Aerosol extinction coefficient inversion method without overlap factor calibration

Abstract

This invention discloses an aerosol extinction coefficient inversion method without overlap factor calibration. It utilizes high-spectral-resolution lidar or Raman lidar to obtain aerosol backscattering coefficients and particle linear depolarization ratios. A priori database containing lidar ratios and particle linear depolarization ratios for different aerosol types is established based on long-term regional lidar observations. The backscattering coefficients of mixed aerosols are separated into components according to the prior values ​​of particle linear depolarization ratios. The backscattering coefficients of each component are then multiplied by their corresponding prior lidar ratio values ​​and summed to obtain the aerosol extinction coefficients within the overlap transition region. This invention eliminates the need for real-time overlap factor profiles and additional calibration hardware, reducing system complexity and extending the effective inversion range of aerosol extinction coefficients to the lowest altitude at which lidar can receive backscattered signals. It is suitable for monitoring the optical properties of near-ground aerosols and for lidar network observations.

Description

Technical Field

[0001] This invention belongs to the field of atmospheric lidar remote sensing and aerosol optical property inversion technology, and in particular relates to a method for inverting aerosol extinction coefficient without the need for overlap factor calibration. Background Technology

[0002] The aerosol extinction coefficient is an important optical parameter characterizing the ability of atmospheric aerosols to attenuate light transmission, and it directly reflects the impact of aerosols on visibility, radiative transmission, and regional air quality. Ground-based lidar can perform continuous, real-time, and vertical profile observations of aerosols, and is an important active remote sensing method for obtaining the aerosol extinction coefficient.

[0003] However, in near-ground detection using ground-based lidar, there exists a geometric overlap relationship between the laser beam and the field of view of the receiving telescope, known as the overlap factor. O ( z This can lead to a high lower boundary in the effective inversion range of conventional aerosol extinction coefficient inversion. Overlap factor calibration errors directly affect the accuracy of extinction coefficient inversion near the ground or in the overlap factor transition region. Chinese patent document CN121901681A proposes an aerosol extinction coefficient inversion method and product, and explains the limitations of current lidar systems in terms of detection height and inversion accuracy.

[0004] Existing methods typically calibrate the overlap factor through independent or joint measurements. Chinese patent document CN119064949A proposes a low-blind-zone lidar system and detection method based on dual-beam transmission and multi-fiber core reception. While this method can improve near-ground inversion results to some extent, it requires additional optical structures or auxiliary equipment, significantly increasing system cost, assembly complexity, and operation and maintenance difficulty. Chinese patent document CN121477157A proposes calibrating the overlap factor of Mie scattering lidar by assuming that the atmospheric backscattering coefficient and extinction coefficient are approximately uniform within a certain range. However, this assumption is not applicable to most scenarios and real-time calibration is difficult, limiting its widespread application in operational lidar networks.

[0005] Therefore, there is an urgent need for a method that can utilize backscattering information obtained from lidar inversion, long-term aerosol classification priors, and component separation models to expand the effective inversion range of aerosol extinction coefficients without increasing overlap factor calibration hardware or relying on real-time overlap factor profiles. Summary of the Invention

[0006] To address the problems of existing technologies, such as reliance on overlap factor calibration, high system complexity, insufficient real-time performance, and noise sensitivity in the inversion of aerosol extinction coefficients in the near-ground overlapping transition region, this invention provides an aerosol extinction coefficient inversion method that does not require overlap factor calibration. This method can expand the effective inversion range of aerosol extinction coefficients without adding overlap factor calibration hardware or relying on real-time overlap factor profiles.

