Meteorologische Zeitschrift Vol 18 No 2 155-162 (April 2009) Open Access Articleccopy by Gebruder Borntraeger 2009
Doppler lidar observations of sensible heat flux andintercomparisons with a ground-based energy balance stationand WRF model output
JENNY CLARE DAVIS1lowast CHRISTOPHER G COLLIER1 FAY DAVIES1 GUY N PEARSON1 RALPH
BURTON2 and ANDREW RUSSELL3
1The University of Salford Salford Greater Manchester UK2School of Earth and Environment The University of Leeds Leeds UK3School of Earth Atmospheric and Environmental Sciences The University of Manchester Manchester UK
(Manuscript received September 9 2008 in revised form February 12 2009 accepted February 12 2009)
Abstract
During the Convective and Orographically induced Precipitation Study (COPS) a scanning Doppler lidar
was deployed at Achern Baden-Wuttemberg Germany from 13th June to 16th August 2007 Vertical velocityprofiles (lsquoraysrsquo) through the boundary layer were measured every 3 seconds with vertical profiles of horizontalwind velocity being derived from performing azimuth scans every 30 minutes During Intense ObservationPeriods radiosondes were launched from the site In this paper a case study of convective boundary layer
development on 15th July 2007 is investigated Estimates of eddy dissipation rate are made from the verticallypointing lidar data and used as one input to the velocity-temperature co-variance equation to estimate sensibleheat flux The sensible heat flux values calculated from Doppler lidar data are compared with a surface basedenergy balance station and output from the Weather Research and Forecasting (WRF) model
Zusammenfassung
Wahrend der Studie zu konvektiven und orographisch induzierten Niederschlagen (Convective and Orograph-ically induced Precipitation Study - COPS) wurde das 15 microm scannende Doppler-Lidar der OrganisationUFAM (Universitiesrsquo Facility for Atmospheric Measurement) an seinem Messstandort in Achern Baden-Wurttemberg Deutschland vom 13 Juni bis zum 16 August 2007 von der Universitat Salford UK betriebenVertikale Windgeschwindigkeitsprofile der atmospharischen Grenzschicht wurden alle 3 Sekunden gemessenAus 30-minutigen Azimut-Scans wurden die Horizontalgeschwindigkeitsprofile abgeleitet Wahrend der in-tensiven Beobachtungsperioden wurden zusatzlich Radiosonden vom Messplatz gestartet Im vorliegendenArtikel wird eine konvektive Fallstudie am 15 Juli 2007 naher untersucht Vom vertikal gerichteten Lidar wirddie Eddy-Dissipationsrate abgeschatzt Sie wird in der Kovarianzgleichung zwischen Windgeschwindigkeitund Temperatur zur Abschatzung des fuhlbaren Warmeflusses verwendet Die verwendeten Annahmen wer-den explizit erwahnt Der fur die verwendete Methode benotigte Grad der atmospharischen Instabilitat wirddurch den Vergleich mit den Radiosondendaten bestatigt Die abgeleiteten Ergebnisse stimmen gut mit denDaten der Energiebilanzstation und des WRF Modells uberein
1 Introduction
The Convective and Orographically-Induced Precipita-tion Study (COPS) field trial was conducted in theBlack Forest region of Germany during the summer of2007 Its aim was to advance the quality of forecasts oforographically-induced convective precipitation usingextensive observations and modelling (COPS 20071)The University of Salford own and operate a scan-ning 15 microm Doppler lidar system and were funded bythe Universitiesrsquo Facility for Atmospheric Measurement(UFAM) through the UK Natural Environment ResearchCouncil to participate in the project The Doppler lidarwas mounted in a mobile laboratory and has a full hemi-
lowastCorresponding author Jenny Clare Davis The University of Salford Sal-
ford Greater Manchester M5 4WT UK e-mail CDavissalfordacuk1Convective and Orographically-Induced Precipitation Project 2007 avail-
able at wwwcops2007de
spheric scanning capability using a dual-mirror scan-ning system Additionally a 14 channel microwave ra-diometer and a Campbell Scientific automatic weatherstation (AWS) were deployed at the Achern site Allinstruments were set up to run continuously from 13th
June to 16th August 2007 The lidar built by Halo Pho-tonics was deployed and being autonomous was leftunattended and fully controlled over the internet It hadbeen previously deployed during the World Weather Re-search Project (WWRP) Helsinki Testbed (BOZIER etal 2007)
The Salford Autonomous Doppler Lidar System op-erating at a wavelength of 15 microm employs novel fi-bre optic technology and the design approach has ledto a new type of eye-safe Doppler lidar providing a highlevel of performance and exhibiting exceptional stabil-ity which was demonstrated by this successful deploy-ment A detailed analysis of wind measurements pre-
0941-294820090367 $ 360
DOI 1011270941-294820090367 ccopy Gebruder Borntraeger Berlin Stuttgart 2009
156 Davis et al Doppler lidar observations Meteorol Z 18 2009
Figure 1 The Salford University Doppler lidar with antenna (inset)
Table 1 Autonomous Doppler lidar system parameters
Parameter Value
Operating Wavelength 15 microm
Pulse Repetition Frequency 20 kHz
Energy per pulse 10 microJ
Pulse duration 150 ns
Beam divergence 50 microrad
Range gate length 30 m
Lens diameter 75 mm
Focal length Infinity
Minimum range 100 m
Maximum Range 7 km
Temporal resolution 01ndash30 s
viously made with this lidar system has been given inPEARSON et al (2009) Figure 1 illustrates the Univer-sity of Salford Doppler lidar system in position at theCOPS field site in Achern Germany
The Doppler lidar data was selected to test a method-ology discussed in depth by DAVIS et al (2008) andGAL-CHEN et al (1992) for calculating QH in an ur-ban area This involves using an estimation of the ki-netic energy dissipation and the vertical gradient ofthe third moment of the vertical velocity During theCOPS field trial there were few instruments measur-ing QH and none of these were located at the Achernsite However the conditions and scanning strategy em-ployed meant that the data were suitable for testing thismethodology in a semi-rural area and since QH is animportant variable when investigating the developmentof thermals (THIELEN et al 2000) and associated rain-fall (BENNETT et al 2006 ROZOFF et al 2003) theCOPS field trial was considered an good opportunity forthis methodology to be tested
2 The Salford University HaloPhotonics Doppler lidar
The lidar system was developed to meet the require-ments of the University of Salford being capable ofunattended autonomous operation The University ofSalford specification required a Doppler lidar systemcapable of providing high temporal and spatial resolu-tion measurements of the wind velocity and backscatterwithin the atmospheric boundary layer The system isportable and rugged and capable of being used for fieldwork and long term measurements (BOZIER et al 2007PEARSON et al 2009) The lidar works by transmit-ting a short laser pulse of approximately 1x10minus7 s andcollecting the backscattered signal from the illuminatedaerosol targets along the path of the laser beam ndash see Ta-ble 1 for more details of the lidar The primary scatterersare small atmospheric aerosol particles whose diametersare within an order of magnitude of the lidar wavelengthAt optical wavelengths scattering within the lower at-mosphere is primarily by particles with diameters lessthan 3 microm which are sufficiently small to be advected bythe wind and serve as an effective tracer of wind veloc-ity (DAVIES et al 2003 HARDESTY et al 1992) FRE-LICH 1995 SCHWIESOW 1986) Accurate estimates ofthe radial component of the velocity (along the line ofsight of the laser beam) are produced as a spatial aver-age over the sensing volume of the transmitted pulse
The lidar system has a modular design arranged inthree separate units the optical base unit the weather-proof antenna consisting of the telescope and associ-ated commercially sensitive scanning electronics andthe signal processing and data acquisition unit The baseunit has approximate dimensions 56 cm x 54 cm x 18 cmand contains the optical source interferometer receiverand the electronics The weather-proof antenna is at-tached to the base unit via a 1rdquo diameter 1 m long opticalfibre conduit The antenna can be deployed permanently
Meteorol Z 18 2009 Davis et al Doppler lidar observations 157
Figure 2 Wind vector plot for 150707 0800ndash2200 h UTC arrows
represent relative horizontal wind speed and direction Due to some
wind velocities being very small the arrows appear as dots The
legend (bottom left) shows an arrow representing a 2 m sminus1 wind
speed with westerly direction
outside whilst the base unit and data acquisition sys-tem are housed within a mobile laboratory environmentThe signal processing has been developed with a viewto providing a high level of flexibility with respect to thedata acquisition parameters Users are able to set param-eters such as the range gate length maximum range andnumber of pulses accumulated for each measurementas detailed in Table 1 During the field campaign mea-surements were limited only by atmospheric moisture(rainfog) a lightning strike and a mains power failureThe Doppler lidar system is described in greater detailin PEARSON et al (2009) HALO PHOTONICS (20082)
3 Measurements
The UFAM pulsed Doppler lidar is capable of measur-ing
bull Directly Radial wind velocity relative backscatterintensity
bull Indirectly Horizontal velocities their variances andcovariances atmospheric backscatter coefficient (β)and turbulence kinetic energy dissipation rate (ε)
During much of the field campaign the lidar systemperformed a series of scan patterns pre-programmedinto the system software Errors are given in Table 1and described in more detail in PEARSON et al (2009)Generally the system carried out
bull 5 minute azimuth scan at 5 different elevations (2030 40 45 and 60) but on the case discussedhere azimuth scans were performed at 60
2The Halo Photonics pulsed Doppler LiDAR system available at wwwhalo-
photonicscom
Figure 3 Full day view 150707 showing thermals building
throughout the day
Figure 4 Doppler lidar and radiosonde horizontal wind speeds to
800 m
bull 25 minute vertically fixed stare consisting of 334radial velocity lsquorayrsquo measurements
The azimuth scan was performed to determine windvelocity profiles derived from VAD analysis (BROWN-ING and WEXLER 1968) and these data were concate-nated enabling a half-hourly vector profile as illustratedin Figure 2 The vertically pointing stares were concate-nated to produce a complete overview of vertical veloc-ities measured by the Doppler lidar each day as illus-trated in Figure 3 Both Figure 2 and Figure 3 showdata up to a height of 1100 m Beyond this range onthis day there was no signal return The data illustratedcan be used to calculate a variety of products includ-ing mixed layer height (COLLIER et al 2005 MOK
and RUDOWICZ 2004) and turbulence statistics DAVIES
et al (2004) LENSCHOW et al (2000) FRELICH andCORNMAN (2002) and BANAKH et al (1999)
The data from the 15th July 2007 show that thermalswere building throughout the day and the synoptic situa-
158 Davis et al Doppler lidar observations Meteorol Z 18 2009
tion indicated that a convergence line passed from westto east to the north of Achern High convectivity partic-ularly at around 1500 UTC was measured by the lidarThe convective conditions suggested a suitable day touse lidar data to calculate (QH ) a parameter importantin monitoring the development of thermals A methoddiscussed in detail by DAVIS et al (2008) is outlinedand tested here
It was first necessary to confirm the reliability ofthe Doppler lidar data Comparisons were made be-tween horizontal velocity data measured by the lidar ra-diosonde and a 1290 MHz wind profiler and are shownhere
Figure 4 illustrates similarities between horizontal ve-locities measured by Doppler lidar and radiosonde theradiosonde was launched at 1000 UTC and the azimuthscan was performed from 1000 UTC to 1005 UTCWhen this data was taken it is considered that the bound-ary layer top was at around 800 m and the lidar data be-comes noisy and therefore un-useable above this level
Another comparison was made between the UFAMDoppler lidar data and the University of Manchester op-erated UFAM funded 1290 MHz wind profiler Thisis illustrated in Figure 5 As with the radiosonde data(Figure 4) it can be seen that there are similarities be-tween the two datasets There are some areas of smallscale differences but generally the two remote sensinginstruments are providing comparable data Having es-tablished that the Doppler lidar is a suitable instrumentto make atmospheric measurements and that the ambi-ent conditions were appropriate more complex calcula-tions could be performed to investigate other parameterssuch as QH
The chosen method assumes unstable conditions so toconfirm whether this was the situation for the 15th July2007 case potential temperature profiles were derivedfrom two radiosonde ascents launched from the Achernsite These profiles suggesting unstable conditions areshown in Figure 6
4 Methodology
Having established unstable conditions it was possibleto calculate QH from the vertical velocity-potential tem-
perature covariance term (wprimeθprime) in (41) following theprocedure outlined in GAL-CHEN et al (1992)
g
θ0wprimeθprime =
1
ρ0wprime
partpprime
partz+
ε
3+
part
partz
(
1
2wprime3
)
(41)
WYNGAARD and COTE (1971) and DAVIS et al
(2008) found that the pressure term partpprime
partz was small com-pared with the other terms of (42) for unstable condi-tions and will therefore be ignored in the subsequentanalysis which can be applied only to cases of convec-tion hence
wprimeθprime asympθ0
g
[
part
partz
(
1
2wprime3
)
+ε
3
]
(42)
Figure 5 Vector plot from UFAM 1290 MHz wind profiler showing
similar features to the lidar (Figure 2) with easterlies early in the
morning becoming southerly as the morning progresses and veering
westerly in the afternoon
Figure 6 Potential temperature profiles on 150707 suggesting
unstable conditions
The mean vertical velocity w is calculated for eachrange gate and averaged over the 25 minute scan periodThis is done for each of the nineteen 30 m-long rangegates from 135 m to 705 m using data from a verti-cally pointing stare scan The wrsquo is the deviation fromthe mean vertical velocity Then wrsquo3 is calculated foreach point and averaged over the 25 minute scan dura-
tion to form a profile of wprime3
In order to use (42) to calculate wprimeθprime it is also neces-sary to estimate ε One way of estimating ε is to exam-ine the line spectra of the longitudinal velocity correla-tion In the inertial subrange the expected relationshipis (BATCHELOR 1967)
f(κ) = αε23κminus53 (43)
where κ is the wave number α is a universal constant(05) and f(κ) is the Fourier transform of the longitudi-nal velocity correlation A -53 slope indicates the pres-
Meteorol Z 18 2009 Davis et al Doppler lidar observations 159
Table 2 Errors associated with Doppler lidar calculations of ε and wprime2
Error type and extent
Vertical velocity variancewprime2 Bias = 002 m2 sminus2 PEARSON et al 2008
Random = plusmn068 m2 sminus2
Turbulent kinetic energy Error = 10 FRELICH and CORNMAN 2002
dissipation rate ε
Figure 7 Spectrum from the GAL-CHEN et al (1992) spatial
method A -53 gradient indicates the presence of the inertial sub-
range
Figure 8 Temporal spectrum as reported by CHAMPAGNE et al
