airborne high spectral resolution lidar for measuring aerosol extinction and backscatter...

Airborne high spectral resolution lidar for measuringaerosol extinction and backscatter coefficients

Michael Esselborn,1,* Martin Wirth,1 Andreas Fix,1 Matthias Tesche,2 and Gerhard Ehret1

1Institut für Physik der Atmosphäre, Deutsches Zentrum für Luft- und Raumfahrt (DLR), Oberpfaffenhofen,82234 Wessling, Germany

2Leibniz-Institut für Troposphärenforschung (IFT), 04318 Leipzig, Germany

*Corresponding author: [email protected]

Received 15 August 2007; revised 3 December 2007; accepted 3 December 2007;posted 4 December 2007 (Doc. ID 86422); published 14 January 2008

An airborne high spectral resolution lidar (HSRL) based on an iodine absorption filter and a high-powerfrequency-doubled Nd:YAG laser has been developed to measure backscatter and extinction coefficientsof aerosols and clouds. The instrument was operated aboard the Falcon 20 research aircraft of theGerman Aerospace Center (DLR) during the Saharan Mineral Dust Experiment in May–June 2006 tomeasure optical properties of Saharan dust. A detailed description of the lidar system, the analysis of itsdata products, and measurements of backscatter and extinction coefficients of Saharan dust are pre-sented. The system errors are discussed and airborne HSRL results are compared to ground-basedRaman lidar and sunphotometer measurements. © 2008 Optical Society of America

OCIS codes: 010.0280, 010.1280, 280.1100, 280.3640, 290.2200, 290.5850.

1. Introduction

Aerosols play a key role in the Earth’s radiative bud-get because they directly influence the fluxes of solarand terrestrial radiation within the atmosphere byabsorption and scattering of light [1]. The quantifi-cation of this effect accounts for accurate and highlyresolved measurements of aerosol extinction and ver-tical distribution. Using a conventional backscatterlidar, aerosol extinction cannot be measured directly.Only with assumption of the aerosol extinction-to-backscatter ratio, the so-called lidar ratio, aerosolextinction coefficients can be retrieved by means ofinversion algorithms [2]. However, the aerosol lidarratio is a highly variable quantity, so that large errorsin the retrieval must be expected if the lidar ratio isnot known exactly. The only alternate lidar method tomeasure aerosol extinction profiles directly is Ramanlidar [3]. The long averaging times associated withthe Raman method, however, have prevented thistechnique from being used in airborne operation sofar. Recently, ground-based demonstration measure-ments of the lidar ratio and other quantities with a

Raman lidar specially developed for airborne opera-tion have been reported [4]. A high spectral resolutionlidar (HSRL) takes advantage of the different spec-tral broadening of light, backscattered by moleculesand aerosols. By means of a narrow bandwidth opti-cal filter the aerosol contribution is separated fromthe molecular backscatter. Thus, aerosol backscatterand extinction coefficients can be measured directlyand no assumption about the lidar ratio is required.Up to now, only a few HSRL instruments have beenproposed and successfully implemented to measureatmospheric parameters or particle properties. Theseutilize either narrow bandwidth interferometers [5]to reject aerosol scattering or atomic [6–8] and mo-lecular [8–10] vapor filters. The advantages of iodinevapor filters are the strong rejection of aerosol back-scatter at low cell temperatures and the marginalsensitivity to optical alignment and angular diver-gence of the backscattered light. Furthermore, theycan be designed in compact dimensions and do nothave to be pressure stabilized. Thus, iodine vaporfilters are ideal candidates for airborne HSRL oper-ation. Airborne HSRL measurements could be dem-onstrated by our group during the LindenbergAerosol Characterization Experiment (LACE) fieldcampaign in 1998 using a 10 Hz single-mode Nd:YAG

0003-6935/08/030346-13$15.00/0© 2008 Optical Society of America

20 January 2008 � Vol. 47, No. 3 � APPLIED OPTICS 346

laser system aboard the German Aerospace Center(DLR) Falcon research aircraft [11]. This system waslimited by the relatively low horizontal sampling ca-pability, low aerosol attenuation within the single-pass iodine vapor cell and unstable laser frequencycontrol. Recently, another HSRL based on a 200 Hzlow-power Nd:YAG laser has been deployed duringtwo field experiments aboard the NASA B200 KingAir [12].

Here, we present a new airborne HSRL, which wasapplied during the Saharan Mineral Dust Experi-ment (SAMUM) field campaign in May–June 2006 tomeasure cross sections of extinction and backscattercoefficients of pure Saharan dust close to its sourceregions for the first time to the best of our knowledge.This system is operated at 100 Hz using a high-powerNd:YAG laser, which is frequency controlled bymeans of a new robust laser frequency stabilization.For stronger aerosol suppression a new iodine vaporfilter in dual-pass configuration has been designed.The development of the HSRL system was partiallydriven by two forthcoming satellite missions of theEuropean Space Agency (ESA). The Atmospheric Dy-namics Mission (ADM [13]) will use a Doppler lidar,which is expected to also have potential for extinctionprofiling [14]. The EarthCARE mission [15] will em-ploy a HSRL instrument at 355 nm. The airborneHSRL measurements can be of great help to evaluatethe potential and performance of the spaceborne in-struments in advance. For validation purposes, air-borne measurements can be coordinated and allownearly real-time spatial overlap along the satellitefootprint.

The purpose of this paper is a detailed descriptionof our new airborne HSRL system, the retrievalmethod of its data products, an error analysis, and apresentation of measurement examples together witha comparison to ground-based Raman lidar andsunphotometer data. A detailed scientific analysis ofthe campaign’s results will be the subject of a forth-coming paper.

2. HSRL Theory and Data Retrieval

The elastic scattering of light by particles small com-pared to the wavelength is usually described by Ray-leigh scattering theory [16]. Besides the elastic part,atmospheric Rayleigh scattering [17] includes the fre-quency shifted rotational Raman bands associatedwith �J � �2 transitions, where J is the rotationalquantum number of the scattering molecule. Thedominant central part of the backscatter spectrum,which arises from elastic scattering together with theunshifted �J � 0 rotational Raman branch make upthe Cabannes line [18], which shows pressure andtemperature-dependent Brillouin sidebands [19]. Atatmospheric temperatures close to 300 K the Dopplerbroadening of the Cabannes line amounts to 2.6 GHzfor green light with a wavelength of 532 nm. In con-trast, aerosol backscatter is hardly broadened due tothe relatively slow motion of aerosol particles, so thatit can be characterized by the laser frequency distri-bution.

