1. Introduction
The atmospheric boundary layer (ABL) constitutes the environment of the human activity. In addition to that, it plays a fundamental role in the energy exchanges between the earth surface and the free atmosphere. That is why so much attention has been turned to it for numerous decades. Boundary layer meteorology is a specific research field, the experimental and theoretical approaches of which can be found in Stull's (1988) and in Garratt's (1992) textbooks.
An important contribution to the understanding of the finestructure of the ABL was brought by the acoustic soundings with Doppler sodars that appeared 30 years ago (Little 1969; Eymard and Weill 1982; Lenschow 1986). This easy to use and inexpensive sensing device was in fact the first well-suited ABL wind profiler. Unfortunately, this instrument makes noise and cannot be used in urban environments. On the other hand, its vertical coverage is limited to less than 1 km by the strong sound atmospheric attenuation and by the wind that tends to blow away the acoustic wave. Sodars are often able to depict the entire ABL depth development but only in early morning hours.
UHF-band radar wind profilers, the operational development of which began in the last decade, do not have such limitations (Ecklund et al. 1988; Carter et al. 1995). They can supply, in clear air and rainy conditions, the same information as can sodars on the ABL dynamics and its finestructure (through the analysis of reflectivity), with a vertical coverage usually exceeding the ABL top Zi (Angevine et al. 1994; Hashiguchi et al. 1995; Grimsdell and Angevine 1998). An ultimate advantage over sodar is that, when equipped with a radio acoustic sounding system (RASS), they can also provide virtual temperature profiles (May and Wilczak 1993; Angevine et al. 1998). They tend to supplant sodars in ABL surveys, except in the surface layer (the first 100 m of the atmosphere), where UHF soundings still suffer from strong ground clutter (May and Strauch 1998).
Finally, vertical profiles in the ABL of pollutant concentration and gases, such as water vapor, ozone, and wind measurements, can be inferred from optical soundings made by Doppler lidars, the utilization of which is increasing in ABL research and monitoring (Kunkel et al. 1980; Schols and Eloranta 1992; Cohn et al. 1998; Grund et al. 2001).
Acoustic, electromagnetic, and optical remote sensing techniques, along with numerical modeling such as large eddy simulation (Stevens and Lenschow 2001), offer an ensemble of tools that facilitate understanding of the finestructure of the ABL circulation, traditionally investigated through a statistical approach. Toward this goal, the Turbulence Radar Aircraft Cells (TRAC) experiment, which took place in France during three weeks in June–July 1998, was intended to study the coherent structures, such as horizontal roll vortices and thermal cells that modulate at large scale the turbulent vertical transfers within the ABL (Campistron et al. 1999). As the key instrument of this experiment, a C-band Doppler radar was used to provide, through the reflectivity fields, the three-dimensional (3D) description of the clear ABL organization and its temporal evolution.
The present work is concerned with the clear air ABL wind profiling aspect of the C-band meteorological Doppler radar. A comparison with UHF wind profilers used during the TRAC experiment is presented. Although unwieldy, meteorological Doppler radars with scanning capability have the advantage over the conventional UHF wind profilers used here, to observe the atmosphere in three dimensions. Consequently, they enable, in particular, deduction of the ABL structure and the 2D topography of the ABL top, whereas a UHF profiler can only provide the temporal evolution above itself. However, this latter assertion has to be attenuated with the recent development of volume-imaging wind profilers with static steering ability (Mead et al. 1998). The last part of the paper is focused on the vertical velocity measurements in the context of the downward bias problem usually observed by UHF or VHF wind profilers.
2. Instrument characteristics and data collection
a. TRAC experiment
The data presented here were collected during the TRAC-98 experiment, which is the second field campaign of the TRAC program launched in 1990 and primarily designed to study the coherent structures, such as horizontal roll vortices and thermal cells, that modulate at a large scale (∼5 km) the turbulent vertical transfers within the ABL (Lohou et al. 1998a,b). This experiment took place during three weeks in June and July 1998 over the particularly flat and homogeneous region of the Beauce plain in central France. The scientific objective of the campaign as well as the experimental facilities, including remote sensing devices (a C-band Doppler radar, two UHF wind profilers, and two sodars), surface flux stations, instrumented masts, and airborne measurements, are described in Campistron et al. (1999).
b. C-band Doppler radar
The C-band Doppler radar RONSARD (Recherche sur les Orages et les Nuages par un Système Associé de Radar Doppler), used in the TRAC experiment, has the following main characteristics: 5.4-cm wavelength, 200-m pulse length, 1464-Hz pulse repetition frequency, 250-kW peak transmitter power, and 0.9° antenna beamwidth. It was operated every day, usually from 0600 to 1600 UTC, in a continuous scanning mode based on a repetitive full volume exploration lasting about 10 min. An exploration was composed of 35 panoramic conical scannings (PPI) at different elevation angles (from 0.5° up to the zenith), with a 1° azimuth resolution, and a 100-m radial range resolution. At each range gate, the three first moments of the Doppler spectra are calculated with an FFT algorithm bearing on 64 data points.
