A method and system for inversion of aerosol optical and microphysical parameters

Abstract

The present application discloses an aerosol optical and microphysical parameter inversion method and system, which solves the problem of high uncertainty in the inversion of single multi-angle polarization load parameters in the prior art. An aerosol optical and microphysical parameter inversion method, comprising: obtaining multi-source data; uniformly matching multi-source data according to space and time; determining the initial value of the inversion parameter; substituting the initial value of the inversion parameter into the atmospheric forward radiation transfer model and calculating the first difference with the actual observation value of the multi-source satellite; calculating the second difference between the inversion parameter and the initial value of the inversion parameter; calculating the third difference between the AAE observation value and the simulation value; iterating the inversion parameter to make the sum of the first difference, the second difference and the third difference the minimum value, stop iteration, and obtain the current inversion parameter. The present application breaks through the constraints of a single multi-angle polarization load on the limited detection capability and detection accuracy of aerosol optics and microphysics, and improves the aerosol optical and microphysical inversion accuracy and signal freedom.

Description

Technical Field

[0001] The present application relates to the field of satellite remote sensing technology, and in particular to an aerosol optical and microphysical parameter inversion method and system. Background Art

[0002] Satellites have been used for global / regional aerosol monitoring for more than 40 years, but most satellite sensors (such as AVHRR, MODIS, AHI, ABI, etc.) only obtain mixed information of aerosols (such as aerosol optical thickness), and do not have the ability to detect optical and microphysical parameters such as aerosol particle spectrum distribution, single scattering albedo, and scattering phase function. With the development of satellite multi-angle and polarization detection technology, multi-angle observations have effectively increased satellite observations, and polarization observations are more sensitive to aerosols. By combining polarization observations with multi-spectral and multi-angle observations, more aerosol observation information is provided on the basis of radiation intensity, which provides the possibility for satellite inversion of aerosol optical and microphysical parameters. However, at present, the inversion of aerosol optical and microphysical parameters is mainly based on ground-based remote sensing. The research on aerosol optical and microphysical remote sensing based on satellite data is still in its infancy. The existing technology only uses single multi-angle polarization payloads such as POLDER-3 and DPC. The constraints on aerosol optical and microphysical parameters in the inversion algorithm are relatively weak, and the uncertainty of parameter inversion is high. Summary of the invention

[0003] The present application proposes an aerosol optical and microphysical parameter inversion method and system to solve the problem of high uncertainty in the inversion of single multi-angle polarization load parameters in the prior art.

[0004] In a first aspect, the present application provides an aerosol optical and microphysical parameter inversion method, comprising:

[0005] Acquire multi-source data; the multi-source data is a collection of multiple data, including satellite data; the satellite data includes multi-angle polarization data, ultraviolet-visible light channel data and thermal infrared channel data;

[0006] Uniformly match multi-source data in space and time;

[0007] Determining initial values ​​of inversion parameters; the inversion parameters include aerosol optical microphysical parameters;

[0008] Iterate the inversion parameters to make the sum of the first difference, the second difference and the third difference the minimum, stop the iteration and obtain the current inversion parameters;

[0009] The first difference is calculated by substituting the initial value of the inversion parameter into the atmospheric forward radiation transfer model and then comparing it with the actual observation value of the multi-source satellite; the second difference is calculated by the inversion parameter after iteration and the initial value of the inversion parameter; the third difference is calculated by the absorption Observed and simulated values ​​of the index are calculated.

[0010] Furthermore, the multi-source data also includes model data and ground-based data. The model data includes meteorological reanalysis data and aerosol component data. The ground-based data includes aerosol optical and microphysical parameter data sets.

[0011] In one embodiment, determining the initial value of the inversion parameter specifically comprises the steps of:

[0012] Establish training sample sets through multi-source data;

[0013] The training sample set obtains the initial values ​​of the inversion parameters through machine learning.

[0014] In one embodiment, the inversion parameter further includes the optical thickness contribution ratio of the aerosol component;

[0015] Also includes the steps:

[0016] Obtaining optical thickness contribution ratios of aerosol components through machine learning; the aerosol components include black carbon;

[0017] The optical thickness contribution ratio of brown carbon is calculated by the optical thickness contribution ratio of black carbon and the absorption coefficient ratio of brown carbon to black carbon.

