Open Access
Add to BibSonomy
Abstract
A pure rotational Raman lidar (PRRL) for full-day troposphere temperature measurement was deployed in February 2020 at Zhongshan Station (69.37°S, 76.37°E), Antarctica, by the 36th Chinese National Antarctic Research Expedition. The PRRL emits a 532.23-nm laser light and employs a 203.2-mm telescope to collect atmospheric backscatter. Cubic nonpolarizing beam splitters are introduced to yield a compact optics arrangement. A quasi-single-line-extraction technique is proposed for extracting the molecular Stokes line signals. A lidar container with a window system is customized to house the whole PRRL system for long-term stable operation. An approach using a laser plummet is developed for fast and convenient adjustment of the telescope zenithward. A home-made calibration module is utilized for straightforward visual optics adjustment with ∼35.3-μrad angular positioning accuracy. Both typical daytime and nighttime temperature measurement examples are presented to verify the lidar performance. From a 30-h continuous temperature measurement result, it is found the tropopause is located at ∼10.8 km above ground level with a mean temperature of ∼203 K; significant temperature variability occurs only at the inversion areas, while off which the 1-h temperature profiles are relatively similar in form with an average lapse rate of -8.3 K/km.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
The polar atmosphere is an important part of the global atmospheric system. The atmosphere over the Polar Regions is more sensitive to global climate change than other areas [1], making it the best place to understand the climate change and climate disruption. The polar atmosphere is also the subsidence area of the atmospheric circulation and acts as a modulator of mass/energy entering and leaving the geospace environment. The structure and change of atmospheric temperature in the polar region is related to the movement of global atmospheric system, as well as the basic processes of sea-ice-air interaction under polar environment, imposing a profound impact on global climate change [2–6]. Consequently, knowledge of polar atmospheric temperature is of great significance for studying the polar atmosphere.
Currently, because of the inherent difficulty in exploring the polar region, the polar atmosphere is still poorly understood. As one of the most powerful active remote sensing technologies, lidar has been widely used in the observation and research of the earth's atmosphere in recent decades, yielding outstanding contributions to the development of atmospheric science. Among its various applications, lidar has proven itself an effective and reliable tool for profiling atmospheric temperature, exhibiting the advantages of continuous measurement ability with high spatiotemporal resolution. By utilizing the fluorescence tracers (such as free metal Fe and Na atoms), the atmospheric temperature in the polar mesosphere and lower thermosphere has been monitored by fluorescence resonant lidars [7–12]. Using the Rayleigh integration technique [13], the polar atmospheric temperature from lower stratosphere to mesosphere has been probed by Rayleigh lidars [14–18]. With the help of the vib-rotational Raman spectroscopy technique, temperatures in the troposphere and lower stratosphere have also been retrieved. At Japanese Syowa Station (69.0°S, 39.6°E), a Rayleigh-Raman lidar system was installed in January 2011 and observed the atmospheric temperature in the height range of 10-80 km by simultaneous detection of Rayleigh signals at 355 nm and N2 Raman signals at 387 nm [18].
However, so far there are hardly any reported Antarctic lidars that can obtain temperature profiles in the troposphere. Since there are no metallic atom tracers and aerosols and clouds may be present, both fluorescence resonant lidar and Rayleigh lidar fail to retrieve temperature results in the troposphere. In contrast, a pure rotational Raman lidar (PRRL) can obtain accurate temperatures in the troposphere despite the presence of optically-thin aerosols and clouds. Nowadays, the reported PPRLs in the middle and low latitude regions can measure atmospheric temperatures throughout the troposphere and lower stratosphere [19–32] and some of them achieve to operate in daytime [24–26,28,29,31]. The PRRL generally extracts pure rotational Raman signals from atmospheric N2 and O2 molecules that are separated from elastic signal in spectrum. With the help of the-state-of-the-art optical extraction devices (e.g., interference filters), it can eliminate the interferences due to contaminating elastic signals from aerosols and clouds. Besides, the theoretical basis for the PRRL to retrieve temperature profiles is that under the conditions of local thermodynamic equilibrium, the signal intensity ratio of the two portions of PRR line signals extracted from atmospheric (N2 and O2) molecules has a temperature dependence strictly defined by the Boltzmann distribution [33]. In the troposphere and lower stratosphere, the molecule collision frequency is high enough so that the local thermodynamic equilibrium is generally valid, enabling the PRRL to have solid theoretical foundation for temperature retrieval.
In this paper we report on the newly-built PRRL at Zhongshan Station (69.37°S, 76.37°E and 15 m above sea level), Antarctica during the 36th Chinese National Antarctic Research Expedition (CHINARE) in 2019-2020. This PRRL, to our knowledge, is the first Antarctic lidar system to measure tropospheric temperature during both daytime and nighttime, which can contribute to understanding the thermal structure and evolution of the troposphere over Antarctica. The details of the lidar system instrumental setup, methodology, measurement examples, and summary are presented successively in following sections.
2. Instrumental setup
The PRRL was first developed and tested at the lidar observation station (30.5°N, 114.4°E) located on the campus of Wuhan University based on our formal work on temperature Raman lidars [27,31,32]. After inspection and domestic acceptance, it was then packed and shipped by the Chinese polar scientific research ship “Xuelong” in October 2019 and arrived at Zhongshan Station about one month later. It finished on-site assembly and re-adjustment on February 6, 2020 and started routine operation from then on. The Zhongshan Station (69.37°S, 76.37°E) is situated in the southeast coastal area of Antarctica with annual average temperature of about -10℃, annual mean humidity of 54% and prevailing southeast wind (days of gale above grade 8 can reach 174 days a year). Each year at Zhongshan Station, there are 54 and 58 continuous days of polar day and polar night, respectively. Since the PRRL was intended to perform both nighttime and daytime temperature measurements mainly in the troposphere, special technical issues were considered to compromise between lidar performance and convenience of on-site lidar maintenance. Besides, a customized container for polar observation was manufactured to house the PRRL and guarantee the long-term stable operation of the lidar under local harsh meteorological environments.
