the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 18 Aug 2025
Emerging mobile lidar technology to study boundary-layer winds influenced by operating turbines
Abstract. The development of a microjoule-class pulsed Doppler lidar and deployment of this compact system on mobile platforms such as aircraft, ships, or trucks has opened a new opportunity to characterize the dynamics of complex mesoscale wind flows. The PickUp-based Mobile Atmospheric Sounder (PUMAS) truck-based lidar system was recently used during the American Wake Experiment (AWAKEN) to assess the general structure of boundary-layer wind and turbulence around wind turbines in central Oklahoma.
Wind speed profiles averaged over PUMAS transects influenced by the operating turbines (waked flow) show a 1–2 m s-1 reduction compared to mean undisturbed (free flow) wind speed profiles. Spatial variability of wind speed was observed in time-height cross sections at different distances from turbines. The wind speeds were about 9–12 m s-1 at 6 km distance compared to 5–7 m s-1 at the transects near the turbines.
The PUMAS dataset from AWAKEN demonstrated the capability of the mobile Doppler lidar system to document spatial variability of wind flows at different distances from wind turbines and obtain quantitative estimates of wind speed reduction in the waked flow. The high-frequency, simultaneous measurements of the horizontal and vertical winds provide a new approach for characterizing dynamic processes critical for wind farm wake analyses.
- Preprint
(5468 KB) - Metadata XML
-
Supplement
(591 KB) - BibTeX
- EndNote
Status: open (extended)
-
CC1: 'Comment on wes-2025-79', Etienne Cheynet, 19 Aug 2025
reply
-
AC1: 'Reply on CC1', Yelena Pichugina, 21 Aug 2025
reply
Dear Etienne Cheynet,
Thank you for your perfect questions!
There are three components of motion consideration that actually use all three words.
- We use a pitch/roll active stabilization platform to maintain the lidar level in real-time. This enables vertical velocity measurement without needing correction for horizontal wind speed in post-processing step as described in Malekmohammadi et all, 2025.
- We optionally shift our detection window to account for platform motion when the truck is moving at high speeds, [SB1]to keep the measured Doppler velocities within our detection bandwidth.
- Platform and lidar orientation and velocities along all axes are tracked and we are putting them into the “world reference frame” for horizontal wind retrieval. I would consider this to technically be a post-processing step, though we do this in our real-time processing and plotting as well when in the field.
We made serious efforts towards quality control of data during AWAKEN using all our experience of mobile-platform measurements (see Table A1 in Appendix A) with the first implementation of the lidar motion corrected system deployment at a NOAA research vessel in the Gulf of Maine (Pichugina et al. 2012, see the manuscript reference).
In addition, during AWAKEN, we provided 5 min measurements in the stationary position at the beginning and the end of each track, which allows us to check the system performance during the post-processing of data.
Regarding your question about terminology, we refer you to lines 112-119 of the manuscript repeated here for your convenience:
“Two significant obstacles to obtaining accurate wind profiles from the high precision lidar measurements using these techniques are correcting platform motions that get projected into the lidar measurements and maintaining accurate control and tracking of the lidar pointing. To address platform velocity, the platform and lidar orientations and velocities are tracked along all axes so that the measured platform-relative line-of-sight velocities can be transformed into the world reference frame. Motion compensation is also performed at high platform velocities by shifting the output laser frequency to keep the Doppler shifted return signal within the detection bandwidth. To further minimize pointing error and enable a vertically pointing beam, the lidar optics are housed in an active motion stabilization frame that keeps the lidar level in the world reference frame to within 1° at all times.
Sincerely,
Yelena Pichugina
Citation: https://doi.org/10.5194/wes-2025-79-AC1
-
AC1: 'Reply on CC1', Yelena Pichugina, 21 Aug 2025
reply
-
RC1: 'Comment on wes-2025-79', Anonymous Referee #1, 03 Sep 2025
reply
Comments on the manuscript entitled "Emerging mobile lidar technology to study boundary-layer winds influenced by operating turbines" submitted to WES
The authors reported a mobile lidar technology (PUMAS) for characterizing wind flows, which has the potential to become a powerful tool for wind assessment in the design and operation of wind energy projects. Measurement results were presented in the paper, demonstrating the capability of the proposed technology. It is an interesting and important work.
The following concerns require proper attention before the manuscript can be accepted for publication.
1. The atmospheric flow evolves as the measurement is taken on a moving truck. There is a characteristic timescale for a specific atmospheric flow event, while the measurement introduced another timescale determined by the speed of the moving truck and the flow event of interest. Choosing a proper path and driving speed seems to be important to ensure the effectiveness of the measurement. What are the authors' thoughts on this issue? What approaches are being taken to cover enough spatiotemporal range in the proposed measurement system? Are there any best practices for measuring typical flow phenomena, e.g., LLJ, wind turbine wakes, and atmospheric flows in complex terrain?
