Networking laser radar data assimilation method based on three-dimensional variation

By employing a three-dimensional variational networked lidar data assimilation method, the uncertainty problem in PM2.5 forecasting of numerical models was solved, realizing data assimilation from ground-based two-dimensional to three-dimensional data, thus improving the accuracy of atmospheric pollutant forecasts.

CN118981584AActive Publication Date: 2024-11-19ANHUI UNIV
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202411058937.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-02
Publication Date
2024-11-19
Estimated Expiration
2044-08-02

AI Technical Summary

Technical Problem

In the current technology, the international mainstream numerical models still have a large degree of uncertainty in the forecast of PM2.5 in various regions of my country, especially during periods of heavy pollution, the forecast deviation is large. This is mainly due to the uncertainty of emission sources, meteorological fields, initial conditions and the model itself, and the lack of three-dimensional detection data leads to insufficient data assimilation accuracy.

Method used

A three-dimensional variational networked lidar data assimilation method is adopted. By performing quality control and preprocessing on the networked lidar data, calculating the background error covariance matrix and objective function value, and updating the background field data using the iterative conjugate gradient method, the data assimilation from two-dimensional to three-dimensional ground data is achieved.

Benefits of technology

It has improved the accuracy and precision of air pollutant forecasts, and achieved data assimilation from two-dimensional to three-dimensional data, maximizing the use of measured data in three-dimensional space to correct model simulation results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118981584A_ABST
    Figure CN118981584A_ABST
Patent Text Reader

Abstract

The invention discloses a networking laser radar data assimilation method based on three-dimensional variation. Comprising the steps of networking laser radar data quality control and data preprocessing; calculating a background covariance matrix; calculating a target function value; calculating a target function gradient value; calculating a minimum value of an objective function; updating background field data, and determining a final data assimilation analysis field as an assimilation result at the current moment; according to the method, the problem that high-level data cannot be provided by ground conventional observation data is effectively solved, the precision and the accuracy of air quality mode simulation are improved, a reanalysis field of particulate matters constrained by three-dimensional observation data is given, the regional heavy pollution process is restored, and the feature rule of spatial and temporal distribution and transportation of regional pollutants is clarified.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the fields of atmospheric environmental science and laser radar, and in particular to a three-dimensional variation-based networked laser radar data assimilation method. Background Art

[0002] Frequent severe air pollution has brought important impacts on human health, atmospheric visibility, climate, etc. Therefore, taking emergency prevention and control measures during severe pollution and issuing early warning and forecast information to the public in a timely manner are key links in dealing with the current severe air pollution problem.

[0003] Numerical forecasting based on air quality numerical models is an important technical method to deal with the current serious haze pollution problem. Numerical models simplify the real atmospheric physical process into an ideal mathematical model, and use meteorological and mathematical methods to simulate horizontally and vertically to describe the process of pollutant transport, reaction, and removal in the atmosphere. However, the current mainstream numerical models in the world still have great uncertainty in the forecast of PM2.5 in various regions of my country, especially during heavy pollution periods, the forecast deviation can reach 30-60%, and the error is mainly due to factors such as emission sources, meteorological fields, initial conditions, and uncertainty of the model itself.

[0004] As an important methodology for integrating models and observations in the Earth System Science Strategy, data assimilation has long been proven to be effective in improving forecast results in the meteorological field. Data assimilation of meteorological data has been put into operation in many meteorological forecast centers at home and abroad. In recent years, data assimilation of atmospheric pollutants has gradually gained attention and development in the field of air quality forecasting. Although data assimilation can improve forecast accuracy to a certain extent, the current data assimilation of atmospheric pollutants is mainly based on the assimilation of the ground layer, lacking data assimilation at the spatial multidimensional level. This is mainly due to the serious lack of three-dimensional detection data of pollutants, the lack of data density, and the insufficient time resolution and spatial resolution to meet the data quality standards of data assimilation. For example, conventional data only have ground detection, while sounding data have low time resolution and low point density. Summary of the invention

[0005] In view of the defects in the prior art, the present invention provides a data assimilation method capable of realizing three-dimensional level.

[0006] The present invention provides a three-dimensional variational based networked lidar data assimilation method, the specific process is as follows:

[0007] A. Perform quality control and preprocessing on the networked lidar data; first query the lidar measured data in the current time period, then check the statistical characteristics of the lidar data to determine whether the data conforms to the statistical law, and finally convert the normal data into the data format required by the stereo assimilation system.

