Method and device for collaborative inversion of anthropogenic carbon emissions by active lidar and ground-based observation

Abstract

The application provides a method and device for cooperatively inverting artificial carbon emission by active laser radar and ground observation, and relates to the technical field of carbon emission inversion. The method comprises the following steps: constructing a multi-scale observation vector corresponding to a target region based on active concentration observation data corresponding to an active spaceborne laser radar and ground concentration observation data corresponding to a ground observation station; constructing a joint observation error covariance matrix based on attribute information corresponding to the active concentration observation data and attribute information corresponding to the ground concentration observation data; cooperatively inverting the artificial carbon emission amount of the target region based on a prior multi-scale state vector, the multi-scale observation vector and the joint observation error covariance matrix, so as to minimize the value of a joint cost function, and obtaining an inversion result of the artificial carbon emission amount. The application can significantly improve the accuracy of the inversion of the artificial carbon emission.

Description

Technical Field

[0001] This invention relates to the field of carbon emission inversion technology, and in particular to a method and apparatus for co-inverting anthropogenic carbon emissions using active lidar and ground-based observation. Background Technology

[0002] Currently, the "top-down" atmospheric Source-sink change inversion is a powerful tool for assessing the effectiveness of carbon neutrality efforts. The key to this inversion is utilizing measured atmospheric data. The optimal global or regional near-surface carbon flux is estimated by combining concentration with atmospheric transport models and data assimilation techniques. This method is widely used to assess the effectiveness of carbon neutrality actions. Although the "top-down" carbon source-sink estimation method has become the mainstream direction in carbon cycle research, due to limitations such as the amount of observational data and model efficiency, there is still very little work on regional-scale assimilation inversion, and the inversion results are mostly low-resolution and highly uncertain, with global-scale assimilation being the most common. Against this backdrop, how to accurately characterize anthropogenic carbon emissions has become one of the key issues in carbon cycle research. Currently, carbon assimilation inversion systems and research mainly focus on the inversion of natural carbon fluxes, while atmospheric assimilation research on anthropogenic carbon emissions is relatively limited. Most carbon flux assimilation inversion methods optimize one side while assuming that there are no errors in either natural or anthropogenic carbon fluxes. This assumption reduces the effectiveness of carbon assimilation inversion to some extent. The accuracy and application value of flux inversion are insufficient to meet the needs of monitoring and verifying the effectiveness of carbon neutrality actions at the provincial and municipal levels. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a method and apparatus for the collaborative inversion of anthropogenic carbon emissions using active lidar and ground-based observation, which can significantly improve the accuracy of anthropogenic carbon emission inversion.

[0004] In a first aspect, the present invention provides a method for co-inverting anthropogenic carbon emissions using active lidar and ground-based observation, comprising: Based on the active spaceborne lidar Concentration observation data, and the foundation corresponding to the ground-based observation station. Concentration observation data are used to construct a multi-scale observation vector corresponding to the target area. Based on active Attribute information and foundation corresponding to concentration observation data Based on the attribute information corresponding to the concentration observation data, construct the joint observation error covariance matrix, which includes the active observation error covariance matrix, the ground-based observation error covariance matrix, and the cross-observation error covariance matrix. Based on prior multi-scale state vectors, multi-scale observation vectors, and joint observation error covariance matrix, the anthropogenic carbon emissions in the target area are collaboratively inverted with the goal of minimizing the joint cost function, resulting in the anthropogenic carbon emission inversion results.

[0005] In one implementation, based on active Attribute information and foundation corresponding to concentration observation data The attribute information corresponding to the concentration observation data is used to construct the joint observation error covariance matrix, including: Based on the column density kernel function weights, laser echo noise parameters, and spatial distance between active spaceborne lidars, an active observation error covariance matrix is ​​constructed. Based on the standard deviation of the observation error of the ground observation station, the representative correlation scale of the ground, and the station distance between the ground observation stations, the ground observation error covariance matrix is ​​constructed. Based on the correlation coefficient of observation errors between the active spaceborne lidar and the ground-based observation station, the standard deviation of the observation error of the active spaceborne lidar, and the standard deviation of the observation error of the ground-based observation station, a cross-observation error covariance matrix is ​​constructed. Based on the active observation error covariance matrix, the ground-based observation error covariance matrix, and the cross-observation error covariance matrix, a joint observation error covariance matrix is ​​constructed.

[0006] In one implementation, an active observation error covariance matrix is ​​constructed based on the column density kernel function weights of the active spaceborne lidar, the laser echo noise parameters, and the spatial distance between the active spaceborne lidars, including: The following operations are performed for the first and second active spaceborne lidar: Based on the laser echo noise estimate and atmospheric vertical height coordinates of the first active spaceborne lidar, and the laser echo noise estimate and atmospheric vertical height coordinates of the second active spaceborne lidar, the hierarchical error covariance between the first and second active spaceborne lidars is determined. Based on the ratio of the spatial distance between the first and second active spaceborne lidars to the horizontal error correlation scale, the active attenuation factor between the first and second active spaceborne lidars is determined. The horizontal error correlation scale is the spatial characteristic length characterizing the error correlation between the observations of the first and second active spaceborne lidars. The radar error term between the active spaceborne lidar observation and the second active spaceborne lidar is determined based on the column concentration kernel function weight, the stratification error covariance, and the active attenuation factor. Based on the laser echo noise parameters, the radar noise term between the active spaceborne lidar observation and the second active spaceborne lidar is determined. The sum of the radar error term and the radar noise term is used as the active observation error value between the active spaceborne lidar observation and the second active spaceborne lidar to construct the active observation error covariance matrix.

[0007] In one implementation, a ground observation error covariance matrix is ​​constructed based on the standard deviation of the observation error of the ground observation stations, the representative correlation scale of the ground, and the station distance between the ground observation stations, including: The following operations will be performed for the first and second ground-based observation stations: Determine the product of standard deviations based on the standard deviation of the observation error of the first ground observation station and the standard deviation of the observation error of the second ground observation station; The ground attenuation factor is determined by the ratio of the distance between the first and second ground observation stations to the representative correlation scale of the ground. The representative correlation scale of the ground is the spatial characteristic length that characterizes the error correlation between the observations of the first and second ground observation stations. Multiply the standard deviation product by the foundation attenuation factor, and the resulting product is used as the foundation observation error covariance between the first and second foundation observation stations to construct the foundation observation error covariance matrix.

