Abstract

A Rayleigh–Mie–Raman lidar has been installed and is operating in the Polar Environment Atmospheric Research Laboratory at Eureka in the High Arctic (79°59′N, 85°56′W) as part of the Canadian Network for the Detection of Atmospheric Change. The lidar operates in both the visible and ultraviolet and measures aerosol backscatter and extinction coefficients, depolarization ratio, tropospheric temperature, and water vapor mixing ratio. Variable field of view, aperture, and filtering allow fine-tuning of the instrument for different atmospheric conditions. Because of the remote location, operations are carried out via a satellite link. The instrument is introduced along with the measurement techniques utilized and interference filter specifications. The temperature dependence of the water vapor signal depends on the filter specifications, and this is discussed in terms of minimizing the uncertainty of the water vapor mixing ratio product. Finally, an example measurement is presented to illustrate the potential of this instrument for studying the Arctic atmosphere.

1. Introduction

The Canadian Network for the Detection of Atmospheric Change (CANDAC), a collaboration between eight Canadian university departments, the Canadian Space Agency, and Environment Canada, has established a suite of atmospheric instruments at Eureka, Nunavut, Canada (79°59′N, 85°56′W). Instruments are distributed among three facilities, which combined form the Polar Environment Atmospheric Research Laboratory (PEARL). The Eureka weather station, associated infrastructure, and research laboratories facilitate year-round study of the polar atmosphere.

Disproportionate climatic change in the Arctic has garnered significant attention, particularly the positive feedback effect of falling surface albedo, due to the shrinking ice pack (Serreze and Francis 2006 and references therein). Recent work has shown that water vapor provides a strong feedback mechanism for regional atmospheric change (Soden and Held 2006; Winton 2006). For example, in a coupled climate model that allowed the surface albedo component to be fixed, Graversen and Wang (2009) show that water vapor constitutes a more significant feedback mechanism in the changing Arctic climate than the albedo. Knowledge of the vertical distribution of the water vapor is important in understanding the associated feedback processes (Bony et al. 2006). Measurements of water vapor profiles are currently provided up to twice daily by radiosondes; however, to capture the dynamic nature of this field and elucidate the key processes, measurements of higher temporal resolution are required. Lidar measurements of tropospheric water vapor have been collected at Ny-Ålesund since June 2001 (Gerding et al. 2004), and the first results from a Raman lidar at Sondrestrom, Greenland, have also recently been published (Neely and Thayer 2011).

During all seasons but summer, the boundary layer tends to be very stable with radiative cooling of the surface producing a strong temperature inversion (Curry et al. 1996 and references therein). During winter at Eureka, it extends from the ground to approximately 1-km altitude with an average lapse rate of approximately 15 K km−1 (Kahl et al. 1992; Lesins et al. 2010). This inversion has a significant influence on radiatively important species; the stable air inhibits the removal of Arctic haze aerosols (Shaw 1995), contains pervasive boundary layer ice crystals (Bourdages et al. 2009), and is a source of moisture for the formation of clouds (Tjernström et al. 2004). Liquid water and mixed-phase clouds are a significant contributor to surface warming in all but high summer, when their albedo effect outweighs the longwave positive forcing (Shupe and Intrieri 2004). Characterizing the inversion layer structure and processes will improve our understanding of the polar radiation budget.

Lidars have been used for some time to quantify some or all of these properties, although the technological issues of operating in the High Arctic are not inconsiderable. Lidar-based investigations at Eureka commenced in 1993 with the installation of an ozone elastic backscatter differential absorption lidar (DIAL) (Whiteway and Carswell 1994) in the then Arctic Stratospheric Ozone (ASTRO) Observatory. This is now the primary PEARL facility, located at 610 m MSL. This system has been used to measure wintertime and springtime ozone (Donovan et al. 1995), gravity wave activity (Duck et al. 2001), middle atmosphere temperatures (Duck et al. 2000), and more recently some water vapor profiles (Moss 2010). From 1993 to 1997, wintertime aerosol measurements were also made with a Rayleigh–Mie depolarization lidar at ASTRO, looking at tropospheric clouds and Arctic haze (Ishii et al. 1999). Starting in 2005 a high-spectral-resolution lidar has been operating at the Zero Altitude PEARL Auxiliary Laboratory (ØPAL), at 10 m MSL, nearly continuously measuring backscatter and depolarization (Eloranta et al. 2006). A tropospheric ozone DIAL with full diurnal capabilities has been installed at ØPAL since 2008.

To address the need for high-resolution profile measurements of water vapor, temperature, and aerosols through the troposphere to the lower stratosphere, the new CANDAC Rayleigh–Mie–Raman lidar (CRL) was built and installed at the ØPAL facility. Collocated with the CRL are a millimeter cloud radar (Moran et al. 1998), an Atmospheric Emitted Radiance Interferometer (Knuteson et al. 2004), a microwave radiometer (Turner et al. 2007), and sun and star photometers (Holben et al. 1998), along with the aforementioned high-spectral-resolution lidar and ozone DIAL. The high-spectral-resolution lidar and microwave radiometer are operated under the Study of Environmental Arctic Change program (SEARCH; http://www.arcus.org/search). At 80°N, Eureka sees a high number of overpasses by the Cloud–Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) satellite due to its orbit inclination of 98.2° (Winker et al. 2009). This means the CRL is well located for combined temporal–spatial comparisons with measurements by the onboard Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP). Eureka is an Environment Canada weather station and so provides meteorological data and twice-daily radiosondes. This suite of instruments allows a significant number of complementary datasets in the lower and middle atmosphere to be taken at this one site. The installation of the CRL further strengthens these measurement capabilities.

The CRL has been constructed around two transmitted wavelengths, 532 and 355 nm, and eight measurement channels. Table 1 lists the primary specifications of the lidar. Elastic measurements are taken in both the visible and ultraviolet along with Raman backscatter from nitrogen. The weak Raman profiles of water vapor are measured in the ultraviolet to maximize the scattering cross section and thus signal levels (Avila et al. 2004). Two channels close to 532 nm are used to measure the anti-Stokes branch of the overlapping rotational Raman spectra of nitrogen and oxygen, which are used for tropospheric temperature retrievals. Linear depolarization measurements are made in the visible. A deliberate decision was made to use mature technologies when constructing the CRL to minimize the risks associated with the installation and operation of such an instrument in the High Arctic. The discussion of these technologies will thus be brief, and the interested reader is directed to the references given for further information. The instrument was designed to operate by remote control in a semiautonomous fashion.

Table 1.

System specifications.

System specifications.
System specifications.

In this paper we present the new instrument, its capabilities, and some example measurements. A description of the laboratory and infrastructure is given in section 2, with details of the transmitter in section 3 and the receiver in section 4. Section 5 discusses the measurements the instrument is capable of and the selection of interference filter characteristics. Particular emphasis is given to the influence of interference filter specifications on the temperature-induced uncertainty in the water vapor retrieval. Data acquisition and control software are presented in section 6. Section 7 discusses an example measurement, both in terms of the instrument and also the atmospheric processes shown. We shall recap with a brief summary in section 8.

2. Laboratory

ØPAL is a series of container-based laboratories linked by breezeways, [see Fig. 2 of Bourdages et al. (2009) for a photograph]. The CRL laboratory is a customized 6.1 m (20 ft.), high-cube shipping container built by Container and Trailer Services (Dartmouth, Nova Scotia, Canada). Three-quarters of the interior is occupied by the optical table. This holds all the optical components, including the 1-m-diameter telescope, ensuring all are vibrationally coupled. Directly above the telescope is a 1-m-diameter roof window for access to the sky with a laser-quality insert in the center for transmission of the outgoing beams. The window is protected from direct sunlight by a 1.5-m-high box, on top of which is a motorized hatch. At the other end of the laboratory is a small office space that is separated from the optical table by an instrument rack holding the seed laser, data acquisition and detector modules, and power and communication switches. To cope with widely varying environmental conditions, the laboratory is equipped with a filtered air exchange system (continuous summertime solar irradiation can result in interior temperatures reaching levels detrimental to instrument operation) and electric heaters.

