Sensitivity of coherent range-resolved differentialabsorption lidar
Takayasu Fukuda, Yoshio Matsuura, and Tadatami Mori
A range-resolved DIAL (differential absorption lidar) system with heterodyne detection has been developed.A hybrid TEA CO2 laser was employed as the transmitter oscillator, which emitted single-frequency pulsesof 140 mJ. The heterodyne receiver, which had a minimum detectable power of 2 X 10-11 W, could detectthe echo signals backscattered from atmospheric aerosols at a 5-km or greater range. The system sensitivityto the target gas, defined as the product of the minimum detectable concentration and the difference in theabsorption coefficients, was experimentally found to be 3.7 X 10-4 m-1 for a range resolution of 300 m afteraveraging over fifty backscattered signals.
lidar technique offers an alternative to in situ sensorsfor the measurement of gaseous constituents in the at-mosphere. RRDA lidar systems use a pulsed lasertransmitter with atmospheric aerosol particles as thebackscattering medium to perform the range-resolvedmeasurements. Because of its range resolution, thistechnique has been attracting much attention and re-cently systems with different types of laser have beenreported. 1 -3
Among these, the RRDA lidar with CO2 laser hasadvantages in weather penetration capability and ver-satility; in the wavelength region of lasing transition ofCO2 there are absorption lines from a number of gasspecies, 03, H20, NH3, and so forth. In spite of theseadvantages, few such systems have been reported, be-cause the scattering cross section of aerosols is extremelysmall and a direct-detection system would require atransmitter oscillator of high energy, several joules, forexample.3 In this wavelength region, however, sensi-tivity with heterodyne (coherent) detection can be mademuch higher than with direct detection. Kobayasi andInaba4 proposed the RRDA lidar with heterodyne de-tection, which has high range performance with a low
The authors are with Fujitsu System Integration Laboratories, Ltd.,Electronics System Department, 1333, Kamikodanaka, Nakahara-ku,Kawasaki 211, Japan.
energy laser. Although their proposal seemed prom-ising, technical difficulties have prevented its realizationso far.
As the minimum detectable power (MDP) of a het-erodyne receiver is proportional to the IF bandwidth,the optical bandwidth of the transmitted laser pulsemust be narrow and its center frequency must be stable.Because of a wide (1-GHz) bandwidth, the conven-tional TEA CO2 laser, with all its high energy output,is not suitable for use as the transmitter oscillator of acoherent lidar. Even if we could obtain a highly co-herent transmitter, the narrow bandwidth of its outputpulse, which is comparable with that of the receiver,would cause another problem: return signal fluctua-tion, 5 which is referred to as speckle noise. Since theaccuracy of the gas concentration inferred from RRDAmeasurements is impaired by errors in determining theamplitude of the signal, speckle fluctuation, dominatingthe errors in the large signal limit,6 should be suppressedto improve the system sensitivity to the target gas.
We have developed a coherent RRDA lidar whichdeals with these problems. A hybrid TEA CO2 laserwas employed as the highly coherent transmitter.Among the various mode-selection methods, we choseto use the hybrid approach because it is reliable andbecause it provides a continuous reference laser outputfor adjusting the output pulse frequency. To reducethe speckle fluctuation in return signals, we incorpo-rated interpulse averaging of signals.
