Low-Altitude Sensing of Urban Atmospheric Turbulence with UAV

Round 1

Reviewer 1 Report

Manuscripts can be accepted.

Reviewer 2 Report

The authors improved the manuscript and satisfactorily included all my suggestions from my first review. I recommend publication.

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Specific comments are attached.

Comments for author File: Comments.pdf

Author Response

The authors are grateful to you for the detailed consideration of the manuscript, interesting questions, and valuable comments. Your comments and recommendations are constructive, which allows us to improve the quality of the presented research and to maximally structure the manuscript according to the requirements of the journal.  

 

Reviewer comments for "Low-altitude sensing of urban atmospheric turbulence with UAV".

This study focuses on the capabilities of a quadcopter in the hover mode for low-altitude sensing of atmospheric turbulence with high spatial resolution in urban areas characterized by complex orography. It provides a new idea for the observation of atmospheric turbulence in regions with complex topography and under severe weather conditions. But in my opinion, there are still some problems to be solved before publishing.

Specific concerns:

 

1. Figure: The axes, labels, and legends of Figures 4-9 are blurry, and clearer Figures need to be provided.

Response: New edition of the article includes figures in the different format.

 

2. What is the innovation of the article relative to other UAV studies on sensing atmospheric turbulence should be highlighted in the introduction section.

Response: When atmospheric turbulence is studied over a territory with complex orography, diagnostic methods providing data with high spatial resolution are preferable.  In contrast to fixed-wing UAV, a quadcopter allows us to obtain atmospheric data with high spatial resolution. The main innovation of the article consists in studying quadcopter capabilities when sensing atmospheric turbulence over a territory with complex orography. The corresponding revision is made in Introduction.

 

3. It is advisable for authors to categorize and integrate the 12 sections of the article so that the structure of the article is clearer. For example, put the modeling methods used in the article into one section and all the experimental results into the order section.

Response: New structure of the article:

1. Introduction
2. Theory of a quadcopter in a turbulent atmosphere
3. Experiment
4. Discussion

Sections 2 and 3 are divided into subsections.

 

4. Sections 3-6: Have the authors made any innovative improvements to the methods or models described in Sections 3-6? If so, please highlight the description, if not, a brief description of the method or model can be provided without listing all the formulas.

Response: These Sections describe the theory of turbulence as applied to the problem of sensing the turbulence spectrum using UAV. The theory of ideal hovering of a quadcopter in a turbulent atmosphere is not only equations, but also the correct choice of a coordinate system and the solution of the equations in this coordinate system. From the viewpoint of the theory of turbulence, the correct approach to the analysis of experimental data is possible in the coordinate system associated with the mean wind. 

Unfortunately, there are a number of papers, in which turbulent fluctuations are studied in the coordinate system used in meteorology. This choice of coordinate system is not considered erroneous. However, in the theoretical sense, it is not a good choice, because in this case it is unclear how to use the Taylor hypothesis, modern turbulence models, etc. In our article, we wanted to most fully present the theoretical results that would allow a wide range of readers to understand the problem of low-altitude sensing of turbulence with copter-type UAVs and the choice of the ways to solve this problem. The corresponding revision is made in Introduction.

 

5. P14-L348-351: The slight differences are observed in the high-frequency range of the spectrum in Figure 9. However, the authors do not explain the reasons for the discrepancy.

Response: These differences are likely caused by at least the following factors:

1. The actual behavior of the quadcopter is described by nonlinear equations. Nonlinearity may change the turbulence spectrum.
2. Quadcopter control includes noise filtering, which can lead to spectrum distortion in the high-frequency range.
3. Different vortices in the atmosphere have different energy. A vortex with the scales  or has maximal energy, while the small-scale part of the spectrum has low energy. Thus, the low-frequency part of the spectrum is measured more accurately, while the high-frequency part may be invisible for UAV.  Our studies show [48] that light-weight and small DJI Mini quadcopter allows obtaining smaller differences in the turbulence spectrum than heavy and big DJI Phantom 4 Pro does.

The issue of differences in the turbulence spectrum is a separate problem and we plan to investigate it in the future.