[0007] A method for inverting aerosol extinction coefficients without overlap factor calibration includes the following steps: (1) The multi-channel echo signal of the atmosphere to be measured is acquired by lidar and preprocessed to obtain the aerosol optical property parameters of the atmosphere to be measured, including the aerosol backscattering coefficient. Particle linear debiasing ratio and lidar ratio ; (2) Establish a prior database of aerosol optical properties, and perform aerosol clustering based on the lidar ratio and particle linear depolarization ratio in the prior database to obtain the lidar ratio prior values ​​corresponding to different types of aerosols. And the prior value of linear debiasing ratio of particles ; (3) Based on the aerosol optical property parameters obtained in step (1) and the aerosol clustering results obtained in step (2), determine the aerosol mixing type in the atmosphere to be tested; (4) Using the aerosol backscattering coefficient obtained in step (1) And particle linear debias ratio Combined with the prior value of the particle linear debiasing ratio obtained in step (2) The total backscattering coefficient of the mixed aerosol was separated into components to obtain the backscattering coefficients of different components of the aerosol. ; (5) The backscattering coefficients of each aerosol component obtained in step (4) Compare with the corresponding radar prior value Multiplying these components yields the extinction coefficients of each aerosol component. The extinction coefficients of each aerosol component were also analyzed. Summing yields the aerosol extinction coefficient of the atmosphere under test, including the overlap factor transition region. ; (6) Output the profile or spatiotemporal distribution of aerosol extinction coefficient, thereby expanding the effective inversion range of aerosol extinction coefficient without calibration of overlap factor.

[0008] In step (1), the lidar is a lidar whose backscattering coefficient inversion result is not affected by the overlap factor, specifically a high spectral resolution lidar or a Raman lidar.

[0009] Furthermore, when the lidar is a high-spectral-resolution lidar, the multi-channel echo signal includes two polarization channel signals and one molecular channel signal; the two polarization channel signals include a mixed vertical polarization channel signal and a mixed parallel polarization channel signal; the molecular channel signal is the molecular elastic backscattering signal retained after the aerosol mi scattering components have been filtered out or suppressed by the spectral discriminator.

[0010] LiDAR overlap state according to overlap factor It is divided into blind zone, transition zone and full zone, among which Blind spot As a transition zone, The method of the present invention is applied to the overlapping transition region where a portion of the backscattered signal can be received.

[0011] In step (2), when performing aerosol clustering, Mahalanobis distance is used to classify aerosol samples. The covariance matrix introduced by Mahalanobis distance includes at least the aerosol radar ratio and the particle linear debiasing ratio, and may further include one or more of the backscattering color ratio and debiasing ratio.

[0012] In step (4), when the mixed aerosol is composed of a first aerosol component and a second aerosol component, and the prior value of the linear debiasing ratio of the particles of the first aerosol component is... The prior value of the linear debiasing ratio of particles greater than that of the second aerosol component At that time, the backscattering coefficient of the first aerosol component Calculate using the following formula: ; Backscattering coefficient of the second aerosol component Calculate using the following formula: .

[0013] To avoid non-physical values ​​in the separation results, when season , ;when season , .

[0014] In step (4), when the mixed aerosol contains three or more aerosol components, the total backscattering coefficient is separated into two components in order of the prior value of the linear depolarization ratio of the particles from high to low or from low to high. Each additional aerosol type is separated into one additional separation step until the backscattering coefficient of each aerosol component is obtained.

[0015] In step (5), the extinction coefficient of each aerosol component and total aerosol extinction coefficient Calculate using the following formula: ; ; In the formula, For the first Radar ratio prior values ​​for various aerosol components For the first The backscattering coefficient of aerosol component.

[0016] In step (5), the aerosol extinction coefficient The calculation does not use the vertical gradient or optical thickness integral term of the molecular channel signal, thus avoiding the influence of overlap factor error and numerical differential noise on the extinction inversion of the overlap transition region.

[0017] In step (6), the effective inversion lower boundary of the aerosol extinction coefficient profile or spatiotemporal distribution result can be extended to the lowest effective backscattered signal height that satisfies the signal-to-noise ratio and quality control conditions.

