July 1, 2008 / Vol. 33, No. 13 / OPTICS LETTERS 1485

High spatial and temporal resolution mobileincoherent Doppler lidar for sea surface

wind measurements

Zhi-Shen Liu,* Bing-Yi Liu, Song-Hua Wu, Zhi-Gang Li, and Zhang-Jun WangKey Laboratory of Ocean Remote Sensing, Ministry of Education of China,

Ocean University of China, 5 Yushan Road, Qingdao 266003, China*Corresponding author: zsliu@orsi.ouc.edu.cn

Received May 12, 2008; accepted May 22, 2008;posted May 29, 2008 (Doc. ID 96028); published June 26, 2008

A mobile Doppler lidar based on an injection-seeded diode-pumped Nd:YAG pulsed laser with a high repeti-tion rate was developed to measure the sea surface wind (SSW) with high spatial and temporal resolution.The system was operated during the 2007 Qingdao International Regatta to measure the distribution ofSSW in the racing area in real time with 50–100 m horizontal resolution and 2–10 min temporal resolution.An observation of nonuniform distribution of SSW is presented. The lidar results are compared with bothbuoy and wind tower measurements, which show good agreement. This lidar can be used advantageously forthe 2008 Olympic sailing games as well as for observing mesoscale and microscale meteorology processes.© 2008 Optical Society of America

OCIS codes: 010.3640, 120.0280, 010.3920.

A mobile Doppler wind lidar with the ability for themeasurement of sea surface wind (SSW) was devel-oped by the Key Laboratory of Ocean Remote Sensingof Ministry of Education of China, Ocean Universityof China, as shown in Fig. 1.

The unique feature of this lidar is its capability forscanning measurements with high spatial and tem-poral resolution. This is carried out by using a diode-pumped Nd:YAG pulsed laser with a high repetitionrate tunable up to 10 kHz. Considering the averagepower and the limitation of data acquisition, the la-ser is operated at 2.8 kHz (Table 1), yielding an aver-age power of 6 W. Consequently, enough laser shotsand photon counts can be integrated at each directionduring the scanning measurement. To obtain laserpulses in single longitudinal mode, the pulsed laser isinjection seeded by a cw seed laser by minimizing thepulse built-up time. A small part of the frequency-doubled pulsed laser radiation is reflected to afrequency-locking module (Fig. 2) that continuouslymonitors the frequency deviation and controls thetemperature and cavity length of the seed laser tolock the pulsed green laser radiation to the edge of aDoppler-broadened iodine absorption line. The ad-vantage of monitoring and stabilizing the pulsed la-ser frequency instead of seed laser frequency [1,2] isthat the frequency jitter of the pulsed laser can be de-tected and restrained in real time. The long-term fre-quency stability is controlled within 3 MHz. In addi-tion, an elevation-over-azimuth scanner achieves thedirectional transmitting and receiving of the laserbeam with a continuous scanning speed of 1° –10° /sand directional accuracy of 0.1°. A second iodine ab-sorption filter is utilized for discriminating the Dop-pler frequency shift from the backscattering light.Compared with the wind measurements from mo-lecular backscattering in high altitudes, the mea-surements using Doppler lidar based on an iodine ab-

sorption filter in aerosol backscattering dominated

0146-9592/08/131485-3/$15.00 ©

low altitudes has a higher sensitivity because of ahigh aerosol content [1,3], which enables the lidar tomeasure the SSW with a higher accuracy.

The line-of-sight (LOS) velocity at direction i canbe expressed as [2]

VLOS,i =RW,i − r0

r0S, �1�

where RW,i is the wind ratio at direction i defined asthe ratio of the photon counts in the measurementand reference channel; zero wind ratio r0 is the windratio when the LOS velocity is zero; and S is the sen-sitivity defined as the fractional change in the windratio per unit LOS velocity, S= �1/r0��dRW /dVLOS�.The value of r0 is obtained by measuring RW with afixed-elevation � at three directions in 120° azimuth

Fig. 1. (Color online) Mobile Doppler lidar during scan-

ning measurement (exposure time of photo: 16 s).

