Open Access
Add to BibSonomy
Abstract
Vertical distributions of ambient aerosols and their corresponding optical properties are crucial to the assessment of aerosol radiative effects. Traditionally, ambient aerosol phase function is assumed as a constant of input parameter in the retrieval of the vertical distribution of aerosol optical characteristics from remote sensing measurements (e.g. lidar or camera-laser based instruments). In this work, sensitivity studies revealed that using constant aerosol phase function assumptions in the algorithm would cause large uncertainties. Therefore, an improved retrieval method was established to simultaneously measure ambient aerosol scattering phase functions and aerosol scattering function profiles with a modified charge-coupled device-laser aerosol detection system (CLADS), which are then combined to yield vertical profiles of aerosol extinction coefficients. This method was applied and evaluated in a comprehensive field campaign in the North China Plain during January 2016. The algorithm showed robust performance and was able to capture temporal variations in ambient aerosol scattering phase functions and aerosol scattering function profiles. Aerosol extinction coefficients derived with simultaneously measured aerosol phase functions agreed well with in-situ measurements, indicating that uncertainties in the retrieval of aerosol extinction vertical profiles have been significantly reduced by using the proposed method with the modified CLADS. The advantage of this modified CLADS is that it can accomplish these aerosol measurements independent of other supplementary instruments. Benefiting from its low cost and high spatial resolution (∼1 m on average) in the boundary layer, this measurement system can play an important role in the research of aerosol vertical distributions and its impacts on environmental and climatic studies.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
The effect of aerosols on cloud albedo, also called the aerosol indirect climate effect, is one of the largest sources of uncertainties in the estimation of radiative forcing impact on global climate [1,2]. Over the past years, China, especially eastern China, has experienced rapid economic development, accompanied by increased aerosol emissions [3]. To alleviate the heavy air pollution, the Chinese government has already introduced many policies to reduce the aerosol loading, which successfully decreased the PM level in Beijing. Still, particulate pollution prevails in large parts of eastern China. Under the implementation of different regulations, particulate pollution is changing both in amounts and composition, which presents an urgent need to follow up on the changes in aerosol optical properties, and potential changes in their radiative forcing effects [4]. Many instruments were developed to measure in-situ aerosol optical properties, such as nephelometer or aethalometer [5,6]. However, ground observations alone are insufficient in the assessment of aerosol climate effects, wherein vertical distributions of aerosols and their optical properties are highly needed [7].
There are several ways to measure the vertical distribution of aerosols, including aircraft measurement, balloon and remote sensing. In-situ instruments can be deployed on aircrafts or balloons to measure the aerosol properties vertically [8]. Considering the expensive costs of both manpower and resources, remote sensing instruments are seen as the main method to detect the vertical profiles of aerosols on a continuous basis [9]. Lidar is the most commonly used remote sensing instrument to measure the vertical profiles of aerosols [10–12]. Nevertheless, due to the geometric structure of lidar, an incomplete overlap zone exists near the surface, in which aerosol properties cannot be characterized [13]. This zone (usually 200-500m high) severely hinders the comparison and combination of lidar and in-situ ground measurements [14,15].
To solve this problem, a novel technique has been used to measure the aerosol optical properties in the near surface layer. In this technique, charged couple device (CCD) is used as detector to capture the backscattering signals of laser beam by imaging principle to avoid the incomplete overlap zone [16]. Meki et al. first applied the CCD detector with laser emitter for qualitative atmospheric measurements [17]. Barnes et al. developed the CCD camera–based lidar using wide-angle optics instead of a scanning system to cover the entire altitude range [18,19]. After that, several instruments based on similar techniques were developed to measure the vertical distribution of aerosol concentration and aerosol optical depth [20,21]. The Scheimpflug lidar with CCD sensors as detectors based on the Scheimpflug principle was also demonstrated to retrieve the aerosol extinction coefficient profiles [22–24]. Bian et al. built a CCD-laser atmospheric detection system to retrieve the nocturnal boundary layer structure [25]. The results showed that the fine characteristics and patterns of the nocturnal boundary layer structures could be captured. Wang et al. used measurements of CCD-laser instruments to correct the lidar overlap factor [26]. The inter-comparison of various experimental results with other instruments shown in these studies demonstrated the CCD–laser based method to be reliable.
One important input parameter in the retrieval of aerosol optical properties measured by CCD-laser based instruments is the aerosol scattering phase function, which describes the angular distribution of the aerosol scattering intensity [27]. Sharma et al. used different types of aerosol phase function in the retrieval method of the CCD-laser instrument, which resulted in significant changes in retrieved extinction coefficients [16]. In the sensitivity study made by Lian et al. [28], the aerosol phase function can lead up to 462% uncertainties within the retrieved profile from CCD-laser instrument. Direct measurements of aerosol phase function are still in lack. While aerosol phase function can be theoretically calculated using the Mie theory with the size and complex refractive index of particles under spherical shape assumption [29], Calculations are much more complicated for nonspherical particle. The Aurora 4000 polar nephelometer (Ecotech Pty Ltd., Australia) is currently the only commercial instrument that can measure the aerosol phase function, however, only in a scattering angle range of 0-90° for dry aerosols. Backscattering phase functions of ambient aerosols are needed in the retrieval method of CCD-laser instruments, which is a dilemma. In past years, several versions of polar nephelometers have been developed by individual research groups to measure the scattering phase function of aerosol particles, cloud droplets and ice crystals. Some instruments measured the angular distribution of scattering signals by using the rotational mechanism [30–32]. The main uncertainty of this design comes from the fact that signals were not measured simultaneously. Barkey et al. measured scattering signals with different scattering angles at the same time with many photomultiplier tubes (PMTs) mounted at different scattering angles [33]. Curtis et al. used an ellipsoidal mirror to reflect the scattering light to a CCD to detect the aerosol phase function [34]. By using CCD as detector, this method can offer better angular and temporal resolution at a wider range of scattering angles than the other methods above. Espinosa et al. built a polarized imaging nephelometer by deploying a CCD in the cavity of the nephelometer to measure the scattering signals at different angles [35]. Bian et al. developed a measurement system, which has the advantages of wider detection range and better stability, to measure the scattering phase function of ambient aerosols applying CCD-laser detection techniques [36].