(1977) The -23 slope highlights the presence of the inertial sub-
range
ence of an inertial subrange and is shown in Figure 7This will be referred to henceforth as the spatial spec-tra method We calculate the spectra for each ray from135 to 705 m and average the spectra over the 25 minutetime period
An alternative method for calculating spectra is pro-posed by CHAMPAGNE et al (1977) Here Taylorrsquos Hy-
pothesis (STULL 1988) is assumed and is used to cal-culate the wave number The u is calculated from the li-dar data using VAD analysis The u is the average meanwind over the 135ndash705 m height range The spectrumcan be calculated from
nS (n) = 068ε23
(
2πn
u
)minus23
(44)
where S(n) is the spectral energy of frequency n andu is the mean wind speed Here the inertial subrangeis characterized by a ndash23 gradient and is shown in Fig-ure 8 This will be referred to henceforth as the temporalspectra method
When performing spectral analysis to obtain ε thedata were detrended to remove any nonstationary be-haviour However errors are still expected and these arelisted in Table 2
With unstable conditions QH can thus be calculatedfrom
QH = ρ0CP wprimeθprime (45)
taking CP the heat capacity of air at constant pressureas 1004 J kminus1 kgminus1 Kminus1 ρ0 the air density as 1275kg mminus3 and where is the potential temperature-verticalvelocity covariance calculated in (42) (GAL-CHEN etal 1992) (43) also assumes unstable conditions whichwere present on this day as illustrated in Figure 6
5 Results
The Salford Autonomous Doppler Lidar System was de-ployed for COPS in Achern Germany where it col-lected data almost continuously for 3 months A casestudy of convective development is described here andthe results include turbulence spectra and some time-series of ε and QH Figure 7 shows a spectrum producedfrom a vertically-pointing stare taken at 1100 UTC onthe 15th July 2007 in order to obtain a single value of εThe spectrum in Figure 7 was calculated using the spa-tial method presented by GAL-CHEN et al (1992) Inaddition in order to confirm the results shown in Fig-ure 7 spectra were also plotted using the same datasetand the method proposed by CHAMPAGNE et al (1977)as shown in Figure 8 The points deviating from the ex-pected ndash23 gradient at high wavenumbers are consid-ered to be due to noise (MAYOR et al 1997)
160 Davis et al Doppler lidar observations Meteorol Z 18 2009
Figure 9 Eddy dissipation rates calculated from Doppler lidar data
using both spatial and temporal methods on 150707
Figure 10 QH timeseries calculated for 150707 showing an esti-
mation of the expected errors Half-hourly averages of QH as mea-
sured by the IMK-FZK energy balance station for that day are also
shown
The time series in Figure 9 illustrates ε derived fromboth temporal and spatial methods developing through-out the day as would be expected ndash increasing towardsnoon peaking at 1500 UTC and then decreasing towardsthe evening as convection is developing in the morn-ing and then dissipating in the evening (STULL 1988)With the estimates for ε and vertical velocity fluctua-tions (wprime3) QH was calculated as illustrated in Figure10 It was decided that although the ε values were sim-ilar for both spectra the values derived from the spatialspectra would be utilised to calculate QH as per GAL-CHEN et al (1992) Figure 10 also shows a timeseriesof QH measured using an Energy Balance Station atthe COPS operation centre Baden Airpark 10 km northof Achern where the Doppler lidar was located but stillin the Rhine Valley
Figure 11 QH from Doppler lidar data and WRF model data
When considering the results achieved using the datacollected with the Doppler lidar certain factors need tobe considered for example how did the results com-pare with other measurements Figure 10 shows the li-dar data plotted against data collected from an energybalance station located at Baden Airpark 10 km to thenorth of the Doppler lidar The energy balance stationmade point measurements of QH at a single point intime at the surface and averaged over half an hour Themeasurements made by the lidar are averages over thedepth of the boundary layer up to 705 m and while show-ing a general trend cannot be the same for the followingreasons
bull The two instruments were not collocated
bull The lidar as a remote sensor provides a volumeaverage over a pencil-shaped beam over a range of120 m to 705 m above the surface
The QH values measured by the lidar are higher thanthose measured at Baden Airpark It is considered thatthe relatively high values of QH calculated from thelidar data compared with those from the energy bal-ance station result from the nature of the instrumentsand their associated lsquofootprintsrsquo (SCHMID 1994) Li-dar measurements are averaged over the lowest sim700m of the boundary layer and so have larger footprintsthan ground based instruments in the order of severalkilometres WOOD and MASON (1993) BELCHER andWOOD (1996) and WOOD et al (2001) discuss the ef-fects of increasing form drag and roughness length whenflow is directed across orography such as the mountainsto the east of the Achern site Since the mountains aresome 8 km away within the fetch of the lidar it is con-sidered that it the high eddy dissipation rates are due tothis orography These high values greatly influence themagnitude of the QH The instruments at Baden Air-park are at surface level so have lower fetch and are un-likely to be affected by the orography It should also be
Meteorol Z 18 2009 Davis et al Doppler lidar observations 161
noted that on this occasion the synoptic situation indi-cated that the convergence line mentioned in section 3was more active to the north of Achern
Despite these differences the results are comparablehaving a reasonable order of magnitude and showing adiurnal trend as expected
Output from the NERC WRF model runs for this dayis shown in Figure 11 These data also show a diurnaltrend similar to those shown in Figure 10 It is not real-istic to expect remotely sensed data to replicate the re-sults from a computer model but as with the surface val-ues the results are within the same order of magnitudethere is a similar trend in the model output with valuesrising towards noon and dropping towards sunset It isinteresting to note that where the model output suggestshigh heat flux values at 0800 UTC the lidar does notAlthough the measured QH values were comparativelyhigh at around 1500 UTC these values were unrealisti-cally so It is thought that this may be due to a mesoscaleevent that occurred at about 1500 UTC that day as men-tioned previously Investigations into this are ongoing
6 Conclusions and future work
It has been shown that by using a method outlined byDAVIS et al (2008) GAL-CHEN et al (1992) it is pos-sible to calculate values of QH and it is considered thatthese values fall within an acceptable range It is notedthat the diurnal cycle of the lidar-derived QH is closer tothe diurnal cycle of ε whereas both the model and sur-face measurements of QH peak earlier in the day QH
calculated from Doppler lidar data using the method de-scribed here would be expected to differ from that cal-culated using a surface based lsquopointrsquo sensor such as anenergy balance instrument or a sonic anemometer sincethe data collected with a Doppler lidar is integrated overa lsquovolumersquo of atmosphere aloft advecting through thelaser beam However these measurements are a usefuladdition to the other measurements provided by the li-dar and require no extra equipment
It is considered that further investigations of employ-ing this method to calculate QH are needed to establishthe overall applicability of the method and build on thepreliminary calculations It is expected that such inves-tigations will include
bull Calculation of ε using alternative methods
bull investigation of ε at different heights and atmo-spheric conditions
bull verification of Doppler lidar estimations of ε and wprime3and
bull investigation over different surface types
Acknowledgments
The authors would like to acknowledge Dr N KALT-HOFF at IMK FZK Karlsruhe for supplying QH datafrom their energy balance station located at Baden Air-field Germany collected as part of COPS July 2007Also Dr A GADIAN from the University of Leeds forhis help in securing the WRF model output Dr E NOR-TON from the University of Manchester for making theUFAM 1290 MHz wind profiler quicklook plots avail-able online HP SCHMID for the use of the flux source-area model (FSAM) finally UK COPS for fundingthrough NERC
References
BANAKH VA IN SMALIKHO F KOPP C WERNER1999 Measurements of turbulent energy dissipation ratewith a CW Doppler lidar in the atmospheric boundarylayer J Atmos Oceanic Technol 16 1044 mdash1061
BATCHELOR GK 1967 The theory of homogeneous tur-bulence 5th ed ndash Cambridge University Press London
BELCHER SE A WOOD 1996 Form and wave drag dueto stably stratified turbulent flow over low ridges ndash QuartJ Roy Meteor Soc 122 863mdash902
BENNETT LJ KA BROWNING AM BLYTH DJPARKER PA CLARK 2006 A review of the initiation ofprecipitating convection in the United Kingdom ndash QuartJ Roy Meteor Soc 132 1001mdash1020
BOZIER KE GN PEARSON CG COLLIER 2007Doppler lidar observations of Russian forest fire plumesover Helsinki ndash Weather 62 203ndash208
BROWNING KA R WEXLER 1968 The Determinationof Kinematic Properties of a Wind Field using DopplerRadar ndash Quart J Roy Meteor Soc 97 283ndash299
CHAMPAGNE FH CA FRIEHE JC LARUE JC WYN-GAARD 1977 Flux measurements flux estimation tech-niques and fine scale turbulence measurements in the un-stable surface layer over land ndash J Atmos Sci 34 515ndash530
COLLIER CG F DAVIES KE BOZIER AR HOLTDR MIDDLETON GN PEARSON S SIEMEN DVWILLETS GJG UPTON RI YOUNG 2005 DualDoppler lidar measurements for improving dispersionmodels ndash Bull Amer Meter Soc 86 825mdash838
DAVIES F CG COLLIERKE BOZIER GN PEARSON2003 On the accuracy of retrieved wind information fromDoppler lidar observations ndash Quart J Roy Meteor Soc129 321ndash334
DAVIES F CG COLLIER GN PEARSON KE BOZIER2004 Doppler lidar measurements of turbulent structurefunction over an urban area ndash J Atmos Oceanic Technol21 753mdash761
DAVIS JC KE BOZIER CG COLLIER F DAVIES2008 Spatial variations of sensible heat flux over an urbanarea ndash Meteor Appl 15 367mdash380
FREHLICH RG 1995 Comparison of 2- and 10- microm coher-ent Doppler lidar performance ndash J Atmos Oceanic Tech-nol 14 54ndash75
FREHLICH R L CORNMAN 2002 Estimating spatial ve-locity statistics with coherent Doppler lidar ndash J AtmosOceanic Technol 19 355ndash366
162 Davis et al Doppler lidar observations Meteorol Z 18 2009
GAL-CHEN T M XU WL EBERHARD 1992 Estima-tions of atmospheric boundary layer fluxes and other tur-bulence parameters from Doppler lidar data ndash J GeophysRes 97 D17 18409ndash18423
HARDESTY RM CJ GRUND MJ POST BJ RYEGN PEARSON 1992 Measurements of winds and cloudcharacteristics a comparison of Doppler lidar systemsPaper 2 Session TA-P ndash International Geoscience andRemote Sensing Symposium Houston Texas USA
LENSCHOW DH V WULFMEYER C SENFF 2000 Mea-suring Second- through Fourth-Order Moments in NoisyData ndash J Atmos Oceanic Technol 17 1330ndash1347
MAYOR SD SA COHN DH LENSCHOW CJGRUND RM HARDESTY 1997 Convective Bound-ary Layer Vertical Velocity Statistics Observed by 2-micromDoppler Lidar ndash 9th Conference on Coherent Laser RadarJune 23ndash27 Linkoping Sweden
MOK TM CZ RUDOWICZ 2004 A lidar study of theatmospheric entrainment zone and mixed layer over HongKong ndash Atmos Res 69 147ndash163
PEARSON G F DAVIES C COLLIER 2009 An analysisof the performance of the UFAM pulsed Doppler lidar forobserving the boundary layer ndash J Atmos Oceanic Technol2 240ndash250
ROZOFF CM WR COTTON JO ADEGOKE 2003 Sim-ulation of St Louis Missouri Land Use Impacts on Thun-
derstorms ndash J Appl Meteor 42 716ndash738
SCHMID HP 1994 Source areas for scalars and scalarfluxes ndash Bound-Layer Meteor 67 293ndash318
SCHWIESOW RL 1986 A comparative overview of activeremote-sensing techniques in Lenschow DH (Ed) Prob-ing the atmospheric boundary layer Amer Meteorol SocBoston
STULL RB 1988 An introduction to Boundary Layer Me-teorology ndash Kluwer Academic Publishers The Nether-lands
THIELEN J W WOBROCK A GADIAN P METSTAYERJ CREUTIN 2000 The possible influence of urban sur-faces on rain fall development a sensitivity study in 2D inthe meso-γ-scale Atmos Res 54 15ndash39
WOOD N AR BROWN FE HEWER 2001 Parametriz-ing the effects of orography on the boundary layer an al-ternative to effective roughness lengths ndash Quart J RoyMeteor Soc 127 759ndash777
WOOD N PJ MASON 1993 The pressure force inducedby neutral turbulent flow over hills ndash Quart J Roy MeteorSoc 119 1233ndash1267
WYNGAARD JC OR COTE 1971 The budgets of turbu-lent kinetic energy and temperature variance in the atmo-spheric surface layer ndash J Atmos Sci 28 190ndash201
156 Davis et al Doppler lidar observations Meteorol Z 18 2009
Figure 1 The Salford University Doppler lidar with antenna (inset)
Table 1 Autonomous Doppler lidar system parameters
Parameter Value
Operating Wavelength 15 microm
Pulse Repetition Frequency 20 kHz
Energy per pulse 10 microJ
Pulse duration 150 ns
Beam divergence 50 microrad
Range gate length 30 m
Lens diameter 75 mm
Focal length Infinity
Minimum range 100 m
Maximum Range 7 km
Temporal resolution 01ndash30 s
viously made with this lidar system has been given inPEARSON et al (2009) Figure 1 illustrates the Univer-sity of Salford Doppler lidar system in position at theCOPS field site in Achern Germany
The Doppler lidar data was selected to test a method-ology discussed in depth by DAVIS et al (2008) andGAL-CHEN et al (1992) for calculating QH in an ur-ban area This involves using an estimation of the ki-netic energy dissipation and the vertical gradient ofthe third moment of the vertical velocity During theCOPS field trial there were few instruments measur-ing QH and none of these were located at the Achernsite However the conditions and scanning strategy em-ployed meant that the data were suitable for testing thismethodology in a semi-rural area and since QH is animportant variable when investigating the developmentof thermals (THIELEN et al 2000) and associated rain-fall (BENNETT et al 2006 ROZOFF et al 2003) theCOPS field trial was considered an good opportunity forthis methodology to be tested
2 The Salford University HaloPhotonics Doppler lidar
The lidar system was developed to meet the require-ments of the University of Salford being capable ofunattended autonomous operation The University ofSalford specification required a Doppler lidar systemcapable of providing high temporal and spatial resolu-tion measurements of the wind velocity and backscatterwithin the atmospheric boundary layer The system isportable and rugged and capable of being used for fieldwork and long term measurements (BOZIER et al 2007PEARSON et al 2009) The lidar works by transmit-ting a short laser pulse of approximately 1x10minus7 s andcollecting the backscattered signal from the illuminatedaerosol targets along the path of the laser beam ndash see Ta-ble 1 for more details of the lidar The primary scatterersare small atmospheric aerosol particles whose diametersare within an order of magnitude of the lidar wavelengthAt optical wavelengths scattering within the lower at-mosphere is primarily by particles with diameters lessthan 3 microm which are sufficiently small to be advected bythe wind and serve as an effective tracer of wind veloc-ity (DAVIES et al 2003 HARDESTY et al 1992) FRE-LICH 1995 SCHWIESOW 1986) Accurate estimates ofthe radial component of the velocity (along the line ofsight of the laser beam) are produced as a spatial aver-age over the sensing volume of the transmitted pulse