[0018] In one embodiment, the method further comprises the steps of:

[0019] Acquire multi-source data and analyze the sensitivity of inversion parameters of each data in the multi-source data;

[0020] If the sensitivity of the inversion parameters is lower than the set threshold, the corresponding data will be eliminated.

[0021] In one embodiment, the method further comprises the steps of:

[0022] Acquire multi-source data, and calculate the first signal degree of freedom of each data in the multi-source data;

[0023] Sort the first signal degrees of freedom of all data from large to small;

[0024] The data with the largest degree of freedom of the first signal is selected, and the remaining data are combined with the data with the largest degree of freedom of the first signal one by one to form a multi-source satellite actual observation value set to recalculate the second signal degree of freedom;

[0025] Selecting a multi-source satellite actual observation value set with the largest second signal degree of freedom, and recalculating the signal degrees of freedom by comparing the remaining multi-source satellite actual observation value sets with the multi-source satellite actual observation value set with the second signal degree of freedom;

[0026] Repeat the above steps until the number of actual observation value sets of multi-source satellites reaches the set threshold.

[0027] In one embodiment, the method further comprises the steps of:

[0028] Acquiring ground-based data; the ground-based data comprising aerosol optical and microphysical parameter data sets;

[0029] The inversion parameters are uniformly matched with the foundation data in space and time, and then compared and verified.

[0030] In a second aspect, the embodiments of the present application further provide an aerosol optical and microphysical parameter inversion system, which is used to implement the aerosol optical and microphysical parameter inversion method described in any one of the embodiments of the first aspect, and includes: an acquisition module, which is used to acquire multi-source data. A determination module, which is used to uniformly match multi-source data in space and time; it is also used to determine the initial value of the inversion parameter; it is also used to obtain the current inversion parameter after the inversion parameter stops iterating. A calculation module, which is used to calculate the first difference, the second difference, and the third difference; it is also used to iterate the inversion parameters so that the sum of the first difference, the second difference, and the third difference is the minimum value.

[0031] In a third aspect, an embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any embodiment of the first aspect.

[0032] In a fourth aspect, an embodiment of the present application further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method described in any embodiment of the first aspect is implemented.

[0033] At least one of the above technical solutions adopted in the embodiments of the present application can achieve the following beneficial effects:

[0034] This application breaks through the constraints of a single multi-angle polarization payload on the limited aerosol optical and microphysical detection capabilities and detection accuracy, and the difficulty in separating various aerosol components. A multi-angle polarization aerosol optical and microphysical parameter synchronous inversion method constrained by ultraviolet and thermal infrared observations is proposed. By establishing a multi-source satellite cross-payload joint inversion model, ultraviolet and thermal infrared observation spectral information is introduced to strengthen the constraints on the cost function, optimize the separation of aerosol components with different absorption characteristics, and improve the aerosol optical and microphysical inversion accuracy and signal freedom. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0036] Figure 1 This is a flow chart of an aerosol optical and microphysical parameter inversion method according to an embodiment of the present application;

[0037] Figure 2 This is a flow chart of another aerosol optical and microphysical parameter inversion method according to an embodiment of the present application;

[0038] Figure 3 This is a flow chart of the third aerosol optical and microphysical parameter inversion method of the embodiment of the present application;

[0039] Figure 4 This is a flow chart of the fourth aerosol optical and microphysical parameter inversion method of the embodiment of the present application;

[0040] Figure 5 This is a flow chart of the fifth aerosol optical and microphysical parameter inversion method of the embodiment of the present application;

[0041] Figure 6 This is the overall technical circuit diagram of the embodiment of this application;

[0042] Figure 7 This is a structural diagram of an aerosol optical and microphysical parameter inversion system according to an embodiment of the present application. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0044] In view of the limitations of the existing single satellite payload for aerosol optics (aerosol optical thickness AOD, single scattering albedo SSA, scattering phase function, etc.) and microphysical parameter (particle spectrum distribution, complex refractive index, etc.) detection technology, a multi-source satellite cross-payload joint inversion strategy was established based on the optimal estimation theory using the onboard multi-angle polarization, ultraviolet and thermal infrared payload observation information, to achieve remote sensing synchronous inversion of aerosol optics and microphysical parameters based on multi-source satellite observations, providing technical and data support for the research on refined aerosol information in the fields of climate change, ecological environment, public health, etc.

[0045] The technical solutions provided by various embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

[0046] Figure 1 This is a flow chart of an aerosol optical and microphysical parameter inversion method according to an embodiment of the present application, including steps 110 to 140.