Actually, this moderately sensitive Doppler radar was constructed in the early 1970s for the purpose of wind field analysis in precipitating environments. Its application to clear boundary layer observation, where meteorological echoes are usually lower than 0 dBZ, required an improvement in online data processing in order to increase its sensitivity and to efficiently remove ground clutter contamination. Prior to the spectral processing, each 64-point time series is weighted with a Hanning window, and a band-rejected filter is applied to eliminate the dc spectral line and to strongly attenuate the two adjacent lines. Once the Doppler spectra are obtained, the segment method (Petitdidier et al. 1997) is used to find the spectral noise level. The first three moments of the Doppler spectra are calculated with the stronger spectral peak formed with adjacent lines exceeding the noise level. During offline editing, remaining spurious echoes are eliminated through thresholding and continuity tests. The updated RONSARD was operated for the first time during the 1993 campaign of the TRAC project. During this experiment it was proved that this radar was able to provide, in daytime clear air ABL, coherent and continuous velocity fields from 0.1 km up to 3 km above ground, and over a horizontal range reaching at least 30 km in the ABL, depending on the atmospheric conditions (Lohou et al. 1998b).
c. 5- and 3-beam UHF wind profilers
The 5-beam UHF wind profiler (called here UHF5), of which measurements will be mainly compared to the C-band observations, has the following characteristics: 1238-MHz transmitted frequency, 4-kW peak power, 20-kHz pulse repetition frequency, and 150-m pulse width. In order to get the three components of the wind, the profiler uses alternatively five beams, one vertical and four oblique, with a one-way half-power aperture of 8.5°. The oblique beams, with an off-zenith angle of 17°, are disposed every 90° in azimuth. Vertical profiles of the radial velocity, reflectivity, and spectral width are obtained from 75 m up to a height of about 3 km, with a 75-m vertical resolution and a 5-min temporal resolution. The data are processed with a 30-min consensus. The profiler was located at a distance of about 15 km from the C-band radar site. More detailed technical description can be found in Campistron et al. (1997). The other UHF wind profiler (called UHF3) utilized during the campaign was a 3-beam profiler, 37 km away from the first.
Data quality control and processing are realized through a consensus algorithm based on time and height continuity of the edited spectra. The consensus works over a 30-min period and provides, every 5 min, a weighted temporal average of the first three moments associated with the resolved meteorological peak. The data presented here have been processed with a software version that only includes the measurements made on three beams (the vertical and two oblique ones). The zenith-pointing beam radial velocity provides the air vertical velocity. The horizontal wind components are inferred from the measurements of the oblique and vertical beams under the assumption of horizontal wind local homogeneity.
3. Profiling with the C-band Doppler radar
a. Method of analysis
For each C-band radar sequence, parameters of the mean flow were derived using the velocity volume processing (VVP) technique (Waldteufel and Corbin 1979) based on the hypothesis of local linearity of the mean wind field, as for the more traditional velocity azimuth display (VAD) method (Browning and Wexler 1968). This assumption is supposed to be verified independently within staggered horizontal cylindrical slices equidistant in height, centered on the vertical of the radar (see Fig. 1). For the scanning geometry of the radar data analyzed in this paper, a 100-m vertical depth and a 30-km horizontal diameter have been chosen for the dimensions of these elementary processing volumes. Usually, in each slice, several thousands of data points are involved in the calculation of the wind field parameters.
These parameters are representative of the mean flow at the scale equal to the horizontal slice diameter, that is, 30 km in our case, or greater. Note that horizontal divergence is deduced and that vertical vorticity cannot be measured with such a technique. The vertical velocity w0 is also directly provided by the VVP method but, because this component has a weak impact on the radial velocity mainly obtained at low elevation angles, experience showed that this quantity was poorly retrieved. With the anelastic continuity equation, the vertical velocity can be computed through the vertical integration of the horizontal divergence with realistic assumptions on the vertical motion at the upper or/and lower boundaries of integration. Finally, an equivalent technique is used to obtain vertical profiles of the reflectivity at the processing slice scale.
b. A case study of clear air boundary layer C-band profiling
The observations collected on 20 June 1998 are the main source of the detailed analysis presented in the remaining of the paper. This case was selected because it is characteristic of the capability of the C-band radar in the clear air sensing of the daytime summer ABL. This day is synoptically associated with a high pressure system (1005 hPa on the experiment area) that started to form over western Europe three days earlier. It is the warmest day of that period with the temperature at the ground reaching 32°C in the afternoon. It is also marked by the absence of upper and lower-level clouds. Time–height sections of the reflectivity factor and wind velocity measured between 0600 and 1600 UTC by the C-band and UHF radar are displayed in Fig. 2.
The circulation in the first 3 km of the atmosphere observed by both radars (Figs. 2a,b) appears relatively complex with regard to the simple surface pressure field associated to the anticyclonic situation. On average, the wind is southeasterly at low levels and turns southwesterly with height, indicating, in reference to the thermal wind relationship, a warm air advection. The main particularity is a low-level jet reaching 13 m s−1 in the early morning at a 500-m height. With time and turbulence development, this wind maximum zone is eroded in the bottom and only a thin branch of about 10 m s−1 rising above the ABL subsists in the afternoon. The second particular feature is a 500-m-depth layer of minimum wind intensity centered at about 1.6-km height. The wind becomes null at about 1000 UTC, whereas the wind magnitude above and below reaches 10 m s−1.
At a first glance there is a very close resemblance between these two wind fields. However, the UHF wind profiler offers many more details on the wind field finestructure than the C-band radar, for which the VVP technique smoothed out atmospheric scales smaller than the processing slice dimension. The main discrepancy between the two instruments is found below 1 km, where C-band winds have the largest magnitude by about 2 m s−1 in the worst case. The C-band wind measurements are confirmed by sodar observations and rawinsoundings (Jakoby-Koaly et al. 2001, hereafter JK). From these comparisons, we concluded that the meteorological radar with the VVP technique covers the region from 0.1 up to 2 or 3 km, depending on the atmospheric conditions, and works better than the UHF wind profiler at low levels, where ground clutter contamination is the main source of wind measurement errors.