Using HSRL, the received atmospheric backscatteris split into two channels. The narrow bandwidthoptical filter in the molecular channel suppresses theaerosol backscatter, whereas the combined channeldetects the intensity of both aerosol and molecularbackscatter. Therefore the emitted laser frequencymust be tuned to match the filter absorption line. Theiodine absorption filter eliminates the aerosol back-scatter and transmits the wings of the Doppler broad-ened molecular backscatter spectrum. To determinethe amount of molecular backscatter absorbed by theiodine filter, the HSRL system needs to be calibrated.This is done by measuring the filter transmissionspectrum and calculating the atmospheric tempera-ture and pressure-dependent filter transmission withan appropriate molecular backscatter model. Formeasuring the iodine filter transmission spectrum, ahighly attenuated reflection of the pulsed green laseremission is directed through the receiver assemblyand the laser frequency is scanned. The filter trans-mission is determined by the product of the iodinefilter transmission and the calculated molecularbackscatter spectrum. In our experiments the trans-mitted fraction of the molecular backscatter exceeds30% for atmospheric temperatures higher than 200 Kusing iodine absorption line 1109 (line notation fol-lows Ref. [20]). The partial transmission �m�r� ofthe temperature- and pressure-dependent Cabannesspectrum R��, T, p� through the iodine vapor filterwith transmission function �� may be written as

�m�T, p� �

� ��R��, T, p�l�� d��d�

�R��, T, p�l�� d��

, (1)

with l�� being the laser spectrum. Figure 1 shows themeasured transmission of the iodine filter cell oper-ated at a vapor pressure of 53 Pa as a function offrequency offset from the absorption maximum. Thecentral absorption line is iodine line 1109. The mo-

Fig. 1. Measured iodine transmission spectrum at 563.244 THztogether with molecular backscatter spectrum calculated with theS6-Tenti model for an atmospheric temperature of 300 K andpressure of 1000 hPa.

347 APPLIED OPTICS � Vol. 47, No. 3 � 20 January 2008

lecular backscatter spectrum was calculated with theS6 code provided by Tenti et al. [21] using a molecularmass of 28.8 kg�kmol for air molecules. Due to thelack of experimentally verified data for air molecules,the molecular parameters for the ratio of shear vis-cosity to bulk viscosity and the ratio of shear viscosityto thermal conductivity of nitrogen [22] have beenused. The result is given for an atmospheric temper-ature of 300 K and pressure of 1000 hPa. The trans-mitted fraction of molecular backscatter is the ratioof the integrals of the molecular spectrum beforeand after filtering. In the case of T � 300 K, p �1000 hPa the fraction amounts to �m � 0.430.

Depending on the bandwidth of the backgroundfilter, spectral components arising from pure rota-tional Raman scattering have to be considered. Fig-ure 2 shows the measured transmission function ofthe interference filter and the rotational Raman spec-trum as a function of the frequency offset from theincident laser wavelength. The redshifted Stokes andthe blueshifted anti-Stokes branch are denoted by Sand O, respectively. The full width at half-maximum(FWHM) of the interference filter is 1095 GHz. Ascan be seen, the first Stokes and anti-Stokes Ramanline of oxygen and the first two lines of nitrogen areincluded within the FWHM of the filter. The centralQ branch of the rotational Raman spectrum adds tothe Cabannes line R��, T, p� and has the same spec-tral broadening. Thus, it does not affect the calcula-tion of �m. The shifted O and S branches contributeonly 2.5% of the whole Rayleigh cross section. Theirpartial transmission through the background inter-ference filter can be calculated. In our case, 5.7% ofthe shifted rotational Raman scattering intensity istransmitted through the filter. Raman scatteringcontributes less than 0.5% to the molecular channelintensity assuming that at least 1�3 of the totalCabannes scattering is detected and the Ramanbands are not attenuated by the iodine absorptionspectrum. Therefore the small impact of Raman scat-tering on the backscatter calculation is neglected.

The other quantity measured through calibrationis the transmission of aerosol backscatter in the mo-lecular channel. This quantity decreases with in-creasing vapor pressure and optical path lengthwithin the filter cell. Assuming no spectral broaden-

ing, the aerosol backscatter spectrum can be de-scribed by the laser frequency distribution. Thelaser’s spectral width of �75 MHz is small comparedto the bandwidth of the iodine filter, which is �2 GHzat typical filter conditions. Thus, the transmittedfraction of aerosol backscatter �a through the filtercan be regarded as the filter transmission at linecenter:

�a � ��0�. (2)

The absorption within the iodine filter cell at linecenter is usually very strong and the experimentaldetermination of �a is restricted by detector noise orstray light. With a model provided by Forkey et al.[23,24], iodine absorption spectra at 532 nm can becalculated for different cell lengths and tempera-tures. For our HSRL experiments the calculatedaerosol transmission is typically of the order of 10�6

� �a � 10�5. As will be shown in Subsection 3.B, ourreceiver module is laid out for polarization sensitivedetection. At first, the received atmospheric back-scatter is split into its parallel (superscript �) andcross-polarized (superscript �) components. Theparallel-polarized channel is split again into thecombined channel (subscript C) and the molecularchannel (subscript M). The received power from dis-tance � in the three measurement channels can bewritten as

PC��r� � C

�E0�

c2

A

r2 �2�r��m��r� � a

��r��, (3)

PM��r� � M

�E0�

c2

A

r2 �2�r��m�r�m��r� � �aa

��r��, (4)

P��r� � �E0�

c2

A

r2 �2�r��m��r� � a

��r��, (5)

where � denotes the channel efficiency, c is the speedof light, A is the telescope aperture; E0

� is the pulseenergy, �2�r� is the two-way atmospheric transmis-sion over range r from the lidar transmitter to thescattering event, and m,a

�,��r� denotes the paralleland perpendicular backscatter coefficients of mole-cules and aerosols. Note that �m�r� is a function ofheight due to its dependence on atmospheric temper-ature and pressure. The atmospheric transmission iscomposed of a molecular and an aerosol contribution,

�2�r� � �m2�r� �a

2�r�

� exp�2�0

r

��m�r�� a�r��dr�, (6)

where �m,a�r� denotes the molecular and aerosol ex-tinction coefficient. Molecular and aerosol absorptioneffects can be neglected at the used lidar wavelengthof 532 nm. Following Rayleigh scattering theory [17],�m�r� is strictly proportional to m�r�,

Fig. 2. Measured interference filter transmission and calculatedrotational Raman spectrum at 563.244 THz.


�m�r� �8

3�45 � 10��45 � 7��

m�r�, (7)

where ��532 nm� � 0.222 accounts for molecular an-isotropy [25]. m�r� is given by the product of thedifferential backscattering cross section �� and the number density N�r� of the scatterers [26]

m�r� ��

��N�r�, (8)

where N�r� can be calculated from measured tem-perature and pressure data or from standardatmosphere profiles. The Rayleigh scattering proper-ties of the atmospheric gas composition are well-known and tabulated in the literature [25,27]. Thetotal Rayleigh scattering cross section at 532 nmamounts to � � 5.16 � 10�31 m2 as interpolated fromdata given in Ref. [27]. The differential backscatter-ing cross section of the Cabannes line only amounts to�� 5.93 � 10�32 m2 sr�1 at 532 nm. How-ever, the effectively detected cross section depends onthe spectral and polarization properties of the re-ceiver because of the depolarization of the rotationalRaman lines and their spectral separation. As al-ready stated the rotational Raman lines partiallytransmit through the interference filter. For thebackground interference filter and the polarizationbeam splitters in our receiver the value of the differ-ential backscattering cross section reduces to 5.91� 10�32 m2 sr�1.