2008 Optical Society of America

1486 OPTICS LETTERS / Vol. 33, No. 13 / July 1, 2008

range with 60° azimuth intervals, denoted as RW,−60,RW,0, and RW,+60. Given that the wind and atmo-spheric state is uniform at each range circle r, r0�r�=RW,−60�r�+RW,+60�r�−RW,0�r� can be derived fromtrigonometry. After the aerosol backscattering ratioRb�r� is obtained at the direction of the RW,0 measure-ment, the value of S�r� can be derived [1]. During ourmeasurements for SSW, the value of Rb was usuallymore than 2 over the sea surface, so the value of Swas around 0.015–0.020. Then, VLOS at elevation �can be obtained from Eq. (1). Consequently, the two-dimensional wind field can be retrieved by the algo-rithm of velocity–azimuth processing under the as-sumption that the wind vectors of contiguousazimuths are uniform [4,5].

For comparison, both lidar and buoy data have tobe adjusted to a 10 m reference height above the seasurface. In the atmospheric surface layer, at a heightz, the wind velocity u�z� is expressed by [6]

u�z� =u*

k �ln� z

z0� − �m� z

L�� , �2�

where u* is friction wind velocity; k is Von Karman’sconstant (0.4); the roughness length z0=0.0185u*2 /g+ �� /u*�exp�−5.5k�, where � is the kinematic viscos-ity of air (1.56�10−5 m2/s at 25°C); and g is gravita-tional acceleration. �m�z /L� is a modification to log-arithmic profiles depending on the atmospheric sta-bility [7] and can be determined by the wind towermeasurements at different heights. During our mea-surements most values of �m�z /L� were between −1and 1. After �m�z /L� is obtained, with a measuredu�z� at the height of z by either lidar or buoy, u* canbe calculated from Eq. (2) with iterations. Therefore,both the lidar and buoy wind can be corrected to 10 m

Table 1. Main Parameters of the Lidar System

TransmitterWavelengtha 532 nmRepetition ratea 2.8 kHzPulse energya 2.2 mJPulse widtha 30 nsPulse laser linewidth �35 MHzSpectral purity �99.9%Far-field divergence (beam expanded)a 70 �radBeam pointing stability �5 �rad

ReceiverTelescope aperturea 305 mmField of view 120 �radInterference filter bandwidtha 0.11 nmInterference filter peak transmissiona 76%Sampling range resolution 10 m

ScannerAzimuth–elevation scanning speeda 1–10° /sAzimuth–elevation scanning accuracya 0.1°aThe parameter was measured.

reference height.

During the field campaign for the 2007 Qingdao In-ternational Regatta the mobile Doppler lidar was lo-cated at the south coast of Maidao Island (36.053°N,120.426°E) and measured at a 1.25° elevation froman altitude of 27 m above sea level. The azimuth cov-ered from 100° to 220° with a 2° resolution. One scanover 120° was achieved in 2 min with a continuousazimuth scanning speed of 1° /s. Four or five 2 minscans measured within 10 min were averaged to ob-tain the 10 min averaged SSW. The detection rangevaried from 3 to 8 km depending on the atmosphericconditions. Assuming Rb is 2, the expected uncer-tainty was typically 0.2–1.9 m/s, with the detectionrange of 1–6 km. The buoy was moored at 36.044°N,120.431°E within the lidar scanning area. The an-emometer of the buoy at 3 m altitude acquired10 min averaged wind continuously. In addition, thewind tower at Chidao Island (36.066°N, 120.462°E)established by Qingdao Meteorological Administra-tion measured wind at six heights between 2 and20 m. As shown in Fig. 3, an example of nonuniformdistribution of the SSW was observed by lidar at8:10–8:20 a.m. on August 22, 2007. The dominatingwind direction was northeast, and the wind speedvaried obviously from up to 10 m/s at the southeastarea to only 1–3 m/s in the rest of the measurementarea.