 

6. The article lacks a detailed discussion and explanation of the findings, and some of the conclusions of the study are skimmed over, such as those reached in Figures 7 and 8.

There is a lack of comparison with studies done by others to support their findings.

Response: Figures 7 and 8 are presented to prove the reliability of the results obtained: the resulting spectra have a form characteristic of a turbulent atmosphere. Publication of this kind of figures is a common practice in research of this type, and they are provided for the article to be clear to a wide range of readers.

The results of comparison of the two turbulence measurement methods based on UAV and acoustic anemometry are added to Subsection 3.3. The results of calculation of the variances and correlation coefficients are presented in Tables 2 and 3. 

The detailed description of the measured spectra in the inertial range is included in Subsection 3.5.  The exponents and frequency ranges for the longitudinal and lateral turbulence spectra are given in Table 5. 

From the viewpoint of remote sensing, we discuss in detail the comparison of turbulence spectra in two cases:  uniform surface and territory with complex orography. Generalizing data in the case of uniform surface are given in the papers [Monin, Kaimal], which are in the list of references.

The main difficulty, we faced in the study, is the lack of general theory and generalizing experimental results concerning atmospheric turbulence over a territory with complex orography and having high spatial resolution. As a result, detailed comparison of our findings with results of other authors is unfeasible.

Reviewer 2 Report

The manuscript (drones-1544926) “Low-altitude sensing of urban atmospheric turbulence with 3 UAV” by A. Shelekhov et al. describes new results demonstrating the capability a UAV in hoovering mode for turbulence sensing in complex terrain and for different seasons. This are a new results compared to their earlier published work where they test the capability a UAV in hoovering mode for turbulence sensing only for flat terrain. It is shown that the results obtained with the DJI Phantom 4 Pro agree well with results from ultrasonic weather stations. In particular, the longitudinal and lateral spectra and the ratio of turbulence scales are in goo agreement.

In summary, I recommend the paper for acceptance. The information provided by the paper is relevant for UAV traffic management systems as well as very useful for researchers that are interested in the turbulence of the lower atmospheric boundary layer.

I have the following suggestions for the authors:

• Consider citing the following publications in the introduction where an overview of different methods to characterize atmospheric turbulence using UAVs is given:

Sucher et al.: "Investigation of optical turbulence from an unmanned aerial system," Proc. SPIE 10787, Environmental Effects on Light Propagation and Adaptive Systems, 1078706 (10 October 2018); doi: 10.1117/12.2325626

Sprung et al.: “Using ultrasonic anemometers for temperature measurements and implications on Cn2," Proc. SPIE 11153, Environmental Effects on Light Propagation and Adaptive Systems II, 111530B (11 October 2019); doi: 10.1117/12.2534023

Maybe the methods described in these publications also suggest a way for further evaluation and cross checking of the technique described in the manuscript in future research.

• Maybe it should be in more detail explained (2 or 3 sentences) in what exact format the UAVs telemetry data are available and how they are processed.
• Figures (in particular Fig.9) are low quality. Some text is barely readable. In Fig. 9 e-h what is probably supposed to be a “=” looks more like an “-“ on my screen.

Author Response

The authors are grateful to you for the detailed consideration of the manuscript, interesting questions, and valuable comments. Your comments and recommendations are constructive, which allows us to improve the quality of the presented research and to maximally structure the manuscript according to the requirements of the journal.  

 

The manuscript (drones-1544926) “Low-altitude sensing of urban atmospheric turbulence with 3 UAV” by A. Shelekhov et al. describes new results demonstrating the capability a UAV in hoovering mode for turbulence sensing in complex terrain and for different seasons. This are a new results compared to their earlier published work where they test the capability a UAV in hoovering mode for turbulence sensing only for flat terrain. It is shown that the results obtained with the DJI Phantom 4 Pro agree well with results from ultrasonic weather stations. In particular, the longitudinal and lateral spectra and the ratio of turbulence scales are in goo agreement.

In summary, I recommend the paper for acceptance. The information provided by the paper is relevant for UAV traffic management systems as well as very useful for researchers that are interested in the turbulence of the lower atmospheric boundary layer.