[0008] In one implementation, a cross-observation error covariance matrix is ​​constructed based on the correlation coefficient of observation errors between the active spaceborne lidar and the ground-based observation station, the standard deviation of the observation errors of the active spaceborne lidar, and the standard deviation of the observation errors of the ground-based observation station. This includes: The following operations should be performed for active spaceborne lidar and ground-based observation stations: The spatial attenuation factor is determined by the ratio of the spatial distance between the active spaceborne lidar and the ground-based observation station to the horizontal error correlation scale. The horizontal error correlation scale is the spatial characteristic length that characterizes the error correlation between the active spaceborne lidar observation and the ground-based observation station observation. The time decay factor is determined by the ratio of the absolute value of the time difference between the observation time of the active spaceborne lidar and the observation time of the ground-based observation station to the time error correlation scale. The time error correlation scale is the time characteristic length that characterizes the error correlation between the observation of the active spaceborne lidar and the observation of the ground-based observation station. Multiply the spatial decay factor by the temporal decay factor, and the product is used as the correlation coefficient of the observation error between the active spaceborne lidar and the ground-based observation station; The observation error correlation coefficient, the standard deviation of the observation error of the active spaceborne lidar, and the standard deviation of the observation error of the ground-based observation station are multiplied together. The resulting product is used as the cross-observation error covariance between the active spaceborne lidar and the ground-based observation station to construct the cross-observation error covariance matrix.

[0009] In one implementation, based on prior multi-scale state vectors, multi-scale observation vectors, and joint observation error covariance matrix, and with the objective of minimizing the joint cost function, a collaborative inversion of anthropogenic carbon emissions in the target region is performed to obtain anthropogenic carbon emission inversion results, including: Using a multi-scale observation operator, the prior multi-scale state vector is transformed into the observation space, resulting in... Concentration simulation data, and on Concentration simulation data and active Concentration observation data, foundation Spatiotemporal matching of concentration observation data; By using a pre-constructed multi-scale joint cost function, based on the matched... Concentration simulation data, multi-scale observation vectors and Concentration simulation data are used to perform a collaborative inversion of anthropogenic carbon emissions in the target area, minimizing the value of the multi-scale joint cost function. The multi-scale state vectors corresponding to the concentration simulation data are used as the results of anthropogenic carbon emission inversion.

[0010] In one implementation, the multi-scale observation vector, multi-scale state vector, multi-scale observation operator, and multi-scale joint cost function are all obtained by decomposing the large-scale background transport, regional medium-scale transport, and local anthropogenic emission disturbances.

[0011] Secondly, the present invention also provides a device for co-inverting anthropogenic carbon emissions using active lidar and ground-based observation, comprising: The observation vector construction module is based on the active spaceborne lidar corresponding to the active... Concentration observation data, and the foundation corresponding to the ground-based observation station. Concentration observation data are used to construct a multi-scale observation vector corresponding to the target area. Covariance construction module, used for active... Attribute information and foundation corresponding to concentration observation data Based on the attribute information corresponding to the concentration observation data, construct the joint observation error covariance matrix, which includes the active observation error covariance matrix, the ground-based observation error covariance matrix, and the cross-observation error covariance matrix. The inversion module is used to perform a collaborative inversion of anthropogenic carbon emissions in the target area based on prior multi-scale state vectors, multi-scale observation vectors, and joint observation error covariance matrix, with the goal of minimizing the value of the joint cost function, and obtain the anthropogenic carbon emission inversion results.

[0012] Thirdly, the present invention also provides an electronic device including a processor and a memory, the memory storing computer-executable instructions executable by the processor, the processor executing the computer-executable instructions to implement any of the methods provided in the first aspect.

[0013] Fourthly, the present invention also provides a computer-readable storage medium storing computer-executable instructions, which, when invoked and executed by a processor, cause the processor to implement any of the methods provided in the first aspect.

[0014] The present invention provides a method and apparatus for co-inverting anthropogenic carbon emissions using active lidar and ground-based observation. Firstly, it is based on the active lidar corresponding to spaceborne lidar. Concentration observation data, and the foundation corresponding to the ground-based observation station. Concentration observation data are used to construct a multi-scale observation vector corresponding to the target area; then, based on active... Attribute information and foundation corresponding to concentration observation data Based on the attribute information corresponding to the concentration observation data, a joint observation error covariance matrix is ​​constructed, which includes the joint observation error covariance matrix, the active observation error covariance matrix, the ground-based observation error covariance matrix, and the cross-observation error covariance matrix. Finally, based on the prior multi-scale state vector, multi-scale observation vector, and joint observation error covariance matrix, with the goal of minimizing the joint cost function, the anthropogenic carbon emissions in the target area are collaboratively inverted to obtain the anthropogenic carbon emission inversion results. The aforementioned method constructs a multi-scale observation vector by fusing multi-source observation data from active spaceborne lidar and ground-based observation stations. Based on the attribute information of the observation data, it constructs a joint observation error covariance matrix that includes active observation errors, ground-based observation errors, and cross-observation errors, thus achieving a complete characterization of the multi-source observation error structure. On this basis, the multi-scale state vector, multi-scale observation vector, and joint observation error covariance matrix are uniformly incorporated into the collaborative inversion framework. The inversion solution is performed with the goal of minimizing the joint cost function, so that information from different observation sources can be reasonably weighted according to their respective error characteristics, effectively reducing the uncertainty in the inversion process and thus significantly improving the accuracy of the inversion results of anthropogenic carbon emissions in the target area.

[0015] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained through the structures particularly pointed out in the description and the drawings.

[0016] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0017] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 A schematic flowchart illustrating a method for collaborative inversion of anthropogenic carbon emissions using active lidar and ground-based observation, provided in an embodiment of the present invention; Figure 2 This invention provides a schematic diagram illustrating the functional relationship between anthropogenic carbon emissions and observed CO2 concentrations. Figure 3 This is a schematic diagram of a device for co-inverting anthropogenic carbon emissions using active lidar and ground-based observation, provided in an embodiment of the present invention. Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Most current carbon flux assimilation and inversion methods optimize the other side under the assumption that natural or anthropogenic carbon fluxes are error-free. This assumption reduces the effectiveness of carbon flux assimilation to some extent. The accuracy and application value of flux inversion are insufficient to meet the needs of monitoring and verifying the effectiveness of carbon neutrality actions at the provincial and municipal levels. Therefore, this invention provides a method and apparatus for collaborative inversion of anthropogenic carbon emissions using active lidar and ground-based observations, which can significantly improve the accuracy of anthropogenic carbon emission inversion.

[0021] To facilitate understanding of this embodiment, a detailed description of the method for co-inverting anthropogenic carbon emissions using active lidar and ground-based observation, as disclosed in this embodiment of the invention, will be provided first. (See [link to relevant documentation]). Figure 1 The diagram shows a flowchart of a method for co-derived inversion of anthropogenic carbon emissions using active lidar and ground-based observation. This method mainly includes the following steps S102 to S106: Step S102, based on the active spaceborne lidar corresponding to the active Concentration observation data, and the foundation corresponding to the ground-based observation station. Concentration observation data are used to construct multi-scale observation vectors corresponding to the target area.

