Open Access
Add to BibSonomy
Abstract
The sodium fluorescence lidar utilizes a 589 nm narrowband pulse laser system to measure mesopause region atomic sodium density, atmospheric temperature, and wind. However, this system is complicated and unstable. The continuous-wave (CW) sodium laser system can achieve ultra-narrow bandwidth, all-solid-state, and small compact size, as such it is extremely valuable for mobile, aircraft, and space-borne applications. In this study, we developed the first pseudo-random modulated CW (PMCW) sodium lidar by using an electro-optic modulated narrowband 589 nm CW laser with an output power of ∼1.2W. A pseudorandom M-sequence-code with a length of 127 is used to achieve altitude information by modulating laser and then decoding photon signals. Also, a biaxial structure with 9 m separation between the optical axes of the transmitter and receiver is designed to suppress the strong near-ground signals, which are crucial for improving the signal-to-noise ratio (SNR) of the PMCW lidar system. Nighttime measurements on December 2-4, 2019 show that the SNR at sodium layer peak is more than 10, corresponding to a statistical uncertainty of less than 10% in sodium density with temporal and spatial resolutions of 5 min and 1.05 km respectively. The comparison of vertical profiles of sodium density simultaneously observed by PMCW lidar and collocated pulse lidar shows good agreement.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Lidar has been one of the most effective and high precision instruments for atmospheric detection since the invention of the laser [1,2]. The large fluorescence cross-sections of the sodium (Na) atoms in the mesopause region (80-105 km) [3–5] make the lidar remote sensing measurements in this region possible. The sodium lidar system has been widely used to detect the sodium density and/or atmospheric parameters (narrowband), e.g. temperature and wind in the mesopause region [6–9]. The laser system to generate narrowband sodium pulse light at 589 nm usually includes a liquid dye amplifier pumped by a high-power (∼20W) 532 nm pulse laser beam [8]. However, the ring cavity and the pulsed dye amplifier (PDA) are sensitive to environmental temperature and vibrations, which induces complex operations and instability.
Recent progress in all-solid-state narrowband pulse sodium laser makes the mobile system possible. One practicable approach is the sum-frequency generation (SFG) technique to generate narrowband 589 nm beam from solid-state Nd: YAG lasers at 1064 nm and 1319 nm [10,11]. Several sodium lidar systems have been developed with SFG as seeders into two diode-pumped pulsed Nd: YAG lasers at 1064 nm and 1319 nm, and the narrowband pulse 589 nm beam frequency summed inside the nonlinear crystal [12–15]. However, most 589 nm pulsed lasers still suffer some limitations, e.g. the large volume of the transmitter due to the need for two Nd: YAG lasers and the high cost of customizing the Nd: YAG lasers for specific wavelengths.
Continuous-wave (CW) radar systems were first used in microwave Radar [16], in which frequency-modulated CW (FM-CW) radar successfully measured the hydrometeor and wind around the boundary layer [17,18]. The first CW lidar application was demonstrated by Takeuchi [19] with a ranging-capability aerosol lidar utilizing the CW laser sources modulated by pseudorandom M-sequence-code (M-code). Further development of the CW lidar technique yielded higher-vertical-resolution cloud detection and multi-frequency carbon dioxide detection [20–22]. In the 17 International Laser Radar Conference, Abo and Nagasawa proposed the use of a random modulation CW lidar for probing the mesospheric Na layer [23]. She et al. (2011) simulated a pseudorandom modulation CW (PMCW) sodium lidar that can achieve a similar signal-to-noise ratio comparable to a pulse system [24]. This simulated lidar system directly uses an all-solid-state CW laser as the source of light, which makes the power consumption of the PMCW lidar system less than half that of the same power pulse lidar system. Also, the PMCW lidar, which does not require pumping amplification, makes an all-solid-state and compact system possible. From the perspective of detection accuracy, on the one hand, the PMCW lidar system has better frequency stability than a pulsed system as it does not require a pumping-amplification process [25]. On the other hand, as the energy density per unit time of CW laser is only one-millionth that of the high-power pulsed laser, there is no optical saturation in the sodium layer even when using a high-power CW laser and/or reducing the divergence angle of the laser.