2.1. Lidar system description
Figure 1 provides a schematic layout of the PRRL system. The lidar transmitter utilizes an injection-seeded Nd:YAG laser (Powerlite DLS 9030, Continuum, US) to generate nominal radiation of ∼800 mJ per pulse at 532.23 nm with a repetition of 30 Hz, a pulse width of 8 ns and linewidth of <0.006 cm−1. To protect the subsequent transmitter optics and reduce the maintenance frequency of the flash lamps of the laser, the actual laser output is limited and maintained to ∼400 mJ per pulse. A home-made 5× beam expander (BE) is equipped to compress the divergence of the output laser beam to be ∼0.1 mrad, as well as to decrease the output laser energy density. An electronically controlled 2-D reflecting mirror (RM1) is used to guide the incident expanded beam into atmosphere zenithward. The RM1 has a working angle of 45° and a reflectance of >99.5% around 532 nm.
Fig. 1. Schematic layout of the PRRL system. BE, beam expander; RM, reflecting mirror; L, lens; BPF, bandpass filter; BS, beam splitter; IF, interference filter.
The lidar receiver is biaxially configured with the transmitter. For reducing the lidar blind area, a commercial Cassegrain telescope (Meade, US) with a clear aperture of 203.2 mm is employed to collect atmospheric backscatter. The separate distance between the optical axes of the telescope and the outgoing laser beam is ∼200 mm. The telescope has a focal length of 2 m and a subsequent iris to set the telescope field of view (FOV) to be ∼1.0 mrad. After the iris, the backscattered signal is redirected by a reflecting mirror (RM2) to be horizontal before being collimated by a collimating lens (L0). A bandpass filter (BPF) is first positioned to filter the collimated parallel signal lights. The BPF has a full-width at half-maximum (FWHM) of 18 nm centered at 532 nm and an average transmittance of >93% in the passband, while rejecting signals out of band with a suppression ratio of >6 orders. After the BPF, two cubic nonpolarizing beam splitters (BS1 and BS2; Thorlabs, US) are used to separate and divide the signal lights into three parts which are then detected by three independent detection channels. The usage of cubic beam splitters decreases the overall signal transmitting efficiency for each detection channel, but enables a larger FOV (e.g., 1.0 mrad). Since the intensities of backscatter signals at lower heights are relatively strong and lidar blind area is the most concerned issue here, the spectral extracting scheme based on the cheaper cubic nonpolarizing beam splitters was finally selected. In detail, the first cube beam splitter (BS1) has a reflectance–transmittance ratio (RE:TR) of 1:9; the reflected light by the BS1 is guided into the first detection channel, while the transmitted light by the BS1 enters the second cube beam splitter (BS2). The BS2 has a RE:TR of 1:1; the reflected light by the BS2 enters the second detection channel, while the transmitted light enters the third detection channel. Both BS1 and BS2 have a side length of 25.4 mm. Additional advantage of the beam-splitting optics is that it greatly reduces the optical path of all the three detection channels, which yields a compact optical arrangement.
All the three detection channels utilize a modular design and each includes interference filter (IF), a focusing lens (L) and a photomultiplier tube (PMT). The focusing lens (L1, L2 or L3) has a focal length of 40 mm and transmittance of >99% around 532 nm. The PMT (H10721-210, Hamamatsu, JP) in each detection channel has an effective photosensitive surface diameter of 8 mm and a quantum efficiency of ∼20% around 532 nm. However, the three detection channels differ in the optical parameters for the corresponding IFs. The IF1 has a peak transmittance of >73% at the central wavelength (CWL) of 532.23 nm, a FWHM of 0.3 nm, and a rejection ratio of >6 orders of magnitude to signals out of band. The IF2 has a peak transmittance of >73% at the CWL of 533.92 nm, a FWHM of 0.2 nm, and a rejection ratio of >6 orders of magnitude to signals out of band. The IF3 has a peak transmittance of >72% at the CWL of 535.30 nm, a FWHM of 0.2 nm, and a rejection ratio of >6 orders of magnitude to signals out of band. All utilized IFs are customized (Barr, US) and have working angles of 0°. Note for a laser radiation of 532.23 nm, the atmospheric N2 Stokes line signal with rotational quantum number of J=6 (12) has a wavelength of 533.93 (535.29) nm. So that the IF2 (IF3) with a FWHM of 0.2 nm centered at 533.92 (535.30) nm mainly extract the N2 Stokes J=6 (12) line signal, as well as the O2 Stokes J=9 (17) line signal (see Fig. 6). Accordingly, the two Raman detection channels with IF2 and IF3 are designated as the low-J and high-J Raman channels here, respectively. The first detection channel with IF1 records elastic Mie/Rayleigh signals and thus be named as the elastic channel. Neutral density filters (NF) with a total transmittance of ∼0.5‰ are added before the IF1 (not shown in Fig. 1) to avoid saturation of the PMT. Besides, in each Raman detection channel, two identical IFs are used to improve the suppression of the strong elastic signal.
After the photoelectric conversion, the signals detected by the three PMTs in the three detection channels are recorded simultaneously in both analog (AD) and photon counting (PC) modes by a three-channel Licel transient digitizer (TR40-160, Licel, DE). The Licel digitizer has 40-MHz analog sampling rate and 800-MHz PC rate. Each data sampling has a bin width of 25 ns (corresponding to a height resolution of 3.75 m) and an integration time of 60 s (corresponding to 1650 laser shots plus a 5-s data transfer). A home-made trigger circuit controls the entire PRRL to operate orderly and automatically. A computer is utilized for data acquisition, system control, and real-time monitoring of the lidar operation status. Table 1 summarizes the main system parameters for the overall PRRL.
Table 1. Main system parameters for the PRRL.a
2.2. Lidar container for polar observation
To enable the long-term stable operation of the PRRL at Zhongshan Station under severe polar meteorological environments, a specially-customized container for polar observation was manufactured and deployed to house the PRRL. Figure 2 gives pictures of the lidar container.
Fig. 2. Pictures of (a) inside of the container with the PRRL system under operation; (b) the window system on the roof of the container; (c) outer scene of the PRRL working at Zhongshan Station, Antarctica.