In Figs. 2c and 2d, the reflectivity fields observed by the two instruments present an evident similarity. During that day the strongest echoes are mainly concentrated below 1.5-km height. Within this layer the main feature observed by both radars is a bright band of enhanced reflectivity sloping upward with time from around 0.5 km at 1000 UTC up to 1.2 km at 1600 UTC. According to the radiosoundings launched from the UHF site, this band is positioned in the inversion layer capping the ABL. As a consequence, a C-band radar is able to depict the daytime vertical development of the boundary layer like a UHF wind profiler (Angevine et al. 1994). In the early morning until about 0900 UTC a series of thin layered echoes staggered vertically is clearly apparent in the UHF reflectivity section (Fig. 2d). Except for the one centered around 2.4-km height, these horizontal structures are not very marked in the C-band reflectivity field (Fig. 2c). In the next section a three-dimensional analysis of these layered echoes is presented.
During the experiment, the UHF radar was calibrated during light rain episodes in comparison with reflectivity factors measured by a disdrometer at the ground [this procedure is presented in Campistron et al. (1997)]. The same technique was also used for the C-band radar. The dispersion in the calibrating factor was found to be less than 3 dBZ for the UHF radar but as large as 6 dBZ for the C-band radar. On average, the reflectivity values in Figs. 2c and 2d, presented here in dBZ units, are stronger for the UHF radar. The maximum reflectivity ratio between UHF and C-band radar reaches 15 dBZ in the ABL. This indicates a different backscattering source for each radar. The origin of clear air radar echoes in the lower atmosphere was in the past a matter of a long debate (see, e.g., Battan 1973; Gossard and Strauch 1983; Gage et al. 1999, Muschinski and Lenschow 2001). It seems now well established, in particular from multiwavelength comparisons such as those made by Rogers et al. (1992) or Wilson et al. (1994), that the origin of the clear air echoes is dependent on the wavelength λ of the radar used. Scattering by particles, such as insects, dominates a meteorological radar with wavelength ranging from the millimetric to decimetric band (Rayleigh scattering with a λ−4 dependence) in the ABL, whereas for UHF and VHF radar, scattering from air refractive index irregularities is the principal origin of the returns (Bragg scattering with a λ−1/3 dependence). During the TRAC-93 experiment a comparison was made between the air refractive index structure constant
Considering pure Bragg scattering, the reflectivity factor ratio between the UHF band and the C band, given by the expression 10 log[(λUHF/λC)11/3] according to Wilson et al. (1994), should be 24 dBZ. On the other hand, for pure Rayleigh scattering this ratio should be 0 dBZ. The 15-dBZ maximum deviation observed here indicates a mixture of both types of echoes. However, we cannot rule out Rayleigh scattering in the echos intensity observed by the UHF radar, in particular at a 1.7-km height (Figs. 2c,d), where the reflectivity values of the two radars are very close.
c. Comparison between C-band and UHF radar wind velocity over several days of common observations
The smoothed appearance in Fig. 2a of the C-band wind velocity measurements gives a first insight about the quality of the wind retrieval by the meteorological radar. These C-band wind measurements were compared to the UHF wind profiler data during 11 days of common observations of the daytime lower atmosphere. The results are presented as histograms of the wind speed (Fig. 3a) and direction difference (Fig. 3b), corresponding to the wind measurements between 200 and 2500 m. The standard deviation for the difference in speed is 1.4 m s−1; it is 12.7° for the direction. Moreover, a value of −0.1 m s−1 for the mean wind speed difference tends to indicate a weak bias between the two instruments. Nearly the same results were obtained by JK with the same dataset when comparing the UHF radar and the collocated Doppler sodar. This agreement may be seen as very good when considering that errors of both instruments are cumulated, that both radar are not collocated, and that, because of the different wind-retrieval methods, they do not resolve the same atmospheric scales. We may conclude that the different echo sources do not have a noticeable impact on the horizontal wind retrieval.
4. Three-dimensional aspect of the C-band meteorological radar observations
The steering capability of the meteorological radar is used in this section to analyze the spatial distribution of two particular features of the reflectivity field mentioned in the previous section.
a. Layered echo structures
In section 3c it was shown that the C-band radar turns out to be capable of providing reliable wind profiles, although it is more unwieldy than the usual UHF wind profilers. But it has the advantage over the more conventional wind profilers in that it can scan an air volume in three dimensions. This steering capability was used to study the 3D structure of the ABL.
As an example, Fig. 4 displays two panoramic conical scannings (PPI) made at a low elevation angle, at 0900 (Fig. 4a) and 1400 UTC (Fig. 4b) on 20 June 1998. In both cases, concentric rings can clearly be distinguished. They are constituted with a high density of strong reflectivity echoes and correspond to the intersection of the radar beam with layered echo structure. Seven distinct layers between 300- and 3000-m height can thus be observed, labeled “a”–“g.” Five of them are present at 0900 and only four at 1400 UTC. At 0900 UTC, layer a traces the ABL top, which only rises about 300 m. At 1400 UTC, layers a, b, and c have disappeared because of the turbulent mixing in the boundary layer, and the ABL top is now materialized by the deepest layer, d, which is present in both Figs. 4a and 4b and reaches about 1000 m at 1400 UTC. The depth of the different layers is about 50–100 m, but reaches 200 m at certain regions of layer d in the early afternoon. The spatial structure of these layers seems quite homogenous in the morning. At 1400 UTC, layer d shows roughly more echoes in its southern part, as does layer g in its western part. However, this spatial variability has to be interpreted with caution since it might be due, in a large part, to a difference in altitude related to a problem of an inaccuracy in the elevation angle measurement of about 0.2°.
