Available online at www.sciencedirect.com

www.elsevier.com/locate/asr

Advances in Space Research 41 (2008) 1488–1494

A new narrow beam Doppler radar at 3 MHz for studiesof the high-latitude middle atmosphere

W. Singer a,*, R. Latteck a, D.A. Holdsworth b

a Leibniz Institute of Atmospheric Physics, Schloss-Street 6, 18225 Kuehlungsborn, Germanyb Atmospheric Radar Systems, 26 Stirling Street, Thebarton, SA 5031, Australia

Received 2 November 2004; received in revised form 24 September 2007; accepted 10 October 2007

Abstract

A Doppler radar at 3.17 MHz has been installed at Saura close to the Andøya Rocket Range as part of the ALOMAR observatory atAndenes, Norway in summer 2002 to improve the ground based capabilities for measurements of small scale features and turbulence inthe mesosphere. The main feature of the new Saura MF radar is the transmitting/receiving antenna which is arranged as a Mills Cross of29 crossed half-wave dipoles with a minimum beam width of about 7�. Each dipole is fed by its own transceiver, and the individual phasecontrol of the 58 transceiver modules on transmission and reception provides high flexibility in beam forming and pointing as well astransmission switching between ordinary and extraordinary mode circular polarisation. In addition, beams with different widths atthe same pointing angle can be formed. For multiple receiver applications (spaced antenna wind measurements, all-sky meteor detec-tions) four independent receiving channels are available.

The observation capabilities of the radar are demonstrated by wind measurements using various techniques, turbulent kinetic energydissipation rates from spectral width measurements, electron density estimates using differential absorption and phase measurements,and meteor observations. The results are discussed in relation with simultaneous observations from co-located MF/HF/VHF radars.� 2007 COSPAR. Published by Elsevier Ltd. All rights reserved.

Keywords: Instruments and techniques; Radar atmospheric physics; Antenna arrays; Mesospheric dynamics; D-region electron density; Meteors

1. Introduction

Radar observations at frequencies between 2 and 3 MHzallow continuous monitoring of the mesosphere betweenabout 50 and 95 km altitude throughout the year. Mostof these MF radars use a vertical wide beam antenna ontransmission, and the ionospheric returns are typicallyreceived by three crossed horizontal dipoles arranged inan equilateral triangle. These radars provide horizontalwinds and tides at altitudes between 60 km and about92 km using the full correlation analysis (spaced antennawinds). They lack much of the capabilities of the Dopplerradars that operate at higher frequencies such as the esti-

0273-1177/$34 � 2007 COSPAR. Published by Elsevier Ltd. All rights reserv

doi:10.1016/j.asr.2007.10.006

* Corresponding author.E-mail addresses: singer@iap-kborn.de (W. Singer), latteck@iap-kborn.de

(R. Latteck), dholdswo@atrad.com.au (D.A. Holdsworth).

mation of horizontal momentum fluxes and of turbulenceparameters using spectral width measurements. MF Dopp-ler radars require physically large antenna arrays to achieverelatively small beam widths successfully realized with theAdelaide MF Doppler radar (Briggs et al., 1969; Reidet al., 1995) and with a cross-array near Brisbane (Jones,1981).

The ALOMAR facility located nearby the AndoyaRocket Range at Andenes (69.3�N, 16.0�E) provides valu-able tools for observations of the middle atmosphere in anearly identical volume using ground based and in-situmethods (von Zahn, 1997). The ALWIN VHF radar at53.5 MHz (Latteck et al., 1999), the Andenes wide beamMF radar at 1.98 MHz (Singer et al., 1997), the ALOMARall-sky meteor radar at 32.5 MHz (Singer et al., 2003), theALOMAR RMR-lidar (von Zahn et al., 2000) and theWeber sodium lidar (She et al., 2002) as well as rocket-

ed.

W. Singer et al. / Advances in Space Research 41 (2008) 1488–1494 1489

borne in-situ measurements are available for studies ofdynamical processes in the Arctic middle atmosphere. Thiscomplex of complementing experiments close together(Fig. 1) has provided the motivation to build a MF Dopp-ler radar nearby with the aim to improve our ground basedcapabilities for studies of small scale features and turbu-lence. The new system was installed in July 2002 startingwith spaced antenna observations, with Doppler capabili-ties added in April 2003.

Fig. 1. Radar locations at the norther

2. The Saura MF Doppler radar

2.1. The antenna array

The narrow beam transmitting/receiving antenna con-sists of 29 crossed 3.17-MHz half-wave dipoles supportedabout 11 m above the ground, and arranged as a MillsCross. The spacing of the crossed dipoles is 0.7 wavelengthsat 3.17 MHz resulting in a minimum beam width of 6.6�

n part of Andøya island/Norway.