In this paper, a modified CCD camera-laser aerosol detective system (CLADS) based on the camera imaging principle and the optical structure of the fisheye lens is developed to measure ambient scattering aerosol phase functions simultaneously with aerosol extinction coefficient vertical profiles. An improved retrieval method was setup to obtain the vertical profiles of aerosol extinction coefficients using simultaneously obtained phase functions, highly lowering afore-mentioned uncertainties of assumed phase functions, improving the accuracy of retrieved results.
2. Instrument and methodology
2.1 Modified CLADS setup
The modified CLADS includes several main components: the vertical emitting part, the horizontal emitting part and the related receiving parts. These components are mounted at the same height and with certain distances to each other. The geometric structure of the modified CLADS is shown in Fig. 1, while specific parameters of the components used in the modified CLADS are listed in Table 1.
Fig. 1. Sketch map of the geometric structure of modified CLADS: a. top view of the CLADS; b. side view of the CLADS. Dash line shows the field of view of CCD1. Dotted line shows the field of view of CCD2. Laser1 is pointed perpendicular to the paper outward.
Table 1. Instrument parameters of CLADS.
The (vertical/horizontal) emitting parts of the modified CLADS mainly consist of high-power continuous laser emitters. Neodymium-doped yttrium aluminum garnet (Nd: YAG) was selected as the solid laser material to reach a wavelength of 532 nm. To change the polarization state of the laser from linear to circular, a quarter-wave plate was mounted in front of the laser emitter. During the exposure time (few minutes) of the image, the circular-polarization light can be assumed to be unpolarised.
Both receiving parts in CCD-LADS consist of three main components: the CCD cameras, the fisheye lenses and the optical filters. The STF-8300 CCD imaging camera is SBIG's second-generation camera using the KAF-8300 CCD sensor (ON Semiconductor, Phoenix, AZ, USA). The area array (17.96*13.52 mm) of KAF-8300 CCD has 8.3 million (3326*2504) effective pixels, while each pixel is a square 5.4 µm on a side. The exposure time is from 0.1 s to 1 h. The quantum efficiency of the CCD is about 55% at 532 nm, while the linearity error is about 10%. The dark signal of KAF-8300 CCD is less than 200 e-/s.
The fisheye lens (Sigma Corp., Japan) has a maximum aperture of f/2.8 and a focus length of 10 mm. It features a 180° angle view as measured across the diagonal of the frame and provides high quality images from infinity to the closest focusing distance when this lens is used with the STF-8300 camera. The lens uses an equisolid projection, meaning that the solid angle of the object is directly proportional to the area on the CCD arrays [37]. The modulation transfer function of the lens shows that, according to the size of the CCD sensor, the difference of the sensitivities from the center to the corner is less than 5% (http://www.sigma-photo.co.jp/english/lens/wide/10_28/#/data). The angle resolution can reach 0.1° per pixel.
To filter out the background noise from the sky radiation, an optical filter (Thorlabs, Newton, NJ, USA) is mounted between the CCD camera and the lens. The central wavelength of this filter is 532 ± 0.2 nm, and the full width at half maximum is 1 ± 0.2 nm, while the maximum transmission at the peak is around 40%.
The laser emitters and CCD cameras are mounted on tripods and controlled by laptops. As can be seen in the sketch map of the geometric relationship in the modified CLADS (Fig. 1), Laser1 belongs to the vertical emitting part and its beam is emitted vertical to the ground. Laser2 belongs to the horizontal emitting part and is emitted horizontally with a beam trap at the end of the laser beam. Two CCD cameras (CCD1 and CCD2) with fisheye lenses are installed at the same height with the lasers to capture the scattering signal from the vertical and horizontal laser beam, respectively. According to the image formation principle of the fisheye lens, there is a one-to-one correspondence between the image of the laser beam captured by the CCDs and the laser beam object. The scattering lights from two laser beams are focused by the fisheye lenses, filtered by the optical filters, and then captured by the CCD sensors. Then the CCD cameras convert the optical signals to the electric signals and transport the signals to the laptops.
After the instrument has been set up, the first step is to measure the distances among the CCD cameras, the laser beams and the laser emitters. With the CCD and lense models we applied, the distance between Laser1 and CCD1 should be around 100m to get a suitable vertical resolution curve, while that between CCD2 and the laser beam emitted from Laser2 should be around 1m to reduce the optical depth between CCD2 and the laser beam. With different CCD and lense models, the pixel array resolution of the CCDs and the focal length of lenses are different, thus the distances should be adjusted accordingly. The light scattered at different positions on the laser beam will be collected by different pixels on CCD2, so that the scattering light at different angles can be retrieved from the captured image. Due to the open structure of CLADS, the background noise is much higher during daytime than nighttime, which is why currently the CLADS system was only used during nighttime.
2.2 Improved retrieval method of modified CLADS
The data collection and processing routine of the modified CLADS takes three steps: image processing to obtain the horizontal angular resolved scattering signal and the vertical profile; retrieval method to derive the aerosol scattering phase function and the vertical profile of scattering function; substituting aerosol phase function into the retrieval method to derive the aerosol extinction coefficient profile.
