A direct assimilation method and system for aerosol extinction backscatter coefficient

Abstract

The application relates to a direct assimilation method and system of aerosol attenuation backscattering coefficient, and relates to the technical field of data assimilation of atmospheric aerosol. The method comprises the following steps: based on a laser radar equation, a nonlinear observation operator for directly assimilating aerosol attenuation backscattering coefficient is constructed by using Mie scattering theory; according to a three-dimensional variation assimilation theory and the nonlinear observation operator, an assimilation model for solving a minimum solution of a target function is constructed; laser radar aerosol observation data for assimilation are preprocessed, and the preprocessed observation data are input into the assimilation model; the assimilation model is run, the minimum solution obtained by solving is taken as an aerosol increment field and is superimposed on a background field to obtain an analysis field of the aerosol. The application realizes direct assimilation of first-level laser radar aerosol observation data, provides a more accurate initial distribution of the aerosol after assimilation, and can obtain more accurate analysis and prediction results of the aerosol.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of data assimilation of atmospheric aerosols, in particular to a direct assimilation method and system of aerosol attenuation backscattering coefficients. BACKGROUND

[0002] Aerosols are tiny solid or liquid particles suspended in the atmosphere, and their sources include natural processes (such as volcanic eruptions, sandstorms, and sea salt) and human activities (such as industrial emissions, vehicle exhaust, and biomass burning). Aerosols have a significant impact on climate, environment, and human health, and are one of the main pollutants causing air quality deterioration.

[0003] Ground-based or space-borne lidar plays a crucial role in aerosol monitoring, as it can provide data on the vertical profile of aerosols and accurately and highly resolvedly obtain the vertical distribution, optical properties, and dynamic changes of aerosols. Lidar obtains vertical information of aerosols by emitting laser pulses and receiving backscattering signals. Due to the light intensity attenuation caused by the scattering and absorption of aerosols and atmospheric molecules during atmospheric transmission, the aerosol attenuation backscattering coefficient at each height layer is directly retrieved from the received signals. According to the radar equation, the aerosol attenuation backscattering coefficient does not introduce additional assumed parameters in the inversion process, has high accuracy, and can be used as primary data to calculate other secondary data, such as aerosol extinction coefficient, backscattering coefficient, and aerosol optical depth (AOD) by introducing the lidar ratio (ratio of backscattering coefficient to extinction coefficient). Therefore, the aerosol attenuation backscattering coefficient, which uses primary data, has more important significance for the study of aerosols.

[0004] Air quality models are important tools for predicting aerosols, but due to the uncertainty of the initial field of aerosols, the prediction of the model has a large error. The purpose of data assimilation is to improve the initial field of the model, that is, to provide a more accurate initial field for the model, so as to improve the analysis and prediction quality of aerosols. Data assimilation technology can integrate multiple observation data, and in addition to introducing conventional aerosol observation data into the model, the processing of aerosol optical property data also has advantages. However, existing research usually assimilates secondary data such as aerosol mass concentration, extinction coefficient, and satellite AOD, and the assimilation method is relatively simple, but compared with the more accurate aerosol attenuation backscattering coefficient, there is still a certain deficiency in the accuracy of aerosol analysis and prediction. SUMMARY

[0005] Therefore, it is necessary to provide a direct assimilation method and system of aerosol attenuation backscattering coefficients to improve the accuracy of aerosol analysis and prediction.

[0006] A direct assimilation method of aerosol attenuation backscattering coefficient, the method comprising:

[0007] Based on the laser radar equation, a nonlinear observation operator of direct assimilation of aerosol attenuation backscattering coefficient is constructed by using Mie scattering theory;

[0008] According to the three-dimensional variational assimilation theory and the nonlinear observation operator, an assimilation model of aerosol attenuation backscattering coefficient for solving the minimum solution of the objective function is constructed;

[0009] The assimilated lidar aerosol observation data are preprocessed, and the preprocessed observation data are input into the assimilation model;

[0010] The assimilation model is run, and the minimum solution obtained is taken as the aerosol increment field and superimposed on the background field to obtain the analysis field of the aerosol.

[0011] In one of the embodiments, the laser radar equation is represented as:

[0012] ;

[0013] Wherein, is the return signal received by the laser radar, is the laser radar constant, is the distance between the target aerosol and the laser radar, is the radius of the aerosol particle, is the backscattering coefficient, is the extinction coefficient, subscript and represent the atmospheric molecules and aerosols respectively; the laser is exponentially attenuated by the atmospheric molecules and aerosols during transmission, and this attenuation process corresponds to the exp function term in the laser radar equation, and the remaining terms on the right side of the laser radar equation except the coefficient The remaining terms on the right side of the laser radar equation except the coefficient The atmospheric molecule term with subscript is calculated by an empirical formula, and the value of the atmospheric molecule term is much smaller than the aerosol term with subscript Therefore, the aerosol attenuation backscattering coefficient mainly reflects the characteristics of the aerosol.

