the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 07 Nov 2024
Turbulence kinetic energy dissipation rate estimated from a WindCube Doppler Lidar and the LQ7 1.3 GHz radar wind profiler in the convective boundary layer
Abstract. From 21 August to 15 September 2022, a WindCube v2 Infrared coherent Doppler Lidar (DL) supplied by EKO Co. (Japan) was deployed at the Shigaraki MU Observatory (Japan) near the LQ7 UHF (1.357 GHz) wind profiler in routine operation. Horizontal and vertical velocity measurements from DL were reliably obtained in the [40–300] m height range with vertical and temporal resolutions of 20 m and 4 seconds, respectively. The LQ7 wind measurements are collected with range and temporal resolutions of 100 m and 59 s, respectively, and 10-min average profiles are calculated after data quality control. Reliable LQ7 Doppler data are collected from the height of 400 m. Despite the lack of overlap in the height range, we compared the Turbulence Kinetic Energy (TKE) dissipation rate ε in the daytime planetary boundary layer estimated by the two instruments. A method based on the calculation of the one-dimensional transverse line spectrum of the vertical velocity W from mean W time series (TS method) was applied to DL ( εDL). The same method was also applied to 1-min LQ7 data (εLQ7TS) to assess its performance with respect to DL despite the poorer time resolution. A more standard method based on the Doppler Spectral width (DS) was also applied to LQ7 ( εLQ7DS) from the 10-min average profiles. We tested recently proposed models of the form ε=σ3/L where σ is half the spectral width corrected for non-turbulent effects and L is assumed to be a constant or a fraction of the depth D of the Convective Boundary Layer (CBL). The main results are: (1) For the deepest CBLs (max(D) >~1.0 km ) that develop under high atmospheric pressure, the time-height cross-sections of εLQ7DS and εDL show very consistent patterns and do not show any substantial gaps in the transition region of 300–400 m when εLQ7DS is evaluated with L~70 m, which is found to be about one tenth of the average of the CBL depth (L~0.1D) . (2) Hourly mean εDL averaged over the [100–300] m height range is on average about twice the hourly mean εLQ7TS averaged over the [400–500] m height range when D >~1.0 km. (3) Hourly mean εDL averaged over the [100–300] m height range and hourly mean εLQ7DS averaged over the [400–500] m height range with L~0.1 D are identical on average. Consistent with the fact that ε is expected to decrease slightly with height in the mixed layer, (2) and (3) imply an uncertainty as to the exact value of the L / D ratio: ~0.1 D < L <~0.2 D. We have also studied in detail the case of a shallow (D <~0.6 km) convective boundary layer that developed under low atmospheric pressure and cloudy conditions. Despite the fact that hourly mean εDL averaged over the [100–300] m height range and hourly mean εLQ7TS averaged over the [400–500] m height range show more significant discrepancies, maybe due to the different properties of the shallow convection, the time-height cross-sections of εDL and εLQ7DS show more consistent patterns and levels.
- Preprint
(2414 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on amt-2024-111', Anonymous Referee #2, 06 Dec 2024
This manuscript builds on previous papers (by Luce and various co-authors) that have developed a novel formulation for extracting turbulence kinetic energy dissipation rates from wind profiling radar observations. The earlier papers looked at shear generated turbulence within the free troposphere as observed by a lower-VHF radar. The present manuscript extends the analysis to turbulence occurring within the Convective Boundary Layer (CBL). It compares observations made by a UHF radar and a Doppler Lidar. Despite the fact that there is no altitude overlap between the two sets of measurements, the manuscript shows that the data are consistent when the CBL is well developed, but differ when the CBL is less well developed.
This is an interesting manuscript. The analysis and discussion are sufficiently thorough that I do not have an substantial questions about the science. There are, however, a few points that would benefit from clarification. Most of my comments relate to the figures, since some aspects were not entirely clear when the manuscript was printed out at A4 size.
POINTS REQUIRING CLARIFICATION
1) On line 53, it is stated that LQ7 observations are made at a time resolution of 59 s and that the data are also processed to provide deliverables at a time resolution of 10 minutes. It would be helpful if the authors could clarify when they are using 59 s data (presumably for the time series analysis) and when they are using 10 minute data.
2) Do references, on line 90, to the mean wind (U bar) and the angle between the beam axis and mean wind direction (α) imply three-dimensional measures or just horizontal ones?
3) What does the word "minimum" imply on line 105 in the sentence, "However, we did not observe such a characteristic, but rather a very clear -5/3 slope when the horizontal wind is minimum."? Does this imply slow wind speeds or actually to the minimum values?
