Airborne Raman lidar

Airborne Raman lidar

Wm. S. Heaps and J. Burris

We designed and tested an airborne lidar system using Raman scattering to make simultaneous mea-surements of methane, water vapor, and temperature in a series of flights on a NASA-operated C-130aircraft. We present the results for methane detection, which show that the instrument has therequisite sensitivity to atmospheric trace gases. Ultimately these measurements can be used to exam-ine the transport of chemically processed air from within the polar vortex to mid-latitudinal regions andthe exchange of stratospheric air between tropical and mid-latitudinal regions. © 1996 Optical Societyof America

Key words: Lidar, ozone, remote sensing.

1. Introduction

An unintended consequence of using certain highlystable compounds consisting of chlorine, fluorine, andbromine has been the introduction of these atomicspecies into the stratosphere and the resulting an-nual destruction of ozone over the southern polarregions. Ozone depletion in the northern polar re-gions has now been observed, raising concerns aboutthe possible destruction of ozone over populated re-gions of Asia, Europe, and North America.1–3 Al-though winter stratospheric temperatures abovethese areas are rarely cold enough to generate thepolar stratospheric clouds ~PSC’s! required for theheterogeneous chemical destruction of ozone, the pos-sibility exists for the transport of chemically pro-cessed air from inside colder regions in which PSC’sdo form to the mid-latitudinal regions.The winter circulation in the polar regions is char-

acterized by a reasonably stable band of winds cir-cling the poles from west to east and defining thepolar vortex boundary. The breaking of large-scaleatmospheric waves against this boundary peels offlong, thin filaments of air that can then move tomid-latitudes. Their formation, spatial extent, andmovement are of great interest in the understandingof transport processes across the vortex boundary inthe time period before the final vortex breakup.Some observations suggest that filamentary dimen-

The authors are with the Laboratory of Atmospheres, NASAGoddard Space Flight Center, Greenbelt, Maryland 20771.Received 22 December 1994; revisedmanuscript received 2 April

1996.0003-6935y96y367128-08$10.00y0© 1996 Optical Society of America

7128 APPLIED OPTICS y Vol. 35, No. 36 y 20 December 1996

sions can vary from hundreds of kilometers to belowseveral kilometers. These filaments transportchemically processed air that can vary from severalpercent upward to 30% of the polar vortex volumeover time scales of 10 days to 2 weeks.4–6To track the motion of a filament, it is desirable to

employ a tracer whose concentration is representa-tive of the filament’s origin and whose lifetime ismuch longer than the relevant time scale for atmo-spheric motion. Two suitable gases are methaneand nitrous oxide, neither of which has a strato-spheric source. In the stratosphere, nitrous oxide isphotolyzed by UV radiation, creating a concentrationgradient that decreases with altitude. Methane alsoexhibits a negative vertical gradient that is due to itsdestruction by hydroxyl. During winter, diabaticcooling results in a downward vertical displacementof several kilometers within the vortex. This meansthat parcels within the polar vortex have lower meth-ane and nitrous oxide concentrations than those par-cels just outside the boundary. Thus filaments of airpeeled off the vortex edge can be identified by theirrelatively lower CH4 ~or N2O! mixing ratio ~the ratioof CH4 to atmospheric number density!.The ozone loss within the polar vortex is due to the

appearance of PSC’s. Clouds do not usually form inthe stratosphere because the water vapor concentra-tion is too low for condensation to occur at typicalstratospheric temperatures. However, within thevortex extremely low temperatures permit the forma-tion of clouds of nitric acid trihydrate near 195 K andwater ice crystals at 188 K. It is now believed thatheterogeneous chemical reactions occurring at thesurface of these clouds are responsible for the chem-istry that produces polar ozone depletion. Becausethe boreal polar vortex ismoremobile than its austral

counterpart, the average temperatures are not quiteas low as those found at the South Pole. However,PSC’s are observed regularly every winter within thenorthern polar vortex.7 Measurements of water andtemperature within the polar vortex are of greatvalue in the assessment of the formation probabilityof PSC’s and, thereby, the potential for the chemicaldestruction of ozone. To facilitate the investigationof all these phenomena related to PSC formation andmass transport across the vortex boundary, we de-signed an airborne lidar system that employs Ramanscattering to take simultaneous measurements ofmethane, water vapor, and temperature.