[0022] Active spaceborne lidar refers to the Atmospheric Environmental Monitoring Satellite (AEMS), the world's first active lidar satellite for CO2 detection, launched on April 16, 2022. Ground-based observation stations are fixed, continuous greenhouse gas observation stations deployed on the ground. Concentration observation data refers to the data collected by active spaceborne lidar targeting the target area. Concentration observation data, foundation Concentration observation data refers to the data collected by ground-based observation stations for the target area. Concentration observation data. Multi-scale observation vectors refer to the observation vectors obtained by decomposing the data of all active spaceborne lidar observation pixels and all ground-based observation stations within the target area according to three scales: large-scale background transport, regional medium-scale transport, and local anthropogenic emission disturbances.

[0023] Step S104, based on active Attribute information and foundation corresponding to concentration observation data Based on the attribute information corresponding to the concentration observation data, construct the joint observation error covariance matrix, which includes the active observation error covariance matrix, the ground-based observation error covariance matrix, and the cross-observation error covariance matrix.

[0024] The attribute information corresponding to active CO2 concentration observation data refers to a set of descriptive parameters directly related to the carbon dioxide column concentration values ​​acquired by active spaceborne lidar, including the geographical location of the observed pixel, the weight of the column concentration kernel function, and the laser echo noise parameters. The attribute information corresponding to ground-based CO2 concentration observation data refers to a set of descriptive parameters directly related to the carbon dioxide concentration values ​​acquired by ground observation stations, including the station's precise geographic coordinates, the standard deviation of the observation error, and the ground-based representativeness correlation scale.

[0025] The joint observation error covariance matrix is ​​a square matrix formed by unifying the observation errors of all active spaceborne lidar observation pixels and all ground-based observation stations within the target area into a single mathematical structure. This matrix characterizes the statistical correlation of observation errors between different observation points. Its diagonal elements represent the error variance of each observation point itself, while the off-diagonal elements represent the error covariance between similar observation points (e.g., lidar versus lidar) or dissimilar observation points (e.g., lidar versus ground-based station). The active observation error covariance matrix is ​​a submatrix within the joint observation error covariance matrix composed solely of all active spaceborne lidar observation pixels; the ground-based observation error covariance matrix is ​​a submatrix within the joint observation error covariance matrix composed solely of all ground-based observation stations; and the cross-observation error covariance matrix is ​​a rectangular sub-block within the joint observation error covariance matrix formed by pairwise combinations of active spaceborne lidar observation pixels and ground-based observation stations.

[0026] Step S106: Based on the prior multi-scale state vector, multi-scale observation vector, and joint observation error covariance matrix, the anthropogenic carbon emissions in the target area are collaboratively inverted with the goal of minimizing the value of the joint cost function, and the anthropogenic carbon emission inversion results are obtained.

[0027] The method for co-inverting anthropogenic carbon emissions using active lidar and ground-based observation provided in this invention constructs a multi-scale observation vector by fusing multi-source observation data from active spaceborne lidar and ground-based observation stations. Based on the attribute information of the observation data, a joint observation error covariance matrix is ​​constructed, which includes active observation errors, ground-based observation errors, and cross-observation errors, thus achieving a complete characterization of the multi-source observation error structure. On this basis, the multi-scale state vector, multi-scale observation vector, and joint observation error covariance matrix are uniformly incorporated into the co-inversion framework. The inversion solution is performed with the goal of minimizing the joint cost function, so that information from different observation sources can be reasonably weighted according to their respective error characteristics, effectively reducing the uncertainty in the inversion process and significantly improving the accuracy of the anthropogenic carbon emissions inversion results for the target area.

[0028] For ease of understanding, this embodiment of the invention provides a specific implementation method for the collaborative inversion of anthropogenic carbon emissions using active lidar and ground-based observation.

[0029] (I) Error dynamic construction method of active lidar and ground-based observation.

[0030] Most traditional assimilation systems construct their observation error covariance matrix using a diagonal matrix, where the elements on the main diagonal correspond to the error variance of each observation point, while all off-diagonal elements are set to 0. This implies that the errors at all observation points are completely independent, which is overly simplistic and does not reflect the actual physical processes of observation. This invention introduces satellite off-diagonal observation error covariance. This method considers the vertical error propagation characteristics of the atmosphere, the horizontal atmospheric transport correlation, and the spatial modulation effect of the error caused by the kernel function path integral during column concentration inversion in satellite lidar observations. This makes the construction of the observation data covariance introduced into the carbon assimilation system more consistent with reality.

[0031] In one implementation, it includes: (1.1) Based on the cylinder density kernel function weights, laser echo noise parameters, and spatial distances between active spaceborne lidars, the active observation error covariance matrix is ​​constructed. The specific implementation process is as follows: The following operations are performed for the first and second active spaceborne lidar: (1.11) Based on the laser echo noise estimate and atmospheric vertical height coordinates of the first active spaceborne lidar, and the laser echo noise estimate and atmospheric vertical height coordinates of the second active spaceborne lidar, determine the layered error covariance between the first active spaceborne lidar and the second active spaceborne lidar.

[0032] The laser echo noise estimate refers to the quantized noise value corresponding to the intensity fluctuation of the laser echo signal received by an active spaceborne lidar at a specific atmospheric vertical altitude layer. Atmospheric vertical altitude coordinates refer to the geometric center height of each layer in the atmospheric vertical stratification covered by the active spaceborne lidar observations. Layer error covariance refers to the statistical covariance of errors between the first and second active spaceborne lidars at the same atmospheric vertical altitude layer due to the spatial correlation of their laser echo noise estimates.

[0033] In one implementation, the atmospheric vertical altitude coordinates of the first active spaceborne lidar observation point are subtracted from the atmospheric vertical altitude coordinates of the second active spaceborne lidar observation point, and the absolute value of this difference is taken as the absolute altitude difference. Dividing this absolute altitude difference by a preset vertical error correlation length yields a dimensionless ratio, denoted as the vertical direction ratio. Taking the negative of the vertical direction ratio and performing a natural exponential operation on it yields a dimensionless attenuation coefficient between 0 and 1. Multiplying the estimated laser echo noise of the first active spaceborne lidar observation point, the estimated laser echo noise of the second active spaceborne lidar observation point, and the aforementioned dimensionless attenuation coefficient gives the resulting value as the layer error covariance between the first and second active spaceborne lidar observation points at that atmospheric vertical altitude layer.

[0034] (1.12) Based on the ratio of the spatial distance between the first active spaceborne lidar and the second active spaceborne lidar to the horizontal error correlation scale, the active attenuation factor between the first active spaceborne lidar and the second active spaceborne lidar is determined. The horizontal error correlation scale is the spatial characteristic length characterizing the error correlation between the observations of the first active spaceborne lidar and the second active spaceborne lidar, which is determined by the horizontal transport process of the atmospheric boundary layer.

[0035] The active attenuation factor is a dimensionless attenuation coefficient used to characterize the degree to which the error correlation between the first and second active spaceborne lidar observations weakens as the spatial distance increases.

