1s”? I always thought there is some additional processing time after each cycle instead. Please check.
[l 123, same for following equations] Please insert same space between the single sine/cosine terms, for better readability.
[section 2.2.] I am bit surprised that you get the (good) yaw measurement from the IMU – are you sure that the IMU is the only motion measurement device on the platform?
As mentioned above, please also discuss the impact of the platform in more detail in this section.
(I have carefully read your explanation in between but no comments to them.)
[l 325] Why have you used the FLS “wind data” as a reference here and not the data of the mast or fixed lidar? Could you say a few words on this?
[l 334] Could you give a reasoning for the criterion “TI > 0.4”? why this value?
[ll 375-376] “to predict” what? Please review this sentence – I think, it is incomplete.
[Figure 12(b)] For me it is obvious that the data do not longer follow a linear relationship – please comment on this and think about another way to analysis the data (e.g. bin averages instead of linear fit).
[l 663] I don’t agree with the conclusion “which means that the error is distributed almost indepently” – I think it just indicates an offset (if the relation would be still linear).
[ll 668-670] “We also note…” I do not fully agree with this conclusion – but it is ok as an assumption. Please review and revise the formulation.
[Figure 13] Please add the dashed lines for the quartiles to the legend.
[References] Please add DOIs to the literature references.
Author Response
Dear Reviewer,
Please find the answer to your comments and questions in the attached file.
Regards,
Thibault Désert
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
I would like to thank the authors for answering all my comments and suggestions. I think that they have provided sufficient answers and clarification that they have improved the quality of the document. However, I think that there are still few things that need to be explained better. Please find below 5 points that I think that they require correction or improvement.
1. Section 2.1
As far as I know the Windcube v2 does not have an optical scanner capable of emitting the line-of-sight into five fixed directions. Unless the authors get more information about this from Leosphere, I suggest that the first sentence of this section is written:
“The WindCube v2® emits laser pulses into five fixed directions, four inclined …”
2. In Equation (1) the wind vector components have a positive sign. In Equation (6) the components of the unit vector have negative. Their inner product of the two vectors cannot be the result presented in Equation (7)
3. Section 2.2.1
The authors have added a paragraph to explain the probe volume. However:
The estimation of the probe length is not estimated by the multiplication of the pulse duration to the speed of light. See more at https://orbit.dtu.dk/en/publications/on-mean-wind-and-turbulence-profile-measurements-from-ground-base-2
The calculation of the authors lead to values that of the order of 50 m, which are higher than the ones already published. For example in https://doi.org/10.5194/amt-11-4291-2018 it is stated that the range gate resolution for a 175 ns is 10 – 20 m
4. Section 3.1.1 - Equation 21.
When the wind lidar is not moving then a radial wind speed is described by the equation 7
Vr= u sin?cos?+ v sin?sin?+ wcos?
However, as the authors also state in their reply, when the wind lidar has a translation velocity this has to be projected to the line-of-sight over which the radial wind speed is measured. Assuming that the translation wind speed Vt is equal to {ut,0,0} then Equation 7, I think, should be written
Vr= (u+ut) sin?cos?+ v sin?sin?+ wcos?
When using the above equation to estimate the wind component along the N and S axes of the coordinate system used from the lidar, then the terms that describe the contribution to the translational wind speed should not divided by 2 times the sine of the tilt angle of the line-of-sight.
5. Section 4.3.1
Figures 12b, 15 a and c
I think that the current version of the figures is confusing. I understand why the authors use normalized values but there must be a consistency between the data presented in the plots and the results of the regression analysis. I suggest that the authors should either:
a. remove completely the regression analysis results from the figures and the present the results of the unnormalized measurements in the text.
b. or include the results of the regression analysis and the corresponding graphical representation (black line) of both the unnormalized and normalized values in the graphs.
Furthermore, in the caption of Figure 12 it is written that “the ticks are normalized by the anemometer mean value”. Does this correspond to the mean TI over all the data?
Figure 13
If one tries to extent the black line to see where does it meet the y-axis when the “anemometer normalized horizontal TI” is zero then the value of the “floating lidar normalized horizontal TI” is approximately 0.5. This is offset is one order of magnitude higher than the values of the “30 min absolute FLS TI error” presented in the Figure 13 and the bias (0.029) presented in figure 12b. I guess that if one divides the bias 0.029 with the offset in Figure 12b (approximately 0.5) should get the mean TI used to normalize the data. The resulting value is 0.058, is this correct?
Author Response
Dear reviewer,
Please find attached the response to your 5 comments.
Regards,
Thibault Désert
Author Response File: Author Response.docx
