A 14c-based urban carbon emission verification and inventory checking system and method

Abstract

This invention relates to the field of carbon emission monitoring technology, specifically disclosing a method based on... 14 The present invention relates to a city carbon emission verification and inventory validation system and method, comprising a data input layer, an inversion calculation layer, an audit verification layer, an uncertainty assessment layer, and a visualization output layer. The data input layer receives and preprocesses multi-source input data from the target area. The inversion calculation layer, connected to the data input layer, performs Bayesian inversion optimization of emission fluxes. The audit verification layer, connected to the inversion calculation layer, performs independent verification and attribution analysis of emission results. The invention introduces... 14 C isotope tracing technology constructs an observation constraint system independent of the traditional bottom-up statistical path. It can objectively verify the spatial allocation rationality and total accuracy of existing emission inventories, overcome the inherent defects of a single inventory that cannot be self-verified due to differences in statistical caliber and proxy variables, and free the audit results from excessive reliance on the prior assumptions of the inventory.

Description

Technical Field

[0001] This invention relates to the field of carbon emission monitoring technology, and more particularly to a method based on... 14 C. Urban carbon emission verification and inventory validation system and method. Background Technology

[0002] Currently, the carbon emission accounting system supporting target tracking still relies heavily on bottom-up statistical data and emission factor methods, lacking objective verification methods independent of emission inventories. At the same time, the application of atmospheric inversion technology at the urban scale is limited by factors such as sparse observation networks, large background signal interference, and strong spatial heterogeneity of emissions, making it difficult to effectively distinguish between fossil and biological emissions.

[0003] The existing urban carbon emission audits and inventory verification mainly face the following technical challenges: First, bottom-up emission inventories exhibit significant systematic biases at the regional scale due to spatial allocation proxy variables and differences in statistical methods. Furthermore, the inventories themselves lack independent verification capabilities and cannot determine the authenticity of the spatial distribution of emissions. Secondly, although radiocarbon isotope tracing can effectively distinguish between fossil-derived and biogenic carbon dioxide, the existing integrated schemes for observation and inversion have not yet formed a systematic process for fine-scale urban environments. This includes the lack of standardized methods for observation network design, high-resolution transmission modeling, and prior error construction. Third, the inversion results are highly sensitive to the error covariance parameter and spatial correlation assumptions. Existing studies are mostly single-parameter configurations and lack a systematic uncertainty quantification framework, which makes it difficult for audit results to provide traceable confidence intervals, which is not conducive to the generation and decision-making application of standardized audit reports. Therefore, a method based on 14 C's carbon emission audit and inventory verification system and method are proposed to address the issues raised above. Summary of the Invention

[0004] To overcome the shortcomings of the prior art, the present invention provides a method based on... 14 The urban carbon emission verification and inventory validation system and method of C can effectively solve the problems involved in the above background technology.