The presence of different layers within the atmosphere is not a new finding, since some stable layers often associated with gravity waves had already been revealed, 30 years ago. The observations of a frequency-modulated continuous-wave radar sounder (Richter 1969), those of a powerful pulse radar (Browning 1972), or those of an acoustic sounding system (McAllister 1968) give examples of wave structure under statically stable tropospheric conditions. Gossard et al. (1985) also observed such thin stable layers with radar, and the effects of thin layers on clear air backscatter are discussed in Muschinski and Lenschow (2001) and Muschinski and Wode (1998).
Figure 5 shows 1-h-averaged vertical profiles, centered at 0700 UTC 20 June 1998, of the reflectivity factor measured by the C-band radar and the UHF profiler (Fig. 5a), and the vertical profiles of temperature and relative humidity given by the radiosounding at 0658 UTC, launched from the site of the UHF profiler (Fig. 5b). Because of the receiver saturation, the measurements of the UHF wind profiler are not plotted below 200 m. Figure 5a shows that the different layers are observed by both instruments, even if they are more easily distinguished in the UHF profile. The more smoothed appearance of the C-band profile can be due in part to its coarser vertical resolution (100 m instead of 75 m for the UHF radar), along with the spatial average performed in each vertical processing slice. According to the radiosounding, layers a and d are associated with temperature inversions. Layer b stretches at the base of an isothermal layer, as do layers c and e, much less apparently, with a change in relative humidity. However, layers f and g, observed at 1400 UTC, cannot be explained by the sole radiosoundings. Moreover, the strong variations of stability observed with the radiosounding at about 1500 m are not associated with any high-reflectivity ring. This shows that radiosoundings are not sufficient to explain the high-reflectivity layers observed by the radar and that there is no determined rule to link atmospheric stability and radar echoes.
The panoramic conical scannings acquired at the launching time of the radiosounding presented nearly the same reflectivity rings, but these were less distinguishable than in the 0900 PPI presented for clarity.
b. Topography of the atmospheric boundary layer top
Figure 6 gives an example of a vertical plane of reflectivity deduced with a Cressman's interpolation from the C-band observation sequence centered at 1400 UTC 20 June 1998. Note that the original image was bell-shaped in the proximity of the radar. This deformation was corrected, taking into account a slight error in the measurement of the echoes distance. The inversion layer capping the ABL appears clearly around 1-km height as a reflectivity bright band. From the horizontal distribution of the height of this layer of enhanced echo intensity, it is thus possible to draw the topography of the ABL lid. This information on the spatial variability of the ABL top gives the C-band radar an advantage over standard UHF wind profiler that is limited to a small number of fixed beam directions. It is particularly useful when ABL observations are collected over complex terrain.
5. Measurement of vertical motion with UHF profiler and C-band radar
Air vertical velocity is a key parameter in the understanding of the vertical transfer and fluxes within the ABL. It remains the most difficult component of the ABL wind to retrieve because its intensity is usually much smaller than the horizontal wind components, and thus it is much more sensitive to ground clutter contamination in its measurement. Presently, there is serious doubt concerning the accuracy of air vertical velocity measurement made not only with UHF profilers but also with VHF profilers, since there is experimental evidence that these instruments give, on average, a systematic downward bias. Angevine's work (1997), based on 47 days of clear air convective ABL observation, was one of the first to report this bias with UHF measurements. In his daily average, the vertical velocity has a diurnal cycle with persistent downdraft observed during daytime, reaching −0.3 m s−1 around 1400 UTC when the convective ABL is fully turbulent, and with nearly null values at night. At VHF band the vertical velocity error in the free troposphere is not so dramatic and amounts to a few centimeters per second, attributed by Nastrom and VanZandt (1996) to gravity waves, developed in a stable atmosphere, which tend to induce stronger reflectivity values for downward motions.
At UHF band there is not yet a firm explanation of the origin of the erroneous vertical velocity measurement. Angevine (1997) proposes, as a likely source of error, the intrinsic fall speed of particulate scatterers as partly responsible of the echoes. More recently, Muschinski [1998, see his Eq. (6.45), p. 60] has derived a theoretical formula on the effect of the nonzero correlation between wind velocity and refractive-index fluctuations on the first moment of the Doppler spectrum, which could explain radial velocity measurement errors. His result has been confirmed recently in Tatarskii and Muschinski (2001). In this section we present, with C-band measurements, some material that might be useful in the understanding of this open question.
a. UHF vertical velocity
Figure 7 presents an ensemble-averaged diurnal cycle obtained in the period 18–25 June 1998 of three types of mean vertical velocity UHF measurements taken between the heights 0.2 and 1 km during 11 fine weather days. The time series labeled UHF5 comes from the measurements taken by the vertical beam of the UHF profiler whose data are used in the previous section. The time series labeled UHF3 was acquired by the vertical beam of the 3-beam UHF profiler, 37 km away from the UHF5 radar. The last vertical velocity estimate Wop was deduced from the combination of the radial velocities measured by the four oblique beams of the UHF5 radar. The vertical velocity measurement made by the UHF5 profiler was corrected for a slight beam pointing offset of 0.1° of the vertical beam.