[0014] In one of the embodiments, the nonlinear observation operator includes two calculation processes, including: first, calculating the aerosol attenuation backscattering coefficient at each grid point in the simulation area according to the aerosol control variable; then interpolating the aerosol attenuation backscattering coefficient at the grid point to the actual observation position, comparing with the observation value, and calculating the observation increment; wherein, the process of calculating the aerosol attenuation backscattering coefficient from the aerosol control variable includes:

[0015] the average wet radius and the average complex refractive index of the aerosol particles in the aerosol of the certain size section are calculated;

[0016] The size parameters of the aerosol particles are calculated by using the incident wavelength and the average wet radius of the lidar, and according to the Mie scattering theory, the size parameters and the average complex refractive index of the aerosol particles are inputted to calculate and output the aerosol extinction efficiency and the aerosol backscattering efficiency;

[0017] The aerosol extinction coefficient and the aerosol backscattering coefficient are obtained by calculation from the aerosol extinction efficiency, the aerosol backscattering efficiency and the wet particle number concentration of the aerosol;

[0018] The atmospheric molecule backscattering coefficient and the atmospheric molecule extinction coefficient are further calculated according to the incident wavelength and by using an empirical formula, and the aerosol attenuation backscattering coefficient is constructed by comprehensively considering the atmospheric molecule backscattering coefficient, the atmospheric molecule extinction coefficient, the aerosol backscattering coefficient and the aerosol extinction coefficient.

[0019] In one of the embodiments, the average wet radius and the average complex refractive index of the aerosol particles in the aerosol of the certain size section are calculated, including:

[0020] It is assumed that each aerosol particle and water substance are internally mixed, the total volume of the aerosol of the certain size section is obtained by volume summation, the average volume of the aerosol particles is obtained by dividing the total volume of the aerosol by the dry particle number concentration of the aerosol, and the average wet radius of the aerosol particles corresponding to the certain size section is calculated from the average volume according to the volume calculation formula of a sphere, on the assumption that the aerosol particles are spherical particles ; wherein the subscript i represents the size section number;

[0021] The average complex refractive index of the aerosol particles corresponding to the certain size section is calculated by using the volume weighting method according to the complex refractive index of each aerosol particle .

[0022] In one of the embodiments, the size parameters of the aerosol particles are calculated by using the incident wavelength and the average wet radius of the lidar, and according to the Mie scattering theory, the size parameters and the average complex refractive index of the aerosol particles are inputted to calculate and output the aerosol extinction efficiency and the aerosol backscattering efficiency, including:

[0023] The size parameters of the aerosol particles are calculated by using the incident wavelength and the average wet radius of the lidar, and are represented as ;

[0024] According to the Mie scattering theory, the size parameters of the aerosol particles are inputted and average complex refractive index and the output aerosol extinction efficiency is calculated by using a polynomial fitting method and aerosol backscattering efficiency .

[0025] In one embodiment, the aerosol extinction coefficient and the aerosol backscattering coefficient are calculated from the aerosol extinction efficiency, the aerosol backscattering efficiency and the wet particle number concentration of the aerosol, including:

[0026] The aerosol extinction coefficient and the aerosol backscattering coefficient are calculated from the aerosol extinction efficiency , the aerosol backscattering efficiency and the wet particle number concentration of the aerosol, and are respectively expressed as:

[0027] ;

[0028] ;

[0029] wherein subscript i represents the particle size segment number, the aerosol has four particle size segments in total, and the wet particle number concentration of the aerosol in a certain particle size segment.

[0030] In one embodiment, according to the three-dimensional variational assimilation theory and the nonlinear observation operator, an assimilation model for solving the minimum solution of the objective function of the aerosol attenuation backscattering coefficient is constructed, including:

[0031] According to the three-dimensional variational assimilation theory and the nonlinear observation operator constructed based on the Mie scattering theory, an assimilation model for the aerosol attenuation backscattering coefficient is constructed, and the purpose of the assimilation model is to solve the minimum solution of an objective function determined by the measurement error and the model error. The incremental form of the objective function is:

[0032] ;

[0033] wherein, represents the optimization target, is the increment of the aerosol control variable in the assimilation model, is the observation increment; and are the background error covariance matrix and the observation error covariance matrix, respectively; ​​a nonlinear observation operator of the direct assimilation aerosol extinction backscatter coefficient is constructed based on the Mie scattering theory, which is used to convert the aerosol control variables into observation values and compare them with the actual observation values, so as to adjust the aerosol control variables; the superscript T denotes transposition.