[0005] The objective of this invention can be achieved through the following technical solution: Firstly, this invention provides a method based on... 14 C's urban carbon emission verification and inventory validation system includes a data input layer, an inversion calculation layer, an audit validation layer, an uncertainty assessment layer, and a visualization output layer; The data input layer is used to receive and preprocess multi-source input data of the target region, providing high-quality input data in a unified format for subsequent inversion calculations, and ensuring data consistency and reliability. The inversion calculation layer, connected to the data input layer, is used to perform Bayesian inversion optimization of emission flux, achieve optimal estimation of fossil fuel emission flux, and improve the estimation accuracy of emissions. The audit verification layer, connected to the inversion calculation layer, is used to perform independent verification and attribution analysis of emission results, providing objective verification results that do not depend on the inventory, and enhancing the credibility of emission data; The uncertainty assessment layer, connected to the inversion calculation layer, is used to quantify the uncertainty of the inversion results, clarify the reliability of the emission estimation, and avoid deviations caused by single parameter assumptions. The visualization output layer, connected to the audit verification layer and the uncertainty assessment layer, is used to generate audit result products, intuitively presenting the spatial distribution of emissions and audit conclusions, and supporting carbon governance decision-making needs. The uncertainty assessment layer includes: The observation uncertainty assessment module is configured to be based on 14 C. Measurement error and atmospheric transport model error are used to construct the observation error covariance matrix R, which quantifies the random errors in the observation and transport process and ensures the accuracy of the inversion input. The prior sensitivity testing module is configured to perform multiple inversion comparisons by changing the source of the prior list and adjusting the scaling factor of the error covariance matrix B, thereby quantifying the impact of prior hypotheses on posterior results, identifying the sensitivity of prior hypotheses to posterior results, and improving the interpretability of inversion results. The spatial correlation sensitivity test module is configured to set different spatial correlation length parameters for the observation error covariance matrix R and the prior error covariance matrix B, execute multiple sets of inversion experiments, evaluate the impact of the spatial correlation assumption on the inversion stability, clarify the impact of the spatial correlation assumption on the inversion stability, and optimize the selection of spatial smoothness. The aggregate inversion statistics module is configured to summarize the posterior emission results of all sensitivity tests, calculate their mean and standard deviation, output the probability distribution and confidence interval of emissions, provide traceable statistical confidence intervals, and enhance the scientific rigor and credibility of audit reports.

[0006] Preferably, the data input layer includes 14 C. Observation data module, emission inventory data interface module, and meteorological and environmental data interface module; Among them, the 14 The C observation data module is configured to receive data from the target area. 14 C observation data, execute Δ 14 C calculation, baseline deduction and Signal extraction and output target area By observing concentrations, interference from biological sources can be effectively eliminated, thus improving the accuracy of identifying CO2 concentrations from fossil sources. The emission inventory data interface module is configured to receive at least two types of bottom-up emission inventory data, perform format unification, spatial reprojection and grid interpolation, output a priori emission field under a unified grid resolution, and integrate multi-source inventory information to provide a diverse data foundation for the construction of a priori emission field. The meteorological and environmental data interface module is configured to receive atmospheric reanalysis meteorological field data and regional air quality monitoring data, perform data quality checks and time matching, output meteorological driving fields and collaborative pollutant verification datasets, ensure the driving accuracy of atmospheric transport simulation, and provide data support for spatial collaborative verification.

[0007] Preferably, the data input layer specifically performs the following steps: pass 14 The C observation data module receives at least one of the following: plant samples, atmospheric samples, or tree-ring samples from the target area. 14 C was observed in the medium and determined by accelerator mass spectrometry. 14 C / 12 Δ is calculated after the C ratio. 14 C value, based on the background Δ 14 The C value and background CO2 concentration are calculated using the following formula: Extract each sampling point Observed concentrations, among which This represents the background CO2 concentration. The atmospheric Δ reflected by the observation medium 14 C value, The base Δ 14 C value, fossil fuel Δ 14 The C-value (taken as -1000‰) effectively separates fossil-derived and biogenic CO2, achieving the target area. Precise quantification of concentration; The system receives at least two bottom-up emission inventory data through the emission inventory data interface module, performs format unification, spatial reprojection and bilinear interpolation on each inventory, unifies each inventory to the same spatial resolution grid, outputs at least two sets of prior emission flux fields, and provides emission values ​​and spatial coordinate indexes for each grid cell, eliminates format and coordinate system differences between multi-source inventories, and constructs a spatially consistent multi-scenario prior emission field. The system receives atmospheric reanalysis meteorological field data and regional air quality monitoring data through a meteorological and environmental data interface module. It performs data quality checks and time matching on the meteorological field data to output a meteorological driving field, and extracts PM2.5 from the regional air quality monitoring data. 2.5 PM 10Concentration fields of at least one co-pollutant among NO2, CO, and SO2 are used to construct a co-pollutant validation dataset, ensuring the spatiotemporal continuity of the meteorological driving field and providing multi-dimensional cross-validation data to support auxiliary judgment of emission sources.

