Aerosols play a pivotal role in natural systems, impacting air pollution^[1], cloud nucleation^[2], precipitation distribution^[3], and variations^[4] by modulating solar radiation^[5,6]. Thorough insights into aerosol transport dynamics can greatly advance the design and implementation of effective air pollution mitigation measures^[7]. Thus, obtaining synchronized datasets on aerosols and wind is crucial. Rigorous observation and analysis of these factors are essential for tracking climate patterns and understanding their implications for both environmental and societal well-being^[8,9].

Relative to traditional lidar, coherent lidar (CL) offers distinct benefits, including a more compact design, lower power consumption, and increased adaptability in detection modalities. As a result, coherent Doppler wind lidar technology has become the predominant technique for wind field detection^[10–14] and has the ability to meet the requirements of high-precision measurements. Prolific research and subsequent advancements underscore the vital and efficacious role of CL in diverse fields, encompassing atmospheric studies, wind energy sectors, and aviation safety^[15–19]. Nonetheless, it is imperative to recognize that most contemporary CL systems predominantly utilize Doppler frequency shift data derived from the echo signal spectrum to determine wind velocities^[20–24], or integrate additional devices to capture meteorological parameters^[25–27].

Several studies have demonstrated the correlation between CLs and aerosols or boundary layer height. Menzies and Tratt^[28] (1994) utilized a CO2 airborne CL (near 9 µm) to obtain the vertical distribution of backscatter in the troposphere and lower stratosphere after hard target calibration. Chouza et al.^[29] (2015) employed a 2 µm airborne Doppler wind lidar and calibrated a ground-based 532 nm direct aerosol lidar (AL) using a sun photometer for comparison, with error results within 20% for the backscatter and extinction profiles. Abdelazim et al.^[30] (2015) used a 1.5 µm all-fiber coherent Doppler lidar, with the carrier-to-noise ratio (CNR) as a distance-squared correction signal, to obtain aerosol backscatter coefficient and wind speed. Dong et al.^[31] (2018) simulated the process of aerosol particle backscattering of a horizontally polarized laser using a semi-analytical Monte Carlo method, analyzing the characteristics of laser backscatter under various meteorological conditions and providing reference for CL systems. Wang et al.^[32] (2019) combined the direct detection lidar for PM2.5 with the coherent Doppler wind lidar, both at 1.5 µm, and used the CNR along with vertical wind speed to determine the boundary layer height. It indicates that CL can acquire relevant aerosol information through calibration with other instruments. However, further comparative analysis is still needed to obtain more information about measured variables beyond aerosols under various weather conditions.

Departing from traditional methodologies, this research aims to highlight the potential of compact CL technology for detecting a range of meteorological parameters, encompassing cloud height, aerosol extinction coefficient, and aerosol concentration, coupled with wind field evaluation. Especially, experimental comparisons of aerosols were conducted under multiple meteorological conditions. In light of this, the study introduces a comprehensive data-processing approach to extract these elements, leveraging a compact CL with a 1.55 µm operational wavelength.

Figure 1 illustrates the layout of our CL system. The 1550 nm seed laser with a linewidth of 10 kHz is split into two beams using a beam splitter. One beam enters the local oscillator, while the other is amplified by an erbium-doped fiber amplifier (EDFA), controlled to achieve a pulse energy of 150 µJ after passing through an acousto-optic modulator (AOM) and modulated by the control system with a radio frequency (RF) tone before being transmitted into the atmosphere. The atmospheric echo signal is received by a 100 mm telescope, part of a scanning system with an integrated transmitter and receiver controlled to avoid the blind zone. The aerosol scattering and local oscillator signals are then converted into a heterodyne electrical signal through coherent homodyne mixing, followed by signal processing.

The intermediate frequency current signals, emanating from the balanced detector, are generated by mixing echo signals with local oscillator signals, IRF=2ℜPLPScos(2πωdt+Δφ),(1)where PL is the power of the local oscillator signal, PS is the power of the echo signal, R is the detector responsivity, ωd is the beat frequency, and Δφ is the phase difference between the local oscillator signal and the echo signal.

The current signal is converted to a voltage signal through impedance amplification in the detector. By applying a fast Fourier transform (FFT), it is possible to obtain the Doppler frequency shift of the aerosol, which can then be used to calculate the wind speed. Concurrently, the signal is processed to yield a sampling rate of SR and a sample size of N from the detector, and this power is subsequently utilized for analyzing the spectral energy. The CL ultimately outputs a frequency-domain signal derived from time-frequency conversion. According to Parseval’s theorem^[33], the signal’s energy remains constant during this conversion. Consequently, the power of the echo signal PS is PS=SRRL2PLℜ2G2NEs=SRRL2PLR2G2NEf,(2)where RL is the impedance of the signal acquisition and processing board, G is the amplification ratio of the balanced detector, Es is the energy of the acquired signal, and Ef is the signal energy from the spectral signal.