4) In section 2.2, it would be useful to add a paragraph that summarises the spectral width method, e.g. indicating that vertical beam observations corrected for the effects of beam broadening have been used. At present, the reader must refer to the earlier papers for these details.
5) For completeness, the following abbreviations should be given in full where they first appear in the main manuscript. I note that two of them are given in the abstract.
5a) "CBL" on line 46; it is given on line 47 (and on line 21 in the abstract)
5b) "UHF" on line 48
5c) "TS" on line 61 (given on line 17 in the abstract)
5d) "VHF" on line 62
5e) "rms" on line 94FIGURES
6) The axis labels in Figure 2 are difficult to read since they are so small. Using a non bold font might help. Otherwise it would be useful to mark the location of the observation site with a marker.
7) The tick labels on the time axis of Figures 4, 7, and 9 are quite far apart (5 hour intervals). It would be helpful for the reader if these could be shown at shorter intervals (c.f. Figure 5) since the manuscript refers to features that occur at specific times. It would also help if the tick marks could be shown pointing outwards from the plot area.
8) For the right hand side of the bottom panel of Figure 4, the wind direction arrows overlap each other making it difficult to see if they represent easterly or westerly winds (I note that this is clarified in the main text).
9) For clarity, it would be helpful to state the approximate backscatter values corresponding to the "red" and "brown" colours in Figure 4, which are referred to on lines 178 and 179. I think what the authors refer to as "brown" is what I would describe as dark red.
10) It can be difficult to distinguish between the blue and black lines on the upper two panels of Figure 4 and in Figure 6a. Using higher contrast colours could help. It is also difficult to distinguish between the grey and blue dots in Figure 10. Using larger dots might help.
11) The black dashed line indicating the proxy for the CBL top height in Figure 5 is a little difficult to see. Could the line be made thicker?
12) The thin red lines in right hand plots of Figure 9 showing the availability of LQ7 data (and to some extent the thin blue lines showing the availability of DL data) are initially difficult to see. Creating separate plots for these lines could help.
13) It would be useful to consistently use the same colours to represent DL and LQ7 data in Figures 5, 6, 7, 9 (right hand plots), and 13. Moreover, it would be helpful to use these colours for the vertical lines in Figure 6 to distinguish between the scale sizes used for the different instruments.
TECHNICAL CORRECTIONS
14) The name O'Connor, in the reference to O'Connor et al. (2010) should be shown with the letter "c" in upper case rather than in lower case.
15) The DOI https://doi.org/10.5194/amt-2023-38 for Luce et al. (2023a) shown in the manuscript refers to the discussion version of the paper. The DOI https://doi.org/10.5194/amt-16-3561-2023, which refers to the accepted version, should be used instead.
Citation: https://doi.org/10.5194/amt-2024-111-RC1 -
RC2: 'Comment on amt-2024-111', Anonymous Referee #1, 06 Dec 2024
General comments and recommendation for disposition of the paper
This work presents an analysis on the observations made jointly by a Doppler lidar (DL) and a
UHF band wind profiler radar (LQ7) in convective boundary layers (CBL) in clear air. The database
was built during a measurement campaign lasting about one month. The research mainly uses the
vertical profiles of the time series of the vertical air velocity W measured continuously by the two
instruments. The information on the height Zi of the atmospheric boundary layer is deduced from
the vertical reflectivity profiles of the LQ7 (local maximum of reflectivity). This important
information is used in the discussions. As the title indicates, the authors focused the results on ε
which represents the dissipation rate of the turbulent kinetic energy (TKE). It is a fundamental
parameter in the description of the boundary layer. Given the role that W plays in the restitution of ε
and in the demonstrations related to the inter-instrument comparison, it would be appropriate to
include the vertical velocity in the title.
The difficulty of the comparison comes from the fact that the measurements of the two instruments
do not overlap in altitude. The lidar produces observations below 300 m while the radar begins its
usable observations from 400 m. The analysis then calls upon considerations of vertical continuity
of W and ε and the known vertical evolution of ε in the CBLs. A true quantitative comparison
remains to be made but from a qualitative point of view the results obtained are very convincing.
The dissipation rate of the turbulent kinetic energy ε is obtained from W in two different ways.