[0047] Step 110, acquiring multi-source data; the multi-source data is a collection of multiple data, including satellite data; the satellite data includes multi-angle polarization data, ultraviolet-visible light channel data, and thermal infrared channel data;

[0048] Satellite data:

[0049] The multi-angle polarization payload POLDER-3 and DPC data are selected to provide observation information at different angles and polarization directions for the inversion of aerosol optical and microphysical parameters.

[0050] OMI, TROPOMI and EMI payloads are selected to provide UV-visible channel observation information, which is used to constrain and optimize the separation of optical and microphysical parameters of absorbing brown carbon and black carbon aerosols.

[0051] The thermal infrared channel observation information provided by AIRS, CrIS, Himawari-8 / 9 and other payloads, as well as the dust distribution data calculated by AIRS, CrIS, Himawari-8 / 9 and other payloads are selected to constrain and optimize the separation of brown carbon and dust optical and microphysical parameters.

[0052] According to the on-orbit time of POLDER-3 and DPC data, appropriate ultraviolet-visible light and thermal infrared payload data are selected to match them in time and space.

[0053] Furthermore, the multi-source data also includes model data and ground-based data. The model data includes meteorological reanalysis data and aerosol component data. The ground-based data includes aerosol optical and microphysical parameter data sets.

[0054] Mode data:

[0055] ERA-5 or MERRA-2 meteorological reanalysis data, as well as MERRA-2 aerosol component data (such as sulfate, black carbon, organic carbon, dust, sea salt, etc.) are selected to construct a dynamic aerosol component prior model.

[0056] Ground data:

[0057] Obtain the L2-level aerosol optical and microphysical parameter dataset of the AERONET station network to construct a dataset of aerosol optical and microphysical parameter prior information. Obtain multi-band black carbon instrument observation data to calculate parameters such as the brown carbon absorption contribution ratio and establish a relationship model between brown carbon and black carbon under different emission source conditions.

[0058] Step 120: uniformly matching multi-source data in space and time;

[0059] For example, taking into account the differences in the absorption characteristics of black carbon and dust aerosols in the ultraviolet, visible and thermal infrared bands, the scalar and vector radiation information in the ultraviolet-visible and thermal infrared spectral ranges provided by satellite payloads such as POLDER-3 (or DPC), OMI (or TROPOMI, EMI), and AIRS (or CrIS, Himawari-8 / 9) will be collected, combined with ERA5 meteorological reanalysis data, MERRA-2 aerosol composition model data, satellite dust distribution, ground-based black carbon observation data, AERONET ground-based aerosol optical and microphysical data, etc., to achieve unified matching in space and time.

[0060] Step 130, determining initial values ​​of inversion parameters; the inversion parameters include aerosol optical and microphysical parameters;

[0061] For example, the multi-source data only includes satellite data. Since there are multiple optical and microphysical parameters to be inverted, the initial values ​​of the inversion parameters are a set of values ​​set based on experience.

[0062] The multi-source data also includes model data and foundation data. The initial values ​​of the inversion parameters are obtained by machine learning after training the sample set. The specific process is as follows: Figure 2 shown.

[0063] Step 140, iterate the inversion parameters to make the sum of the first difference, the second difference and the third difference a minimum value, stop the iteration, and obtain the current inversion parameters.

[0064] The first difference is calculated by substituting the initial value of the inversion parameter into the atmospheric forward radiation transfer model and then comparing it with the actual observation value of the multi-source satellite; the second difference is calculated by the inversion parameter after iteration and the initial value of the inversion parameter; the third difference is calculated by the absorption Observed and simulated values ​​of the index are calculated.

[0065] Specifically: Multi-angle polarization aerosol inversion model constrained by ultraviolet and thermal infrared observations

[0066] By calculating the sum of the norms of the first difference, the second difference and the third difference as the minimum value, the iteration is stopped and the current inversion parameters are obtained.