[0034] In one of the embodiments, the aerosol control variables are designed based on the multi-species multi-particle size section aerosol scheme MOSAIC in the air quality model WRF-Chem, and the total number of the aerosol control variables is 20, including the mass concentrations of 5 kinds of aerosol particles in 4 particle size sections; wherein the 5 kinds of aerosol particles include three independent species of black carbon, organic carbon and unclassified other inorganic salt, a combined species composed of sulfate, nitrate and ammonium salt, and a combined species composed of chloride and sodium salt.

[0035] In one of the embodiments, the assimilated lidar aerosol observation data are preprocessed, and the preprocessed observation data are input into the assimilation model, including:

[0036] The assimilated lidar aerosol observation data are subjected to quality control, including extreme value control, outlier removal, data thinning and noise reduction processing;

[0037] After the quality control, the data format of the observation data is further converted into a format matched with the assimilation model, and the observation data after the format conversion are input into the assimilation model.

[0038] A direct assimilation system of aerosol extinction backscatter coefficient, the system comprising:

[0039] An observation operator construction module is configured to construct a nonlinear observation operator of the direct assimilation aerosol extinction backscatter coefficient based on the lidar equation and the Mie scattering theory;

[0040] An assimilation model construction module is configured to construct an assimilation model of the aerosol extinction backscatter coefficient for solving the minimum solution of the objective function according to the three-dimensional variational assimilation theory and the nonlinear observation operator;

[0041] A data processing module is configured to preprocess the assimilated lidar aerosol observation data and input the preprocessed observation data into the assimilation model;

[0042] An assimilation analysis module is configured to run the assimilation model, take the minimum solution obtained by solving as an aerosol increment field and superimpose it on a background field to obtain an analysis field of the aerosol.

[0043] The direct assimilation method and system of aerosol attenuation backscattering coefficient described above, on the basis of the laser radar equation, first constructs a nonlinear observation operator of direct assimilation of aerosol attenuation backscattering coefficient based on Mie scattering theory, and then constructs an assimilation model of aerosol attenuation backscattering coefficient, wherein the aerosol attenuation backscattering coefficient is directly retrieved from the radar receiving signal and can be used as first-level data to retrieve other second-level data, and has high accuracy. Compared with other methods of assimilating aerosol second-level data, the aerosol observation data is reasonably applied to the air quality model in the present application, which plays a role in optimizing the initial field of the model, provides a more accurate aerosol initial distribution after assimilation, and can obtain more accurate aerosol analysis and prediction results. BRIEF DESCRIPTION OF DRAWINGS

[0044] Figure 1 It is a flowchart of a direct assimilation method of aerosol attenuation backscattering coefficient in one embodiment;

[0045] Figure 2 It is a flowchart of constructing aerosol attenuation backscattering coefficient in one embodiment;

[0046] Figure 3 It is a schematic diagram of the background field, the analysis field simulated aerosol attenuation backscattering coefficient profile and the increment field generated by assimilation in a simulation region at different times in one embodiment; wherein, Figure 3 (a) is the aerosol attenuation backscattering coefficient profile simulated by the background field at time one; Figure 3 (b) is the aerosol attenuation backscattering coefficient profile simulated by the analysis field at time one; Figure 3 (c) is the increment field generated by assimilation at time one; Figure 3 (d) is the aerosol attenuation backscattering coefficient profile simulated by the background field at time two; Figure 3 (e) is the aerosol attenuation backscattering coefficient profile simulated by the analysis field at time two; Figure 3 (f) is the increment field generated by assimilation at time two;

[0047] Figure 4 It is a schematic diagram of the background field and the analysis field simulated attenuation backscattering coefficient profile at any three points in a simulation region and the corresponding observation profile in one embodiment; wherein, Figure 4 (a) is the profile schematic diagram at the first point; Figure 4 (b) is the profile schematic diagram at the second point; Figure 4 (c) is the profile schematic diagram at the third point;

[0048] Figure 5 It is a schematic diagram of the background field, the analysis field simulated extinction coefficient profile and the increment field generated by assimilation in a simulation region at time one in one embodiment; wherein, Figure 5(a) the extinction coefficient profile simulated for the background field, Figure 5 (b) the extinction coefficient profile simulated for the analysis field, Figure 5 (c) the increment field generated by assimilation;

[0049] Figure 6 Fig. 1 is a schematic diagram of the AOD and statistical indicators simulated in the background field and the analysis field at time one and time two in an embodiment;

[0050] Figure 7 Fig. 2 is a schematic diagram of the variation of the prediction result and the observed value of the 24-hour prediction of the average mass concentration of PM2.5 with the prediction time, in which the time two is taken as the initial time in an embodiment. DETAILED DESCRIPTION

[0051] In order to make the objects, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not used to limit the present application.