In this study, we develop the first PMCW sodium lidar at the University of Science and Technology of China. Section 2 discusses the principle of the measurements and the PMCW lidar equation. The detailed components and specific parameters of the system are presented in Section 3. Section 4 presents the initial observation results and error analysis, followed by a summary in Section 5.
2. Lidar equation and signal-to-noise ratio
The transmitting laser in the PMCW lidar system is usually modulated with a random code and the received signal is then able to be decoded to achieve distance resolution. In our system, the pseudorandom modulation M-code is selected as the modulation code [26,27]. The acceptable length of the M-code is $N = {2^n} - 1$, e.g. 127 with n = 7 in this system. The M-code can be either one or zero, i.e., ${a_i} = 1\; \textrm{or}\; 0$. The time sequence of the inverse code satisfies the equation, ${a^{\prime}_i} = 2{a_i} - 1$. Then the cross-correlation of M-code, ${a_i}$ and inverse code, ${a^{\prime}_i}$ can be calculated as [19]
That is, only when the inverse code is in one-to-one correspondence with the M-code, the products of M-code and its inverse code are summed to obtain a non-zero value ($k = 0$), otherwise a zero ($k \ne 0$). This unique property allows the backscattering signal received by the PMCW lidar to be decoded to obtain altitude information. Table 1 shows the M-code with 127 lengths used in this system.
Table 1. Time Series of the Transmitting and Receiving M-Code and Their Correlation for N = 127
The PMCW lidar received raw signal ${y_i}$ can be expressed in a convolution form as [19]
Where ${P_0}$ is the CW laser power, $\Delta t$ is the length of time for a single code ${\alpha _i}$, $G({{h_j}} )$ is the atmospheric response of the lidar system at altitude ${h_j}$ (${h_j} = j\cdot c\cdot \Delta t/2$, c is the speed of light) and B is the average intensity of background noise in $\Delta t$.
Using the properties of M-code in Eq. (1), we can retrieve the signal with altitude information via the cross-correlation between the raw signal and inverse code as:
Since the retrieved signals with altitude information involve all signals received during each code within one period, the statistic noise of ${S_n}$ is:
Where ${\sigma _{{y_i}}}$ is the photon noise of the raw signal ${y_i}$.
Therefore, the signal-to-noise ratio (SNR) calculation formula of the whole system in an M-code cycle is as follows:
If the number of cycles (m) are integrated, the SNR can be rewritten as:
Unlike the pulse lidar, the noise term in the SNR formula of PMCW lidar includes all the signals in each code. To effectively improve the SNR of PMCW lidar, we need to increase the CW laser output power and reduce background noise and near-ground signals.
3. Lidar system
The schematic diagram of the PMCW sodium lidar is shown in Fig. 1, and the main parameters of the lidar system are summarized in Table 2.
Table 2. Main parameters of the PMCW lidar system
3.1 Transmitter
The lidar transmitter includes a diode laser with a digital laser controller, a feedback centralizer, a sodium cell for Doppler-free spectroscopy [8] that transmits laser wavelength information to the feedback centralizer, a modulation system, and a laser collimation system. The CW diode laser has a working adjustable wavelength near 589 nm and output power of ∼1.2W. A small portion (<1%) of light is sent to the sodium cell for frequency locking, the rest of this light beam is directed through a series of lenses and mirrors then injected into the modulation system. Based on the Doppler-free spectroscopy and the feedback centralizer, the laser-emitted light can be dynamically locked at 589.158 nm with 2 MHz frequency uncertainty.
3.1.1 Modulation
The modulation system includes an Electro-Optic Modulator (EOM), an arbitrary function generator, and a high voltage push-pull power amplifier capable of swinging output voltage in the order of 175 V push-pull. The arbitrary function generator outputs the M-code waveform signal of Transistor-Transistor-Logic (TTL) voltage level with both the rising and falling edges times of less than 5 ns. The TTL signal is amplified and offset by the amplifier to control the EOM crystal. Here the voltage amplification and offset are controllable. The parameters of EOM are given in Table 1. Some CW lidar systems used an Acousto-Optic Modulator (AOM) instead of EOM to modulate the CW laser beam. However, the EOM has higher energy conversion efficiency and does not change the laser frequency and direction, making the optical path simpler and more stable.