[0067] The cost function under the constraint of adding ultraviolet and thermal infrared information in the optimization inversion is redefined as:

[0068]

[0069] Among them, || || represents the vector norm, vector X is the inversion parameters such as aerosol optics, microphysics and mixing ratio factors of each component, and vector X a is the initial value of vector X, vector Y mis the normalized radiation value of m input observations (including multi-angle polarization, ultraviolet-visible light, and thermal infrared load observations), Y m is the normalized radiation value vector of multi-source satellite observations (i.e., the actual observation value of multi-source satellites); γ is the Lagrange multiplier, f(X) is the value obtained by bringing the initial value of the inversion parameter into the atmospheric forward radiation transfer model, AAE i,obs is the AAE calculated from the POLDER-3 / DPC 865nm, 565nm and OMI / TROPOMI / EMI 388nm, TROPOMI 710nm observations (i.e., AAE observations), i,cal is the AAE value of n corresponding wavelength pairs (i.e., AAE simulation value) calculated based on the aerosol inversion parameters and the forward radiation transfer model, where AAE is calculated after dust interference correction based on infrared observation information.

[0070] Among them, it is worth noting that The term is not considered in the current optimization inversion of aerosol composition, and cannot be constrained by single POLDER-3 or DPC load inversion. The norm here is the Mahalanobis distance, and the above formula can be further converted to

[0071]

[0072] Where R is Y m The error covariance vector of , which mainly includes the errors caused by the instrument and the system; T is the vector transpose; B represents the background covariance vector. After defining the cost function, the iterative formula is derived and the constraints on the iterative formula are strengthened according to the Levenberg-Marquardt model. The positive definite vector and the γ factor in the Levenberg-Marquardt method are adaptively adjusted through the diagonal elements and nonlinearity of the Hessian vector.

[0073] Based on the above-mentioned redefined cost function and the Bayesian optimization estimation inversion framework, we conducted research on multi-pixel joint inversion strategies and adaptive iterative step-size optimization schemes, and established a synchronous inversion model for aerosol components, optics, and microphysics.

[0074] Figure 2 This is a flow chart of another aerosol optical and microphysical parameter inversion method according to an embodiment of the present application, including steps 110 to 120 , 210 to 220 , and 140 .

[0075] Step 110, acquiring multi-source data; the multi-source data is a collection of multiple data, including satellite data; the satellite data includes multi-angle polarization data, ultraviolet-visible light channel data, and thermal infrared channel data;

[0076] It should be noted that the observations described below are the multi-source data. Each observation is one of the multiple data.

[0077] Step 120: uniformly matching multi-source data in space and time;

[0078] Step 130, determining initial values ​​of inversion parameters; the inversion parameters include aerosol optical microphysical parameters;

[0079] In one embodiment, step 130 determines the initial value of the inversion parameter, specifically comprising the steps of:

[0080] Step 210: Establish a training sample set through multi-source data;

[0081] Data matching and cleaning

[0082] The scalar and vector radiation information in the ultraviolet-visible and thermal infrared spectral ranges provided by satellite payloads such as POLDER-3 (or DPC), OMI (or TROPOMI, EMI), and AIRS (or CrIS, Himawari-8 / 9) will be collected, combined with multi-source data such as ERA5 meteorological reanalysis data, MERRA-2 aerosol component model data, satellite dust distribution, ground-based black carbon observation data, AERONET ground-based aerosol optical and microphysical data, etc., to perform unified matching in space and time, eliminate invalid and abnormal data, and establish a training sample set.

[0083] Step 220: The training sample set obtains initial values ​​of inversion parameters through machine learning.

[0084] Construction of prior model of conventional aerosol components:

[0085] The sample training set is resampled to the same vertical grid; the random forest method is used to analyze the feature importance of the collected data to obtain the important features of the model input; the long short-term memory network LSTM is used to train and test the data, and the appropriate hidden layer and number of nodes are selected to obtain the optimal network structure, thereby obtaining the initial value of the inversion parameter.

[0086] Step 140, iterate the inversion parameters to make the sum of the first difference, the second difference and the third difference a minimum value, stop the iteration, and obtain the current inversion parameters.

[0087] Furthermore, in one of the embodiments, the inversion parameters also include optical thickness contribution ratios of aerosol components;

[0088] Also includes the steps:

[0089] Obtaining optical thickness contribution ratios of aerosol components through machine learning; the aerosol components include black carbon;

[0090] The sample training set is resampled to the same vertical grid; the random forest method is used to perform feature importance analysis on the collected data to obtain important features of the model input; the long short-term memory network LSTM is used to train and test the data, and the appropriate hidden layer and number of nodes are selected to obtain the optimal network structure, thereby obtaining a priori model of the optical thickness contribution ratio (the ratio of the optical thickness of each aerosol component to the total optical thickness of the aerosol) of conventional aerosol components (such as black carbon, organic carbon, dust, sulfate, nitrate, sea salt, etc.) with better generalization.