[0052] In an embodiment, as shown in Fig. 1, a direct assimilation method of aerosol extinction backscatter coefficient is provided, which comprises the following steps: Figure 1

[0053] Step S1: Based on the laser radar equation, a nonlinear observation operator for direct assimilation of aerosol extinction backscatter coefficient is constructed by using Mie scattering theory.

[0054] The laser radar equation is expressed as:

[0055] ;

[0056] wherein, is the return signal received by the laser radar, is the laser radar constant, is the distance between the target aerosol and the laser radar, and generally the laser is emitted vertically upward from the ground, so that is the vertical height from the emission point; is the radius of the aerosol particle, is the backscatter coefficient, is the extinction coefficient, and the subscripts and respectively represent the atmospheric molecules and the aerosol; the laser is exponentially attenuated by the atmospheric molecules and the aerosol during the transmission process, and the attenuation process corresponds to the exp function term in the laser radar equation, and the remaining terms on the right side of the laser radar equation except the coefficient is the aerosol extinction backscatter coefficient, wherein the subscript of the aerosol extinction backscatter coefficient is ​The atmospheric molecule term is calculated by an empirical formula, and the value of the atmospheric molecule term is much smaller than the aerosol term with subscript , so the aerosol attenuation backscattering coefficient mainly reflects the characteristics of the aerosol.

[0057] The nonlinear observation operator includes two calculation processes, including: first, the aerosol attenuation backscattering coefficient at each grid point in the simulation area is calculated according to the aerosol control variable; then the aerosol attenuation backscattering coefficient at the grid point is interpolated to the actual observation position, compared with the observation value, and the observation increment is calculated. The process of calculating the aerosol attenuation backscattering coefficient from the aerosol control variable is shown in Figure 2 , including the following steps:

[0058] The aerosol control variable, the aerosol water content and the aerosol dry particle number concentration are obtained. First, it is assumed that each aerosol particle and water substance is internally mixed, and the total volume of the aerosol in a certain size range is obtained by volume summation. The average volume of the aerosol particles is obtained by dividing the total volume of the aerosol by the dry particle number concentration of the aerosol. Further, it is assumed that the aerosol particles are spherical particles, and the average wet radius of the aerosol particles corresponding to a certain size range is calculated according to the spherical volume calculation formula ; wherein subscript i represents the size range number.

[0059] Then, according to the complex refractive index of each aerosol particle, the volume weighting method is used to calculate the average complex refractive index of the aerosol particles corresponding to a certain size range .

[0060] Then, the incident wavelength of the laser radar and the average wet radius are used to calculate the size parameter of the aerosol particles, which is represented as . According to the Mie scattering theory, the size parameter and the average complex refractive index of the aerosol particles are input, and the calculation efficiency is high. The polynomial fitting method is used to calculate the output aerosol extinction efficiency and the aerosol backscattering efficiency . For example, the aerosol extinction efficiency is represented as:

[0061] ;

[0062] wherein is the normalized logarithmic value of the average wet radius , is the i order Chebyshev polynomial, and , is the total order of Chebyshev polynomials; is the expansion coefficient, and the value is determined by the average complex refractive index and the incident wavelength .

[0063] Then, the aerosol extinction coefficient and the aerosol backscattering coefficient are calculated by the aerosol extinction efficiency , the aerosol backscattering efficiency and the wet particle number concentration of the aerosol, which are expressed as:

[0064] ;

[0065] ;

[0066] wherein the subscript i represents the particle size segment number, indicates that the aerosol has four particle size segments in common, is the wet particle number concentration of the aerosol in a certain particle size segment.

[0067] Finally, the atmospheric molecular backscattering coefficient and the atmospheric molecular extinction coefficient are calculated according to the incident wavelength and using an empirical formula, and the aerosol attenuation backscattering coefficient is constructed by comprehensively considering the atmospheric molecular backscattering coefficient, the atmospheric molecular extinction coefficient, the aerosol backscattering coefficient and the aerosol extinction coefficient.

[0068] Step S2: According to the three-dimensional variational assimilation theory and the nonlinear observation operator, an assimilation model of the aerosol attenuation backscattering coefficient for solving the minimum solution of the objective function is constructed.