[0036] In one implementation, the Euclidean straight-line distance between the two observation points on the Earth's surface projection plane is calculated in the WGS84 coordinate system based on their geographic latitude and longitude coordinates. This distance is denoted as the spatial distance between the two points. The horizontal error correlation scale is retrieved from the system's preset parameter library. This parameter is a constant, with units of kilometers, and its value is calibrated based on typical atmospheric boundary layer horizontal transport characteristics. The dimensionless ratio between the spatial distance and the horizontal error correlation scale is determined. After inverting this ratio, a natural exponentiation operation is performed with the natural constant as the base and the inverted ratio as the exponent to obtain the active attenuation factor.

[0037] (1.13) Determine the radar error term between the active spaceborne lidar observation and the second active spaceborne lidar based on the column concentration kernel function weight, stratification error covariance and active attenuation factor.

[0038] The column concentration kernel function weight refers to the integral weight assigned to the concentration contribution of the active spaceborne lidar to different atmospheric vertical layers during the carbon dioxide column concentration inversion process, reflecting the spatial response characteristics of the kernel function. This weight matrix is ​​generated by the sequential interaction of the column integral weight matrix, the satellite average kernel function, and the layer interpolation matrix. The layer interpolation matrix maps the carbon dioxide concentration output from the model from the original vertical layers to the vertical resolution layers required for lidar inversion. The satellite average kernel function characterizes the average response characteristics of the lidar to each vertical layer under standard atmospheric conditions. The column integral weight matrix is ​​used to weighted summation of the interpolated and kernel function-modulated layer responses to obtain the total column response.

[0039] In one implementation, an atmospheric transport model is run, using a priori emission field as input, to drive the model to complete a forward simulation and output a three-dimensional carbon dioxide stratified concentration field simulated by the model within the target time period. Its vertical dimension follows the original vertical stratification definition of the model. Based on the vertical stratification scheme used in lidar inversion (e.g., 500-meter equidistant stratification), coordinates are aligned with the original vertical stratification of the atmospheric transport model. Following the principle of conservation interpolation, the area-weighted allocation coefficients of each target grid on the original model layer are calculated to form a layer interpolation matrix. The satellite average kernel function is read from the lidar on-orbit calibration database. This function is expressed in discrete form, with the independent variable being the atmospheric vertical height and the dependent variable being the normalized response intensity of the corresponding layer, and has been pre-interpolated according to the lidar inversion vertical stratification. Based on the pressure difference distribution between the surface pressure and the top of the atmosphere, the column integral weights are calculated according to the proportion of each vertical layer's pressure range, forming a column vector with the same number of rows as the lidar inversion vertical stratification. Multiply the layer interpolation matrix by the corresponding elements of the satellite average kernel function to obtain the intermediate response matrix; then sum the intermediate response matrix with the column integral weight matrix column-weighted to finally generate a column vector, which is the column concentration kernel function weight matrix.

[0040] According to the rules of matrix multiplication, multiplying the column concentration kernel function weight matrix with the transpose of the column concentration kernel function weight matrix yields a square matrix, the number of rows and columns of which are equal to the number of vertical atmospheric strata.

[0041] The matrix, the aforementioned layered error covariance, and the aforementioned active attenuation factor are multiplied sequentially according to the scalar and matrix multiplication rules: first, the layered error covariance is multiplied by each element of the matrix, and then the resulting new matrix is ​​multiplied by the active attenuation factor; the element values ​​in the final matrix corresponding to the row and column positions of the first and second active spaceborne lidar observation point numbers are taken as the radar error term.

[0042] (1.14) Based on the laser echo noise parameters, determine the radar noise term between the active spaceborne lidar observation and the second active spaceborne lidar; Laser echo noise parameters refer to a set of fundamental parameters describing the stability of the echo signal of an active spaceborne lidar, obtained from factory calibration and on-orbit verification. These parameters include the instrument noise variance at each observation point. The instrument noise variance is the square of the random error intensity of the active spaceborne lidar's own detection system within the laser echo noise parameters; it represents the inherent error variance of a single observation point. The radar noise term refers to the error component caused solely by the random noise of the active spaceborne lidar instrument itself, which does not change with spatial distance. Its value is equal to the instrument noise variance at the corresponding observation point (when the two observation points are the same), or 0 (when the two observation points are different).

[0043] In one implementation, the radar noise term is determined by the instrument noise variance of the first active spaceborne lidar observation point: when the first active spaceborne lidar observation point and the second active spaceborne lidar observation point are the same observation point, the radar noise term is taken as the instrument noise variance corresponding to that observation point; when the two are different observation points, the radar noise term is 0.

[0044] (1.15) The sum of the radar error term and the radar noise term is used as the active observation error value between the active spaceborne lidar observation and the second active spaceborne lidar to construct the active observation error covariance matrix.

[0045] The active observation error value refers to the sum of the radar error term and the radar noise term, which serves as the unique filling value for the corresponding row and column positions in the active observation error covariance matrix, representing the total error covariance between the pair of observation points. The active observation error covariance matrix is ​​a symmetrical square matrix composed of pairwise combinations of all active spaceborne lidar observation pixels, with each element being an active observation error value.

[0046] Based on this, the first active spaceborne lidar With the second active spaceborne laser radar The active observation error covariance matrix between The expression is as follows: ; in, The column concentration kernel function weight matrix; The stratified concentration error covariance; For the first The and the first Spatial distance between observations; The relative length of the horizontal error; For the noise variance of the active lidar instrument; The Kronecker delta function is used to determine whether the first active spaceborne lidar observation point and the second active spaceborne lidar observation point are the same observation point: when the observation cell numbers of the two observation points are exactly the same, the coefficient is 1; when the observation cell numbers of the two observation points are different, the coefficient is 0.

[0047] In this embodiment of the invention, considering that the concentration of passive remote sensing satellite columns is mainly affected by spectral inversion errors, while the concentration error of active lidar columns is simultaneously affected by factors such as laser path attenuation, boundary layer aerosol accumulation, and echo energy fluctuations; and considering that although two radar monitoring stations are located at different locations, they may pass through the same boundary layer, be subject to the same aerosol disturbances, and be affected by factors such as using similar kernel function weights, their errors are not independent. Therefore, this embodiment of the invention proposes a specific implementation method for constructing the active observation error covariance matrix, breaking through the limitations of traditional fixed errors and providing a more flexible error calculation method to improve the inversion accuracy of carbon source and sink.

[0048] (1.2) Based on the standard deviation of the observation error of the ground observation stations, the representative correlation scale of the ground, and the station distance between the ground observation stations, the ground observation error covariance matrix is ​​constructed. The specific implementation process is as follows: The following operations will be performed for the first and second ground-based observation stations: (1.21) Determine the product of standard deviations based on the standard deviation of the observation error of the first ground observation station and the standard deviation of the observation error of the second ground observation station.

[0049] The standard deviation of observation error is used to characterize the statistical dispersion of the observed carbon dioxide concentration values ​​at ground-based observation stations around the true value. The product of standard deviations is a non-negative real number obtained by multiplying the standard deviations of the observation errors at the first ground-based observation station by the standard deviations of the observation errors at the second ground-based observation station.