First, the images captured by these two CCD cameras should be processed to obtain angular-resolved scattering signals. Images of the vertical and horizontal laser beam captured by the two CCDs are transported to the computer. Scattering signals without CCD current noises are obtained by subtracting the intensities of the dark image from the measurement images. The central axis of the laser beams are fitted using linear regressions and a relationship is setup between pixels on the axis and corresponding scattering angles (θ). A normal distribution was then used to fit the signal of each point on the central axis to obtain the intensities of atmospheric (aerosol and air molecule) scattering signals for both vertical and horizontal laser beams. Combining the intensities and the relationship between scattering angles and pixels, the angular resolved signal intensities can be derived on vertical/horizontal laser beams respectively. Detailed description of the data collecting and pre-processing process can be found in Bian et al. [36].
After the calculation of angular resolved scattering signals, the signals captured by the two CCD cameras need to be processed respectively. Figure 2 shows the flow chart of the retrieval algorithm for determining aerosol extinction coefficient profiles from the signals. Depending on the geometric relationship between CCD1 (vertical) and Laser1, the corresponding relationship between scattering angle θ1 and scattering altitude z can be build. The angular resolved scattering signals captured by CCD1 can then be translated into vertical signal profiles. The return signal from CCD1 can be expressed as:
where EL1 is the total laser energy emitted from Laser1 during the exposure time, Sp the area of each pixel on the CCD sensor, Tz the transmittance of the laser light from the emitting point to the altitude z, TR the transmittance of scattering light from altitude z to CCD1 across a distance R, β(z) the scattering function at altitude z and D the distance between the CCD and Laser1. K1 is a calibration factor, which depends on the optical efficiency of the receiving part with CCD1.Fig. 2. Flow chart of the retrieval algorithm for determining aerosol extinction coefficient profile from modified CLADS measurements.
Under the assumption that there are no aerosols above 4 km altitude, scattering function profiles of air molecules and the ratios between gas and aerosol scattering signals can be obtained based on Eq. (1). Then, vertical profiles of aerosol scattering function at a single scattering angle can be calculated as follows:
Here, the scattering function of air molecules βair(z) does not have a significant temporal variation. To simplify the equation, parameters unrelated to atmospheric scattering and altitude (K1, EL1, Sp and D) are collectively defined as N.After deriving βaerosol(z), aerosol single scattering albedo ω and aerosol phase function paerosol(θ1) are required in Eq. (3) to retrieve the aerosol extinction coefficient profiles kex-aerosol(z).
Substituting molecular scattering phase function into Eq. (4), the angle-resolved aerosol scattering signal Eaerosol(θ) can be derived,
Finally, the aerosol extinction coefficient profile kex-aerosol(z) can be retrieved by substituting the paerosol(θ) derived with Eq. (5) and the βaerosol(z) derived with Eq. (2) into the Eq. (3). Simultaneous single scattering albedo measurements could also be substituted into Eq. (3) if there are any. If not, the single scattering albedo from CALIPSO aerosol classification can be used as substitute. This substitution has lead to an uncertainty up to 25% in our previous studies [28].
3. Field measurements and results
3.1 Measurements in Beijing 2016
During January 2016, a comprehensive field campaign focused on wintertime air pollution in the North China Plain (NCP) was conducted on the Yanqi campus of the University of Chinese Academy of Sciences (UCAS, 40°24′ N, 116°40′ E, 91 m a.s.l.) located in Huairou district, Beijing, China. This site is 60 km northeast of downtown Beijing and is at the edge of the NCP, which makes it suitable for the study of regional pollution properties. In this campaign, the modified CLADS introduced above was mounted next to a mobile laboratory and an automatic weather station on the roof of an academic building (Fig. 3). Aerosol particle number size distributions (PNSD), scattering/hemi-backscattering coefficients and absorption coefficients were measured in the mobile laboratory using a scanning mobility particle sizer (SMPS; Model 3936, TSI, Inc., Shoreview, MN, USA)/aerodynamic particle sizer (APS; Model 3321, TSI, Inc., Shoreview, MN, USA), a nephelometer (Aurora 3000, Ecotech Pty Ltd., Australia) and a multi angle absorption photometer (MAAP, Thermo Fisher Scientific Inc., USA), respecitvely, with sample flow relative humidity (RH) dried to below 20% [38]. 522 profiles were obtained with the modified CLADS in the time period from 5th to 21th January.