1490 W. Singer et al. / Advances in Space Research 41 (2008) 1488–1494

(full width at half-power, one way). The positions of thecrossed dipoles are shown in Fig. 2. The crossed dipolesZ1 and Z2 are additional receiving antennas for spacedantenna observations connected to the receive path oftransceivers 59–62. Fig. 3 presents the radiation patternof a vertically directed beam calculated with the Numerical

DDS

DDS

DDS

DDS

Envelope

Con

trol

ler

cntrl 1...16

Controler module 1

DDS

DDS

DDS

DDS

Envelope

Con

trol

ler

ref

ref

ref

Controler module 16

1 2 31…

....

Receiver(4 channel)

Controller

Synthesiser& Reference

Generator

IF CombinerRx4

Rx3

Rx2

Rx1

DA

DA

DA

DA

cntrl

Control Analysis

Transmitter

Inte

rfac

e

IP1

IP2

IP3

IP4

QP1

QP2

QP3

QP4

Fig. 2. Saura MF radar block diagram and Mills Cross

← W sin(θ) sin(φ)

← s

in(

Sθ)

cos

(φ)

N→

Saura M

–1 –0.5 0–1

–0.5

0

0.5

1

Fig. 3. Radiation pattern of a vertically pointed beam with a

Electromagnetics Code (NEC) with the Norton–Sommer-feld approximation, making allowance for the imperfectground.

Each of the dipoles is accessible at a patch board in thecentral hut through low loss coaxial cables of equal lengthof multiples of half-wavelengths. The estimated phase dif-

Tx/Rx2kW

Tx/Rx2kW

Tx/Rx2kW

Tx/Rx2kW

+

+

1

2

3

4

Tx/Rx2kW

Tx/Rx2kW

Tx/Rx2kW

Tx/Rx2kW

+

+

61

62

....

..

Transceivermodules

5.5˚

N

70 m

Mills CrossAntenna

W2W4W6

N2

N4

N6

S2

S4

S6

E6E4E2

Z1 Z2

antenna (spacing 70 m, length of each arm 1030 m).

E →

F radar

0.5 1

Gai

n / d

Bi

–21

–18

–15

–12

–9

–6

–3

0

3

6

9

12

15

18

beam width of 6.6� (full width at half-power, one way).

Table 1System parameters of the Saura MF radar (standard operation parametersin bold)

Location 69.3�N, 16.0�EFrequency 3.17 MHzPeak power 116 kWFeeding system loss 3.6 dBTransceiver modules 58 (62), each of 2 kWPulse width (Gaussian pulse) 7 ls, 10 ls, 13 lsSampling resolution 1000 m, 1500 m, 2000 mPulse repetition frequency 80–100 Hz (max. 200 Hz)Length of time series 120–180 sHeight range 40–103 kmCoherent integrations 2

Nyquist frequency/ resolution 10 Hz/0.0057 Hz

Table 2Antenna characteristics

Polarisation Left/right circularCrossed dipoles 29–on transmission: all modes 1. . .29–on reception: DBS mode 1. . .28–on reception: SA mode 4–on reception: meteor mode 4Gain 19 dBBeam width (FWHP) 6.6�/13.8�Beam directions: azimuth NW, SE, NE, SWBeam directions: zenith Vertical, 7.2�, 17.2�Interleaving modes NW/SE, NE/SW beams

Left/right polarisation

W. Singer et al. / Advances in Space Research 41 (2008) 1488–1494 1491

ferences between the individual dipoles including feedingcables and matching transformers are less than 15�. Thesedifferences are corrected by adjusting the cables connectingthe patch board with the transceiver input ports.

2.2. The transmitting and receiving system

Each dipole of the array is fed by its own transceiverunit with a peak power of 2 kW which is individually con-trolled in phase on transmission and reception (Fig. 4).Four transceiver modules are controlled by a controllermodule, and each transceiver has its own synthesiser forthe transmitting and receiving branch with a phase controlin increments of 11.25�. In addition, the modules can becontrolled in amplitude on transmission. The system isoperated with a peak power of 116 kW and has a maxi-mum height resolution of 1 km. The antenna and trans-ceiver design provide high flexibility in beam forming andbeam pointing as well as the capability of forming rightand left circular polarisation (ordinary and extraordinarymagneto-ionic component). All 58 transmitting and receiv-ing modules are adjusted in phase with accuracy betterthan 20�. The intermediate frequency (IF) outputs of thetwo transceivers feeding one crossed dipole arrangementare combined and led to a combining unit of the IF signals.The IF combiner provides the input signals for the 4-chan-nel receiver (down converter) and their base band signalsare digitized at a maximum of 64 height channels.