At such a remote location, having a dedicated operator on-site for long-term data accumulation is prohibitively expensive. This installation has been designed to be remotely operated and monitored over a low-bandwidth satellite data link. The entire laboratory and lidar—from the roof hatch to laser operation and alignment, data collection, and safety systems—are controlled by the remote operator in real time. An on-site technician carries out scheduled facility inspections and maintenance.

3. Transmitter

This is a dual-wavelength system operating at 532 and 355 nm. To ensure high power and good beam quality at both wavelengths, two separate Surelite III-10 lasers from Continuum (Santa Clara, California) are used. This also provides some redundancy in case of a laser failure. The lasers have a repetition rate of 10 Hz and produce 380- and 240-mJ pulses of approximately 5-ns duration at 532 and 355 nm, respectively. They do not require external water cooling and consume a modest 2.1 kW each, important considerations when operating at such a remote site. To ensure a stable, narrow line width for the rotational Raman temperature measurements, the visible wavelength laser is injection seeded with a temperature-stabilized fiber laser.

Figure 1 shows the layout of the optical transmitter. The frequency-doubled and -tripled lasers produce the required 532- and 355-nm pulses, and the residual 1064- and 532-nm beams are terminated in beam dumps (BD in Fig. 1), using long-wave pass (LWP) and short-wave pass (SWP) dichroic mirrors. These are stock neodymium-doped yttrium aluminum garnet (Nd:YAG) harmonic separation dichroics with a reflectivity of 99.5% and a transmissivity of 90%. The visible and ultraviolet beams from the two lasers are nominally coaligned with a short-wave pass dichroic, labeled A in Fig. 1. Both beams are then expanded with a ×6 spherical beam expander, the angles of which are set to minimize astigmatism in the outgoing beam. After expansion the beam divergence is approximately 0.1 mrad. Although the system has been set up to be easily extended into the near infrared, this facility is not currently being used. High-reflectivity laser mirrors (labeled M in Fig. 1) are dielectric coated for 532 nm or double coated for 532 and 355 nm with reflectivities >99%.

Fig. 1.

Schematic of the lidar transmitter and telescope. LWP and SWP dichroics separate the residual wavelengths from the visible and UV beams, respectively, and these are directed into BDs. Steering mirrors A and B are used to align the visible and UV beams to the telescope’s field of view. A motorized wheel with 532- and 355-nm dichroic BSs direct light onto the power meter when required. Laser line dielectric mirrors M are used for the remainder of the beam redirection. See section 3 for details.

Fig. 1.

Schematic of the lidar transmitter and telescope. LWP and SWP dichroics separate the residual wavelengths from the visible and UV beams, respectively, and these are directed into BDs. Steering mirrors A and B are used to align the visible and UV beams to the telescope’s field of view. A motorized wheel with 532- and 355-nm dichroic BSs direct light onto the power meter when required. Laser line dielectric mirrors M are used for the remainder of the beam redirection. See section 3 for details.

The output of each laser can be measured individually with a power meter. A motorized filter wheel with 532- and 355-nm dichroic beam splitters directs each beam in turn onto the meter for periodically monitoring laser performance while angle tuning the laser second- and third-harmonic crystals. During normal operations the wheel is positioned such that both beams pass straight through unimpeded. To provide a real-time relative measure of laser output, two spectrally filtered photodiodes detect light scattered off a mirror with the peak voltages of each being saved as metadata along with the profile data.

The lidar is of coaxial design with three 100-mm turning mirrors used to direct the beams skyward, the last of which is fitted with computer-controlled actuators for beam steering. The two beams are nominally coaligned on the table and then independently aligned to the telescope’s field of view using the data acquisition system. First the ultraviolet beam is aligned using the final steering mirror, labeled B in Fig. 1, and then the visible beam is aligned using mirror A. Because of a slight beam translation as mirror A is moved, a second iteration of the alignment procedure may occasionally be required.

4. Receiver

The receiving telescope is of Dall–Kirkham design with a 1-m primary mirror, 0.4-m secondary (giving an effective area of 0.66 m2), and 9.265-m focal length. It was fabricated by Optical Structures, Inc. (Rancho Cordova, California) and was designed to be transported in a small aircraft and then assembled and aligned on-site. The telescope is suspended though a cutout in the optical table with a tertiary-fold mirror that directs the collected light into the polychromator. Given the optical efficiency of the transmitter optics discussed above, the power aperture product of the system is 2.37 and 1.51 W m2 at 532 and 355 nm, respectively, for the full 1-m aperture.

The polychromator was built by Spectral Applied Research (Richmond Hill, Ontario, Canada). There are eight measurement channels, with five in the visible and three in the ultraviolet; Fig. 2 shows the functional layout of the polychromator. Although shown here in a single plane, the visible and ultraviolet receiver chains are optically isolated from each other to minimize the possibility of cross talk. Every channel is also optically isolated downstream of the interference filters with concertina-like baffles that allow for angle tuning of the filters.

Fig. 2.

Schematic of the polychromator receiver chain. Light from the telescope is focused through the FS, collimated, and goes through the AS. LWP dichroics split the light into UV and visible arms and separate the N2 Raman line from the shorter wavelength visible channels. A beam splitter (BS) directs residual 532-nm light through the spinning polarotor. UV-enhanced Al BB mirrors are used elsewhere in the chain for redirection. NDs are used to control light levels on the elastic channels. Each channel has an IF [and a second IF (IF2) for additional blocking on two channels], and L to focus light onto each of the PMTs. Each channel is labeled; see Table 2 for wavelengths.

Fig. 2.

Schematic of the polychromator receiver chain. Light from the telescope is focused through the FS, collimated, and goes through the AS. LWP dichroics split the light into UV and visible arms and separate the N2 Raman line from the shorter wavelength visible channels. A beam splitter (BS) directs residual 532-nm light through the spinning polarotor. UV-enhanced Al BB mirrors are used elsewhere in the chain for redirection. NDs are used to control light levels on the elastic channels. Each channel has an IF [and a second IF (IF2) for additional blocking on two channels], and L to focus light onto each of the PMTs. Each channel is labeled; see Table 2 for wavelengths.

Four redirecting mirrors, two of which are on a translation stage such that they form an optical delay stage, are used to align light from the telescope to the optical path of the polychromator. Tilt adjust allows angular alignment, while translation allows for the longitudinal positioning of the image plane relative to the field stop. Currently, the field stop is located a focal length from the telescope, optimizing the receiver for large object distances. However, for atmospheric scatterers close to the ground, the image plane moves behind the field stop (Campbell et al. 2002) thus the optical delay stage allows for commensurate changes in distance from the primary to the field stop. This may be used to compensate for the reduction of field of view as scatterer altitudes approach the ground. The stage has a travel of 0.3 m, so the image plane can be moved by 0.6 m in total, which translates to reducing the object distance from infinity to approximately 150 m. This is labeled as the focus stage in Fig. 2, and characterization of this facility and its effect on measurements is still ongoing.

The focus stage directs the light from the telescope into the polychromator and through the field stop, labeled FS. A motorized iris is used as the field stop, which allows the telescope field of view to be varied between 0.3 and 2 mrad. Reducing the field of view over this range lowers the background count rate by a factor of 20 which, combined with narrowband filters, allows for daytime measurements. Routine nighttime measurements are taken with a field of view of 1.5 mrad, while during the polar day 0.3 mrad is used. The light is then collimated and passed through a motorized aperture stop (AS), which controls the effective telescope aperture and the amount of light, both background and lidar return, on all channels. The light is then split into the visible and ultraviolet beam paths with a LWP dichroic.