The lidar system has a modular design arranged inthree separate units the optical base unit the weather-proof antenna consisting of the telescope and associ-ated commercially sensitive scanning electronics andthe signal processing and data acquisition unit The baseunit has approximate dimensions 56 cm x 54 cm x 18 cmand contains the optical source interferometer receiverand the electronics The weather-proof antenna is at-tached to the base unit via a 1rdquo diameter 1 m long opticalfibre conduit The antenna can be deployed permanently
Meteorol Z 18 2009 Davis et al Doppler lidar observations 157
Figure 2 Wind vector plot for 150707 0800ndash2200 h UTC arrows
represent relative horizontal wind speed and direction Due to some
wind velocities being very small the arrows appear as dots The
legend (bottom left) shows an arrow representing a 2 m sminus1 wind
speed with westerly direction
outside whilst the base unit and data acquisition sys-tem are housed within a mobile laboratory environmentThe signal processing has been developed with a viewto providing a high level of flexibility with respect to thedata acquisition parameters Users are able to set param-eters such as the range gate length maximum range andnumber of pulses accumulated for each measurementas detailed in Table 1 During the field campaign mea-surements were limited only by atmospheric moisture(rainfog) a lightning strike and a mains power failureThe Doppler lidar system is described in greater detailin PEARSON et al (2009) HALO PHOTONICS (20082)
3 Measurements
The UFAM pulsed Doppler lidar is capable of measur-ing
bull Directly Radial wind velocity relative backscatterintensity
bull Indirectly Horizontal velocities their variances andcovariances atmospheric backscatter coefficient (β)and turbulence kinetic energy dissipation rate (ε)
During much of the field campaign the lidar systemperformed a series of scan patterns pre-programmedinto the system software Errors are given in Table 1and described in more detail in PEARSON et al (2009)Generally the system carried out
bull 5 minute azimuth scan at 5 different elevations (2030 40 45 and 60) but on the case discussedhere azimuth scans were performed at 60
2The Halo Photonics pulsed Doppler LiDAR system available at wwwhalo-
photonicscom
Figure 3 Full day view 150707 showing thermals building
throughout the day
Figure 4 Doppler lidar and radiosonde horizontal wind speeds to
800 m
bull 25 minute vertically fixed stare consisting of 334radial velocity lsquorayrsquo measurements
The azimuth scan was performed to determine windvelocity profiles derived from VAD analysis (BROWN-ING and WEXLER 1968) and these data were concate-nated enabling a half-hourly vector profile as illustratedin Figure 2 The vertically pointing stares were concate-nated to produce a complete overview of vertical veloc-ities measured by the Doppler lidar each day as illus-trated in Figure 3 Both Figure 2 and Figure 3 showdata up to a height of 1100 m Beyond this range onthis day there was no signal return The data illustratedcan be used to calculate a variety of products includ-ing mixed layer height (COLLIER et al 2005 MOK
and RUDOWICZ 2004) and turbulence statistics DAVIES
et al (2004) LENSCHOW et al (2000) FRELICH andCORNMAN (2002) and BANAKH et al (1999)
The data from the 15th July 2007 show that thermalswere building throughout the day and the synoptic situa-
158 Davis et al Doppler lidar observations Meteorol Z 18 2009
tion indicated that a convergence line passed from westto east to the north of Achern High convectivity partic-ularly at around 1500 UTC was measured by the lidarThe convective conditions suggested a suitable day touse lidar data to calculate (QH ) a parameter importantin monitoring the development of thermals A methoddiscussed in detail by DAVIS et al (2008) is outlinedand tested here
It was first necessary to confirm the reliability ofthe Doppler lidar data Comparisons were made be-tween horizontal velocity data measured by the lidar ra-diosonde and a 1290 MHz wind profiler and are shownhere
Figure 4 illustrates similarities between horizontal ve-locities measured by Doppler lidar and radiosonde theradiosonde was launched at 1000 UTC and the azimuthscan was performed from 1000 UTC to 1005 UTCWhen this data was taken it is considered that the bound-ary layer top was at around 800 m and the lidar data be-comes noisy and therefore un-useable above this level
Another comparison was made between the UFAMDoppler lidar data and the University of Manchester op-erated UFAM funded 1290 MHz wind profiler Thisis illustrated in Figure 5 As with the radiosonde data(Figure 4) it can be seen that there are similarities be-tween the two datasets There are some areas of smallscale differences but generally the two remote sensinginstruments are providing comparable data Having es-tablished that the Doppler lidar is a suitable instrumentto make atmospheric measurements and that the ambi-ent conditions were appropriate more complex calcula-tions could be performed to investigate other parameterssuch as QH
The chosen method assumes unstable conditions so toconfirm whether this was the situation for the 15th July2007 case potential temperature profiles were derivedfrom two radiosonde ascents launched from the Achernsite These profiles suggesting unstable conditions areshown in Figure 6
4 Methodology
Having established unstable conditions it was possibleto calculate QH from the vertical velocity-potential tem-
perature covariance term (wprimeθprime) in (41) following theprocedure outlined in GAL-CHEN et al (1992)
g
θ0wprimeθprime =
1
ρ0wprime
partpprime
partz+
ε
3+
part
partz
(
1
2wprime3
)
(41)
WYNGAARD and COTE (1971) and DAVIS et al
(2008) found that the pressure term partpprime
partz was small com-pared with the other terms of (42) for unstable condi-tions and will therefore be ignored in the subsequentanalysis which can be applied only to cases of convec-tion hence
wprimeθprime asympθ0
g
[
part
partz
(
1
2wprime3
)
+ε
3
]
(42)
Figure 5 Vector plot from UFAM 1290 MHz wind profiler showing
similar features to the lidar (Figure 2) with easterlies early in the
morning becoming southerly as the morning progresses and veering
westerly in the afternoon
Figure 6 Potential temperature profiles on 150707 suggesting
unstable conditions
The mean vertical velocity w is calculated for eachrange gate and averaged over the 25 minute scan periodThis is done for each of the nineteen 30 m-long rangegates from 135 m to 705 m using data from a verti-cally pointing stare scan The wrsquo is the deviation fromthe mean vertical velocity Then wrsquo3 is calculated foreach point and averaged over the 25 minute scan dura-
tion to form a profile of wprime3
In order to use (42) to calculate wprimeθprime it is also neces-sary to estimate ε One way of estimating ε is to exam-ine the line spectra of the longitudinal velocity correla-tion In the inertial subrange the expected relationshipis (BATCHELOR 1967)
f(κ) = αε23κminus53 (43)
where κ is the wave number α is a universal constant(05) and f(κ) is the Fourier transform of the longitudi-nal velocity correlation A -53 slope indicates the pres-
Meteorol Z 18 2009 Davis et al Doppler lidar observations 159
Table 2 Errors associated with Doppler lidar calculations of ε and wprime2
Error type and extent
Vertical velocity variancewprime2 Bias = 002 m2 sminus2 PEARSON et al 2008
Random = plusmn068 m2 sminus2
Turbulent kinetic energy Error = 10 FRELICH and CORNMAN 2002
dissipation rate ε
Figure 7 Spectrum from the GAL-CHEN et al (1992) spatial
method A -53 gradient indicates the presence of the inertial sub-
range
Figure 8 Temporal spectrum as reported by CHAMPAGNE et al
(1977) The -23 slope highlights the presence of the inertial sub-
range
ence of an inertial subrange and is shown in Figure 7This will be referred to henceforth as the spatial spec-tra method We calculate the spectra for each ray from135 to 705 m and average the spectra over the 25 minutetime period
An alternative method for calculating spectra is pro-posed by CHAMPAGNE et al (1977) Here Taylorrsquos Hy-
pothesis (STULL 1988) is assumed and is used to cal-culate the wave number The u is calculated from the li-dar data using VAD analysis The u is the average meanwind over the 135ndash705 m height range The spectrumcan be calculated from
nS (n) = 068ε23
(
2πn
u
)minus23
(44)
where S(n) is the spectral energy of frequency n andu is the mean wind speed Here the inertial subrangeis characterized by a ndash23 gradient and is shown in Fig-ure 8 This will be referred to henceforth as the temporalspectra method
When performing spectral analysis to obtain ε thedata were detrended to remove any nonstationary be-haviour However errors are still expected and these arelisted in Table 2
With unstable conditions QH can thus be calculatedfrom
QH = ρ0CP wprimeθprime (45)
taking CP the heat capacity of air at constant pressureas 1004 J kminus1 kgminus1 Kminus1 ρ0 the air density as 1275kg mminus3 and where is the potential temperature-verticalvelocity covariance calculated in (42) (GAL-CHEN etal 1992) (43) also assumes unstable conditions whichwere present on this day as illustrated in Figure 6
5 Results
The Salford Autonomous Doppler Lidar System was de-ployed for COPS in Achern Germany where it col-lected data almost continuously for 3 months A casestudy of convective development is described here andthe results include turbulence spectra and some time-series of ε and QH Figure 7 shows a spectrum producedfrom a vertically-pointing stare taken at 1100 UTC onthe 15th July 2007 in order to obtain a single value of εThe spectrum in Figure 7 was calculated using the spa-tial method presented by GAL-CHEN et al (1992) Inaddition in order to confirm the results shown in Fig-ure 7 spectra were also plotted using the same datasetand the method proposed by CHAMPAGNE et al (1977)as shown in Figure 8 The points deviating from the ex-pected ndash23 gradient at high wavenumbers are consid-ered to be due to noise (MAYOR et al 1997)
160 Davis et al Doppler lidar observations Meteorol Z 18 2009
Figure 9 Eddy dissipation rates calculated from Doppler lidar data
using both spatial and temporal methods on 150707
Figure 10 QH timeseries calculated for 150707 showing an esti-
mation of the expected errors Half-hourly averages of QH as mea-
sured by the IMK-FZK energy balance station for that day are also
shown
The time series in Figure 9 illustrates ε derived fromboth temporal and spatial methods developing through-out the day as would be expected ndash increasing towardsnoon peaking at 1500 UTC and then decreasing towardsthe evening as convection is developing in the morn-ing and then dissipating in the evening (STULL 1988)With the estimates for ε and vertical velocity fluctua-tions (wprime3) QH was calculated as illustrated in Figure10 It was decided that although the ε values were sim-ilar for both spectra the values derived from the spatialspectra would be utilised to calculate QH as per GAL-CHEN et al (1992) Figure 10 also shows a timeseriesof QH measured using an Energy Balance Station atthe COPS operation centre Baden Airpark 10 km northof Achern where the Doppler lidar was located but stillin the Rhine Valley
Figure 11 QH from Doppler lidar data and WRF model data
When considering the results achieved using the datacollected with the Doppler lidar certain factors need tobe considered for example how did the results com-pare with other measurements Figure 10 shows the li-dar data plotted against data collected from an energybalance station located at Baden Airpark 10 km to thenorth of the Doppler lidar The energy balance stationmade point measurements of QH at a single point intime at the surface and averaged over half an hour Themeasurements made by the lidar are averages over thedepth of the boundary layer up to 705 m and while show-ing a general trend cannot be the same for the followingreasons
bull The two instruments were not collocated
bull The lidar as a remote sensor provides a volumeaverage over a pencil-shaped beam over a range of120 m to 705 m above the surface
The QH values measured by the lidar are higher thanthose measured at Baden Airpark It is considered thatthe relatively high values of QH calculated from thelidar data compared with those from the energy bal-ance station result from the nature of the instrumentsand their associated lsquofootprintsrsquo (SCHMID 1994) Li-dar measurements are averaged over the lowest sim700m of the boundary layer and so have larger footprintsthan ground based instruments in the order of severalkilometres WOOD and MASON (1993) BELCHER andWOOD (1996) and WOOD et al (2001) discuss the ef-fects of increasing form drag and roughness length whenflow is directed across orography such as the mountainsto the east of the Achern site Since the mountains aresome 8 km away within the fetch of the lidar it is con-sidered that it the high eddy dissipation rates are due tothis orography These high values greatly influence themagnitude of the QH The instruments at Baden Air-park are at surface level so have lower fetch and are un-likely to be affected by the orography It should also be
Meteorol Z 18 2009 Davis et al Doppler lidar observations 161
noted that on this occasion the synoptic situation indi-cated that the convergence line mentioned in section 3was more active to the north of Achern
Despite these differences the results are comparablehaving a reasonable order of magnitude and showing adiurnal trend as expected
Output from the NERC WRF model runs for this dayis shown in Figure 11 These data also show a diurnaltrend similar to those shown in Figure 10 It is not real-istic to expect remotely sensed data to replicate the re-sults from a computer model but as with the surface val-ues the results are within the same order of magnitudethere is a similar trend in the model output with valuesrising towards noon and dropping towards sunset It isinteresting to note that where the model output suggestshigh heat flux values at 0800 UTC the lidar does notAlthough the measured QH values were comparativelyhigh at around 1500 UTC these values were unrealisti-cally so It is thought that this may be due to a mesoscaleevent that occurred at about 1500 UTC that day as men-tioned previously Investigations into this are ongoing
6 Conclusions and future work
It has been shown that by using a method outlined byDAVIS et al (2008) GAL-CHEN et al (1992) it is pos-sible to calculate values of QH and it is considered thatthese values fall within an acceptable range It is notedthat the diurnal cycle of the lidar-derived QH is closer tothe diurnal cycle of ε whereas both the model and sur-face measurements of QH peak earlier in the day QH