[0050] (1.22) Determine the foundation attenuation factor based on the ratio of the distance between the first foundation observation station and the second foundation observation station to the representative relevant scale of the foundation.

[0051] The ground-based representative correlation scale is the spatial characteristic length characterizing the error correlation between observations from the first and second ground-based observation stations. It is jointly determined by the atmospheric boundary layer transport scale, topographic shielding effect, and spatial scale of the local meteorological system in the region where the ground-based observation stations are located. The ground-based attenuation factor is a dimensionless coefficient used to characterize the degree to which the error correlation between observations from the first and second ground-based observation stations weakens with increasing station distance.

[0052] Based on the geographic latitude and longitude coordinates of the first and second ground-based observation stations, the Euclidean straight-line distance between them on the surface projection plane in the WGS84 coordinate system is calculated and denoted as the station distance. The representative ground-related scale is retrieved from the system's preset parameter library. This parameter is a constant, with units of kilometers, and its value is determined based on the calibration results of the typical land surface observation network in the target area and the spatial characteristics of the boundary layer. The inverse of the ratio between the station distance and the representative ground-related scale is taken, and a natural exponentiation operation is performed with the natural constant as the base and the inverse as the exponent. A positive real number between 0 and 1 is output, which is the ground attenuation factor between the first and second ground-based observation stations.

[0053] (1.23) Multiply the standard deviation product with the foundation attenuation factor, and use the resulting product as the foundation observation error covariance between the first foundation observation station and the second foundation observation station to construct the foundation observation error covariance matrix.

[0054] Ground-based observation error covariance: refers to the statistical correlation of errors between the first and second ground-based observation stations caused by common atmospheric processes (such as boundary layer transport and regional weather systems). It is a fundamental element constituting the ground-based observation error covariance matrix. The ground-based observation error covariance matrix is ​​a symmetric matrix formed by pairwise combinations of all ground-based observation stations, where each element represents a ground-based observation error covariance.

[0055] Based on this, the embodiments of the present invention consider that the ground-based observation errors mainly originate from instrument errors, subgrid transport errors, and topographic representativeness errors; adjacent ground-based observation stations are affected by boundary layer transport processes and regional weather systems, and their observation errors have spatial correlations. Therefore, the ground-based observation error covariance can be determined according to the implementation method provided in (1.2) above. Specifically, the first ground-based observation station... With the second ground-based observation station Covariance of ground observation errors between The expression is as follows: ; in Distance between stations; For representative relevant scales of the foundation; The corresponding number is Standard deviation of observation error for each foundation observation point The corresponding number is The standard deviation of the observation error at each ground observation point.

[0056] (1.3) Based on the correlation coefficient of observation errors between the active spaceborne lidar and the ground-based observation station, the standard deviation of the observation error of the active spaceborne lidar, and the standard deviation of the observation error of the ground-based observation station, a cross-observation error covariance matrix is ​​constructed. The specific implementation process is as follows: The following operations should be performed for active spaceborne lidar and ground-based observation stations: (1.31) Determine the spatial attenuation factor based on the ratio of the spatial distance between the active spaceborne lidar and the ground-based observation station to the scale of the horizontal error.

[0057] The horizontal error correlation scale is the spatial characteristic length characterizing the error correlation between active spaceborne lidar observations and ground-based observation station observations. It is determined by the regional-scale atmospheric horizontal transport process and is measured in kilometers. The spatial attenuation factor is a dimensionless coefficient used to characterize the degree to which the spatial error correlation between active spaceborne lidar observations and ground-based observation station observations weakens with increasing spatial distance.

[0058] In one implementation, based on the geographic latitude and longitude coordinates of the active spaceborne lidar observation point and the ground-based observation station, the Euclidean straight-line distance between them on the surface projection plane in the WGS84 geodetic coordinate system is calculated and denoted as the spatial distance between the lidar and the station. The horizontal error correlation scale is retrieved from the system's preset parameter library. This parameter is a calibrated constant in kilometers, and its value reflects the length of the spatial correlation characteristic of errors dominated by the regional-scale atmospheric horizontal transport process. The calculated spatial distance is divided by this horizontal error correlation scale to obtain a dimensionless ratio. The negative of this ratio is then taken. An exponential operation is performed with the natural constant as the base and the negative of this ratio as the exponent. The result is the spatial attenuation factor between the active spaceborne lidar and the ground-based observation station. This factor is a positive real number between 0 and 1, used to quantitatively characterize the degree of correlation attenuation of the observation errors in the spatial dimension: when the spatial distance is much smaller than the horizontal error correlation scale, the factor approaches 1, indicating a high degree of error correlation; when the spatial distance is significantly larger than this scale, the factor approaches 0, indicating that the errors are essentially unrelated.

[0059] (1.32) Determine the time decay factor by the ratio of the absolute value of the time difference between the observation time of the active spaceborne lidar and the observation time of the ground-based observation station to the time error related scale.

[0060] The observation time refers to the timestamp when an active spaceborne lidar completes a valid observation, or the timestamp when a ground-based observation station acquires a valid concentration observation. The time error correlation scale is the time characteristic length characterizing the error correlation between active spaceborne lidar observations and ground-based observation station observations. It is jointly determined by the diurnal variation period of the atmospheric boundary layer and the time scale of regional weather system evolution, and is measured in hours. The time decay factor is a dimensionless coefficient used to characterize the degree to which the time correlation between active spaceborne lidar observations and ground-based observation station observations weakens as the time difference increases.

[0061] In one implementation, the timestamp corresponding to a single effective carbon dioxide concentration observation completed by the active spaceborne lidar is extracted and recorded as the lidar's observation time; simultaneously, the timestamp corresponding to a single effective concentration observation data acquired by a ground-based observation station is extracted and recorded as the station's observation time. The difference between these two timestamps is calculated, and their absolute value is taken to obtain the deviation of the two observations on the time axis, in hours. A time error correlation scale is read from the system's preset parameter library. This parameter is a calibrated constant, in hours, whose value comprehensively reflects the error time correlation characteristic length determined by the diurnal variation cycle of the atmospheric boundary layer and the evolution process of the regional weather system. The absolute value of the aforementioned time difference is divided by this time error correlation scale to obtain a dimensionless ratio. The negative of this ratio is then taken. An exponential operation is performed with the natural constant e as the base and the negative of this ratio as the exponent. The result is the time decay factor between the active spaceborne lidar and the ground-based observation station. This factor is a positive real number between 0 and 1, used to quantitatively characterize the degree of correlation decay of the observation errors of the two in the time dimension: when the time difference is much smaller than the time error correlation scale, the factor is close to 1, indicating that the error has strong time consistency; when the time difference is significantly larger than the scale, the factor approaches 0, indicating that the error is basically decoupled in time.