The polychromator layout adopted is largely based on that developed by Behrendt and Reichardt (2000) and uses a cascade of interference filters (IFs) that transmit each wavelength in turn while reflecting the remainder to subsequent optics in the chain. Interference filters within the cascade were designed for an angle of incidence of 10°, whereas those at the end of the chains are at normal incidence. Broadband (BB) UV-enhanced Al fold mirrors have been added to the cascade to ease alignment and locate all the photomultiplier tubes (PMTs) along the top of the polychromator. Lenses (Ls) focus the light onto the active area of each Hamamatsu R7400-03 PMT (R7400-20 for the visible nitrogen channel). Quantum efficiency of the R7400-03 in the ultraviolet is approximately 20% and ~6.5% at around 532 nm, while the R7400-20 is ~16% at 607 nm. The visible nitrogen and depolarization channels are not part of the filter cascade, the light being separated by a 550-nm long-wave pass dichroic. Leakage of light at 532 nm through this dichroic—nominally 0.3%—is directed through a spinning Glan–Thompson polarizer [Licel GmbH (Berlin, Germany) polarotor], which acts as the system’s master trigger and provides linear depolarization measurements.

The spinning polarization analyzer allows for the measurement of depolarization with a single PMT and narrowband interference filter. This offers some advantages to the more common scheme where a polarizing beam splitter directs the parallel and perpendicular light to individual, though ideally identical, filters and PMTs. Once the polarotor has been aligned, any aging of these components will affect both perpendicular and parallel channels equally, meaning a more consistent calibration. Characterization of the polarization preference of the receiver optics was done by measuring the signal from a depolarized white-light source. Then, by comparing the clear-sky lidar return to the known molecular depolarization (Bodhaine et al. 1999); the offset from the transmitter plane of polarization was determined.

Neutral density filters are used to equalize the Raman channel count rates as much as possible, while five-position neutral density filter wheels (NDs in Fig. 2) control the elastic channel count rates. As with all instruments in this system, the polychromator is completely computer controlled. Specifications for the lidar are given in Table 1.

5. Data products and filter specification

To determine appropriate filter center wavelengths for the Raman channel interference filters, the rotational nitrogen and oxygen and rotational–vibrational nitrogen and water vapor Raman backscatter spectra were calculated for laser wavelengths of 354.72 and 532.08 nm (Wandinger 2005; Behrendt and Reichardt 2000; Avila et al. 2004). Nitrogen filter center wavelengths were set to maximize signal levels at 386.67 and 607.46 nm. Filter bandwidths were based on existing daytime lidar systems (Behrendt et al. 2004; Goldsmith et al. 1998; Whiteman et al. 2006): 0.35 nm for both elastic channels and 0.4 nm for both nitrogen channels. These, along with other specifications of the interference filters, are given in Table 2.

Table 2.

Receiver optical specifications.

Receiver optical specifications.
Receiver optical specifications.

In addition to the center wavelength and bandwidth, the passband transmission and stopband rejection and reflectivity are very important. The peak transmission attainable in modern narrowband filters is impressive with the measured transmission of the ultraviolet filters ranging from 54% to 75% and for the visible filters, 75%–83%. Stopband reflectivities of the cascaded filters were greater than 80%. The optical density of the filters at the laser wavelengths was specified to be >6 for the rotational Raman filters that are close to the 532-nm laser line and 8 for those filters farther removed from the laser lines. For the rotational Raman filter centered at 531.16 nm, it was necessary to use two narrowband filters in series to adequately reject the elastic return when there is significant backscatter from optically thick clouds (Behrendt and Reichardt 2000). A second filter (bandwidth of 25 nm, peak transmission of 50%, and out-of-band blocking ≥3) was also used on the visible N2 channel. It was found that the narrowband interference filter suffered from skylight leakage that led to an unacceptably high background signal.

a. Aerosols and clouds

Measurements of elastic and nitrogen Raman profiles allow for the determination of both aerosol extinction and backscatter coefficients, assuming that the molecular density is known (Ansmann and Müller 2005). The molecular density can be obtained from radiosonde temperature and pressure profiles or a standard atmosphere. The calculation of aerosol extinction depends on knowledge of the Ångström exponent, which describes the change in extinction coefficient of a scatterer with wavelength. Values typically range from 0 to 2 (Ferrare et al. 1998) and Bokoye et al. (2002) report a year-round average Ångström exponent at Barrow, Alaska (71°18′N, 156°39′W), for 1997–2000 of 1.341 ±0.557 from sun photometer measurements (440–870 nm). It is possible to negate the effect of the differential transmission by using the sum of the rotational Raman channels instead of the nitrogen channel (Behrendt et al. 2002). By calculating the extinction using both the nitrogen and sum of the rotational Raman wavelengths, it will be possible to calculate the Ångström exponent in the visible. The lidar is collocated with sun and star photometers, which measure aerosol extinction at additional wavelengths, thus allowing for the comparison of aerosol properties. Pahlow et al. (2006) detail a retrieval method for particle effective radius, surface area, and volume concentrations using combined lidar and sun photometer data. The color ratio, which here is the ratio of backscatter coefficients at 532 and 355 nm, gives a measure of effective particle radius (Tao et al. 2008).

The use of a polarized laser in conjunction with polarization sensitive detection, with the polarotor, for example, allows the separate measurement of return signals polarized parallel and perpendicular to the plane of polarization of the transmitted light. The ratio of perpendicular to parallel polarized backscatter coefficients, the depolarization ratio, can be used to identify the phase state, and sometimes the ice crystal type, of cloud constituents (Sassen 2005). Combining the color ratio with the depolarization ratio can be used for a statistical characterization of cloud, ice crystal, and aerosol occurrences (Grenier et al. 2009).

The CRL is nominally zenith pointing and so complications in the retrieval due to specular reflections from horizontally aligned ice crystals are to be expected (Noel and Sassen 2005). Sassen and Zhu (2009) find an 11% increase in the CALIPSO-derived depolarization ratio during the Arctic winter night when moving from nominally on nadir to off nadir. The collocation of the CRL with the off-zenith high-spectral-resolution lidar may present the opportunity for further comparative studies of ice crystal habits.

b. Rotational Raman temperature

Profiles of tropospheric temperatures in the presence of aerosols can be obtained by measuring the backscatter from the pure rotational Raman lines of nitrogen and oxygen (the lines of the two species overlap). The population distribution of these lines has a Boltzmann distribution and so changes shape with temperature. Probing appropriate low rotational quantum number (low-J channel) and high rotational quantum number (high-J channel) regions of the spectrum results in a measurement ratio that is proportional to temperature (Cooney 1972). The proportionality factor can be found by calibrating the ratio of high-J to low-J channels with radiosonde profiles as described by Behrendt (2005).