[0091] The optical thickness contribution ratio of brown carbon is calculated by the optical thickness contribution ratio of black carbon and the absorption coefficient ratio of brown carbon to black carbon.

[0092] Calculate the brown carbon absorption contribution ratio, specifically:

[0093] In atmospheric remote sensing models and applications, black carbon is usually assumed to be the only absorbing aerosol. However, a large number of field observations and laboratory simulation studies in recent years have shown that in the ultraviolet-near visible light band range (200-550nm), black carbon is not the only aerosol component with absorption. A part of the organic carbon component, namely brown carbon (BrC) aerosol. Therefore, in order to further constrain the uncertainty of black carbon absorption, this technology, based on the calculation and difference of optical parameters of black carbon and brown carbon particles under different particle sizes, morphological structures, and mixed states (internal mixing and semi-mixed states), intends to use the dual-wavelength method to simulate and calculate the AAE value of black carbon, and improve the algorithm of the conventional assumption that AAE = 1.0, and use the following formula to calculate the actual absorption coefficient (Abs) of black carbon at short wavelength (λ):

[0094]

[0095] The absorption coefficient of brown carbon is the difference between the total absorption coefficient and the absorption coefficient of black carbon at each wavelength.

[0096] It should be noted that the data needs to be preprocessed before calculation. According to the different wavelength dependence of carbonaceous aerosol and dust aerosol, the interference of dust absorption is eliminated. Based on the improved brown carbon estimation method, the absorption coefficients of black carbon and brown carbon in ground-based data under long-term series observation and their contribution to the total carbonaceous aerosol absorption are calculated.

[0097] Construction of a priori model for brown carbon aerosol

[0098] Based on the brown carbon absorption contribution estimation model and observation data, combined with meteorological and black carbon model data, and referring to the above steps for calculating the brown carbon absorption contribution ratio, a small sample machine learning method is used to construct a dynamic estimation model for the brown carbon absorption contribution ratio. In the inversion process, the black carbon optical thickness contribution ratio and the brown carbon absorption contribution ratio are obtained through dynamic calculation, and then the dynamic prior information of the brown carbon optical thickness contribution ratio can be obtained.

[0099] It should also be noted that due to the instability of aerosol optical and microphysical parameters inverted from satellite data, and this instability is closely related to the composition of atmospheric aerosols, the background prior covariance vector is required to constrain the solution; if the variance of the background field is underestimated, it will lead to overly strict constraints, making the solution obtained unstable. As shown in the following formula, B represents the background field covariance vector or the initial prior error covariance. The background covariance vector is obtained by statistics of the above prior knowledge base:

[0100]

[0101] Among them, X componet are the optical and microphysical parameters of aerosol components (black carbon, organic carbon, brown carbon, dust, sulfate, nitrate, sea salt, etc.) in the prior knowledge base, is the statistical mean of the optical and microphysical parameters of the corresponding aerosol components in the prior knowledge base.

[0102] Figure 3 This is a flow chart of a third aerosol optical and microphysical parameter inversion method according to an embodiment of the present application, including steps 110, 310-320, and 120-140.

[0103] Step 110, acquiring multi-source data; the multi-source data is a collection of multiple data, including satellite data; the satellite data includes multi-angle polarization data, ultraviolet-visible light channel data, and thermal infrared channel data;

[0104] Since the present application involves the joint inversion of observation data of multiple multi-angle polarization loads, ultraviolet-visible light loads, and thermal infrared loads, the number of observations is large and the amount of redundant information is large. The goal of multi-load joint inversion is to examine under what kind of observation set the inversion capability reaches the highest. Therefore, in order to ensure the inversion accuracy and improve the calculation speed at the same time, it is necessary to select the optimal channel, angle, and polarization observation combination from the observation set that each load can provide by calculating the degrees of freedom of the inversion parameter signal.

[0105] In one embodiment, the method further comprises the steps of:

[0106] Step 310, acquiring multi-source data, and analyzing the sensitivity of the inversion parameters of each data in the multi-source data;

[0107] Step 320: If the sensitivity of the inversion parameter is lower than the set threshold, the corresponding data is discarded.

[0108] Step 120: uniformly matching multi-source data in space and time;

[0109] Step 130, determining initial values ​​of inversion parameters; the inversion parameters include aerosol optical microphysical parameters;

[0110] Step 140, iterate the inversion parameters to make the sum of the first difference, the second difference and the third difference a minimum value, stop the iteration, and obtain the current inversion parameters.