[0062] (1.33) Multiply the spatial decay factor by the temporal decay factor, and use the product as the correlation coefficient of the observation error between the active spaceborne lidar and the ground-based observation station; The observation error correlation coefficient is a dimensionless comprehensive measure that comprehensively reflects the spatial and temporal correlation of errors between active spaceborne lidar observations and ground-based observation stations.

[0063] (1.34) Multiply the observation error correlation coefficient, the observation error standard deviation of the active spaceborne lidar, and the observation error standard deviation of the ground-based observation station, and use the product as the cross-observation error covariance between the active spaceborne lidar and the ground-based observation station to construct the cross-observation error covariance matrix.

[0064] Cross-observation error covariance refers to the statistical correlation of errors between active spaceborne lidar observations and ground-based observations caused by common atmospheric processes (such as anthropogenic plumes, boundary layer accumulation, and regional transport). It is a fundamental element constituting the cross-observation error covariance matrix. The cross-observation error covariance matrix is ​​a rectangular sub-block matrix formed by pairwise combinations of all active spaceborne lidar observation pixels and all ground-based observation stations, with each element representing a cross-observation error covariance.

[0065] Based on this, active spaceborne lidar observation Observations with ground-based observation stations Cross-observation error covariance matrix The expression is as follows: , ; in Represents spatial distance. For horizontal conveying dimensions, Represents time difference, Represents a time-related scale. The correlation coefficient represents the observation error. This represents the standard deviation of the observation error of the active spaceborne lidar. This represents the standard deviation of the observation error at the ground-based observation station.

[0066] (1.4) Construct the joint observation error covariance matrix based on the active observation error covariance matrix, the ground observation error covariance matrix, and the cross-observation error covariance matrix. The joint observation error covariance matrix It consists of four logical sub-blocks, forming a symmetrical block structure: the upper left sub-block is the "active observation error covariance matrix". ", its rows and columns correspond to all observed pixels of the active spaceborne lidar, the first in the matrix Line number The elements of the column represent the first... The lidar pixel and the first The covariance of observation errors between individual lidar pixels; the lower right sub-block represents the "covariance matrix of ground-based observation errors". ", its rows and columns correspond to all ground-based observation stations, the first in the matrix Line number The elements of the column represent the first... The first base station and the first The covariance of observation errors between individual ground stations; the upper right sub-block is the "cross-observation error covariance matrix". ", its rows correspond to all active spaceborne lidar observation pixels, and its columns correspond to all ground-based observation stations, the first in the matrix..." Line number The elements of the column represent the first... The lidar pixel and the first The covariance of observation errors among the ground stations; the lower left sub-block represents the aforementioned cross-observation error covariance matrix. The transpose of, i.e., its first... Line number The elements of the column and the top right sub-block Line number The elements in the column are exactly the same.

[0067] This invention achieves joint assimilation of active lidar observations and ground-based data. Therefore, there is a correlation between the observed column concentrations from ground-based and satellite sources. Traditional joint assimilation typically assumes... This means that while the two types of observations are completely independent, they are actually affected by factors such as urban emission plumes, boundary layer accumulation, and weather conditions, which does not hold true in regional carbon assimilation. Therefore, this invention uses multi-source observation error coupling modeling to form a joint covariance that differs from the traditional approach, thereby further improving the accuracy of carbon source-sink assimilation and inversion.

[0068] (ii) The multi-scale observation vector, multi-scale state vector, multi-scale observation operator, and multi-scale joint cost function are all obtained by decomposing the large-scale background transport, regional medium-scale transport, and local anthropogenic emission disturbance.

[0069] Current assimilation methods perform assimilation and inversion at a single scale. This invention not only performs scale conversion between model concentration and satellite column concentration, but also proposes a "joint assimilation method based on multi-scale decomposition." This method can effectively enhance the sensitivity of active lidar to boundary layer anthropogenic emissions, improve the ability to identify spatial details of emissions at the urban scale, and reduce inversion uncertainties in sparsely observed regions. Based on large-scale background transport, regional medium-scale transport, and local anthropogenic emission disturbances, CO2 concentration observations are... Decomposed into CO2 concentration observations at a large-scale background transport scale CO2 concentration observation at regional medium-scale transport scale CO2 concentration observation at the scale of local anthropogenic emissions disturbance Among them, the ground-based station More sensitive and More sensitive to active lidar column concentration, these two types of observation data complement each other in scale, each leveraging its own advantages.

[0070] Similarly, based on large-scale background transport, regional medium-scale transport, and local anthropogenic emission disturbances, the state vector (anthropogenic carbon emissions) is transformed from the traditional... Expanded to deliver state vectors at a large-scale background. State vectors at the regional scale of transmission State vector at the scale of local anthropogenic emission disturbance .

[0071] Similarly, based on large-scale background transport, regional medium-scale transport, and local anthropogenic emission disturbances, the observation operator is extended to an observation operator at the large-scale background transport scale. Observation operators at regional medium-scale transport scale Observation operators at the scale of local anthropogenic emission disturbances The operator exhibits a strictly diagonally dominant structure: non-zero values ​​are set only at the positions of the first row, first column, second row, second column, and third row, third column. These values ​​represent the physical response intensity or transformation weights corresponding to the large-scale background transport process, the regional medium-scale transport process, and the local anthropogenic emission disturbance process, respectively. All other positions, including the first row, second column, first row, third column, second row, first column, second row, third column, third row, first column, and third row, second column, are explicitly set to 0. This structure indicates that emission state variables at each scale only affect the concentration response at the same scale through their corresponding physical mechanisms, and there is no direct cross-mapping or aliasing effect between different scales. This rigorously ensures the orthogonality of the multi-scale decomposition and the consistency of the physical interpretation at the mathematical modeling level.

[0072] Similarly, based on large-scale background transport, regional medium-scale transport, and local anthropogenic emission disturbances, the cost function is extended to the scale of large-scale background transport. Cost function under regional scale transmission Cost function under the scale of local anthropogenic emission disturbance The final cost function is the sum of the three cost functions mentioned above.

[0073] (iii) Joint assimilation of anthropogenic carbon emissions based on active lidar and ground-based systems.

[0074] In one implementation: Using a multi-scale observation operator, the prior multi-scale state vector is transformed into the observation space, resulting in... Concentration simulation data, and on Concentration simulation data and active Concentration observation data, foundation Spatiotemporal matching of concentration observation data; By using a pre-constructed multi-scale joint cost function, based on the matched... Concentration simulation data, multi-scale observation vectors and Concentration simulation data are used to perform a collaborative inversion of anthropogenic carbon emissions in the target area, minimizing the value of the multi-scale joint cost function. The multi-scale state vectors corresponding to the concentration simulation data are used as the results of anthropogenic carbon emission inversion.

[0075] Specifically: The "error dynamic construction method" and "multi-scale decomposition joint assimilation method" proposed in this invention are introduced into the regional carbon source and sink monitoring verification support system. The cost function is constructed by the difference between the simulated value and the observed value of the model, and anthropogenic carbon emissions are calculated by the local ensemble Kalman filter method.