In this paper we shall describe an experimental studyof the achieved performance of a newly developed co-herent DIAL which includes range capability, hetero-dyne efficiency, and detection sensitivity to the targetgas. This seems to be the first work which has given anexperimental basis to a sensitivity analysis of coherentRRDA lidar. 6 ,7
11. Detection Sensitivity of Coherent RRDA LidarIn the RRDA technique, echo signal measurements
are made at two wavelengths Xo and X1, respectively, offand on the absorption resonance of the target gas. Thegas concentration C (R) at range R is deduced from themeasured received power using the following equa-tion:
where we assumed that samples of Pr are not correlatedwith each other.8 Granted that the range dependenceof Pr is not significant, Eq. (1) can be reduced to
2 1 var(Pr) (2)Oc (Ak)2 (A1?)2 [(Pr) 2 2
In Eq. (2), the ratio (Pr)/[var(Pr)] 1/2 is a measure of therelative amplitude accuracy. We refer to this ratio asinverse relative root variance of the measured receivedpower [IRRV(Pr)] in accordance with Jakeman etal.5
For a coherent lidar, IRRV of the IF output,IRRV(PIF), is given as9
IRRV(PIF) = CNR/2 11/2[1 + CNR/2IRRV + (2CNR)-1j
where the carrier-to-noise ratio (CNR),CNR= (P,)/MDP,
(3)
is the ratio of the mean signal and mean noise contri-butions of IF output, and IRRVo is the IRRV of thesignal component. In the large signal limit (CNR > 5),Eq. (3) is reduced to IRRV(PIF) IRRVO, and specklefluctuation dominates IRRVO, which is VIM7 if the re-ceiver views M coherent areas. Because the video signalextracted from the IF output is smoothed by the videoamplifier, IRRV(Pr) for the single-pulse echo, IRRV1,is given as
IRRV1 = VMB 1T, (5)
where BI = intensity fluctuation bandwidth, and T =integration time of the video amplifier. If the echoesare not correlated with each other, by averaging theirN video signals, IRRV(Pr) will be improved to
IRRV(P,) = VN IRRV 1 = NMBT. (6)
Thus, the standard deviation of the deduced gas con-centration is given as o = 1/(AkAR MNBiT).Strictly speaking, the experimental estimation of -should be done from samples of C(R) for a particular
TRANSMTTER OSCLLATOR
iI w BF Bnid
i
DETECTOR B.PF RF DET. VIDEOASSEMBLY AMP.
Fig. 1. RRDA lidar with heterodyne detection.
DESKTOPCOMPUTER
range. Since it would take a long time to derive a, withaccuracy in this manner, we calculated it from thesamples of C in one distribution profile, assuming thatIRRV was independent of range.
The value Ak o- is an estimate of the measurementaccuracy which is independent of the target gas species.We will refer to this as the system sensitivity S, thatis,
S Akcr = 1/(ARgMNB 1 T). (7)
Similarly we will define the generalized gas concentra-tion Co as Ak C. The dimension of these quantities, bydefinition, is (m- 1 ).
In a heterodyne receiver, MDP in Eq. (4) is givenas
MDP= hBIF77D 71sys
(8)
where hv = photon energy,BIF = IF bandwidth,'1 D = heterodyne equivalent quantum
efficiency of the detector, andwqys = efficiency of the receiver optics.
The antenna efficiency in heterodyne reception of in-coherent backscatter signals is incorporated in .sys
11. System DescriptionA block diagram of the experimental coherent DIAL
system is shown in Fig. 1. The system consisted of thetransmitter oscillator, detectors, local oscillator (LO),desktop computer, transmitter frequency controller,and transmitting and receiving antennas.
The transmitter oscillator was a hybrid TEA CO2laser consisting of a TEA gain section and a low-pres-sure gain section in a common optical cavity formedfrom a Littrow mounted concave grating and a flatoutput coupler. The output coupler was mounted ona piezoelectric translator. The transmitter oscillatorproduced a single-frequency output with a cw outputpower of -1 W and a pulsed output energy of 140mJ/pulse at 5 pps.
The detectors, one for the receiver and the other forthe frequency controller, were wide-bandwidth HgCdTe
photodiodes1 0 whose performance is shown in Table I.The received signal was mixed with LO output at thereceiver photodiode, and the IF signal was fed to apreamplifier through a bias insertion unit. The pho-todiode and the preamplifier were assembled togetherin a shielded housing so that noise caused by the TEAgain section would not interfere with the low level IFsignal.
The local oscillator was a conventional cw laser whosecooling water temperature was controlled to an accuracyof 0.20C to stabilize the LO frequency. The LO outputbeam was expanded to -1 cm in diameter, which wasequal to the diameter of the receiving telescope exitpupil.
The IF signal regulated by a step attenuator was de-modulated by the linear envelope detector to extract thevideo signal. The 8-bit A-D converter (transient re-corder) sampled and digitized the video signal andstored it in its memory. The digitized samples of thereturn signal were transferred to the computer by directmemory access (DMA). The computer integrated thesignals and processed them to display a backscatteredsignal waveform or a gas distribution profile on a CRT.The computer made a hard copy of the CRT display.
A portion of the transmitter cw output and the LOoutput was mixed at the other detector, and the resul-tant IF signals were fed to the frequency controller,where a frequency discriminator detected the deviationof the beat frequency from a set frequency (12 MHz),and the error signal was fed back to the piezoelectrictranslator on which the output coupler of the trans-mitter was mounted. The TEA gain section wastriggered at the very moment when the fluctuating IFcoincided with the set point. Although a rapid (1-kHz) fluctuation of the cw output frequency was causedby mechanical vibration of the transmitter laser reso-nator, this triggering technique adequately reduced itseffect on the pulse output frequency.