[0069] Specifically, according to the three-dimensional variational assimilation theory and the nonlinear observation operator constructed based on the Mie scattering theory, an assimilation model of the aerosol attenuation backscattering coefficient is constructed, and the purpose of the assimilation model is to solve the minimum solution of an objective function determined by the measurement error and the model error. The incremental form of the objective function is:

[0070] ;

[0071] wherein, represents the optimization target, is the increment of the aerosol control variable in the assimilation model, is the observation increment, and are both one-dimensional vectors, but have different lengths. The length N of The length K depends on the number of effective observation data, i.e., the number of aerosol extinction backscatter coefficients input into the assimilation model; and are the background error covariance matrix and the observation error covariance matrix, respectively, and and are N×N and K×K matrices, respectively, representing the proportion of the background value and the observation value in the analysis value; is a nonlinear observation operator of the direct assimilation of the aerosol extinction backscatter coefficient constructed by using the Mie scattering theory, which functions to convert the aerosol control variable into an observation and compare it with the actual observation value, so as to adjust the aerosol control variable; the superscript T represents transposition.

[0072] The aerosol control variable is designed based on the multi-species and multi-particle size section aerosol scheme MOSAIC in the air quality model WRF-Chem, and the total number of the aerosol control variable is 20, including the mass concentration of 5 kinds of aerosol particles in 4 particle size sections; the 5 kinds of aerosol particles include three independent species of black carbon (BC), organic carbon (EC) and other unclassified inorganic salts (OIN), a combined species composed of sulfate (SO4), nitrate (NO3) and ammonium (NH4), and a combined species composed of chloride (CL) and sodium salt (NA).

[0073] Further, in order to obtain the minimum solution of the objective function, it is also necessary to calculate the gradient of the objective function by using the adjoint operator. That is, after completing the programming of the forward calculation process from the aerosol control variable to the observation, the calculation code of the adjoint program is also constructed by using the principle of adjoint programming.

[0074] Step S3: obtaining the assimilated laser radar aerosol observation data for preprocessing, and inputting the preprocessed observation data into the assimilation model.

[0075] Specifically, in the embodiment, the CALIPSO satellite-borne radar data is selected, and the aerosol extinction backscatter coefficient is downloaded as the first-level data. Since the data has large noise, quality control is first performed, including extreme value control, abnormal value elimination, data dilution and noise reduction processing; then the data format is processed into a format matched with the assimilation model, specifically, the satellite data is converted from the HDF (hierarchical data) format to the text format.

[0076] In addition, according to the needs of assimilation effect verification, the aerosol hourly mass concentration of the national control site can also be collected, and the secondary data of the CALIPSO satellite such as the extinction coefficient, the backscatter coefficient and the AOD observation data are downloaded.

[0077] Step S4: running assimilation model, taking the minimum solution as the increment field of aerosol and superimposing it on the background field to obtain the analysis field of aerosol.

[0078] The core content of the present application is to construct an assimilation model, which is written in Fortran 90 language, compiled and run on a Linux server. The model contains five directories, wherein the first directory is a source program directory (source) for storing Fortran 90 source code program files. The second directory is an executable program directory (bin) for storing the executable file da.exe generated after compilation and linking, and there is also a text format parameter file da_files.in in it, which is used to record the absolute path of each input and output file. The third directory is a data directory (data) for storing assimilated observation data files and background error covariance files, and the observation data is the aerosol attenuation backscatter coefficient. The fourth directory stores aerosol background field data (background). The fifth directory is an analysis directory (analysis) for storing the increment file (netcdf format) of the aerosol control variable generated by the executable file da.exe, and the subdirectory dx2wrf in it stores the executable file dx2wrf.exe, which superimposes the increment field on the background field to generate the analysis field. The last analysis field is placed in the analysis directory. Running the assimilation model requires the following steps:

[0079] First step: compiling the program. A Makefile file is written to automatically compile each source code file. Use the make command in the command line to compile, and the executable file da.exe is generated in the bin directory after compilation. In the same way, the executable file dx2wrf.exe is generated in the analysis directory.

[0080] Second step: modifying the parameter file da_files.in. Modify the da_files.in file according to the path of the input and output files.

[0081] Third step: running da.exe. Write a shell script to submit to the background operation, and generate the increment field file after the running is completed.

[0082] Fourth step: running dx2wrf.exe to generate the analysis field file. When running dx2wrf.exe, input 3 parameters in the command line, in the format of. / dx2wrf.exe parameter1 parameter2 parameter3, parameter1 is the background field file directory, parameter2 is the analysis field file directory, and parameter3 is the increment field file directory.

[0083] The direct assimilation method of the aerosol extinction backscatter coefficient first constructs a nonlinear observation operator and assimilation model of the direct assimilation of the aerosol extinction backscatter coefficient based on Mie scattering theory, and realizes the direct assimilation of the first-level observation data of the laser radar aerosol. The aerosol extinction backscatter coefficient is directly retrieved from the radar receiving signal, and can be used as the first-level data to retrieve other second-level data, has high accuracy, compared with other methods of assimilating aerosol second-level data, the application realizes the direct assimilation of the first-level data, effectively integrates the radar observation data into the air quality model, improves the analysis accuracy of the aerosol and has important application value for predicting the change of the aerosol, and can provide scientific suggestions for regional air pollution analysis, early warning and governance.