To obtain a ratio of rotational Raman signals that varies with atmospheric temperature, the filter specifications must be chosen carefully. The center wavelength and width of the interference filters used determines the temperature dependence of each signal and so contributes to the uncertainty of the retrieved temperature. The temperature uncertainty was calculated using the method described by Behrendt et al. (2004) for the anti-Stokes branch over a temperature range of 210–260 K. These temperatures were chosen as most important based on a climatology of local radiosondes. We modeled the filter passband characteristics with the convolution of a top hat (or boxcar) and a raised cosine function. With Hanning window lengths between 0.3 and 0.5 nm, this convolution model respectably fits the steep sides that are possible with modern interference filter designs over the range of bandwidths and center wavelengths required for the uncertainty calculations. The characteristics of the rotation Raman interference filters are given in Table 2.

c. Water vapor mixing ratio

A measure of the atmospheric water vapor mixing ratio can be obtained from the ratio of the Raman backscatter of water vapor and of nitrogen—nitrogen being a proxy for dry air. As has been shown previously (Whiteman 2003a; Adam 2009), care must be taken to account for the temperature dependence of the water vapor measurements, and this is done with the addition of a temperature dependence factor F(T) to the usual Raman lidar equation for both the nitrogen and water vapor channels (Whiteman 2003a). The temperature dependence factor is the ratio of the signal through the filter and the sum of all Raman lines multiplied by the transmission of the filter at the wavelength of interest. The water vapor mixing ratio is described by Adam (2009) as

 
formula

where subscripts N and H indicate nitrogen and water vapor terms, respectively. The proportionality constant, k ≈ 0.485, relates the number density of water and nitrogen to the mass density of water and dry air (Whiteman 2003b). Channel efficiencies have been divided into ξ(λx), the filter transmission, which changes rapidly over the wavelengths of interest; and κ(λx), which is constant over the spectral range of the filter. The κ(λN, λH) is a system efficiency constant, κ(λN)/κ(λH) while Ox(r) are the overlap functions, P(λx, r) are the detected powers, and Δτ(λN, λH, r) is the differential atmospheric transmission for the two wavelengths. The spectral range over which the total Raman backscatter cross sections, x(π)/dΩ, is calculated must be the same as that used for calculating Fx(T).

Equation (1) may be written more succinctly with the use of a single system factor,

 
formula

becoming

 
formula

The system factor, k*, may be found by carefully calculating the spectral response of the instrument (Vaughan et al. 1988), either calibrating the integrated profile with a column water vapor measurement (Turner and Goldsmith 1999) or calibrating with an in situ measurement such as that from a radiosonde (Ferrare et al. 1995). The system factor here includes the differential overlap of the system that can be corrected by comparing with radiosonde profiles in the partial-overlap region (Adam et al. 2010). We intend to primarily calibrate our measurements with water vapor profiles from the twice-daily radiosondes launched on-site.

The water vapor retrieval is modified by temperature dependence factors FN(T) and FH(T), for the nitrogen and water vapor channels, respectively. For a very wide filter, Fx(T) approaches one; however, for the narrow filter bandwidths required for daytime measurements, the filter center wavelength must be carefully selected to minimize the sensitivity of the water vapor mixing ratio, w, to temperature fluctuations. A calculation of filter specifications is presented here that is based, in a similar manner to the rotational Raman temperature filter optimization procedure, on minimizing the uncertainty in the water vapor mixing ratio.

Uncertainty in the retrieval is a summation of the relative uncertainties in the system, the differential transmission, the ratio of return signals due to photon statistics, and the ratio of temperature dependence factors. Unlike the discussion of uncertainty in Whiteman (2003b), here we are interested in minimizing the temperature sensitivity only, so all uncertainties are neglected except for that in FN(T)/FH(T). The relative variance of w is given by the relative variances of the individual temperature dependence factors and the covariance; thus,

 
formula

Assuming a linear change in Fx for small changes in temperature, and so ignoring higher-order terms of the Taylor expansion, the variance of Fx can be given in terms of the variance of the temperature as

 
formula

Thus, the relative uncertainty in the water vapor mixing ratio can then be written as

 
formula

It can be seen that if both FN(T) and FH(T) are constant at all temperatures of interest, then, as expected, the uncertainty induced by temperature fluctuations will be zero. Additionally, if the ratio FN/FH is independent of temperature, then σw will likewise be zero.

The temperature dependence factors were calculated for a range of filter characteristics using the modeled filter transmission curves and the calculated Raman spectra for water vapor and nitrogen for an excitation wavelength of 354.72 nm. Calculations were performed for temperatures 200–300 K; however, based on radiosonde temperature profiles for 2005–09, temperatures 210–260 K were deemed most important. The nitrogen filter center wavelength was set to that which gives a maximum in the integrated backscatter cross section. For a filter bandwidth of 0.4 nm, this is 386.67 nm for all relevant temperatures and is equivalent to experimentally angle tuning the interference filter to maximize the signal. A small change in the nitrogen signal of 0.4% across these temperatures was included in Eq. (6). Calculations were done using finite differences with the temperature difference, ΔT, being ±2.5 K around the temperature of interest. The relative water vapor uncertainty per kelvin, σw/(wΔT), is shown in Fig. 3 for 240 K as a function of the water vapor filter center wavelength and bandwidth.

Fig. 3.

Relative uncertainty of the water vapor mixing ratio per kelvin (Δw) at 240 ±2.5 K as a function of water vapor filter center wavelength and full-width half-maximum bandwidth. The actual interference filter is marked with the open circle.

Fig. 3.

Relative uncertainty of the water vapor mixing ratio per kelvin (Δw) at 240 ±2.5 K as a function of water vapor filter center wavelength and full-width half-maximum bandwidth. The actual interference filter is marked with the open circle.

For larger bandwidths—that is, filters spanning a significant proportion of the water vapor lines—there exists a wide range of center wavelengths with negligible temperature sensitivity. However, when using narrowband filters, the choice of center wavelength becomes critical. Based on previous daytime water vapor systems (Whiteman et al. 2007; Adam et al. 2010), a 0.25-nm bandwidth was chosen as a compromise between background suppression and signal strength. A center wavelength of 407.55 nm was specified to facilitate angle tuning to blue shift the filter transmission passband appropriately.

Adam (2009) presents an analysis of the effect of filter center wavelength—given in terms of the angle of incidence—and bandwidth on how close the ratio FN/FH is to unity, denoted by RMSE, and its relative error, RE. These two quantities are shown in Figs. 16a and 16b in Adam (2009) and, although related to the temperature dependence factors rather than Δw, are similar in form to Fig. 3 in this work as our analysis includes the effect of both Adam’s RMS and RE quantities. One subtlety to note in the comparison is the nonlinear nature of filter center wavelength with angle of incidence.

Figure 4 shows the relative uncertainty as a function of filter center wavelength for a single bandwidth of 0.25 nm (and so is a horizontal slice through Fig. 3) and temperatures ranging from 210 to 260 K. The temperature uncertainty was set at ±2.5 K but changing Raman line strengths means that the Δw curve is different for each average temperature. The water vapor backscatter cross section was calculated for these temperatures—a representative spectrum at 240 K is shown in the figure—and convolved with the filter transmission curve over wavelengths covered by the Raman spectrum. Because of the width of the filter, the backscatter cross section was calculated over both ν1 and ν3 water vapor bands and Q, S, and O branches, not just the Q branch of the ν1 band as is more usual. The figure shows the average of these signal curves as a dashed line and has a maximum signal for a filter center wavelength of 407.516 nm.

Fig. 4.

Calculated Raman spectrum of water vapor at 240 K. The dashed line is the convolution of the water vapor filter and spectrum averaged over 210–260 K and divided by 10. The graded lines are the water vapor mixing ratio uncertainties for 210–260 K in steps of 10 K. The nitrogen filter is fixed at 386.67 nm, filter bandwidths are 0.25 and 0.4 nm for H2O and N2, respectively.

Fig. 4.

Calculated Raman spectrum of water vapor at 240 K. The dashed line is the convolution of the water vapor filter and spectrum averaged over 210–260 K and divided by 10. The graded lines are the water vapor mixing ratio uncertainties for 210–260 K in steps of 10 K. The nitrogen filter is fixed at 386.67 nm, filter bandwidths are 0.25 and 0.4 nm for H2O and N2, respectively.