Figure 4 shows the time series of meteorological parameters and aerosol properties measured during the field campaign. The aerosol extinction coefficient profiles and scattering phase functions were retrieved using the modified CLADS measurements. The gap of data on January 13th resulted from a power blackout, while CLADS measurements were put on hold during a snow event from January 16th to 17th. During the campaign, ambient temperatures were mostly below 0 °C and relative humidity maintained at a low level (<50%) [Fig. 4(a)]. Under these circumstances, little change can be expected in the characteristics of ambient aerosols before and after passing the dryer of the mobile laboratory. Mountain-valley circulations were frequently observed with prevailing wind directions turning from southeast during daytime to southwest during nighttime [(Fig. 4(b)]. Figure 4(c) shows the scattering/absorption coefficient of dry aerosols measured in the mobile laboratory. During the observation period of CLADS, dry aerosol scattering coefficients (blue line) were mostly below 150 Mm−1 except at the midnight on January 21th, indicating that the atmosphere was relatively clean during this campaign, especially during nighttime when winds came from the mountain areas. From the afternoon of January 20th, air pollutants began to accumulate on the site. Several new particle formation events were detected in the PNSD measurements [Fig. 4(d)]. Comparing the measured PVSD [Fig. 4(e)] with the PVSDs of CALIPSO aerosol types, our measurements obviously fit into the biomass burning aerosol category [39]. The Mie model [27] was used together with PNSD measurements to simulate the scattering phase function of dry aerosols in this campaign. Due to the low ambient RH, the simulated scattering phase function can be used to approximate ambient scattering phase function in the retrieval of ambient aerosol extinction coefficient profiles measured by the vertical CLADS (CCD1). Figure 4(f) displays the profiles retrieved using simulated phase functions. The boundary layer heights shown in this figure were estimated by using a “gradient method” introduced in [25] with the retrieved profiles. The results show that the nocturnal boundary layers were below 200 m during the clean period. Nocturnal boundary layer height significantly increased since January 20th with accumulation of aerosols. A residual layer indicated by high aerosol loading at ∼1 km altitude was detected after sunset on January 5th, 8th, 9th and 19th. The aerosol scattering phase functions measured with horizontal CLADS (CCD2) are depicted in Fig. 4(g). Aerosol scattering phase function only displayed significant variations when there were distinct changes in PNSD. The retrieved aerosol vertical profile and phase functions are discussed in detail in the following sections.
Fig. 4. Time series of (a) temperature and relative humidity, (b) wind speed and wind direction, (c) scattering coefficient (Aurora3000) and absorption coefficient (MAAP) at 525nm, (d) PNSD, (e) particle volume size distribution (PVSD), (f) extinction coefficient profile at 532nm, (g) aerosol scattering phase function at 532nm in Beijing 2016 field campaign.
3.2 Aerosol scattering function profiles derived from CCD1 measurements
From the retrieval algorithm for the aerosol extinction coefficient profile (Fig. 2) it can be seen that the most crucial part before applying the aerosol phase function is the derivation of the aerosol scattering function profile. The probability distribution of the aerosol scattering functions at each altitude are shown in Fig. 5. It should be noted that scattering angles vary with altitude due to the geometric structure of vertical CLADS, being 90° at surface level and rising up to 180° with increasing altitude. Most of the atmospheric scattering function are below 4 Mm−1 sr−1 near the surface, again indicating low aerosol loading. Several profiles revealed high values from ground up to ∼1km. Another group of profiles revealed peaks between 500 m to 1 km, with peak values often exceeding the aerosol scattering function at surface level. According to these characteristics, atmospheric scattering function profiles were classified into three categories: “clean”, “polluted” and “residual” case.
Fig. 5. The probability distribution of aerosol scattering function at each altitude measured with the modified CLADS in Beijing 2016 field campaign (total sample number of 522).
The profiles classified into the three case categories are shown in Fig. 6, with their respective average profiles depicted as the black lines. The “clean” case (Case 1) has 440 profiles, featuring low aerosol scattering functions from ground to 2 km altitude, with a very low average ground level aerosol scattering function (scattering angle=90°) of 3.04 Mm−1 sr−1. The “polluted” case (Case 2) with only 24 samples all observed on January 20th, shows largest aerosol scattering functions (12.38 Mm−1 sr−1 on average) at surface level, which decrease rapidly with altitude within 100 m, due to the fast change in scattering angles in this altitude range. The standard deviation of this case is very small on the ground, gradually increasing with altitude. The “residual” case (Case 3) grouped 58 aerosol scattering function profiles that decreased quickly with altitude from ground up, featuring an additional peak between 500 m – 1 km that was comparable or even higher than the surface aerosol scattering function.
Fig. 6. Three categories of atmospheric scattering functions measured with modified CLADS in Beijing 2016 field campaign. Gray lines show the measured profiles; thick black lines show the average profiles of the 3 categories; error bars represent the standard deviations of the average profile on different altitudes.
3.3 Aerosol extinction coefficient profiles retrieved with CALIPSO aerosol phase function
Following the methodology introduced in Section 2.2, the aerosol extinction coefficient profile can be retrieved with aerosol scattering function profile and aerosol phase function. To evaluate the influence of aerosol phase function to the retrieval results, three kinds of aerosol phase functions from CALIPSO aerosol classification were used to retrieve the aerosol extinction coefficient for the three case categories introduced in Section 4.2.
Results are displayed in Fig. 7, which shows the retrieved aerosol extinction coefficient profiles (gray lines) and the average extinction coefficients profiles and their standard deviation at various altitudes. In all the three cases, significant differences were detected in the vertical profiles retrieved with different aerosol phase functions, especially in Case 2. Differences were typically larger in the near surface layer. Aerosol extinction coefficients of the ground layer (0-10 m) and of the 100 m height level retrieved with different types of aerosol phase functions respectively for the three cases were compared in Table 2. Although the vertical pattern of retrieved profiles in Case 1 and Case 3 changed little with different aerosol phase function assumptions, the absolute values varied greatly, with largest and lowest results respectively coming from the “biomass burning” and “dust” aerosol assumptions. The relative differences between the results using these two assumptions are 20.07% and 15.45% near the ground, 44.25% and 39.52% at 100m altitude for Case 1 and Case 3, respectively. For the “polluted” case, even the vertical distribution pattern changed remarkably with aerosol phase function. The average extinction coefficient retrieved with “biomass burning” aerosol phase function at 100 m is higher than within the ground layer. However, when “polluted dust” and “dust” aerosol phase functions were used, the aerosol extinction coefficients at 100 m are lower than that in the ground layer. This sensitivity study shows that, if ambient aerosol phase function changed from CALIPSO “biomass burning” to “dust” aerosol phase function, using a constant aerosol phase function would result in large uncertainties.
