double-edge molecular measurement of lidar wind profiles at 355 nm

3
1466 OPTICS LETTERS / Vol. 25, No. 19 / October 1, 2000 Double-edge molecular measurement of lidar wind profiles at 355 nm Cristina Flesia* Groupe de Physique Appliquée, Université de Genève, 20, Rue de l’Ecole de Medecine, 1211 Geneva 4, Switzerland C. Laurence Korb Laboratory for Atmospheres, 912, NASA Goddard Space Flight Center, Greenbelt, Maryland 20771 Christian Hirt Groupe de Physique Appliquée, Université de Genève, 20, Rue de l’Ecole de Medecine, 1211 Geneva 4, Switzerland Received May 10, 2000 We built a direct-detection Doppler lidar based on the double-edge molecular technique and made the what we believe to be the first molecular-based wind measurements using the eye-safe 355-nm wavelength. Three etalon bandpasses are obtained with step etalons on a single pair of etalon plates. We eliminate long-term frequency drift of the laser and the capacitively stabilized etalon by locking the etalon to the laser fre- quency. We use a low-angle design to avoid polarization effects. Wind measurements of 1 2-ms accuracy are obtained to 10-km altitude with 5 mJ of laser energy, a 750-s integration, and a 25-cm telescope. Good agreement is obtained between lidar and rawinsonde measurements. © 2000 Optical Society of America OCIS codes: 280.3340, 280.3640, 120.2230, 120.0280, 300.6320, 010.1290. Wind is one of the basic atmospheric-state variables. It has long been recognized that global tropospheric wind measurements are required for understand- ing and predicting weather and climate on all time scales. 1 Despite its importance, there is no current global capability for measuring winds over the tro- posphere. We built a molecular-based double-edge system 2 at 355 nm as a follow-up to our work on single-edge aerosol systems. 3,4 The double-edge sys- tem is compact, uses the eye-safe 355-nm wavelength, and is designed as a spaceborne representative system. The NASA Zephyr Shuttle–based program selected this system. We made the atmospheric wind mea- surements presented here in March 1998 to provide an experimental demonstration of the feasibility of the technique for the Zephyr program. Direct-detection lidar techniques can use either aerosol 3–8 or molecular 2,3,9,10 backscatter for measuring the atmospheric wind field. Aerosol-based wind mea- surements can use high spectral resolution because the aerosol-backscattered spectrum is narrow with respect to the laser width and as a result has the same spectral width as the outgoing laser. Thus they offer the possibility of high-sensitivity measure- ments in areas of high aerosol backscatter. However, large regions of the Southern Hemisphere as well as mid-oceanic regions have low aerosol concentrations in the free troposphere. As a result, current aerosol sys- tems can provide only partial measurement capability. On the other hand, the molecular-backscattered spectrum is broad, which limits the sensitivity of the measurements. However, molecular scattering provides a dependable and reasonably uniform source of scattering on a global basis. This is particularly important for satellite-based wind measurements. We utilize a double-edge lidar technique to measure the wind, using the molecular signal backscattered from the atmosphere, as shown in Fig. 1. The molecular signal is spectrally broadened by Doppler shifts that are due to the random thermal motion of molecules and by Brillouin scattering. 11 The method is similar to the single-edge molecular method for determining the wind. 3 With the single edge we measure the Doppler shift of the spectrum by locating it on a moderately sharp edge of a high-spectral- resolution f ilter. Relatively large changes in signal are observed for small frequency shifts, owing to the steep slope of the edge. We sample a small portion of the outgoing beam to determine the frequency of the outgoing laser signal by measuring its location on the edge of the filter. Likewise, the frequency of the laser energy that is backscattered from the atmosphere is measured for each range element. The Doppler shift, and thus the wind, is determined from a differential measurement of the frequency of the outgoing laser pulse and the frequency of the laser re- turn backscattered from the atmosphere. This makes the measurement insensitive to the laser- and fil- ter-frequency jitter and drift. A detailed description Fig. 1. Double-edge measurement of Doppler shifts Dn of the Rayleigh –Brillouin profile with the laser at frequency n l and the two edge filters at frequencies n 1 and n 2 . 0146-9592/00/191466-03$15.00/0 © 2000 Optical Society of America

Upload: christian

Post on 05-Oct-2016

212 views

Category:

Documents


0 download

TRANSCRIPT

1466 OPTICS LETTERS / Vol. 25, No. 19 / October 1, 2000

Double-edge molecular measurement of lidar windprofiles at 355 nm

Cristina Flesia*

Groupe de Physique Appliquée, Université de Genève, 20, Rue de l’Ecole de Medecine, 1211 Geneva 4, Switzerland