To accurately obtain the aforementioned parameters, it is crucial to make distance corrections to PS: S(r)=ln(PSIz),(3)where S(r) is the correction of PS and Iz is the correction factor related to the range r and the correction coefficient Δz, which is given by Iz=Δz(r−Δz)2+Δz2,(4)where Δz depends on the focal length F, telescope radius at, and laser wavelength λ, as described by^[34]{Δz=F2zrF2+zr2zr=2πλ(Fat)2.(5)

The range corrected signal S(r) is utilized for subsequent inversion calculations to determine the aerosol extinction coefficient, then the cloud height, visibility, and aerosol concentration are all converted by the extinction coefficient. The parameters of the CL are shown in Table 1.

Since the accuracy of estimation of the frequency shift depends on the echo signal and noise power, the signal-to-noise ratio (SNR)^[35] is an important characteristic in terms of the possibility of sensing the atmosphere using a CL: SNR=ηEchvBT2(r)β(r)Sc(r)r2,(6)where η is the quantum efficiency of the detector, E is the pulse energy, c is the speed of light, h is the Plank’s constant, υ is the laser’s frequency, B is the detector’s bandwidth, T is the atmospheric transmittance, β is the backscatter coefficient of aerosol particles, and Sc(r) is the area of the coherent plate, which are related to the coherence radius of the backscattered wave in the telescope plane^[36]. Consequently, the SNR is influenced by several parameters, including aerosol characteristics, coherence radius of the echo signal, detector quantum efficiency, and wavelength. Among them, the strength of refractive turbulence represents the atmospheric inhomogeneity, which compromises the spatial coherence of the scattered waves in the detection beam and the telescope plane, leading to the deterioration of the SNR. Under standard atmospheric conditions, the SNR at the distance of 1 km is approximately 11 to 26 dB (affected by turbulence intensity).

The operational principle and algorithmic procedure of the CL, applied to the detection of wind field, cloud height, visibility, and aerosol concentration, are illustrated in Fig. 2. Eight steps for extracting the mentioned key observable associated with winds and aerosols are as follows:

1. (a)In order to calculate the power of the signal PS and the SNR, it is necessary to convert the signal from the time domain to the frequency domain. This includes performing a Fourier transform on the signal, which breaks it down into its frequency components. By analyzing the frequency domain information, it is possible to determine the frequency shift, total energy of the signal, and quality of the signal.
2. (b)The wind speed and direction are calculated using the zenith angle and azimuth angle of the lidar, which provides valuable information on the atmospheric flow field.
3. (c)The distance correction of PS yields the signal S(r), which is then calibrated using a linear regression equation through piecewise linear fitting.
4. (d)The maximum coefficient of correlation in the linear regression equation is then used to calculate the extinction coefficient σ0 and position r0 (reference distance) using the Collis slope method.
5. (e)The extinction coefficient is calculated using the Klett method^[37], which involves performing backward integration for r−r0 and forward integration for r0−r, providing an accurate estimate of the extinction coefficient.
6. (f)For cloud height measurement, a comprehensive approach is employed, involving the threshold method, differentiation method, and aerosol extinction coefficient method to determine the cloud base height from the echo signal.
7. (g)Visibility measurement involves converting the extinction coefficient to visibility and using the wavelength corrected q for the corresponding wavelength^[38].
8. (h)Finally, aerosol concentration measurement is achieved using an exponential model^[39], where calibration parameters, a,b, and c, are obtained by measuring multiple sets of aerosol concentration values using standard equipment at the same location.

Between March and April 2022, a side-by-side experiment was executed in Chengdu, China. The study employed CL with an operational wavelength of 1550 nm and a dual-polarization aerosol detection lidar set to a 532 nm wavelength. The evaluation spanned multiple meteorological scenarios, encompassing sunny, cloudy, rainy, and foggy conditions. The experimental setup is illustrated in Fig. 3.

The systems of the two different devices were compared in Table 2, revealing significant differences between the two lidar types. The CL is different from the AL in the principle of heterodyne coherent detection. The CL, using a 1550 nm wavelength, is capable of measuring the wind field while also capturing extinction coefficients, cloud height, and visibility. On the other hand, the AL is a direct detection lidar based on the Mie scattering principle, using a 532 nm wavelength to measure aerosol related parameters. It can measure extinction coefficients, visibility, PM10 and PM2.5 aerosol concentrations, and depolarization ratio.

Compared to the AL, the CL presents superior applicative advantages, including enhanced detection efficiency, a streamlined structure, and a more comprehensive range of data products. This technology can concurrently measure cloud height, extinction coefficient, visibility, aerosol concentration, and wind field. Employing a singular apparatus diminishes the intricacy of the detection system, bolsters system stability, and augments adaptability for mobile detection tasks, specifically in domains such as meteorological support and environmental pollution monitoring.

The aerosol concentration was measured under various meteorological conditions, including sunny, cloudy, rainy, and foggy days. An analysis of aerosol extinction coefficients under these distinct weather conditions is presented in Figs. 4 and 5, and is compared in Table 3.