[0008] Preferably, the inversion calculation layer includes: The prior emission field construction module is configured to construct a prior emission flux field and its error covariance matrix B based on emission inventory data, quantify the spatial uncertainty of prior emissions, and provide reasonable initial constraints for Bayesian inversion. The atmospheric transport operator construction module is configured to use the Lagrange particle diffusion model to release virtual particles at each observation point and backtrack them to generate a source-receptor relationship sensitivity field, construct the Jacobian matrix H, establish a quantitative response relationship between the emission source and the observation point, and realize the effective mapping of the observation signal to the emission space. The Bayesian assimilation engine module is configured to solve for the posterior emission flux x by minimizing the cost function, integrating prior information and observation constraints to obtain the optimal emission flux estimate that takes into account the advantages of both.

[0009] Preferably, the inversion calculation layer specifically performs the following steps: A priori emission flux field is constructed using a priori emission field construction module based on at least two emission inventory data at a uniform grid resolution. Calculate the absolute value of the emission difference between the two inventories in each grid cell, and multiply this absolute value by the scaling factor SF. B The prior error covariance matrix B is constructed by squaring the squares, i.e., B = (SF B ×|List 1-List 2|)², objectively reflecting prior uncertainty with the differences between multiple lists, and avoiding the bias caused by subjectively setting the error distribution; By using the Lagrange particle diffusion model through the atmospheric transport operator construction module, virtual particles are released at preset time intervals at each observation point and tracked backward for a preset number of days. The residence time of the virtual particles in the near-surface footprint layer is statistically analyzed to generate a source-receptor sensitivity field. Based on the source-receptor sensitivity field, a Jacobian matrix H is constructed. The near-surface footprint layer is the range of 0 to 100 meters above the ground. This finely characterizes the near-surface atmospheric transport process and ensures that the response relationship between the emission source and the observation point is accurate and reliable. The observation error covariance matrix R is constructed using the Bayesian assimilation engine module. R is a diagonal matrix, where the diagonal elements are the sum of the squares of measurement uncertainty and transmission error. This is achieved by minimizing the cost function. Solving for posterior emission fluxes The total regional emissions are obtained by multiplying the posterior emission flux by the grid area and the calculation period length, where... For the a priori emission flux field, For based on14 C observation value estimation concentration, The observation error covariance matrix, Given the prior error covariance matrix, we can coordinate the balance between observation constraints and prior constraints to obtain a posterior emission flux estimate that is both accurate and stable.

[0010] Preferably, the audit verification layer includes a multi-list cross-comparison module and a spatial collaborative verification module; The multi-list cross-comparison module is configured to compare the total posterior emissions with the total of each prior list grid by grid, calculate the percentage deviation of the total and the spatial deviation distribution, output the difference analysis results from top to bottom and bottom to top, reveal the systematic deviation characteristics between the posterior emissions and each prior list, and identify the areas where the lists are overestimated or underestimated. The spatial collaborative verification module is configured to perform a post-verification... Spatial correlation analysis is performed on the spatial field and the density distribution of industrial facilities in the target area and the concentration field of at least one co-polluting pollutant. The spatial consistency between emission hotspots and industrial clusters is quantified by kernel density estimation. Pearson correlation coefficient and inverse variance weighted correlation coefficient are calculated, and a consistency score is output to provide multidimensional independent evidence of the rationality of emission spatial distribution and enhance the credibility of audit conclusions.