Fig. 7. Aerosol extinction coefficient profiles retrieved with three different aerosol phase functions: CALIPSO-Biomass burning (a, b, c), CALIPSO-Polluted Dust (d, e, f), CALIPSO-Dust (g, h, i). Different rows represent different case categories as specified in Sect. 4.2: Case 1 (a, d, g), Case 2 (b, e, h) and Case 3 (c, f, i). Gray lines are the retrieved profiles; black lines and error bars are the average profiles and standard deviations.
Table 2. Average values of retrieved aerosol extinction coefficient (unit: Mm−1)
Asymmetry factor is the commonly used single-valued representation of scattering phase function [27]. The asymmetry factors (g) of these phase functions used here are 0.638, 0.653, 0.662, respectively. According to previous study [40], the diurnal variation of ambient aerosol g values ranged from 0.62 to 0.71 in NCP, indicating significant variation in ambient aerosol phase function. Differences of aerosol phase function from sunset to sunrise may even be larger than the differences between the three CALIPSO aerosol phase functions used here, indicating the assumption of a constant aerosol phase function might even induce larger uncertainties than in our sensitivity study. Real-time measured ambient aerosol phase functions can significantly reduce such uncertainties.
3.4 Ambient aerosol phase function derived from CCD2 measurement
As discussed in Section 4.1, the Mie model together with the PNSD measurements can provide information on dry state aerosol phase functions, however, ambient aerosol phase functions are necessary in the retrieval algorithm to obtain aerosol extinction coefficient profiles, since CLADS measures ambient aerosol scattering function profiles. Using simulated dry state aerosol phase functions would induce large uncertainties under higher RH conditions. Horizontal CLADS has the ability to measure the ambient aerosol phase function near the ground during night time accurately [36]. During the Beijing 2016 field campaign, ambient aerosol phase function was measured by the horizontal CLADS (CCD2) simultaneous to the vertical CLADS measurements. It already came to attention that aerosol phase functions were varying with particle size distributions [Fig. 4(g)]. To analyze this phenomenon in detail, measurement data at two nights are selected as examples (Fig. 8).
Fig. 8. Time series of (a/e) scattering coefficient and absorption coefficient at 525nm, (b/f) particle surface area size distribution (PASD), (c/g) extinction coefficient profile at 532nm, (d/h) aerosol scattering phase function at 532nm on January 10th/18th, respectively.
Figure 8(c) shows an obvious aerosol layer descending to the ground from ∼700 m height after sunset on January 9th. When the high aerosol loading reached surface at about 21:00, the aerosol scattering/absorption coefficients increased significantly on ground level [Fig. 8(a)]. At the same time, the particle surface area size distribution (PASD) also changed because of the different sources and aging level between aerosols from ground and upper layers. Aerosol phase function changes with particle scattering cross-section, which is expressed in units of area and is more susceptible to changes in PASD rather than PNSD or PVSD [27]. In the abrupt change of PASD, the proportion of particles smaller than 100 nm rapidly increased [Fig. 8(b)], leading to sudden variations in aerosol phase function [Fig. 8(d)]. Figures 8(e)–8(h) show another case on January 18th. Aerosol concentrations continuously decreased during the whole night from 18:00 on January 17th to 6:00 on January 18th. Before midnight, although aerosol concentrations and optical properties kept changing, the measured aerosol phase function varied little, since the PASD preserved its shape during the overall decrease in aerosol concentrations. After midnight, the shape of PASD changed and aerosol scattering phase function varied correspondingly. This phenomenon indicates that the shape of aerosol surface area size distribution rather than the absolute aerosol surface area concentration was the determining the aerosol scattering phase function.
3.5 Aerosol extinction coefficient profiles retrieved with measured phase function
To further evaluate the retrieval method of modified CLADS, a comparison between remote sensing retrieved and in-situ measured aerosol extinction coefficients was made. The in-situ measured aerosol extinction coefficient is the summation of scattering coefficient and absorption coefficient respectively measured with Aurora 3000 nephelometer and MAAP. Note that these instruments measured dry state aerosol properties in the laboratory. However, due to the relatively low ambient RH in this campaign, the in-situ measurement can be used to compare with ambient aerosol observations.
Figure 9 shows the comparison results. Here, aerosol phase function retrieved with real time modified CLADS, aerosol phase function of “biomass burning” aerosol introduced in the CALIPSO aerosol classification and aerosol phase function simulated with Mie model and PNSD measurements are used to retrieve the aerosol extinction coefficient profiles and compared to in-situ measurements. Due to the lack of extinction coefficient profile measurements to compare with CLADS measurements, the retrieved extinction coefficient can only be compared with in-situ ground measurements. The extinction coefficient profiles were averaged over 0-10 m height to represent ground level remote sensing measurements. The slope of the fitting lines reached 0.99, 0.85 and 0.97, respectively. Results obtained using measured and modelled aerosol phase function were closer to in-situ measurements than those retrieved using “biomass burning” aerosol phase function. Although our PVSD measurements resembled that in the CALIPSO “biomass burning” category, using a constant “biomass burning” aerosol phase function that does not reflect the variation of real ambient aerosol phase function still induced the largest discrepancies between retrieved and in-situ measured extinction coefficients. The coefficients of determination (R2 in Fig. 9) were 0.87, 0.89 and 0.90, respectively, indicating overall good linearity, with the results retrieved with modelled aerosol phase function agreeing slightly better with in-situ measurements. However, the modelled aerosol phase function was only applicable because the relative humidity during this campaign remained mostly below 50%. Under higher environmental humidity, the optical properties of ambient aerosols will differ more from that of dry state aerosols measured in the laboratory due to aerosol hygroscopic growth. Under such conditions, modelled aerosol phase function would be inapplicable in the retrieval of ambient aerosol extinction coefficient profiles.