2. Technique

The lidar technique has been applied in a variety offorms to take a broad range of atmospheric measure-ments. In the most general terms, a pulse of light,usually from a laser, is transmitted into the atmo-sphere to scatter the light both elastically and inelas-tically. The backscattered light is collected by atelescope and its intensity measured as a function oftime after the emission of the light pulse. The tim-ing information is used to calculate the range withinthe atmosphere to the scattering site.For the determination of trace species concentra-

tion, three approaches are in common use: fluores-cence lidar, differential absorption lidar, and Ramanlidar. Fluorescence lidar is applicable if the speciesto bemeasured fluoresces under excitation at anotherwavelength. If the species has either a unique ab-sorption or fluorescence spectrum, then these fea-tures can be used to differentiate the desired speciesagainst signals arising from other scattering pro-cesses. Fluorescence lidar has been used success-fully tomeasure hydroxyl radicals in the stratosphereand sodium atoms in the ionosphere.8,9 Differentialabsorption is an applicable technique for species thatexhibit a significantly strong, unique absorption spec-trum. The technique makes use of elastic scatteringin the atmosphere to produce the return signal. Theelastic-scattering signal at two different wavelengths,one of which is absorbed strongly by the species underinvestigation, is compared as a function of range.Concentration of the absorbing species can be de-duced from changes in the relative strength of thebackscatter at the two wavelengths. Differential ab-sorption has been applied to many species with mea-surements of ozone and water vapor as theoutstanding examples.10,11 In Raman scattering,the scattered light is shifted in frequency from theincident beam by an amount that is characteristic ofthe scattering species. Raman lidar has had partic-ular success in the measurement of atmospheric wa-ter vapor.12In the design of a lidar to measure stratospheric

methane, we considered all three approaches.Methane has no fluorescence features of sufficientstrength to permit one to construct a lidar systememploying the fluorescence technique. Althoughmethane does have a unique absorption signature inthe infrared, significant absorption would not be ex-

pected over the desired spatial ranges at the concen-tration levels anticipated for the stratosphere. Inaddition, the Rayleigh scattering responsible for gen-erating the backscatter signal in the differential ab-sorption lidar technique is weak in the infraredwhere methane absorption occurs. Strong Mie scat-tering from aerosols would complicate seriously theanalysis of a spectrum. Fortunately methane doeshave a relatively strong vibrational Raman cross sec-tion with a shift of 2914 cm21 that produces a detect-able scattering feature in the stratosphere.When it is applicable, the Raman technique has a

number of advantages that make it an attractivechoice. In particular, Raman scattering does not de-pend on particular spectroscopic features that areunique to a given molecular species. Unlike the sit-uation with fluorescence and differential absorptionlidar systems in which considerable effort must begiven to the generation of a narrow-band output, witha Raman system one can use a spectrally unmodifiedlaser source. This eliminates problems associatedwith wavelength stability and usually means thatmore laser power will be available than with fluores-cence or differential absorption lidar. In addition,this means that a single laser source can be used toconstruct a measurement system capable of detectinga number of species simultaneously. In our case, wedeveloped a system employing a single excimer lasersource with multiple detectors tuned for the Raman-shifted wavelengths of nitrogen, methane, and watervapor. The Rayleigh- and Mie-scattered light at thelaser wavelength are also observed.As discussed above, the methane measurement is

desired because methane concentration can be usedas a tracer of atmospheric motions. The water vapormeasurement is interesting because it permits a cal-culation of the potential for the formation of PSC’s.The nitrogen and oxygen measurements, along withthe Rayleigh measurement, can be used to calculateboth temperature and pressure. This useful dataset is obtained from a single instrument. The prin-cipal design difficulty arises from the following facts.First, typical Rayleigh-scattering cross sections areapproximately 3 orders of magnitude greater thanRaman cross sections. Second, the abundance of ni-trogen and oxygen in the atmosphere is approxi-mately 6 orders of magnitude greater than that ofmethane or stratospheric water vapor. This meansthat the water vapor and methane detectors must becapable of detecting signals in the appropriate wave-length region while rejecting signals that are be-tween 5 and 8 orders of magnitude stronger only afew nanometers away in the spectrum. Table 1 il-lustrates the problem.