After beam expanding, the CW laser sends ∼1.05 W laser beam into EOM to be modulated with M-code. The final modulated beam is ∼ 840 mW at M-code 1 and below 14 mW at code 0. The extinction ratio of the EOM is ∼ 60:1. (To reduce power loss, Polarizing Beam Splitter (PBS) was not placed in front of EOM, thus the incomplete polarization of the laser makes the extinction ratio slightly lower than the factory standard of EOM). The rising and falling edge time of the modulated laser is ∼20 ns. The modulation noise comes from input voltage oscillation due to rapid voltage change in the circuit. The modulation signal frequency fed to the electro-optic crystal should avoid its resonance frequency (35 kHz and 482 kHz for this EOM) [28]. The errors due to the modulation code will be described in detail in Section 6.
3.1.2 Collimation
Since the PMCW lidar system needs to suppress the strong near-ground signal by increasing the blind range of field of view (FOV) with a biaxial structure [23], the requirement for alignment precision is extremely strict. To achieve high precision alignment, we use scanning stepping motors to control the laser reflector in azimuth and zenith directions, so that the laser can be precisely aligned with the FOV of the receiving telescope. This automatic alignment solution had been well demonstrated on the narrowband sodium temperature/wind lidar and ozone lidar developed at the University of Science and Technology of China (USTC) [8,29].
For alignment, we first use EOM to cut the CW beam into 20µs wide pulse for easy detection and use an oscilloscope to search for low-altitude signals similar to pulsed lidar. When the signals near 20 km appear on the oscilloscope, the pulse beam is replaced with M-code modulated beam for auto scan alignment. For each step of the motor, the signal is integrated for 10s and then retrieved to the photon count profile with altitude information. Similar to the pulsed lidar, Fig. 2 shows the signal curves at different heights (21.0 km, 25.2 km, 31.5 km, and 88.2 km) scanned in the tilt and rotation directions by the automatic collimation system. The signals at rotation direction present a typical trapezoid function, and the middle position of the trapezoid (the blue area in Fig. 2(a)) is selected as the best rotational position. Because of the 9 m distance biaxial structure, the maximum signal values in the tilt direction at different heights are not exact at the same position. Since we focus on the sodium layer signals at 88.2 km and the Rayleigh signals at 31.5 km, we set the motor at the middle position of these two maximum signals (the blue area in Fig. 2(b)). Under this alignment scheme, the laser beam is likely not fully overlap with the telescope FOV below 30 km, so the denominator likely reaches the minimum in the SNR ratio function in Eq. (6).
3.2 Receiver
The backscattering photons are collected by a vertical pointing telescope with a 760 mm diameter and then coupled into a 10 m length optical fiber with a core diameter of 1.5 mm and a numerical aperture of 0.37. Achromatic double cemented lenses are used to collimate the light output from the optical fiber. After the bandpass filtered through the interference filter, the photons are focused onto the target surface of the photon multiplier by a convex lens. After photoelectric conversion, the electrical digital signals are recorded by a photon-counting card inside a computer.
As mentioned in Section 2, to improve SNR, the system needs to eliminate the lower atmosphere backscattering signals, while to ensure the signals between 30 km and 120 km are fully received by the telescope. Since overlap Functions have been discussed in detail by She et al. (2011), we will focus on system parameters and the possibility of a more compact system.