Fig. 9. Comparison results between extinction coefficients measured by CLADS and in-situ instruments (nephelometer and MAAP) with different aerosol phase functions: a. CLADS measured phase function; b. aerosol phase function of CALIPSO-biomass burning aerosols; c. aerosol phase function simulated with Mie model. (Dash lines are the 1:1 line while solid lines are the fitting lines).
Overall, retrieval results of the modified CLADS could well represent the in-situ measured aerosol extinction coefficients, with a slope close to 1. Although retrieval results using modelled phase functions seemed to have outperformed real-time measured ones by the horizontal CLADS (CCD2) in terms of R2, it was only applicable due to the low RH conditions within this campaign. The modified CLADS system has the advantage of measuring aerosol phase functions under ambient RH conditions and deriving simultaneous aerosol extinction vertical profiles independent of other measurements. Since the aerosol phase function used here was retrieved from the horizontal configuration at surface level, its application to levels far away from the surface, where the result could not be validated in this study, might lead to higher uncertainties than those on ground level.
4. Conclusions and discussions
Aerosols play an important role in air pollution and climate change. Many instruments were developed to measure ground level aerosol physical and chemical properties under dry state conditions. However, the vertical distribution of ambient aerosols might be of higher importance for the assessment of their climate impacts. A charged couple device camera-laser atmospheric detection system (CLADS) was developed to measure the vertical distribution of aerosol optical characteristics, especially under 1 km altitude, however, lacking direct measurements of ambient aerosol phase function, which is an important parameter in the CLADS retrieval algorithm. In this study, an improved method along with a modified CLADS was developed to simultaneously measure the ambient scattering aerosol phase function with vertical profiles of aerosol extinction coefficients. The retrieval algorithm first calculates the real time ground level aerosol phase function and aerosol scattering function profiles separately, then combines them to yield aerosol extinction coefficient profiles.
Performance of the improved retrieval method with modified CLADS was evaluated in a comprehensive field campaign on wintertime air pollution in the North China Plain during January 2016. 522 aerosol scattering function profiles were obtained by the vertical CLADS, which were classified into three case categories (“clean”, “polluted” and “residual”) according to their vertical distribution characteristics. Different aerosol phase functions provided in the CALIPSO aerosol classification were used to retrieve aerosol extinction coefficient profiles. Results show that time-resolved aerosol phase functions are necessary in retrieval of aerosol extinction coefficient profiles. Assuming constant aerosol phase functions would lead to significant uncertainties. Real-time measured ambient aerosol phase function by the horizontal CLADS and Mie code modelled ones were also used to retrieve the aerosol extinction coefficient profiles, which both agreed better with in-situ measurements than those retrieved with a constant CALIPSO aerosol phase function assumption, again emphasizing the importance to use time-resolved phase functions. However, the Mie model approach is only applicable to low humidity conditions and depends on PNSD and aerosol absorption property measurements, whereas the modified CLADS can derive ambient aerosol phase functions with relatively high accuracy and use them to simultaneously retrieve vertical profiles of aerosol extinction coefficients with low uncertainties and independent of other measurement instruments.
It should be noted that, the aerosol scattering phase function used in this study, is measured at surface level. Based on the assumption that the aerosol phase function does not change with altitude when aerosols are well mixed, the measured aerosol phase functions on the ground was used to retrieve the vertical profile of aerosol properties. To date, we have no means to measure the real-time vertical structures of aerosol phase function, which is why such kind of assumptions are typically used in the retrieval algorithm of remote sensing instruments (i.e. CLADS and lidar systems). However, this assumption has its limitation. The uncertainty in derived aerosol extinction coefficients will increased with height. Therefore, future measurements of aerosol phase function vertical profiles are very crucial for further lowering uncertainties in the retrieval of remote sensing instruments. CLADS has the potential to derive the aerosol phase functions at different height with several CCDs. Furthermore, the vertical structure of the aerosol asymmetry factor, which is also an important parameter in the radiative transfer model, could be calculated with the vertical structure of aerosol phase functions.
Funding
National Key Research and Development Program of China (2017YFC0212803); National Natural Science Foundation of China (41875159, 41975043); Natural Science Foundation of Beijing Municipality (8184076, 8194080); China Meteorological Administration (2018B07).
References
1. K. S. Carslaw, L. A. Lee, C. L. Reddington, K. J. Pringle, A. Rap, P. M. Forster, G. W. Mann, D. V. Spracklen, M. T. Woodhouse, L. A. Regayre, and J. R. Pierce, “Large contribution of natural aerosols to uncertainty in indirect forcing,” Nature 503(7474), 67–71 (2013). [CrossRef]
2. R. K. Pachauri, M. R. Allen, V. R. Barros, J. Broome, W. Cramer, R. Christ, J. A. Church, L. Clarke, Q. Dahe, P. Dasgupta, N. K. Dubash, O. Edenhofer, I. Elgizouli, C. B. Field, P. Forster, P. Friedlingstein, J. Fuglestvedt, L. Gomez-Echeverri, S. Hallegatte, G. Hegerl, M. Howden, K. Jiang, B. Jimenez Cisneroz, V. Kattsov, H. Lee, K. J. Mach, J. Marotzke, M. D. Mastrandrea, L. Meyer, J. Minx, Y. Mulugetta, K. O’Brien, M. Oppenheimer, J. J. Pereira, R. Pichs-Madruga, G.-K. Plattner, H.-O. Pörtner, S. B. Power, B. Preston, N. H. Ravindranath, A. Reisinger, K. Riahi, M. Rusticucci, R. Scholes, K. Seyboth, Y. Sokona, R. Stavins, T. F. Stocker, P. Tschakert, D. van Vuuren, and J.-P. van Ypserle, Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC, 2014), p. 151.