To determine the amount of total (parallel andcross-polarized) backscatter, which is required for de-polarization data processing, the measured powers inthe parallel-polarized combined channel and thecross-polarized channel has to be calculated:

PT�r� � PC� �

C�

�P��r�. (9)

To determine the relative sensitivity factor C�� a

calibration during the measurement flight is re-quired. Therefore the receiver module is rotated byan angle of 45°, such that the plane of polarization ofboth the parallel and the cross-polarized channel isrotated by 45° relative to the transmitter’s plane ofpolarization. C

�� is then calculated from the ratioof the signal intensities in the parallel and cross-polarized channel. To obtain the aerosol extinctionand backscatter coefficients from the measured sig-nals given in Eqs. (3)–(5) and (9), the signals aremultiplied by r2 for range correction and divided by��m

2�r� � m�r�� to account for molecular extinctionand backscatter, respectively. Moreover, the signalsare normalized to the power received from a region atdistance r0 where the aerosol concentration can beneglected, i.e., a�r0� � 0, whereby all constants can-cel out. Hence, an attenuated backscatter ratio R,which is the backscatter ratio �1 � a�r��m�r�� mul-tiplied by aerosol transmission is defined as

RC��r� �

PC��r�

CC�

r2

�m2�r�m

��r�� a

2�r�1 �a

��r�m

��r�, (10)

RM��r� �

PM��r�

CM�

r2

�m2�r�m

��r�� a

2�r��m�r� � �a

a��r�

m��r�,

(11)

RT�r� �PT�r�

CT

r2

�m2�r�m

T�r�� a

2�r�1 �a

T�r�m

T�r�, (12)

where m,aT�r� denotes the total (parallel and cross-

polarized) molecular and aerosol backscatter coeffi-cients. The constant CC,M

�,T for each channel is chosento set RC

��r0� and RT�r0� close to 1 and RM��r0� to the

value of �m�r0�. The constant is composed of CC,M�,T

� C,M�,TE0

�A�c�2� and calculated for each profile.Equations (10) and (11) are used to express the two-way aerosol transmission by the two measured quan-tities and the calibration constants:

�a2�r� �

RM��r� � �aRC

��r��m�r� � �a

. (13)

The aerosol extinction coefficient �a�r� follows fromaerosol optical thickness (AOT) ta�r� by differentia-tion:

ta�r� � �12 ln��a

2�r��, (14)

�a�r� ��

�r ta�r�. (15)

To calculate the numerical derivative of the AOT pro-file, a Savitzky–Golay filter [28] of first order is used.This is equivalent to a linear regression on a specifiednumber of data points within a moving window. Theslope of the fitted straight line in each point is takenas the value of the derivative. This differentiationmethod tends to preserve slopes and edges in theAOT profile, which would be flattened by adjacentaveraging techniques.

To determine the total aerosol backscatter coeffi-cient a

T�r� � a��r� � a

��r� from the three measuredsignals given in Eqs. (3)–(5), the aerosol depolariza-tion ratio �a�r� has to be analyzed first. Therefore theequation given in [29] is used:

�a�r� �a

��r�a

��r��

�1 � �m��v�r�RT�r��a2�r� � �1 � �v�r��m

�1 � �m�RT�r��a2�r� � �1 � �v�r��

,

(16)

where �m denotes the volume depolarization ratio ofmolecular backscatter and �v�r� the volume depolar-ization ratio of total backscatter defined as


�v�r� �C

�

�

P��r�PC

��r�. (17)

The value of �m is highly dependent on the amount ofdetected pure rotational Raman scattering and variesbetween 3.63 � 10�3 and 1.43 � 10�2 depending onthe detection of the Cabannes line alone or the fullinclusion of the pure rotational Raman bands, res-pectively [30]. As has been discussed before, our in-terference filters partially transmit the rotationalRaman spectrum and the value of �m has been calcu-lated to 6.8 � 10�3. Finally, the total aerosol back-scatter coefficient can be expressed by

aT�r� � RC

��r��a

2�r�� 1m

��r��1 � �a�r��. (18)

The aerosol lidar ratio Sa�r� is defined as the ratioof the aerosol extinction coefficient to the backscattercoefficient:

Sa�r� ��a�r�a

T�r�. (19)

Unlike the constant molecular lidar ratio, the aerosollidar ratio generally varies with height because itdepends on the aerosol size distribution, its shape,and chemical composition [31,32].

3. System Description

A. Transmitter

The HSRL described in this paper has been devel-oped as an extension of the existing airborne DLRwater vapor differential absorption lidar (DIAL) sys-tem [33]. The laser transmitter used for the HSRLmeasurements is a high-power, Q-switched Nd:YAGlaser in a master-oscillator power amplifier configu-ration. It consists of a low-power master oscillatorand three power amplifiers yielding a pulse energy of220 mJ at 1064 nm and a repetition rate of 100 Hz. Aschematic of the HSRL transmitter is shown in Fig. 3.

A detailed description of the laser system can befound in Ehret et al. [34]. To obtain single longitudi-

nal mode operation the master oscillator is injectionseeded by a monolithic Nd:YAG ring laser and stabi-lized by minimizing the Q-switch build-up time [35].The seed laser frequency is tuned by changing theNd:YAG crystal temperature. The fundamental radi-ation of the pulsed laser is frequency-doubled using atemperature stabilized KTP crystal yielding a pulseenergy of 110 mJ. An attenuated reflection of thegreen radiation is directed to a frequency stabilizerthat continuously controls the temperature of theseed laser. The stabilization method is based onacousto-optic modulation and locks the green laserradiation to a Doppler-broadened iodine absorptionline. The advantage of this frequency stabilizationtechnique is the monitoring and stabilization of theamplified laser output instead of the seed laser fre-quency. With this new method long-term frequencyfluctuations have been reduced to �� 4.8 MHz or inrelative terms �� 8.5 � 10�9 during airborneoperation. Table 1 summarizes the system proper-ties.

B. Receiver

The atmospheric backscatter is collected by means ofa 350 mm Cassegrain telescope. A field stop withinthe focal plane of the telescope limits the acceptanceangle to 1 mrad. Dichroic beam splitters are used tospectrally separate the backscatter signals at 1064and 532 nm. The part of the receiver that detectsatmospheric backscatter at 532 nm is schematicallyshown in Fig. 4. The received light is split into itspolarized components with a polarizing beam split-ter cube. The cross-polarized component is directedto photomultiplier PMT3 whereas the parallel-polarized component is transmitted. In both paths asecond polarizing beam splitter is placed to reducepolarization cross talk. The transmitted parallel com-ponent of the backscatter signal is split again, suchthat half of its energy is directed into the molecularchannel.Fig. 3. Schematic of the transmitter system.