calculated from Doppler lidar data using the method de-scribed here would be expected to differ from that cal-culated using a surface based lsquopointrsquo sensor such as anenergy balance instrument or a sonic anemometer sincethe data collected with a Doppler lidar is integrated overa lsquovolumersquo of atmosphere aloft advecting through thelaser beam However these measurements are a usefuladdition to the other measurements provided by the li-dar and require no extra equipment
It is considered that further investigations of employ-ing this method to calculate QH are needed to establishthe overall applicability of the method and build on thepreliminary calculations It is expected that such inves-tigations will include
bull Calculation of ε using alternative methods
bull investigation of ε at different heights and atmo-spheric conditions
bull verification of Doppler lidar estimations of ε and wprime3and
bull investigation over different surface types
Acknowledgments
The authors would like to acknowledge Dr N KALT-HOFF at IMK FZK Karlsruhe for supplying QH datafrom their energy balance station located at Baden Air-field Germany collected as part of COPS July 2007Also Dr A GADIAN from the University of Leeds forhis help in securing the WRF model output Dr E NOR-TON from the University of Manchester for making theUFAM 1290 MHz wind profiler quicklook plots avail-able online HP SCHMID for the use of the flux source-area model (FSAM) finally UK COPS for fundingthrough NERC
References
BANAKH VA IN SMALIKHO F KOPP C WERNER1999 Measurements of turbulent energy dissipation ratewith a CW Doppler lidar in the atmospheric boundarylayer J Atmos Oceanic Technol 16 1044 mdash1061
BATCHELOR GK 1967 The theory of homogeneous tur-bulence 5th ed ndash Cambridge University Press London
BELCHER SE A WOOD 1996 Form and wave drag dueto stably stratified turbulent flow over low ridges ndash QuartJ Roy Meteor Soc 122 863mdash902
BENNETT LJ KA BROWNING AM BLYTH DJPARKER PA CLARK 2006 A review of the initiation ofprecipitating convection in the United Kingdom ndash QuartJ Roy Meteor Soc 132 1001mdash1020
BOZIER KE GN PEARSON CG COLLIER 2007Doppler lidar observations of Russian forest fire plumesover Helsinki ndash Weather 62 203ndash208
BROWNING KA R WEXLER 1968 The Determinationof Kinematic Properties of a Wind Field using DopplerRadar ndash Quart J Roy Meteor Soc 97 283ndash299
CHAMPAGNE FH CA FRIEHE JC LARUE JC WYN-GAARD 1977 Flux measurements flux estimation tech-niques and fine scale turbulence measurements in the un-stable surface layer over land ndash J Atmos Sci 34 515ndash530
COLLIER CG F DAVIES KE BOZIER AR HOLTDR MIDDLETON GN PEARSON S SIEMEN DVWILLETS GJG UPTON RI YOUNG 2005 DualDoppler lidar measurements for improving dispersionmodels ndash Bull Amer Meter Soc 86 825mdash838
DAVIES F CG COLLIERKE BOZIER GN PEARSON2003 On the accuracy of retrieved wind information fromDoppler lidar observations ndash Quart J Roy Meteor Soc129 321ndash334
DAVIES F CG COLLIER GN PEARSON KE BOZIER2004 Doppler lidar measurements of turbulent structurefunction over an urban area ndash J Atmos Oceanic Technol21 753mdash761
DAVIS JC KE BOZIER CG COLLIER F DAVIES2008 Spatial variations of sensible heat flux over an urbanarea ndash Meteor Appl 15 367mdash380
FREHLICH RG 1995 Comparison of 2- and 10- microm coher-ent Doppler lidar performance ndash J Atmos Oceanic Tech-nol 14 54ndash75
FREHLICH R L CORNMAN 2002 Estimating spatial ve-locity statistics with coherent Doppler lidar ndash J AtmosOceanic Technol 19 355ndash366
162 Davis et al Doppler lidar observations Meteorol Z 18 2009
GAL-CHEN T M XU WL EBERHARD 1992 Estima-tions of atmospheric boundary layer fluxes and other tur-bulence parameters from Doppler lidar data ndash J GeophysRes 97 D17 18409ndash18423
HARDESTY RM CJ GRUND MJ POST BJ RYEGN PEARSON 1992 Measurements of winds and cloudcharacteristics a comparison of Doppler lidar systemsPaper 2 Session TA-P ndash International Geoscience andRemote Sensing Symposium Houston Texas USA
LENSCHOW DH V WULFMEYER C SENFF 2000 Mea-suring Second- through Fourth-Order Moments in NoisyData ndash J Atmos Oceanic Technol 17 1330ndash1347
MAYOR SD SA COHN DH LENSCHOW CJGRUND RM HARDESTY 1997 Convective Bound-ary Layer Vertical Velocity Statistics Observed by 2-micromDoppler Lidar ndash 9th Conference on Coherent Laser RadarJune 23ndash27 Linkoping Sweden
MOK TM CZ RUDOWICZ 2004 A lidar study of theatmospheric entrainment zone and mixed layer over HongKong ndash Atmos Res 69 147ndash163
PEARSON G F DAVIES C COLLIER 2009 An analysisof the performance of the UFAM pulsed Doppler lidar forobserving the boundary layer ndash J Atmos Oceanic Technol2 240ndash250
ROZOFF CM WR COTTON JO ADEGOKE 2003 Sim-ulation of St Louis Missouri Land Use Impacts on Thun-
derstorms ndash J Appl Meteor 42 716ndash738
SCHMID HP 1994 Source areas for scalars and scalarfluxes ndash Bound-Layer Meteor 67 293ndash318
SCHWIESOW RL 1986 A comparative overview of activeremote-sensing techniques in Lenschow DH (Ed) Prob-ing the atmospheric boundary layer Amer Meteorol SocBoston
STULL RB 1988 An introduction to Boundary Layer Me-teorology ndash Kluwer Academic Publishers The Nether-lands
THIELEN J W WOBROCK A GADIAN P METSTAYERJ CREUTIN 2000 The possible influence of urban sur-faces on rain fall development a sensitivity study in 2D inthe meso-γ-scale Atmos Res 54 15ndash39
WOOD N AR BROWN FE HEWER 2001 Parametriz-ing the effects of orography on the boundary layer an al-ternative to effective roughness lengths ndash Quart J RoyMeteor Soc 127 759ndash777
WOOD N PJ MASON 1993 The pressure force inducedby neutral turbulent flow over hills ndash Quart J Roy MeteorSoc 119 1233ndash1267
WYNGAARD JC OR COTE 1971 The budgets of turbu-lent kinetic energy and temperature variance in the atmo-spheric surface layer ndash J Atmos Sci 28 190ndash201
Meteorol Z 18 2009 Davis et al Doppler lidar observations 157
Figure 2 Wind vector plot for 150707 0800ndash2200 h UTC arrows
represent relative horizontal wind speed and direction Due to some
wind velocities being very small the arrows appear as dots The
legend (bottom left) shows an arrow representing a 2 m sminus1 wind
speed with westerly direction
outside whilst the base unit and data acquisition sys-tem are housed within a mobile laboratory environmentThe signal processing has been developed with a viewto providing a high level of flexibility with respect to thedata acquisition parameters Users are able to set param-eters such as the range gate length maximum range andnumber of pulses accumulated for each measurementas detailed in Table 1 During the field campaign mea-surements were limited only by atmospheric moisture(rainfog) a lightning strike and a mains power failureThe Doppler lidar system is described in greater detailin PEARSON et al (2009) HALO PHOTONICS (20082)
3 Measurements
The UFAM pulsed Doppler lidar is capable of measur-ing
bull Directly Radial wind velocity relative backscatterintensity
bull Indirectly Horizontal velocities their variances andcovariances atmospheric backscatter coefficient (β)and turbulence kinetic energy dissipation rate (ε)
During much of the field campaign the lidar systemperformed a series of scan patterns pre-programmedinto the system software Errors are given in Table 1and described in more detail in PEARSON et al (2009)Generally the system carried out
bull 5 minute azimuth scan at 5 different elevations (2030 40 45 and 60) but on the case discussedhere azimuth scans were performed at 60
2The Halo Photonics pulsed Doppler LiDAR system available at wwwhalo-
photonicscom
Figure 3 Full day view 150707 showing thermals building
throughout the day
Figure 4 Doppler lidar and radiosonde horizontal wind speeds to
800 m
bull 25 minute vertically fixed stare consisting of 334radial velocity lsquorayrsquo measurements
The azimuth scan was performed to determine windvelocity profiles derived from VAD analysis (BROWN-ING and WEXLER 1968) and these data were concate-nated enabling a half-hourly vector profile as illustratedin Figure 2 The vertically pointing stares were concate-nated to produce a complete overview of vertical veloc-ities measured by the Doppler lidar each day as illus-trated in Figure 3 Both Figure 2 and Figure 3 showdata up to a height of 1100 m Beyond this range onthis day there was no signal return The data illustratedcan be used to calculate a variety of products includ-ing mixed layer height (COLLIER et al 2005 MOK
and RUDOWICZ 2004) and turbulence statistics DAVIES
et al (2004) LENSCHOW et al (2000) FRELICH andCORNMAN (2002) and BANAKH et al (1999)
The data from the 15th July 2007 show that thermalswere building throughout the day and the synoptic situa-
158 Davis et al Doppler lidar observations Meteorol Z 18 2009
tion indicated that a convergence line passed from westto east to the north of Achern High convectivity partic-ularly at around 1500 UTC was measured by the lidarThe convective conditions suggested a suitable day touse lidar data to calculate (QH ) a parameter importantin monitoring the development of thermals A methoddiscussed in detail by DAVIS et al (2008) is outlinedand tested here
It was first necessary to confirm the reliability ofthe Doppler lidar data Comparisons were made be-tween horizontal velocity data measured by the lidar ra-diosonde and a 1290 MHz wind profiler and are shownhere
Figure 4 illustrates similarities between horizontal ve-locities measured by Doppler lidar and radiosonde theradiosonde was launched at 1000 UTC and the azimuthscan was performed from 1000 UTC to 1005 UTCWhen this data was taken it is considered that the bound-ary layer top was at around 800 m and the lidar data be-comes noisy and therefore un-useable above this level
Another comparison was made between the UFAMDoppler lidar data and the University of Manchester op-erated UFAM funded 1290 MHz wind profiler Thisis illustrated in Figure 5 As with the radiosonde data(Figure 4) it can be seen that there are similarities be-tween the two datasets There are some areas of smallscale differences but generally the two remote sensinginstruments are providing comparable data Having es-tablished that the Doppler lidar is a suitable instrumentto make atmospheric measurements and that the ambi-ent conditions were appropriate more complex calcula-tions could be performed to investigate other parameterssuch as QH
The chosen method assumes unstable conditions so toconfirm whether this was the situation for the 15th July2007 case potential temperature profiles were derivedfrom two radiosonde ascents launched from the Achernsite These profiles suggesting unstable conditions areshown in Figure 6
4 Methodology
Having established unstable conditions it was possibleto calculate QH from the vertical velocity-potential tem-
perature covariance term (wprimeθprime) in (41) following theprocedure outlined in GAL-CHEN et al (1992)
g
θ0wprimeθprime =
1
ρ0wprime
partpprime
partz+
ε
3+
part
partz
(
1
2wprime3
)
(41)
WYNGAARD and COTE (1971) and DAVIS et al
(2008) found that the pressure term partpprime
partz was small com-pared with the other terms of (42) for unstable condi-tions and will therefore be ignored in the subsequentanalysis which can be applied only to cases of convec-tion hence
wprimeθprime asympθ0
g
[
part
partz
(
1
2wprime3
)
+ε
3
]
(42)
Figure 5 Vector plot from UFAM 1290 MHz wind profiler showing
similar features to the lidar (Figure 2) with easterlies early in the
morning becoming southerly as the morning progresses and veering
westerly in the afternoon
Figure 6 Potential temperature profiles on 150707 suggesting
unstable conditions
The mean vertical velocity w is calculated for eachrange gate and averaged over the 25 minute scan periodThis is done for each of the nineteen 30 m-long rangegates from 135 m to 705 m using data from a verti-cally pointing stare scan The wrsquo is the deviation fromthe mean vertical velocity Then wrsquo3 is calculated foreach point and averaged over the 25 minute scan dura-
tion to form a profile of wprime3
In order to use (42) to calculate wprimeθprime it is also neces-sary to estimate ε One way of estimating ε is to exam-ine the line spectra of the longitudinal velocity correla-tion In the inertial subrange the expected relationshipis (BATCHELOR 1967)
f(κ) = αε23κminus53 (43)
where κ is the wave number α is a universal constant(05) and f(κ) is the Fourier transform of the longitudi-nal velocity correlation A -53 slope indicates the pres-
Meteorol Z 18 2009 Davis et al Doppler lidar observations 159
Table 2 Errors associated with Doppler lidar calculations of ε and wprime2
Error type and extent
Vertical velocity variancewprime2 Bias = 002 m2 sminus2 PEARSON et al 2008
Random = plusmn068 m2 sminus2
Turbulent kinetic energy Error = 10 FRELICH and CORNMAN 2002
dissipation rate ε
Figure 7 Spectrum from the GAL-CHEN et al (1992) spatial
method A -53 gradient indicates the presence of the inertial sub-
range
Figure 8 Temporal spectrum as reported by CHAMPAGNE et al
(1977) The -23 slope highlights the presence of the inertial sub-
range
ence of an inertial subrange and is shown in Figure 7This will be referred to henceforth as the spatial spec-tra method We calculate the spectra for each ray from135 to 705 m and average the spectra over the 25 minutetime period
An alternative method for calculating spectra is pro-posed by CHAMPAGNE et al (1977) Here Taylorrsquos Hy-
pothesis (STULL 1988) is assumed and is used to cal-culate the wave number The u is calculated from the li-dar data using VAD analysis The u is the average meanwind over the 135ndash705 m height range The spectrumcan be calculated from
nS (n) = 068ε23
(
2πn
u
)minus23
(44)
where S(n) is the spectral energy of frequency n andu is the mean wind speed Here the inertial subrangeis characterized by a ndash23 gradient and is shown in Fig-ure 8 This will be referred to henceforth as the temporalspectra method
When performing spectral analysis to obtain ε thedata were detrended to remove any nonstationary be-haviour However errors are still expected and these arelisted in Table 2
With unstable conditions QH can thus be calculatedfrom
QH = ρ0CP wprimeθprime (45)
taking CP the heat capacity of air at constant pressureas 1004 J kminus1 kgminus1 Kminus1 ρ0 the air density as 1275kg mminus3 and where is the potential temperature-verticalvelocity covariance calculated in (42) (GAL-CHEN etal 1992) (43) also assumes unstable conditions whichwere present on this day as illustrated in Figure 6
5 Results
The Salford Autonomous Doppler Lidar System was de-ployed for COPS in Achern Germany where it col-lected data almost continuously for 3 months A casestudy of convective development is described here andthe results include turbulence spectra and some time-series of ε and QH Figure 7 shows a spectrum producedfrom a vertically-pointing stare taken at 1100 UTC onthe 15th July 2007 in order to obtain a single value of εThe spectrum in Figure 7 was calculated using the spa-tial method presented by GAL-CHEN et al (1992) Inaddition in order to confirm the results shown in Fig-ure 7 spectra were also plotted using the same datasetand the method proposed by CHAMPAGNE et al (1977)as shown in Figure 8 The points deviating from the ex-pected ndash23 gradient at high wavenumbers are consid-ered to be due to noise (MAYOR et al 1997)
160 Davis et al Doppler lidar observations Meteorol Z 18 2009
Figure 9 Eddy dissipation rates calculated from Doppler lidar data