With Figs. 3 and 4, the center wavelength of the water vapor filter can be set such that the temperature-induced uncertainty in the water vapor mixing ratio is zero at a single temperature or minimized over a range of temperatures. The CRL does not have the facility to accurately measure the transmitted spectrum through the water vapor filter or the angle of incidence. This imposes the practical issue of determining and setting the center wavelength for minimum uncertainty.

A solution is to use what can be easily and accurately measured—signal count rate averaged over a range of altitudes/temperatures—to determine the angle of incidence that results in maximum signal. However, the filter center wavelength that maximizes signal strength is not equal to that which minimizes uncertainty. At all temperatures, the filter wavelength that gives minimum uncertainty is shorter than that which gives peak signal. However, this difference is quite small, ranging from 0.036 to 0.055 nm for temperatures 210–260 K.

If the signal is maximized, then there is a small residual uncertainty—shown in Fig. 5 as the dashed line—that ranges from 0.07 to 0.09% K−1 for temperatures 210–260 K. The ratio of nitrogen to water vapor temperature dependence factors required to correct for this uncertainty was calculated for our filters and is shown as the solid line. For the spectral range used here, the ratio ranges from 1.23 to 1.28, although in Fig. 5 the curve has been normalized in the same way as Fig. 3 in Sakai et al. (2007). Our ratio values are different from those presented previously (Whiteman 2003a; Adam 2009) because of the different spectral ranges used in the calculation.

Fig. 5.

Normalized residual ratio of temperature dependence factors associated with maximizing the water vapor signal is shown as the solid line. Plotted with the dashed line is the relative uncertainty in the water vapor mixing ratio when the signal is maximized but the temperature dependence is not accounted for.

Fig. 5.

Normalized residual ratio of temperature dependence factors associated with maximizing the water vapor signal is shown as the solid line. Plotted with the dashed line is the relative uncertainty in the water vapor mixing ratio when the signal is maximized but the temperature dependence is not accounted for.

After an iterative analysis of the compromises to ideal specifications due to manufacturing tolerances, the interference filter specifications for the system were chosen and are given in Table 2. Upon installation in the polychromator, the filters were roughly angle tuned by monitoring the throughput of each filter with a fiber-coupled spectrometer. Fine-tuning of the angle of incidence of each filter was done using the lidar return.

6. Data acquisition and software

Data are acquired with Licel GmbH transient recorders with combined analog and photon counting modules for the elastic channels and photon counting modules for the Raman channels. The combined signals from the analog–photon counting channels are “glued” together to increase the dynamic range of the elastic signals (Newsom et al. 2009). Although analog recorders on the Raman channels may be advantageous for daytime operation, the variable field and aperture stops are used to control the photon counting rates during daylight hours instead. Bin width is 50 ns, resulting in an ultimate vertical resolution of 7.5 m. The modules have onboard averaging, with data typically being collected with a temporal resolution of 60 s. The data are then downloaded onto a computer via the laboratory Ethernet network.

The lidar and laboratory were designed to be remotely operable over the satellite communications link. Custom software was written in the Python programming language to control the facility from any remote location. An overview of the program structure is shown in Fig. 6.

Fig. 6.

Schematic of the lidar control software structure showing the server side at Eureka and client side at Dalhousie University communicating over the satellite link.

Fig. 6.

Schematic of the lidar control software structure showing the server side at Eureka and client side at Dalhousie University communicating over the satellite link.

The server side handles all instrument interaction and communicates with the client side using the Extensible Markup Language remote procedure call (XML-RPC) protocol. Data are saved as hierarchical data format (HDF version 5) files that also contain measurement metadata. Rudimentary integration may be performed so that near-real-time profiles of reduced data size can be passed to the client side in “quick look” form for quality control. Full-resolution data are archived on-site and transferred off-site to the main CANDAC data servers once a day.

For safety and security, the communications link is actively monitored by the server and any disconnection will result in the system going into a safe state. To monitor the laboratory independently of the computer, a stand-alone webserver [AKCess Pro (AKCP) CameraProbe 8] monitors laboratory environmental conditions and has security cameras connected for visual inspection. With the satellite link shared by all the CANDAC instruments, all software was designed to minimize bandwidth requirements; operating the lidar generally requires less than 30 kb s−1, although added video feeds may increase this up to 100 kb s−1. This is a significant proportion of the 300 kb s−1 total capacity of the link and so it is generally avoided.

7. Example measurement

A set of example nighttime measurements from 1340 UTC 17 December to 0800 UTC 18 December 2010 is shown in Fig. 7. These were taken with a field of view of 1.5 mrad and the aperture stop was set to pass light from the full 1-m telescope diameter. Figure 7a shows the visible particulate backscatter coefficient with a vertical resolution of 105 m (14 bins) and temporal resolution of 10 min, obtained with the Raman technique (Ansmann and Müller 2005). The ultraviolet measurement displays the same features with poorer noise characteristics due to the enhanced molecular return (see Fig. 8). The clear-air normalization region was set between 15 and 20 km, averaged over the entire measurement period to minimize the influence of the optically thick clouds in the latter part of the measurement. The elastic measurements are combined analog and photon counting signals with gluing between 0.5 and 5 MHz. Dead-time correction has not yet been performed, thus the 5-MHz limit to ensure linear counting.

Fig. 7.

Measurement from 1340 UTC 17 Dec to 0800 UTC 18 Dec 2010: (a) visible particulate backscatter coefficient, (b) linear depolarization, and (c) water vapor mixing ratio.

Fig. 7.

Measurement from 1340 UTC 17 Dec to 0800 UTC 18 Dec 2010: (a) visible particulate backscatter coefficient, (b) linear depolarization, and (c) water vapor mixing ratio.

Fig. 8.

Profiles of backscatter coefficient in the (a) visible and (b) ultraviolet integrated over the period from 2000 UTC 17 Dec to 0000 UTC 18 Dec 2010. Vertical resolution is 105 m. Statistical uncertainties are shown as dashed lines with scale on top axis.

Fig. 8.

Profiles of backscatter coefficient in the (a) visible and (b) ultraviolet integrated over the period from 2000 UTC 17 Dec to 0000 UTC 18 Dec 2010. Vertical resolution is 105 m. Statistical uncertainties are shown as dashed lines with scale on top axis.

Linear depolarization is shown in Fig. 7b, also for integration over 105 m and 10 min. Because of technical difficulties with the polarotor, this has been calculated using the perpendicular and total elastic (being the sum of perpendicular and parallel polarizations) returns. This is inferior to using the rotating analyzer discussed in section 4 because of the significant mismatch in signal levels, the differential overlap, and the requirement of an additional term in the calibration function that removes the perpendicular component from the total elastic return. These factors will increase the uncertainty of the depolarization ratio although, because of the temporary nature of the current configuration, a complete error analysis has yet to be performed. The spinning analyzer will be repaired and replaced after the 2010/11 winter. The water vapor mixing ratio for the same period, integrated over 105 m and 15 min, is shown in Fig. 7c. The longer integration time was required because of the poorer signal-to-noise ratio of these channels.

Profiles of particulate backscatter coefficient measurements leading up to 0000 UTC 18 December 2010 are shown for the visible (Fig. 8a) and ultraviolet (Fig. 8b) channels for a vertical integration of 105 m and a temporal integration of 4 h. The statistical uncertainty is also shown; note that the scale is along the top axis. The increase in uncertainty at low altitudes is due to the incomplete overlap that dominates the error below 1 km. Analysis in the boundary layer has been complicated as the instrument has been found to have both an overlap region (that changes with alignment) and a differential overlap (that is very stable). A correction to the differential overlap of the aerosol products is applied, based on obtaining a backscatter ratio of unity during a very clean-air episode. Correction is applied to both the water vapor mixing ratio and temperature profiles by applying an altitude-dependent calibration. We are currently unable to satisfactorily correct for the overlap, which means that extinction for the near-ground aerosols shown in Fig. 7a cannot be confidently calculated.