Table 1. System Parameters of the HSRL System

MOPA Laser systemPulse energy at 1064 nm 220 mJRepetition rate 100 HzPulse width at 1064 nm 18 nsLinewidth at 1064 nm �75 MHzLaser divergence at 1064 nm 0.5 mrad

SHGPulse energy at 532 nm 100 mJFrequency fluctuations �5 MHzSpectral purity 99.96%

ReceiverTelescope diameter 350 mmBackground filter bandwidth 1 THzIodine absorption bandwidth 2 GHzOptical iodine filter length 380 mmDetector divergence 1 mradPMT quantization 14 bitsPMT sampling rate 10 MHz


The other part is guided to PMT1 to measure theintensity of the combined backscatter. Within the mo-lecular channel light travels through the iodine vaporcell and is reflected at a 0° mirror. The plane of po-larization is rotated by 90° using a quarter-waveplate in dual pass, so that a polarizing beam splitterplate allows the reflected light to be transmitted ontoPMT2. By means of the two-way pass, the opticalpath inside the iodine vapor cell is doubled, so thatthe geometric dimensions are kept small and stron-ger absorption is achieved without increasing the io-dine vapor pressure. Thus, self-pressure broadeningof the iodine absorption line is reduced and less of thewings of the molecular backscatter spectrum isblocked. The geometric length of the filter cell is190 mm. It is temperature stabilized and a cool fingerkept at lower temperatures than the ambient cell isused to control the iodine vapor pressure. Becausethe amount of absorbing iodine in the gas phase sen-sitively depends on the cool finger’s temperature, ithas to be stabilized with high accuracy. The HSRLexperiments were conducted with cool finger temper-atures ranging from 28 °C to 45 °C resulting in vaporpressures between 53 and 159 Pa. The ambient cellwas kept at temperatures 5° higher than the coolfinger to avoid sublimation of iodine vapor. The cellwindows are wedged and antireflection coated toavoid interference effects between the window sur-faces. At each end of the cell there is an additionalwindow for thermal isolation.

The receiver module also provides avalanche pho-todiodes (APDs) for the detection of 1064 nm paralleland cross-polarized backscatter (not shown in Fig. 4).A further APD is installed for measurements at935 nm in case water vapor measurements are con-ducted simultaneously using the DIAL technique[33]. Here we focus only on HSRL operation. Thequantities measured with this instrument are pro-files of extinction and backscatter coefficients at532 nm, linear depolarization at 532 and 1064 nm,and attenuated backscatter at 1064 nm with highhorizontal and vertical resolution. Thus, the atmo-spheric aerosol can be characterized and classified byits lidar ratio, its depolarization ratio, and its green-

to-infrared backscatter ratio. All these quantities donot depend on aerosol concentration. Another uniquefeature of an airborne HSRL is its ability to measurethe horizontal variation of aerosol optical thicknessalong the flight path.

4. Case Study of Results from the Saharan MineralDust Experiment

The SAMUM is a joint research project of severalGerman institutes aiming at the characterization ofSaharan dust referring to its optical, microphysical,chemical, and radiative properties. The first fieldphase of the SAMUM took place in May–June 2006 insouthern Morocco including ground stations in Ouar-zazate (30.9°N, 6.9°W) and Zagora (30.3°N, 5.8°W)from where ground-based measurements with activeand passive remote sensing as well as in situ instru-ments were conducted. Two research aircraft wereengaged during the first SAMUM field phase.Equipped with the nadir-viewing HSRL and an ex-tensive set of aerosol in situ probing instruments, theDLR Falcon research aircraft performed airbornemeasurements of various dust properties. During thefirst SAMUM field phase, the DLR Falcon aircraftwas stationed in Casablanca from where scienceflights started either southbound to sound the atmo-sphere over the ground stations or northbound forobservation of long-range transport of desert dust. Intotal eight measurement flights were performed dur-ing the first SAMUM field phase. For validation pur-poses, airborne HSRL measurements were comparedto ground-based Raman lidar and sunphotometermeasurements conducted by the Leibniz Institute forTropospheric Research (IFT) at Ouarzazate. In thefollowing, the performance of the HSRL will be shownon the basis of two case studies.

Figure 5 shows the route of the DLR Falcon re-search aircraft on 4 June with takeoff at 09:15 UTC inCasablanca. The flight altitude is color coded in thediagram. The main objectives of this flight missionwere to measure the vertical and horizontal distribu-tion of Saharan dust south of the High Atlas moun-

Fig. 5. Flight path of the DLR Falcon on 4 June between 09:15and 12:35 UTC. The flight level is color coded.

Fig. 4. Schematic of the receiver module used for detection at532 nm. BS, beam splitter; CF, cool finger; IF, interference filter; L,lens; M, mirror; PBS, polarization beam splitter; PMT, photomul-tiplier; WP, quarter-wave plate.


tain range at high altitude over Ouarzazate andZagora and to conduct in situ measurements at sev-eral lower levels within the dust layer. Moreover, theflight was coordinated with satellite overpassesof the multiangle imaging spectroradiometer (MISR)and the medium resolution imaging spectrometer(MERIS) for validation purposes. Figure 6 shows thecross sections of the HSRL backscatter ratio andaerosol extinction coefficient measured during thefirst flight leg at high altitude over Ouarzazate. Themean flight altitude on this leg extending over 165km from north to south was 9.3 km above sea level(ASL). The dust distribution south of the High Atlasmountain range shows a complex stratified structure.In the backscatter ratio cross section several separatelayers can be distinguished. The uppermost layer ex-ceeds a height of 4.5 km ASL. The southern part ofthe cross section shows a distinct dust plume with amaximum aerosol concentration at 3 km ASL. Thebackscatter ratio of the parallel-polarized combinedchannel ranges from 2 up to 6 within the differentlayers. The extinction coefficient cross section showscharacteristics similar to the backscatter cross sec-tion. The aerosol extinction coefficient reaches valuesof up to 0.25 km�1 within the layers of high dustloading. The comparatively high noise level in thisdiagram is due to the numerical differentiation nec-essary to derive the extinction coefficient from theAOT.

Figure 7 shows an average of the lidar signals mea-sured in the parallel-polarized combined and molec-ular channel as a function of distance from thetransmitter. These signals were averaged over 13 saround 09:48 UTC at 30.85°N, �6.90°E. A total of 25profiles were averaged from which each profile is the

mean profile of 52 laser shots. For background lightcorrection the mean pretrigger value is subtractedfrom the measured raw data profiles. Both signalsshow the quadratic decay with distance. The signal inthe combined channel features the structures of theaerosol layers and the ground echo, whereas the sig-nal in the molecular channel decreases monotoni-cally. These two signals are the basis of retrieving theaerosol backscatter and extinction coefficients. Thetransmitted fraction of aerosol backscatter throughthe iodine filter strongly depends on the injectionseeding conditions of the master oscillator and gen-erally can vary during airborne operation. Hence, amore appropriate way to determine �a is to analyzethe ground return detected in both channels. For thispurpose the two channels are cross calibrated to ad-just the same sensitivity and the ratio of the groundreturn detected in both channels is analyzed. In thiscase �a is determined to be �5 � 10�4.