[0011] Preferably, the audit verification layer specifically performs the following steps: The multi-list cross-comparison module compares the total posterior emissions with the total emissions of each prior list grid by grid, calculates the difference between the posterior emissions of each grid cell and the emissions of each prior list, as well as the percentage deviation of the total amount, generates a spatial deviation distribution map, and outputs the difference analysis results between the top-down inversion results and the bottom-up list, intuitively presenting the spatial offset pattern of posterior emissions relative to the prior list, and providing directional reference for list improvement. The spatial collaborative verification module employs kernel density estimation to perform posterior verification. The spatial field and the density distribution of industrial facilities in the target area are spatially matched and quantified to identify the degree of spatial overlap between emission hotspots and industrial clusters. The first verification consistency score based on the distribution of industrial facilities is output. Based on the spatial distribution of industrial facilities, the spatial matching degree between the inverted emission hotspots and industrial clusters is quantitatively evaluated. The posterior is calculated using the spatial collaborative verification module. Pearson correlation coefficient and inverse variance-weighted correlation coefficient between the spatial field and the concentration field of co-pollutants, wherein the co-pollutants are selected from PM2.5. 2.5 PM 10At least one of NO2, CO, and SO2 is used to output a second verification consistency score based on the correlation coefficient, complete the emission source attribution judgment, and use the pollutant homology characteristics to assist in emission source identification, so as to achieve multi-dimensional cross-validation attribution judgment.

[0012] Preferably, the uncertainty assessment layer specifically performs the following steps: Based on the observation uncertainty assessment module 14 The measurement uncertainty is determined by the standard deviation of parallel samples measured by C-accelerator mass spectrometry. The transmission error is determined based on the atmospheric transmission model error. The sum of the squares of the measurement uncertainty and the transmission error is used as the diagonal element of the observation error covariance matrix R. The transmission error is taken as the observed... The preset ratio of values ​​comprehensively quantifies the random errors in the observation and transmission process, ensuring that the weight configuration of observation constraints is reasonable and reliable. Multi-level scaling factor SF is set through the prior sensitivity test module. B The values ​​were selected, and multiple sets of inversion comparisons were performed with each prior list source as the prior emission field. The spatial correlation sensitivity test module was used to set the observation error covariance matrix R and the prior error covariance matrix B to three spatial correlation lengths of 0, 1 and 2 times the spatial grid resolution, respectively, to form multiple sets of sensitivity experiment matrices to perform inversion experiments. The system evaluated the impact of prior assumptions and spatial correlation settings on the inversion results and identified the optimal configuration range of key parameters. The ensemble inversion statistics module summarizes the posterior emission results of all sensitivity tests, calculates the mean and standard deviation of the posterior emissions for each experiment, constructs a probability distribution function of emissions based on the mean and standard deviation, and outputs emission estimation results with a pre-set confidence interval. The pre-set confidence interval is a 95% confidence interval, which comprehensively reflects the degree of dispersion of emission estimation caused by parameter changes and provides traceable statistical confidence intervals.

[0013] Preferably, the visualization output layer includes a spatial interpolation mapping module, an emission spatial distribution map generation module, an automatic verification report generation module, and an uncertainty visualization module; The spatial interpolation mapping module is configured to perform... The observed concentrations are interpolated using Kriging space to generate a high-resolution spatial distribution map of the concentration, transforming discrete sampling point concentration observations into continuous spatial distribution information, revealing... Regional distribution characteristics of concentration; The emission spatial distribution map generation module is configured to render a graded color map based on the posterior emission field, supporting gridded display and hotspot annotation, intuitively presenting the spatial differences and clustering characteristics of emission fluxes, and highlighting the spatial location of high emission areas. The automatic verification report generation module is configured to automatically generate carbon emission audit reports based on multi-list comparison, spatial collaborative verification, and uncertainty assessment results, according to a preset template. This standardizes the presentation of audit conclusions throughout the entire process and improves the consistency and readability of audit results. The uncertainty visualization module is configured to output the probability distribution of emissions in the form of an error bar chart or a confidence interval rendering chart, clearly conveying the reliability range of the emission estimate and assisting decision-makers in scientifically assessing the confidence level of the results.