[0018] Compared with the prior art, the present invention has the following beneficial effects: First, this invention can invert the extinction coefficient of aerosols in the overlapping transition region without relying on overlap factor calibration, thus avoiding the increased costs and difficulties in assembly and adjustment caused by additional hardware, auxiliary instruments, or complex scanning structures.

[0019] Second, this invention utilizes the characteristic that high-spectral-resolution lidar or Raman lidar is insensitive to the overlap factor when inverting backscattering coefficients, transforming the near-ground extinction inversion problem into a component extinction estimation problem based on the backscattering coefficient and the prior value of the radar ratio, thereby reducing the impact of overlap factor error on the inversion results.

[0020] Third, this invention establishes a priori database of aerosol optical properties through long-term regional observations, making the prior radar ratio and particle linear depolarization ratio regionally representative. Compared with directly using empirical parameters, it can better adapt to local aerosol types and observation conditions.

[0021] Fourth, this invention separates the components of the backscattering coefficient of mixed aerosols by using the prior value of the linear depolarization ratio of particles, and avoids non-physical inversion results by using threshold constraints. It is applicable to observation scenarios that include various types of aerosols such as dust, polluting aerosols, and dust-polluting mixed aerosols.

[0022] Fifth, this invention does not directly rely on the height differentiation of molecular channel signals for extinction calculation in the overlapping transition region, which can reduce the influence of numerical differentiation, smooth window boundary effects and background noise on the near-ground extinction inversion results.

[0023] Sixth, the method of the present invention has clear steps, low computational load, and is easy to implement in a program. It can be used for ground-based lidar operational observation, regional air quality monitoring, and near-surface aerosol extinction coefficient inversion in lidar network observation. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a flowchart of an aerosol extinction coefficient inversion method that does not require overlap factor calibration, according to an embodiment of the present invention.

[0026] Figure 2 This is a schematic diagram of aerosol clustering results based on an optical property database.

[0027] Figure 3 This is a schematic diagram illustrating the process of separating components for the backscattering coefficient of a mixed aerosol and calculating the extinction coefficient.

[0028] Figure 4 This is a schematic diagram comparing the extinction coefficient inversion results of the method of the present invention with those of conventional uncalibrated methods and dual-field calibration methods. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] It should be noted that, unless otherwise specified, the features in the following embodiments and implementation methods can be combined with each other.

[0031] like Figure 1 As shown, a method for inverting aerosol extinction coefficient without overlap factor calibration mainly includes steps such as multi-channel signal acquisition, long-term prior database construction, aerosol classification, backscattering coefficient component separation, extinction coefficient calculation, and result output. Specifically, it includes the following steps: (1) The backscattering coefficient inversion results are not affected by the overlap factor. The multi-channel echo signal of the atmosphere to be measured is obtained by using lidar. The multi-channel echo signal is preprocessed to obtain the aerosol backscattering coefficient, particle linear depolarization ratio and lidar ratio.

[0032] This embodiment employs a 532 nm hyperspectral resolution lidar system for observation. The system includes a hybrid vertical polarization channel, a hybrid parallel polarization channel, and a molecular channel. The hybrid vertical and hybrid parallel polarization channels are used to acquire echo signals in orthogonal polarization directions; the molecular channel uses an iodine molecular absorption cell as a spectral discriminator to suppress aerosol Mie scattering components while retaining the Brillouin component in molecular elastic backscattering.

[0033] In this embodiment, the multi-channel echo signal acquired by the hyperspectral resolution lidar is first preprocessed, including background subtraction, range-squared correction, time averaging, and height registration. Then, the aerosol backscattering coefficient is inverted. Particle linear debiasing ratio and lidar ratio Optical characteristic parameters. The specific formula is: ; ; ; ; ; in, The linear departition ratio of the molecule. For volume linear deflection ratio; The backscattering coefficient of atmospheric molecules; The aerosol extinction coefficient; and These represent the transmittance of the iodine molecule cell for aerosol scattering and molecular scattering signals, respectively. (Superscript ‖ and...) These represent signals in parallel polarization and vertical polarization states, respectively. and These are the backscattered signals received by the hybrid parallel channel and the molecular channel, respectively. Scattering ratio Defined as the total backscattering coefficient With molecular backscattering coefficient The ratio of .