2. Interpretation of PUMAS measurements is not straightforward. The measurements contain distributions of physical quantities in both space and time directions. The authors plot most of their results in the time-height cross sections. I understand that this is partly for comparison with the measurements from the stationary lidar. On the other hand, one advantage of PUMAS is that it provides variations of wind in space (in both vertical and horizontal directions). Can the obtained measurements be employed to show the spatial variation (not in the vertical direction) of a flow phenomenon, say, the downwind variation of wind turbine wakes (the authors showed some results, but the capability does not seem to be well demonstrated)?
3. On lines 657-660, the authors stated the capability of the proposed PUMAS technology in predicting flow statistics of different orders. However, the paper mostly focused on the first-order statistics. It is necessary to examine the proposed PUMAS system in measuring higher-order statistics, like variance, skewness, and energy spectrum.
4. Some discussions are necessary on the uncertainty of the measurements, like how the measurement accuracy depends on the atmospheric conditions, terrains, and the measuring conditions of the PUMAS itself.
5. In the conclusions section, it is suggested to discuss the limitations of the proposed PUMAS technology and potential issues to be addressed.
Citation: https://doi.org/10.5194/wes-2025-79-RC1 -
AC2: 'Reply on RC1', Yelena Pichugina, 11 Sep 2025
reply
Comments on the manuscript entitled "Emerging mobile lidar technology to study boundary-layer winds influenced by operating turbines" submitted to WES
The authors reported a mobile lidar technology (PUMAS) for characterizing wind flows, which has the potential to become a powerful tool for wind assessment in the design and operation of wind energy projects. Measurement results were presented in the paper, demonstrating the capability of the proposed technology. It is an interesting and important work.
The following concerns require proper attention before the manuscript can be accepted for publication.
1. The atmospheric flow evolves as the measurement is taken on a moving truck. There is a characteristic timescale for a specific atmospheric flow event, while the measurement introduced another timescale determined by the speed of the moving truck and the flow event of interest. Choosing a proper path and driving speed seems to be important to ensure the effectiveness of the measurement. What are the authors' thoughts on this issue? What approaches are being taken to cover enough spatiotemporal range in the proposed measurement system? Are there any best practices for measuring typical flow phenomena, e.g., LLJ, wind turbine wakes, and atmospheric flows in complex terrain?A1). The reviewer points out an important interpretation issue, the parsing of along-track variations into spatial and temporal components. We have discussed this in previous papers (Pichugina et a. 2012; Banta et al. 2013b) but did not go into enough detail here, in part because our daytime operations and the relative constancy of the well-mixed boundary layer flow made these issues less of a problem for these analyses. They are, however, important considerations for mobile-platform sampling in general, so we have added a paragraph to the Introduction addressing this point. Specifically for this study, when operating the mobile lidar we used 10–20-minute repeat legs as a guideline to minimize confusion from spatial and temporal evolution. I.e., driving on the King Plains roads eastward and then back westward to the same point took less than 20 minutes so that the atmosphere did not have much time to change temporally, and our measurements were thus more representative of spatial changes rather than temporal. Of course, this assumes some stationarity of the atmosphere, but this is a reasonable assumption for most of the AWAKEN cases given the meteorology and terrain of the area. The longer transects, usually 50-60 min were from the round trips between the hotel, and the AWAKEN study area also consist of two parts: influenced by wind turbines and a free flow (see Fig. 3). That being said, the overarching temporal evolution across hours spent measuring in the wind farms on a given day was much more likely to be subject to large-scale changes that may mask the smaller-scale processes that we are interested in. This is where the other AWAKEN measurements, such as wind profiles upwind of all wind farms from stationary Doppler lidars as well as from the long-term lidar measurements from the nearby ARM SGP sites C1 (15-min data) and E37 (10-min data), are useful for context. Besides, all available NWP models were monitored before the final planning of PUMAS measurement during each day.
In some cases, if any, larger-scale temporal changes may be removed as a time-dependent mean flow, and turbine wakes may be represented as anomalies.