The lidar performed 273 two-minute scans of SSWsmainly in the daytime during August 15 to 23. Con-sequently, 63 sets of 10 min averaged SSWs were ob-tained. For comparison, the lidar wind at the buoyposition was corrected from a 48 m height to a 10 mreference height, while the buoy wind was correctedfrom 3 to 10 m. The mean values of the height correc-

Fig. 2. (Color online) Schematic of the mobile Doppler

lidar.

July 1, 2008 / Vol. 33, No. 13 / OPTICS LETTERS 1487

tion for the lidar and buoy wind were −0.59 and+0.39 m/s, respectively, which were comparable withthe results obtained by Daniault et al. [8] in 1988.The time series of the lidar, buoy, and wind towermeasurements are shown in Fig. 4. Since wind towerdata for only one 10 min average period were avail-able in every 60 min, the data closest in time were se-lected for comparison. As can be seen in Fig. 4, theresults agree very well.

The comparison between lidar and buoy wind re-sults in correlation coefficients �R2� of wind directionand speed of 0.85 and 0.87, respectively. The root-mean-square deviation of wind direction and speedare 20° and 1.1 m/s, which is partly due to the mea-surement uncertainty of both instruments (�0.2 m/sfor the lidar and �1 m/s for the buoy), the uncer-tainty of height correction, and the difference of av-eraging methods for lidar and buoy data. The biasesof lidar wind over buoy and wind tower data are+0.69 and −0.10 m/s, respectively. It indicates thatthe bias between the lidar and buoy wind probablyresults from the underestimated buoy data ratherthan lidar measurements. A possible reason is thatthe buoy data are undercorrected considering�m�z /L� is calculated from the wind tower data mea-sured �3.7 km away from the buoy position.

Fig. 3. (Color online) SSW measured with 1.25° elevationat 8:10–8:20 a.m. on August 22, 2007. The height is (a) in-dicated by the contour lines and (b) corrected to the refer-ence height of 10 m. Wind direction and speed with 50 mresolution are denoted by arrows and colors.

In conclusion, a new mobile Doppler lidar based ona high repetition rate diode-pumped laser was pre-sented for SSW measurements, and it was appliedduring the 2007 Qingdao International Regatta tomeasure the distribution of the SSW in real time forthe first time to the best of our knowledge. The non-uniform distribution of the SSW was observed withhigh spatial and temporal resolution. The lidar mea-surements with high accuracy matched well withboth buoy and wind tower data. The mobile Dopplerlidar is proved to be a reliable instrument for real-time SSW measurements and it can be used advan-tageously for the 2008 Olympic sailing games. It alsohas a great potential for observing and researchingmesoscale and microscale meteorology processes.

This work was supported by the National NaturalScience Foundation of China (NSFC) projects60578038, 40427001, and 40505003. We thank twoanonymous reviewers for providing helpful sugges-tions and comments.

References

1. Z. S. Liu, D. Wu, J. T. Liu, K. L. Zhang, W. B. Chen, X.Q. Song, J. W. Hair, and C. Y. She, Appl. Opt. 41, 7079(2002).

2. Z. S. Liu, B. Y. Liu, Z. G. Li, Z. A. Yan, S. H. Wu, and Z.B. Sun, Appl. Phys. B 88, 327 (2007).

3. C. Y. She, J. Yue, Z. A. Yan, J. W. Hair, J. J. Guo, S. H.Wu, and Z. S. Liu, Appl. Opt. 46, 4444 (2007).

4. Z. Y. Tao, Acta Metrologia Sinica 50, 81 (1992) (inChinese).

5. Y. G. Zheng, S. Y. Liu, J. Bai, and Z. Y. Tao, in 31stConference on Radar Meteorology (AmericanMeteorological Society, 2003), http://ams.confex.com/ams/pdfpapers/64095.pdf.

6. C. Paulson, J. Appl. Meteorol. 9, 857 (1970).7. W. Large and S. Pond, J. Phys. Oceanogr. 11, 324

(1981).8. N. Daniault, M. Champagne-Philippe, M. Camblan,

and J. Thépaut, J. Atmos. Ocean. Technol. 5, 385(1988).

Fig. 4. (Color online) Comparison of 10 min averaged SSWamong lidar, buoy, and wind tower from August 15 to 23,2007.