I have the following suggestions for the authors:

Consider citing the following publications in the introduction where an overview of different methods to characterize atmospheric turbulence using UAVs is given:

Sucher et al.: "Investigation of optical turbulence from an unmanned aerial system," Proc. SPIE 10787, Environmental Effects on Light Propagation and Adaptive Systems, 1078706 (10 October 2018); doi: 10.1117/12.2325626

Sprung et al.: “Using ultrasonic anemometers for temperature measurements and implications on Cn2," Proc. SPIE 11153, Environmental Effects on Light Propagation and Adaptive Systems II, 111530B (11 October 2019); doi: 10.1117/12.2534023

Response: The reference to [Sucher, Sprung] is included in Introduction. 

 

Maybe the methods described in these publications also suggest a way for further evaluation and cross checking of the technique described in the manuscript in future research.

Response:  It is a good idea to consider the relation between temperature fluctuations and turbulent vortices of wind velocity. This is especially important for development of models that predict the state of atmospheric turbulence under urban conditions like Yamada-Melor models. We plan experiments of this kind in the future.

 

Maybe it should be in more detail explained (2 or 3 sentences) in what exact format the UAVs telemetry data are available and how they are processed.

Response: Information about the format and processing of telemetry data is included in the text (see P.9, L286)

 

Figures (in particular Fig.9) are low quality. Some text is barely readable. In Fig. 9 e-h what is probably supposed to be a “=” looks more like an “-“ on my screen.

Response: New edition of the article includes figures in the different format.

Author Response File: Author Response.docx

Reviewer 3 Report

Low-altitude sensing of urban atmospheric turbulence with UAV

The authors provide an interesting comparison of meteorological observations from a quasi-stationary quadcopter and nearby meteorological station measurements. They show comparisons in both the temporal and spectral domains and reveal a certain degree of agreement and provide encouraging results for using quadcopters in complex terrain for this purpose in the future.

 

The manuscript needs improvement in terms of the English. The readability and clarity would be improved with copyediting or some other improvements in the grammar and sentence structures. Some examples are below. My main issue with the manuscript is the lack of quantitative comparisons between the two data sets. The time series comparisons in section 9 are only by eye-ball metrics, which is insufficient. There needs to be rigorous statistical comparisons done there. As well, the spectral comparisons need to be significantly improved upon with any means of spectral coherence analysis or statistical tests. The conclusions would be significantly strengthened with a rigorous quantitative analysis between these time series, on which multiple textbooks have been written.

 

L15: what does “main regularities” mean?

L34-36: Rephrase this sentence. Meaning is unclear.

L41-43: Rephrase this sentence and fix grammar.

L55: “Data on the atmosphere” is an awkward phrasing

L62: UAVs?

L64-69: It is unclear to me why UAV uncertainty should be the same as lidar uncertainty.  I don’t understand the meaning of this comparison.

L73: “bordered a cottage village…”

L227-231: The same location was not used for each season? So, you will need to account for both seasonal and geographical variability?

L286: How are they neglected? Do you remove the times, or do you include the inaccuracies during these times?

L289-293: I believe this was supposed to be informative about the focus of the following section, but it is actually phrased as a command

Author Response

The authors are grateful to you for the detailed consideration of the manuscript, interesting questions, and valuable comments. Your comments and recommendations are constructive, which allows us to improve the quality of the presented research and to maximally structure the manuscript according to the requirements of the journal.  

 

Low-altitude sensing of urban atmospheric turbulence with UAV

The authors provide an interesting comparison of meteorological observations from a quasi-stationary quadcopter and nearby meteorological station measurements. They show comparisons in both the temporal and spectral domains and reveal a certain degree of agreement and provide encouraging results for using quadcopters in complex terrain for this purpose in the future.

The manuscript needs improvement in terms of the English. The readability and clarity would be improved with copyediting or some other improvements in the grammar and sentence structures. Some examples are below.