[0111] Figure 4 This is a flow chart of a fourth aerosol optical and microphysical parameter inversion method according to an embodiment of the present application, including steps 110, 410-450, and 120-140.

[0112] Step 110, acquiring multi-source data; the multi-source data is a collection of multiple data, including satellite data; the satellite data includes multi-angle polarization data, ultraviolet-visible light channel data, and thermal infrared channel data;

[0113] In one embodiment, the method further comprises the steps of:

[0114] Step 410: Acquire multi-source data and calculate the first signal degree of freedom of each data in the multi-source data;

[0115] Signal degrees of freedom (DFS) is a quantitative description of the target information contained in the observation. The higher the DFS, the more sufficient the target information contained in the observation. Usually, when the DFS of a parameter is greater than 0.5, it can be considered that the parameter can be constrained by observation in theory. The specific formula can be expressed as

[0116]

[0117] Among them, vector A is the average kernel function, which is the partial derivative of the posterior estimate with respect to the state parameter in the optimal estimation theory, and its diagonal elements, namely the DFS components, are the partial sensitivities of a single inversion parameter relative to its true value.

[0118] Step 420, sorting the first signal degrees of freedom of all data from large to small;

[0119] Step 430: Select the data with the largest degree of freedom of the first signal, and combine the remaining data with the data with the largest degree of freedom of the first signal one by one to form a multi-source satellite actual observation value set to recalculate the second signal degree of freedom;

[0120] Step 440: Select the multi-source satellite actual observation value set with the largest second signal degree of freedom, and recalculate the signal degrees of freedom by comparing the remaining multi-source satellite actual observation value sets with the multi-source satellite actual observation value set with the largest second signal degree of freedom one by one;

[0121] Step 450: Repeat the above steps until the number of actual observation value sets of multi-source satellites reaches a set threshold.

[0122] Step 120: uniformly matching multi-source data in space and time;

[0123] Step 130, determining initial values ​​of inversion parameters; the inversion parameters include aerosol optical microphysical parameters;

[0124] Step 140, iterate the inversion parameters to make the sum of the first difference, the second difference and the third difference a minimum value, stop the iteration, and obtain the current inversion parameters.

[0125] For example, the above Figure 3 and Figure 4 The steps can be used simultaneously. For example, for multi-angle polarization, ultraviolet and thermal infrared satellite payloads, the sensitivity of remote sensing detection to aerosol composition, mixing state, optical and microphysical information under different surface conditions is analyzed based on the forward radiation transfer model. All observations that are sensitive to aerosol composition and are less affected by gas absorption in each payload are selected as the candidate observation set.

[0126] Forward simulation is performed on a single observation, and the total signal degrees of freedom of aerosol optics, microphysics and the mixing ratio of each component under each observation are calculated, sorted, and the observation with the largest signal degrees of freedom is selected to be added to the inversion input observation set; the remaining observations are combined with the inversion input observation set one by one to recalculate the signal degrees of freedom, and the observation that can increase the signal degrees of freedom the most is added to the inversion input observation set; the above steps are repeated until the signal degrees of freedom are no longer increased. The final inversion input observation set will be used as the input observation for the cross-load joint inversion of brown carbon.

[0127] Figure 5 This is a flow chart of a fifth aerosol optical and microphysical parameter inversion method according to an embodiment of the present application, including steps 110 to 140 and 510 to 520.

[0128] Step 110, acquiring multi-source data; the multi-source data is a collection of multiple data, including satellite data; the satellite data includes multi-angle polarization data, ultraviolet-visible light channel data, and thermal infrared channel data;

[0129] Step 120: uniformly matching multi-source data in space and time;

[0130] Step 130, determining initial values ​​of inversion parameters; the inversion parameters include aerosol optical microphysical parameters;

[0131] Step 140, iterate the inversion parameters to make the sum of the first difference, the second difference and the third difference a minimum value, stop the iteration, and obtain the current inversion parameters.

[0132] In one embodiment, the method further comprises the steps of:

[0133] Step 510: Acquire ground-based data; the ground-based data includes aerosol optical and microphysical parameter data sets;

[0134] Step 520: uniformly match the inversion parameters with the foundation data in space and time, and then compare and verify.