Regarding the “best practices for measuring typical flow phenomena, e.g., LLJ, wind turbine wakes, and atmospheric flows in complex terrain?”. The new text in the Introduction has recommended repeating sampling tracks and mixing in fixed-sensor data into the analyses and interpretation of the data, and these should be incorporated as key aspects of best practices for the use of mobile sensors for reasons described. The PUMAS operated during AWAKEN over a short period of mid-August-mid-September due to the involvement of the instrument into other NOAA projects. Besides measurements were taken mostly on the late morning-daytime hours to obtain the communication and support (if needed) from the engineers in the office located in a different time zone. The measurement conditions were characterized by low (3-4 m/s) to moderate (10-12 m/s) winds and a rare late-morning remnant of the nocturnal LLJ was observed on Aug 5 and analyzed in the paper. As mentioned in the paper, our goal was to test the instrument along the performance of a motion compensating system in the real conditions driving within wind farm, develop a better driving pattern, and prepare for the future wind experiments.2. Interpretation of PUMAS measurements is not straightforward. The measurements contain distributions of physical quantities in both space and time directions. The authors plot most of their results in the time-height cross sections. I understand that this is partly for comparison with the measurements from the stationary lidar. On the other hand, one advantage of PUMAS is that it provides variations of wind in space (in both vertical and horizontal directions). Can the obtained measurements be employed to show the spatial variation (not in the vertical direction) of a flow phenomenon, say, the downwind variation of wind turbine wakes (the authors showed some results, but the capability does not seem to be well demonstrated)?
A2). Yes, as pointed in the paper, these time-height cross sections represent 3d wind variability, vertical, temporal and spatial. In this paper we did not emphasize the spatial variability along each transect due to the short length of the transects, flat terrain, and more generally due to daytime turbulence masking the signature of turbine wakes. We have done this successfully for other projects in the past, and as in the replies to the previous comment, we have added a paragraph to the Introduction to address these issues.
The added paragraph is “Profile measurements from a moving platform document the horizontal variability of the flow, which could (for example) be due to turbine wakes or terrain-related flows, within a curtain of data along the track. But also included is variability due to temporal changes during the transect (Pichugina et al. 2012). For instance, a frontal passage halfway through a sampling leg will appear as a difference between the first and second half of the leg. Lacking additional information, one cannot determine whether these measurements show a genuine, persistent difference in the flow between the two regions. Other small-scale phenomena over the sampling track at timescales smaller than the sampling time interval of the leg may similarly appear to be horizontal variations. One approach for clarification is to retrace the path, as in the offshore LLJ example of Pichugina et al. (2012: see their Fig. 15 and accompanying text), to look for persistence of flow structures, indicating stationarity. Another is to use a mix of mobile and fixed-platform sensors to sort out the spatial and temporal variabilities, as proposed by Banta et al. (2013). In the following we use both approaches”3. On lines 657-660, the authors stated the capability of the proposed PUMAS technology in predicting flow statistics of different orders. However, the paper mostly focused on the first-order statistics. It is necessary to examine the proposed PUMAS system in measuring higher-order statistics, like variance, skewness, and energy spectrum.
A3). The PUMAS lidar measurements can be processed to examine higher orders, similar to a stationary Doppler lidar. We have not examined the obtained higher orders of turbulence such as variance and skewness in this article as we expect or already know that turbine wake effect would just be masked by daytime turbulence. The rich dataset obtained can be used for more analysis and future research papers
4. Some discussions are necessary on the uncertainty of the measurements, like how the measurement accuracy depends on the atmospheric conditions, terrains, and the measuring conditions of the PUMAS itself.
A4). The general information on the uncertainty of the measurements was provided in the paper. As mentioned, not much variations in wind conditions and terrain were observed during PUMAS measurements. But all data including as time-series of pitch, roll, lidar height (ASL), measured and motion corrected vertical velocity are available for a future detailed analysis.
5. In the conclusions section, it is suggested to discuss the limitations of the proposed PUMAS technology and potential issues to be addressed.
A5). Ideally it would be great to obtain long-term measurements over various seasons and atmospheric events. We do not see much limitation for PUMAS measurements except the very nasty road conditions for driving the truck.
-
AC2: 'Reply on RC1', Yelena Pichugina, 11 Sep 2025
reply
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 601 | 51 | 19 | 671 | 26 | 32 | 32 |
- HTML: 601
- PDF: 51
- XML: 19
- Total: 671
- Supplement: 26
- BibTeX: 32
- EndNote: 32
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Dear authors,
Thank you very much for this interesting draft! I have just a short question regarding the use of motion-compensated (or corrected) lidar data. Were the data corrected in post-processing, as in the Lollex experiment [1], where Doppler wind lidars were deployed on a vessel moving in an offshore wind farm. Or was the motion corrected in real time using an active system, similar to the gyroscopic self-levelling tables found on cruise ships? Is it possible that both active compensation and post-processing correction were used simultaneously? I believe this information is provided around lines 112–120, but I am not sure I fully understand it.
I am also uncertain which term would be most appropriate:compensated, corrected, or stabilized. Please feel free to propose the terminology you consider most accurate.
Best regards,
Etienne Cheynet
Reference
[1] Malekmohammadi, S., Cheynet, E., & Reuder, J. (2025). Observation of Kelvin–Helmholtz billows in the marine atmospheric boundary layer by a ship-borne Doppler wind lidar. Scientific Reports, 15(1), 5245.