Figure 8(a) shows the profiles of the attenuatedbackscatter ratio determined from these signals. Theparallel-polarized combined channel indicates thedifferent aerosol layers with a top layer height of 4.7km as well as the ground level at 1.3 km ASL. Thebackscatter ratio of the parallel-polarized molecularchannel slightly increases along with temperature inthe free troposphere and declines monotonicallywithin the dust layer due to aerosol extinction. Forcalibration of the molecular channel, the iodine ab-sorption spectrum of the filter was measured duringthe flight and the partial transmission of molecularbackscatter through the iodine cell �m�T, p� was cal-culated using the Tenti code [21]. The profiles of at-mospheric temperature and pressure were takenfrom a radiosonde launched at Ouarzazate the sameday at 10:39 UTC. For normalization of the combinedchannel signal, a background backscatter coefficientof 1 � 10�5 sr�1 km�1 was assumed in the free tropo-sphere at 8.3 km ASL. This value corresponds to abackscatter ratio of 1.005. At this height completeoverlap of the transmitter’s and the laser’s field ofview is assured and no stratification due to the pres-ence of aerosols was observed in the backscatter sig-

Fig. 6. North–south cross section of the HSRL backscatter ratioand the aerosol extinction coefficient during the measurementflight on 4 June 2006 at 09:45 UTC. The altitude is above sealevel.

Fig. 7. Measured atmospheric backscatter in the parallel-polarized combined and molecular channel as a function of dis-tance from transmitter averaged over 13 s on 4 June 2006.


nals. The assumption of the background backscattercoefficient was based on the analysis of the in situmeasurement of the aerosol size distribution. Thein situ instrumentation included a set of condensa-tion particle counters (CPC) partially connected todiffusion screens (size range � 5 nm), a passive cavityaerosol spectrometer probe (PCASP 100�, size range0.1–3 �m) and a forward scattering spectrometerprobe (FSSP 300, size range 0.3–20 �m) to measurethe aerosol size distribution with particle diametersranging from 0.005–20 �m. The background back-

scatter coefficient was calculated using the aerosolsize distribution measured at flight altitude during09:38 and 09:50 UTC by means of an inversion algo-rithm [36] to be 0.6 � 10�5 sr�1 km�1.

Figure 8(b) shows the total backscatter coefficientas calculated using Eq. (18). The vertical resolution is15 m. The error bars indicate the 3� statistical devi-ation from the average. The statistical deviation isbased on the variance of the signals including theatmospheric variability among shot and electronicnoise. The backscatter coefficient of the distinct

Fig. 8. (a) Backscatter ratio of the parallel-polarized molecular and combined channel, (b) total aerosol backscatter coefficient, (c) aerosoloptical thickness, (d) aerosol extinction coefficient, (e) aerosol depolarization ratio, (f) lidar ratio. Profiles averaged over 13 s on 4 June 2006.Vertical resolution is 15 m except 540 m for (d) and (f). The error bars denote the 3� statistical deviation.


dust plume at 3 km ASL amounts to �5.9 � 0.5�� 10�3 sr�1 km�1. The relatively large statistical er-ror within this dust layer is attributed to atmosphericvariations within the distance of approximately 3 kmtraveled with the aircraft during the averaging time.Figure 8(c) shows the increase of AOT from the freetroposphere toward the ground with the error barsindicating the 3� deviation from the mean. Withinthe dust plume at 3 km ASL the AOT profile showsthe steepest increase with height, indicating maxi-mum aerosol extinction at this altitude. At groundlevel the AOT amounts to 0.37 � 0.02, i.e., approxi-mately 1�3 of the impinging green radiation is scat-tered and absorbed by the dust loaded atmosphere.For calculation of the extinction coefficient the AOTprofile is differentiated with respect to height using aSavitzky–Golay filter of first order. A total of 51 rangebins is used for the filter resulting in a FWHM of thefilter kernel of 540 m. Figure 8(d) shows the corre-sponding extinction coefficient profile. The error barsindicate the 3� deviation from the mean with regardto the noise reducing property of the smoothing ker-nel.

Figure 8(e) shows the aerosol depolarization ratioof the dust from ground to 4.5 km ASL. As can beseen, the aerosol depolarization ratio shows only alittle variation over the entire dust layer from groundto 4.7 km ASL. The values range from 28% to 32%and the error bars denote the 3� statistical deviation.For calculating the lidar ratio, the vertically highlyresolved backscatter coefficient profile was smoothedin order to adjust its vertical resolution to that of theextinction coefficient profile. The vertical variation ofthe lidar ratio is shown in Fig. 8(f). The lidar ratioprofile has a minimum value of 41 sr within the dis-tinct dust layer at 3 km ASL. Toward the ground andthe upper layers the lidar ratio increases to 55 srwithin the uppermost layer and around 60 sr at thelowest layer. Both the aerosol depolarization and thelidar ratio do not depend on concentration and canthus be used for aerosol characterization.

5. Error Analysis

Systematic errors in the measurement of the back-scatter coefficient may arise from uncertainties of themeasured quantities RC

��r�, RM��r�, �a�r� and the cal-

culated quantities �m�r�, �a�r�, m��r�. The uncertainty

�RC��r�, is mainly due to normalization, where the

magnitude of background aerosol at a referenceheight within the free troposphere has to be esti-mated. For error calculation, the background aerosolcoefficient was estimated to range from zero (puremolecular scattering) to the double value used for theanalysis in Section 4. As stated before, the assump-tion of the background backscatter coefficient wasbased on the analysis of in situ measured aerosol sizedistributions. Comparisons with lidar measurementshave shown that the deviation of calculated frommeasured data is �30% [37]. However, the back-ground aerosol is close to the detection limit of theinversion method. Thus, the rather large errorbounds were chosen to include the error of the calcu-

lated backscatter coefficient. The uncertainty ��m�r�is mainly induced by laser frequency fluctuations andvariations in the atmospheric temperature. Figure 9shows the relative deviation of ��m�T� for laser fre-quency offsets from �15 to �15 MHz. As can be seen,the relative deviations of �m�T� decrease with tem-perature, i.e., the frequency fluctuations become lesssignificant with the increasing Doppler broadeningof the molecular backscatter spectrum. Though thelong-term laser frequency fluctuations were keptbelow �15 MHz, these bounds include short-termfrequency variations inherent to the Q-switch build-up-time reduction technique. The variation of the at-mospheric temperature profile was determined bythe difference of the radiosonde measurement at10:39 UTC and the aircraft temperature measure-ment during a dive at 09:50 UTC. The mean temper-ature difference was 0.8 K over the measurementrange down to 1.9 km ASL. For error calculation atemperature uncertainty of 1 K was chosen.

The uncertainties ��a and RM��r� are mainly caused

by changes in the laser’s spectral purity due to mul-timode laser shots. As already described, the spectralpurity is derived from the average transmission ofthe ground return through the iodine filter during themeasurement interval. The error ��a is determinedfrom the statistical deviation of the filter transmis-sion of each of the 25 profiles from the average valueand amounts to 2 � 10�4. The uncertainty of theaerosol depolarization ratio ��r� is mainly due tothe accuracy of the 45° calibration measurement. Themechanical precision and the reproducibility of theinitial adjustment is estimated with a tolerance of0.6° resulting in a nearly constant error of the aerosoldepolarization ratio profile of 4 percentage points.Table 2 summarizes the error sources and their esti-mated uncertainties.