In Doppler (DBS) mode, the IF signals of the sevencrossed dipoles of each half-arm of the Mills cross (e.g.N1, . . .,N7) are combined into one receiver. Off-zenithbeams towards N, S, E, W and NW, NE, SE, SW at zenithangles of 7.3�/17.2� can be formed. In addition, beams withdifferent widths at the same pointing angle can be realized.For multiple receiver applications (spaced antenna windmeasurements, meteor observations) one crossed dipolecan be selected from each half-arm and two additionalcrossed dipole arrangements (Z1, Z2 in Fig. 2) are avail-able. The parameters of the radar system and of the Millscross antenna are summarized in Tables 1 and 2.

Andenes April 27, 2003

Hei

ght(

km)

90

80

70

-20 -200 020 20Vzon ( m/s)

0 UT 12 UT

Fig. 4. Hourly means of Doppler (thin lines) winds and spaced antennawinds (thick lines) estimated with the narrow beam Saura MF radar incomparison with spaced antenna winds from the wide beam Andenes MFradar (dotted lines).

3. Applications of the Saura MF radar

3.1. Wind and turbulence measurements

The great flexibility of the radar provides the opportu-nity for numerous Doppler beam steering (DBS) experi-ments and spaced antenna (SA) applications. Spacedantenna wind measurements are performed using the wholearray on transmission and four crossed dipoles in a ‘‘Y’’configuration (antennas Z1, Z2, N2, and N4 in Fig. 2) on

-80 -60 -40 -20 0 20 40

Zonal wind (m/s)

70

80

90

100

Hei

ght (

km)

-60 -40 -20 20 40 60

04-Jul-200205:00-06:00 UT

28-Jan-200312:00-13:00 UT

0

Fig. 5. Hourly means of spaced antenna winds obtained with the narrowbeam Saura MF radar (full lines) and meteor winds from the AndenesSKiYMET meteor radar (dashed lines) at Andenes under summer andwinter conditions.

1 10 100

(mW/kg)

50

60

70

80

90

100

Hei

ght(

km)

rocketclimatology

summer winter

radar

Fig. 6. Turbulent energy dissipation rates estimated from spectral widthobservations on 29 December 2005 and 14 June 2006 in comparison withclimatological data for summer and winter conditions (dashed lines) fromprevious rocket soundings at Andenes.

1492 W. Singer et al. / Advances in Space Research 41 (2008) 1488–1494

reception. Simultaneously measured zonal winds using thespaced antenna technique at the Andenes and Saura MFradars are in reasonable agreement both with the Dopplerwinds obtained by the Saura radar (Fig. 4). There are indi-cations that the spaced antenna winds are slightly underes-timated at the upper most altitudes. The spaced antennawinds obtained with the Saura MF radar are also in agree-ment with the winds determined by the ALOMAR meteorradar (Fig. 5) taking into account the different samplingstrategies of both instruments.

← sin (W θ) sin(

← s

in(θ

) co

s(φ)

NS

→

Saura MF

Fig. 7. Radiation pattern of a donut-l

The high flexibility of the Doppler experiments allowsmeasurements at various off-zenith directions with variousbeam widths. Radial velocities and fluxes of horizontalmomentum can be determined from radial velocities mea-sured in co-planar planes with interleaved beam swingingfrom data point to data point.

Radar observations of the spectral width allow an esti-mation of turbulent energy dissipation rates by differenttechniques, and the results can be compared with rocketmeasurements from the nearby Andoya Rocket Range.The application of the nested beam technique (VanZandtet al., 2002) which is based on soundings with a narrowand a broad beam at an off-zenith angle of 17� providesestimates of the turbulent kinetic energy dissipation ratescomparable to rocket soundings for low wind speeds (Lat-teck et al., 2003).

The spectral broadening has been precisely correctedtaking into account the antenna radiation pattern and themeasured background wind field (Hocking, 1983; Lattecket al., 2005) since September 2003. Fig. 6 presents meanturbulent energy dissipation rates for winter and summerconditions. In June low dissipation rates between 2 and20 mW/kg were found below 75 km and larger rates occu-red in December. Above 80 km larger energy dissipationrates were observed in June with about 180 mW/kg at90 km. These data are in qualitative agreement with clima-tological means from previous rocket measurements atAndenes (Lubken, 1997) with in general lower values insummer below about 82–85 km.