After expanding the beam, the laser beam divergence angle is ∼0.38mrad, and the beam size on the transmitter mirror is ∼8.5 mm. The FOV of the receiving telescope is equal to $d/f$ (∼0.82mrad), with d being core diameter of the fiber (1.5 mm) and f being the focal length of the telescope (183 cm). The distance between the laser beam and the telescope center is ∼9 m. The estimated overlap function suggests that the laser beam starts to enter the telescope FOV at ∼14 km, fully overlaps with the telescope at ∼25 km. There is one possibility to reduce the distance between the laser beam and the receiving telescope, which is to reduce the beam divergence and telescope FOV. We could further expand beam size and reduce the divergence to ∼0.2mrad, and decrease the fiber core size to reduce the FOV of the telescope to ∼0.3mrad. The estimated overlap function suggests that a 4 m separation between the beam and the telescope center could effectively block the backscattering photons below 10 km. In the future, we will try to build a smaller biaxial structure, but in this system, we adopt 9 m separation setup. Another possible solution is to deploy the system on a high-altitude ground observatory. Compared to low-altitude area, the air in high-altitude ground is thinner, which is good for us to reduce strong near-ground signal. However, the high-altitude ground will also create new challenges to the PMCW system.
3.3 Timing
Since the range information of PMCW lidar is obtained through M-code modulation of a laser beam and then decoding the received signal, the timings of laser emission and signal receiving systems must be strictly synchronized to avoid the height deviation in the decoded signal. Figure 3 shows the timing diagram of our PMCW lidar. The waveform generator generates an M-code TTL signal for EOM modulation of the CW laser beam and a synchronization start signal for triggering photon-counting cards. The minimum gate time of the counters corresponds to the time of a single code ‘0’ or ‘1’ in M-code, i.e., 7.0µs. Within one trigger cycle of 10.0 ms, there are a total of 9 periods of 127 M-code with each 0.889 ms, 2 periods of zero-set M-code, and a short empty time. The number of collected photons will be recorded in one-to-one correspondence with the serial number of the M-code. For the first period of 127 M-code, the laser beam does not reach the maximum detection distance and returns to the telescope, as such the first M-code time will be discarded in the calculation. The subsequent 8 periods of 127 M-code are used for retrieval calculation (the blue part in Fig. 3). Similarly, the sky background signal is collected during the second period of zero-set M-code to avoid detecting the laser light that is still propagating in space.
4. Results and discussions
4.1 Raw signal profiles
The backscattering signals received by the telescope are binned for every single code (7.0µs, ∼ 1.05 km) and integrated for 5 min, and then stored as raw photon count profiles. Figure 4(a) shows a typical preliminary photon signal with a 127 M-code at 23:07-23:12 local time (LT) on December 3, 2019, and Fig. 4(b) shows its corresponding vertical profiles decoded from raw photon counts. The maximum measurement range is 133.3 km under 7.0µs single code time and 127 M codes. The vertical profile of the retrieved signal shows that the backscattering signal is successfully eliminated below 10 km (the blind area of the receiving telescope) and completely received by the telescope above 25 km. The exponential decay of the signal between 20 and 40 km is due to Rayleigh and/or aerosol backscattering. We see sodium backscattering signals between 80 and 100 km with a peak of ∼105 K photons at 91 km. There are two periods of zero-set M-code in the timing sequence to collect sky background signals in real-time. In the calculation of the preliminary signal, raw photon counts in each group are subtracted from the average value of the second period of zero-set M-code signals to remove the background. Figure 4(c) shows more details of the decoded signal between 80 and 100 km that recalculated after subtracting the background. Figure 4(d) shows the signal-to-noise ratio at most of the sodium layer (∼ 83-97 km) is equal to or greater than 10.
Fig. 4. The (a) raw signal arranged in m-code sequence with 1.05 km vertical and 5 min temporal resolutions on December 3, 2019, (b) sodium signal results with background subtracted, (c) the error ratio of the sodium signal, and (d) the signal to noise ratio.
4.2 Sodium density profile
Figure 5 shows the comparison of sodium density profile retrieved from PMCW lidar with that simultaneously observed by a collocated broadband sodium lidar [30]. Long-term observations have shown systematic differences between sodium densities derived from broad-band and narrow-band sodium lidars [31]. The broad-band sodium densities shown in Fig. 5 have been corrected. The two profiles are in very good agreement. We estimate that the statistic uncertainty of the typical measurement is less than 10% at the sodium peak (e.g., 90 km) for our PMCW lidar with 1.2W laser power, 1.05 km vertical, and 5 min temporal resolutions, but increases rapidly from 10% to 30% at the sodium layer edges (e.g., 82 km and 97 km) due to lower sodium density.