3. Laser Selection

Because of the low mixing ratios for stratosphericmethane and water vapor, we gave careful consider-ation to the requirements for a laser. Because Ra-man scattering is independent of any specificmolecular transition, the wavelength can be selectedbased on other criteria. The scattering cross section

20 December 1996 y Vol. 35, No. 36 y APPLIED OPTICS 7129

Table 1. Signal Parameters for the Various Channels in the Raman Lidara

Detector ChannelFrequency Shift

~cm21!Wavelength

~351-nm source! Relative Strength

RayleighyMie 0 351.0 1,000,000,000Oxygen 1555 and 3089 371.3 and 393.7 200,000 and 50Nitrogen 2331 and 4631 382.3 and 419.1 800,000 and 650Methane 2914 391.0 16Water vapor 3652 402.6 3

aThe tremendous range of signal strength exhibited by the different channels represents a formidable challenge in the design of adetector for the lidar.

does depend on the fourth power of the frequency, soshorter wavelengths generate more scattering.However, as wavelengths drop into the UV, absorp-tion by ozone ~,330 nm! and oxygen ~,240 nm!causes significant signal attenuation. To maximizethe signal-to-noise ratio, high peak powers and spec-trally narrow outputs are desired, thus restrictingthe choice to either Nd:YAG’s ~third harmonic! orexcimers. The relatively small size of a diode-pumped Nd:YAG, its high electrical wall-plug effi-ciency, and its ability to be line narrowed easily madethe YAG an attractive choice. Its output at 355 nmis not absorbed by ozone, and, in addition, the inertlasingmedia presents no health or safety problems asdo the halogens associated with excimers. At thepresent time, these lasers have serious problemswiththermally induced birefringence distorting the opti-cal phase front and thereby significantly degradingthe harmonic conversion efficiency. Excimers pro-duce intense UV light directly without the need forharmonic conversion. The xenon chloride excimer~XeCl!with emission near 308 nm is a good candidatefor a Raman lidar source. It possesses high peakpower, high average power, can be narrowed spec-trally for narrow-band detection, and uses provedtechnology with a long, successful history of fieldwork. Because of significant ozone absorption at308 nm, however, it would be difficult to do a temper-ature measurement with the 308-nm channel. Theother suitable excimer, xenon fluoride ~XeF!, alsogenerates high peak power, high average power, anduses proved technology. It was chosen ultimatelyfor the prototype instrument. Its major advantageis that ozone absorption does not interfere with thetemperature measurement made with the 351-nmsignal. The major disadvantage is that the spectraloutput consists of three narrow lines centered at 349,351, and 353 nm that cannot be line narrowed with-out significant energy loss. The detector’s band-width then must be ;5 nm, which increases thesystem’s sensitivity to background light.

4. Detector

The separation of multiple signals requires a detec-tion scheme that relies on either gratings or dielectricfilters to provide the necessary spectral separation.Because of the large difference in signal strengthsbetween the Rayleigh and the methane returns, adouble monochromator would be required to provide


a suitable stray light rejection ratio. We used thestray light rejection ratio to measure the fraction ofthe incident light that is scattered isotropically.Good holographic gratings typically have a stray lightrejection ratio of ;1025, which still leaves a Rayleighsignal that is 3 orders of magnitude stronger than theexpected methane and water vapor returns. A gooddouble monochromator can increase this to 10210,thereby providing enough rejection for the measure-ment to be made. Unfortunately, the throughputefficiency is not high ~;10%!, and it is difficult tolocate the photomultiplier tubes ~PMT’s! used for sig-nal detection at the focal plane of the instrumentbecause of the narrow separation between adjacentchannels ~several millimeters!. This obstacle can besurmounted by the use of optical fibers. However,alignment becomes critical and overall efficiency isextremely low ~,,1%!. A detector based on filters,unlike gratings and fiber optics, is both simple toconstruct and to align, and it possesses high effi-ciency.Filters consist of dielectric materials laid down on