[0076] Furthermore, during assimilation and inversion, a spatial scale mismatch exists between the CO2 concentration simulated by the model and the observed CO2 concentration data from multiple sources. Therefore, spatial matching processing is required before assimilation and inversion. Matching for ground-based observation stations is relatively simple, as the simulated values ​​can be interpolated based on the station locations. However, for Active Spaceborne Radar (ACDL) observations, a satellite kernel function scaling transformation method is needed to achieve scale matching between the simulated concentration and the satellite column concentration, according to the required assimilation accuracy. The scaling transformation process is as follows: (1) Interpolate the CO2 stratification concentration simulated by the model to the stratification height concentration observed by the active spaceborne radar.

[0077] (2) According to the column concentration conversion method, the simulated stratified concentration is converted into the simulated satellite column concentration.

[0078] Finally, greenhouse gas observation network data and active spaceborne radar column concentration observation data for XCO2 are simultaneously introduced into the regional carbon assimilation system to perform assimilation and inversion of anthropogenic carbon emissions, such as... Figure 2 This diagram illustrates the functional relationship between anthropogenic carbon emissions and observed CO2 concentrations. This represents a pixel (region). Site Concentration Observation operators; active observation This represents a pixel (region). Satellite column concentration The observation operator. Different from the station concentration observation operator. Active spaceborne radar column concentration observation operator The flux data needs to be converted into CO2 stratified concentration values, and the simulated stratified concentration needs to be converted into simulated CO2 column concentration using the active spaceborne radar column concentration conversion formula. Represents a pixel (region) Anthropogenic carbon emissions; This represents atmospheric CO2 concentration observations from ground-based and active spaceborne radar.

[0079] Optionally, other algorithms can also be used to achieve the coordinated inversion of anthropogenic carbon emissions. This embodiment of the invention does not limit the specific algorithm used.

[0080] Based on the foregoing embodiments, this invention provides a device for the coordinated inversion of anthropogenic carbon emissions using active lidar and ground-based observation. See [link to related documentation]. Figure 3 The schematic diagram shown illustrates a device for the coordinated inversion of anthropogenic carbon emissions using active lidar and ground-based observation, comprising: Observation vector construction module 402, based on the active spaceborne lidar corresponding to the active... Concentration observation data, and the foundation corresponding to the ground-based observation station. Concentration observation data are used to construct a multi-scale observation vector corresponding to the target area. Covariance construction module 404, used for active... Attribute information and foundation corresponding to concentration observation data Based on the attribute information corresponding to the concentration observation data, construct the joint observation error covariance matrix, which includes the active observation error covariance matrix, the ground-based observation error covariance matrix, and the cross-observation error covariance matrix. The inversion module 406 is used to perform a collaborative inversion of anthropogenic carbon emissions in the target area based on the prior multi-scale state vector, multi-scale observation vector, and joint observation error covariance matrix, with the goal of minimizing the value of the joint cost function, and obtain the anthropogenic carbon emission inversion result.

[0081] The active lidar and ground-based observation collaborative inversion device for anthropogenic carbon emissions provided in this invention constructs a multi-scale observation vector by fusing multi-source observation data from active spaceborne lidar and ground-based observation stations. Based on the attribute information of the observation data, it constructs a joint observation error covariance matrix that includes active observation errors, ground-based observation errors, and cross-observation errors, thus achieving a complete characterization of the multi-source observation error structure. On this basis, the multi-scale state vector, multi-scale observation vector, and joint observation error covariance matrix are uniformly incorporated into the collaborative inversion framework. The inversion solution is performed with the goal of minimizing the joint cost function, so that information from different observation sources can be reasonably weighted according to their respective error characteristics, effectively reducing the uncertainty in the inversion process and significantly improving the accuracy of the anthropogenic carbon emissions inversion results for the target area.

[0082] The device provided in this embodiment of the invention has the same implementation principle and technical effect as the aforementioned method embodiment. For the sake of brevity, any parts not mentioned in the device embodiment can be referred to the corresponding content in the aforementioned method embodiment.

[0083] This invention provides an electronic device, specifically, the electronic device includes a processor and a memory; the memory stores a computer program, which, when run by the processor, executes the method described in any of the above embodiments.

[0084] Figure 4 The present invention provides a schematic diagram of the structure of an electronic device 100, which includes a processor 50, a memory 51, a bus 52 and a communication interface 53. The processor 50, the communication interface 53 and the memory 51 are connected through the bus 52. The processor 50 is used to execute executable modules, such as computer programs, stored in the memory 51.

[0085] The memory 51 may include high-speed random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Communication between this system network element and at least one other network element is achieved through at least one communication interface 53 (which can be wired or wireless), such as the Internet, wide area network, local area network, metropolitan area network, etc.

[0086] Bus 52 can be an ISA bus, PCI bus, or EISA bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 4 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.

[0087] The memory 51 is used to store programs. After receiving an execution instruction, the processor 50 executes the programs. The method executed by the device for defining the flow process disclosed in any of the foregoing embodiments of the present invention can be applied to the processor 50 or implemented by the processor 50.

[0088] Processor 50 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of processor 50 or by instructions in software form. Processor 50 can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this invention can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory 51. The processor 50 reads the information in memory 51 and, in conjunction with its hardware, completes the steps of the above method.

[0089] The computer program product of the readable storage medium provided in the embodiments of the present invention includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the methods described in the foregoing method embodiments. For specific implementation, please refer to the foregoing method embodiments, which will not be repeated here.

[0090] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0091] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for co-inverting anthropogenic carbon emissions using active lidar and ground-based observation, characterized in that, include: Based on the active spaceborne lidar Concentration observation data, and the foundation corresponding to the ground-based observation station. Concentration observation data are used to construct a multi-scale observation vector corresponding to the target area. Based on the active The attribute information corresponding to the concentration observation data and the foundation The attribute information corresponding to the concentration observation data is used to construct a joint observation error covariance matrix, which includes the active observation error covariance matrix, the ground observation error covariance matrix, and the cross-observation error covariance matrix. Based on the prior multi-scale state vector, the multi-scale observation vector, and the joint observation error covariance matrix, the anthropogenic carbon emissions in the target region are collaboratively inverted with the goal of minimizing the value of the joint cost function, and the anthropogenic carbon emission inversion result is obtained.