[0135] The uncertainty of the inversion results mainly comes from the input atmospheric and surface parameters, observation noise and model parameters. The uncertainty caused by the prior information is quantitatively analyzed by two situations: the prior information is not updated and the prior information is updated; the uncertainty caused by the observation noise is quantitatively analyzed by changing the observation covariance vector during inversion. Since the inversion model of this technology involves multiple loads, the signal-to-noise ratio of different loads needs to be considered in the observation covariance vector, and at the same time, the impact of the error transmission across loads on the inversion of aerosol optical and microphysical parameters needs to be considered.

[0136] The verification of the inversion results includes two aspects. On the one hand, the inversion results are compared and verified with the aerosol optical and microphysical data obtained from observations such as AERONET CE318 and black carbon instruments; on the other hand, for absorbing carbonaceous aerosols, the black carbon inversion results obtained by synchronous inversion are compared and verified with the OE / EC observation results, and combined with the equipment observation results of the online mass spectrometer, the rationality and applicability of the brown carbon inversion results are indirectly verified.

[0137] This application establishes dynamic prior models of different components based on various factors that affect the distribution of aerosol optical and microphysical information, obtains prior information and background covariance vectors of aerosol components that are more in line with actual conditions, and provides a prerequisite for more accurate inversion of aerosol optical and microphysical parameters.

[0138] For multiple historical and on-orbit multi-angle polarization, ultraviolet and thermal infrared payload observations, sensitivity and signal freedom analysis under different observation and atmospheric state scenarios are carried out, and different payload channel, angle, and polarization observation combinations that are sensitive to aerosol components and have the highest signal freedom are selected as inversion inputs; VLIDORT and LBLRTM radiation transfer models are coupled, and on the basis of the multi-angle polarization aerosol optimization inversion algorithm, dynamic prior models are coupled to introduce ultraviolet and thermal infrared satellite observation information to enhance inversion constraints and optimize the separation of brown carbon from black carbon and dust aerosols. Specifically, Figure 6shown.

[0139] Figure 7 This is a structural diagram of an aerosol optical and microphysical parameter inversion system according to an embodiment of the present application, which is used to implement the aerosol optical and microphysical parameter inversion method described in any one of the embodiments of the first aspect, comprising:

[0140] The acquisition module 71 is used to acquire multi-source data.

[0141] The determination module 72 is used to uniformly match the multi-source data in space and time; it is also used to determine the initial value of the inversion parameter; it is also used to obtain the current inversion parameter after the inversion parameter stops iterating.

[0142] The calculation module 73 is used to calculate the first difference, the second difference and the third difference; and is also used to iterate the inversion parameters so that the sum of the first difference, the second difference and the third difference is the minimum.

[0143] In one of the embodiments, the determination module further includes a machine learning unit for establishing a training sample set using multi-source data and obtaining initial values ​​of inversion parameters based on the training sample set using machine learning.

[0144] In one embodiment, the determination module also includes an optical thickness contribution ratio inversion unit, which is used to obtain the optical thickness contribution ratio of aerosol components through machine learning; the aerosol components include black carbon; and is also used to calculate the optical thickness contribution ratio of brown carbon through the optical thickness contribution ratio of black carbon and the absorption coefficient ratio of brown carbon to black carbon.

[0145] In one of the embodiments, the acquisition module further comprises a sensitivity screening unit for analyzing the sensitivity of the inversion parameters of each data in the multi-source data; and eliminating the data whose inversion parameter sensitivity is lower than a set threshold.

[0146] In one of the embodiments, the acquisition module further includes a degree of freedom screening unit, which is used to calculate the first signal degree of freedom of each data in the multi-source data, and sort the first signal degrees of freedom of all the data from large to small; select the data with the largest first signal degree of freedom, and combine the remaining data with the data with the largest first signal degree of freedom one by one to form a multi-source satellite actual observation value set to recalculate the second signal degree of freedom; select the multi-source satellite actual observation value set with the largest second signal degree of freedom, and combine the remaining multi-source satellite actual observation value sets with the multi-source satellite actual observation value set with the multi-source satellite actual observation value set with the second signal degree of freedom to recalculate the signal degrees of freedom; repeat the above steps until the number of multi-source satellite actual observation value sets reaches a set threshold.

[0147] In one embodiment, a verification module 74 is further included, which is used to compare and verify the inversion parameters and the foundation data after uniformly matching them in space and time.

[0148] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0149] Therefore, the present application also proposes a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method described in any embodiment of the present application is implemented.