a borosilicate or quartz substrate. By varying thecoating thickness, refractive index ~through thechoice of different materials! and number of layers, avendor can produce a product with the required spec-ifications. Because borosilicate glass will fluoresceunder excitation by 351-nm radiation, care must beexercised to ensure the use of a quartz substrate.Unlike a grating instrument, a large aperture is pos-sible, which would reduce problems associated withalignment, allow a wider choice of PMT’s, and help tominimize problems associated with inhomogeneitiesin the quantum efficiency of the photocathodes whenonly a small region is illuminated. The filters, madeby Barr Associates, had a relatively flat transmission~60%! over the 5-nm bandpass with out-of-bandblocking of 105. A 5-nm bandpass was required be-cause the XeF excimer’s output is generated on threenarrow lines and cannot be narrowed to a single linewithout a significant loss in output energy.Blocking out-of-band signals can be achieved throughthe use of either an UV transmitting glass ~SchottUG-11! or additional layers of dielectric material.Initially, the glass was chosen because of cost consid-erations. However, UG-11 can fluoresce under351-nm excitation, and this proved to be a problemwith certain filter combinations. Because of thelarge dynamic range in signal seen on the Rayleigh

Fig. 1. Schematic of the optical layout of the lidar system with particular emphasis on the function of the mirrors and filters used in thewavelength separation. BP, bandpass; OSC, oscillator.

and N2 Raman channels, the detector chain wasmod-ified in two significant ways. First, the nitrogenchannel was split into two channels: a high channelfor use at altitudes .10 km and a low channel for usefrom the ground to ;10 km. The low channel wasattenuated heavily to prevent large signals from driv-ing the PMT nonlinear. All channels of the detectorwere blocked mechanically through the use of a chop-per, which is basically a propeller blade rotating at6000 rpm. Blocking not only prevents nonlinear op-eration initially, it also eliminates signal-inducednoise, a behavior in which the long-term noise char-acteristics of the PMT are no longer linear andthereby makes it difficult to separate the high-altitude weak signals from PMT dark counts.Figure 1 shows a schematic of the system. Scat-

tered light is collected by a Newtonian telescope~40-cm primarywith a 120-cm focal length! supportedby a 120 cm 3 240 cm optical table. The telescope’ssecondary positions the focus 10 cm beyond the tele-scope tube at a height of 6.25 cm above the surface ofthe table. All the optical components are thenmounted on the table. This design facilitateschanges in the optical layout so that different ap-proaches can be tried without a major redesign of thedetector. A pinhole, located at the focus of the tele-scope, defines the detector field of view ~FOV!. Be-yond the pinhole the light expands to a diameter of;2.5 cm at which point it is collimated by twoantireflection-coated fused silica lenses. The colli-mated beam is refocused to a diameter of ;0.1 cm,and through this focus a blade rotating at high speedperiodically blocks ~chops! the optical path from thetelescope to the detectors. When the chopper bladeand laser are synchronized, the blocking period canbe chosen with high precision and great stability,currently less than 1 ms. Chopping was found to be

necessary to protect the PMT’s from the extremelyintense Rayleigh returns within the first 30 ms. An-other technique commonly used to protect PMT’s in-volves the rapid switching of the voltages on twoadjacent dynodes in the voltage divider network toprevent any amplification of the initial pulse. Al-though effective at blocking the initial current spike,the photocathode was still exposed to the intense ini-tial pulse, and signal-induced noise was still a severeproblem. After chopping, light enters the first beamsplitter box; this box contains a dichroic mirror thatreflects greater than 99% of all the Raman-scatteredwavelengths while transmitting 60% of the laser fun-damental wavelengths. The reflection is made at a15° angle of incidence to minimize polarization effectson reflection. Light transmitted by the dichroic mir-ror is attenuated by a factor of 100 and detected by aPMT behind a 5.0-nm bandpass filter centered at 351nm. Light reflected out of the first box enters thesecond beam splitter box. Here again the light isseparated by a dichroic mirror with wavelengthslonger than 385 nm ~consisting of the methane andwater vapor Raman signals! that are transmittedwhile the nitrogen and oxygen Raman wavelengthsare reflected. The transmitted long wavelengthlight enters a third beam splitter box that contains adichroic mirror to separate the water vapor andmethane signals. These two signals are detected byPMT’s located behind appropriate bandpass filters.Light reflected by the second beam splitter box issimilarly separated into the nitrogen and oxygencomponents that are then detected by suitably fil-tered PMT’s. One additional complication associ-ated with themethane channel is that the signal fromthe oxygen Raman overtone ~second Stokes shift! in-terferes with the methane measurement. This sig-nal is approximately three times the strength of the


methane and too close to the methane returns to beselected against by dichroic filters; as a consequenceit must be removed with an atmospheric model.