[0008] B. Calculate the background error covariance matrix; the background error covariance matrix is ​​the product of the background error matrix and the background correlation coefficient matrix, where the background error is the difference between the two sets of forecast fields at the same time of two forecasts, and the background correlation coefficient is calculated using the Gaussian model method.

[0009] C. Calculate the objective function value; use the background field data simulated by the model and the measured data of the networked lidar to calculate the objective function value through the three-dimensional variational method.

[0010] D. Calculate the gradient value of the objective function; use the objective function value processed in the previous step to calculate the gradient value of the objective function.

[0011] E. Calculate the minimum value of the objective function; use the objective function value and the minimum value of the objective function calculated in the previous two steps to input the step-by-step iteration method, and calculate the minimum value of the objective function through the iterative conjugate gradient process.

[0012] F, update the background field data; when the objective function of the previous step reaches the minimum state, the corresponding state quantity at this time is the optimal estimate, forming the final analysis field of networked lidar data assimilation.

[0013] Furthermore, the laser radar data quality control specifically includes:

[0014] A. Obtaining networked LiDAR data: Based on the latitude and longitude range area covered by the model, find the number of networked LiDARs in the area, and query all data of the required time period through the database.

[0015] B. LiDAR data quality control: The stiffness check is also called the minimum frequency change check. The stiffness is caused by the failure of the observation instrument and other reasons during the transmission or recording process, which makes the observation value unchanged for a long time, causing the record to be untrue. The detection method is as follows: Z is the average absolute difference of the intensity change of the echo signal, that is, the change of the intensity of the adjacent echo signal data:

[0016]

[0017] r is the correlation coefficient of adjacent radar echo signals, that is, the similarity of the echo waveforms:

[0018]

[0019] Wherein, A and B represent two echo signals respectively.

[0020] Judgment criteria: If continuous data appears, Z between the adjacent data is 0, and its corresponding r is 1. Then the adjacent data is judged to be continuous dead values.

[0021] C. LiDAR data preprocessing: The fine particulate matter concentration (PM2.5) profile obtained by direct inversion of the LiDAR needs to be converted into a file in the Universal Binary Meteorological Data (BUFR) format required by the assimilation system. The following data formats are established: PTID, CLONH, CLATH, CSIGLEV, TPHR, TYPO, and COPOPM. Among them: PTID represents the data ID, which is a self-increasing integer; CLONH and CLATH represent the longitude and latitude information of the current radar point, respectively; CSIGLEV represents the vertical layer information corresponding to the LiDAR data; TPHR represents the offset from the current simulation time; TYPO represents the data type ID number of the assimilation parameter PM2.5, which is currently fixed to 11; COPOPM represents the concentration information of PM2.5 above the current radar at the current spatial position;

[0022] Furthermore, the background error covariance matrix specifically includes:

[0023] A. Calculate the background error matrix: Use two different sets of starting forecast times to forecast the same time in the future, and use the difference between the two sets of resulting forecast fields as the approximate value of the forecast error.

[0024] In this way, the background error covariance matrix B can be expressed as a block diagonal matrix, which is usually obtained by statistics through the NMC (the National Meteorological Center) method (Parish DF, Derber J., 1992). The core of this method is to use the difference between the forecast values ​​of different time periods at the same time as the approximation of the forecast error.

[0025]

[0026] Here, the subscripts i and l represent the grid points in the horizontal and vertical directions, respectively. and It indicates the forecast value of a certain factor at a certain moment with a validity period of 24 hours and 12 hours respectively.

[0027] Among them, the background error covariance matrix of chemical field variables for statistical analysis mainly includes OC1, OC2, BC1, BC2, Sulfate, Sea salt1, and Dust1-2 in the GOCART model.

[0028] B. Calculate the background correlation coefficient matrix: Using the Gaussian model method, define x, x 0 The background field correlation coefficient b(x, x) between two points 0 ) is the distance between two points dist(xx 0 ), the specific calculation formula is as follows

[0029]

[0030] C. Calculate the background error covariance matrix: Multiply the results of step 3.1 and step 3.2 to obtain the background error covariance matrix.

[0031] Furthermore, the calculation of the objective function value specifically includes:

[0032] A, define the objective function:

[0033]

[0034] Among them, X a is the analysis vector, or analysis field, which represents the result of the model simulation after assimilation of the observed data, that is, the final output value of the assimilation system. b It is the forecast field or background field, which represents the unassimilated simulated value of the model and is a set of input values ​​of the assimilation system. o is the observed value, which means the value observed by the instrument, and is also a set of input values ​​for the assimilation system. In addition, B and O in the formula both represent the error covariance matrix, where B is the background error covariance matrix. O is the observation error covariance matrix, which also represents the observation error. H represents the observation operator, which means the transformation method or function from the observation value to the model value. The superscript "T" represents the transposition, and the superscript "-1" represents the inverse matrix.