Heavy haze resulted in reduced visibility, with the measurements from both instruments closely aligned, as demonstrated in Figs. 4 and 5(a), which present a comparison of detected aerosol extinction coefficient and SNR. During daytime, the haze induced reduced visibility and SNR, with the cloud layer observed at a relatively low altitude.

On gloomy and rainy days, the aerosol extinction coefficient and SNR detected by both instruments aligned with the atmospheric conditions, as depicted in Figs. 4 and 5(b). Both devices registered analogous trends in their measurements, with the identification of cloud layers aligning consistently. Throughout the day, from dawn to dusk, the atmospheric structure in the lower troposphere remained relatively unchanged, maintaining consistent aerosol concentrations and SNRs. However, during precipitation events, data from the CL were absent due to rain-induced interference. In contrast, the AL continued to yield data, although its accuracy was compromised by significant errors stemming from the rainfall, thus reducing its reliability.

The extremely low visibility weather conditions such as heavy fog significantly constrained the detection range of two lidar systems tasked with monitoring the aerosol extinction coefficient and SNR, as delineated in Figs. 4 and 5(c). Under conditions of dense fog characterized by remarkably low visibility, the detection range for both systems was curtailed, often restricted to less than 1 km. The data gap in the AL between 0:00 and 9:00 could be attributed to geometric factors. Moreover, the SNR value for the AL at altitudes below 6 km is notably high. The inconsistency in SNR could stem from unreliable AL measurements. Geometric constraints and high aerosol concentrations might cause AL saturation, leading to data loss and yielding incorrect SNR data.

Under most weather conditions, the results of the aerosol extinction coefficient and SNR by the two lidars are generally consistent. However, during certain weather conditions such as a rainy day and heavy fog, some factors such as laser wavelengths, polarization characteristics, and geometric factors can lead to varying results in the data products of the CL and AL.

In October 2022, a series of experiments were conducted based on the CL in Chengdu to collect comprehensive data products. The experiment yielded results for extinction coefficients, cloud height, wind field, and SNR data over a continuous period of 48 h, as illustrated in Fig. 6.

Figure 6(a) demonstrates a strong correlation among the aerosol extinction coefficient, cloud height, and wind field with the SNR depicted in Fig. 6(b). This correlation implies the accuracy and reliability of the measurements. On November 11, 2022, from noon to evening, a significant presence of aerosols was detected in the atmosphere, predominantly concentrated around a height of 1.5 km. These high concentrations of aerosols facilitated the formation of cloud clusters, as indicated by elevated extinction coefficient values. During this period, the wind remained relatively calm, with speeds ranging from 1 to 2 m/s. In the evening, at an altitude of 2 km, wind speeds increased to 6–8 m/s, concurrent with cloud formation at the same altitude. These stronger winds caused the clouds to move more rapidly, resulting in a shorter duration of the high aerosol concentration area. From the 12th, low visibility and wind speeds of approximately 1–2 m/s contributed to the accumulation of aerosols, leading to high aerosol levels at a distance of 1 km. In the afternoon, there was a low SNR due to the rain, and then cloud clusters formed at altitudes of 1.5 km and 500 m, persisting until midnight. From 00:00 on the 13th, the wind speed escalated to 8 m/s, while clusters of clouds emerged at an altitude of 1.5 km.

Simultaneous measurement of meteorological factors such as wind, aerosols, and clouds enables more precise atmospheric monitoring. This approach provides accurate insights into the formation and dissipation of aerosol clouds and supports meteorological disaster warnings. Additionally, it serves as a basis for environmental preservation and pollution control measures.

In this investigation, the newly developed CL facilitates the determination of diverse meteorological parameters, including cloud height, extinction coefficient, visibility, and aerosol concentration, while concurrently assessing the wind field. This study examines the characteristics of the CL signal echoes, establishing a technique for calculating the echo signal power in the time domain. The Klett algorithm is employed to determine the atmospheric extinction coefficient, after which methods involving differential thresholds and aerosol extinction coefficients are used to derive cloud heights. Additionally, atmospheric visibility and aerosol concentration are inferred based on their relationship with the extinction coefficient.

The developed CL consistently demonstrates its capability to measure a range of meteorological parameters, such as horizontal wind speed and direction, cloud height, visibility, extinction coefficient, and aerosol concentration under diverse weather conditions including sunny, cloudy, rainy, and foggy scenarios. Integrating this CL for comprehensive meteorological observations offers benefits in terms of simplicity, compactness, and system robustness while also enhancing mobile detection capabilities. The amalgamation of these meteorological parameters results in a rich set of secondary data products, notably aerosol flux products. Beyond merely tracking pollutant locations and concentrations, the CL provides tools for pinpointing pollutant origins, gauging dispersion velocities and trajectories, and forecasting pollutant movement patterns, thereby aiding in the implementation of effective interventions.