Figure 2(a) provides a picture of the container inside with the PRRL under operation after its on-site installation and adjustment in the container. The inside of the container has a length of 5.66 m, a width of 2.10 m and a height of 2.45 m. Special measures are taken to enhance its windproof, waterproof and shockproof abilities. Temperature control is also considered to keep the room temperature to be around 20 ℃ in the container. To help room temperature and cleanliness maintenance during lidar operation periods, a window system is constructed on the roof of one inner side of the container (see Fig. 2(a)). The window system contains one laser exit hole and three echo entrance holes mechanically drilled on an aluminum plate (Fig. 2(b)). The laser exit hole enables a glass window to be positioned with a tilted angle of ∼5° off zenith. The glass window in the laser exit hole has a diameter of 89 mm and a thickness of 12 mm. The tilted positioning angle of ∼5° off zenith is to prevent reflection of propagating laser beam by the bottom flat surface of the glass window back into the lidar transmitting optics. Currently, only one echo entrance hole is in use and the residual two are closed for future extending application (Fig. 2(b)). The used echo entrance hole also has a glass window in place but without tilting. The glass window in the echo entrance hole has a diameter of 239 mm and a thickness of 25 mm. Note both glass windows are made of fused silica and coated to allow lights in 527-537 nm range to transmit with averaging transmittance of >99%. Moreover, to avoid possible condensations of water vapor onto the outer surfaces of the glass windows, an air-blower (not shown) is equipped. When necessary, the air-blower blows out warm air flows which first pass through the black rubber pipes (see Fig. 2(b)) and then get out of the blast holes to blow across the outer surfaces of the glass windows. Under unsuitable observational meteorological conditions (such as heavy cloud, strong wind, snowstorm etc.), an electronically controlled mechanical cover is closed so that the window system can be covered and isolated from the outside. Figure 2(c) also provides an outer scene of the PRRL working at Zhongshan Station, Antarctica.
2.3. Adjustment for the lidar optics
After shipment and arrival at Zhongshan Station, the transmitter and receiver of the PRRL had to be reassembled into the lidar container for polar observation. Procedures were taken to guarantee all the optics of the PRRL in position.
2.3.1 Adjustment for the lidar transmitter
Figure 3 presents the optical principle for adjusting the lidar transmitter. As shown in Fig. 3(a), a removable plate (RP) was first screwed onto the laser exit hole of the window system. The RP has a pinhole with a diameter of ∼1.5 mm in the center. The mechanical design makes sure that the geometric central axis of the RP coincides with that of the laser exit hole of the window system after installation of the RP. Then a laser plummet (LP) was placed above the RP. The LP emitted a visible (red) downward propagating laser light that passed through the central pinhole of the RP. The LP was pre-calibrated so that the downward laser light propagated with an angle of <0.1 mrad off zenith. A first optical bench (not shown) was subsequently placed into the container with its surface center illuminated by the visible downward laser light. Note the bench surface was adjusted to be horizontal with the help of a horizontal bubble. Later, the RM1 was placed onto the optical bench by positioning the RM1 until its surface center was illuminated by the visible downward laser light. Since the mechanical design yielded an initial working angle of 45° for the RM1 so that the reflected laser light propagated nearly horizontally. At this stage a second optical bench for loading the solid laser was positioned into the container with its bench surface adjusted to be horizontal with the help of the horizontal bubble. Then the solid laser was placed onto the second optical bench with its exit hole illuminated by the reflected horizontal laser light. Last, the LP was removed and the solid laser operated with attenuated 532.23-nm laser light output (e.g., 2-5 mJ per pulse). By adjusting the mechanical mounting holders of the solid laser, the attenuated 532.23-nm laser light got through the central pinhole of the RP after being reflected to vertical by the surface center of the RM1 (see Fig. 3(b)). Note these above adjustment procedures located the two optical benches, the RM1 and the solid laser. Besides, the output laser beam exiting the laser exit hole of the window system was nearly zenithward. For routine operation, the RP was replaced by the glass window, allowing the 532.23-nm laser beam to transmit. The BE was subsequently positioned between the solid laser and the RM1 on the second optical bench. So that the expanded laser beam illuminated the central surface area of the RM1 and passed through the central parts of the laser exit hole of the window system after being reflected by the RM1. The electronically controlled RM1 finally adjusted slightly and precisely the expanded laser beam direction to be zenithward.
Fig. 3. Principle for the PRRL transmitter adjustment: (a) The downward propagating laser light from the laser plummet (LP) passes through the central pinhole of the removable plate (RP) and then enters the exit hole of the solid laser after being reflected to horizontal by the RM1; (b) The horizontal laser light from the solid laser gets through the central pinhole of the RP after being reflected to vertical by the RM1.
2.3.2 Adjustment for the lidar telescope
The telescope needed to be located on the first optical bench with its optical axis being adjusted to be zenithward. The location procedures were similar to those of locating the RM1: another larger RP with a pinhole in the center was screwed onto the echo entrance hole of the window system and the downward laser light from the LP passed through this pinhole; then the Cassegrain telescope was placed onto the first optical bench by positioning the telescope until the top surface center of its secondary mirror (see Fig. 4(a)) was illuminated by the visible downward laser light. At this stage, the RP was removed, but the LP was repositioned above the echo entrance hole so that the downward laser light illuminated the marginal part of the telescope primary mirror (Fig. 4(a)). Then following successive procedures were taken to align the optical axis of the telescope to be zenithward:
Fig. 4. Principle for adjusting the telescope optical axis to be zenithward: (a) Side view. The downward propagating laser light from the LP illuminates the marginal part of the telescope primary mirror and passes through the iris. A white blank paper screen is placed behind the iris for viewing the light spots. Dashed line indicates the optical axis of the telescope; (b) Top view. While rotating the LP, the determined circle center (the cross point of the two thick dashed lines) by the corresponding light spots is recorded on the paper screen, but off that of the background circle when the telescope optical axis is not zenithward; (c) Top view. Adjusting the telescope system pointing until the determined circle center coincides with that of the background circle, indicating that the telescope optical axis is zenithward.
Step 1: Set the telescope iris with maximum diameter. Adjust the telescope pointing so that the laser light from the LP could pass through the iris. Then place a white blank paper behind the iris to serve as a paper screen and record the transmitted laser light spot (Fig. 4(a)). Note a much larger background circle (due to illumination of background sky light through the telescope; see Fig. 4(b)) is always discernible on the paper screen even if the LP is removed;
Step 2: Rotate the LP a full circle with a fixed step angle (90° for this case) and record positions of the corresponding (four) light spots on the paper screen. Adjust the telescope pointing so that all (four) light spots can be found in the background circle (Fig. 4(b));
Step 3: Reduce the iris diameter. Repeat Step 2 and Step 3 until the iris diameter can no longer be decreased;
Step 4: Use geometrical centers of all the (four) recorded light spots to determine a circle (indicated by the dashed circle). The circle center is further located (e.g., by the cross point of the two thick dashed lines connecting the centers of two opposite spots, see Fig. 4(b)). Adjust the telescope pointing finely and repeat Step 2 so that this located circle center coincides with that of the background circle (Fig. 4(c)). Lock the mechanical mounting holders so that the telescope pointing no longer varies.