using both spatial and temporal methods on 150707
Figure 10 QH timeseries calculated for 150707 showing an esti-
mation of the expected errors Half-hourly averages of QH as mea-
sured by the IMK-FZK energy balance station for that day are also
shown
The time series in Figure 9 illustrates ε derived fromboth temporal and spatial methods developing through-out the day as would be expected ndash increasing towardsnoon peaking at 1500 UTC and then decreasing towardsthe evening as convection is developing in the morn-ing and then dissipating in the evening (STULL 1988)With the estimates for ε and vertical velocity fluctua-tions (wprime3) QH was calculated as illustrated in Figure10 It was decided that although the ε values were sim-ilar for both spectra the values derived from the spatialspectra would be utilised to calculate QH as per GAL-CHEN et al (1992) Figure 10 also shows a timeseriesof QH measured using an Energy Balance Station atthe COPS operation centre Baden Airpark 10 km northof Achern where the Doppler lidar was located but stillin the Rhine Valley
Figure 11 QH from Doppler lidar data and WRF model data
When considering the results achieved using the datacollected with the Doppler lidar certain factors need tobe considered for example how did the results com-pare with other measurements Figure 10 shows the li-dar data plotted against data collected from an energybalance station located at Baden Airpark 10 km to thenorth of the Doppler lidar The energy balance stationmade point measurements of QH at a single point intime at the surface and averaged over half an hour Themeasurements made by the lidar are averages over thedepth of the boundary layer up to 705 m and while show-ing a general trend cannot be the same for the followingreasons
bull The two instruments were not collocated
bull The lidar as a remote sensor provides a volumeaverage over a pencil-shaped beam over a range of120 m to 705 m above the surface
The QH values measured by the lidar are higher thanthose measured at Baden Airpark It is considered thatthe relatively high values of QH calculated from thelidar data compared with those from the energy bal-ance station result from the nature of the instrumentsand their associated lsquofootprintsrsquo (SCHMID 1994) Li-dar measurements are averaged over the lowest sim700m of the boundary layer and so have larger footprintsthan ground based instruments in the order of severalkilometres WOOD and MASON (1993) BELCHER andWOOD (1996) and WOOD et al (2001) discuss the ef-fects of increasing form drag and roughness length whenflow is directed across orography such as the mountainsto the east of the Achern site Since the mountains aresome 8 km away within the fetch of the lidar it is con-sidered that it the high eddy dissipation rates are due tothis orography These high values greatly influence themagnitude of the QH The instruments at Baden Air-park are at surface level so have lower fetch and are un-likely to be affected by the orography It should also be
Meteorol Z 18 2009 Davis et al Doppler lidar observations 161
noted that on this occasion the synoptic situation indi-cated that the convergence line mentioned in section 3was more active to the north of Achern
Despite these differences the results are comparablehaving a reasonable order of magnitude and showing adiurnal trend as expected
Output from the NERC WRF model runs for this dayis shown in Figure 11 These data also show a diurnaltrend similar to those shown in Figure 10 It is not real-istic to expect remotely sensed data to replicate the re-sults from a computer model but as with the surface val-ues the results are within the same order of magnitudethere is a similar trend in the model output with valuesrising towards noon and dropping towards sunset It isinteresting to note that where the model output suggestshigh heat flux values at 0800 UTC the lidar does notAlthough the measured QH values were comparativelyhigh at around 1500 UTC these values were unrealisti-cally so It is thought that this may be due to a mesoscaleevent that occurred at about 1500 UTC that day as men-tioned previously Investigations into this are ongoing
6 Conclusions and future work
It has been shown that by using a method outlined byDAVIS et al (2008) GAL-CHEN et al (1992) it is pos-sible to calculate values of QH and it is considered thatthese values fall within an acceptable range It is notedthat the diurnal cycle of the lidar-derived QH is closer tothe diurnal cycle of ε whereas both the model and sur-face measurements of QH peak earlier in the day QH
calculated from Doppler lidar data using the method de-scribed here would be expected to differ from that cal-culated using a surface based lsquopointrsquo sensor such as anenergy balance instrument or a sonic anemometer sincethe data collected with a Doppler lidar is integrated overa lsquovolumersquo of atmosphere aloft advecting through thelaser beam However these measurements are a usefuladdition to the other measurements provided by the li-dar and require no extra equipment
It is considered that further investigations of employ-ing this method to calculate QH are needed to establishthe overall applicability of the method and build on thepreliminary calculations It is expected that such inves-tigations will include
bull Calculation of ε using alternative methods
bull investigation of ε at different heights and atmo-spheric conditions
bull verification of Doppler lidar estimations of ε and wprime3and
bull investigation over different surface types
Acknowledgments
The authors would like to acknowledge Dr N KALT-HOFF at IMK FZK Karlsruhe for supplying QH datafrom their energy balance station located at Baden Air-field Germany collected as part of COPS July 2007Also Dr A GADIAN from the University of Leeds forhis help in securing the WRF model output Dr E NOR-TON from the University of Manchester for making theUFAM 1290 MHz wind profiler quicklook plots avail-able online HP SCHMID for the use of the flux source-area model (FSAM) finally UK COPS for fundingthrough NERC
References
BANAKH VA IN SMALIKHO F KOPP C WERNER1999 Measurements of turbulent energy dissipation ratewith a CW Doppler lidar in the atmospheric boundarylayer J Atmos Oceanic Technol 16 1044 mdash1061
BATCHELOR GK 1967 The theory of homogeneous tur-bulence 5th ed ndash Cambridge University Press London
BELCHER SE A WOOD 1996 Form and wave drag dueto stably stratified turbulent flow over low ridges ndash QuartJ Roy Meteor Soc 122 863mdash902
BENNETT LJ KA BROWNING AM BLYTH DJPARKER PA CLARK 2006 A review of the initiation ofprecipitating convection in the United Kingdom ndash QuartJ Roy Meteor Soc 132 1001mdash1020
BOZIER KE GN PEARSON CG COLLIER 2007Doppler lidar observations of Russian forest fire plumesover Helsinki ndash Weather 62 203ndash208
BROWNING KA R WEXLER 1968 The Determinationof Kinematic Properties of a Wind Field using DopplerRadar ndash Quart J Roy Meteor Soc 97 283ndash299
CHAMPAGNE FH CA FRIEHE JC LARUE JC WYN-GAARD 1977 Flux measurements flux estimation tech-niques and fine scale turbulence measurements in the un-stable surface layer over land ndash J Atmos Sci 34 515ndash530
COLLIER CG F DAVIES KE BOZIER AR HOLTDR MIDDLETON GN PEARSON S SIEMEN DVWILLETS GJG UPTON RI YOUNG 2005 DualDoppler lidar measurements for improving dispersionmodels ndash Bull Amer Meter Soc 86 825mdash838
DAVIES F CG COLLIERKE BOZIER GN PEARSON2003 On the accuracy of retrieved wind information fromDoppler lidar observations ndash Quart J Roy Meteor Soc129 321ndash334
DAVIES F CG COLLIER GN PEARSON KE BOZIER2004 Doppler lidar measurements of turbulent structurefunction over an urban area ndash J Atmos Oceanic Technol21 753mdash761
DAVIS JC KE BOZIER CG COLLIER F DAVIES2008 Spatial variations of sensible heat flux over an urbanarea ndash Meteor Appl 15 367mdash380
FREHLICH RG 1995 Comparison of 2- and 10- microm coher-ent Doppler lidar performance ndash J Atmos Oceanic Tech-nol 14 54ndash75
FREHLICH R L CORNMAN 2002 Estimating spatial ve-locity statistics with coherent Doppler lidar ndash J AtmosOceanic Technol 19 355ndash366
162 Davis et al Doppler lidar observations Meteorol Z 18 2009
GAL-CHEN T M XU WL EBERHARD 1992 Estima-tions of atmospheric boundary layer fluxes and other tur-bulence parameters from Doppler lidar data ndash J GeophysRes 97 D17 18409ndash18423
HARDESTY RM CJ GRUND MJ POST BJ RYEGN PEARSON 1992 Measurements of winds and cloudcharacteristics a comparison of Doppler lidar systemsPaper 2 Session TA-P ndash International Geoscience andRemote Sensing Symposium Houston Texas USA
LENSCHOW DH V WULFMEYER C SENFF 2000 Mea-suring Second- through Fourth-Order Moments in NoisyData ndash J Atmos Oceanic Technol 17 1330ndash1347
MAYOR SD SA COHN DH LENSCHOW CJGRUND RM HARDESTY 1997 Convective Bound-ary Layer Vertical Velocity Statistics Observed by 2-micromDoppler Lidar ndash 9th Conference on Coherent Laser RadarJune 23ndash27 Linkoping Sweden
MOK TM CZ RUDOWICZ 2004 A lidar study of theatmospheric entrainment zone and mixed layer over HongKong ndash Atmos Res 69 147ndash163
PEARSON G F DAVIES C COLLIER 2009 An analysisof the performance of the UFAM pulsed Doppler lidar forobserving the boundary layer ndash J Atmos Oceanic Technol2 240ndash250
ROZOFF CM WR COTTON JO ADEGOKE 2003 Sim-ulation of St Louis Missouri Land Use Impacts on Thun-
derstorms ndash J Appl Meteor 42 716ndash738
SCHMID HP 1994 Source areas for scalars and scalarfluxes ndash Bound-Layer Meteor 67 293ndash318
SCHWIESOW RL 1986 A comparative overview of activeremote-sensing techniques in Lenschow DH (Ed) Prob-ing the atmospheric boundary layer Amer Meteorol SocBoston
STULL RB 1988 An introduction to Boundary Layer Me-teorology ndash Kluwer Academic Publishers The Nether-lands
THIELEN J W WOBROCK A GADIAN P METSTAYERJ CREUTIN 2000 The possible influence of urban sur-faces on rain fall development a sensitivity study in 2D inthe meso-γ-scale Atmos Res 54 15ndash39
WOOD N AR BROWN FE HEWER 2001 Parametriz-ing the effects of orography on the boundary layer an al-ternative to effective roughness lengths ndash Quart J RoyMeteor Soc 127 759ndash777
WOOD N PJ MASON 1993 The pressure force inducedby neutral turbulent flow over hills ndash Quart J Roy MeteorSoc 119 1233ndash1267
WYNGAARD JC OR COTE 1971 The budgets of turbu-lent kinetic energy and temperature variance in the atmo-spheric surface layer ndash J Atmos Sci 28 190ndash201
158 Davis et al Doppler lidar observations Meteorol Z 18 2009
tion indicated that a convergence line passed from westto east to the north of Achern High convectivity partic-ularly at around 1500 UTC was measured by the lidarThe convective conditions suggested a suitable day touse lidar data to calculate (QH ) a parameter importantin monitoring the development of thermals A methoddiscussed in detail by DAVIS et al (2008) is outlinedand tested here
It was first necessary to confirm the reliability ofthe Doppler lidar data Comparisons were made be-tween horizontal velocity data measured by the lidar ra-diosonde and a 1290 MHz wind profiler and are shownhere
Figure 4 illustrates similarities between horizontal ve-locities measured by Doppler lidar and radiosonde theradiosonde was launched at 1000 UTC and the azimuthscan was performed from 1000 UTC to 1005 UTCWhen this data was taken it is considered that the bound-ary layer top was at around 800 m and the lidar data be-comes noisy and therefore un-useable above this level
Another comparison was made between the UFAMDoppler lidar data and the University of Manchester op-erated UFAM funded 1290 MHz wind profiler Thisis illustrated in Figure 5 As with the radiosonde data(Figure 4) it can be seen that there are similarities be-tween the two datasets There are some areas of smallscale differences but generally the two remote sensinginstruments are providing comparable data Having es-tablished that the Doppler lidar is a suitable instrumentto make atmospheric measurements and that the ambi-ent conditions were appropriate more complex calcula-tions could be performed to investigate other parameterssuch as QH
The chosen method assumes unstable conditions so toconfirm whether this was the situation for the 15th July2007 case potential temperature profiles were derivedfrom two radiosonde ascents launched from the Achernsite These profiles suggesting unstable conditions areshown in Figure 6
4 Methodology
Having established unstable conditions it was possibleto calculate QH from the vertical velocity-potential tem-
perature covariance term (wprimeθprime) in (41) following theprocedure outlined in GAL-CHEN et al (1992)
g
θ0wprimeθprime =
1
ρ0wprime
partpprime
partz+
ε
3+
part
partz
(
1
2wprime3
)
(41)
WYNGAARD and COTE (1971) and DAVIS et al
(2008) found that the pressure term partpprime
partz was small com-pared with the other terms of (42) for unstable condi-tions and will therefore be ignored in the subsequentanalysis which can be applied only to cases of convec-tion hence
wprimeθprime asympθ0
g
[
part
partz
(
1
2wprime3
)
+ε
3
]
(42)
Figure 5 Vector plot from UFAM 1290 MHz wind profiler showing
similar features to the lidar (Figure 2) with easterlies early in the
morning becoming southerly as the morning progresses and veering
westerly in the afternoon
Figure 6 Potential temperature profiles on 150707 suggesting
unstable conditions
The mean vertical velocity w is calculated for eachrange gate and averaged over the 25 minute scan periodThis is done for each of the nineteen 30 m-long rangegates from 135 m to 705 m using data from a verti-cally pointing stare scan The wrsquo is the deviation fromthe mean vertical velocity Then wrsquo3 is calculated foreach point and averaged over the 25 minute scan dura-
tion to form a profile of wprime3
In order to use (42) to calculate wprimeθprime it is also neces-sary to estimate ε One way of estimating ε is to exam-ine the line spectra of the longitudinal velocity correla-tion In the inertial subrange the expected relationshipis (BATCHELOR 1967)
f(κ) = αε23κminus53 (43)
where κ is the wave number α is a universal constant(05) and f(κ) is the Fourier transform of the longitudi-nal velocity correlation A -53 slope indicates the pres-
Meteorol Z 18 2009 Davis et al Doppler lidar observations 159
Table 2 Errors associated with Doppler lidar calculations of ε and wprime2
Error type and extent
Vertical velocity variancewprime2 Bias = 002 m2 sminus2 PEARSON et al 2008
Random = plusmn068 m2 sminus2
Turbulent kinetic energy Error = 10 FRELICH and CORNMAN 2002
dissipation rate ε
Figure 7 Spectrum from the GAL-CHEN et al (1992) spatial
method A -53 gradient indicates the presence of the inertial sub-
range
Figure 8 Temporal spectrum as reported by CHAMPAGNE et al
(1977) The -23 slope highlights the presence of the inertial sub-
range
ence of an inertial subrange and is shown in Figure 7This will be referred to henceforth as the spatial spec-tra method We calculate the spectra for each ray from135 to 705 m and average the spectra over the 25 minutetime period
An alternative method for calculating spectra is pro-posed by CHAMPAGNE et al (1977) Here Taylorrsquos Hy-
pothesis (STULL 1988) is assumed and is used to cal-culate the wave number The u is calculated from the li-dar data using VAD analysis The u is the average meanwind over the 135ndash705 m height range The spectrumcan be calculated from