This figure clearly shows that the diurnal variation offered by these three different vertical velocity estimates are nearly the same. These time series reproduce and confirm Angevine's (1997) results, with persistent mean subsiding motion during the daytime and the maximum downward velocity of about −0.3 m s−1 in the middle of the day. The presence of two peaks of upward velocity at sunrise and of downward velocity at sunset is noteworthy, since it was a frequent feature observed during the experiment. The agreement between the vertical velocity derived directly from the vertical beam and with the oblique beams allows us to exclude an effect of ground clutter contamination in the explanation of vertical measurement bias, since oblique radial velocity are usually far from the dc line. We may conclude also that if there is a vertical velocity measurement error, this error also affects radial velocity acquired at oblique incidence and so the horizontal components of the wind derived with all the beams measurements.
Further details about the mean vertical evolution and the statistics of the negative bias observed by the vertical beam of UHF5 radar are presented in Figs. 8a and 8b, respectively. The dataset corresponds to 11 selected fine weather days of the period 18 June–3 July 1998, and measurements were made between 1000 and 1400 UTC. For the histogram of Fig. 8b the data are restricted to the boundary layer depth Zi deduced from radiosoundings and UHF reflectivity time series. Negative vertical velocities are predominant in the mean vertical profile of Fig. 8a normalized in height by Zi. The most intense downdraft of about −0.2 m s−1 is found at 0.4 Zi. Weak positive vertical motions are only observed below 0.1 Zi and above 1.3 Zi. The statistical distribution of the vertical velocity presented in Fig. 8b gives a mean value of −0.2 m s−1, a standard deviation of 0.3 m s−1, and a maximum probability at −0.1 m s−1. In this histogram strong negative values appear more probable than strong positive values. These results are in contradiction with the general accepted vertical velocity distribution in convective ABL, in which vertical velocity has negligible value on average, and intense updrafts have higher probability than intense downdrafts (Stull 1988, 461–467).
b. C-band radar vertical velocity
The VVP technique provides the horizontal divergence of the flow representative of atmospheric scales larger than the processing slice diameter (30 km). In Fig. 9, we confront the vertical profiles of the horizontal divergence given by the VVP analysis from the C-band measurements to the vertical velocity measured by the UHF5 wind profiler. These mean profiles are an average of the data acquired between 0630 and 0730 UTC 20 June 1998, a time period already discussed in the previous section. This figure shows clearly that updrafts are associated with convergence (negative values) and downdraft with divergence (positive values). In this particular time period, where the averaged vertical velocities from Fig. 7 are slightly positive, the measurements of both instruments thus appear consistent. Vertical velocities deduced from the vertical integration of the VVP horizontal divergence (not presented) have values with, on average, the same sign as the UHF vertical velocity, but with an order of magnitude reduction, compatible with the resolved atmospheric scale of the VVP analysis (>30 km).
To conclude the discussion on the vertical velocity measurement, a mean view of the vertical motion within convective ABL offered by the C-band radar is presented. A composite day, formed using the previously selected 11 days, of the mean divergence and vertical velocity measured by the C-band radar is displayed in Figs. 10a and 10b, respectively. These mean time series, taken during the daytime development of the ABL between 0600 and 1500 UTC, represent an average of these parameters between 0 and Zi. Figure 10a shows convergence prevailing during the daytime ABL with a maximum magnitude of −1.5 × 10−4 s−1 in the early morning around 0600 UTC. The convergence decreases quasi linearly in the morning until 1030 UTC, corresponding to the time when the maximum ABL top height is nearly reached. After that hour the convergence evolution is approximately stationary with weak negative values greater than −0.5 × 10−4 s−1. Correspondingly, the vertical velocity deduced from the upward integration of the horizontal divergence presents, in Fig. 10b, positive values giving a mean vertical ascent of about 3 cm s−1 for the entire analyzed period. The maximum upward motion reaches an intensity of about 5 cm s−1, which corresponds to half of the vertical rise of the ABL depth development observed in the morning. Consequently, even when relatively weak, these upward movements have to be considered for the understanding of the vertical transfer affecting the ABL. The presence of weak upward motion in the convective ABL seems more physically justified than the strong downward motion observed by UHF profiler. However, more experimental work must be done in order to confirm the present C-band results.
6. Conclusions
The past decade has seen the rapid operational development of UHF radar for the wind profiling and vertical depth survey of the atmospheric boundary layer. During the TRAC experiment, held in June–July 1998, two UHF wind profilers and a C-band meteorological Doppler radar with scanning capability were operated. We took this opportunity to investigate the potentiality of lower-atmosphere clear air profiling at C band compared to UHF band.
Originally built to study precipitating systems, the C-band Doppler radar considered here was upgraded in 1993 for the first field campaign of the TRAC project for the detection of the clear air coherent organizations within the daytime summer ABL. This campaign demonstrated that this radar was able to provide, in daytime clear air ABL, coherent and continuous velocity fields from 0.1 km up to 3 km above ground, and over a horizontal range reaching at least 30 km in the ABL, depending on the atmospheric conditions (Lohou et al. 1998b). C-band vertical profiles of wind variables and reflectivity were retrieved with a 100-m vertical resolution and 10-min time resolution with the VVP technique, using the volume scan data continuously collected by this radar. The comparison, made over 11 fair weather days, of common observations with the UHF wind profilers during the 0600–1600 UTC time period has produced the following results.
