† Corresponding author. E-mail:
A method for aerosol extinction profile retrieval using ground-based multi-axis differential optical absorption spectroscopy (MAX-DOAS) is studied, which is based on a look-up table algorithm. The algorithm uses parametric method to represent aerosol extinction profiles and simulate different atmospheric aerosol states through atmospheric radiation transfer model. Based on the method, aerosol extinction profile was obtained during six cloud-free days. The O4 differential air mass factor (dAMF) measured by MAX-DOAS is compared with the corresponding model results under different atmospheric conditions (
With the development of industry and the acceleration of urbanization, the concentration of particulate matter and secondary particulate matter in the atmosphere has increased substantially, which has an important impact on atmospheric environmental quality and terrestrial gas system radiation budget.[1,2] Due to the uncertainty of aerosol in morphology, composition, physical and chemical properties, and optical properties, it is difficult to accurately assess the impact of aerosol on the environment and global climate. Therefore, the spatial and temporal distribution, optical properties, and spectral distribution characteristics of aerosol have become a research hotspot in atmospheric environmental monitoring and related climate effects. It is very important to obtain aerosol optical properties for understanding the evolution process of atmospheric pollution, calibration of model, and satellite data.[3]
In recent years, differential optical absorption spectroscopy (DOAS) is used in the field of atmospheric environmental monitoring.[4] The multi-axis DOAS (MAX-DOAS) composed of zenith direction and low observation elevation greatly improves the observation sensitivity of the tropospheric atmosphere. Trace gas profiles in the troposphere can be obtained by combining with the atmospheric radiation transfer model. In 2004, Wagner et al. proposed to retrieve the atmospheric scattering light path by using the characteristics of oxygen dimer (O4) vertical profile which is basically unchanged in the atmosphere and evenly distributed horizontally.[5] Since the scattering energy of aerosol strongly affects the photonic optical path of the atmosphere, the aerosol optical thickness and profile information can be obtained by observing the slant column density (SCD) of O4 through ground-based MAX-DOAS. In order to achieve the rapid, accurate, and real-time acquisition of aerosol profile information, a series of studies have been carried out by different research groups.[6–8] At present, two main algorithms of the optimal estimation method and look-up table method are used to obtain aerosol optical properties based on the SCD of O4 through multi-elevation angles with MAX-DOAS. The optimal estimation method can obtain a continuous profile with a high vertical resolution. However, the algorithm relies on prior information and singular values could appear in the retrieval.[9] The look-up table method carries out parameter quantization assumption for aerosol profile, and the retrieval of aerosol profile is relatively fast in look-up table algorithm after the establishment of database. The results are relatively stable because they come from the given situation of database.[10]
In 2010, look-up table method was first used to retrieve the extinction profile of aerosol by Li et al.[11] The results of different time periods were compared with the ground-based in situ data. However, the complete weight factor was not included in the profile hypothesis (0.5–1 proportion of aerosols in the boundary layer) and the effect of extraneous transport on aerosol was not considered. In 2013, Wu et al. used look-up table method to retrieve the aerosol extinction profile and compared it with aerosol optical thickness obtained by the Lidar.[12] However, only four hours of measurement results were obtained in the noon period of a day (9:00–13:00).
The look-up table method including complete weight factor (0.1–1) and different time periods (8:00–16:00) is used to retrieve the aerosol profile in this paper. Field experiments are carried out to study the ability of the algorithm to obtain the profile of aerosol extinction under different atmospheric aerosol conditions. The boundary layer aerosol extinction by MAX-DOAS is compared with PM2.5 data obtained by ground-based point instruments.
The one-dimensional MAX-DOAS system was used in the field, which was developed by Anhui Institute of Optics and Fine Mechanics.[13] The experimental site was located in the northern suburb of Huaibei, and continuous observation was carried out in the cloud-free days from September 29, 2018 to October 4, 2018. The telescope pointed downtown (azimuth angle 177°, north is 0°). Figure
For the aerosol profile, a two-layer setup for vertical aerosol distribution is used in the troposphere (0 km–15 km), which can be described by a limited set of parameters. The aerosol is assumed to be homogeneous in a layer near the ground surface (called boundary layer), and the aerosol extinction coefficient (unit: km−1) in the layer aloft (i.e., the free troposphere) decreases exponentially with height. This type of aerosol distribution is frequently observed in polluted regions and verified by Lidar observation. The extinction profile
![]() | Table 1.