nS (n) = 068ε23
(
2πn
u
)minus23
(44)
where S(n) is the spectral energy of frequency n andu is the mean wind speed Here the inertial subrangeis characterized by a ndash23 gradient and is shown in Fig-ure 8 This will be referred to henceforth as the temporalspectra method
When performing spectral analysis to obtain ε thedata were detrended to remove any nonstationary be-haviour However errors are still expected and these arelisted in Table 2
With unstable conditions QH can thus be calculatedfrom
QH = ρ0CP wprimeθprime (45)
taking CP the heat capacity of air at constant pressureas 1004 J kminus1 kgminus1 Kminus1 ρ0 the air density as 1275kg mminus3 and where is the potential temperature-verticalvelocity covariance calculated in (42) (GAL-CHEN etal 1992) (43) also assumes unstable conditions whichwere present on this day as illustrated in Figure 6
5 Results
The Salford Autonomous Doppler Lidar System was de-ployed for COPS in Achern Germany where it col-lected data almost continuously for 3 months A casestudy of convective development is described here andthe results include turbulence spectra and some time-series of ε and QH Figure 7 shows a spectrum producedfrom a vertically-pointing stare taken at 1100 UTC onthe 15th July 2007 in order to obtain a single value of εThe spectrum in Figure 7 was calculated using the spa-tial method presented by GAL-CHEN et al (1992) Inaddition in order to confirm the results shown in Fig-ure 7 spectra were also plotted using the same datasetand the method proposed by CHAMPAGNE et al (1977)as shown in Figure 8 The points deviating from the ex-pected ndash23 gradient at high wavenumbers are consid-ered to be due to noise (MAYOR et al 1997)
160 Davis et al Doppler lidar observations Meteorol Z 18 2009
Figure 9 Eddy dissipation rates calculated from Doppler lidar data
using both spatial and temporal methods on 150707
Figure 10 QH timeseries calculated for 150707 showing an esti-
mation of the expected errors Half-hourly averages of QH as mea-
sured by the IMK-FZK energy balance station for that day are also
shown
The time series in Figure 9 illustrates ε derived fromboth temporal and spatial methods developing through-out the day as would be expected ndash increasing towardsnoon peaking at 1500 UTC and then decreasing towardsthe evening as convection is developing in the morn-ing and then dissipating in the evening (STULL 1988)With the estimates for ε and vertical velocity fluctua-tions (wprime3) QH was calculated as illustrated in Figure10 It was decided that although the ε values were sim-ilar for both spectra the values derived from the spatialspectra would be utilised to calculate QH as per GAL-CHEN et al (1992) Figure 10 also shows a timeseriesof QH measured using an Energy Balance Station atthe COPS operation centre Baden Airpark 10 km northof Achern where the Doppler lidar was located but stillin the Rhine Valley
Figure 11 QH from Doppler lidar data and WRF model data
When considering the results achieved using the datacollected with the Doppler lidar certain factors need tobe considered for example how did the results com-pare with other measurements Figure 10 shows the li-dar data plotted against data collected from an energybalance station located at Baden Airpark 10 km to thenorth of the Doppler lidar The energy balance stationmade point measurements of QH at a single point intime at the surface and averaged over half an hour Themeasurements made by the lidar are averages over thedepth of the boundary layer up to 705 m and while show-ing a general trend cannot be the same for the followingreasons
bull The two instruments were not collocated
bull The lidar as a remote sensor provides a volumeaverage over a pencil-shaped beam over a range of120 m to 705 m above the surface
The QH values measured by the lidar are higher thanthose measured at Baden Airpark It is considered thatthe relatively high values of QH calculated from thelidar data compared with those from the energy bal-ance station result from the nature of the instrumentsand their associated lsquofootprintsrsquo (SCHMID 1994) Li-dar measurements are averaged over the lowest sim700m of the boundary layer and so have larger footprintsthan ground based instruments in the order of severalkilometres WOOD and MASON (1993) BELCHER andWOOD (1996) and WOOD et al (2001) discuss the ef-fects of increasing form drag and roughness length whenflow is directed across orography such as the mountainsto the east of the Achern site Since the mountains aresome 8 km away within the fetch of the lidar it is con-sidered that it the high eddy dissipation rates are due tothis orography These high values greatly influence themagnitude of the QH The instruments at Baden Air-park are at surface level so have lower fetch and are un-likely to be affected by the orography It should also be
Meteorol Z 18 2009 Davis et al Doppler lidar observations 161
noted that on this occasion the synoptic situation indi-cated that the convergence line mentioned in section 3was more active to the north of Achern
Despite these differences the results are comparablehaving a reasonable order of magnitude and showing adiurnal trend as expected
Output from the NERC WRF model runs for this dayis shown in Figure 11 These data also show a diurnaltrend similar to those shown in Figure 10 It is not real-istic to expect remotely sensed data to replicate the re-sults from a computer model but as with the surface val-ues the results are within the same order of magnitudethere is a similar trend in the model output with valuesrising towards noon and dropping towards sunset It isinteresting to note that where the model output suggestshigh heat flux values at 0800 UTC the lidar does notAlthough the measured QH values were comparativelyhigh at around 1500 UTC these values were unrealisti-cally so It is thought that this may be due to a mesoscaleevent that occurred at about 1500 UTC that day as men-tioned previously Investigations into this are ongoing
6 Conclusions and future work
It has been shown that by using a method outlined byDAVIS et al (2008) GAL-CHEN et al (1992) it is pos-sible to calculate values of QH and it is considered thatthese values fall within an acceptable range It is notedthat the diurnal cycle of the lidar-derived QH is closer tothe diurnal cycle of ε whereas both the model and sur-face measurements of QH peak earlier in the day QH
calculated from Doppler lidar data using the method de-scribed here would be expected to differ from that cal-culated using a surface based lsquopointrsquo sensor such as anenergy balance instrument or a sonic anemometer sincethe data collected with a Doppler lidar is integrated overa lsquovolumersquo of atmosphere aloft advecting through thelaser beam However these measurements are a usefuladdition to the other measurements provided by the li-dar and require no extra equipment
It is considered that further investigations of employ-ing this method to calculate QH are needed to establishthe overall applicability of the method and build on thepreliminary calculations It is expected that such inves-tigations will include
bull Calculation of ε using alternative methods
bull investigation of ε at different heights and atmo-spheric conditions
bull verification of Doppler lidar estimations of ε and wprime3and
bull investigation over different surface types
Acknowledgments
The authors would like to acknowledge Dr N KALT-HOFF at IMK FZK Karlsruhe for supplying QH datafrom their energy balance station located at Baden Air-field Germany collected as part of COPS July 2007Also Dr A GADIAN from the University of Leeds forhis help in securing the WRF model output Dr E NOR-TON from the University of Manchester for making theUFAM 1290 MHz wind profiler quicklook plots avail-able online HP SCHMID for the use of the flux source-area model (FSAM) finally UK COPS for fundingthrough NERC
References
BANAKH VA IN SMALIKHO F KOPP C WERNER1999 Measurements of turbulent energy dissipation ratewith a CW Doppler lidar in the atmospheric boundarylayer J Atmos Oceanic Technol 16 1044 mdash1061
BATCHELOR GK 1967 The theory of homogeneous tur-bulence 5th ed ndash Cambridge University Press London
BELCHER SE A WOOD 1996 Form and wave drag dueto stably stratified turbulent flow over low ridges ndash QuartJ Roy Meteor Soc 122 863mdash902
BENNETT LJ KA BROWNING AM BLYTH DJPARKER PA CLARK 2006 A review of the initiation ofprecipitating convection in the United Kingdom ndash QuartJ Roy Meteor Soc 132 1001mdash1020
BOZIER KE GN PEARSON CG COLLIER 2007Doppler lidar observations of Russian forest fire plumesover Helsinki ndash Weather 62 203ndash208
BROWNING KA R WEXLER 1968 The Determinationof Kinematic Properties of a Wind Field using DopplerRadar ndash Quart J Roy Meteor Soc 97 283ndash299
CHAMPAGNE FH CA FRIEHE JC LARUE JC WYN-GAARD 1977 Flux measurements flux estimation tech-niques and fine scale turbulence measurements in the un-stable surface layer over land ndash J Atmos Sci 34 515ndash530
COLLIER CG F DAVIES KE BOZIER AR HOLTDR MIDDLETON GN PEARSON S SIEMEN DVWILLETS GJG UPTON RI YOUNG 2005 DualDoppler lidar measurements for improving dispersionmodels ndash Bull Amer Meter Soc 86 825mdash838
DAVIES F CG COLLIERKE BOZIER GN PEARSON2003 On the accuracy of retrieved wind information fromDoppler lidar observations ndash Quart J Roy Meteor Soc129 321ndash334
DAVIES F CG COLLIER GN PEARSON KE BOZIER2004 Doppler lidar measurements of turbulent structurefunction over an urban area ndash J Atmos Oceanic Technol21 753mdash761
DAVIS JC KE BOZIER CG COLLIER F DAVIES2008 Spatial variations of sensible heat flux over an urbanarea ndash Meteor Appl 15 367mdash380
FREHLICH RG 1995 Comparison of 2- and 10- microm coher-ent Doppler lidar performance ndash J Atmos Oceanic Tech-nol 14 54ndash75
FREHLICH R L CORNMAN 2002 Estimating spatial ve-locity statistics with coherent Doppler lidar ndash J AtmosOceanic Technol 19 355ndash366
162 Davis et al Doppler lidar observations Meteorol Z 18 2009
GAL-CHEN T M XU WL EBERHARD 1992 Estima-tions of atmospheric boundary layer fluxes and other tur-bulence parameters from Doppler lidar data ndash J GeophysRes 97 D17 18409ndash18423
HARDESTY RM CJ GRUND MJ POST BJ RYEGN PEARSON 1992 Measurements of winds and cloudcharacteristics a comparison of Doppler lidar systemsPaper 2 Session TA-P ndash International Geoscience andRemote Sensing Symposium Houston Texas USA
LENSCHOW DH V WULFMEYER C SENFF 2000 Mea-suring Second- through Fourth-Order Moments in NoisyData ndash J Atmos Oceanic Technol 17 1330ndash1347
MAYOR SD SA COHN DH LENSCHOW CJGRUND RM HARDESTY 1997 Convective Bound-ary Layer Vertical Velocity Statistics Observed by 2-micromDoppler Lidar ndash 9th Conference on Coherent Laser RadarJune 23ndash27 Linkoping Sweden
MOK TM CZ RUDOWICZ 2004 A lidar study of theatmospheric entrainment zone and mixed layer over HongKong ndash Atmos Res 69 147ndash163
PEARSON G F DAVIES C COLLIER 2009 An analysisof the performance of the UFAM pulsed Doppler lidar forobserving the boundary layer ndash J Atmos Oceanic Technol2 240ndash250
ROZOFF CM WR COTTON JO ADEGOKE 2003 Sim-ulation of St Louis Missouri Land Use Impacts on Thun-
derstorms ndash J Appl Meteor 42 716ndash738
SCHMID HP 1994 Source areas for scalars and scalarfluxes ndash Bound-Layer Meteor 67 293ndash318
SCHWIESOW RL 1986 A comparative overview of activeremote-sensing techniques in Lenschow DH (Ed) Prob-ing the atmospheric boundary layer Amer Meteorol SocBoston
STULL RB 1988 An introduction to Boundary Layer Me-teorology ndash Kluwer Academic Publishers The Nether-lands
THIELEN J W WOBROCK A GADIAN P METSTAYERJ CREUTIN 2000 The possible influence of urban sur-faces on rain fall development a sensitivity study in 2D inthe meso-γ-scale Atmos Res 54 15ndash39
WOOD N AR BROWN FE HEWER 2001 Parametriz-ing the effects of orography on the boundary layer an al-ternative to effective roughness lengths ndash Quart J RoyMeteor Soc 127 759ndash777
WOOD N PJ MASON 1993 The pressure force inducedby neutral turbulent flow over hills ndash Quart J Roy MeteorSoc 119 1233ndash1267
WYNGAARD JC OR COTE 1971 The budgets of turbu-lent kinetic energy and temperature variance in the atmo-spheric surface layer ndash J Atmos Sci 28 190ndash201
Meteorol Z 18 2009 Davis et al Doppler lidar observations 159
Table 2 Errors associated with Doppler lidar calculations of ε and wprime2
Error type and extent
Vertical velocity variancewprime2 Bias = 002 m2 sminus2 PEARSON et al 2008
Random = plusmn068 m2 sminus2
Turbulent kinetic energy Error = 10 FRELICH and CORNMAN 2002
dissipation rate ε
Figure 7 Spectrum from the GAL-CHEN et al (1992) spatial
method A -53 gradient indicates the presence of the inertial sub-
range
Figure 8 Temporal spectrum as reported by CHAMPAGNE et al
(1977) The -23 slope highlights the presence of the inertial sub-
range
ence of an inertial subrange and is shown in Figure 7This will be referred to henceforth as the spatial spec-tra method We calculate the spectra for each ray from135 to 705 m and average the spectra over the 25 minutetime period
An alternative method for calculating spectra is pro-posed by CHAMPAGNE et al (1977) Here Taylorrsquos Hy-
pothesis (STULL 1988) is assumed and is used to cal-culate the wave number The u is calculated from the li-dar data using VAD analysis The u is the average meanwind over the 135ndash705 m height range The spectrumcan be calculated from
nS (n) = 068ε23
(
2πn
u
)minus23
(44)
where S(n) is the spectral energy of frequency n andu is the mean wind speed Here the inertial subrangeis characterized by a ndash23 gradient and is shown in Fig-ure 8 This will be referred to henceforth as the temporalspectra method
When performing spectral analysis to obtain ε thedata were detrended to remove any nonstationary be-haviour However errors are still expected and these arelisted in Table 2
With unstable conditions QH can thus be calculatedfrom
QH = ρ0CP wprimeθprime (45)
taking CP the heat capacity of air at constant pressureas 1004 J kminus1 kgminus1 Kminus1 ρ0 the air density as 1275kg mminus3 and where is the potential temperature-verticalvelocity covariance calculated in (42) (GAL-CHEN etal 1992) (43) also assumes unstable conditions whichwere present on this day as illustrated in Figure 6
5 Results
The Salford Autonomous Doppler Lidar System was de-ployed for COPS in Achern Germany where it col-lected data almost continuously for 3 months A casestudy of convective development is described here andthe results include turbulence spectra and some time-series of ε and QH Figure 7 shows a spectrum producedfrom a vertically-pointing stare taken at 1100 UTC onthe 15th July 2007 in order to obtain a single value of εThe spectrum in Figure 7 was calculated using the spa-tial method presented by GAL-CHEN et al (1992) Inaddition in order to confirm the results shown in Fig-ure 7 spectra were also plotted using the same datasetand the method proposed by CHAMPAGNE et al (1977)as shown in Figure 8 The points deviating from the ex-pected ndash23 gradient at high wavenumbers are consid-ered to be due to noise (MAYOR et al 1997)
160 Davis et al Doppler lidar observations Meteorol Z 18 2009
Figure 9 Eddy dissipation rates calculated from Doppler lidar data
using both spatial and temporal methods on 150707