To show the specific influence of the individual er-ror sources, their contribution to the relative system-atic error has been calculated separately. Figure10(a) shows the profiles of the relative systematicerrors of the backscatter coefficient shown in Fig. 8(b)caused by the individual error sources. For error cal-culation, the uncertainties listed in Table 2 wereused. As can be seen the uncertainty in the depolar-

Fig. 9. Relative deviation of �m��, T� induced by laser frequencyfluctuations of �5, �10, and �15 MHz, respectively.


ization measurement induces the largest individualerror of �3% over the dust layer. Laser frequencyfluctuations, uncertainties associated with the nor-malization and the atmospheric temperature causerelative errors between 1% and 2.5%. Under theseconditions the variations of the laser’s spectral purityhave the least impact on the backscatter measure-ment. To calculate the total systematic error the in-dividual errors have been added in quadrature, i.e.,as the square root of the sum of the squares. This hasbeen done assuming statistically independent influ-ence of the background aerosol, the laser’s frequencydrift and spectral purity, the depolarization measure-ment, and the atmospheric temperature. The totalsystematic error of the backscatter measurement is4% to 5% within the dust layer from the ground up to4.5 km ASL and comparable to the statistical error.

In the case of the aerosol extinction coefficient, thesystematic error was derived from the error of theAOT measurement by means of differentiation. Fig-ure 10(b) shows the profiles of the relative systematic

error of the aerosol extinction coefficient caused bythe different error sources. In contrast to the back-scatter measurement, spectral impurity caused bymultimode laser shots influences the extinction mea-surement considerably and errors due to normaliza-tion, atmospheric temperature variations, and laserfrequency fluctuations can be neglected. The magni-tude of the error induced by spectral impurity signif-icantly depends on the ratio of �a��m�r�. The smallvalue of �a in our experiments allows for the relativeimprecise determination of ��a. Again, the individualerrors haven been added in quadrature to calculatethe total error. The relative systematic error of theextinction coefficient is less than 3% within the dustlayer, so that the statistical deviation with 10%–20%is dominant. However, the statistical error can fur-ther be reduced by means of longer averaging timesand increasing length of the smoothing window. Theerror analysis of the AOT measurement using thediscussed uncertainties achieves an AOT detectionlimit of 0.008.

6. Comparison with Raman Lidar and Sun PhotometerMeasurements

During the SAMUM field campaign both Raman li-dar and sunphotometer measurements as well as ra-dio soundings were conducted at the Ouarzazateground station by IFT. The Raman lidar BackscatterExtinction lidar-Ratio Temperature Humidity Pro-filing Apparatus (BERTHA) utilizes two frequency

Fig. 11. Comparison of IFT Raman and DLR HSRL (a) backscat-ter, (b) extinction, and (c) lidar ratio profiles at 532 nm over Ouar-zazate on 3 June 2006 at 04:14 UTC. The vertical resolution is300 m for the Raman measurements and the HSRL extinctioncoefficient and lidar ratio. The vertical resolution is 15 m for theHSRL backscatter coefficient profile. The averaging time is 11 s forthe HSRL and 1800 s for the Raman lidar, respectively. The errorbars indicate the 3 � statistical error.

Table 2. Error Sources and Their Estimated Uncertainties

Parameter Estimate

Background aerosol backscatter coefficient 0 � �a(r0) � 2 � 10�5 km�1 sr�1

Laser frequency drift 15 MHzLaser spectral purity 99.96% 0.03%Angular tolerance of depolarization calibration 0.6°Atmospheric temperature 1 K

Fig. 10. Relative systematic errors of the (a) aerosol backscatterand (b) extinction measurement arising from the uncertainties in(1) spectral purity, (2) atmospheric temperature, (3) normalization,(4) laser frequency, (5) depolarization ratio, and (6) total error.


stabilized Nd:YAG and two Ti:sapphire lasers togenerate laser emission at six wavelengths. Thereceiver provides 14 channels altogether to detectatmospheric backscatter separated in elastic, vibra-tional, and rotational Raman shifted and depolar-ized backscatter. A detailed description of thelidar system can be found in [38]. During SAMUMthe vibrational (387 and 607 nm) and rotational�532 nm� Raman channels were used to directly mea-sure profiles of the aerosol extinction and backscattercoefficients. For intercomparison with the airborneHSRL, the Raman lidar measurements were coordi-nated with the aircraft overpasses. Because of atmo-spheric variability, the averaging in time and spacewas chosen to be as short as possible for both instru-ments. Figure 11 shows a comparison of airborneHSRL and ground-based Raman lidar measurementstaken at night on 3 June 2006. The HSRL profileswere averaged over 11 s at 04:14 UTC during thedirect overpass of the aircraft over the Raman lidar atOuarzazate. The error bars indicate the 3� statisticalerror. For intercomparison, the Raman lidar mea-surements were averaged from 03:56 to 04:26 UTC.The vertical resolution is 300 m in case of the Ramanlidar measurements and the HSRL extinction coeffi-cient and lidar ratio profile. The HSRL backscattercoefficient was measured with 15 m vertical resolu-

tion. As can be seen, the results of HSRL and Ramanlidar agree very well within the measurement uncer-tainty. Because of the different averaging times ofboth instruments, temporal and spatial variations ofthe dust layers over Ouarzazate must be taken intoaccount.

To provide a more quantitative comparison of bothinstruments, Fig. 12 shows a correlation of the ex-tinction coefficient profiles derived from measure-ments with both instruments for six aircraftoverpasses during the SAMUM campaign. Typicalaveraging times for the HSRL range from 5–10 s inorder to keep spatial atmospheric variations small.The averaging intervals for the Raman lidar wereusually 60–120 min depending on the homogeneity ofthe dust layer. Assuming a typical wind speed of5 m�s within the dust layers the averaging periodscorrespond to a horizontal range of 1–2 km in case ofthe HSRL measurements and of 18–36 km for theRaman lidar measurements. A linear least-squaresfit of the 98 data points leads to a slope value of 0.97with an error of 0.03 and a y-axis intercept of 0.004indicating an excellent correlation of the data. A lin-ear “robust” fit using a least-absolute deviationmethod, which can be more applicable in case of out-liers and strong scattering of the data, results in aslope value of 0.98 and a y-axis intercept of 0.003.