Fig. 5. Comparison of one profile with (a) sodium density and (b) its uncertainty observed by the PMCW sodium lidar (black) with 1.05 km vertical and 5 min temporal resolutions and the broadband sodium lidar [30] (blue) with 1.0 km vertical and 4 min temporal resolutions on December 3, 2019.
Figure 6 shows the contours of sodium density observed by PMCW and pulse lidars on December 3-4, 2019. Both observations agree very well in spatial-temporal distributions. Both observations on December 3 show multiple small perturbations at 85-90 km likely due to gravity waves. On the night of December 4, 2019, both lidars observed a significant density increase around 24:00 local time. Both observed peaks of sodium density were more than 7,200 cm-3, with an error of less than 10%. We also simulate the uncertainty of a 20W power PMCW lidar and compare it with pulsed lidar in Fig. 5(b). Limited by different detection techniques, the uncertainty is not as small as pulsed lidar, but it is accurate enough to carry out the study. It can be concluded that the sodium densities observed by the PMCW lidar are reliable.
Fig. 6. The contour map of sodium density observed by PMCW lidar system with 1.05 km vertical and 5 min temporal resolutions on December 3, 2019 (a) and December 4, 2019 (b); The contour map of sodium density observed by the broadband sodium lidar of Meridian Project with 1.0 km vertical and 4 min temporal resolutions on December 3, 2019 (c) and December 4, 2019 (d).
4.3 Error analysis
The previous study has demonstrated that the SNR of PMCW lidar depends on the sum of all signals received in a full-length coding cycle [20]. Therefore, PMCW lidar must take advantage of the blind zone outside the receiving telescope FOV to minimize the strong Mie and Rayleigh backscattering signals near the ground. We tested 3 different distances (2.0, 4.5, and 9.0 m) between the emitted laser beam and the center of the receiving telescope to find out the best setup. Figure 7 shows the testing signals for three different biaxis setups. To reduce statistic uncertainty, the signals were integrated for 1 h. We see strong anomalous noise peaks near 30 km and 60-80 km (marked as red) for 2.0 and 4.5 m setups, but not clear for 9.0 m setup. The noise signals at 60-80 km decrease 5 and 20 times when the distance between the laser beam and telescope axis increases from 2.0 to 4.5 and 9.0 m, suggesting that this noise signal is likely induced by strong near-ground multiple scattering. Meanwhile, the backgrounds in both 4.5 and 9.0 m setup are significantly dropped to 1% of the sodium peak signal. For our system, we finally select a 9.0 m setup to minimize this anomalous noise signal.
Fig. 7. Three typical detection results in some containing anomalous noise, with the distance between the transmitter and the telescope is 2.0 meters (a), 4.5 meters (b), and 9.0 meters (c) respectively.