[0084] In a specific embodiment, taking an aerosol pollution process occurring in a simulation region at time one and time two as the object, the beneficial effects of the method in the aerosol data assimilation and prediction are verified, and the specific implementation steps include:

[0085] Step 1: Compile the assimilation model. Including the aerosol control variable, the observation operator, the forward calculation process and the adjoint process. The program is stable and correct through repeated tests in the early stage. For the two pollution cases of time one and time two, WRF-Chem is set to double grid, and the grid resolution is 27km and 9km respectively, and the innermost layer region covers the simulation region. The parameters of the assimilation model are set, and then the Makefile file is used to compile the program, and after the compilation is successful, the executable program file is generated.

[0086] Step 2: Prepare the observation data for assimilation. Collect two CALIPSO satellite observation cases in the simulation region, that is, the aerosol extinction backscatter coefficient profiles obtained by scanning at time one and time two, and perform quality control and format conversion processing on the collected observation data, and accordingly collect various observation data required for verifying the assimilation effect.

[0087] Step 3: Run the assimilation model to generate the increment field. By inputting the prepared aerosol extinction backscatter coefficient observation data into the assimilation model, the increment field is generated after the running is completed.

[0088] Step 4: Superimpose the increment field on the background field to obtain the analysis field of the aerosol. By cold starting WRF-Chem, two background fields of time one and time two are output, the increment field generated by the assimilation model is superimposed on the background field, and the analysis field is obtained.

[0089] The background field does not assimilate the observation data, in order to verify the assimilation effect, the simulated aerosol extinction backscatter coefficient, extinction coefficient and AOD in the background field and the analysis field are directly compared, and the comparison results are as shown in Figures 3 to 6 Figure 3 ​The time one and the time two are respectively displayed, the aerosol extinction backscattering coefficient profile simulated by the background field simulation in the simulation area, the aerosol extinction backscattering coefficient profile simulated by the analysis field simulation and the increment field generated by assimilation are displayed. Figure 3 It can be seen that compared with the observation value of the aerosol extinction backscattering coefficient, the simulation value of the aerosol extinction backscattering coefficient of the analysis field is closer to the observation value, which indicates that the analysis field assimilated by the assimilation model of the application can effectively simulate the analysis of the aerosol extinction backscattering coefficient. Figure 4 The background field and the analysis field simulated extinction backscattering coefficient profiles at optional three points in the simulation area and the corresponding observation profiles are displayed, Figure 4 The middle blue line represents the extinction backscattering coefficient profile simulated by the background field, the red line represents the extinction backscattering coefficient profile simulated by the analysis field after assimilation, and the green line represents the observation profile. Figure 4 It is not difficult to see that the extinction backscattering coefficient profile simulated by the analysis field after assimilation is closer to the observation value, which indicates that the assimilation plays a significant improvement effect. Figure 5 The extinction coefficient profile simulated by the background field, the analysis field and the increment field generated by assimilation in the simulation area at time one are respectively displayed, Figure 6 The AOD simulated by the background field and the analysis field and the statistical index at time one and time two are respectively displayed, Figure 6 The middle blue dot represents the control test, and no observation data is assimilated, Figure 6 The middle red dot represents the forecast after assimilating the aerosol extinction backscattering coefficient, the solid line is the 1:1 line, indicating that the simulation value is equal to the observation value, the dashed line corresponds to the 1:2 and 2:1 ratio line, and CORR, RMSE and BIAS respectively represent the correlation coefficient, the root mean square error and the average deviation between the simulation value and the observation value. From Figure 5 And Figure 6 It can also be seen that the analysis field assimilated by the assimilation model of the application can effectively simulate the analysis of the aerosol extinction coefficient and the AOD.

[0090] Meanwhile, the 24-hour forecast of PM2.5 is carried out by taking the background field and the analysis field as the initial field respectively, and the effect of assimilation on the PM2.5 forecast is verified, Figure 7 The variation curve of the forecast result and the observation value with the forecast time of the 24-hour forecast of the average mass concentration of PM2.5 by taking time two as the initial time is displayed, Figure 7 The middle blue line represents the control test, and no observation data is assimilated in the initial field of the mode, and the red line represents the forecast after assimilating the aerosol extinction backscattering coefficient, which is obviously close to the black curve representing the observation value. With the extension of time, the red line coincides with the blue line, the improvement effect of assimilation weakens until it disappears, and overall the improvement effect of assimilation on the PM2.5 forecast can last for more than 12 hours.