The transmitting and receiving antennas were twoidentical folded off-axis beam expanders, 15 cm in di-ameter, which expanded the output beam and con-densed the collected echo signal beam by a factor of 15.
The field of view of the system could be steered by amirror to aim at the target. The field of view of thereceiver and the transmitted beam could be crossed atany range from -100 m to infinity.
The pulse repetition rate, IF bandwidth, IF signalattenuation, sampling rate, and number of integrationscould be set through the computer keyboard. Laser lineselection of the transmitter and local oscillator wasperformed manually and took -1 min.
IV. Performance
A. Transmitter OutputFigure 2(a) is a typical waveform of the transmitter
output pulse with duration of -250 nsec. The wave-form was observed by a HgCdTe photodiode whosebandwidth was -250 MHz, and seems to have no modebeating.
Figure 2(b) shows the output pulse frequency char-acteristics observed in an IF waveform of the returnpulse from a hard target. The optical frequency of theoutput pulse was precisely controlled to the set pointby the technique described above. Pulse-to-pulse de-viation of the optical frequency was within -0.5 MHz.No intrapulse chirp was found for 1 ,usec from a pulsebuildup.
Neither beam profile nor beam divergence for pulseoperation was measured, but they can be estimatedfrom those obtained in cw operation. The beam profileof the cw laser output at the telescope aperture is shownin Fig. 3. The Gaussian-like beam profiles suggest thatthe transmitter laser oscillates on the lowest transverseresonator mode. Defining beam radius wo as the dis-tance at which the power is i/e2 times that on the axis,
(a) (b)Fig. 2. Pulse shape and frequency characteristics: (a) transmitter
output pulse, (b) received pulse at IF amplifier output.
P
Z)n
.EW
XA
Cs
-10 -8 -6 -4 -2 2 4 6 8 10
RADIAL STANCE (cm)
Fig. 3. Beam power profile of the cw output at the telescope aper-ture; wo is l/e 2 power radius.
Fig. 4. Backscattered signal vs range; N = 50, Bid= 1.5 MHz, =3 X 10-8 m-1 sr-1 . The sharp peak is an echo signal from a hillside
at -9 km.
the diameters or 2wo were estimated to be 8.0 and 7.1cm from the horizontal and vertical scan profiles, re-spectively. The difference in wo may be due to theaberration caused by the Littrow mounted concavegrating. Beam divergence is estimated to be 0.19 mrad,assuming that the beam waist is at the aperture.
B. System Efficiency DeterminationThe heterodyne receiver was calibrated to determine
the system MDP. For this calibration, we used the cwoutput of the hybrid TEA laser, because the availableCNR is moderate and optical attenuators, which wouldcause measurement errors, are not required. Targetswere sandblasted aluminum plates, 50 cm in diameter,whose reflection coefficient p had been determined inadvance of the field test. The reflected power Pr inci-dent on the receiving aperture can be estimated usingthe following equation:
Pr r exp(-2aR),7rR 2
where Pcw = transmitted cw power,Ar = receiver aperture area,R = target range, anda = extinction coefficient.
We evaluated a from the relative humidity and thetemperature at that time.1 The received signals re-turned from the target were integrated for 100 sec andcompared with the noise level to obtain the CNR. Thesystem MDP and the overall efficiency 71D sys weredetermined from the estimated Pr and measured CNR.For BIF = 10 MHz, MDP was determined to be 1.0-2.0X 10-11 W. The value varied depending on the heter-odyne efficiency of the detector employed or 71D. Thelowest value, 1.0 X 10-11 W for qD = 0.3, correspondsto ijsys = 0.063. We calibrated the receiver time aftertime to correct the system MDP value which was usedto derive Pr from the CNR. The scatter in the mea-sured MDP values was 10% at the most.
We measured IRRV of the signal returned from thetarget at the 430-nm range to estimate the spatial av-eraging factor M. In this case, the fluctuation band-width was <1 kHz, and the postdetection integrationhad no effect. So, it follows that IRRV1 = \/1A. Frommeasured IRRV values, which ranged from 0.80 to 1.05,we concluded that M was unity and that the effect of
spatial averaging was negligible at any range greaterthan -500 m. This conclusion also means that thereis no need to take account of the range dependence ofthe antenna efficiency in heterodyne reception of in-coherent backscattered signals.C. Aerosol Backscatter Measurement and AvailableCNR
A waveform of the averaged echo returned from at-mospheric aerosols is shown in Fig. 4. To obtain thiswaveform, fifty echoes were integrated. The dashedline is the MDP of the system as determined by thecalibration described above. The sharp peak was anecho signal returned from a hillside -9 km from thelidar.