A close similarity was observed in the reflectivity pattern of the time–height sections provided by both types of wavelength. In particular, the inversion capping the ABL top is materialized with a branch of enhanced reflectivity at UHF band, which is now well known, but also in the C-band data. This feature was used to derive a 2D topographical map of the ABL top using the steering facility of the C-band radar (an advantage over the conventional UHF wind profiler). Moreover, on a particular day, both radars detected at the same levels seven layered echoes above the ABL evolving with time. Five of them were positioned close to singularities in the humidity or temperature profiles provided by a radiosounding. The two remaining thin echo sheets appear to have nothing to do with the mean atmospheric thermodynamics. The maximum reflectivity ratio between UHF- and C-band radar reached 15 dBZ in the ABL. This result agrees with several previous works that showed that at centimetric wavelength the main origin of clear air ABL echoes is material (e.g., insects), whereas at UHF backscattering from air refractive index irregularities is predominant.
The mean standard deviation of the difference in wind speed and wind direction measured by both types of radar reached 1.4 m s−1 and 12.7°, respectively. This agreement may be seen as very good when considering that statistical errors of both instruments are cumulative, and that they do not resolve the same atmospheric scales due to their different types of data collection. This agreement tends to show that the different echo sources at the two wavelengths used do not have a noticeable impact on the horizontal wind measurement. On the other hand, wind velocity at C band below 200 m was usually greater than at UHF band. This result can be attributed to ground clutter contamination being more effective for the UHF wind profilers, which are operated with a larger beamwidth.
The mean daily evolution of the vertical velocity measured by the two UHF wind profilers confirmed Angevine's (1997) observations of a relatively strong downward motion reaching, on average, −0.3 m s−1 during the most energetic phase of the ABL. This result, which does not seem physical (one expects, on average, negligible vertical motion), was obtained not only with the vertical velocity measured directly by the zenith beam but also with the vertical velocity deduced from opposite oblique beams. This latter method tends to ensure that the apparent vertical bias is not linked to the processing of ground echo removal near the dc line. A different and more probable view of the vertical transport is offered by the C-band radar on the mesoscale. On average, the horizontal divergence deduced by the VVP method is negative in the daytime ABL. The convergence is maximum in the early morning and decreases with time during the ABL development, with an average value of about −0.5 × 10−4 s−1. The vertical integration of the divergence leads to a mean vertical velocity of 3 cm s−1. This value, which corresponds to half the morning deepening velocity of the ABL, is not negligible for the understanding of the lower-atmosphere vertical transport. However, this result has to be confirmed by other observations because it was difficult to assess the accuracy of the vertical velocity measurements.
The present results concerning the ABL profiling capability at centimetric wavelength are quite encouraging. However, it is a first step because the observations were limited to daytime and to a particular period of the year. Observations in less favorable conditions (nighttime, winter, etc.) are needed in order to conclude. One of the advantages of centimetric wavelength is the possibility of getting a smaller beamwidth with reduced antenna dimension and so decreasing ground clutter contamination, particularly at levels below 200 m, where presently wind measurements by UHF profilers are difficult, especially when the atmospheric signal is weak compared to clutter echoes.
Acknowledgments
TRAC was funded by the French organization CNRS/INSU, EDF, LA, and Météo-France. The C-band RONSARD radar is the property of the CETP laboratory, which operated it during the experiment. We particularly acknowledge A. Muschinski and the two other reviewers whose comments and suggestions have improved the content and clearness of this paper.
REFERENCES
Angevine, W. M., 1997: Errors in mean vertical velocities measured by boundary layer wind profilers. J. Atmos. Oceanic Technol., 14 , 565–569.
Angevine, W. M., White A. B. , and Avery S. K. , 1994: Boundary-layer depth and entrainment zone characterization with a boundary-layer profiler. Bound.-Layer Meteor., 68 , 375–385.
Angevine, W. M., Bakwin P. S. , and Davis K. J. , 1998: Wind profiler and RASS measurements compared with measurements from a 450-m-tall tower. J. Atmos. Oceanic Technol., 15 , 818–825.
Battan, L. J., 1973: Radar Observation of the Atmosphere. University of Chicago Press, 324 pp.
Browning, K. A., 1972: Atmospheric research using the Defford radar facility. Weather, 27 , 1–6.
Browning, K. A., and Wexler R. , 1968: The determination of kinematic properties of a wind field using Doppler radar. J. Appl. Meteor., 7 , 105–113.
Campistron, B., Bénech B. , Dessens J. , Jacoby-Koaly S. , Dupont E. , and Carissimo B. , 1997: Performance evaluation of a UHF boundary layer radar in raining conditions based on disdrometer measurements. Eighth Int. Workshop on Technical and Scientific Aspects of MST Radar, Bangalore, India, Solar–Terrestrial Energy Program, 334–337.
Campistron, B., and Coauthors. 1999: The Turbulence Radar Aircraft Cells, TRAC-98 experiment. Preprints, 13th Symp. on Boundary Layers and Turbulence, Dallas, TX, Amer. Meteor. Soc., 620–623.
Carter, D. A., Gage K. S. , Ecklund W. L. , Angevine W. M. , Johnston P. E. , Riddle A. C. , Wilson J. , and Williams C. R. , 1995: Developments in UHF lower tropospheric wind profiling at NOAA's Aeronomy Laboratory. Radio Sci., 30 , 977–1001.
Cohn, S. A., Mayor S. D. , Grund C. J. , Weckwerth T. M. , and Senff C. , 1998: The Lidars in Flat Terrain (LIFT) experiment. Bull. Amer. Meteor. Soc., 79 , 1329–1343.