5. Calibration and Data Reduction

The data reduction for the lidar follows from the lidarequation that can be written as follows:

S 5 Inse~Ay4pz2!TDz, (1)

where S is the number of photons detected, I is thenumber of photons emitted by the laser, n is thenumber density of the scattering species, s is thescattering cross section for the species, ε is the photondetection efficiency of the detector, A is the area of thetelescope, T is the round-trip transmission of the at-mosphere, and z and Dz are the range and length ofthe scattering volume, respectively. Putting in rel-evant numbers for methane in the stratosphereyields an expected count rate of less than a photonper shot. Calibration of a lidar system can, in prin-ciple, be accomplished if one measures the photondetection efficiency of the system and simply invertsEq. ~1! to solve for the species number density, as-suming that the scattering cross section is known andthe atmospheric transmission can be approximatedto a sufficient degree of accuracy. For the methanechannel, the equation is complicated by the fact thatthe oxygen overtone signal is also present in thebandpass of the detector. In this case the simpleexpression nsε must be replaced by n1s1ε1 1 n2s2ε2,where subscript 1 applies to the methane Ramanscattering and subscript 2 applies to the oxygen over-tone Raman scattering. Therefore, when one in-verts the equation, one is solving the expression forthe sum of two signals. Fortunately the atmo-spheric number density of oxygen is related strictly tothe nitrogen number density over the range of alti-tudes that we are considering here. This meansthat the value of n2 can be inferred from the signalsin the lidar nitrogen channel. Similarly, we knowthe relative instrumental efficiency at the methaneand oxygen wavelengths ~this is just the passband ofthe filter that defines this channel!. Finally, thecross sections for the two processes are available inthe literature.13,14 For typical values of methaneand oxygen in the lower stratosphere, the signal de-tected is composed of approximately 75% from oxygenand 25% from methane.The ultimate calibration of the instrument is then

obtained when one takes a ground-based measure-ment of methane. The tropospheric mixing ratio ofmethane is 1.7 parts in 106. By taking a measure-ment in the troposphere where both the methane andoxygen number densities are known, one can deducethe overall efficiency of the system. This factor en-ables one to calculate the methane number density atany other point in the atmosphere from the signalsobtained in the methane and nitrogen channels of theinstrument.A few remarks about the achievable measurement

precision are in order. The signal levels for methane


~and for the oxygen overtone! are quite low in theupper troposphere and stratosphere. The nitrogensignals are 5 orders of magnitude stronger, however,so that the oxygen overtone signal, which must besubtracted from the total signal, will be known togreat precision. Because the precision of the meth-ane measurement depends on the total number ofsignal counts received, the count uncertainty intro-duced by the oxygen overtone represents a factor of 2reduction in the signal-to-noise ratio achieved if therewere no oxygen overtone. Consequently, at the up-per end of the range of usefulness for the instrument,the measurement precision may be only of the orderof 50%. It is important to remember the use towhich the instrument is to be put when this numberis evaluated. The instrument will be used to locatesharp discontinuities in the methane concentrationthat manifest themselves over a considerable rangeof altitudes in the upper troposphere and lowerstratosphere. Changes in methane of 20% andmorewill occur at these edges. When the precision at alidar range of 8–10 km is only 50%, the precision at3–4 km will be substantially higher. The location ofa discontinuity at the lower range will serve as noticethat additional signal averaging in the range andtemporal regimes will likely resolve the discontinuityat higher altitudes. The instrument performance todate appears to be adequate to perform the task athand.