C. Laurence Korb

Laboratory for Atmospheres, 912, NASA Goddard Space Flight Center, Greenbelt, Maryland 20771

Christian Hirt

Groupe de Physique Appliquée, Université de Genève, 20, Rue de l’Ecole de Medecine, 1211 Geneva 4, Switzerland

Received May 10, 2000

We built a direct-detection Doppler lidar based on the double-edge molecular technique and made the whatwe believe to be the first molecular-based wind measurements using the eye-safe 355-nm wavelength. Threeetalon bandpasses are obtained with step etalons on a single pair of etalon plates. We eliminate long-termfrequency drift of the laser and the capacitively stabilized etalon by locking the etalon to the laser fre-quency. We use a low-angle design to avoid polarization effects. Wind measurements of 1 2-m�s accuracyare obtained to 10-km altitude with 5 mJ of laser energy, a 750-s integration, and a 25-cm telescope. Goodagreement is obtained between lidar and rawinsonde measurements. © 2000 Optical Society of America

OCIS codes: 280.3340, 280.3640, 120.2230, 120.0280, 300.6320, 010.1290.

Wind is one of the basic atmospheric-state variables.It has long been recognized that global troposphericwind measurements are required for understand-ing and predicting weather and climate on all timescales.1 Despite its importance, there is no currentglobal capability for measuring winds over the tro-posphere. We built a molecular-based double-edgesystem2 at 355 nm as a follow-up to our work onsingle-edge aerosol systems.3,4 The double-edge sys-tem is compact, uses the eye-safe 355-nm wavelength,and is designed as a spaceborne representative system.The NASA Zephyr Shuttle–based program selectedthis system. We made the atmospheric wind mea-surements presented here in March 1998 to providean experimental demonstration of the feasibility of thetechnique for the Zephyr program.

Direct-detection lidar techniques can use eitheraerosol3 – 8 or molecular2,3,9,10 backscatter for measuringthe atmospheric wind field. Aerosol-based wind mea-surements can use high spectral resolution becausethe aerosol-backscattered spectrum is narrow withrespect to the laser width and as a result has thesame spectral width as the outgoing laser. Thusthey offer the possibility of high-sensitivity measure-ments in areas of high aerosol backscatter. However,large regions of the Southern Hemisphere as well asmid-oceanic regions have low aerosol concentrations inthe free troposphere. As a result, current aerosol sys-tems can provide only partial measurement capability.On the other hand, the molecular-backscatteredspectrum is broad, which limits the sensitivity ofthe measurements. However, molecular scatteringprovides a dependable and reasonably uniform sourceof scattering on a global basis. This is particularlyimportant for satellite-based wind measurements.

We utilize a double-edge lidar technique to measurethe wind, using the molecular signal backscatteredfrom the atmosphere, as shown in Fig. 1. Themolecular signal is spectrally broadened by Doppler

0146-9592/00/191466-03$15.00/0 ©

shifts that are due to the random thermal motion ofmolecules and by Brillouin scattering.11 The methodis similar to the single-edge molecular method fordetermining the wind.3 With the single edge wemeasure the Doppler shift of the spectrum by locatingit on a moderately sharp edge of a high-spectral-resolution f ilter. Relatively large changes in signalare observed for small frequency shifts, owing to thesteep slope of the edge. We sample a small portionof the outgoing beam to determine the frequency ofthe outgoing laser signal by measuring its locationon the edge of the f ilter. Likewise, the frequencyof the laser energy that is backscattered from theatmosphere is measured for each range element. TheDoppler shift, and thus the wind, is determined froma differential measurement of the frequency of theoutgoing laser pulse and the frequency of the laser re-turn backscattered from the atmosphere. This makesthe measurement insensitive to the laser- and fil-ter-frequency jitter and drift. A detailed description

Fig. 1. Double-edge measurement of Doppler shifts Dn ofthe Rayleigh–Brillouin profile with the laser at frequencynl and the two edge f ilters at frequencies n1 and n2.

2000 Optical Society of America

October 1, 2000 / Vol. 25, No. 19 / OPTICS LETTERS 1467

of the theory for the single-edge method is given inRef. 3.

We recently described double-edge lidar techniquesfor measuring the wind by use of the signal backscat-tered from either molecules2 or aerosols8 in theatmosphere. With these techniques we use a secondedge measurement for signal normalization insteadof an energy-monitor measurement as is used in thesingle-edge method. This second edge measurementdoubles the signal change per unit Doppler shift andthe sensitivity of the measurement, and it leads to animprovement of almost a factor of 2 in the measure-ment accuracy.