Ecklund, W. L., Carter D. A. , and Basley B. B. , 1988: A UHF wind profiler for the boundary layer: Brief description and initial results. J. Atmos. Oceanic Technol., 5 , 432–441.
Eymard, L., and Weill A. , 1982: Investigation of clear air convective structures in the PBL using a dual Doppler radar and a Doppler sodar. J. Appl. Meteor., 21 , 1891–1906.
Gage, K. S., Williams C. R. , Ecklung W. L. , and Johnston P. E. , 1999: Use of two profilers during MCTEX for unambiguous identification of Bragg scattering and Rayleigh scattering. J. Atmos. Sci., 56 , 3679–3691.
Garratt, J. R., 1992: The Atmospheric Boundary Layer. Cambridge University Press, 316 pp.
Gossard, E. E., and Strauch R. G. , 1983: Radar Observation of Clear Air and Clouds. Elsevier, 280 pp.
Gossard, E. E., Gaynor J. E. , Zamora R. J. , and Neff W. D. , 1985: Finestructure of elevated stable layers observed by sounder and in situ tower sensors. J. Atmos. Sci., 42 , 2156–2169.
Grimsdell, A. W., and Angevine W. M. , 1998: Convective boundary layer height measurement with wind profilers and comparison to cloud base. J. Atmos. Oceanic Technol., 15 , 1331–1338.
Grund, C. J., Banta R. M. , George J. L. , Howell J. N. , Post M. J. , Richter R. A. , and Weickmann A. M. , 2001: High-resolution Doppler lidar for boundary layer and cloud research. J. Atmos. Oceanic Technol., 18 , 376–393.
Hashiguchi, H., Fukao S. , Tsuda T. , Yamanaka M. D. , Tobing D. L. , Sribimawati T. , Harijono S. W. B. , and Wiryosumarto H. , 1995: Observations of the planetary boundary layer over equatorial Indonesia with an L band clear-air Doppler radar: Initial results. Radio Sci., 30 , 1043–1054.
Jacoby-Koaly, S., Campistron B. , Bernard S. , Bénech B. , Girard F. , Dessens J. , Dupont E. , and Carissimo B. , 2001: Turbulent dissipation rate in the boundary layer via UHF wind profiler Doppler spectral width measurement. Bound.-Layer Meteor., in press.
Kunkel, K. E., Eloranta E. W. , and Weinman J. A. , 1980: Remote determination of winds, turbulence spectra and energy dissipation rates in the boundary layer from lidar measurements. J. Atmos. Sci., 37 , 978–985.
Lenschow, D. H., 1986: Probing the Atmospheric Boundary Layer. Amer. Meteor. Soc., 269 pp.
Little, C. G., 1969: Acoustic methods for remote probing of the lower atmosphere. Proc. IEEE, 57 , 571–578.
Lohou, F., Druilhet A. , Foster P. , Campistron B. , Gervaise C. , Loubet A. , and Delahaye J. Y. , 1995: Measurement of C2 n in the lower atmosphere with an airborne refractometer. Comparison with simultaneous Doppler radar observations. Preprints, 27th Conf. on Radar Meteorology, Vail, CO, Amer. Meteor. Soc., 290–292.
Lohou, F., Campistron B. , Druilhet A. , Foster P. , and Pagès J. P. , 1998a: Turbulence and coherent organizations in the atmospheric boundary layer: A radar-aircraft experimental approach. Bound.-Layer Meteor., 86 , 147–179.
Lohou, F., Druilhet A. , and Campistron B. , 1998b: Spatial and temporal characteristics of horizontal rolls and cells in the atmospheric boundary layer based on radar and in situ observations. Bound.-Layer Meteor., 89 , 407–444.
May, P. T., and Wilczak J. M. , 1993: Diurnal and seasonal variations of boundary layer structure observed with a radar wind profiler and RASS. Mon. Wea. Rev., 121 , 673–682.
May, P. T., and Strauch R. G. , 1998: Reducing the effect of ground clutter on wind profiler velocity measurements. J. Atmos. Oceanic Technol., 15 , 579–586.
McAllister, L. G., 1968: Acoustic sounding of the lower troposphere. J. Atmos. Terr. Phys., 30 , 1439–1440.
Mead, J. B., Hopcraft G. , Frasier S. J. , Pollard B. D. , Cherry C. D. , Schaubert D. H. , and McIntosh R. E. , 1998: A volume-imaging radar wind profiler for atmospheric boundary layer turbulence studies. J. Atmos. Oceanic Technol., 15 , 849–859.
Muschinski, A., 1998: The first moments of the variance and cross-spectra of standard and interferometric clear-air Doppler radar signals. NCAR Tech. Note 441, 102 pp.
Muschinski, A., and Wode C. , 1998: First in situ evidence for coexisting submeter temperature and humidity sheets in the lower free troposphere. J. Atmos. Sci., 55 , 2893–2906.
Muschinski, A., and Lenschow D. H. , 2001: Future directions for research on meter- and submeter-scale, atmospheric turbulence. Bull. Amer. Meteor. Soc., 82 , 2831–2843.
Nastrom, G. D., and VanZandt T. E. , 1996: Biases due to gravity waves in wind profiler measurements of winds. J. Appl. Meteor., 35 , 243–257.
Petitdidier, M., Sy A. , Garrouste A. , and Delcourt J. , 1997: Statistical characteristics of the noise power spectral density in UHF and VHF wind profilers. Radio Sci., 32 , 1229–1247.
Richter, J. H., 1969: High-resolution tropospheric radar sounding. Radio Sci., 4 , 1261–1268.