6. Results

The measurement of atmospheric water vapor andtemperature by Raman-scattering lidar is a well-established technology. The performance of this in-strument is consistent with that demonstrated bynumerous other researchers for ground-based in-strumements ~see, for example, Melfi et al.12 andMeasures15!. Methane detection by Raman scatter-ing has not been demonstrated previously and is theprincipal focus of this section. Although the instru-ment is intended ultimately for operations on theNASA DC-8, for the first flight tests it was installedon the C-130 aircraft operating out of NASA’s Wal-lops Island Flight Facility. The C-130 is a militarytransport powered by four turboprop engines and hasa rear access ramp that greatly facilitated the instal-lation of the rather large lidar system onto the air-craft. Extended duration flights of 12 h permittedtests of the long-term operating stability of the lidar.Maximum operational altitudes were 6 kmwith a fullfuel load and 9 km with a half fuel load, both some-what under the instrument’s intended altitude of 12km.A total of seven test flights was conducted. These

consisted of four short flights in the Wallops Islandvicinity for shakedown purposes, a ferry flight fromWallops Island to Bangor, Maine, a long missionnorth from Bangor to 70 °N ~Baffin Island! where weattempted to penetrate a remnant of the polar vortex,and a ferry flight back to Wallops Island. The in-strument performed well on all flights. On the firstflight it was found that, in turbulence, the overlap

between the telescope and the laser beam tended todrift. The telescope mounting was stiffened, and areal-time alignment monitoring channel was added toalleviate this problem. Nevertheless, slight adjust-ments to the alignment were required at intervals ofapproximately 30 min. Misalignment does not affectthe calibration of the instrument because signal-ratioingmethods are used in the data analysis. How-ever, it does affect the signal-to-noise ratio because thesignal is lost when the overlap between the laser beamand telescope FOV is not optimum. A more seriousproblem involved ice formation on the interior surfaceof the window used to transmit the laser light andsignal returns. We solved this problem by directingwarm air from the aircraft heating system onto thewindow. A slow, gradual buildup of film on the win-dow’s surface was noticed, which significantly reducedUV transmission. This film remained on the windowafter the flight was concluded and the window hadwarmed up to ambient temperatures. A hydraulicfluid leak in one engine suggested that the contami-nant originated there through a defective heat ex-changer that was used for cabin heating. In anyfuture operations on the C-130, a slow purge of warm,dry nitrogen through carefully cleaned hoses wouldeliminate the contamination problem.Figure 2 illustrates the performance of the system.

These plots depict the raw counts obtained in themethane and water vapor channels as a function ofrange above the aircraft. The three plots were ob-tained in 5 min of 50-Hz operation as the aircraft de-scended to Wallops Island after a flight. The signalarising from water ~the thin curve on the plots! in-creases dramatically as the aircraft descends becauseof the strong vertical gradient in the water vapor mix-ing ratio. Signals from the methane channel, whichcontains counts from both methane Raman and oxy-gen overtone Raman, show no such dramatic change.This is expected because methane is well mixed in the

Fig. 2. Comparison of the raw counts in the methane and watervapor channels obtained as the aircraft descended. The properbehavior of the signals with altitude suggests that the detector isworking as it should, and, particularly, that filter fluorescence isnot a problem.

mid-latitudinal troposphere. In addition to demon-strating the water channel’s large dynamic range,these data set upper limits on both the backgroundcounts that were due to fluorescence in the dichroicfilters and the 351-nm leakage through the filters intothe methane and water vapor channels. Additionalresearchwith 351-nm radiation, 6 orders ofmagnitudeabove expected levels, produced no detectable fluores-cence or punch through on these channels.Figure 3 presents the raw counts obtained in the

methane channel versus the range above the aircraftobtained shortly after takeoff on the poleward flightout of Bangor, Maine. These counts were obtainedover a 20-min period with the laser system operatingat 50Hz, which is one fourth of its normal operationalrate. The salient features of this plot are that thebackground count is only approximately 250 ~approx-imately 0.004 per shot! and that the signal in excessof the background is present as far as 10 km abovethe aircraft.Figure 4 presents the methane altitude profile cal-

culated from the data in Fig. 3. Data taken later inthe flight are of lower overall quality for two reasons:first, oil accumulation on the window reduced signallevels during the course of the mission, and second,an unusually spectacular aurora occurred on thenight of the flight. Because of the aurora, the back-ground signals in the methane channel were ob-served to increase dramatically from 0.004 per shot tomore than 0.04 per shot. Because the laser powerand the system alignment held steady over the 12-hduration of the flight ~the excimer gas fill waschanged after approximately 6 h!, we believe thatdata taken early in the mission, before the problemswith the aurora and oil film, are most representativeof the system capabilities.