[0034] (2) An a priori database of aerosol optical properties was established based on long-term regional lidar observation data. The aerosol optical property database includes lidar comparison data. And particle linear debias ratio Aerosol clustering was performed based on the optical property data in the database to obtain the prior radar ratio values ​​corresponding to different types of aerosols in the region. And the prior value of linear debiasing ratio of particles .

[0035] Specifically, the aerosol radar ratio and particle linear debiasing ratio obtained from the inversion of long-term observation data are screened to remove outliers; then, clustering methods are used to determine the main aerosol types in the region. In one implementation, a 2×2 covariance matrix composed of the aerosol radar ratio and particle linear debiasing ratio is used to describe the statistical distribution of each aerosol type, and Mahalanobis distance is used to classify the observed samples.

[0036] Assuming the selected single-type aerosol optical property data dimensions are: n A corresponding measurement matrix can be constructed. A For use in subsequent analysis.

[0037] ; Each row of the matrix represents a sample, and each column represents a feature, namely, the ratio of LiDAR to... And particle linear debias ratio The covariance matrix between these two types of features needs to be calculated. : ; in, express The standard deviation of the corresponding column data represent and The Pearson correlation coefficient between them.

[0038] These well-defined samples allow for the classification of a wider range of aerosol optical property data. Specifically, the Mahalanobis distance between the samples and measured data is used as the criterion, and a threshold is set to reclassify the aerosol data within this range. The Mahalanobis distance expression is: ; in, The vector of measured data to be classified. This represents the sample vector. The threshold value affects the proportion of optical property data that can be classified into a specific category of aerosol to the total data.

[0039] (3) Based on the aerosol optical property parameters obtained in step (1) and the aerosol clustering results obtained in step (2), determine the aerosol type or aerosol mixture type in the atmosphere to be measured; specifically, the clustering results established by the database can be verified by comparing the range of the two parameters, lidar ratio and particle linear depolarization ratio, in the aerosol cluster.

[0040] (4) Using the aerosol backscattering coefficient obtained in step (1) And particle linear debias ratio Combined with the prior value of the particle linear debiasing ratio obtained in step (2) The total backscattering coefficient of the mixed aerosol was separated into components to obtain the backscattering coefficients of different components of the aerosol. .

[0041] When the mixed aerosol consists of a first aerosol component and a second aerosol component, and the prior value of the linear debiasing ratio of the particles of the first aerosol component is... The prior value of the linear debiasing ratio of particles greater than that of the second aerosol component At that time, the backscattering coefficient of the first aerosol component The calculation formula is: ; Backscattering coefficient of the second aerosol component The calculation formula is: ; To avoid non-physical values ​​in the separation results, when season , ;when season , For three or more aerosol types, a stepwise binary separation can be performed according to the prior value of the particle linear debias ratio, with an additional separation step for each additional aerosol type.

[0042] (5) The backscattering coefficients of each aerosol component obtained in step (4) Compare with the corresponding radar prior value Multiplying these components yields the extinction coefficients of each aerosol component. The extinction coefficients of each aerosol component were also analyzed. Summing yields the aerosol extinction coefficient of the atmosphere under test, including the overlap factor transition region. The specific formula is: ; ; (6) Output the profile or spatiotemporal distribution of aerosol extinction coefficient, thereby expanding the effective inversion range of aerosol extinction coefficient without calibration of overlap factor.