The edge-filter measurement generally depends onthe magnitude of the atmospheric backscattered molec-ular, Rayleigh–Brillouin, and aerosol signals. Thusin general the molecular and the aerosol signals haveto be spectrally separated, measured separately, andtreated independently in the analysis. We can, how-ever, greatly simplify the measurement by locating theedge-filter measurement in a crossover region wherethe fractional change in the measured molecular andthe aerosol signals are equal for a given frequencyshift. That is,2

1Ri�ni, nl�

ddn

Ri�ni,nl� �1

Ti�nl 2 ni�d

dnTi�nl 2 ni�

(1)

for each edge filter, where

Ri�ni, nl 1 Dn� �Z `

2`

Ti�n 2 ni�IR �n 2 nl 2 Dn�dn .

(2)

In Eqs. (1) and (2), Ti is the transmission of theith edge filter centered at ni for the laser, IR is thebackscattered Rayleigh–Brillouin spectrum, nl isthe laser frequency, and Dn is the Doppler shift.Thus the measurement sensitivities are equal for themolecular and the aerosol portions of the signal. Inthis case the aerosol signal acts in a manner identicalto the molecular signal, and the measurement isdesensitized to the effects of aerosol scattering.2

The detector package layout is shown in Fig. 2.Frequency tripling a single-longitudinal-mode Nd:YAGlaser provides the required 355-nm output. The laserenergy is 80 mJ�pulse at 355 nm, with a 30-Hzrepetition rate. The laser has an angular output of0.5 mrad (see Table 1). A small portion of the outgo-ing laser signal is picked off and sent to the telescopeand from there travels through the optical system tothe etalon, allowing the outgoing laser frequency to bemeasured as described below.

The major portion of the laser energy is sent to theatmosphere. A portion of this energy is backscatteredfrom each range element of the atmosphere. The sig-nal is collected by a 25-cm-diameter telescope. It isfocused into a multimode fiber-optic cable that sets thefield of view of the telescope at 0.125 mrad. The sig-nal from the fiber is collimated and directed into aFabry–Perot etalon, which serves as the double-edgefilter. The laser energy of 80 mJ�pulse is too largefor tropospheric measurements even for our relatively

small 25-cm-diameter telescope. As a result we utilizethe mismatch between the 0.125-mrad field of view ofthe telescope and the 0.5-mrad divergence of the laserto reduce the laser energy to 5 mJ�pulse. We notethat, even for this reduced laser energy, saturation ef-fects still occur for altitudes less than 2 km.

The high-resolution etalon is capacitively stabilized.It is composed of three bandpasses. One bandpassuses a circular region on the two basic etalon plates,and the two other bandpasses use coatings with athickness of a fraction of l�2 on circular regions of oneplate. These coatings form step etalons that shiftthe frequency of the basic bandpass by a fraction ofa free spectral range. Two of these bandpasses areused as double-edge etalons for the atmospheric mea-surements, as shown in Fig. 1. The third bandpassis used as a locking etalon. It is located with onehalf-width point, HWHM, centered between the twoedge etalons. We use a servo-control system to lockthe half-width point of the locking etalon to the pulsedNd:YAG laser frequency to remove laser and etalonfrequency drift. The outgoing laser wavelength canbe measured with high precision, since the half-widthis the point of maximum sensitivity. A low-angledesign with a 5± incidence angle is used to split the sig-nal among the two edge channels, the locking channel,and the energy-moniter channel to eliminate polariza-tion effects. The etalons have a spectral bandwidthof 1.56 GHz, FWHM. The double-edge etalons arelocated about the laser frequency with a separation of62.60 GHz. This corresponds to the crossover pointwhere the sensitivity, the percent change in signalper meter per second, is the same for the molecularand the aerosol signals. These parameters also allowus to obtain the minimum measurement error whilesimultaneously achieving aerosol desensitization.2

High-speed photomultiplier detectors with a quan-tum efficiency of 28% are used in a photon-countingmode for the edge and in an analog mode for theenergy-monitor and the locking-channel detectors.The photon-counting signals are collected in a multi-channel scalar and integrated over a variable numberof shots. We typically used integration times of 10 s,which correspond to 300 shots at 30 Hz. The systemrecords the signal for each of the two atmospheric-edgeetalon channels averaged over n shots as a function of

Fig. 2. Optical layout of the detector package: EM, en-ergy monitor; PMT’s, photomultiplier detectors; BS’s, beamsplitters.