[0091] In one embodiment, a direct assimilation system of aerosol attenuation backscattering coefficient is provided, comprising:

[0092] An observation operator construction module is configured to construct a nonlinear observation operator of direct assimilation of aerosol attenuation backscattering coefficient based on a laser radar equation and Mie scattering theory;

[0093] An assimilation model construction module is configured to construct an assimilation model of aerosol attenuation backscattering coefficient for solving a minimum solution of a target function according to a three-dimensional variation assimilation theory and the nonlinear observation operator;

[0094] A data processing module is configured to preprocess assimilated laser radar aerosol observation data and input the preprocessed observation data into the assimilation model;

[0095] An assimilation analysis module is configured to run the assimilation model, take the solved minimum solution as an aerosol increment field, superimpose the aerosol increment field on a background field, and obtain an analysis field of the aerosol.

[0096] The specific limitations of the direct assimilation system of aerosol attenuation backscattering coefficient can refer to the limitations of the direct assimilation method of aerosol attenuation backscattering coefficient described above, which will not be repeated here. Each module in the direct assimilation system of aerosol attenuation backscattering coefficient described above can be realized by software, hardware, and combinations thereof, in whole or in part. The above modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so as to be called and executed by the processor to perform the operations corresponding to each module.

[0097] Each technical feature of the above embodiments can be combined arbitrarily. To make the description concise, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combinations of the technical features do not exist contradictions, they should be considered as the scope of the present application.

[0098] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the present application. It should be noted that for those skilled in the art, without departing from the concept of the present application, some modifications and improvements can be made, which are within the scope of the present application.

Claims

1. A method for directly assimilating the attenuation backscattering coefficient of aerosols, characterized in that, The direct assimilation method includes: Based on the lidar equation, a nonlinear observation operator for directly assimilating the aerosol attenuation backscattering coefficient is constructed using Mie scattering theory. Based on the three-dimensional variational assimilation theory and the nonlinear observation operator, an assimilation model for solving the minimum solution of the objective function and the aerosol attenuation backscattering coefficient is constructed. The assimilated lidar aerosol observation data is acquired, preprocessed, and then the preprocessed assimilated lidar aerosol observation data is input into the assimilation model. Running the assimilation model, the minimum solution obtained is used as the aerosol increment field and superimposed on the background field to obtain the aerosol analysis field; The lidar equation is expressed as: ； in, The returned signal received by the lidar. For lidar constants, It is the distance between the target aerosol and the lidar. Let be the radius of the aerosol particle. It is the backscattering coefficient. It is the extinction coefficient, subscript and Let represent atmospheric molecules and aerosols, respectively. During transmission, the laser light decays exponentially due to the combined effects of atmospheric molecules and aerosols. This exponential decay corresponds to the exp function term in the lidar equation, and the right side of the lidar equation is divided by coefficients. The remaining terms represent the aerosol attenuation backscattering coefficient, where the subscript of the aerosol attenuation backscattering coefficient is... The atmospheric molecular term is calculated using an empirical formula, and its value is much smaller than the subscript . The aerosol term, therefore the aerosol attenuation backscattering coefficient mainly reflects the characteristics of aerosols; The nonlinear observation operator comprises two computational processes: first, calculating the aerosol attenuation backscattering coefficient at each grid point within the simulation region based on the aerosol control variables; then, interpolating the aerosol attenuation backscattering coefficient at the grid points to the actual observation locations, comparing it with the observed values, and calculating the observation increment; wherein, the process of calculating the aerosol attenuation backscattering coefficient from the aerosol control variables includes: Calculate and obtain the average wet radius and average complex refractive index of aerosol particles in a certain particle size range; The size parameters of aerosol particles are calculated using the incident wavelength of the lidar and the average wet radius. Based on the Mie scattering theory, the size parameters and average complex refractive index of the aerosol particles are input, and the aerosol extinction efficiency and aerosol backscattering efficiency are calculated and output. The aerosol extinction coefficient and aerosol backscattering coefficient are calculated from the aerosol extinction efficiency, aerosol backscattering efficiency, and aerosol wet particle number concentration. The atmospheric molecule backscattering coefficient and atmospheric molecule extinction coefficient are further calculated based on the incident wavelength and using empirical formulas. The aerosol attenuated backscattering coefficient is then constructed by combining the atmospheric molecule backscattering coefficient, atmospheric molecule extinction coefficient, aerosol backscattering coefficient, and aerosol extinction coefficient.

2. The direct assimilation method for aerosol attenuation backscattering coefficient according to claim 1, characterized in that, Calculate and obtain the average wet radius and average complex refractive index of aerosol particles in a certain particle size range, including: Assuming that the aerosol particles and water are internally mixed, the total aerosol volume for a certain particle size range is obtained by summing the volumes. Dividing this total aerosol volume by the dry particle number concentration yields the average volume of the aerosol particles. Further assuming the aerosol particles are spherical, the average wet radius of the aerosol particles corresponding to a certain particle size range is calculated from the average volume using the spherical volume formula. ; where subscript i Indicates the particle size range number; Based on the complex refractive index of each aerosol particle, the average complex refractive index of the aerosol particles corresponding to a certain particle size range is calculated using the volume-weighted method. .