[0043] like Figure 2As shown, taking long-term observations in the Beijing area as an example, the main aerosol types can be divided into pollutant aerosols, dust-pollution mixed aerosols, and dust aerosols. Based on long-term clustering statistics, the prior values ​​for radar ratio and particle linear debiasing ratio of pollutant aerosols can be taken as 48 sr and 0.05, respectively; for dust-pollution mixed aerosols, 46 sr and 0.20, respectively; and for dust aerosols, 36 sr and 0.32, respectively. These values ​​are only preferred values ​​in this embodiment; in other regions or other observation seasons, they can be re-determined based on corresponding long-term observation results.

[0044] like Figure 3 As shown, for mixed aerosols obtained through real-time observation, the particle linear depolarization ratio is used. and the linear debiasing ratio of prior particles For the total backscattering coefficient Component separation is performed. Taking a mixture of dust and non-dust components as an example, let the backscattering coefficient of the dust aerosol be... The backscattering coefficient of non-dust aerosols is The prior values ​​of the linear debiasing ratios of dust and non-dust aerosols are respectively... and ,and The backscattering coefficient of dust aerosols... Calculate using the following formula: ; Non-dust aerosol backscattering coefficient Calculate using the following formula: ; Among them, when At that time, it was assumed that the contribution of dust in the observed volume was negligible, so... , ;when At that time, it was assumed that the backscattering in the observed volume mainly came from dust aerosols, leading to... , The aforementioned threshold constraint can prevent negative values ​​of the backscattering coefficient or non-physical results exceeding the total backscattering coefficient due to prior value bias or observation noise.

[0045] For complex aerosol environments with three or more components, the high depolarization ratio component can be separated into two components first to obtain the backscattering coefficient of the first component and the backscattering coefficient of the remaining mixed components. Then, the remaining mixed components can be separated step by step using the same depolarization ratio separation equation until the backscattering coefficient of each aerosol component is obtained.

[0046] After obtaining the backscattering coefficients of each aerosol component, the extinction coefficient is calculated based on the prior radar ratio value for the corresponding aerosol type. Taking dust and non-dust components as an example, the total aerosol extinction coefficient... The expression is: ; in, The radar ratio prior value for dust aerosols. This is the prior value for radar ratio of non-dust aerosols. For N The total aerosol extinction coefficient can be written as: ; In this embodiment, the method of the present invention is used to process hyperspectral resolution lidar observation data, and the results are compared with conventional extinction inversion results without overlap factor calibration and dual-field-of-view hyperspectral resolution lidar inversion results. Conventional extinction inversion without overlap factor calibration has obvious ineffective inversion regions near the ground, while the method of the present invention can continuously output aerosol extinction coefficients in the overlapping transition region where the lidar still has effective backscattering signals, thereby significantly reducing the lower boundary of the effective inversion range.

[0047] Under a specific observation condition, such as Figure 4 As shown, the lower boundary of the blind zone of the hyperspectral resolution lidar system used is approximately 50 m, while the lower boundary of the effective inversion range of traditional uncalibrated extinction inversion is approximately above 1 km. The extinction coefficient profile obtained using dual-field-of-view calibration limits the blind zone to below 300 meters. However, the method of this invention can extend the effective inversion range of aerosol extinction coefficients to approximately 50 m. The consistency of the inversion results obtained by the method of this invention was verified using long-term observation data measured by the dual-field-of-view hyperspectral resolution lidar system from December 2021 to April 2022. The dataset contains 4,258 measurement points. The aerosol extinction coefficients inverted by the dual-field-of-view method and the results obtained based on the method in this paper show high consistency, with a fitting slope of 0.98, a relative deviation of 0.01, and a root mean square error (RMSE) of 0.03. R 2 It reached 0.95.

[0048] This invention is not limited to high spectral resolution lidar. When using other lidar systems capable of obtaining aerosol backscattering coefficients that are insensitive to overlap factors, such as Raman lidar, the long-term prior database construction, aerosol component separation, and extinction coefficient calculation steps of this invention can also be used to extend the effective inversion range of aerosol extinction coefficients without performing overlap factor calibration.