[0035] The first term on the right side of the equal sign in the above formula is used to constrain the optimal parameters to make the model state as close as possible, and the second term is used to adjust the variables to make the model output as close as possible to the continuous observation value. The third term Jc represents the constant term, such as the changes caused by the kinetic constant, water vapor constant, etc.

[0036] Define the analysis increment (ΔX) X = X a -X b , the objective function is rewritten as:

[0037]

[0038] Assuming H is a linear operator, the objective function is rewritten as:

[0039]

[0040] Define the observation increment as O=O 0 -H Xb , the objective function is finally rewritten as:

[0041]

[0042] Furthermore, the calculation of the minimum value of the objective function specifically includes:

[0043] After determining the objective function of data assimilation, the next task is to minimize the objective function. Minimization means changing the independent variable to minimize the function value. At this time, solving the objective function to minimize it under a series of external constraints is the cost function to be solved, or it is called the loss function or error function.

[0044] The minimum value of the objective function can be determined by step-by-step iteration. Define a new variable y = B -1 X, then (16) becomes:

[0045] J=y T B T y+(HBy-0) T O -1 (HBy-0)+J c (17)

[0046] According to the chain rule in calculus derivation, the gradients of the cost function J for the background field x and the observation field y are:

[0047] ▽ X J=B -1 X+H T O -1 (HX-0) (18)

[0048]

[0049] Equations (6) and (7) are simultaneously minimized by an iterative conjugate gradient procedure. To find the optimal solution, the iterative steps are as follows:

[0050] Starting assumption:

[0051] x 0 =y 0 =0

[0052] Iteration step n:

[0053]

[0054] x n =x n-1 +αDir·x n

[0055] y n =y n-1 +αDir·y n

[0056] The above iterations are terminated until the maximum number of iterations is reached or the two gradients (18) and (19) reach the pre-given minimization conditions. When the objective function reaches the minimum state, the corresponding state quantity is the optimal estimate, forming the final analysis field of networked lidar data assimilation.

[0057] Beneficial Effects

[0058] It can be seen from the above technical scheme that the present invention provides a networked lidar data assimilation method based on three-dimensional variation. Compared with the traditional data assimilation method, the salient feature of the present invention is that based on the networked lidar stereo data, the traditional three-dimensional variational method is expanded from ground two-dimensional to spatial three-dimensional, and the assimilation function of the traditional assimilation system from ground two-dimensional to stereo three-dimensional is realized, and more stereoscopic space measured data is used to correct the model simulation results as much as possible. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is a flow chart of networked lidar data assimilation;

[0060] Figure 2 It is a single point test verification of the assimilation system;

[0061] Figure 3 It is a time series comparison chart of assimilated and non-assimilated results. DETAILED DESCRIPTION

[0062] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0063] Embodiment 1

[0064] Combination Figure 1 , a three-dimensional variational based networked lidar data assimilation method, mainly including the following steps:

[0065] Step 1: Perform quality control and preprocessing on the networked lidar data; first query the lidar measured data in the current time period, then check the statistical characteristics of the lidar data to determine whether the data conforms to statistical laws, and finally convert the normal data into the data format required by the stereo assimilation system.

[0066] Step 2: Calculate the background error covariance matrix; the background error covariance matrix is ​​the product of the background error matrix and the background correlation coefficient matrix, where the background error is the difference between the two sets of forecast fields at the same time of two forecasts, and the background correlation coefficient is calculated using the Gaussian model method.

[0067] Step 3: Calculate the objective function value; use the background field data simulated by the model and the measured data of the networked lidar to calculate the objective function value through the three-dimensional variational method.

[0068] Step 4: Calculate the gradient value of the objective function; use the objective function value processed in the previous step to calculate the gradient value of the objective function.

[0069] Step 5: Calculate the minimum value of the objective function; use the objective function values ​​and the minimum value of the objective function calculated in the previous two steps to input the stepwise iteration method, and calculate the minimum value of the objective function through the iterative conjugate gradient process.

[0070] Step 6: Update the background field data; when the objective function in the previous step reaches the minimum state, the corresponding state quantity is the optimal estimate, forming the final analysis field for networked lidar data assimilation.