φ) E →

radarG

ain

/ dB

i

–12

–9

–6

–3

0

3

6

9

ike beam for meteor observations.

W. Singer et al. / Advances in Space Research 41 (2008) 1488–1494 1493

3.2. Meteor observations

Meteor observations can be performed in the MF fre-quency range during night time only. On transmission adonut-like antenna beam is formed by radial phasing ofthe array with maximum radiation at zenith angles between30� and 40� (Fig. 7) where most of the meteors are detected.Four crossed dipoles (W1, N1, E2, and S3 in Fig. 2) of theMills Cross are used as an interferometer on reception. Theheight distribution of meteors observed at 3 MHz inNovember 2002 peaks at around 105 km, about 15 kmhigher than in the VHF frequency range but with consider-

Fig. 9. Hourly mean winds obtained from meteor observations with theAndenes SKiYMET meteor radar (dashed lines) and the Saura MF radar(full lines) on 14 December 2004, 05:30 UT.

0 50 100 150 200 250 300

Meteors/day

70

80

90

100

110

120

130

140

Hei

ght [

km]

53.5 MHz

32.5 MHz

3.17 MHz

Andenes 20-Nov-2002

Fig. 8. Meteor flux observed at 3.17 MHz, 32.5 MHz, and 53.5 MHz onNovember 20, 2002 at Andenes.

able lower count rates (Fig. 8). The combination of co-located and simultaneous meteor observations (Singeret al., 2003) in the MF/VHF range allows a substantialextension of the height range of meteor observations forinvestigations of atmospheric dynamics and for astronom-ical studies (e.g. meteor showers with high entry velocitiessuch as the Leonids where the burn up of the meteoroidsstarts at greater altitudes). Meteor winds obtained withthe Saura MF radar extend the height range of wind obser-vations up to altitudes of about 110 km. The SKiYMETmeteor radar at 32.55 MHz covers the height range 82–98 km and the Saura MF radar provides data above about96 km. Both observations agree well in the region of over-lap. (See Fig. 9)

3.3. Electron density measurements

Electron number density profiles can be estimated fromradar soundings using different polarisations of the trans-mitted radio waves. Interleaved transmission of pulses withordinary and extraordinary polarisation provides reliablemeasurements of the differential absorption (DAE) andthe differential phase (DPE) which the radar signals receiveon their passage through the atmosphere. The electronnumber densities obtained from both data sets between60 and 85 km are in remarkably good agreement(Fig. 10) and are combined to a mean profile provided thatthe DAE and DPE results do not differ by more than afactor of two (Singer et al., 2005). The radar data are ingeneral agreement with electron densities derived fromrocket-borne radio wave propagation experiments (Fried-rich, 2000) for comparable levels of solar activity and solarzenith angles (Fig. 11).

4. Summary

The Saura MF radar has high flexibility in forming ofDoppler beams with various beam widths at different off-zenith angles and azimuths. Spaced antenna and Doppler

Fig. 10. Electron density profiles estimated from differential absorption(DAE) and differential phase (DPA) measurements on 25 July 2004,03:11–03:13 UT and the derived mean profile.

Fig. 11. Electron density profile from DAE/DPE measurements with theSaura MF radar on 25 July 2004 at solar zenith angle 80.8� (sunspotnumber 130) and electron density profile from a rocket-borne radio wavepropagation experiment at Andenes on 5 July 1999 under comparableconditions (solar zenith angle 87.8�, sunspot number 130).

1494 W. Singer et al. / Advances in Space Research 41 (2008) 1488–1494

beam swinging measurements provide 3D winds, estimatesof horizontal momentum fluxes and of turbulent kineticenergy dissipation rates. Height profiles of the electronnumber density are obtained from differential absorptionand differential phase measurements. The Saura MF radar,ALWIN MST radar, Andenes MF radar, Andenes meteorradar and the lidars of the ALOMAR facility (ALOMARRMR lidar, ALOMAR Weber sodium lidar) are locatedwithin 16 km allowing common volume observation. Cor-relative experiments with the co-located radars, lidars,and in-situ measurements from sounding and meteorolog-ical rockets launched from the Andoya Rocket Rangeallow intercomparisons of different remote and in-situ tech-niques as well as complementary studies of mesospheric/lower thermospheric processes.