Such experiments were repeated at all seasons todetermine the maximum useful range of the lidar.When a was not extremely high (a S 0.4 km-'), theCNR at the 5-km range was higher than 7 dB even forthe lowest volume backscatter coefficient ( = 3 X 10-8m-1 sr-'). Since a is usually smaller than 0.4 km-1 fora clear atmosphere, we can properly conclude that themaximum range Rmax of our lidar is 5 km. In mid-summer, a sometimes exceeded 0.4 km-', which de-graded Rmax. The highest a obtained from the mete-orological measurement was -0.7 km-1 , which reducedRmax to -3 km. When the visibility was better than 3km, the aerosol extinction was negligible, and the rangecapability of the lidar was not affected by it.
We estimated 3 using the coherent lidar at an altitudelower than -30 m where it will be in practical use. Asthe lidar receiver had been calibrated, /3 was derivedfrom the measured CNR value using the followingequation:
1 MDP * CNR E exp(2aR),cEtAr
where Et is the pulse energy and c is the light ve-locity.
Figure 5 shows a typical / distribution and a back-scattered signal from which was derived. Because the
-40 1
1982 FEB04.10:43
-60 -
-70 -
-80. 1-6
1-7
108 -8
10 0 1 2 3 4 5
RANGE km)
Fig. 5. Backscattered signal and derived distribution; N = 30, Bid= 1.5 MHz.
Fig. 7. Histogram of the normalized re-ceived power; for N = 1, IRRV is 1.8. Byintegrating fifty echoes, IRRV is improved to
-12. Bd = 1.5 MHz.
Fig. 8.
2 5 10 20 50 100
NUMBER OF PULSES INTEGRATED N
IRRV improvement by integration;Bvid = 1.5 MHz, PRF = 5 Hz.
field of view of the receiver and the section of thetransmitted beam did not overlap completely at closerange, the estimated : values are too small. So, /3samples at ranges >2 km were averaged to get : at thattime. The results, ranging from 3 X 10-8 to 8 X 10-7m-1 sr- 1, are consistent with data integrated byPost.1 2
D. Speckle Suppression and System SensitivityThe system sensitivity to the target gas is dominated
by the accuracy of the backscattered signal measure-ments. We measured IRRV of the received power andrelated it to the detection sensitivity.
The echo of a single pulse fluctuates greatly as shownby the solid curve in Fig. 6. The dotted curve shows abackscattered signal averaged by integrating fiftyechoes. The strong fluctuation was reduced by inter-pulse averaging, and the accuracy of the measurementwas improved as shown in Fig. 7, which shows that anIRRV of -1.8 for a single-pulse echo increases to 12when fifty echoes are integrated.
The dependence of IRRV(Pr) on the number ofpulses integrated, N, is shown in Fig. 8. The solid linerepresents the dependence predicted by Eq. (6). Thefigure shows that the backscatters from aerosols are notcorrelated with each other for a pulse rate of 5 pps. Onan average, IRRV(Pr) is -12 for N 50, which corre-sponds to 2.8 X 10-4 m- 1 for S for AR = 300 m accord-ing to Eq. (7). By curve fitting, the IRRV for the sin-gle-pulse echo, IRRVI, is determined to be 1.6 when Bvid= 1.5 MHz. For Bvid = 0.55 MHz, IRRVI was measuredas 2.6. The results agree with the values obtained byusing Eq. (5) from the system parameters, when we as-sume that BIT in the equation can be estimated by(rpBvid)Y1 and that M = 1.
The system sensitivity to the target gas S was deter-mined by carrying out RRDA measurement withouttarget gas. Figure 9 shows a derived gas concentrationprofile. By averaging the values, S was estimated to be-3.7 X 10-4 m- 1 for AR = 300 m and 7.4 X 10-4 m- 1 forAR = 150 m when N = 50 and PRF = 5 pps.