Note the downward laser light from the LP propagates with an angle of <0.1 mrad off zenith, the Step 3 only aligns the optical axis of the telescope to be approximately zenithward. Then the Step 4 eliminates the influence of pointing deviation of the LP off zenith and guides the optical axis of the telescope to be accurately zenithward. The advantage of the current approach is that it uses only one cheap commercial LP to adjust the telescope zenithward; besides, it is rather convenient for on-site adjustment during (polar) day periods.
2.3.3 Adjustment for the subsequent receiving optics
The subsequent receiving optics utilizes two cubic nonpolarizing beam splitters to separate the wanted elastic and Raman signals. Figure 5(a) shows a top view of the subsequent receiving optics. The signal light from the iris is first redirected to horizontal by the RM2 and then collimated by the L0. After the BPF, the parallel light is divided by the BS1 into two parts: the reflected light is delivered to the EX1 and extracted for the elastic signal, while the transmitted light is incident on the BS2. The reflected light by the BS2 is guided to EX2 and extracted for the low-J Raman signal, while the transmitted light by the BS2 is guided to EX3 and extracted for the high-J Raman signal. The components of RM2, L0, BPF, BS1 and BS2 are all installed on adjustable mounts, which can be fixed with clamping screws. The total subsequent receiving optics is compact with a length of 267 mm, a width of 92 mm, and a height of 85 mm. It is convenient to connect to the telescope system. In addition, its exits (EX1, EX2 and EX3) are internally SM1-threaded holes.
Fig. 5. (a) Top view of the subsequent receiving optics. Central solid black line indicates the mechanical central axis of light propagating channel; (b) Schematic of the calibration module; (c) Schematic of the detection module. AL, achromatic lens. CMOS, the CMOS camera.
For a convenient visual adjustment of the optical components in the subsequent receiving optics, a home-made calibration module (see Fig. 5(b)) is utilized for the adjustment [32]. It consists mainly of an achromatic lens (AL) and a camera with a complementary metal-oxide-semiconductor (CMOS) sensor. The AL (AC254-150-A-ML, Thorlabs, US) is mounted on an axially movable lens tube (SM1ZM, Thorlabs, US) and has a focal length of ∼150 mm. The CMOS camera (DCC1240C, Thorlabs, US) is positioned through a c-mount adaptor. A controlling computer is employed to show real-time images captured by the camera. The entrance side of the calibration module is machined to have SM1 external thread so that it can be easily screwed into EXs (EX1, EX2 and EX3) of the subsequent receiving optics. The calibration module has been pre-regulated so that the photosensitive surface of the CMOS camera lies on the focal plane of the AL.
The regulated calibration module is used to accomplish the following adjustment to the optical components in the subsequent receiving optics. First, the RM2, L0 and BPF are mounted and the calibration module is screwed into the EX3 through the SM1 external thread. By adjusting the RM2, the iris on the focal plane of the telescope is imaged on the center of the CMOS camera so that the working angle of RM2 is determined. Second, moving the L0 along the optical axis until the iris is clearly imaged on the center of the CMOS camera, thus the positioning of L0 is finished. Note that the subsequent insertion of the BS1 and BS2 is unlikely to change the image location of the iris on the camera. Thus the calibration of the high-J Raman channel is accomplished. Third, the elastic channel is calibrated. The BS1 is installed and the calibration module is screwed into the EX1. The BS1 is adjusted until the iris is imaged on the center of the CMOS camera. This represents the accomplishment of the adjustment of the elastic channel. Forth, the low-J Raman channel is calibrated. The BS2 is mounted and the calibration module is screwed into the EX2. The BS2 is adjusted until the iris is imaged on the center of the CMOS camera, indicating the accomplishment of the adjustment of the low-J Raman channel. At this time, adjustments for all the optical components in the receiving optics are complete. After that, the calibration module is removed. Since the introduced CMOS camera has a pixel side length of 5.3 μm, the calibration module can guarantee that the angle deviation between the optical axis of the signal light beam (reflected by RM2, BS1 or BS2) and the mechanical central axis of light-propagating channel (indicated by central solid black line in Fig. 5(a)) is not larger than ∼35.3 μrad [32].
For extracting the wanted signals, three home-made detection modules that are corresponding to the three detection channels are utilized. As shown in Fig. 5(c), each detection module has SM1 external thread at the entrance side. In the detection module, the NFs and/or IFs can be conveniently installed or removed. The focusing lens (FL) with a focus length of 40 mm is used to focus the transmitted signal light on photosensitive surface of subsequent PMT. The PMT is connected through a home-made auxiliary mechanical device. The three detection modules are directly screwed into the EX1, EX2, and EX3 through the SM1 external thread, respectively. When the PMT in the detection module is replaced by the CMOS camera via the c-mount adaptor, a small light spot with a diameter of <1.5 mm is visible on the center of the CMOS camera. Considering the PMT has an effective photosensitive surface with a diameter of 8 mm, the signal light is entirely detected.