Figure 9 shows the water vapor mixing ratio (Fig. 9a) and temperature profiles (Fig. 9b), along with the statistical uncertainties. Integrations are 1 h and 105 m for the water vapor profile and 4 h and 105 m for the temperature. Despite the increased noise, a shorter integration time was used for the water vapor mixing ratio as it was changing rapidly over this period. For comparison, the profiles from the radiosonde launched at 2315 UTC are also shown. It should be noted that this radiosonde was not used for direct calibration of these lidar profiles, so an exact fit is not expected. Radiosonde profiles are only used for calibration when there are no clouds, that is, when differential extinction is not an issue, and conditions are stable so that the radiosonde and lidar are sampling the same air mass. With time, and many suitable radiosonde profiles, the calibration will converge to a well-quantified function for a given set of instrument settings. The dearth of calibration sondes is also the reason for the poor differential overlap correction of the temperature profile below 2 km. We have also found that the calibration of the rotational Raman temperature is currently susceptible to changes in laboratory temperature. It is believed that the quality of the laser seeding or temperature control of the seed laser is the cause. Work is continuing to isolate and correct the source of this variability.

Fig. 9.

Profiles of (a) water vapor mixing ratio and (b) tropospheric temperature integrated from 2300 UTC 17 Dec to 0000 UTC 18 Dec and from 2000 UTC 17 Dec to 0000 UTC 18 Dec 2010, respectively. Vertical resolution is 210 m. The 2315 UTC radiosonde profiles are shown as circles and statistical uncertainties as dashed lines with scale on top axis.

Fig. 9.

Profiles of (a) water vapor mixing ratio and (b) tropospheric temperature integrated from 2300 UTC 17 Dec to 0000 UTC 18 Dec and from 2000 UTC 17 Dec to 0000 UTC 18 Dec 2010, respectively. Vertical resolution is 210 m. The 2315 UTC radiosonde profiles are shown as circles and statistical uncertainties as dashed lines with scale on top axis.

The measurement presented in Fig. 7 illustrates the benefits of a multichannel lidar in studying polar atmospheric processes, particularly those involving water in all its forms and its interaction with aerosols. During the measurement there are two different cloud types, a mixed-phase cloud with some hanging precipitation at 0600–0800 UTC and ice clouds covering much of the latter part of the measurement at altitudes 3–9 km. The distinction can be made when comparing the depolarization ratios of the two types of cloud. The increased noise above the optically thick clouds is due to the extinction of the weaker Raman signals by the clouds and also a significant ice crystal event in the lowest 100 m from 0000 to 0330 UTC (not shown in Fig. 7a because of overlap region uncertainty). Across the entire measurement there is a pervasive aerosol layer between the ground and 1 km. Taking the ratio of the 532- and 355-nm backscatter coefficients gives the color ratio, although with only two wavelengths it is not possible to obtain unambiguous size information; the particles can be separated into coarse and fine modes. The cloud has a color ratio of 0.92 ±0.14, while that of the aerosol layer is 0.54 ±0.11. The lidar ratio of the ice cloud was found to be 20 ±9 sr in the ultraviolet and 25 ±12 sr in the visible. These are very close to the 21 ±6 sr at 532 nm reported by Lampert et al. (2009) for a subvisible Arctic ice cloud. A lower value of 13.8 sr was found by Gayet et al. (2009) for Arctic ice crystals, while measurements of midlatitude cirrus cloud indicate that the lidar ratio can vary significantly (Ansmann et al. 1992; Chen et al. 2002).

The water vapor mixing ratio shows an influx of moist air, peaking at 0700 UTC and 3 km at approximately 1.2 g kg−1. This maximum corresponds to the thick mixed-phase cloud seen in Fig. 7a. Overlaying the aerosol and water vapor plots (not shown) reveals that the descending streamers of aerosol in the first half of the measurement correspond to wetter air masses. The stable wintertime lower troposphere can give rise to such highly stratified aerosol layers (Radke et al. 1989). The subsidence of the moist air is blocked by the temperature inversion that extends from the surface to approximately 2 km.

8. Summary

This paper describes a new lidar that was recently installed in the Canadian High Arctic that considerably extends the capability of atmospheric profiling measurements in the region. Aerosols are measured in both the ultraviolet and visible, enabling the color ratio to be calculated while cloud phase measurements are possible with a depolarization channel. Tropospheric temperature and water vapor mixing ratio profiling can now be carried out at the resolutions possible using the lidar technique, enabling better understanding of the dynamic processes of these fields. The remoteness of the facility has meant a significant investment in computer-controlled instrumentation for operation over a satellite link. The capability to operate remotely should enable significantly extended hours of data collection compared to a system requiring on-site supervision.

The lidar is currently operational and an example measurement has been presented. Both the lidar and the analysis packages are under development and although there are still significant improvements that are being made to both, the potential of the system for enhancing our understanding of Arctic atmospheric processes—particularly those involving water vapor—is evident. Initial analyses of data taken to date can be viewed online (http://aolab.phys.dal.ca/data/).

Acknowledgments

Canadian Network for the Detection of Atmospheric Change (CANDAC) staff members A. Harrett, A. Khmel, P. Loewen, K. MacQuarrie, O. Mikhailov, and M. Okraszewski have been instrumental in the installation of this system along with staff of the Environment Canada Eureka weather station. GJN thanks M. Adam, D. D. Turner, G. Vaughan, and D. N. Whiteman for their helpful discussions and the anonymous reviewers for their improvements to the manuscript. The Polar Environment Atmospheric Research Laboratory (PEARL) is operated by CANDAC. CANDAC/PEARL funding partners are the Arctic Research Infrastructure Fund, Atlantic Innovation Fund/Nova Scotia Research Innovation Trust, Canadian Foundation for Climate and Atmospheric Science, Canadian Foundation for Innovation, Canadian Space Agency, Environment Canada, Government of Canada International Polar Year, Natural Sciences and Engineering Research Council, Ontario Innovation Trust, Ontario Research Fund, Indian and Northern Affairs Canada, and the Polar Continental Shelf Program.