[0049] The embodiments described above provide a detailed explanation of the technical solutions and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, additions, and equivalent substitutions made within the scope of the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An aerosol extinction coefficient retrieval method without the need for an overlap factor calibration, characterized in that, Includes the following steps: (1) Adopting laser radar to obtain multi-channel echo signals of the measured atmosphere and to pre-process, to obtain aerosol optical characteristic parameters of the measured atmosphere, including aerosol backscattering coefficient , particle linear depolarization ratio and laser radar ratio ; the laser radar is hyperspectral resolution laser radar or Raman laser radar; (2) Establish an aerosol optical property prior database, and cluster aerosols according to the lidar ratio and particle linear depolarization ratio in the prior database to obtain prior values of the lidar ratio and the particle linear depolarization ratio corresponding to different types of aerosols and particle linear depolarization ratio ; (3) Based on the aerosol optical property parameters obtained in step (1) and the aerosol clustering results obtained in step (2), determine the aerosol mixing type in the atmosphere to be tested; (4) Using the aerosol backscattering coefficient obtained in step (1) And particle linear debias ratio Combined with the prior value of the particle linear debiasing ratio obtained in step (2) The total backscattering coefficient of the mixed aerosol was separated into components to obtain the backscattering coefficients of different components of the aerosol. ; (5) The backscattering coefficients of each aerosol component obtained in step (4) Compare with the corresponding radar prior value Multiplying these components yields the extinction coefficients of each aerosol component. The extinction coefficients of each aerosol component were also analyzed. Summing yields the aerosol extinction coefficient of the atmosphere under test, including the overlap factor transition region. ; (6) Output the profile or spatiotemporal distribution of aerosol extinction coefficient, thereby expanding the effective inversion range of aerosol extinction coefficient without calibration of overlap factor.

2. The aerosol extinction coefficient inversion method without overlap factor calibration according to claim 1, characterized in that, When the lidar is a high-spectral-resolution lidar, the multi-channel echo signal includes two polarization channel signals and one molecular channel signal; the two polarization channel signals include a mixed vertical polarization channel signal and a mixed parallel polarization channel signal; the molecular channel signal is the molecular elastic backscattering signal retained after the aerosol mi scattering components have been filtered out or suppressed by the spectral discriminator.

3. The aerosol extinction coefficient inversion method without overlap factor calibration according to claim 1, characterized in that, In step (2), Mahalanobis distance is used to classify aerosol samples when performing aerosol clustering.

4. The aerosol extinction coefficient inversion method without overlap factor calibration according to claim 1, characterized in that, In step (4), when the mixed aerosol is composed of a first aerosol component and a second aerosol component, and the prior value of the linear debiasing ratio of the particles of the first aerosol component is... The prior value of the linear debiasing ratio of particles greater than that of the second aerosol component At that time, the backscattering coefficient of the first aerosol component Calculate using the following formula: ； Backscattering coefficient of the second aerosol component Calculate using the following formula: 。 5. The aerosol extinction coefficient inversion method without overlap factor calibration according to claim 4, characterized in that, when season , ;when season , .

6. The aerosol extinction coefficient inversion method without overlap factor calibration according to claim 4, characterized in that, In step (4), when the mixed aerosol contains three or more aerosol components, the total backscattering coefficient is separated into two components in order of the prior value of the linear depolarization ratio of the particles from high to low or from low to high. Each additional aerosol type is separated into one additional separation step until the backscattering coefficient of each aerosol component is obtained.

7. The aerosol extinction coefficient inversion method without overlap factor calibration according to claim 1, characterized in that, In step (5), the extinction coefficient of each aerosol component and total aerosol extinction coefficient Calculate using the following formula: ； ； In the formula, For the first Radar ratio prior values ​​for various aerosol components For the first The backscattering coefficient of aerosol components.

Images