3. B. Li, T. Gasser, P. Ciais, S. Piao, S. Tao, Y. Balkanski, D. Hauglustaine, J.-P. Boisier, Z. Chen, M. Huang, L. Z. Li, Y. Li, H. Liu, J. Liu, S. Peng, Z. Shen, Z. Sun, R. Wang, T. Wang, G. Yin, Y. Yin, H. Zeng, Z. Zeng, and F. Zhou, “The contribution of China’s emissions to global climate forcing,” Nature 531(7594), 357–361 (2016). [CrossRef]
4. B. H. Samset, “How cleaner air changes the climate,” Science 360(6385), 148–150 (2018). [CrossRef]
5. J. Heintzenberg and R. J. Charlson, “Design and Applications of the Integrating Nephelometer: A Review,” J. Atmos. Oceanic Technol. 13(5), 987–1000 (1996). [CrossRef]
6. N. Ma, C. S. Zhao, A. Nowak, T. Müller, S. Pfeifer, Y. F. Cheng, Z. Z. Deng, P. F. Liu, W. Y. Xu, L. Ran, P. Yan, T. Göbel, E. Hallbauer, K. Mildenberger, S. Henning, J. Yu, L. L. Chen, X. J. Zhou, F. Stratmann, and A. Wiedensohler, “Aerosol optical properties in the North China Plain during HaChi campaign: an in-situ optical closure study,” Atmos. Chem. Phys. 11(12), 5959–5973 (2011). [CrossRef]
7. Y. Kuang, C. S. Zhao, J. C. Tao, Y. X. Bian, and N. Ma, “Impact of aerosol hygroscopic growth on the direct aerosol radiative effect in summer on North China Plain,” Atmos. Environ. 147, 224–233 (2016). [CrossRef]
8. P. F. Liu, C. S. Zhao, Q. Zhang, Z. Z. Deng, M. Y. Huang, X. C. Ma, and X. X. Tie, “Aircraft study of aerosol vertical distributions over Beijing and their optical properties,” Tellus B 61(5), 756–767 (2009). [CrossRef]
9. A. A. Kokhanovsky, Aerosol Optics (Springer, 2008).
10. S. Wu, X. Song, B. Liu, G. Dai, J. Liu, K. Zhang, S. Qin, D. Hua, F. Gao, and L. Liu, “Mobile multi-wavelength polarization Raman lidar for water vapor, cloud and aerosol measurement,” Opt. Express 23(26), 33870–33892 (2015). [CrossRef]
11. A. Ansmann, M. Riebesell, and C. Weitkamp, “Measurement of atmospheric aerosol extinction profiles with a Raman lidar,” Opt. Lett. 15(13), 746–748 (1990). [CrossRef]
12. H. Xie, T. Zhou, Q. Fu, J. Huang, Z. Huang, J. Bi, J. Shi, B. Zhang, and J. Ge, “Automated detection of cloud and aerosol features with SACOL micro-pulse lidar in northwest China,” Opt. Express 25(24), 30732–30753 (2017). [CrossRef]
13. C. Weitkamp, Lidar Range-Resolved Optical Remote Sensing of the Atmosphere (Springer, 2005).
14. S. Groß, J. Gasteiger, V. Freudenthaler, T. Müller, D. Sauer, C. Toledano, and A. Ansmann, “Saharan dust contribution to the Caribbean summertime boundary layer – a lidar study during SALTRACE,” Atmos. Chem. Phys. 16(18), 11535–11546 (2016). [CrossRef]
15. T. Nishizawa, N. Sugimoto, I. Matsui, A. Shimizu, Y. Hara, U. Itsushi, K. Yasunaga, R. Kudo, and S.-W. Kim, “Ground-based network observation using Mie–Raman lidars and multi-wavelength Raman lidars and algorithm to retrieve distributions of aerosol components,” J. Quant. Spectrosc. Radiat. Transfer 188, 79–93 (2017). [CrossRef]
16. N. C. P. Sharma, J. E. Barnes, T. B. Kaplan, and A. D. Clarke, “Coastal Aerosol Profiling with a Camera Lidar and Nephelometer,” J. Atmos. Oceanic Technol. 28(3), 418–425 (2011). [CrossRef]
17. K. Meki, K. Yamaguchi, X. Li, Y. Saito, T. D. Kawahara, and A. Nomura, “Range-resolved bistatic imaging lidar for the measurement of the lower atmosphere,” Opt. Lett. 21(17), 1318–1320 (1996). [CrossRef]
18. J. E. Barnes, S. Bronner, R. Beck, and N. C. Parikh, “Boundary Layer Scattering Measurements with a Charge-Coupled Device Camera Lidar,” Appl. Opt. 42(15), 2647–2652 (2003). [CrossRef]
19. J. E. Barnes, N. C. P. Sharma, and T. B. Kaplan, “Atmospheric aerosol profiling with a bistatic imaging lidar system,” Appl. Opt. 46(15), 2922–2929 (2007). [CrossRef]
20. Z. Tao, D. Liu, Z. Wang, X. Ma, Q. Zhang, C. Xie, G. Bo, S. Hu, and Y. Wang, “Measurements of aerosol phase function and vertical backscattering coefficient using a charge-coupled device side-scatter lidar,” Opt. Express 22(1), 1127–1134 (2014). [CrossRef]
21. Y. Bian, C. Zhao, W. Xu, N. Ma, J. Tao, Y. Kuang, G. Zhao, and H. Liu, “Method to retrieve the nocturnal aerosol optical depth with a CCD laser aerosol detective system,” Opt. Lett. 42(22), 4607–4610 (2017). [CrossRef]