For validation purposes, the HSRL measurementof the AOT has been compared to sunphotometermeasurements conducted by IFT at Ouarzazatethroughout the SAMUM field campaign. During oneflight mission toward Portugal the HSRL measure-ment could be compared to an AERONET sunpho-tometer located at Cabo da Roca (38.5°N, 9.5°W). InTable 3 the AOT measured by sunphotometers andHSRL are listed. The AOT measured by HSRL wasdetermined from the AOT profile as the mean of thelast five range bins above ground. The correspondingerror is given by the standard deviation. To minimizethe influence of horizontal atmospheric variations,the HSRL averaging times are less than 20 s. TheAOT measured by the airborne HSRL at 532 nm iscompared to sunphotometer measurements at 440,675, and, if available, 500 nm. Within the limits ofuncertainty the AOT measured with both instru-ments deviate by only 3% in the worst case. Devia-tions between these two instruments can arise fromseveral origins. First, the airborne nadir-viewing

Fig. 12. Correlation of extinction coefficient profiles measured byIFT Raman lidar, BERTHA, and DLR HSRL at six aircraft over-passes during the first SAMUM field phase in May–June 2006.Squares denote the comparison to rotational Raman lidar; circlesindicate vibrational Raman measurements. The error bars indi-cate a relative statistical deviation of 15%.

Table 3. Comparison of Aerosol Optical Thickness Measured with HSRL and Sunphotometers During the SAMUM Field Campaign in 2006a

Date LocationTime(UTC)

SP440 nm

SP500 nm

SP675 nm

HSRL532 nm

NormalizationHeight

(km ASL)

19 May Ouarzazate 10:55:05 0.431 0.423 0.392 0.38 0.03 8.227 May Cabo da Roca 10:20:53 0.439 — 0.411 0.42 0.03 7.33 June Ouarzazate 07:31:40 0.435 0.431 0.413 0.40 0.02 8.54 June Ouarzazate 09:56:50 0.407 0.408 0.368 0.40 0.02 8.44 June Ouarzazate 10:41:51 0.447 0.450 0.429 0.44 0.02 8.4

aSP: aerosol optical thickness measured by sunphotometry at the respective wavelength; normalization height: height ASL chosen fornormalization.


HSRL determines the AOT between the normaliza-tion height, which is close to the flight level, and theground. In contrast, a sunphotometer measures theAOT of the entire atmospheric column. Thus, com-pared to sunphotometry, the HSRL values of the AOTare systematically reduced by the amount of AOTabove flight level. However, this residual amount isusually very small compared to the AOT of the loweratmospheric column and can be neglected in case ofhigh flight altitudes. Second, deviations of the resultsof both instruments can be attributed to the measure-ment geometry. The airborne HSRL strictly mea-sures in nadir configuration, whereas a Sun trackingphotometer has arbitrary hemispherical viewing an-gles. Thus, horizontal variations in the dust stratifi-cation and composition can lead to deviations of themeasured AOT, especially in the early morning orlate evening.

7. Conclusions

We have developed an airborne high spectral resolu-tion lidar based on an iodine vapor absorption filterand a high-power, frequency-doubled Nd:YAG laser.The HSRL separates the aerosol contribution fromthe atmospheric backscatter and thereby measuresaerosol backscatter and extinction coefficients di-rectly. With the use of a dual-pass optical configu-ration, strong suppression of aerosol backscatterassociated with a high degree of transmission of mo-lecular backscatter and compact mechanical dimen-sions is achieved. Due to the system’s favorabletransmitter and receiver properties, atmosphericcross sections of aerosol extinction, optical thickness,and lidar ratio can be measured with high verticaland horizontal resolution. In contrast to Raman li-dars, the averaging times of the HSRL can be as shortas a few seconds. Thus, the HSRL can be used ad-vantageously for airborne operation. Its performanceaboard the DLR Falcon research aircraft during thefirst SAMUM field phase proved to be stable andreliable. With this instrument optical properties ofpure Saharan dust were measured close to its sourceregions for the first time, to the best of our knowledge.The relative systematic error of the backscatter co-efficient has been evaluated to be less than 5%for backscatter coefficients higher than 7.2 � 10�4

sr�1 km�1. For the extinction measurement, the rel-ative systematic error is less than 3% within the dustlayer. The HSRL results were compared to Ramanlidar measurements as well as sunphotometry. In thecase of Raman lidar measurements the comparison ofextinction and backscatter profiles has shown agree-ment of both instruments within their combined lim-its of uncertainty. In the case of sunphotometers, theagreement of the measured AOT is excellent consid-ering the different measurement ranges and geome-tries of both techniques. The airborne HSRL is avaluable instrument to study the optical properties ofvarious aerosol types and their transportation andmixing processes.

This work was supported by the German ResearchFoundation (DFG) under grant FOR 539-PE 632�2,the European Space Agency (ESA) under contract19429�06�NL�AR, and the European Fleet for Air-borne Research (EUFAR) as part of the Desert Dustand Biomass Burning Aerosols over Portugal(DARPO) project. We thank Albert Ansmann (IFT)and Frank Wagner of the Centro de Geofisica, Uni-versidade de Evora, Portugal, for providing the sun-photometer data and Bernadett Weinzierl (DLR) forthe analyses of the aerosol in situ measurements.Two anonymous reviewers provided helpful sugges-tions and comments. We greatly appreciate the scien-tific coordination of the Falcon flight missions byAndreas Petzold (DLR) during the SAMUM field cam-paign in Morocco.

References1. J. A. Coakley, R. D. Cess, and F. B. Yurevich, “The effect of

troposheric aerosols on the earth’s radiation budget: a param-eterization for climate models,” J. Atmos. Sci. 40, 116–138(1983).

2. J. D. Klett, “Stable analytical inversion solution for processinglidar returns,” Appl. Opt. 20, 211–220 (1981).

3. A. Ansmann and D. Müller, “Lidar and atmospheric aerosolparticles,” in Lidar: range-resolved optical remote sensing of theatmosphere, C. Weitkamp, ed. (Springer, 2005), pp. 105–142.

4. D. N. Whiteman, I. Veselovskii, M. Cadirola, K. Rush, J.Comer, J. R. Potter, and R. Tola, “Demonstration measure-ments of water vapor, cirrus clouds, and carbon dioxide usinga high-performance Raman lidar,” J. Atmos. Ocean. Technol.24, 1377–1388 (2007).

5. S. T. Shipley, D. H. Tracy, E. W. Eloranta, J. T. Trauger, J. T.Sroga, F. L. Roesler, and J. A. Weinman, “High spectral reso-lution lidar to measure optical scattering properties of atmo-spheric aerosols. 1: Theory and instrumentation,” Appl. Opt.22, 3716–3724 (1983).

6. H. Shimizu, S. A. Lee, and C. Y. She, “High spectral resolutionlidar system with atomic blocking filters for measuring atmo-spheric parameters,” Appl. Opt. 22, 1373–1381 (1983).

7. R. J. Alvarez, L. M. Caldwell, Y. H. Li, D. A. Krueger, andC. Y. She, “High-spectral-resolution lidar measurement of tro-posheric backscatter-ratio with barium atomic blocking fil-ters,” J. Atmos. Ocean. Technol. 7, 876–881 (1990).

8. J. W. Hair, L. M. Caldwell, D. A. Krueger, and C. Y. She,“High-spectral-resolution lidar with iodine-vapor filters: mea-surement of atmospheric-state and aerosol profiles,” Appl. Opt.40, 5280–5294 (2001).

9. P. Piironen and E. W. Eloranta, “Demonstration of a high-spectral-resolution lidar based on an iodine absorption filter,”Opt. Lett. 19, 234–236 (1994).