Figure 10 QH timeseries calculated for 150707 showing an esti-
mation of the expected errors Half-hourly averages of QH as mea-
sured by the IMK-FZK energy balance station for that day are also
shown
The time series in Figure 9 illustrates ε derived fromboth temporal and spatial methods developing through-out the day as would be expected ndash increasing towardsnoon peaking at 1500 UTC and then decreasing towardsthe evening as convection is developing in the morn-ing and then dissipating in the evening (STULL 1988)With the estimates for ε and vertical velocity fluctua-tions (wprime3) QH was calculated as illustrated in Figure10 It was decided that although the ε values were sim-ilar for both spectra the values derived from the spatialspectra would be utilised to calculate QH as per GAL-CHEN et al (1992) Figure 10 also shows a timeseriesof QH measured using an Energy Balance Station atthe COPS operation centre Baden Airpark 10 km northof Achern where the Doppler lidar was located but stillin the Rhine Valley
Figure 11 QH from Doppler lidar data and WRF model data
When considering the results achieved using the datacollected with the Doppler lidar certain factors need tobe considered for example how did the results com-pare with other measurements Figure 10 shows the li-dar data plotted against data collected from an energybalance station located at Baden Airpark 10 km to thenorth of the Doppler lidar The energy balance stationmade point measurements of QH at a single point intime at the surface and averaged over half an hour Themeasurements made by the lidar are averages over thedepth of the boundary layer up to 705 m and while show-ing a general trend cannot be the same for the followingreasons
bull The two instruments were not collocated
bull The lidar as a remote sensor provides a volumeaverage over a pencil-shaped beam over a range of120 m to 705 m above the surface
The QH values measured by the lidar are higher thanthose measured at Baden Airpark It is considered thatthe relatively high values of QH calculated from thelidar data compared with those from the energy bal-ance station result from the nature of the instrumentsand their associated lsquofootprintsrsquo (SCHMID 1994) Li-dar measurements are averaged over the lowest sim700m of the boundary layer and so have larger footprintsthan ground based instruments in the order of severalkilometres WOOD and MASON (1993) BELCHER andWOOD (1996) and WOOD et al (2001) discuss the ef-fects of increasing form drag and roughness length whenflow is directed across orography such as the mountainsto the east of the Achern site Since the mountains aresome 8 km away within the fetch of the lidar it is con-sidered that it the high eddy dissipation rates are due tothis orography These high values greatly influence themagnitude of the QH The instruments at Baden Air-park are at surface level so have lower fetch and are un-likely to be affected by the orography It should also be
Meteorol Z 18 2009 Davis et al Doppler lidar observations 161
noted that on this occasion the synoptic situation indi-cated that the convergence line mentioned in section 3was more active to the north of Achern
Despite these differences the results are comparablehaving a reasonable order of magnitude and showing adiurnal trend as expected
Output from the NERC WRF model runs for this dayis shown in Figure 11 These data also show a diurnaltrend similar to those shown in Figure 10 It is not real-istic to expect remotely sensed data to replicate the re-sults from a computer model but as with the surface val-ues the results are within the same order of magnitudethere is a similar trend in the model output with valuesrising towards noon and dropping towards sunset It isinteresting to note that where the model output suggestshigh heat flux values at 0800 UTC the lidar does notAlthough the measured QH values were comparativelyhigh at around 1500 UTC these values were unrealisti-cally so It is thought that this may be due to a mesoscaleevent that occurred at about 1500 UTC that day as men-tioned previously Investigations into this are ongoing
6 Conclusions and future work
It has been shown that by using a method outlined byDAVIS et al (2008) GAL-CHEN et al (1992) it is pos-sible to calculate values of QH and it is considered thatthese values fall within an acceptable range It is notedthat the diurnal cycle of the lidar-derived QH is closer tothe diurnal cycle of ε whereas both the model and sur-face measurements of QH peak earlier in the day QH
calculated from Doppler lidar data using the method de-scribed here would be expected to differ from that cal-culated using a surface based lsquopointrsquo sensor such as anenergy balance instrument or a sonic anemometer sincethe data collected with a Doppler lidar is integrated overa lsquovolumersquo of atmosphere aloft advecting through thelaser beam However these measurements are a usefuladdition to the other measurements provided by the li-dar and require no extra equipment
It is considered that further investigations of employ-ing this method to calculate QH are needed to establishthe overall applicability of the method and build on thepreliminary calculations It is expected that such inves-tigations will include
bull Calculation of ε using alternative methods
bull investigation of ε at different heights and atmo-spheric conditions
bull verification of Doppler lidar estimations of ε and wprime3and
bull investigation over different surface types
Acknowledgments
The authors would like to acknowledge Dr N KALT-HOFF at IMK FZK Karlsruhe for supplying QH datafrom their energy balance station located at Baden Air-field Germany collected as part of COPS July 2007Also Dr A GADIAN from the University of Leeds forhis help in securing the WRF model output Dr E NOR-TON from the University of Manchester for making theUFAM 1290 MHz wind profiler quicklook plots avail-able online HP SCHMID for the use of the flux source-area model (FSAM) finally UK COPS for fundingthrough NERC
References
BANAKH VA IN SMALIKHO F KOPP C WERNER1999 Measurements of turbulent energy dissipation ratewith a CW Doppler lidar in the atmospheric boundarylayer J Atmos Oceanic Technol 16 1044 mdash1061
BATCHELOR GK 1967 The theory of homogeneous tur-bulence 5th ed ndash Cambridge University Press London
BELCHER SE A WOOD 1996 Form and wave drag dueto stably stratified turbulent flow over low ridges ndash QuartJ Roy Meteor Soc 122 863mdash902
BENNETT LJ KA BROWNING AM BLYTH DJPARKER PA CLARK 2006 A review of the initiation ofprecipitating convection in the United Kingdom ndash QuartJ Roy Meteor Soc 132 1001mdash1020
BOZIER KE GN PEARSON CG COLLIER 2007Doppler lidar observations of Russian forest fire plumesover Helsinki ndash Weather 62 203ndash208
BROWNING KA R WEXLER 1968 The Determinationof Kinematic Properties of a Wind Field using DopplerRadar ndash Quart J Roy Meteor Soc 97 283ndash299
CHAMPAGNE FH CA FRIEHE JC LARUE JC WYN-GAARD 1977 Flux measurements flux estimation tech-niques and fine scale turbulence measurements in the un-stable surface layer over land ndash J Atmos Sci 34 515ndash530
COLLIER CG F DAVIES KE BOZIER AR HOLTDR MIDDLETON GN PEARSON S SIEMEN DVWILLETS GJG UPTON RI YOUNG 2005 DualDoppler lidar measurements for improving dispersionmodels ndash Bull Amer Meter Soc 86 825mdash838
DAVIES F CG COLLIERKE BOZIER GN PEARSON2003 On the accuracy of retrieved wind information fromDoppler lidar observations ndash Quart J Roy Meteor Soc129 321ndash334
DAVIES F CG COLLIER GN PEARSON KE BOZIER2004 Doppler lidar measurements of turbulent structurefunction over an urban area ndash J Atmos Oceanic Technol21 753mdash761
DAVIS JC KE BOZIER CG COLLIER F DAVIES2008 Spatial variations of sensible heat flux over an urbanarea ndash Meteor Appl 15 367mdash380
FREHLICH RG 1995 Comparison of 2- and 10- microm coher-ent Doppler lidar performance ndash J Atmos Oceanic Tech-nol 14 54ndash75
FREHLICH R L CORNMAN 2002 Estimating spatial ve-locity statistics with coherent Doppler lidar ndash J AtmosOceanic Technol 19 355ndash366
162 Davis et al Doppler lidar observations Meteorol Z 18 2009
GAL-CHEN T M XU WL EBERHARD 1992 Estima-tions of atmospheric boundary layer fluxes and other tur-bulence parameters from Doppler lidar data ndash J GeophysRes 97 D17 18409ndash18423
HARDESTY RM CJ GRUND MJ POST BJ RYEGN PEARSON 1992 Measurements of winds and cloudcharacteristics a comparison of Doppler lidar systemsPaper 2 Session TA-P ndash International Geoscience andRemote Sensing Symposium Houston Texas USA
LENSCHOW DH V WULFMEYER C SENFF 2000 Mea-suring Second- through Fourth-Order Moments in NoisyData ndash J Atmos Oceanic Technol 17 1330ndash1347
MAYOR SD SA COHN DH LENSCHOW CJGRUND RM HARDESTY 1997 Convective Bound-ary Layer Vertical Velocity Statistics Observed by 2-micromDoppler Lidar ndash 9th Conference on Coherent Laser RadarJune 23ndash27 Linkoping Sweden
MOK TM CZ RUDOWICZ 2004 A lidar study of theatmospheric entrainment zone and mixed layer over HongKong ndash Atmos Res 69 147ndash163
PEARSON G F DAVIES C COLLIER 2009 An analysisof the performance of the UFAM pulsed Doppler lidar forobserving the boundary layer ndash J Atmos Oceanic Technol2 240ndash250
ROZOFF CM WR COTTON JO ADEGOKE 2003 Sim-ulation of St Louis Missouri Land Use Impacts on Thun-
derstorms ndash J Appl Meteor 42 716ndash738
SCHMID HP 1994 Source areas for scalars and scalarfluxes ndash Bound-Layer Meteor 67 293ndash318
SCHWIESOW RL 1986 A comparative overview of activeremote-sensing techniques in Lenschow DH (Ed) Prob-ing the atmospheric boundary layer Amer Meteorol SocBoston
STULL RB 1988 An introduction to Boundary Layer Me-teorology ndash Kluwer Academic Publishers The Nether-lands
THIELEN J W WOBROCK A GADIAN P METSTAYERJ CREUTIN 2000 The possible influence of urban sur-faces on rain fall development a sensitivity study in 2D inthe meso-γ-scale Atmos Res 54 15ndash39
WOOD N AR BROWN FE HEWER 2001 Parametriz-ing the effects of orography on the boundary layer an al-ternative to effective roughness lengths ndash Quart J RoyMeteor Soc 127 759ndash777
WOOD N PJ MASON 1993 The pressure force inducedby neutral turbulent flow over hills ndash Quart J Roy MeteorSoc 119 1233ndash1267
WYNGAARD JC OR COTE 1971 The budgets of turbu-lent kinetic energy and temperature variance in the atmo-spheric surface layer ndash J Atmos Sci 28 190ndash201
160 Davis et al Doppler lidar observations Meteorol Z 18 2009
Figure 9 Eddy dissipation rates calculated from Doppler lidar data
using both spatial and temporal methods on 150707
Figure 10 QH timeseries calculated for 150707 showing an esti-
mation of the expected errors Half-hourly averages of QH as mea-
sured by the IMK-FZK energy balance station for that day are also
shown
The time series in Figure 9 illustrates ε derived fromboth temporal and spatial methods developing through-out the day as would be expected ndash increasing towardsnoon peaking at 1500 UTC and then decreasing towardsthe evening as convection is developing in the morn-ing and then dissipating in the evening (STULL 1988)With the estimates for ε and vertical velocity fluctua-tions (wprime3) QH was calculated as illustrated in Figure10 It was decided that although the ε values were sim-ilar for both spectra the values derived from the spatialspectra would be utilised to calculate QH as per GAL-CHEN et al (1992) Figure 10 also shows a timeseriesof QH measured using an Energy Balance Station atthe COPS operation centre Baden Airpark 10 km northof Achern where the Doppler lidar was located but stillin the Rhine Valley
Figure 11 QH from Doppler lidar data and WRF model data
When considering the results achieved using the datacollected with the Doppler lidar certain factors need tobe considered for example how did the results com-pare with other measurements Figure 10 shows the li-dar data plotted against data collected from an energybalance station located at Baden Airpark 10 km to thenorth of the Doppler lidar The energy balance stationmade point measurements of QH at a single point intime at the surface and averaged over half an hour Themeasurements made by the lidar are averages over thedepth of the boundary layer up to 705 m and while show-ing a general trend cannot be the same for the followingreasons
bull The two instruments were not collocated
bull The lidar as a remote sensor provides a volumeaverage over a pencil-shaped beam over a range of120 m to 705 m above the surface
The QH values measured by the lidar are higher thanthose measured at Baden Airpark It is considered thatthe relatively high values of QH calculated from thelidar data compared with those from the energy bal-ance station result from the nature of the instrumentsand their associated lsquofootprintsrsquo (SCHMID 1994) Li-dar measurements are averaged over the lowest sim700m of the boundary layer and so have larger footprintsthan ground based instruments in the order of severalkilometres WOOD and MASON (1993) BELCHER andWOOD (1996) and WOOD et al (2001) discuss the ef-fects of increasing form drag and roughness length whenflow is directed across orography such as the mountainsto the east of the Achern site Since the mountains aresome 8 km away within the fetch of the lidar it is con-sidered that it the high eddy dissipation rates are due tothis orography These high values greatly influence themagnitude of the QH The instruments at Baden Air-park are at surface level so have lower fetch and are un-likely to be affected by the orography It should also be
Meteorol Z 18 2009 Davis et al Doppler lidar observations 161
noted that on this occasion the synoptic situation indi-cated that the convergence line mentioned in section 3was more active to the north of Achern
Despite these differences the results are comparablehaving a reasonable order of magnitude and showing adiurnal trend as expected
Output from the NERC WRF model runs for this dayis shown in Figure 11 These data also show a diurnaltrend similar to those shown in Figure 10 It is not real-istic to expect remotely sensed data to replicate the re-sults from a computer model but as with the surface val-ues the results are within the same order of magnitudethere is a similar trend in the model output with valuesrising towards noon and dropping towards sunset It isinteresting to note that where the model output suggestshigh heat flux values at 0800 UTC the lidar does notAlthough the measured QH values were comparativelyhigh at around 1500 UTC these values were unrealisti-cally so It is thought that this may be due to a mesoscaleevent that occurred at about 1500 UTC that day as men-tioned previously Investigations into this are ongoing
6 Conclusions and future work
It has been shown that by using a method outlined byDAVIS et al (2008) GAL-CHEN et al (1992) it is pos-sible to calculate values of QH and it is considered thatthese values fall within an acceptable range It is notedthat the diurnal cycle of the lidar-derived QH is closer tothe diurnal cycle of ε whereas both the model and sur-face measurements of QH peak earlier in the day QH
calculated from Doppler lidar data using the method de-scribed here would be expected to differ from that cal-culated using a surface based lsquopointrsquo sensor such as anenergy balance instrument or a sonic anemometer sincethe data collected with a Doppler lidar is integrated overa lsquovolumersquo of atmosphere aloft advecting through thelaser beam However these measurements are a usefuladdition to the other measurements provided by the li-dar and require no extra equipment