3. Methodology
3.1. Quasi-single-line-extraction technique
The PRRL was intended to perform both daytime and nighttime temperature measurements of the polar troposphere at Zhongshan Station. For the consideration of improving the daytime performance of the PRRL, narrowband optical devices are favored to extract the interested Raman line signals as well as restrain sky background noise. Figure 6 provides the theoretically calculated pure rotational Raman spectra of atmospheric N2 (red) and O2 (blue) molecules when radiated by a 532.23-nm laser light. One optional technical scheme is the single-line-extraction technique [31] that uses Fabry-Perot interferometers (FPI) to extract the single N2 line signals (either the N2 Anti-Stokes line signals with J=6 and 16, or the N2 Stokes line signals with J=4 and 14). Since the FPIs can have rather narrow spectral bandwidths (e.g., ∼30 pm), the daytime background light is thus greatly rejected. However, the FPIs usually need to work with fine temperature control and the whole subsequent receiving optics becomes rather complicated. To ease the technical difficulty and for convenience of on-site assembly and adjustment of the subsequent receiving optics, narrowband IFs instead of FPIs, are adopted for this case of polar application. As shown in Fig. 6, for a laser radiation of 532.23 nm, the N2 Stokes line signal with J=6 (12) has a wavelength of 533.93 (535.29) nm. So that the IF2 (IF3) with a FWHM of 0.2 nm centered at 533.92 (535.30) nm definitely extract the N2 Stokes J=6 (12) line signal, as indicated by the schematic transmittance curve of the IF (magenta). Note here the Stokes (rather than Anti-Stokes) PRR line signals are extracted due to stronger line signal intensity. Considering that Antarctic atmosphere generally has less abundant aerosols compared to those of middle and low latitude regions and our PPRL emits a 532.23-nm visible laser light, the risk of interference of PRR line signals with atmospheric fluorescence [21,34] eases. This issue is further verified by filed temperature measurement results (see Figs. 7 and 8). It is noticed in Fig. 6 that the O2 Stokes line signal with J=9 (17) also passes through the IF2 (IF3). It is concluded that each of the two Raman channels of the RPPL mainly extracts two line signals from atmospheric N2 and O2 molecules. This spectral extraction scheme is named as quasi-single-line-extraction technique in this work. Compared to the single-line-extraction technique [31], this scheme favors stronger Raman signal intensities. Besides, the choice of a 0.2-nm FWHM for IF2 and IF3 is a compromise between daytime background noise rejection ability and insensitivity of the IFs to variations of room temperature and filter working angle. Compared to FPIs, the 0.2-nm IFs have worse performance of suppressing the daytime background noise. But considering the Polar Regions have less intense solar radiation compared to middle and low latitude areas, the technical demand for narrowing the FWHM of the IFs weakens. Additional advantage of usage of the narrowband IFs is that it enables straightforward and compact optical design for the whole subsequent receiving optics (see Fig. 1).
Fig. 6. Theoretically calculated pure rotational Raman spectra of atmospheric N2 (red) and O2 (blue) molecules given a laser radiation of 532.23 nm. Magenta solid lines simulate the schematic transmittance curves of the interference filters employed in the low-J and high-J Raman channels for extracting the corresponding Raman lines.
Fig. 7. (a) Raw photon count profiles (solid) for the low-J (red) and high-J (blue) Raman channels, respectively. Dotted lines indicate the daytime background level; (b) Lidar temperature profile (blue) as well as accompanying radiosonde data (red); (c) 1-σ statistical uncertainty (magenta) and deviation between lidar and radiosonde (blue). The lidar measurement was made during 0636-0736 UTC on April 27, 2020. The height resolution is 90 m. Note only lidar temperature with 1-σ statistical uncertainty less than 5 K is shown.
Fig. 8. Same as Fig. 7 but during 1425-1525 UTC on May 14, 2020. Note the background levels for this nighttime measurement case are too small to be shown in the leftmost panel.
3.2. Data processing and lidar calibration
The output signals from the PMTs are recorded simultaneously in both AD and PC modes with the Licel transient digitizer. For a raw signal profile registered by the Licel, the AD data correctly represent the actual lidar signal intensity at lower heights, while the PC data exactly reflect the actual lidar signal intensity at higher heights. In order to obtain a complete lidar signal profile, the AD and PC data are glued [35,36] to form a reasonable PC rate profile (in megahertz) with large dynamic range. The glued PC profiles have a range resolution of 30 m and a time resolution of 1 min. The coarser resolution signal profiles can be obtained simply by integrating the glued profiles in range and time.
The atmospheric temperature T(z) at height z can be retrieved from the ratio Q(z) of the two Raman channel signals using the following calibration function [22,27,33]:
Assuming a Poisson statistics, the 1-σ statistical uncertainty of the temperature measurements can be calculated [22]:
Equation (2) indicates that the statistical uncertainty is small when the photon counts of the two Raman channels are large. Considering that the Raman photon counts fall with increasing height, a sliding-smoothing algorithm [32] can be introduced to control the 1-σ statistical uncertainty:
For determination of the three calibration constants (A, B, and C), temperature profile from radiosonde data at Zhongshan Station is utilized. Currently at Zhongshan Station the radiosonde (RS41-SG, Vaisala) was launched approximately once a month (∼200 m away from the PRRL system) when meteorological conditions were suitable and the PPRL had entered into the automatic observation mode. The radiosonde temperature profile is first linearly interpolated to have identical height intervals as the lidar data. Then simultaneous the lidar-derived Raman signal ratio Q(z) and the interpolated radiosonde temperature profile during the same time interval (usually 1 h from releasing the radiosonde), the three calibration constants are thus retrieved from Eq. (1).
4. Measurement examples
After on-site installation and adjustments, the PRRL started routine operations since February 6, 2020 when meteorological conditions were suitable at Zhongshan Station. The tested minimum height with full overlap for the PRRL is ∼500 m above ground level (ABL). Hence temperature measurement results are discussed here only for heights above 500 m.
Figure 7 gives a daytime measurement example performed during 0636-0736 UTC on April 27, 2020. To increase SNR, the glued PC rate profile is first changed to photon count profile and then accumulated in height with a lower height resolution of 90 m. Figure 7(a) shows the background-subtracted photon count profiles (solid lines) of the low-J (red) and high-J (blue) Raman channels, respectively. The two raw Raman signal profiles are smooth at lower heights, but exhibit obvious fluctuations at higher heights (e.g., above 7 km). For comparison, the daytime background levels are also provided (dotted lines). It is clear that the daytime background level is still high even if the narrowband IFs are introduced. SNRs for the low (high)-J Raman channel signals quickly decrease to be <1 at heights above 2.5 (2.7) km. Figure 7(b) plots the lidar-retrieved temperature profile (blue) and concurrent radiosonde temperature results (red). The three calibration constant values are determined to be -45532.40, 139.911 and 0.6431 for A, B and C, respectively by Eq. (1). To control the 1-σ temperature measurement uncertainty, the respective height sliding windows (HSW) are set to be 270 m at height ranges of 2-4 km, 450 m at height ranges of 4-8 km, and 630 m at height ranges of 8-15 km. Note plot of lidar temperature results is limited to heights where the corresponding 1-σ statistical uncertainty is less than 5 K. The lidar-derived temperature generally agrees with that of the concurrent radiosonde at heights of 0.5-5.9 km. As shown in Fig. 7(c), the 1-σ statistical uncertainty (magenta) is <2 K at heights below ∼3.3 km and grows to be <5 K at heights below ∼5.9 km. The absolute temperature deviation (blue) between lidar and concurrent radiosonde measurements is <1 K below ∼2.2 km and <4.1 K below ∼5.9 km. An inversion structure is observable around 2.8 km as indicated by both lidar and radiosonde results, exhibiting an absolute deviation reaching 2.7 K.