2. The method for co-inverting anthropogenic carbon emissions using active lidar and ground-based observation as described in claim 1, characterized in that, Based on the active The attribute information corresponding to the concentration observation data and the foundation The attribute information corresponding to the concentration observation data is used to construct the joint observation error covariance matrix, including: Based on the column density kernel function weights, laser echo noise parameters, and spatial distances between the active spaceborne lidars, an active observation error covariance matrix is ​​constructed. Based on the standard deviation of the observation error of the ground observation stations, the representative correlation scale of the ground, and the station distance between the ground observation stations, a ground observation error covariance matrix is ​​constructed. Based on the correlation coefficient of observation errors between the active spaceborne lidar and the ground-based observation station, the standard deviation of the observation error of the active spaceborne lidar, and the standard deviation of the observation error of the ground-based observation station, a cross-observation error covariance matrix is ​​constructed. A joint observation error covariance matrix is ​​constructed based on the active observation error covariance matrix, the ground-based observation error covariance matrix, and the cross-observation error covariance matrix.

3. The method for co-inverting anthropogenic carbon emissions using active lidar and ground-based observation as described in claim 2, characterized in that, Based on the column density kernel function weights, laser echo noise parameters, and spatial distances between the active spaceborne lidars, an active observation error covariance matrix is ​​constructed, including: The following operations are performed for the first and second active spaceborne lidar: Based on the laser echo noise estimate and atmospheric vertical height coordinates of the first active spaceborne lidar, and the laser echo noise estimate and atmospheric vertical height coordinates of the second active spaceborne lidar, the layered error covariance between the first active spaceborne lidar and the second active spaceborne lidar is determined. Based on the ratio of the spatial distance between the first active spaceborne lidar and the second active spaceborne lidar to the horizontal error correlation scale, the active attenuation factor between the first active spaceborne lidar and the second active spaceborne lidar is determined. The horizontal error correlation scale is the spatial characteristic length characterizing the error correlation between the observations of the first active spaceborne lidar and the second active spaceborne lidar. The radar error term between the active spaceborne lidar observation and the second active spaceborne lidar is determined based on the column concentration kernel function weight, the stratification error covariance, and the active attenuation factor. Based on the laser echo noise parameters, the radar noise term between the observation of the active spaceborne lidar and the second active spaceborne lidar is determined; The sum of the radar error term and the radar noise term is used as the active observation error value between the active spaceborne lidar observation and the second active spaceborne lidar, so as to construct the active observation error covariance matrix.

4. The method for co-inverting anthropogenic carbon emissions using active lidar and ground-based observation as described in claim 2, characterized in that, Based on the standard deviation of the observation error of the ground observation stations, the representative correlation scale of the ground, and the station distance between the ground observation stations, a ground observation error covariance matrix is ​​constructed, including: The following operations will be performed for the first and second ground-based observation stations: The standard deviation product is determined based on the standard deviation of the observation error of the first ground observation station and the standard deviation of the observation error of the second ground observation station. The ground attenuation factor is determined based on the ratio of the distance between the first ground observation station and the second ground observation station to the representative correlation scale of the ground. The representative correlation scale of the ground is the spatial characteristic length that characterizes the error correlation between the observations of the first ground observation station and the observations of the second ground observation station. The product of the standard deviations is multiplied by the foundation attenuation factor, and the resulting product is used as the foundation observation error covariance between the first foundation observation station and the second foundation observation station to construct the foundation observation error covariance matrix.

5. The method for co-inverting anthropogenic carbon emissions using active lidar and ground-based observation as described in claim 2, characterized in that, Based on the correlation coefficient of observation errors between the active spaceborne lidar and the ground-based observation station, the standard deviation of the observation error of the active spaceborne lidar, and the standard deviation of the observation error of the ground-based observation station, a cross-observation error covariance matrix is ​​constructed, including: For active spaceborne lidar and ground-based observation stations, the following operations shall be performed: The spatial attenuation factor is determined based on the ratio of the spatial distance between the active spaceborne lidar and the ground-based observation station to the horizontal error correlation scale. The horizontal error correlation scale is the spatial characteristic length that characterizes the error correlation between the active spaceborne lidar observation and the ground-based observation station observation. The time decay factor is determined by the ratio of the absolute value of the time difference between the observation time of the active spaceborne lidar and the observation time of the ground-based observation station to the time error correlation scale. The time error correlation scale is the time characteristic length that characterizes the error correlation between the observation of the active spaceborne lidar and the observation of the ground-based observation station. Multiply the spatial attenuation factor by the time attenuation factor, and the product is used as the observation error correlation coefficient between the active spaceborne lidar and the ground-based observation station; The observation error correlation coefficient, the observation error standard deviation of the active spaceborne lidar, and the observation error standard deviation of the ground-based observation station are multiplied together, and the resulting product is used as the cross-observation error covariance between the active spaceborne lidar and the ground-based observation station to construct the cross-observation error covariance matrix.

6. The method for co-inverting anthropogenic carbon emissions using active lidar and ground-based observation as described in claim 1, characterized in that, Based on the prior multi-scale state vector, the multi-scale observation vector, and the joint observation error covariance matrix, and with the objective of minimizing the joint cost function, a collaborative inversion of anthropogenic carbon emissions in the target region is performed to obtain the anthropogenic carbon emission inversion results, including: Using a multi-scale observation operator, the prior multi-scale state vector is transformed into the observation space, resulting in... Concentration simulation data, and on the Concentration simulation data and the active Concentration observation data, the foundation Spatiotemporal matching of concentration observation data; By using a pre-constructed multi-scale joint cost function, based on the matched... Concentration simulation data, the multi-scale observation vector, and the Concentration simulation data are used to perform a collaborative inversion of anthropogenic carbon emissions in the target region, and the value of the multi-scale joint cost function is minimized. The multi-scale state vectors corresponding to the concentration simulation data are used as the results of anthropogenic carbon emission inversion.

7. The method for co-inverting anthropogenic carbon emissions using active lidar and ground-based observation according to any one of claims 1-6, characterized in that, The multi-scale observation vector, multi-scale state vector, multi-scale observation operator, and multi-scale joint cost function are all obtained by decomposing the large-scale background transport, regional medium-scale transport, and local anthropogenic emission disturbances.

8. A device for co-inverting anthropogenic carbon emissions using active lidar and ground-based observation, characterized in that, include: The observation vector construction module is based on the active spaceborne lidar corresponding to the active... Concentration observation data, and the foundation corresponding to the ground-based observation station. Concentration observation data are used to construct a multi-scale observation vector corresponding to the target area. Covariance construction module, used to construct based on the active... The attribute information corresponding to the concentration observation data and the foundation The attribute information corresponding to the concentration observation data is used to construct a joint observation error covariance matrix, which includes the active observation error covariance matrix, the ground observation error covariance matrix, and the cross-observation error covariance matrix. The inversion module is used to perform a collaborative inversion of anthropogenic carbon emissions in the target region based on the prior multi-scale state vector, the multi-scale observation vector, and the joint observation error covariance matrix, with the goal of minimizing the value of the joint cost function, to obtain the anthropogenic carbon emission inversion result.

9. An electronic device, characterized in that, The method includes a processor and a memory, the memory storing computer-executable instructions executable by the processor, the processor executing the computer-executable instructions to implement the method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions that, when invoked and executed by a processor, cause the processor to perform the method according to any one of claims 1 to 7.

Images