7. Conclusions

An airborne lidar system was built and operated thatpossesses sufficient sensitivity and specificity tomeasure

Fig. 3. Raw counts obtained in the methane channel for 60,000laser shots are shown as a function of range above the aircraft.These data were collected over a 20-min period early in the flightnorth from Bangor, Maine.


stratospheric methane and water vapor. In partic-ular we have shown that an array of dielectric mir-rors and bandpass filters can provide adequatewavelength separation and transmission to permitthe assembly of a detector capable of recording si-multaneously the atmospheric Raman signal fromnitrogen, oxygen, water vapor, and methane. As aresult of the test program concluded in April 1994,

Fig. 4. Methane mixing ratios at 0.9-km intervals above the air-craft calculated from the signals shown in Fig. 3.

a number of changes in the implementation of thelidar have been made. First, information from theoxygen channel is largely redundant. The signalcontains no information that is not available fromthe nitrogen channel but has a lower signal-to-noiseratio. Second, the blocking against the 351-nmsignal on the methane and water vapor channels ~ataltitude! is adequate protection from the outgoinglaser pulse so that mechanical chopping is not re-quired. Third, it was found that if chopper protec-tion is extended to higher altitudes ~20–25 km! onthe Rayleigh channel, enhanced sensitivity is real-ized, thereby allowing temperature determinationsfrom much greater altitudes. As a consequence ofthese findings the detector will be reconfigured witha number of performance enhancements.The chopper will be repositioned so that it affects

only the Rayleigh and nitrogen channels, thus allow-ing water andmethanemeasurements to be extendedto the zero range. This will facilitate comparisonswith in situ instruments onboard the aircraft. Thechopper timing will be adjusted to turn on at a rangeof 20–25 km above the instrument instead of the2–3-km range used in the tests. With this adjust-ment, approximately 2 orders of attenuation used inthe Rayleigh channel can be removed, extending therange for the Rayleigh channel to 70 km. In regions

Fig. 5. Preliminary methane data as a function of time and altitude obtained on a flight from Fairbanks, Alaska, to the North Pole andback. Data obtained before 4:00 UT were contaminated by twilight and aurora as were data obtained at 9:15 UT. A symmetric patternof tongues of lower methane are seen at approximately 5:00 and 7:45 UT. We believe this to be the signature of a filament of polar aircoming off of the polar vortex.


that are free of aerosols ~above 30 km!, Rayleigh scat-tering can be used to calculate temperature and pres-sure. This change will therefore increase thealtitude coverage for the temperature and pressuremeasurements, as well as reduce the uncertainty atlower altitudes. The oxygen channel will be re-moved altogether in preference for a second nitrogenchannel that is unchopped but attenuated to preventnonlinear response arising from the outgoing laserpulse. This nitrogen channel will be used for thenormalization of the methane and water vapor sig-nals as well as a tie-on for the temperature measure-ments made in the chopped channels. This will, inprinciple, permit temperature and pressuremeasure-ments from the aircraft to the top of the stratosphere.In addition, an improved set of mirrors and filters willbe installed that provides additional rejection of theoxygen overtone signal in the methane channelthereby enhancing the signal-to-noise ratio for thesemeasurements. Finally, some adjustments will beattempted on the FOV of the receiver telescope.Clearly, if the FOV can be reduced then the back-ground count rate for all the channels will be reduced.However, if the FOV is too small, it becomes difficultto retain alignment between the receiver and thetransmitted laser beam under flight conditions in-volving turbulence and aircraft vibration.

8. Epilogue

The airborne lidar was installed onboard the DC-8 inJune 1995, andmeasurement campaigns to character-ize transport in the tropic and polar regions were un-dertaken between December 1995 and February 1996.These measurement campaigns involved multiple in-struments onboard the DC-8 and coordination withsatellite overflights that permit a detailed validation ofthe lidar performance. A preliminary CH4 profile ob-tained in January 1996 is shown in Fig. 5.

The authors express many thanks to David Pierceand the staff of the Aircraft Programs Branch at theWallops Island Flight Facility for their extraordinaryefforts on behalf of this program. Don Silbert of theGoddard Space Flight Center provided technical as-sistance in all areas of instrument development thatis without peer. The Space Science machine shop ofthe University of Maryland provided timely and pro-ficient fabrication support. This research was sup-ported by the Upper Atmosphere Research Programand the Atmospheric Effects of Aviation Project of theNational Aeronautics and Space Administration.