Parameters for look-up table. SZA: solar zenith angle. SRAA: solar relative azimuth angle. . |
The parameters of aerosol optical thickness t, aerosol boundary layer height h, aerosol extinction weight within the boundary layer w, and other atmospheric parameters (such as single scattering albedo) constitute the look-up table.[14] The simulated O4 dAMF of each elevation angle under different atmospheric conditions is calculated through the atmospheric radiation transmission model. The model results and measurement results are compared to obtain the aerosol profile which best conforms to the real atmospheric conditions through the minimization process. The algorithm diagram is shown in Fig.
![]() | Fig. 2. Schematic diagram of aerosol profile retrieved with the absorption of O4 observed by MAX-DOAS. |
The full spherical Monte-Carlo atmospheric radiative transfer model, Monte Carlo atmospheric radiative transfer and inversion model (McArtim), is used in this study.[15] To improve the operational efficiency, the simulation wavelength is selected as 350 nm, which is the central wavelength of the retrieval band (337 nm–365.5 nm). Sensitivity studies have shown single scattering albedo ssa, asymmetric factor g, and surface albedo albedo have a small influence on air mass factor (AMF). Therefore, fixed values of ssa=0.92, g=0.72, and albedo=0.05 are adopted in the simulation. The specific settings of the model are shown in Table
The measurement spectrum is retrieved with DOAS method. The fitting structures are NO2 (298 K, vanDaele), O4 (296 K, Hermans), O3 (223 K, Bogumil), and Ring structure. The Ring structure is calculated by DOASIS software.[16] A fitting example of the 2° spectrum is shown in Fig.
![]() | Fig. 3. Fitting example of the measured spectrum. The black and red curves indicate the absorption structures and the fitted absorption structures, respectively. |
Figure
From the minimization process described in Subsection 2.1, the aerosol optical parameters (t, w, h) can be obtained from the iterative calculation of the O4 dAMF. These three parameters determine the aerosol extinction profile. Figure
![]() | Fig. 8. Surface aerosol extinction E 0 and boundary layer height BH retrieved by MAX-DOAS for 6 days. Also shown are PM2.5 data from surface point instrument. |
![]() | Fig. 9. Comparison between surface aerosol extinction measured by MAX-DOAS and PM2.5 data from point instrument. Different colors represent different aerosol optical thicknesses. |
The difference between the two instruments is largest in the morning. Considering the fact that MAX-DOAS measures the aerosol extinction averaged over a distance while the point instrument detects the aerosol information near the ground, the results of the linear regression demonstrate good agreement between the measurement results of these two instruments. The latter is more easily influenced by local emission source and could produce high values.[11] In addition, fog and high aerosols in the morning will reduce atmospheric visibility and the effective optical path at all elevation angles of the MAX-DOAS telescope, which restricts the measurement process by MAX-DOAS.[9]
The retrieval algorithm of aerosol extinction vertical profile is studied based on the O4 absorption of multi-elevation angles by MAX-DOAS. The look-up table method is used to obtain the optimal aerosol extinction profile by minimizing the cost function. By comparing the measured results of O4 dAMF with model results, it is found that this method can accurately obtain the aerosol information for low and high aerosol situation. The retrieved aerosol extinction in boundary layer is compared with PM2.5 data measured by ground-based point instrument. The diurnal variation trend of the two methods is in good agreement. The results show that the look-up table method can obtain the aerosol state of the troposphere and provide validation for other instrument data. However, due to the limitation of the profile shape, it is difficult to retrieve more complex atmospheric pollution state such as the uplift layer. Therefore, different profile shapes will be added to the model and the results will be compared with other algorithms in the future.
[1] | |
[2] | |
[3] | |
[4] | |
[5] | |
[6] | |
[7] | |
[8] | |
[9] | |
[10] | |
[11] | |
[12] | |
[13] | |
[14] | |
[15] | |
[16] | |
[17] |