My main issue with the manuscript is the lack of quantitative comparisons between the two data sets. The time series comparisons in section 9 are only by eye-ball metrics, which is insufficient. There needs to be rigorous statistical comparisons done there. As well, the spectral comparisons need to be significantly improved upon with any means of spectral coherence analysis or statistical tests. The conclusions would be significantly strengthened with a rigorous quantitative analysis between these time series, on which multiple textbooks have been written.

Response: The results of numerical comparison of the two turbulence measurement methods based on UAV and acoustic anemometry are added to Subsection 3.3. The results of calculation of the variances and correlation coefficients are presented in Tables 2 and 3. These results allow the quantitative comparison of the observation series shown in Fig. 6.

The detailed discussion of the behavior of the measured turbulence spectra in the inertial range is included in Subsection 3.5.  The exponents and frequency ranges for the longitudinal and lateral turbulence spectra are given in Table 5.   

The exponents and frequency ranges characterize the turbulence spectrum in the inertial range, while the turbulence scales characterize it in the energy production range. Thus, these characteristics allow the quantitative comparison of turbulence spectra measured by different methods. 

Unfortunately, the more detailed and objective comparison of acoustic anemometry and quadcopter data is impossible within the scope of this manuscript. Discrepancies in observation series may result from many factors, such as the inhomogeneity of the underlying surface, weather conditions, and others.

The experiment, in which the correct comparison of the data sets would be possible, assumes, first of all, the ideal surface for the air mass motion over this surface to correspond to homogeneous and isotropic turbulence.  Weather conditions during the experiment should be also ideal: no clouds, no wind gusts, and so on. One weather station is insufficient for objective evaluation of the weather situation.  The list of conditions, under which this experiment would be possible, can be continued.  Currently, we are planning experiments of this kind.

 

L15: what does “main regularities” mean?

Response: The corresponding sentence is removed.

 

L34-36: Rephrase this sentence. Meaning is unclear.

Response: The sentence is rephrased.

 

L41-43: Rephrase this sentence and fix grammar.

Response: The sentence is rephrased.

 

L55: “Data on the atmosphere” is an awkward phrasing

Response: The sentence is rephrased.

 

L62: UAVs?

Response: Corrected.

 

L64-69: It is unclear to me why UAV uncertainty should be the same as lidar uncertainty.  I don’t understand the meaning of this comparison.

Response: The resolution of a high-resolution lidar is ~20-30 m. Our analysis of the SUMO sensing scheme shows that the spatial resolution of this UAV is ~1000 m.  As a result, the uncertainty in measuring vortices with scales ~10 m is large for both a lidar and a fixed wing UAV.  According to our estimates, the spatial resolution of DJI Phantom 4 Pro is <1 m.  Thus, vortices with scales ~10 m can be resolved by this quadcopter.

Introduction in the manuscript are cardinally revised.

 

L73: “bordered a cottage village…”

Response: Corrected.

 

L227-231: The same location was not used for each season? So, you will need to account for both seasonal and geographical variability?

Response:  We considered different schemes of the experiment and, as a result, settled on the scheme presented in the article. The main criteria for choosing sites for the experiment were hard-to-reach places and significant changes in orography at the Institute territory. In 2020, we conducted the experiment at the territory close to uniform.  The results of that experiment were published in [48].  The main conclusion is that a rotary wing UAV shows great promise as a tool for diagnostics of the atmospheric boundary layer over a territory with various orography. 

This year, we have conducted a series of experiments on remote sensing of the profile of atmospheric turbulence. Turbulence spectra and profiles at different heights were obtained.  Answering your question about seasonal and geographical variability, we can say the following: we completely agree with you that this is a very urgent and interesting topic. It is worth conducting such experiments within the framework of programs aimed at the study of peculiarities of turbulence formation at some or other territory in different seasons.

 

L286: How are they neglected? Do you remove the times, or do you include the inaccuracies during these times?

Response: When calculating the spectrum, we use the complete measured set of pitch, roll and yaw angles.  The periods when the quadcopter starts to move and its exact positioning in space is distorted are not removed.  

 

L289-293: I believe this was supposed to be informative about the focus of the following section, but it is actually phrased as a command

Response: The sentence is rephrased.

Author Response File: Author Response.docx