1468 OPTICS LETTERS / Vol. 25, No. 19 / October 1, 2000

Table 1. Experiment Parameters

Laser wavelength (nm) 354.7Laser energy per pulse (mJ) 80Pulse repetition rate (Hz) 30Laser divergence (mrad)̀ 0.5Telescope diameter (cm) 25Telescope field of view (mrad) 0.125Effective laser pulse energy (mJ) 5Etalon plate spacing (cm) 1.25Effective finesse 7.7Etalon spectral bandwidth (GHz) 1.56Number of etalon channels 3Laser etalon separation-locking

channel (GHz) 0.78Laser etalon separation atmospheric

channels (GHz) 62.605

Fig. 3. Comparison of double-edge lidar wind measure-ments with 5-mJ laser energy, with standard deviationshown by the bars, and rawinsondes for (a) March 27, 1998,with 22,500 shots and (b) March 29, 1998, with 10,000shots.

altitude. To measure the wind-speed profile alongthe line of sight in a given direction we perform onemeasurement in the chosen direction and anotheralong the vertical as a calibration. We first subtractthe background for each measurement. We thencalculate the ratio between the detected photons ineach edge channel for a given altitude. The windprofile is found as

n �1u

µR 2 Rv

Rv

∂, (3)

where R and Rv represent the ratio of the two edgesfor each altitude in the chosen or vertical direction,respectively, and u is the sensitivity of the system. We

find the standard deviation for each altitude by use ofmeasurements over several data sets at each altitude.

Figures 3(a) and 3(b) show a comparison of ourline-of-sight lidar wind-profile measurements made atthe University of Geneva at 11:00 PM on March 27,1998, and at 7:00 PM on March 29, 1998, respectively,with rawinsonde profile wind measurements madeby the Swiss Meteorological Service at Payerne,Switzerland, at a distance of 50 km. In Fig. 3, therawinsonde wind profiles are projected onto the lineof sight of the lidar measurements. We made ourlidar measurements at an azimuth angle of 53.7± andat zenith angles of 30± and 0±. We did not makemeasurements in a second orthogonal direction, sincethere were nearby buildings. For Fig. 3(a) [3(b)] thestandard deviation was calculated with an averageover 750 (333) s and 22,500 (10,000) shots, with aneffective laser energy of 5 mJ�shot. This energyis equivalent to an average over 47 (21) s and 1400(625) shots, with the full laser energy of 80 mJ�shot.The standard deviation of the lidar measurement ateach altitude is shown by a horizontal bar and variesfrom 1 to 2 m�s in Fig. 3(a). At 10-km altitude, thisis within a factor of 2.5 of the shot-noise limit. Asshown in the f igures, the lidar wind measurementsare in good agreement with the rawinsonde data.

We built a direct-detection Doppler lidar based onthe double-edge molecular technique and made whatwe believe are the first molecular-based wind measure-ments with the eye-safe 355-nm wavelength. The dif-ferential frequency method and the use of a lockingchannel produce a feasible, robust system. Wind mea-surements are obtained to greater than 10-km altitudewith a small system that uses 5 mJ of laser energy.We recently successfully participated in a field cam-paign in which a portable version of this system wasused and will report on this in the future.

C. L. Korb’s e-mail address is [email protected].

*Present address, Space Section, Centro di Eccel-lenza Optronica, Largo Enrico Fermi 6, 50125 Arcetri,Florence, Italy.

References

1. R. Atlas, E. Kalnay, and M. Halem, Opt. Eng. 24, 341(1985).

2. C. Flesia and C. L. Korb, Appl. Opt. 38, 432 (1999).3. C. L. Korb, B. Gentry, and C. Weng, Appl. Opt. 31,

4202 (1992).4. C. L. Korb, B. Gentry, and S. X. Li, Appl. Opt. 36, 5976

(1997).5. V. J. Abreu, Appl. Opt. 18, 2992 (1979).6. M. J. McGill, W. R. Skinner, and T. D. Irgang, Appl.

Opt. 36, 1253 (1997).7. D. Rees and I. S. McDermid, Appl. Opt. 29, 4133 (1990).8. C. L. Korb, B. M. Gentry, S. X. Li, and C. Flesia, Appl.

Opt. 37, 3097 (1998).9. M. L. Chanin, A. Garnier, A. Hauchecorne, and J.

Porteneuve, Geophys. Res. Lett. 16, 1273 (1989).10. A. Garnier and M. L. Chanin, Appl. Phys. B 55, 35

(1992).11. G. Tenti, C. D. Boley, and R. D. Desai, Can. J. Phys.

52, 285 (1974).