References1. R. J. Salawich, S. C. Wofsey, E. W. Gottlieb, L. R. Lait, P. A.

Newman, M. R. Schoeberl, M. Loewenstein, J. R. Podolske,S. E. Strahan, M. H. Proffitt, C. R. Webster, R. D. May, D. W.Fahey, D. Baumgardiner, J. E. Dye, J. C. Wilson, K. K. Kelly,J. W. Elkins, K. R. Chan, and J. G. Anderson, “Chemical lossof ozone in the arctic polar vortex in the winter of 1991–1992,”Science 261, 1146–1149 ~1993!.

2. M. H. Proffitt, K. Aikin, J. J. Margitan, M. Loewenstein, J. R.Podolske, A. Weaver, K. R. Chan, H. Fast, and J. W. Elkins,“Ozone loss inside the northern polar vortex during the 1991–1992 winter,” Science 261, 1150–1154 ~1993!.

3. R. Stolarski, P. Bloomfield, R. D. McPeters, and J. R. Herman,“Total ozone trends deduced from Nimbus 7 toms data,” Geo-phys. Res. Lett. 18, 1015–1018 ~1991!.

4. M. Schoeberl, L. R. Lait, P. A. Newman, and J. E. Rosenfield,“The structure of the polar vortex,” J. Geophys. Res. 97, 7859–7882 ~1992!.

5. D. Waugh, R. A. Plumb, P. A. Newman, M. R. Schoeberl, L. R.Lait, M. Loewenstein, J. R. Podolske, J. W. Elkins, and K. R.Chan, “Fine scale, poleward transport of tropical air duringAASE2,” Geophys. Res. Lett. 21, 2603–2606 ~1994!.

6. R. Plumb, D. W. Waugh, R. J. Atkinson, P. A. Newman, L. R.Lait, M. R. Schoeberl, E. V. Browell, A. S. Simmons, and M.Loewenstein, “Intrusions into the lower stratosphere arcticvortex during the winter of 1991–1992,” J. Geophys. Res. 99,1089–1105 ~1994!.

7. M. P. McCormick, C. R. Trepte, andM. C. Pitts, “Persistence ofpolar stratospheric clouds in the southern polar region,” J.Geophys. Res. 94, 11241–11251 ~1989!.

8. W. Heaps, T. J. McGee, R. D. Hudson, and L. O. Caudill,“Balloon-borne lidar system for the measurement of strato-spheric hydroxyl and ozone,” Appl. Opt. 21, 2265–2274 ~1982!.

9. C. Gardiner, D. G. Velz, C. F. Sechrist, Jr., and A. C. Segal,“Lidar studies of the nighttime sodium layer over Urbana,Illinois. 1. Seasonal and nocturnal variations,” J. Geophys.Res. 91, 13659–13673 ~1986!.

10. T. J. McGee, D. Whiteman, R. Ferrare, J. J. Butler, andJ. Burris, “STROZ LITE: stratospheric ozone lidar trailerexperiment,” Appl. Opt. 30, 29–31 ~1991!.

11. E. V. Browell, T. D. Wilkerson, and T. J. McIlrath, “Watervapor differential absorption lidar development and evalua-tion,” Appl. Opt. 18, 3474–3483 ~1979!.

12. S. H. Melfi, D. Whiteman, and R. Ferrare, “Observation ofatmospheric fronts using Raman lidar moisture measure-ments,” J. Appl. Meteorol. 28, 789–805 ~1989!.

13. W. Knippers, K. VanHelvoort, and S. Stolte, “Vibrational over-tones of the homonuclear diatomics ~N2, O2, D2! observed bythe spontaneous Raman effect,” Chem. Phys. Lett. 121, 279–286 ~1985!.

14. M. A. Buldakov, B. V. Korolev, I. I. Matrosov, and T. N. Pop-ova, “Overtone bands in Raman spectra of nitrogen and oxy-gen,” Opt. Spectrosc. ~USSR!, 63, 460–461 ~1987!.

15. R. M. Measures, Laser Remote Sensing ~Wiley, New York,1983!.


Documents

Airborne Raman lidar