[0071] Embodiment 2

[0072] A networked laser radar data assimilation method based on three-dimensional variation, a networked laser radar data assimilation method based on three-dimensional variation, specifically comprising the following steps;

[0073] S1. Perform quality control and preprocessing on the networked lidar data. First, query the actual lidar data in the current time period, then check the statistical characteristics of the lidar data to determine whether the data conforms to the statistical law, and finally convert the normal data into the data format required by the stereo assimilation system.

[0074] S2. Calculate the background error covariance matrix; the background error covariance matrix is ​​the product of the background error matrix and the background correlation coefficient matrix, wherein the background error is the difference between two sets of forecast fields at the same time of two forecasts, and the background correlation coefficient is calculated using the Gaussian model method;

[0075] S3, calculating the objective function value; using the background field data simulated by the model and the measured data of the networked laser radar, the objective function value is calculated by a three-dimensional variational method;

[0076] S4, calculate the gradient value of the objective function; use the objective function value processed in the previous step to calculate the gradient value of the objective function;

[0077] S5, calculating the minimum value of the objective function; using the objective function values ​​and the minimum value of the objective function calculated in the previous two steps to input the stepwise iteration method, and calculating the minimum value of the objective function through an iterative conjugate gradient process;

[0078] S6. Update the background field data; when the objective function in the previous step reaches the minimum state, the corresponding state quantity at this time is the optimal estimate, forming the final analysis field for networked lidar data assimilation.

[0079] In the present invention, the method for quality control and preprocessing of networked laser radar data specifically includes the following steps:

[0080] A1. Obtaining networked LiDAR data: Based on the latitude and longitude range covered by the model, find the number of networked LiDARs in the area, and query all data for the required time period through the database;

[0081] A2. LiDAR data quality control: The stiffness check is also called the minimum frequency change check. The stiffness is caused by the failure of the observation instrument and other reasons in the transmission or recording process, which makes the observation value unchanged for a long time, causing the record to be untrue. The detection method is as follows: Z is the average absolute difference of the intensity change of the echo signal, that is, the change of the intensity of the adjacent echo signal data;

[0082]

[0083] r is the correlation coefficient of adjacent radar echo signals, that is, the similarity of the echo waveforms;

[0084]

[0085] Among them, A and B represent two echo signals respectively;

[0086] Judgment criteria: If continuous data appears, Z between the adjacent data is 0, and its corresponding r is 1, then the adjacent data is judged to be a continuous dead value;

[0087] A3. Preprocessing of LiDAR data: The fine particle concentration (PM2.5) profile obtained by direct inversion of LiDAR needs to be converted into a file in the Universal Binary Meteorological Data (BUFR) format required by the assimilation system. The following data formats are established: PTID, CLONH, CLATH, CSIGLEV, TPHR, TYPO, COPOPM, where: PTID represents the data ID, which is a self-incrementing integer; CLONH and CLATH represent the longitude and latitude information of the current radar point, respectively; CSIGLEV represents the vertical layer information corresponding to the LiDAR data; TPHR represents the offset from the current simulation time; TYPO represents the data type ID number of the assimilation parameter PM2.5, which is currently fixed to 11; COPOPM represents the PM2.5 concentration information of the current radar at the current spatial position.

[0088] In the present invention, the step of calculating the background error covariance matrix processing specifically includes the following steps:

[0089] B1. Calculate the background error matrix: Use two different starting forecast times to forecast the same time in the future, and use the difference between the two forecast fields as the approximate value of the forecast error. The background error covariance matrix of chemical field variables for statistical analysis mainly includes OC1, OC2, BC1, BC2, Sulfate, Sea salt1, Dust1-2 in the GOCART model;

[0090] B2. Calculate the background correlation coefficient matrix: Using the Gaussian model method, define x, x 0 The background field correlation coefficient b(x, x) between two points 0 ) is the distance between two points dist(xx 0 ), the specific calculation formula is as follows;

[0091]

[0092] B3. Calculate the background error covariance matrix: Multiply the results of step 3.1 and step 3.2 to obtain the background error covariance matrix.