In order to verify the existence of the real targets at near 30 km and 60-80 km altitude, we cut the CW beam into 20µs wide pulse for detection to check the strong anomalous noise peaks but there was no similar anomaly. Ardanuy and Cameron (2017) observed clouds with the output of the cross-correlation process and show some peaks where nothing exists and called these peaks “ghost” targets [21]. They attribute this phenomenon to the effect of nonlinearity and offset on the correlation. However, we found in the experiment that EOM modulation error would also lead to generating “ghost” targets. We detected the laser light intensity after EOM modulation in the case of 127 code length and a single code time of 7.0µs and show the optical signal results in Fig. 8. Due to the rapid voltage change in the modulation process, the signal generator circuit and the voltage amplifier circuit will induce a voltage oscillation; on the other hand, the KD*P crystal used for modulation in EOM has a resonance frequency, which also produces the oscillation of light intensity in time. Both combined effects are the cause of this significant noise. The sum of modulation errors within a single coding time is within 3% of the desired result. The error is mainly caused by the oscillation after the step change of the control voltage, so the position of the error is regular relative to the code. The correlation coefficient of M-code and modulation error is ∼-0.8. This correlation coefficient indicates that the EOM modulation error in the convolution signal will form special peaks in the final inversion of the results. The size of these special peaks is related to the sum of received signals, and the more obvious peaks only appear at a few distinct altitudes. The modulation function $\alpha (t )$ are replaced by the optical signal results in Eq. (2), and the maximum error in simulation can reach 3% of the total signal, and this error may occur at any position including the altitude range without obvious return signals. In addition, we also tested the influence of different M-code lengths and code times on “ghost” targets with 2.0 m separation setup. The different M-code length cannot reduce the occurrence of ghost peaks, and the study of code time shows that the position of “ghost” targets is related to the specific number i of the code $a(i )$ ($i$ and $a(i )$ can be found in Table 1) rather than the signal altitude. However, the study of “ghost” does not yield a definitive result, it will be the focus of our future work. Although this error is hard to eliminate, we can significantly reduce it by following two ways: one is to keep the time of a single code away from the resonant frequency of the crystal during the design process, and the other is to suppress the strong signal near the ground. The effect of “ghost” on sodium signals can be effectively reduced by reducing the near-earth strong signal.
Fig. 8. EOM modulation optical signal results (a) below 100 mW energy and the modulation error (b) between the actual result and the target result.
5. Summary
We successfully developed a PMCW lidar for the observation of mesopause region sodium density. A high-frequency EOM system is implemented into the laser transmitter system to achieve CW beam coding. The laser was modulated by a pseudorandom M-code of length 127, with the minimum code time is 7.0µs which corresponds to 1.05 km range resolution. By increasing the distance between the center of the receiving telescope and the transmitting beam, the near-ground backscatter signals are significantly suppressed via the blind area of the receiving telescope FOV. The vertical profile of photon counts is converted from the received raw photon counts via specific calculation. After the background is removed, the sodium density can be calculated. This system has a spatial and temporal resolution of 1.05 kilometers and 5 min respectively. Compared with the pulse lidar, the PMCW lidar does not require dye pumps and the transmitting system is much simpler and more stable.
We carried out 3 nights of observations with this PMCW sodium lidar system and compared with simultaneous observations of a collocated broadband pulse sodium lidar. There are fairly good agreements between PMCW lidar sodium density and pulse lidar sodium density. We also describe the error factors of the PMCW lidar system. This prototype system demonstrates that a sodium PMCW lidar with high output power could achieve the measurements of mesopause region sodium density, even temperature, and wind with similar precision and resolutions as a pulse system. In the future work, we plan to upgrade the lidar system, use a more powerful continuous-wave laser source, improve the measurement accuracy, and achieve the measurement of wind speed and temperature.
Funding
National Natural Science Foundation of China (41974175, 41974177); The Strategic Priority Research Program of Chinese Academy of Sciences (XDB41000000); The Open Research Project of Large Research Infrastructures of CAS—“Study on the interaction between low/mid‐latitude atmosphere and ionosphere based on the Chinese Meridian Project.”.
Acknowledgments
We acknowledge the narrowband sodium temperature/wind lidar data resources from the National Space Science Data Center, National Science & Technology Infrastructure of China (http://www.nssdc.ac.cn/eng). The sodium density data from the broadband sodium lidar at Hefei is supported by the Chinese Meridian Project (https://data.meridianproject.ac.cn/stations/29/). The authors would like to thank Chiao-Yao She for helpful discussion.
Disclosures
The authors declare no conflicts of interest.
References
1. R. M. Measures, “Laser remote sensing fundamentals and applications,” (1984).
2. A. V. Jelalian, “Laser radar systems,” in EASCON Record: IEEE Electronics and Aerospace Systems Convention (1980).
3. R. S. Lindzen, “Turbulence and stress owing to gravity wave and tidal breakdown,” J. Geophys. Res. 86(C10), 9707–9714 (1981). [CrossRef]
4. R. Bernard, “Tides and planetary waves in the lower thermosphere and the upper mesosphere,” Atmospheric Physics from Spacelab, Eleventh Eslab Symposium Atmospheric Physics from Spacelab, Eleventh ESLAB Symposium, 1976.