[0014] Secondly, the present invention also provides a method based on 14 C's urban carbon emission verification and inventory validation method, based on the aforementioned method... 14 This is achieved through the city's carbon emission verification and inventory validation system, including the following steps: S1: Collect data within the target area. 14 C was observed in the medium and determined by accelerator mass spectrometry. 14 C / 12 Δ is calculated after the C ratio. 14 C value, extracted from each sampling point in combination with baseline data. By observing concentrations, fossil-derived and biogenic CO2 can be effectively separated to obtain high-precision results. Concentration observation data; S2: Receive at least two bottom-up emission inventory data, perform format unification, spatial reprojection and bilinear interpolation processing, unify to the same grid resolution, construct a priori emission flux field, fuse multi-source inventory information to construct a priori emission field, and provide a spatially consistent initial estimate for inversion; S3: Using the Lagrange particle diffusion model, virtual particles are released at each observation point and tracked backward. The residence time of the particles in the near-surface footprint layer is statistically analyzed to generate a source-receptor sensitivity field, construct a Jacobian matrix, and finely characterize the response relationship between emission sources and observation points, thereby improving the accuracy of emission spatial allocation. S4: Based on 14 The observation error covariance matrix is ​​constructed from the measurement error and atmospheric transport model error. By minimizing the cost function composed of observation constraint terms and prior constraint terms, the posterior emission flux field is solved. The observation constraint and prior constraint are weighed in a coordinated manner to obtain a posterior emission estimate that has both accuracy and stability. S5: Compare the total emissions from the post-hoc inventory with the total emissions from each prior inventory grid by grid to generate a spatial deviation distribution map. Then, complete the dual spatial verification of industrial facilities and synergistic pollutants through kernel density estimation and correlation coefficient calculation, providing multi-dimensional independent evidence of the spatial distribution of emissions and enhancing the comprehensiveness and reliability of the audit conclusions. S6: By adjusting the scaling factor and spatial correlation length, multiple sets of sensitivity inversion experiments are performed. All experimental results are summarized to statistically calculate the mean and standard deviation of posterior emissions, and output emission estimation results with confidence intervals. The system quantifies the impact of changes in key assumptions on emission estimation and provides traceable statistical confidence intervals. S7: Based on the post-emission field, render a graded and colored emission distribution map and mark emission hotspots. Integrate the results of multi-inventory comparison, spatial collaborative verification and uncertainty assessment to generate a structured carbon emission verification audit report. The report presents the spatial differences in emissions intuitively with graded coloring and hotspot marking, and outputs the audit conclusions of the entire process in a standardized manner.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. This kind of based on 14 C's urban carbon emission verification and inventory validation system and method, by introducing 14 C isotope tracing technology constructs an observation constraint system independent of the traditional bottom-up statistical path. It can objectively verify the spatial allocation rationality and total accuracy of existing emission inventories, overcome the inherent defect of a single inventory being unable to self-verify due to differences in statistical caliber and proxy variables, and free the audit results from excessive reliance on the prior assumptions of the inventory.

[0016] 2. This is based on 14 C's urban carbon emission verification and inventory validation system and method, through high-density gridded sampling design combined with Lagrange particle diffusion modeling, constructs a refined source-receptor response relationship at the urban scale, effectively capturing the spatial heterogeneity characteristics of emissions, enabling audit results to present the precise location and spatial distribution pattern of emission hotspots, and providing a basis for the formulation of differentiated control measures within the city.

[0017] 3. This type of... 14 The urban carbon emission verification and inventory validation system and method of C, by conducting spatial correlation analysis between the inverted emission field and the spatial distribution of industrial facilities and the concentration fields of multiple synergistic pollutants, forms a multi-level cross-verification mechanism from the spatial attribution of emission sources to the common source of emissions. Multiple pieces of evidence corroborate each other, significantly enhancing the reliability and persuasiveness of audit conclusions.