It is considered that further investigations of employ-ing this method to calculate QH are needed to establishthe overall applicability of the method and build on thepreliminary calculations It is expected that such inves-tigations will include
bull Calculation of ε using alternative methods
bull investigation of ε at different heights and atmo-spheric conditions
bull verification of Doppler lidar estimations of ε and wprime3and
bull investigation over different surface types
Acknowledgments
The authors would like to acknowledge Dr N KALT-HOFF at IMK FZK Karlsruhe for supplying QH datafrom their energy balance station located at Baden Air-field Germany collected as part of COPS July 2007Also Dr A GADIAN from the University of Leeds forhis help in securing the WRF model output Dr E NOR-TON from the University of Manchester for making theUFAM 1290 MHz wind profiler quicklook plots avail-able online HP SCHMID for the use of the flux source-area model (FSAM) finally UK COPS for fundingthrough NERC
References
BANAKH VA IN SMALIKHO F KOPP C WERNER1999 Measurements of turbulent energy dissipation ratewith a CW Doppler lidar in the atmospheric boundarylayer J Atmos Oceanic Technol 16 1044 mdash1061
BATCHELOR GK 1967 The theory of homogeneous tur-bulence 5th ed ndash Cambridge University Press London
BELCHER SE A WOOD 1996 Form and wave drag dueto stably stratified turbulent flow over low ridges ndash QuartJ Roy Meteor Soc 122 863mdash902
BENNETT LJ KA BROWNING AM BLYTH DJPARKER PA CLARK 2006 A review of the initiation ofprecipitating convection in the United Kingdom ndash QuartJ Roy Meteor Soc 132 1001mdash1020
BOZIER KE GN PEARSON CG COLLIER 2007Doppler lidar observations of Russian forest fire plumesover Helsinki ndash Weather 62 203ndash208
BROWNING KA R WEXLER 1968 The Determinationof Kinematic Properties of a Wind Field using DopplerRadar ndash Quart J Roy Meteor Soc 97 283ndash299
CHAMPAGNE FH CA FRIEHE JC LARUE JC WYN-GAARD 1977 Flux measurements flux estimation tech-niques and fine scale turbulence measurements in the un-stable surface layer over land ndash J Atmos Sci 34 515ndash530
COLLIER CG F DAVIES KE BOZIER AR HOLTDR MIDDLETON GN PEARSON S SIEMEN DVWILLETS GJG UPTON RI YOUNG 2005 DualDoppler lidar measurements for improving dispersionmodels ndash Bull Amer Meter Soc 86 825mdash838
DAVIES F CG COLLIERKE BOZIER GN PEARSON2003 On the accuracy of retrieved wind information fromDoppler lidar observations ndash Quart J Roy Meteor Soc129 321ndash334
DAVIES F CG COLLIER GN PEARSON KE BOZIER2004 Doppler lidar measurements of turbulent structurefunction over an urban area ndash J Atmos Oceanic Technol21 753mdash761
DAVIS JC KE BOZIER CG COLLIER F DAVIES2008 Spatial variations of sensible heat flux over an urbanarea ndash Meteor Appl 15 367mdash380
FREHLICH RG 1995 Comparison of 2- and 10- microm coher-ent Doppler lidar performance ndash J Atmos Oceanic Tech-nol 14 54ndash75
FREHLICH R L CORNMAN 2002 Estimating spatial ve-locity statistics with coherent Doppler lidar ndash J AtmosOceanic Technol 19 355ndash366
162 Davis et al Doppler lidar observations Meteorol Z 18 2009
GAL-CHEN T M XU WL EBERHARD 1992 Estima-tions of atmospheric boundary layer fluxes and other tur-bulence parameters from Doppler lidar data ndash J GeophysRes 97 D17 18409ndash18423
HARDESTY RM CJ GRUND MJ POST BJ RYEGN PEARSON 1992 Measurements of winds and cloudcharacteristics a comparison of Doppler lidar systemsPaper 2 Session TA-P ndash International Geoscience andRemote Sensing Symposium Houston Texas USA
LENSCHOW DH V WULFMEYER C SENFF 2000 Mea-suring Second- through Fourth-Order Moments in NoisyData ndash J Atmos Oceanic Technol 17 1330ndash1347
MAYOR SD SA COHN DH LENSCHOW CJGRUND RM HARDESTY 1997 Convective Bound-ary Layer Vertical Velocity Statistics Observed by 2-micromDoppler Lidar ndash 9th Conference on Coherent Laser RadarJune 23ndash27 Linkoping Sweden
MOK TM CZ RUDOWICZ 2004 A lidar study of theatmospheric entrainment zone and mixed layer over HongKong ndash Atmos Res 69 147ndash163
PEARSON G F DAVIES C COLLIER 2009 An analysisof the performance of the UFAM pulsed Doppler lidar forobserving the boundary layer ndash J Atmos Oceanic Technol2 240ndash250
ROZOFF CM WR COTTON JO ADEGOKE 2003 Sim-ulation of St Louis Missouri Land Use Impacts on Thun-
derstorms ndash J Appl Meteor 42 716ndash738
SCHMID HP 1994 Source areas for scalars and scalarfluxes ndash Bound-Layer Meteor 67 293ndash318
SCHWIESOW RL 1986 A comparative overview of activeremote-sensing techniques in Lenschow DH (Ed) Prob-ing the atmospheric boundary layer Amer Meteorol SocBoston
STULL RB 1988 An introduction to Boundary Layer Me-teorology ndash Kluwer Academic Publishers The Nether-lands
THIELEN J W WOBROCK A GADIAN P METSTAYERJ CREUTIN 2000 The possible influence of urban sur-faces on rain fall development a sensitivity study in 2D inthe meso-γ-scale Atmos Res 54 15ndash39
WOOD N AR BROWN FE HEWER 2001 Parametriz-ing the effects of orography on the boundary layer an al-ternative to effective roughness lengths ndash Quart J RoyMeteor Soc 127 759ndash777
WOOD N PJ MASON 1993 The pressure force inducedby neutral turbulent flow over hills ndash Quart J Roy MeteorSoc 119 1233ndash1267
WYNGAARD JC OR COTE 1971 The budgets of turbu-lent kinetic energy and temperature variance in the atmo-spheric surface layer ndash J Atmos Sci 28 190ndash201
Meteorol Z 18 2009 Davis et al Doppler lidar observations 161
noted that on this occasion the synoptic situation indi-cated that the convergence line mentioned in section 3was more active to the north of Achern
Despite these differences the results are comparablehaving a reasonable order of magnitude and showing adiurnal trend as expected
Output from the NERC WRF model runs for this dayis shown in Figure 11 These data also show a diurnaltrend similar to those shown in Figure 10 It is not real-istic to expect remotely sensed data to replicate the re-sults from a computer model but as with the surface val-ues the results are within the same order of magnitudethere is a similar trend in the model output with valuesrising towards noon and dropping towards sunset It isinteresting to note that where the model output suggestshigh heat flux values at 0800 UTC the lidar does notAlthough the measured QH values were comparativelyhigh at around 1500 UTC these values were unrealisti-cally so It is thought that this may be due to a mesoscaleevent that occurred at about 1500 UTC that day as men-tioned previously Investigations into this are ongoing
6 Conclusions and future work
It has been shown that by using a method outlined byDAVIS et al (2008) GAL-CHEN et al (1992) it is pos-sible to calculate values of QH and it is considered thatthese values fall within an acceptable range It is notedthat the diurnal cycle of the lidar-derived QH is closer tothe diurnal cycle of ε whereas both the model and sur-face measurements of QH peak earlier in the day QH
calculated from Doppler lidar data using the method de-scribed here would be expected to differ from that cal-culated using a surface based lsquopointrsquo sensor such as anenergy balance instrument or a sonic anemometer sincethe data collected with a Doppler lidar is integrated overa lsquovolumersquo of atmosphere aloft advecting through thelaser beam However these measurements are a usefuladdition to the other measurements provided by the li-dar and require no extra equipment
It is considered that further investigations of employ-ing this method to calculate QH are needed to establishthe overall applicability of the method and build on thepreliminary calculations It is expected that such inves-tigations will include
bull Calculation of ε using alternative methods
bull investigation of ε at different heights and atmo-spheric conditions
bull verification of Doppler lidar estimations of ε and wprime3and
bull investigation over different surface types
Acknowledgments
The authors would like to acknowledge Dr N KALT-HOFF at IMK FZK Karlsruhe for supplying QH datafrom their energy balance station located at Baden Air-field Germany collected as part of COPS July 2007Also Dr A GADIAN from the University of Leeds forhis help in securing the WRF model output Dr E NOR-TON from the University of Manchester for making theUFAM 1290 MHz wind profiler quicklook plots avail-able online HP SCHMID for the use of the flux source-area model (FSAM) finally UK COPS for fundingthrough NERC
References
BANAKH VA IN SMALIKHO F KOPP C WERNER1999 Measurements of turbulent energy dissipation ratewith a CW Doppler lidar in the atmospheric boundarylayer J Atmos Oceanic Technol 16 1044 mdash1061
BATCHELOR GK 1967 The theory of homogeneous tur-bulence 5th ed ndash Cambridge University Press London
BELCHER SE A WOOD 1996 Form and wave drag dueto stably stratified turbulent flow over low ridges ndash QuartJ Roy Meteor Soc 122 863mdash902
BENNETT LJ KA BROWNING AM BLYTH DJPARKER PA CLARK 2006 A review of the initiation ofprecipitating convection in the United Kingdom ndash QuartJ Roy Meteor Soc 132 1001mdash1020
BOZIER KE GN PEARSON CG COLLIER 2007Doppler lidar observations of Russian forest fire plumesover Helsinki ndash Weather 62 203ndash208
BROWNING KA R WEXLER 1968 The Determinationof Kinematic Properties of a Wind Field using DopplerRadar ndash Quart J Roy Meteor Soc 97 283ndash299
CHAMPAGNE FH CA FRIEHE JC LARUE JC WYN-GAARD 1977 Flux measurements flux estimation tech-niques and fine scale turbulence measurements in the un-stable surface layer over land ndash J Atmos Sci 34 515ndash530
COLLIER CG F DAVIES KE BOZIER AR HOLTDR MIDDLETON GN PEARSON S SIEMEN DVWILLETS GJG UPTON RI YOUNG 2005 DualDoppler lidar measurements for improving dispersionmodels ndash Bull Amer Meter Soc 86 825mdash838
DAVIES F CG COLLIERKE BOZIER GN PEARSON2003 On the accuracy of retrieved wind information fromDoppler lidar observations ndash Quart J Roy Meteor Soc129 321ndash334
DAVIES F CG COLLIER GN PEARSON KE BOZIER2004 Doppler lidar measurements of turbulent structurefunction over an urban area ndash J Atmos Oceanic Technol21 753mdash761
DAVIS JC KE BOZIER CG COLLIER F DAVIES2008 Spatial variations of sensible heat flux over an urbanarea ndash Meteor Appl 15 367mdash380
FREHLICH RG 1995 Comparison of 2- and 10- microm coher-ent Doppler lidar performance ndash J Atmos Oceanic Tech-nol 14 54ndash75
FREHLICH R L CORNMAN 2002 Estimating spatial ve-locity statistics with coherent Doppler lidar ndash J AtmosOceanic Technol 19 355ndash366
162 Davis et al Doppler lidar observations Meteorol Z 18 2009
GAL-CHEN T M XU WL EBERHARD 1992 Estima-tions of atmospheric boundary layer fluxes and other tur-bulence parameters from Doppler lidar data ndash J GeophysRes 97 D17 18409ndash18423
HARDESTY RM CJ GRUND MJ POST BJ RYEGN PEARSON 1992 Measurements of winds and cloudcharacteristics a comparison of Doppler lidar systemsPaper 2 Session TA-P ndash International Geoscience andRemote Sensing Symposium Houston Texas USA
LENSCHOW DH V WULFMEYER C SENFF 2000 Mea-suring Second- through Fourth-Order Moments in NoisyData ndash J Atmos Oceanic Technol 17 1330ndash1347
MAYOR SD SA COHN DH LENSCHOW CJGRUND RM HARDESTY 1997 Convective Bound-ary Layer Vertical Velocity Statistics Observed by 2-micromDoppler Lidar ndash 9th Conference on Coherent Laser RadarJune 23ndash27 Linkoping Sweden
MOK TM CZ RUDOWICZ 2004 A lidar study of theatmospheric entrainment zone and mixed layer over HongKong ndash Atmos Res 69 147ndash163
PEARSON G F DAVIES C COLLIER 2009 An analysisof the performance of the UFAM pulsed Doppler lidar forobserving the boundary layer ndash J Atmos Oceanic Technol2 240ndash250
ROZOFF CM WR COTTON JO ADEGOKE 2003 Sim-ulation of St Louis Missouri Land Use Impacts on Thun-
derstorms ndash J Appl Meteor 42 716ndash738
SCHMID HP 1994 Source areas for scalars and scalarfluxes ndash Bound-Layer Meteor 67 293ndash318
SCHWIESOW RL 1986 A comparative overview of activeremote-sensing techniques in Lenschow DH (Ed) Prob-ing the atmospheric boundary layer Amer Meteorol SocBoston
STULL RB 1988 An introduction to Boundary Layer Me-teorology ndash Kluwer Academic Publishers The Nether-lands
THIELEN J W WOBROCK A GADIAN P METSTAYERJ CREUTIN 2000 The possible influence of urban sur-faces on rain fall development a sensitivity study in 2D inthe meso-γ-scale Atmos Res 54 15ndash39
WOOD N AR BROWN FE HEWER 2001 Parametriz-ing the effects of orography on the boundary layer an al-ternative to effective roughness lengths ndash Quart J RoyMeteor Soc 127 759ndash777
WOOD N PJ MASON 1993 The pressure force inducedby neutral turbulent flow over hills ndash Quart J Roy MeteorSoc 119 1233ndash1267
WYNGAARD JC OR COTE 1971 The budgets of turbu-lent kinetic energy and temperature variance in the atmo-spheric surface layer ndash J Atmos Sci 28 190ndash201
162 Davis et al Doppler lidar observations Meteorol Z 18 2009
GAL-CHEN T M XU WL EBERHARD 1992 Estima-tions of atmospheric boundary layer fluxes and other tur-bulence parameters from Doppler lidar data ndash J GeophysRes 97 D17 18409ndash18423
HARDESTY RM CJ GRUND MJ POST BJ RYEGN PEARSON 1992 Measurements of winds and cloudcharacteristics a comparison of Doppler lidar systemsPaper 2 Session TA-P ndash International Geoscience andRemote Sensing Symposium Houston Texas USA
LENSCHOW DH V WULFMEYER C SENFF 2000 Mea-suring Second- through Fourth-Order Moments in NoisyData ndash J Atmos Oceanic Technol 17 1330ndash1347
MAYOR SD SA COHN DH LENSCHOW CJGRUND RM HARDESTY 1997 Convective Bound-ary Layer Vertical Velocity Statistics Observed by 2-micromDoppler Lidar ndash 9th Conference on Coherent Laser RadarJune 23ndash27 Linkoping Sweden
MOK TM CZ RUDOWICZ 2004 A lidar study of theatmospheric entrainment zone and mixed layer over HongKong ndash Atmos Res 69 147ndash163
PEARSON G F DAVIES C COLLIER 2009 An analysisof the performance of the UFAM pulsed Doppler lidar forobserving the boundary layer ndash J Atmos Oceanic Technol2 240ndash250
ROZOFF CM WR COTTON JO ADEGOKE 2003 Sim-ulation of St Louis Missouri Land Use Impacts on Thun-
derstorms ndash J Appl Meteor 42 716ndash738
SCHMID HP 1994 Source areas for scalars and scalarfluxes ndash Bound-Layer Meteor 67 293ndash318
SCHWIESOW RL 1986 A comparative overview of activeremote-sensing techniques in Lenschow DH (Ed) Prob-ing the atmospheric boundary layer Amer Meteorol SocBoston
STULL RB 1988 An introduction to Boundary Layer Me-teorology ndash Kluwer Academic Publishers The Nether-lands
THIELEN J W WOBROCK A GADIAN P METSTAYERJ CREUTIN 2000 The possible influence of urban sur-faces on rain fall development a sensitivity study in 2D inthe meso-γ-scale Atmos Res 54 15ndash39
WOOD N AR BROWN FE HEWER 2001 Parametriz-ing the effects of orography on the boundary layer an al-ternative to effective roughness lengths ndash Quart J RoyMeteor Soc 127 759ndash777
WOOD N PJ MASON 1993 The pressure force inducedby neutral turbulent flow over hills ndash Quart J Roy MeteorSoc 119 1233ndash1267
WYNGAARD JC OR COTE 1971 The budgets of turbu-lent kinetic energy and temperature variance in the atmo-spheric surface layer ndash J Atmos Sci 28 190ndash201