5. C. Y. She, J. R. Yu, H. Latifi, and R. E. Bills, “High-spectral-resolution fluorescence light detection and ranging for mesospheric sodium temperature measurements,” Appl. Opt. 31(12), 2095–2106 (1992). [CrossRef]
6. C. Y. She, H. Latifi, J. R. Yu, R. J. Alvarez, R. E. Bills, and C. S. Gardner, “Two-frequency Lidar technique for mesospheric Na temperature measurements,” Geophys. Res. Lett. 17(7), 929–932 (1990). [CrossRef]
7. R. E. Bills, C. S. Gardner, and S. J. Franke, “Na Doppler/temperature lidar: initial mesopause region observations and comparison with the Urbana medium frequency radar,” J. Geophys. Res. 96(D12), 22701–22707 (1991). [CrossRef]
8. T. Li, X. Fang, W. Liu, S. Y. Gu, and X. Dou, “Narrowband sodium lidar for the measurements of mesopause region temperature and wind,” Appl. Opt. 51(22), 5401–5411 (2012). [CrossRef]
9. Y. Yang, Y. Yang, Y. Xia, X. Lin, L. Zhang, H. Jiang, X. Cheng, L. Liu, K. Ji, and F. Li, “Solid-state 589 nm seed laser based on Raman fiber amplifier for sodium wind/temperature lidar in Tibet, China,” Opt. Express 26(13), 16226–16235 (2018). [CrossRef]
10. H. Moosmüller and J. D. Vance, “Sum-frequency generation of continuous-wave sodium D_2 resonance radiation,” Opt. Lett. 22(15), 1135–1137 (1997). [CrossRef]
11. J. D. Vance, C.-Y. She, and H. Moosmüller, “Continuous-wave, all-solid-state, single-frequency 400-mW source at 589 nm based on doubly resonant sum-frequency mixing in a monolithic lithium niobate resonator,” Appl. Opt. 37(21), 4891–4896 (1998). [CrossRef]
12. C. Y. She, J. D. Vanee, T. D. Kawahara, B. P. Williams, and Q. Wu, “A proposed all-solid-state transportable narrow-band sodium lidar for mesopause region temperature and horizontal wind measurements,” Can. J. Phys. 85(2), 111–118 (2007). [CrossRef]
13. J. Yue, C.-Y. She, B. P. Williams, J. D. Vance, P. E. Acott, and T. D. Kawahara, “Continuous-wave sodium D_2 resonance radiation generated in single-pass sum-frequency generation with periodically poled lithium niobate,” Opt. Lett. 34(7), 1093–1095 (2009). [CrossRef]
14. T. D. Kawahara, S. Nozawa, N. Saito, T. Kawabata, T. T. Tsuda, and S. Wada, “Sodium temperature/wind lidar based on laser-diode-pumped Nd:YAG lasers deployed at Tromsø, Norway (696°N, 192°E),” Opt. Express 25(12), A491–A501 (2017). [CrossRef]
15. Y. Xia, L. Du, X. Cheng, F. Li, J. Wang, Z. Wang, Y. Yang, X. Lin, Y. Xun, S. Gong, and G. Yang, “Development of a solid-state sodium Doppler lidar using an all-fiber-coupled injection seeding unit for simultaneous temperature and wind measurements in the mesopause region,” Opt. Express 25(5), 5264–5278 (2017). [CrossRef]
16. L. Ridenour, G. N. Glasoe, J. V. Lebacqz, G. B. Collins, I. A. Greenwood, J. V. Holdam, D. Macrae, and J. R. Pierce, “MIT Radiation Laboratory Series,” Phys. Today (1948).