Figure 8 presents a nighttime measurement example performed during 1425-1525 UTC on May 14, 2020. It can be seen in Fig. 8(a) that both the background-subtracted photon count profiles (solid lines) of the low-J (red) and high-J (blue) Raman channels are visually smooth without apparent fluctuations in the height range of 0.5-15 km. Besides, both the background levels are too small (<10) to be shown in the x-axis scaling range. SNRs are larger than 30 even at higher heights around 15 km (not shown here). Figure 8(b) compares the lidar-retrieved temperature profile (blue) and concurrent radiosonde temperature results (red). The three calibrated values are -28845.80, -81.7619 and 1.048320 for A, B, and C, respectively. Identical HSWs as those used in Fig. 7 are applied to yield the lidar temperature profile. It is found that the lidar-derived temperatures are generally coincident with the concurrent radiosonde temperature results. In the height range of 0.5-10 km, the 1-σ statistical uncertainty (magenta) is <1.5 K and the absolute deviation <1.8 K. At heights between 10 and 15 km, the maximum 1-σ statistical uncertainty grows to be ∼3.5 K, while the maximum absolute deviation enlarges to be ∼5.1 K around 14 km; in this height range, several temperature inversion structures are clearly indicated by both lidar and radiosonde results. Temperatures near the inversion areas always show obvious variations [27,32]. Thus the larger temperature discrepancy between lidar and radiosonde in 10-15 km height range reflects the actual temperature variations during the lidar observational period.
Figure 9 shows the time-height contour plot of the lidar-derived temperature at heights of 0.5–15 km with a temporal and spatial resolution of 1 h and 90 m, respectively, spanning from 1100 UTC on May 14 to 1600 UTC on May 15, 2020. Sunrise (SR) and sunset (SS) times are indicated by black dashed lines. Temperature results with 1-σ statistical uncertainty larger than 5 K are excluded. Generally the temperature decreases as a whole with increasing height between 1.5 and 10.8 km, while it increases non-obviously with increasing height above 10.8 km. The temperature variations are visible at lower heights. An inversion layer at heights below 1.5 km began forming around 1100 UTC on the first day and persisted all over the observational period. The temperature at heights below 6 km showed a decreasing trend until 1800 UTC on the first day, and then exhibited an increasing trend until 0800 UTC on the next day.
Fig. 9. Time-height contour plot of the lidar-derived 30-h temperatures between 0.5 and 15 km on the date of May 14-15, 2020. The temporal and spatial resolutions are 1 h and 90 m, respectively. Sunrise (SR) and sunset (SS) times are indicated by black dashed lines. Temperatures with 1-σ statistical uncertainty larger than 5 K are excluded.
Figure 10 investigates 1-h-to-1-h temperature variability in the height ranges of 0.5–15 km on the observational date of May 14-15, 2020. Figure 10(a) shows the 1-h lidar temperature profiles (black) and their mean (magenta). For comparison, the radiosonde temperature is also plotted. As seen from Fig. 10(a), the tropopause is at ∼10.8 km AGL with a mean temperature of ∼203 K. Figure 10(b) plots the mean absolute deviation (blue) and the maximum absolute deviation (red) of the 1-h lidar temperature profiles from the nightly mean temperature profile. It is found that at the inversion heights between 1.0 and 1.5 km, the temperature has a relatively larger mean deviation of 2.0–3.2 K (the maximum deviation is of 4.8–7.2 K), compared to that at upper heights of 1.5–10.0 km with a mean deviation of 1.2-2.1 K (the maximum deviation is of 3.4–7.2 K). At heights between 3.5 and 9.5 km, the 1-h temperature profiles are relatively similar in form (Fig. 10(a)). The fitted average temperature lapse rate at this height range is -8.3 K/km. Around the tropopause region of 10-12 km, the mean deviation grows to be 1.6–4.4 K (the maximum deviation is of 5.6–10.8 K). At upper heights of 12-15 km, the mean deviation decreases to be 2.1–3.4 K (the maximum deviation is of 4.9–11.0 K). It is concluded that significant temperature variability occurs only at the inversion layer heights.
Fig. 10. (a) 1-h lidar temperature profiles (black) and their average (magenta) on the date of May 14-15, 2020. For comparison, the radiosonde temperature profile is also plotted (red). (b) Mean absolute deviation (blue) and maximum absolute deviation (red) of the 1-h lidar temperature profiles from the mean temperature profile.
5. Summary
We have presented the first pure rotational Raman lidar (PRRL) built recently for full-day troposphere temperature measurements at Zhongshan Station (69.37°S, 76.37°E), Antarctica by the 36th Chinese National Antarctic Research Expedition. The lidar emits a 532.23-nm laser light and utilizes a 203.2-mm telescope to collect atmospheric backscatter. Simple cubic nonpolarizing beam splitters are employed to yield compact arrangement of the subsequent receiving optics. Interference filters (IF) centered at the N2 Stokes line wavelengths of J=6 and 12 are used to extract the Raman signals, respectively. Each of the IFs has a FWHM of ∼0.2 nm and mainly extracts two line signals (specifically, the Stokes N2 J=6 and O2 J=9 lines for the low-J Raman channel and the Stokes N2 J=12 and O2 J=17 lines for the high-J Raman channel, respectively). This quasi-single-line-extraction technique enables use of narrowband (0.2 nm) IFs which benefits lidar daytime performance and convenience of lidar optics design. A lidar container is customized to house the whole PRRL system, enabling the long-term stable operation of the PRRL under harsh polar meteorological environments. A self-developed window system is equipped on the roof of the container to help maintaining the temperature and cleanliness inside the container, as well as allowing the outgoing laser and atmospheric backscatter to pass through.