In Fig. 10, the measured values of S are comparedwith the values evaluated from IRRVI using Eq. (7),which are represented by the solid lines. For N smaller
than 30, the sensitivity is dominated by the residualspeckle and agrees with the calculated value. For alarger N, the measured S departs from N- 1 12 behaviorwhile IRRV(Pr) is proportional to \W even for N > 30.And the scatter in S values for N = 50 is larger than thatin the corresponding IRRV values. This means that theadditional error was caused by temporal changes in thelight path condition which become significant duringthe time interval between the backscatter measure-ments for the two wavelengths. Figure 11 shows anextreme case where a large false indication appears atthe 2.8-km range. At altitudes lower than 30 m, back-scatter signals were subject to such irregularities, whichwere presumably attributed to backscatter from aero-sols of urban origin and to extinction by pollutant gasesor water vapor in stack plumes.
The RRDA technique in itself should have eliminatedthe errors caused by the irregularities in backscatter, butthe measurement interval seems to have been too long.If the interval is sufficiently shortened, such additionalerrors can be reduced, as shown in Fig. 12. The returnsignal waveforms were obtained from two sets of fiftyechoes which were measured successively; the intervalwas 10 sec. The waveforms agree with each other inspite of their strong irregularities, and S was not clearlydegraded. In RRDA measurement, however, laser lineselection and transmitter frequency control require acertain period of time, and it remains to be shortened.The simultaneous or alternating measurement of twofrequency backscatters would be advantageous, al-though the lidar must have two transmitter lasers andLOs to do this.
E. Gas Detection ExperimentThe range resolution of the system was determined
by detecting a gas in a sample chamber. It could not beshown whether the measured values of Co were consis-tent with the gas concentration in the chamber, becauseits windows were removed to avoid the extremely strongbackscatter from them.
The chamber is a wooden box 1.8 X 0.9 X 0.9 m lo-cated -1100 m from the lidar. After assuring that thesidewall of the chamber did not scatter the transmittedlaser pulse, target gas (Freon 12) was loaded into the
Fig. 9. Noise in a gas concentration profile;N = 50, Bid = 1.5 MHz, AR = 300 nm. Thestandard deviation of the noise or the sensi-tivity S is 3.6 X 0-4 m-1. The sharp peaksat 3 km are echoes returned from a roof whichwas in the field of view when the gas chamber
was aimed at.
6. 10
2-r
10
'> 5_
10 20 50NUMBER OF PULSES INTEGRATED, N
-50 a.ix
0a.
0
-60
-70
'Eo.
o:
100
Fig. 10. Standard deviation of noise in themeasured concentration or the sensitivity Svs the number of pulses integrated N; Bvid =1.5 MHz, PRF = 5 Hz. Solid lines are S
evaluated from IRRV1 using Eq. (6).
-80
.
0-
0 1 2RANGE (km)
3 4
Fig. 11. False alarm caused by the back-scatter irregularities; N = 50. The mea-
surement interval was -2 min.
chamber. Then the echo signals were measured using10.6-,um P(24) and P(34) lines and the RRDA reductionwas performed to get the gas distribution profile.Figure 13 shows typical waveforms of echo signals anda reduced distribution profile, which definitely indicatesthe presence of the target gas. The system could locatethe gas with accuracy; the range to the peak of the in-dication (1200 m) agrees with the range to the gaschamber. The range resolution or the full width athalf-maximum concentration was 300 m, which agreeswith the range interval AR in RRDA reduction.
The hardware limit of the range resolution (ARm),defined as the length of a range segment in which Prdecreases steeply due to the localized gas as shown inFig. 13, was determined to be 200 m for Bvid = 1.5 MHz,270 m for 0.55 MHz. The range resolution was domi-nated by AR as long as AR was >ARm.
V. ConclusionThe operation of a RRDA lidar with a heterodyne
receiver has been demonstrated and its feasibility wasexperimentally confirmed. The maximum range wasextended by employing heterodyne detection. Inconditions of the usual extinction coefficient (a < 0.4km-'), the system was capable of detecting signalsbackscattered at an -5-km range with a CNR of -7 dBat least, which is enough to perform RRDA reduction.Even in the rather high extinction conditions (a 0.6km-'), the maximum range was -3 km.
Speckle fluctuation caused by the use of a highly co-herent transmitting laser was reduced by integratingreturn signals to get enough IRRV(Pr). With fifty in-tegrations, IRRV(Pr) reached -12, which correspondsto a sensitivity S of 2.8 X 10-4 m-1 when the rangeresolution was set for 300 m. The experiment showedthat fluctuation in the distribution of a and / was an-other source of error in the derived gas concentration.
-40
- -50In
IxrL -60-
0a.