[0093] In the present invention, calculating the objective function value specifically includes the following steps:

[0094] C1. Define the objective function:

[0095]

[0096] Among them, X a is the analysis vector, or analysis field, which represents the result of the model simulation after assimilation of the observed data, that is, the final output value of the assimilation system. b is the forecast field or background field, which represents the unassimilated simulated value of the model. It is a set of input values ​​of the assimilation system. o is the observed value, which means the value observed by the instrument, and is also a set of input values ​​of the assimilation system. In addition, B and O in the formula both represent the error covariance matrix, where B is the background error covariance matrix, O is the observation error covariance matrix, that is, it represents the observation error, H represents the observation operator, which means the transformation method or function from the observation value to the model value, the superscript "T" represents the transpose, and the superscript "-1" represents the inverse matrix;

[0097] The first term on the right side of the equal sign in the above formula is used to constrain the optimal parameters to make the model state as close as possible. The second term is used to adjust the variables so that the model output is as close as possible to the continuous observation value. The third term Jc represents the constant term, such as the changes caused by the kinetic constant, water vapor constant, etc.

[0098] C2. Define the analysis increment (ΔX) X = X a -X b, the objective function is rewritten as:

[0099]

[0100] Assuming H is a linear operator, the objective function is rewritten as:

[0101]

[0102] Define the observation increment as O=O 0 -H Xb , the objective function is finally rewritten as:

[0103]

[0104] In the present invention, calculating the minimum value of the objective function specifically includes the following steps:

[0105] D1. After determining the objective function of data assimilation, the next task is to minimize the objective function. Minimization means changing the independent variable to minimize the function value. At this time, solving the objective function to minimize it under a series of external constraints is the cost function, or loss function or error function.

[0106] Solving the minimum value of the objective function can be determined by step-by-step iteration, defining a new variable y = B -1 X, then the formula becomes;

[0107] J=y T B T y+(HBy-0) T O -1 (HBy-0)+J c

[0108] According to the chain rule in calculus derivation, the gradients of the cost function J for the background field x and the observation field y are:

[0109]

[0110] The system of equations is minimized simultaneously by an iterative conjugate gradient process. To find the optimal solution, the following steps are performed:

[0111] Starting assumption:

[0112] x 0 =y 0 =0

[0113] Iteration step n:

[0114]

[0115] x n =x n-1+αDir·x n

[0116] y n =y n-1 +αDir·y n

[0117] The above iterations are repeated until the maximum number of iterations is reached or the two gradients reach the pre-given minimization conditions.

[0118] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions.

[0119] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A three-dimensional variational based networked lidar data assimilation method, characterized in that: Specifically include the following steps: S1. Perform quality control and preprocessing on the networked lidar data. First, query the actual lidar data in the current time period, then check the statistical characteristics of the lidar data to determine whether the data conforms to the statistical law, and finally convert the normal data into the data format required by the stereo assimilation system. S2, calculate the background error covariance matrix; The background error covariance matrix is ​​the product of the background error matrix and the background correlation coefficient matrix, where the background error is the difference between two sets of forecast fields at the same time of two forecasts, and the background correlation coefficient is calculated using the Gaussian model method; S3, calculating the objective function value; using the background field data simulated by the model and the measured data of the networked laser radar, the objective function value is calculated by a three-dimensional variational method; S4, calculate the gradient value of the objective function; use the objective function value processed in the previous step to calculate the gradient value of the objective function; S5. Calculate the minimum value of the objective function; The objective function value and the minimum value of the objective function calculated in the previous two steps are input into the stepwise iteration method, and the minimum value of the objective function is calculated through the iterative conjugate gradient process; S6. Update the background field data; when the objective function in the previous step reaches the minimum state, the corresponding state quantity at this time is the optimal estimate, forming the final analysis field for networked lidar data assimilation.

2. The method for assimilation of networked lidar data based on three-dimensional variation according to claim 1, characterized in that: The method for quality control and preprocessing of networked laser radar data specifically includes the following steps: A1. Obtaining networked LiDAR data: Based on the latitude and longitude range covered by the model, find the number of networked LiDARs in the area, and query all data for the required time period through the database; A2. LiDAR data quality control: The stiffness check is also called the minimum frequency change check. The stiffness is caused by the failure of the observation instrument and other reasons in the transmission or recording process, which makes the observation value unchanged for a long time, causing the record to be untrue. The detection method is as follows: Z is the average absolute difference of the intensity change of the echo signal, that is, the change of the intensity of the adjacent echo signal data; r is the correlation coefficient of adjacent radar echo signals, that is, the similarity of the echo waveforms; Among them, A and B represent two echo signals respectively; Judgment criteria: If continuous data appears, Z between the adjacent data is 0, and its corresponding r is 1, then the adjacent data is judged to be a continuous dead value; A3. Preprocessing of LiDAR data: The fine particle concentration (PM2.5) profile obtained by direct inversion of LiDAR needs to be converted into a file in the Universal Binary Meteorological Data (BUFR) format required by the assimilation system. The following data formats are established: PTID, CLONH, CLATH, CSIGLEV, TPHR, TYPO, COPOPM, where: PTID represents the data ID, which is a self-incrementing integer; CLONH and CLATH represent the longitude and latitude information of the current radar point, respectively; CSIGLEV represents the vertical layer information corresponding to the LiDAR data; TPHR represents the offset from the current simulation time; TYPO represents the data type ID number of the assimilation parameter PM2.5, which is currently fixed to 11; COPOPM represents the PM2.5 concentration information of the current radar at the current spatial position.