Acknowledgments

The Saura MF radar is a joint experiment of the AndoyaRocket Range, the National Institute of Information andCommunications Technology, Tokyo, Japan and theIAP. The authors thank the Engineers and Techniciansof the Leibniz Institute of Atmospheric Physics and theAndøya Rocket Range for their engagement building upthe new antenna system and installing the radar system.The radar was designed and manufactured by AtmosphericRadar Systems (ATRAD). Radar development was sup-

ported in part by grant D/0127520 of Deutscher Akadem-ischer Austausch Dienst (DAAD).

References

Briggs, B.H., Elford, W.G., Felgate, D.G., Golley, M.G., Rossiter, D.E.,Smith, J.W. Buckland Park aerial array. Nature 223, 1321–1325, 1969.

Friedrich, M., private communication, 2000.Jones, K.L. The interpretation of radar echoes from altitudes near 100 km

using a phased antenna array. J. Atmos. Terr. Phys. 43, 1015–1026,1981.

Hocking, W.K. On the extraction of atmospheric turbulence parametersfrom radar backscatter Doppler spectra – I. Theory. J. Atmos. Terr.Phys. 45, 89–102, 1983.

Latteck, R., Singer, W., Bardey, H. The ALWIN MST Radar - Technicaldesign and performances. In: Kaldeich-Schurmann, B. (Ed.), Proceed-ing of the 14th ESA Symp. On European Rocket and BalloonProgrammes and Related Research. Vol. SP-437, pp. 179–184, 1999.

Latteck, R., Singer, W., Engler, N. Application of the dual beam widthmethod to a narrow beam MF radar for estimation of turbulentspectral width. In: Chau, J., Rottger, J. (Eds.), Proceedings of the 10thInternational Workshop on Technical and Scientific Aspects of MSTRadar (MST10). Radio Observatorio de Jicamarca, Lima, Peru, pp.245–248, 2003.

Latteck, R., Singer, W., Hocking, W.K. Measurement of turbulent kineticenergy dissipation rates in the mesosphere by a 3 MHz Doppler radar.Adv. Space Res. 35, 1905–1910, 2005.

Lubken, F.-J. Seasonal variation of turbulent energy dissipation rates athigh latitudes as determined by in-situ measurements of neutral densityfluctuations. J. Geophys. Res. 102, 13441–13456, 1997.

Reid, I.M., Vandepeer, B.G.W., Dillon, S.D., Fuller, B.G. The newAdelaide MF Doppler radar. Radio Sci. 30, 1177–1189, 1995.

She, C.Y., Vance, J.D., Williams, B.P., Krueger, D.A., Moosuller, H.,Gibson-Wilde, D., Fritts, D.C. Lidar studies of atmospheric dynamicsnear polar mesopause, EOS Transactions. American GeophysicalUnion 83, 289–293, 2002.

Singer, W., Keuer, D., Eriksen, W. The ALOMAR MF radar: Technicaldesign and first results. In: Kaldeich-Schurmann, B. (Ed.), Proceedingof the 13th ESA Symp. On European Rocket and Balloon Pro-grammes and Related Research. Vol. ESA SP-397, pp. 101–104, 1997.

Singer, W., Bremer, J., Hocking, W.K., Weiss, J., Latteck, R., Zecha, M.Temperature and wind tides around the summer mesopause at middleand arctic latitudes. Adv. Space Res. 31 (9), 2055–2060, 2003.

Singer, W., Latteck, R., Friedrich, M., Dalin, P., Kirkwood, S., Engler,N., Holdsworth, D. D-region electron densities obtained by differentialabsorption and phase measurements with a 3-MHz-Doppler radar,Proceeding of the 17th ESA Symposium on European Rocket andBalloon Programmes and Related Research, Sandefjord, Norway, 30May–2 June 2005, ESA SP-590, pp. 233–237, 2005.

VanZandt, T.E., Nastrom, G.D., Furumoto, J., Tsuda, T., Clark, W.L. Adual-beam width radar method for measuring atmospheric turbulentkinetic energy. Geophys. Res. Lett. 29 (12), 13-1–13-3, 2002.

von Zahn, U. Achievements of ALOMAR. In: Kaldeich-Schurmann, B.(Ed.), Proceeding of the 13th ESA Symp. On European Rocket andBalloon Programmes and Related Research. Vol. ESA SP-397, pp.141–162, 1997.

von Zahn, U., von Cossart, G., Fiedler, J., Fricke, K.H., Nelke, G.,Baumgarten, G., Rees, D., Hauchcorne, A., Adolfsen, K. TheALOMAR Rayleigh/Mie/Raman lidar: objectives, configuration andperformance. Ann. Geophys. 18, 815–833, 2000.