17. R. G. Strauch, W. C. Campbell, R. B. Chadwick, and K. P. Moran, “Microwave FM-CW Doppler radar for boundary layer probing,” Geophys. Res. Lett. 3(3), 193–196 (1976). [CrossRef]
18. R. B. Chadwick, K. P. Moran, R. G. Strauch, G. E. Morrison, and W. C. Campbell, “Microwave radar wind measurements in the clear air,” Radio Sci. 11(10), 795–802 (1976). [CrossRef]
19. N. Takeuchi, H. Baba, K. Sakurai, and T. Ueno, “Diode-laser random-modulation cw lidar,” Appl. Opt. 25(1), 63–67 (1986). [CrossRef]
20. R. Matthey and V. Mitev, “Pseudo-random noise-continuous-wave laser radar for surface and cloud measurements,” Opt. Lasers Eng. 43(3-5), 557–571 (2005). [CrossRef]
21. A. Ardanuy and A. Comerón, “Compact lidar system using laser diode, binary continuous wave power modulation, and an avalanche photodiode-based receiver controlled by a digital signal processor,” Opt. Eng. 57(04), 1 (2018). [CrossRef]
22. M. Quatrevalet, X. Ai, A. Perez-Serrano, P. Adamiec, J. Barbero, A. Fix, J. M. G. Tijero, I. Esquivias, J. G. Rarity, and G. Ehret, “Atmospheric CO-2 Sensing with a Random Modulation Continuous Wave Integrated Path Differential Absorption Lidar,” IEEE J. Sel. Top. Quantum Electron. 23(2), 157–167 (2017). [CrossRef]
23. M. Abo and C. Nagasawa, “Random modulation CW dye lidar for measuring mesospheric sodium layer,” presented at the 17th International Laser Radar Conference, Sendai, Japan, 25–29 July 1984, paper 26PA14, p. 258.
24. C. Y. She, M. Abo, J. Yue, B. P. Williams, C. Nagasawa, and T. Nakamura, “Mesopause-region temperature and wind measurements with pseudorandom modulation continuous-wave (PMCW) lidar at 589 nm,” Appl. Opt. 51(12), 1981 (2012). [CrossRef]
25. J. P. Sherman, “Mesopause Region Thermal and Dynamical Studies Based on Simultaneous Temperature, Zonal and Meridional Wind Measurements with an Upgraded Sodium Fluorescence Lidar,” Ph.D. Dissertation (Colorado State University, 2002).
26. N. Takeuchi, N. Sugimoto, H. Baba, and K. Sakurai, “Random modulation cw lidar,” Appl. Opt. 22(9), 1382–1386 (1983). [CrossRef]
27. C. Nagasawa, M. Abo, H. Yamamoto, and O. Uchino, “Random modulation cw lidar using new random sequence,” Appl. Opt. 29(10), 1466–1470 (1990). [CrossRef]
28. P. V. Lenzo, E. G. Spencer, and K. Nassau, “Electro-Optic Coefficients in Single-Domain Ferroelectric Lithium Niobate,” J. Opt. Soc. Am. 56(5), 633–635 (1966). [CrossRef]
29. X. Fang, T. Li, C. Ban, Z. Wu, J. Li, F. Li, Y. Cen, and B. Tian, “A mobile differential absorption lidar for simultaneous observations of tropospheric and stratospheric ozone over Tibet,” Opt. Express 27(4), 4126–4139 (2019). [CrossRef]
30. J. Ma, M. Shangguan, H. Xia, X. Fang, X. Xue, and X. Dou, “Rayleigh and sodium lidar system incorporating time-division and wavelength-division multiplexing,” Opt. Commun. 448, 116–123 (2019). [CrossRef]
31. T. Li, C. Ban, X. Fang, J. Li, Z. Wu, W. Feng, J. M. C. Plane, J. Xiong, D. R. Marsh, M. J. Mills, and X. Dou, “Climatology of mesopause region nocturnal temperature, zonal wind and sodium density observed by sodium lidar over Hefei, China (32°N, 117°E),” Atmos. Chem. Phys. 18(16), 11683–11695 (2018). [CrossRef]