22. L. Mei, P. Guan, Y. Yang, and Z. Kong, “Atmospheric extinction coefficient retrieval and validation for the single-band Mie-scattering Scheimpflug lidar technique,” Opt. Express 25(16), A628–A638 (2017). [CrossRef]
23. G. Sun, L. Qin, Z. Hou, X. Jing, F. He, F. Tan, and S. Zhang, “Small-scale Scheimpflug lidar for aerosol extinction coefficient and vertical atmospheric transmittance detection,” Opt. Express 26(6), 7423–7436 (2018). [CrossRef]
24. G. Zhao, E. Malmqvist, S. Török, P.-E. Bengtsson, S. Svanberg, J. Bood, and M. Brydegaard, “Particle profiling and classification by a dual-band continuous-wave lidar system,” Appl. Opt. 57(35), 10164–10171 (2018). [CrossRef]
25. Y. Bian, C. Zhao, W. Xu, Y. Kuang, J. Tao, W. Wei, N. Ma, G. Zhao, S. Lian, W. Tan, and J. E. Barnes, “A novel method to retrieve the nocturnal boundary layer structure based on CCD laser aerosol detection system measurements,” Remote Sens. Environ 211, 38–47 (2018). [CrossRef]
26. Z. Wang, Z. Tao, D. Liu, D. Wu, C. Xie, and Y. Wang, “New experimental method for lidar overlap factor using a CCD side-scatter technique,” Opt. Lett. 40(8), 1749–1752 (2015). [CrossRef]
27. C. F. Bohren and D. R. Huffman, Absorption and scattering of light by small particles (John Wiley & Sons, 2008).
28. S. Lian, Y. Bian, G. Zhao, W. Li, and C. Zhao, “Dual CCD detection method to retrieve aerosol extinction coefficient profile,” Opt. Express 27(20), A1529–A1543 (2019). [CrossRef]
29. H. Kim, B. Barkey, and S. E. Paulson, “Real refractive indices of α- and β-pinene and toluene secondary organic aerosols generated from ozonolysis and photo-oxidation,” J. Geophys. Res.: Atmos. 115(D24), D24212 (2010). [CrossRef]
30. O. Muñoz, H. Volten, J. F. de Haan, W. Vassen, and J. W. Hovenier, “Experimental determination of scattering matrices of randomly oriented fly ash and clay particles at 442 and 633 nm,” J. Geophys. Res.: Atmos. 106(D19), 22833–22844 (2001). [CrossRef]
31. J.-L. Castagner and I. J. Bigio, “Polar nephelometer based on a rotational confocal imaging setup,” Appl. Opt. 45(10), 2232–2239 (2006). [CrossRef]
32. H. Horvath, L. Alados Arboledas, and F. J. Olmo Reyes, “Angular scattering of the Sahara dust aerosol,” Atmos. Chem. Phys. 18(23), 17735–17744 (2018). [CrossRef]
33. B. Barkey and K. N. Liou, “Polar nephelometer for light-scattering measurements of ice crystals,” Opt. Lett. 26(4), 232–234 (2001). [CrossRef]
34. D. B. Curtis, M. Aycibin, M. A. Young, V. H. Grassian, and P. D. Kleiber, “Simultaneous measurement of light-scattering properties and particle size distribution for aerosols: Application to ammonium sulfate and quartz aerosol particles,” Atmos. Environ. 41(22), 4748–4758 (2007). [CrossRef]
35. W. R. Espinosa, L. A. Remer, O. Dubovik, L. Ziemba, A. Beyersdorf, D. Orozco, G. Schuster, T. Lapyonok, D. Fuertes, and J. V. Martins, “Retrievals of aerosol optical and microphysical properties from Imaging Polar Nephelometer scattering measurements,” Atmos. Meas. Tech. 10(3), 811–824 (2017). [CrossRef]
36. Y. Bian, C. Zhao, W. Xu, G. Zhao, J. Tao, and Y. Kuang, “Development and validation of a CCD-laser aerosol detective system for measuring the ambient aerosol phase function,” Atmos. Meas. Tech. 10(6), 2313–2322 (2017). [CrossRef]
37. K. Miyamoto, “Fish Eye Lens,” J. Opt. Soc. Am. 54(8), 1060–1061 (1964). [CrossRef]
38. T. M. Tuch, A. Haudek, T. Müller, A. Nowak, H. Wex, and A. Wiedensohler, “Design and performance of an automatic regenerating adsorption aerosol dryer for continuous operation at monitoring sites,” Atmos. Meas. Tech. 2(2), 417–422 (2009). [CrossRef]
39. A. H. Omar, J.-G. Won, D. M. Winker, S.-C. Yoon, O. Dubovik, and M. P. McCormick, “Development of global aerosol models using cluster analysis of Aerosol Robotic Network (AERONET) measurements,” J. Geophys. Res.: Atmos. 110(F1), F01014 (2005). [CrossRef]
40. G. Zhao, C. Zhao, Y. Kuang, Y. Bian, J. Tao, C. Shen, and Y. Yu, “Calculating the aerosol asymmetry factor based on measurements from the humidified nephelometer system,” Atmos. Chem. Phys. 18(12), 9049–9060 (2018). [CrossRef]