3. The method for assimilation of networked lidar data based on three-dimensional variational calculus according to claim 1, characterized in that: The step of calculating the background error covariance matrix processing specifically comprises the following steps: B1. Calculate the background error matrix: Use two different starting forecast times to forecast the same time in the future, and use the difference between the two forecast fields as the approximate value of the forecast error. The background error covariance matrix of chemical field variables for statistical analysis mainly includes OC1, OC2, BC1, BC2, Sulfate, Sea salt1, Dust1-2 in the GOCART model; B2. Calculate the background correlation coefficient matrix: Using the Gaussian model method, define the background field correlation coefficient b(x, x0) between two points x and x0 as a function of the distance dist(x-x0) between the two points. The specific calculation formula is as follows; B3. Calculate the background error covariance matrix: Multiply the results of step 3.1 and step 3.2 to obtain the background error covariance matrix.

4. The method for assimilation of networked lidar data based on three-dimensional variational calculus according to claim 1, characterized in that: The calculating objective function value specifically comprises the following steps: C1. Define the objective function: Among them, X a is the analysis vector, or analysis field, which represents the result of the model simulation after assimilation of the observed data, that is, the final output value of the assimilation system. b is the forecast field or background field, which represents the unassimilated simulated value of the model. It is a set of input values ​​of the assimilation system. o is the observed value, which means the value observed by the instrument, and is also a set of input values ​​of the assimilation system. In addition, B and O in the formula both represent the error covariance matrix, where B is the background error covariance matrix, O is the observation error covariance matrix, which also represents the observation error. H represents the observation operator, which means the transformation method or function from the observation value to the model value. The superscript "T" represents the transpose, and the superscript "-1" represents the inverse matrix. The first term on the right side of the equal sign in the above formula is used to constrain the optimal parameters to make the model state as close as possible. The second term is used to adjust the variables so that the model output is as close as possible to the continuous observation value. The third term Jc represents the constant term, such as the changes caused by the kinetic constant, water vapor constant, etc. C2. Define the analysis increment (ΔX) X = X a -X b , the objective function is rewritten as: Assuming H is a linear operator, the objective function is rewritten as: Define the observation increment as O = O0-H Xb , the objective function is finally rewritten as: 。 5. The method for assimilation of networked lidar data based on three-dimensional variation according to claim 1, characterized in that: The step of calculating the minimum value of the objective function specifically comprises the following steps: D1. After determining the objective function of data assimilation, the next task is to minimize the objective function. Minimization means changing the independent variable to minimize the function value. At this time, solving the objective function to minimize it under a series of external constraints is the cost function, or loss function or error function. Solving the minimum value of the objective function can be determined by step-by-step iteration, defining a new variable y = B -1 X, then the formula becomes; J=y T B T y+(HBy-0) T O -1 (HBy-0)+J c According to the chain rule in calculus derivation, the gradients of the cost function J for the background field x and the observation field y are: The system of equations is minimized simultaneously by an iterative conjugate gradient process. To find the optimal solution, the following steps are performed: Starting assumption: x 0 =and 0 =0 Iteration step n: x n =x n-1 +αDir·x n and n =and n-1 +αDir·y n The above iterations are continued until the maximum number of iterations is reached or the two gradients reach the pre-given minimization condition.

Citation Information

Patent Citations

  • WRF-Chem-mode-based three-dimensional variational assimilation method for aerosol laser radar PM2.5 inversion

    CN109709577A

  • Aerosol optical property data assimilation method and device based on three-dimensional variation technology

    CN114112995A

  • Set three-dimensional variation assimilation method and system for radar reflectivity and humidity correlation estimation

    CN115826091A

  • Weather forecast method and system for severe convection rapid update cyclic assimilation

    CN115857056A

  • Spectral transformation method for three-dimension variational data assimilation of numerical weather prediction model using cubed-sphere grid based on spectral element method and hardware device performing the same

    KR101471378B1