10. Z. Liu, I. Matsui, and N. Sugimoto, “High-spectral-resolutionlidar using an iodine absorption filter for atmospheric mea-surements,” Opt. Eng. 38, 1661–1670 (1999).

11. U. Wandinger, D. Müller, C. Böckmann, D. Althausen, V. Mat-thias, J. Bösenberg, V. Weiss, M. Fiebig, M. Wendisch, A.Stohl, and A. Ansmann, “Optical and microphysical character-ization of biomass-burning and industrial-pollution aerosolsfrom multiwavelength lidar and aircraft measurements,” J.Geophys. Res. 107(D21), 8125, doi:10.1029/2000JD000202(2002).

12. R. Ferrare, C. Hostetler, J. Hair, A. Cook, D. Harper, S. Bur-ton, M. Clayton, A. Clarke, P. Russell, and J. Redemann, “Air-borne high spectral resolution lidar aerosol measurementsduring MILAGRO and TEXAQS�GOMACCS,” presented at


the 87th Annual Meeting of the American Meteorological So-ciety, San Antonio, Texas, 14–18 January 2007.

13. A. Stoffelen, J. Pailleux, E. Källn, J. M. Vaughan, L. Isaksen,P. Flamant, W. Wergen, E. Andersson, H. Schyberg, A.Culoma, R. Meynart, M. Endemann, and P. Ingmann, “TheAtmospheric Dynamics Mission for global wind field measure-ment,” Bull. Am. Meteorol. Soc. 86, 73–87 (2005).

14. A. Ansmann, U. Wandinger, O. Le Rille, D. Lajas, and A. G.Straume, “Particle backscatter and extinction profiling with thespaceborne high-spectral-resolution Doppler lidar ALADIN:methodology and simulations,” Appl. Opt. 46, 6606–6622(2007).

15. A. Petzold, B. Weinzierl, M. Esselborn, G. Ehret, M. Fiebig, A.Fix, C. Kiemle, K. Rasp, and M. Wirth, “Technical support ofthe EarthCARE Mission for the validation of spaceborne aero-sol products during the Saharan Mineral Dust Experiment,”European Space Research and Technology Centre (ESTEC)contract 19429�06�NL�AR (2007).

16. H. C. van de Hulst, Light Scattering by Small Particles (Wiley,1975).

17. R. Miles, W. Lempert, and J. Forkey, “Laser Rayleigh scatter-ing,” Meas. Sci. Technol. 12, 33–51 (2001).

18. A. Young, “Rayleigh scattering,” Appl. Opt. 20, 533–535(1981).

19. C. Y. She, “Spectral structure of laser light scattering revisited:bandwidths of nonresonant scattering lidars,” Appl. Opt. 40,4875–4884 (2001).

20. S. Gerstenkorn and P. Luc, Atlas du Spectre D’Absorptionde la Molecule D’Iode. Atlas III (Edition du CNRS, 1978).

21. G. Tenti, C. D. Boley, and R. C. Desai, “On the kinetic modeldescription of Rayleigh–Brillouin scattering from moleculargases,” Can. J. Phys. 52, 285–290 (1974).

22. A. T. Young and G. W. Kattawar, “Rayleigh-scattering lineprofiles,” Appl. Opt. 22, 3668–3670 (1983).

23. J. N. Forkey, Precision Optics Corporation, 22 East Broadway,Gardner, Massachusetts 01440 (personal communication,2005).

24. J. N. Forkey, W. R. Lempert, and R. B. Miles, “Corrected andcalibrated I2 absorption model at frequency-doubled Nd:YAGwavelengths,” Appl. Opt. 36, 6729–6738 (1997).

25. G. W. Kattawar, A. T. Young, and T. J. Humphreys, “Inelasticscattering in planetary atmospheres. 1. The Ring effect, with-out aerosols,” Astrophys. J. 243, 1049–1057 (1981).

26. R. M. Measures, Laser Remote Sensing: Fundamentals andApplications (Wiley, 1984).

27. B. A. Bodhaine, N. B. Wood, E. G. Dutton, and J. R. Slusser,“On Rayleigh optical depth calculations,” J. Atmos. Ocean.Technol. 16, 1854–1861 (1999).

28. A. Savitzky and M. J. E. Golay, “Smoothing and differentiationof data by simplified least square procedures,” Anal. Chem. 36,1627–1639 (1964).

29. J. Biele, G. Beyerle, and G. Baumgarten, “Polarization lidar:corrections of instrumental effects,” Opt. Express 7, 427–435(2000).

30. A. Behrendt and T. Nakamura, “Calculation of the calibrationconstant of polarization lidar and its dependency on atmo-spheric temperature,” Opt. Express 10, 805–817 (2002).

31. R. A. Ferrare, D. D. Turner, L. H. Brasseur, W. F. Feltz, O.Dubovik, and T. P. Tooman, “Raman lidar measurements ofthe aerosol extinction-to-backscatter ratio over the SouthernGreat Plains,” J. Geophys. Res. 106, 20333–20348 (2001).

32. M. Tesche, A. Ansmann, D. Müller, D. Althausen, R.Engelmann, M. Hu, and Y. Zhang, “Particle backscatter, ex-tinction, and lidar ratio profiling with Raman lidar in southand north China,” Appl. Opt. 46, 6302–6308 (2007).

33. G. Poberaj, A. Fix, A. Assion, M. Wirth, C. Kiemle, and G.Ehret, “Airborne all-solid-state DIAL for water vapour mea-surements in the tropopause region: system description andassessment of accuracy,” Appl. Phys. B 75, 165–172 (2002).

34. G. Ehret, H. H. Klingenberg, U. Hefter, A. Assion, A. Fix, G.Poberaj, S. Berger, S. Geiger, and Q. Lü, “High peak andaverage power all-solid-state laser systems for airborne lidarapplications,” LaserOpto 32, 29–37 (2000).

35. R. L. Schmitt and L. A. Rahn, “Diode-laser-pumped Nd:YAGlaser injection seeding system,” Appl. Opt. 25, 629–633 (1986).

36. M. Fiebig, C. Stein, F. Schröder, P. Feldpausch, and A. Petzold,“Inversion of data containing information on the aerosol par-ticle size distribution using multiple instruments,” J. AerosolSci. 36, 1353–1372 (2005).

37. A. Petzold, M. Fiebig, H. Flentje, A. Keil, U. Leiterer, F.Schröder, A. Stifter, M. Wendisch, and P. Wendling, “Verticalvariability of aerosol properties observed at a continental siteduring the Lindenberg Aerosol Characterization Experiment(LACE 98),” J. Geophys. Res. 107(D21), 8125, doi:10.1029/2001JD001043 (2002).

38. D. Althausen, D. Müller, A. Ansmann, U. Wandinger, H. Hube,E. Clauder, and S. Zörner, “Scanning 6-wavelength 11-channelaerosol lidar,” J. Atmos. Ocean. Technol. 17, 1469–1482(2000).


airborne high spectral resolution lidar for measuring aerosol extinction and backscatter...

Documents