REFERENCES

REFERENCES
Adam
,
M.
,
2009
:
Notes on temperature-dependent lidar equations
.
J. Atmos. Oceanic Technol.
,
26
,
1021
1039
.
Adam
,
M.
, and
Coauthors
,
2010
:
Water vapor measurements by Howard University Raman lidar during the WAVES 2006 campaign
.
J. Atmos. Oceanic Technol.
,
27
,
42
60
.
Ansmann
,
A.
, and
D.
Müller
,
2005
:
Lidar and atmospheric aerosol particles
.
Lidar: Range-Resolved Optical Remote Sensing of the Atmosphere, C. Weitkamp, Ed., Springer Series in Optical Sciences, Vol. 102, Springer, 105–141
.
Ansmann
,
A.
,
U.
Wandinger
,
M.
Riebesell
,
C.
Weitkamp
, and
W.
Michaelis
,
1992
:
Independent measurement of extinction and backscatter profiles in cirrus clouds by using a combined Raman elastic-backscatter lidar
.
Appl. Opt.
,
31
,
7113
7131
.
Avila
,
G.
,
J. M.
Fernández
,
G.
Tejeda
, and
S.
Montero
,
2004
:
The Raman spectra and cross-sections of H2O, D2O, and HDO in the OH/OD stretching regions
.
J. Mol. Spectrosc.
,
228
,
38
65
.
Behrendt
,
A.
,
2005
:
Temperature measurements with lidar
.
Lidar: Range-Resolved Optical Remote Sensing of the Atmosphere, C. Weitkamp, Ed., Springer Series in Optical Sciences, Vol. 102, Springer, 273–305
.
Behrendt
,
A.
, and
J.
Reichardt
,
2000
:
Atmospheric temperature profiling in the presence of clouds with a pure rotational Raman lidar by use of an interference-filter-based polychromator
.
Appl. Opt.
,
39
,
1372
1378
.
Behrendt
,
A.
,
T.
Nakamura
,
M.
Onishi
,
R.
Baumgart
, and
T.
Tsuda
,
2002
:
Combined Raman lidar for the measurement of atmospheric temperature, water vapor, particle extinction coefficient, and particle backscatter coefficient
.
Appl. Opt.
,
41
,
7657
7666
.
Behrendt
,
A.
,
T.
Nakamura
, and
T.
Tsuda
,
2004
:
Combined temperature lidar for measurements in the troposphere, stratosphere, and mesosphere
.
Appl. Opt.
,
43
,
2930
2939
.
Bodhaine
,
B. A.
,
N. B.
Wood
,
E. G.
Dutton
, and
J. R.
Slusser
,
1999
:
On Rayleigh optical depth calculations
.
J. Atmos. Oceanic Technol.
,
16
,
1854
1861
.
Bokoye
,
A. I.
,
A.
Royer
,
N. T.
O’Neill
, and
L. J. M.
McArthur
,
2002
:
A North American Arctic aerosol climatology using ground-based sunphotometry
.
Arctic
,
55
,
215
228
.
Bony
,
S.
, and
Coauthors
,
2006
:
How well do we understand and evaluate climate change feedback processes?
J. Climate
,
19
,
3445
3482
.
Bourdages
,
L.
,
T. J.
Duck
,
G.
Lesins
, and
E. W.
Eloranta
,
2009
:
Physical properties of High Arctic tropospheric particles during winter
.
Atmos. Chem. Phys.
,
9
,
6881
6897
.
Campbell
,
J. R.
,
D. L.
Hlavka
,
E. J.
Welton
,
C. J.
Flynn
,
D. D.
Turner
,
J. D.
Spinhirne
,
V.
Stanley Scott
III
, and
I. H.
Wang
,
2002
:
Full-time, eye-safe cloud and aerosol lidar observation at atmospheric radiation measurement program sites: Instruments and data processing
.
J. Atmos. Oceanic Technol.
,
19
,
431
442
.
Chen
,
W.-N.
,
C.-W.
Chiang
, and
J.-B.
Nee
,
2002
:
Lidar ratio and depolarization ratio for cirrus clouds
.
Appl. Opt.
,
41
,
6470
6476
.
Cooney
,
J.
,
1972
:
Measurement of atmospheric temperature profiles by Raman backscatter
.
J. Appl. Meteor.
,
11
,
108
112
.
Curry
,
J. A.
,
W. B.
Rossow
,
D.
Randall
, and
J. L.
Schramm
,
1996
:
Overview of Arctic cloud and radiation characteristics
.
J. Climate
,
9
,
1731
1764
.
Donovan
,
D. P.
,
J. C.
Bird
,
J. A.
Whiteway
,
T. J.
Duck
,
S. R.
Pal
, and
A. I.
Carswell
,
1995
:
Lidar observations of stratospheric ozone and aerosol above the Canadian High Arctic during the 1994–95 winter
.
Geophys. Res. Lett.
,
22
,
3489
3492
.
Duck
,
T. J.
,
J. A.
Whiteway
, and
A. I.
Carswell
,
2000
:
A detailed record of High Arctic middle atmospheric temperatures
.
J. Geophys. Res.
,
105
(
D18
),
22 909
22 918
.
Duck
,
T. J.
,
J. A.
Whiteway
, and
A. I.
Carswell
,
2001
:
The gravity wave–Arctic stratospheric votex interaction
.
J. Atmos. Sci.
,
58
,
3581
3596
.
Eloranta
,
E. W.
,
I. A.
Razenkov
, and
J. P.
Garcia
,
2006
:
Arctic observations with the University of Wisconsin high spectral resolution lidar
.
Reviewed and Revised Papers Presented at the 23rd International Laser Radar Conference, C. Nagasawa and N. Sugimoto, Eds., 399–402
.
Ferrare
,
R. A.
,
S. H.
Melfi
,
D. N.
Whiteman
,
K. D.
Evans
,
F. J.
Schmidlin
, and
D. O.
Starr
,
1995
:
A comparison of water vapor measurements made by Raman lidar and radiosondes
.
J. Atmos. Oceanic Technol.
,
12
,
1177
1195
.
Ferrare
,
R. A.
,
S. H.
Melfi
,
D. N.
Whiteman
,
K. D.
Evans
, and
R.
Leifer
,
1998
:
Raman lidar measurements of aerosol extinction and backscattering 1. Methods and comparisons
.
J. Geophys. Res.
,
103
(
D16
),
19 663
19 672
.
Gayet
,
J.-F.
,
G.
Mioche
,
A.
Dörnbrack
,
A.
Ehrlich
,
A.
Lampert
, and
M.
Wendisch
,
2009
:
Microphysical and optical properties of Arctic mixed-phase clouds. The 9 April 2007 case study
.
Atmos. Chem. Phys.
,
9
,
6581
6595
.
Gerding
,
M.
,
C.
Ritter
,
M.
Müller
, and
R.
Neuber
,
2004
:
Tropospheric water vapour soundings by lidar at high Arctic latitudes
.
Atmos. Res.
,
71
,
289
302
.
Goldsmith
,
J. E. M.
,
F. H.
Blair
,
S. E.
Bisson
, and
D. D.
Turner
,
1998
:
Turn-key Raman lidar for profiling atmospheric water vapor, clouds, and aerosols
.
Appl. Opt.
,
37
,
4979
4990
.
Graversen
,
R. G.
, and
M.
Wang
,
2009
:
Polar amplification in a coupled climate model with locked albedo
.
Climate Dyn.
,
33
,
629
643
.
Grenier
,
P.
,
J.-P.
Blanchet
, and
R.
Muñoz-Alpizar
,
2009
:
Study of polar thin ice clouds and aerosols seen by CloudSat and CALIPSO during midwinter 2007
.
J. Geophys. Res.
,
114
,
D09201
,
doi:10.1029/2008JD010927
.
Holben
,
B. N.
, and
Coauthors
,
1998
:
AERONET—A federated instrument network and data archive for aerosol characterization
.
Remote Sens. Environ.
,
66
,
1
16
.
Ishii
,
S.
,
T.
Shibata
,
T.
Nagai
,
K.
Mizutani
,
T.
Itabe
,
M.
Hirota
,
T.
Fujimoto
, and
O.
Uchino
,
1999
:
Arctic haze and clouds observed by lidar during four winter seasons of 1993–1997, at Eureka, Canada
.
Atmos. Environ.
,
33
,
2459
2470
.
Kahl
,
J. D.
,
M. C.
Serreze
, and
R. C.
Schnell
,
1992
:
Tropospheric low-level temperature inversions in the Canadian Arctic
.
Atmos.–Ocean