The first method (TS method) deduces ε from the spectral analysis of the time series of W under the
assumption of Kolmogorov-Obukov inertial turbulence. The second method (DS method) which is
the most commonly used in the literature uses the spectral variance σt2 of the Doppler spectral peaks
caused by turbulence. This information is only known and recorded by the radar. As expected due to
the low temporal resolution of the radar (more than 10 times lower than DL), the TS method applied
to radar data turns out to be unreliable in comparison with DL retrieval.
In fact the comparison of the ε obtained by DL with the TS method and by LQ7 with the DS
method is the main objective of the work. The various results of this comparison show a good
similarity between these two types of measurement. These results are optimized by using for LQ7
the law ε = σt3 ⁄ (α D). Where D is the thickness of the CBL and α is a parameter, comprised
between 0.1 and 0.2, which depends on the physics of turbulence. This law, the formulation of
which is found in the literature, remains empirical and needs further observations to be validated
and in particular to quantify its uncertainty range.
The measurement of the lidar and radar has a spatial resolution that depends on the resolution
volume as a function of the beam aperture and the pulse length. This induces a spatial filtering
whose effect is mentioned but which is not taken into account here based on a rather weak
justification. This filtering is taken into account in retrieval equations that can be found in the
literature. The resolution volume is a parameter that differs greatly between the two instruments. It
would therefore be important to evaluate the uncertainty brought by ignoring this filtering.
This well-structured document written in a clear and concise style is easy to read and understand.
The demonstrations are supported by well-chosen graphs. The bibliography, which seems well up to
date on the subject, shows that the authors are recognized specialists in the field covered. This work
certainly convinces the reader of the interest of using a radar or a lidar for the measurement of the
rate of dissipation of turbulent kinetic energy ε in the monitoring of CBLs in connection for
example with atmospheric pollution problems.
In conclusion this paper which is well conducted and which presents a good scientific interest is
appropriate for the publication in the journal Atmospheric Measurement Technique. However this
recommendation is subordinated by the consideration by the authors of the previous major and
following minor comments.
Main review points
1) The technical description of the instruments in the document is brief, it is supplemented by
bibliographical references. In order not to tire the reader too much in often tedious research, I
suggest giving directly the essential operating parameters of the instruments: length and coding of
the pulses, number, direction, and aperture of the beams used.... This would allow to quickly have
an idea on the spatial resolution and to understand why the LQ7 has a first gate starting at 300 m
while the given vertical resolution is 100 m.
2) The DS method uses the variance of the spectral Doppler peaks corrected for effects not related
to atmospheric turbulence. It is important to list all of these extra-turbulent causes and to explain
how the correction is made.
3) Figure 1 shows that in addition to the vertical beam, the lidar has four oblique beams and the
radar five. The authors have used only the vertical beam for the direct measurement of W. We can
also use only the oblique beams to deduce W. When the components of the horizontal wind speed
are far from the zero spectral line, the measurement of W will be less sensitive to the presence of
ground echoes which are very penalizing at low levels (mainly for the radar). I would like to know
if the authors have used this possibility and what advantages and disadvantages they found in it.
4) The study offers a database of vertical air speed observed by two different types of instruments of
rare quality but which I find insufficiently exploited. This concerns in particular the problem of the
negative bias of the average vertical speed in the CBL observed by the UHF band wind profiler
radar since their beginning in the seventies. This negative bias can reach several -0.1 ms-1.
Experimental and theoretical explanations have been put forward but to my knowledge without
success and the problem remains. On the contrary in your work you deduce that this bias is almost
zero for the UHF radar and also for the lidar. It is frustrating for an informed reader that you do not
point out this contradiction with the numerous past observations. This paper would therefore benefit
if you developed this particular point even if it does not fit into the main objective of the paper.
5) The localization of the height of the boundary layer Zi by a local maximum of the radar
reflectivity is a method which if well conducted gives precise results. It is therefore not necessary to
associate the word proxy with this radar Zi which in fact tends to discredit this measurement.
6) The observations that are discussed are located in the convective boundary layer. This is a
dynamical and thermodynamic framework that is inefficient for the development and maintenance
of gravity waves. It is therefore unwise to dwell on it too much.
7) For figure 12 use log10 (ε DSLQ7 / εDL) instead of log10 (εDSLQ7)- log10 (εDL) in the caption and instead
of Δ log10 (ε) on the horizontal axis.Citation: https://doi.org/10.5194/amt-2024-111-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 106 | 21 | 9 | 136 | 3 | 4 |
- HTML: 106
- PDF: 21
- XML: 9
- Total: 136
- BibTeX: 3
- EndNote: 4
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1