An approach using a laser plummet (LP) is introduced for fast and convenient location and adjustment of the reflecting mirror (RM1), the laser and the telescope in the container. A unique adjusting method with the help of the LP is developed for convenient and accurate adjustment of guiding the telescope pointing to be zenithward. A home-made calibration module is utilized for straightforward visual adjustment of the subsequent receiving optics with ∼35.3-μrad angular positioning accuracy. A modular design of the detection module provides fast and convenient connection to the subsequent receiving optics.
Both typical daytime and nighttime temperature measurement examples are presented with a height resolution of 90 m and a time resolution of 1 h. For the daytime measurement case, the 1-σ statistical uncertainty is <2 K at heights below ∼3.3 km and <5 K below ∼5.9 km; the absolute temperature deviation between lidar and concurrent radiosonde results is <1 K below ∼2.2 km and <4.1 K below ∼5.9 km. While for the nighttime measurement case, the 1-σ statistical uncertainty is generally <1.5 K and the absolute deviation <1.8 K in the height range of 0.5-10 km; at heights of 10-15 km, the maximum 1-σ statistical uncertainty grows to be ∼3.5 K, while the maximum absolute deviation enlarges to be ∼5.1 K around the inversion areas.
A 30-h continuous temperature measurement results by the PRRL have been investigated. The tropopause is at ∼10.8 km AGL with a mean temperature of ∼203 K. It is found that at the inversion heights between 1.0 and 1.5 km, the 1-h temperature profiles have a relatively larger mean deviation of 2.0-3.2 K from the average temperature profile. Similar characteristics exist around the tropopause region of 10-12 km where the mean deviation grows to be 1.6–4.4 K. While at heights of 1.5–10.0 km the mean deviation decreases to be 1.2-2.1 K; at heights between 3.5 and 9.5 km, the 1-h temperature profiles are relatively similar in form with a fitted average temperature lapse rate of -8.3 K/km. It is concluded that the significant variability occurs only at the inversion layer heights.
It is proven that the PRRL can perform full-day temperature measurement with a time resolution of 1 h and a height resolution of 90 m in the troposphere, enabling studies of the thermal structures and temperature evolution characteristics at Zhongshan Station, Antarctica.
Funding
National Key Research and Development Program of China (2018YFC1407301, 2016YFC1400301); National Natural Science Foundation of China (41831072, 41927804).
Acknowledgments
We sincerely thank the 35th and 36th Chinese National Antarctic Research Expedition (CHINARE) for their contributions to the lidar observatory infrastructure constructions at Zhongshan Station. We are grateful to lidar team (Wentao Huang, Rui Wang, Kaijie Ji, Zhaoling Zeng, Feng Zhang and Hui Li) for on-site lidar installation and data collection. We also appreciate Yunpeng Zhang, Yang Yi and Liang Peng for technical support and development of the lidar, and Zhenping Yin and Yun He for lidar data processing and science discussions.
Disclosures
The authors declare that there are no conflicts of interest related to this article.
References
1. IPCC Report, “Climate change 2007: The physical science basis, contribution of working group I to the fourth assessment report of the intergovernmental panel on climate change,” (Cambridge University, 2007).
2. M. P. Baldwin and T. J. Dunkerton, “Stratospheric harbingers of anomalous weather regimes,” Science 294(5542), 581–584 (2001). [CrossRef]
3. J. W. Meriwether and A. J. Gerrard, “Mesosphere inversion layers and stratosphere temperature enhancements,” Rev. Geophys. 42(3), RG3003 (2004). [CrossRef]
4. K. Mohankumar, “Stratosphere troposphere interactions: An introduction,” (Springer, 2008).
5. A. Gettelman, P. Hoor, L. L. Pan, W. J. Randel, M. I. Hegglin, and T. Birner, “The extratropical upper troposphere and lower stratosphere,” Rev. Geophys. 49(3), RG3003 (2011). [CrossRef]
6. A. K. Smith, “Global Dynamics of the MLT,” Surv. Geophys. 33(6), 1177–1230 (2012). [CrossRef]
7. C. S. Gardner, G. C. Papen, X. Chu, and W. Pan, “First lidar observations of middle atmosphere temperatures, Fe densities, and polar mesospheric clouds over the north and south poles,” Geophys. Res. Lett. 28(7), 1199–1202 (2001). [CrossRef]
8. X. Chu, W. Pan, G. Papen, C. S. Gardner, and J. Gelbwachs, “Fe Boltzmann temperature lidar: design, error analysis, and initial results at the North and South Poles,” Appl. Opt. 41(21), 4400–4410 (2002). [CrossRef]
9. T. D. Kawahara, T. Kitahara, F. Kobayashi, Y. Saito, A. Nomura, C.-Y. She, D. A. Krueger, and M. Tsutsumi, “Wintertime mesopause temperatures observed by lidar measurements over Syowa Station (69°S, 39°E), Antarctica,” Geophys. Res. Lett. 29(15), 4-1–4-4 (2002). [CrossRef]
10. T. D. Kawahara, T. Kitahara, F. Kobayashi, Y. Saito, and A. Nomura, “Sodium temperature lidar based on injection seeded Nd:YAG pulse lasers using a sum-frequency generation technique,” Opt. Express 19(4), 3553–3561 (2011). [CrossRef]
11. F. J. Lübken, J. Hoffner, T. P. Viehl, B. Kaifler, and R. J. Morris, “First measurements of thermal tides in the summer mesopause region at Antarctic latitudes,” Geophys. Res. Lett. 38(24), L24806 (2011). [CrossRef]
12. R. J. Morris, J. Hoffner, F. J. Lubken, T. P. Viehl, B. Kaifler, and A. R. Klekociuk, “Experimental evidence of a stratospheric circulation influence on mesospheric temperatures and ice-particles during the 2010–2011 austral summer at 69◦S,” J. Atmos. Solar Terr. Phys. 89, 54–61 (2012). [CrossRef]
13. A. Hauchecorne and M. Chanin, “Density and temperature profiles obtained by lidar between 35 and 70 km,” Geophys. Res. Lett. 7(8), 565–568 (1980). [CrossRef]
14. G. D. Donfrancesco, A. Adriani, G. P. Gobbi, and F. Congeduti, “Lidar observations of stratospheric temperature above McMurdo Station, Antarctica,” J. Atmos. Solar Terr. Phys. 58(13), 1391–1399 (1996). [CrossRef]
15. W. Pan and C. S. Gardner, “The temperature structure of the winter atmosphere at South Pole,” Geophys. Res. Lett. 29(16), 49-1–49-4 (2002). [CrossRef]
16. A. R. Klekociuk, M. M. Lambert, R. A. Vincent, and A. J. Dowdy, “First year of Rayleigh lidar measurements of middle atmosphere temperatures above Davis, Antarctica,” Adv. Space Res. 32(5), 771–776 (2003). [CrossRef]
17. S. P. Alexander, A. R. Klekociuk, and D. J. Murphy, “Rayleigh lidar observations of gravity wave activity in the winter upper stratosphere and lower mesosphere above Davis, Antarctica (69°S, 78°E),” J. Geophys. Res. 116(D13), D13109 (2011). [CrossRef]
18. H. Suzuki, T. Nakamura, M. K. Ejiri, M. Abo, T. D. Kawahara, Y. Tomikawa, and M. Tsutsumi, “A Rayleigh Raman lidar system for troposphere-mesosphere observations at Syowa station, Antarctica,” the 26th International Laser Radar Conference (ILRC 2012), S9P-18.