,
30
,
511
529
.
Knuteson
,
R. O.
, and
Coauthors
,
2004
:
Atmospheric emitted radiance interferometer. Part I: Instrument design
.
J. Atmos. Oceanic Technol.
,
21
,
1763
1776
.
Lampert
,
A.
, and
Coauthors
,
2009
:
Microphysical and radiative characterization of a subvisible midlevel Arctic ice cloud by airborne observations – A case study
.
Atmos. Chem. Phys.
,
9
,
2647
2661
.
Lesins
,
G.
,
T. J.
Duck
, and
J. R.
Drummond
,
2010
:
Climate trends at Eureka in the Canadian High Arctic
.
Atmos.–Ocean
,
48
,
59
80
.
Moran
,
K. P.
,
B. E.
Martner
,
M. J.
Post
,
R. A.
Kropfli
,
D. C.
Welsh
, and
K. B.
Widener
,
1998
:
An unattended cloud-profiling radar for use in climate research
.
Bull. Amer. Meteor. Soc.
,
79
,
443
455
.
Moss
,
A. L.
,
2010
:
Initial water vapour results from the Eureka, Nunavut lidar
.
M.S. thesis, Dept. of Physics and Astronomy, University of Western Ontario, 82 pp
.
Neely
,
R. R.
, III
, and
J. P.
Thayer
,
2011
:
Raman lidar profiling of tropospheric water vapor over Kangerlussuaq, Greenland
.
J. Atmos. Oceanic Technol.
,
28
,
1141
1148
.
Newsom
,
R. K.
,
D. D.
Turner
,
B.
Mielke
,
M.
Clayton
,
R.
Ferrare
, and
C.
Sivaraman
,
2009
:
Simultaneous analog and photon counting detection for Raman lidar
.
Appl. Opt.
,
48
,
3903
3914
.
Noel
,
V.
, and
K.
Sassen
,
2005
:
Study of planar ice crystal orientations in ice clouds from scanning polarization lidar observations
.
J. Appl. Meteor.
,
44
,
653
664
.
Pahlow
,
M.
,
D.
Müller
,
M.
Tesche
,
H.
Eichler
,
G.
Feingold
,
W. L.
Ederhard
, and
Y.-F.
Cheng
,
2006
:
Retrieval of aerosol properties from combined multiwavelength lidar and sunphotometer measurements
.
Appl. Opt.
,
45
,
7429
7442
.
Radke
,
L. F.
,
C. A.
Brock
,
J. H.
Lyons
,
P. V.
Hobbs
, and
R. C.
Schnell
,
1989
:
Aerosol and lidar measurements of hazes in mid-latitude and polar airmasses
.
Atmos. Environ.
,
23
,
2417
2430
.
Sakai
,
T.
,
T.
Nagai
,
M.
Nakazato
,
T.
Matsumura
,
N.
Orikasa
, and
Y.
Shoji
,
2007
:
Comparisons of Raman lidar measurements of tropospheric water vapor profiles with radiosondes, hygrometers on the meteorological observation tower, and GPS at Tsukuba, Japan
.
J. Atmos. Oceanic Technol.
,
24
,
1407
1423
.
Sassen
,
K.
,
2005
:
Polarization in lidar
.
Lidar: Range-Resolved Optical Remote Sensing of the Atmosphere, C. Weitkamp, Ed., Springer Series in Optical Sciences, Vol. 102, Springer, 19–42
.
Sassen
,
K.
, and
J.
Zhu
,
2009
:
A global survey of CALIPSO linear depolarization ratios in ice clouds: Initial findings
.
J. Geophys. Res.
,
114
,
D00H07
,
doi:10.1029/2009JD012279
.
Serreze
,
M. C.
, and
J. A.
Francis
,
2006
:
The Arctic amplification debate
.
Climatic Change
,
76
,
241
264
.
Shaw
,
G. E.
,
1995
:
The Arctic haze phenomenon
.
Bull. Amer. Meteor. Soc.
,
76
,
2403
2413
.
Shupe
,
M. D.
, and
J. M.
Intrieri
,
2004
:
Cloud radiative forcing of the Arctic surface: The influence of cloud properties, surface albedo, and solar zenith angle
.
J. Climate
,
17
,
616
628
.
Soden
,
B. J.
, and
I. M.
Held
,
2006
:
An assessment of climate feedbacks in coupled ocean–atmosphere models
.
J. Climate
,
19
,
3354
3360
.
Tao
,
Z.
,
M. P.
McCormick
,
D.
Wu
,
Z.
Liu
, and
M. A.
Vaughan
,
2008
:
Measurements of cirrus cloud backscatter color ratio with a two-wavelength lidar
.
Appl. Opt.
,
47
,
1478
1485
.
Tjernström
,
M.
,
C.
Leck
,
P. O. G.
Persson
,
M. L.
Jensen
,
S. P.
Oncley
, and
A.
Targino
,
2004
:
The summertime Arctic atmosphere: Meteorological measurements during the Arctic Ocean Experiment 2001
.
Bull. Amer. Meteor. Soc.
,
85
,
1305
1321
.
Turner
,
D. D.
, and
J. E. M.
Goldsmith
,
1999
:
Twenty-four-hour Raman lidar water vapor measurements during the atmospheric radiation measurement program’s 1996 and 1997 water vapor intensive observation periods
.
J. Atmos. Oceanic Technol.
,
16
,
1062
1076
.
Turner
,
D. D.
,
S. A.
Clough
,
J. C.
Liljegren
,
E. E.
Clothiaux
,
K. E.
Cady-Pereira
, and
K. L.
Gaustad
,
2007
:
Retrieving liquid water path and precipitable water vapor from the Atmospheric Radiation Measurement (ARM) microwave radiometers
.
IEEE Trans. Geosci. Remote Sens.
,
45
,
3680
3690
.
Vaughan
,
G.
,
D. P.
Warfing
,
L.
Thomas
, and
V.
Mitev
,
1988
:
Humidity measurements in the free troposphere using Raman backscatter
.
Quart. J. Roy. Meteor. Soc.
,
114
,
1471
1484
.
Wandinger
,
U.
,
2005
:
Raman lidar
.
Lidar: Range-Resolved Optical Remote Sensing of the Atmosphere, C. Weitkamp, Ed., Springer Series in Optical Sciences, Vol. 102, Springer, 241–271
.
Whiteman
,
D. N.
,
2003a
:
Examination of the traditional Raman lidar technique. I: Evaluating the temperature-dependent lidar equations
.
Appl. Opt.
,
42
,
2571
2592
.
Whiteman
,
D. N.
,
2003b
:
Examination of the traditional Raman lidar technique. II: Evaluating the ratios for water vapor and aerosols
.
Appl. Opt.
,
42
,
2593
2608
.
Whiteman
,
D. N.
, and
Coauthors
,
2006
:
Raman lidar measurements during the International H2O Project. Part I: Instrumentation and analysis technique
.
J. Atmos. Oceanic Technol.
,
23
,
157
169
.
Whiteman
,
D. N.
,
I.
Veselovskii
,
M.
Cadirola
,
K.
Rush
,
J.
Comer
,
J. R.
Potter
, and
R.
Tola
,
2007
:
Demonstration measurements of water vapor, cirrus clouds, and carbon dioxide using a high-performance Raman lidar
.
J. Atmos. Oceanic Technol.
,
24
,
1377
1388
.
Whiteway
,
J. A.
, and
A. I.
Carswell
,
1994
:
Rayleigh lidar observations of thermal structure and gravity wave activity in the High Arctic during a stratospheric warming
.
J. Atmos. Sci.
,
51
,
3122
3136
.
Winker
,
D. M.
,
M. A.
Vaughan
,
A.
Omar
,
Y.
Hu
,
K. A.
Powell
,
Z.
Liu
,
W. H.
Hunt
, and
S. A.
Young
,
2009
:
Overview of the CALIPSO mission and CALIOP data processing algorithms
.
J. Atmos. Oceanic Technol.
,
26
,
2310
2323
.
Winton
,
M.
,
2006
:
Amplified Arctic climate change: What does surface albedo feedback have to do with it?
Geophys. Res. Lett.
,
33
,
L03701
,
doi:10.1029/2005GL025244
.

Footnotes

*

Current affiliation: Facility for Airborne Atmospheric Measurements, Cranfield, United Kingdom.

+

Current affiliation: NATO Undersea Research Centre, La Spezia, Italy.