[0150] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0151] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0152] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0153] Furthermore, the present application also proposes an electronic device (or computing device), comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in any embodiment of the present application when executing the computer program.

[0154] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory. Memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium. Computer-readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. According to the definition in this article, computer-readable media does not include temporary computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0155] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity 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, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0156] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A method for inverting aerosol optical and microphysical parameters, characterized in that: The following steps are involved: Acquire multi-source data, wherein the multi-source data is a collection of multiple types of data, including satellite data; the satellite data includes multi-angle polarization data, ultraviolet-visible light channel data, and thermal infrared channel data; Uniformly match multi-source data in space and time; Determining initial values ​​of inversion parameters, wherein the inversion parameters include aerosol optical microphysical parameters; Iterate the inversion parameters to make the sum of the first difference, the second difference and the third difference the minimum, stop the iteration, and obtain the current inversion parameters; The first difference is calculated by substituting the initial value of the inversion parameter into the atmospheric forward radiation transfer model and the actual observation value of the multi-source satellite; The second difference is calculated by using the iterated inversion parameter and the initial value of the inversion parameter; The third difference is calculated by the observed value and simulated value of the absorptive Ångström index, the observed value is calculated by POLDER-3 / DPC 865nm, 565nm, OMI / TROPOMI / EMI 388nm or TROPOMI 710nm observations; the simulated value is the AAE value of n corresponding wavelengths calculated based on the aerosol inversion parameters and the forward radiation transfer model.

2. The aerosol optical and microphysical parameter inversion method according to claim 1, characterized in that: The multi-source data also includes pattern data and ground-based data; The model data include meteorological reanalysis data and aerosol composition data; The ground-based data include aerosol optical and microphysical parameter datasets.

3. The aerosol optical and microphysical parameter inversion method according to claim 1, characterized in that: Determining the initial value of the inversion parameter specifically comprises the steps of: Establish training sample sets through multi-source data; The training sample set obtains the initial values ​​of the inversion parameters through machine learning.

4. The aerosol optical and microphysical parameter inversion method according to claim 1, characterized in that: The inversion parameters also include the optical thickness contribution ratio of the aerosol components; Also includes the steps: Obtaining optical thickness contribution ratios of aerosol components through machine learning; the aerosol components include black carbon; The optical thickness contribution ratio of brown carbon is calculated by the optical thickness contribution ratio of black carbon and the absorption coefficient ratio of brown carbon to black carbon.

5. The aerosol optical and microphysical parameter inversion method according to claim 1, characterized in that: Also includes the steps: Acquire multi-source data and analyze the sensitivity of inversion parameters of each data in the multi-source data; If the sensitivity of the inversion parameters is lower than the set threshold, the corresponding data will be eliminated.

6. The aerosol optical and microphysical parameter inversion method according to claim 1, characterized in that: Also includes the steps: Acquire multi-source data, and calculate the first signal degree of freedom of each data in the multi-source data; Sort the first signal degrees of freedom of all data from large to small; The data with the largest degree of freedom of the first signal is selected, and the remaining data are combined with the data with the largest degree of freedom of the first signal one by one to form a multi-source satellite actual observation value set to recalculate the second signal degree of freedom; Selecting a multi-source satellite actual observation value set with the largest second signal degree of freedom, and recalculating the signal degrees of freedom by comparing the remaining multi-source satellite actual observation value sets with the multi-source satellite actual observation value set with the second signal degree of freedom; Repeat the above steps until the number of actual observation value sets of multi-source satellites reaches the set threshold.

7. The aerosol optical and microphysical parameter inversion method according to claim 1, characterized in that: Also includes the steps: Acquiring ground-based data; the ground-based data includes aerosol optical and microphysical parameter data sets; The inversion parameters are uniformly matched with the foundation data in space and time and then compared and verified.

8. An aerosol optical and microphysical parameter inversion system, characterized in that: The method for inverting aerosol optical and microphysical parameters according to any one of claims 1 to 7 comprises: Acquisition module, used to acquire multi-source data; A determination module is used to uniformly match multi-source data in space and time; it is also used to determine the initial value of the inversion parameter; it is also used to obtain the current inversion parameter after the inversion parameter stops iterating; The calculation module is used to calculate the first difference, the second difference and the third difference; and is also used to iterate the inversion parameters so that the sum of the first difference, the second difference and the third difference is the minimum.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

10. An electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 8 is implemented.

Images