19. Y. F. Arshinov, S. M. Bobrovnikov, V. E. Zuev, and V. M. Mitev, “Atmospheric temperature measurements using a pure rotational Raman lidar,” Appl. Opt. 22(19), 2984–2990 (1983). [CrossRef]
20. D. Nedeljkovic, A. Hauchecorne, and M.-L. Chanin, “Rotational Raman lidar to measure the atmospheric temperature from the ground to 30 km,” IEEE Trans. Geosci. Remote Sensing 31(1), 90–101 (1993). [CrossRef]
21. A. Behrendt and J. Reichardt, “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(9), 1372–1378 (2000). [CrossRef]
22. A. Behrendt, T. Nakamura, M. Onishi, R. Baumgart, and T. Tsuda, “Combined Raman lidar for the measurement of atmospheric temperature, water vapor, particle extinction coefficient, and particle backscatter coefficient,” Appl. Opt. 41(36), 7657–7666 (2002). [CrossRef]
23. A. Behrendt, T. Nakamura, and T. Tsuda, “Combined temperature lidar for measurements in the troposphere, stratosphere, and mesosphere,” Appl. Opt. 43(14), 2930–2939 (2004). [CrossRef]
24. P. D. Girolamo, R. Marchese, D. N. Whiteman, and B. B. Demoz, “Rotational Raman lidar measurements of atmospheric temperature in the UV,” Geophys. Res. Lett. 31(23), L01106 (2004). [CrossRef]
25. Y. Arshinov, S. Bobrovnikov, I. Serikov, A. Ansmann, U. Wandinger, D. Althausen, I. Mattis, and D. Müller, “Daytime operation of a pure rotational Raman lidar by use of a Fabry–Perot interferometer,” Appl. Opt. 44(17), 3593–3603 (2005). [CrossRef]
26. M. Radlach, A. Behrendt, and V. Wulfmeyer, “Scanning rotational Raman lidar at 355 nm for the measurement of tropospheric temperature fields,” Atmos. Chem. Phys. 8(2), 159–169 (2008). [CrossRef]
27. J. Jia and F. Yi, “Atmospheric temperature measurements at altitudes of 5–30 km with a double-grating-based pure rotational Raman lidar,” Appl. Opt. 53(24), 5330–5343 (2014). [CrossRef]
28. A. Behrendt, V. Wulfmeyer, E. Hammann, S. K. Muppa, and S. Pal, “Profiles of second- to fourth-order moments of turbulent temperature fluctuations in the convective boundary layer: first measurements with rotational Raman lidar,” Atmos. Chem. Phys. 15(10), 5485–5500 (2015). [CrossRef]
29. E. Hammann, A. Behrendt, F. Le Mounier, and V. Wulfmeyer, “Temperature profiling of the atmospheric boundary layer with rotational Raman lidar during the HD(CP)2 observational prototype experiment,” Atmos. Chem. Phys. 15(5), 2867–2881 (2015). [CrossRef]
30. V. Wulfmeyer, R. M. Hardesty, D. D. Turner, A. Behrendt, M. P. Cadeddu, P. D. Girolamo, V. Schlüssel, J. V. Baelen, and F. Zus, “A review of the remote sensing of lower tropospheric thermodynamic profiles and its indispensable role for the understanding and the simulation of water and energy cycles,” Rev. Geophys. 53(3), 819–895 (2015). [CrossRef]
31. M. Weng, F. Yi, F. Liu, Y. Zhang, and X. Pan, “Single-line-extracted pure rotational Raman lidar to measure atmospheric temperature and aerosol profiles,” Opt. Express 26(21), 27555–27571 (2018). [CrossRef]
32. F. Liu, F. Yi, Y. Zhang, and Y. Yi, “Double-receiver-based pure rotational Raman lidar for measuring atmospheric temperature at altitudes between near ground and up to 35 km,” IEEE Trans. Geosci. Remote Sensing 57(12), 10301–10309 (2019). [CrossRef]
33. A. Behrendt, “Temperature measurements with lidar,” in Lidar-Range-Resolved Optical Remote Sensing of the Atmosphere, (Springer, 2015), pp. 273–305.
34. T. Kitada, A. Hori, T. Taira, and T. Kobayashi, “Strange behaviour of the measurement of atmospheric temperature profiles of the rotational Raman lidar,” in Proceedings of the 17th International Laser Radar Conference (National Institute for Environmental Studies, Tsukuba, Japan, 1994), pp. 567–568.
35. R. K. Newsom, D. D. Turner, B. Mielke, M. Clayton, R. Ferrare, and C. Sivaraman, “Simultaneous analog and photon counting detection for Raman lidar,” Appl. Opt. 48(20), 3903–3914 (2009). [CrossRef]
36. Y. Zhang, F. Yi, W. Kong, and Y. Yi, “Slope characterization in combining analog and photon count data from atmospheric lidar measurements,” Appl. Opt. 53(31), 7312–7320 (2014). [CrossRef]