Rogers, R. R., Knight C. A. , Tuttle J. D. , Ecklund W. L. , Carter D. A. , and Ethier S. A. , 1992: Radar reflectivity of the clear air at wavelengths of 5.5 and 33 cm. Radio Sci., 27 , 645–659.
Schols, J. L., and Eloranta E. W. , 1992: Calculation of area-averaged vertical profiles of the horizontal wind velocity from volume-imaging lidar data. J. Geophys. Res., 97 , 18395–18407.
Stevens, B., and Lenschow D. H. , 2001: Observations, experiments, and large-eddy simulation. Bull. Amer. Meteor. Soc., 82 , 283–294.
Stull, R. B., 1988: An Introduction to Boundary Layer Meteorology. Kluwer Academic, 666 pp.
Tatarskii, V. I., and Muschinski A. , 2001: The difference between Doppler velocity and real wind velocity in single scattering from refractive-index fluctuations. Radio Sci., 36 , 1405–1423.
Waldteufel, P., and Corbin H. , 1979: On the analysis of single-Doppler radar data. J. Appl. Meteor., 18 , 532–542.
Wilson, J. W., Weckwerth T. M. , Vivekanandan J. , Wakimoto R. M. , and Russell R. W. , 1994: Boundary layer clear-air radar echoes: Origin of echoes and accuracy of derived winds. J. Atmos. Oceanic Technol., 11 , 1184–1206.
Elementary data-processing volume used in the VVP technique; D and Δh are the horizontal diameter and the vertical depth of the horizontal cylindrical slice, respectively. The C-band radar is positioned at the origin of the coordinate system
Citation: Journal of Atmospheric and Oceanic Technology 19, 6; 10.1175/1520-0426(2002)019<0899:CORRAV>2.0.CO;2
Time–height sections of horizontal wind and reflectivity deduced from observations made by (a), (c) the C-band radar and (b), (d) the UHF wind profiler on 20 Jun 1998. Contours are in m s−1 for the wind speed, and in dBZ for the reflectivity factor. The circles represent the base of the inversion layer capping the ABL
Citation: Journal of Atmospheric and Oceanic Technology 19, 6; 10.1175/1520-0426(2002)019<0899:CORRAV>2.0.CO;2
Histograms of the difference in (a) wind speed and (b) wind direction between C-band and UHF profiler measurements. The observations were collected between 200 and 2500 m in 11 days of clear air ABL
Citation: Journal of Atmospheric and Oceanic Technology 19, 6; 10.1175/1520-0426(2002)019<0899:CORRAV>2.0.CO;2
Reflectivity field observed by the C-band radar on conical scannings (PPI) performed at 7.1° (0900 UTC) and 6.4° (1400 UTC) elevation on 20 Jun 1998
Citation: Journal of Atmospheric and Oceanic Technology 19, 6; 10.1175/1520-0426(2002)019<0899:CORRAV>2.0.CO;2
(a) Vertical profiles of the reflectivity factor measured by the C-band radar and the UHF profiler at 0700 UTC 20 Jun 1998. (b) Vertical profiles of temperature and relative humidity given by the radiosounding at 0658 UTC launched the same day from the site of the UHF profiler
Citation: Journal of Atmospheric and Oceanic Technology 19, 6; 10.1175/1520-0426(2002)019<0899:CORRAV>2.0.CO;2
Vertical section of reflectivity deduced from the interpolation of the C-band observations collected during the full volume exploration sequence of 1400 UTC 20 Jun 1998
Citation: Journal of Atmospheric and Oceanic Technology 19, 6; 10.1175/1520-0426(2002)019<0899:CORRAV>2.0.CO;2
Mean time series, averaged over the period 18–24 Jun 1998 and over the height interval 0.2–1 km, of the vertical velocity measured by the vertical beam of the profilers UHF5 and UHF3, and of the vertical velocity deduced from the radial velocity obtained with the oblique beam of the UHF5 profiler (Wop). A running average over 3600 s was applied. From 18 Jun to 24 Jun, sunrise time was close to 0352 UTC and sunset close to 1957 UTC
Citation: Journal of Atmospheric and Oceanic Technology 19, 6; 10.1175/1520-0426(2002)019<0899:CORRAV>2.0.CO;2
(a) Mean vertical profile of the vertical velocity measured by the vertical beam of the UHF profiler in 11 selected days taken between 18 Jun and 3 Jul 1998, and limited to the time interval 1000–1400 UTC. Height is normalized with the ABL top Zi. (b) Histogram of the same parameter deduced in the same time and date interval but for measurements made between 0 and Zi
Citation: Journal of Atmospheric and Oceanic Technology 19, 6; 10.1175/1520-0426(2002)019<0899:CORRAV>2.0.CO;2
Mean vertical profile, centered at 0700 UTC 20 Jun 1998 and averaged over 1 h, of the vertical velocity measured by the UHF profiler and of the horizontal divergence deduced from C-band observations
Citation: Journal of Atmospheric and Oceanic Technology 19, 6; 10.1175/1520-0426(2002)019<0899:CORRAV>2.0.CO;2
Mean composite time series of the C-band horizontal divergence and vertical velocity obtained from the average of the observations acquired during the 11 selected days taken in the date interval 18 Jun–3 Jul 1998, and within the height layer between 0 and the ABL top Zi. A running average over 3600 s was added
Citation: Journal of Atmospheric and Oceanic Technology 19, 6; 10.1175/1520-0426(2002)019<0899:CORRAV>2.0.CO;2