W _70
UJ
-80 -
-90 L0 1 2 3 4 5
RANGE (km)
Fig. 12. Elimination of the 13 nonhomogeneity effect; N = 50.Without changing the laser line, 100 pulses were measured succes-
sively. The sensitivity S is estimated to be 5.0 X 10-4 m-1 .
-40
E
x)2
To
Or
-50
-60
-70
-804
'0o1
2
0
-2
46o
t i
RANGE (km)
Fig. 13. Gas concentration profile. The range to the chamber is-1100 m. AR = 300 m, Bvid = 0.55 MHz, N = 50, Ak = 4.0 X 10-3
Because of it, S derived from RRDA measurementsincreased to 3.7 X 10-4 m-1 for AR = 300 m. If the in-terval between measurements for two wavelengths isshortened, detection sensitivity would improve. Therange resolution of the system was limited to -200 m.Within this limit, the range resolution was dominatedby AR or Bvid-
From these results we have concluded that RRDAlidar with heterodyne detection is the most promisingscheme for a long-range gas monitoring system.
The authors gratefully acknowledge the technicalassistance provided by M. Tajima and H. Kashiwara.References
1. E. V. Browell, T. D. Wilkerson, and T. J. McIlrath, "Water VaporDifferential Absorption Lidar Development and Evaluation,"Appl. Opt. 18, 3474 (1979).
2. W. B. Grant, R. D. Hake, Jr., E. M. Liston, R. C. Robbins, and E.K. Proktor, Jr., "Calibrated Remote Measurement of NO2 usingthe Differential-Absorption Backscatter Technique," Appl. Phys.Lett. 24, 550 (1974).
3. K. Asai, T. Itabe, and T. Igarashi, "Range-Resolved Measure-ments of Atmospheric Ozone using a Differential-Absorption CO2Laser Radar," Appl. Phys. Lett. 35, 60 (1979).
4. T. Kobayasi and H. Inaba, "Infrared Laser Radar TechniqueHeterodyne Detection for Range-Resolved Sensing of Air Pol-lutants," Opt. Quantum Electron. 7, 319 (1975).
5. E. Jakeman, C. J. Oliver, and E. R. Pike, "Optical HomodyneDetection," Adv. Phys. 24, 349 (1975).
6. P. Brockman, R. V. Hess, L. D. Staton, and C. H. Bair, "DIALwith Heterodyne Detection Including Speckle Noise: Aircraft/Shuttle Measurement of 03, H20, and NH3 with Pulse TunableCO 2 Lasers," NASA CP-2138, International Conference onHeterodyne Systems and Technology, Williamsburg, Va. (Mar.1980).
7. R. M. Hardesty, "A Comparison of Heterodyne and Direct De-tection CO2 DIAL Systems for Ground-Based Humidity Profil-ing," NOAA Technical Memorandum ERL WPL-64 (1980).
8. B. J. Rye, "Differential Absorption Lidar System Sensitivity withHeterodyne Reception," Appl. Opt. 17, 3862 (1978).
9. J. H. Shapiro, B. A. Capron, and R. C. Harney, "Imaging andTarget Detection with a Heterodyne-Reception Optical Radar,"Appl. Opt. 20, 3292 (1981).
10. M. Yoshikawa, T. Fukuda, and T. Akamatsu, "Wide BandwidthHgCdTe Photodiode and Heterodyne Detection," in Proceedings,First Sensor Symposium, Tsukuba, Japan (June 1981), pp.235-239.
11. R. A. McClatchey and A. P. D'Agati, "Atmospheric Transmissionof Laser Radiation: Computer Code LASER," ERP 622, AirForce Geophysics Laboratory (Jan. 1978).
12. M. J. Post, in "Feasibility Study of Satellite-Borne Lidar GlobalWind Monitoring System," R. M. Huffaker, Ed., NOAA Tech-nical Memorandum ERL WPL-37 (1978).
Seventh International Conference onLaser Spectroscopy (SEICOLS)
Maui, Hawaii, USA, June 24-28, 1985
The Seventh International Conference onLaser Spectroscopy will be held fromJune 24-28, 1985 at the Maui Surf Hotel,Maui, Hawaii. Among the topics to bediscussed at this conference arefundamental physical application oflaser spectroscopy, new laserspectroscopic techniques, novelapplications of laser spectroscopy andnew coherent light sources. For moreinformation write to T. W. Hansch,Physics Department, Stanford University,Stanford, CA 94305, USA, or Y. R. Shen,Physics Department, University ofCalifornia, Berkeley, CA 94729, USA.