3. The direct assimilation method for aerosol attenuation backscattering coefficient according to claim 2, characterized in that, The size parameters of aerosol particles are calculated using the incident wavelength of the lidar and the average wet radius. Based on Mie scattering theory, the size parameters and average complex refractive index of the aerosol particles are input, and the aerosol extinction efficiency and aerosol backscattering efficiency are calculated and output, including: Using the incident wavelength of lidar and average wet radius The size parameters of aerosol particles are calculated and expressed as follows: ; Based on Mie scattering theory, the size parameters of aerosol particles are input. and average complex refractive index The aerosol extinction efficiency was calculated using polynomial fitting. and aerosol backscattering efficiency .

4. The direct assimilation method for aerosol attenuation backscattering coefficient according to claim 3, characterized in that, The aerosol extinction coefficient and aerosol backscattering coefficient are calculated from the aerosol extinction efficiency, aerosol backscattering efficiency, and aerosol wet particle number concentration, including: Extinction efficiency of aerosols Aerosol backscattering efficiency The aerosol extinction coefficient is obtained by calculating the wet particle number concentration of the aerosol. and aerosol backscattering coefficient , respectively represented as: ； ； Among them, subscript i Indicates the particle size range number. This indicates that the aerosol has four particle size ranges. This refers to the wet particle number concentration within a certain particle size range of the aerosol.

5. The direct assimilation method for aerosol attenuation backscattering coefficient according to claim 1, characterized in that, Based on the three-dimensional variational assimilation theory and the aforementioned nonlinear observation operator, an assimilation model for solving the aerosol attenuation backscattering coefficient to obtain the minimum solution of the objective function is constructed, including: Based on the three-dimensional variational assimilation theory and the nonlinear observation operator constructed based on Mie scattering theory, an assimilation model for the aerosol attenuation backscattering coefficient is constructed. The purpose of this assimilation model is to find a minimum solution to an objective function determined by measurement error and mode error. The incremental form of this objective function is as follows: ； in, Indicates the optimization objective. It is the increment of the aerosol control variable in the assimilation model. It is the increment of the observation; and These are the background error covariance matrix and the observation error covariance matrix, respectively. This is a nonlinear observation operator for directly assimilating the aerosol attenuation backscattering coefficient, constructed using Mie scattering theory. Its function is to convert the aerosol control variable into an observation and compare it with the actual observation, thereby adjusting the aerosol control variable. T This indicates transpose.

6. The direct assimilation method for aerosol attenuation backscattering coefficient according to claim 1 or 5, characterized in that, The aerosol control variables are designed based on the MOSAIC multi-species, multi-size aerosol scheme in the WRF-Chem air quality model. There are a total of 20 aerosol control variables, including the mass concentrations of 5 aerosol particles in 4 size ranges. Among them, the 5 aerosol particles include three independent species: black carbon, organic carbon, and other unclassified inorganic salts; a combined species composed of sulfates, nitrates, and ammonium salts; and a combined species composed of chlorides and sodium salts.

7. The direct assimilation method for aerosol attenuation backscattering coefficient according to claim 1, characterized in that, Acquire assimilated lidar aerosol observation data, preprocess it, and input the preprocessed observation data into the assimilation model, including: Quality control is performed on assimilated lidar aerosol observation data, including extreme value control, outlier removal, data dilution, and noise reduction. After quality control, the data format of the observation data is further converted into a format that matches the assimilation model, and the converted observation data is then input into the assimilation model.

8. A direct assimilation system for aerosol attenuation backscattering coefficient, implemented using the direct assimilation method for aerosol attenuation backscattering coefficient as described in any one of claims 1-7, characterized in that, The direct assimilation system includes: The observation operator construction module is used to construct a nonlinear observation operator that directly assimilates the aerosol attenuation backscattering coefficient based on the lidar equation and Mie scattering theory. The assimilation model construction module is used to construct an assimilation model for solving the minimum solution of the objective function by using the three-dimensional variational assimilation theory and the nonlinear observation operator to solve the aerosol attenuation backscattering coefficient. The data processing module is used to acquire assimilated lidar aerosol observation data, preprocess it, and input the preprocessed assimilated lidar aerosol observation data into the assimilation model. The assimilation analysis module is used to run the assimilation model, and the minimum solution obtained is used as the aerosol increment field and superimposed on the background field to obtain the aerosol analysis field.

Images