[0018] 4. This type of... 14 The urban carbon emission verification and inventory validation system and method of C establishes a systematic sensitivity testing system covering multiple factors such as observation error, prior assumptions and spatial correlation. It integrates the probability distribution characteristics of emissions from multiple sets of inversion experiments to give the audit results a traceable confidence interval, overcoming the defect that the results cannot be repeated due to single parameter configuration. Attached Figure Description

[0019] Figure 1This invention is based on 14 A schematic diagram of the overall architecture of City C's carbon emission verification and inventory validation system; Figure 2 This invention is based on 14 Overall flowchart of the city carbon emission verification and inventory validation method for city C; Figure 3 For the present invention 14 Data processing flowchart for the C observation data module; Figure 4 The inversion calculation process of the Bayesian assimilation engine module of this invention. Figure 1 ; Figure 5 This is a schematic diagram of the workflow of the audit verification layer of the present invention; Figure 6 The inversion calculation process of the Bayesian assimilation engine module of this invention. Figure 2 ; Figure 7 This is a schematic diagram of the results of multiple sensitivity inversions in an embodiment of the present invention. Detailed Implementation

[0020] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0021] Example 1, please refer to Figures 1 to 7 The present invention provides a technical solution: a method based on 14 C's urban carbon emission verification and inventory validation system includes a data input layer, an inversion calculation layer, an audit validation layer, an uncertainty assessment layer, and a visualization output layer; The data input layer receives and preprocesses multi-source input data from the target region, providing high-quality input data in a unified format for subsequent inversion calculations, ensuring data consistency and reliability. The data input layer includes... 14 C. Observation data module, emission inventory data interface module, and meteorological and environmental data interface module; in, 14 The C observation data module is configured to receive data from the target area. 14 C observation data, execute Δ 14 C calculation, baseline deduction and Signal extraction and output target area By observing concentrations, interference from biological sources can be effectively eliminated, thus improving the accuracy of identifying CO2 concentrations from fossil sources. The emission inventory data interface module is configured to receive at least two types of bottom-up emission inventory data, perform format unification, spatial reprojection and gridded interpolation, output a priori emission field under a unified grid resolution, and integrate multi-source inventory information to provide a diverse data foundation for the construction of a priori emission field. The meteorological and environmental data interface module is configured to receive atmospheric reanalysis meteorological field data and regional air quality monitoring data, perform data quality checks and time matching, and output meteorological driving fields and collaborative pollutant validation datasets to ensure the driving accuracy of atmospheric transport simulation and provide data support for spatial collaborative validation. In this embodiment, the data input layer specifically performs the following steps: through 14 The C observation data module receives at least one of the following: plant samples, atmospheric samples, or tree-ring samples from the target area. 14 C was observed in the medium and determined by accelerator mass spectrometry. 14 C / 12 Δ is calculated after the C ratio. 14 C value, based on the background Δ 14 The C value and background CO2 concentration are calculated using the following formula: Extract each sampling point Observed concentrations, among which This represents the background CO2 concentration. The atmospheric Δ reflected by the observation medium 14 C value, The background Δ 14 C value, fossil fuel Δ 14 The C-value (taken as -1000‰) effectively separates fossil-derived and biogenic CO2, achieving the target area. Precise concentration quantification is achieved by receiving at least two bottom-up emission inventory data sources through the emission inventory data interface module. Format standardization, spatial reprojection, and bilinear interpolation are performed on each inventory to unify them to a grid with the same spatial resolution. At least two sets of prior emission flux fields are output, along with emission values ​​and spatial coordinate indices for each grid cell. This eliminates format and coordinate system differences between multi-source inventories, constructing a spatially consistent multi-scenario prior emission field. The system also receives atmospheric reanalysis meteorological field data and regional air quality monitoring data through the meteorological and environmental data interface module. Data quality checks and time matching are performed on the meteorological field data to output a meteorological driving field. PM2.5 is extracted from the regional air quality monitoring data. 2.5 PM 10 Concentration fields of at least one co-polluting element among NO2, CO, and SO2 are used to construct a co-polluting validation dataset to ensure the spatiotemporal continuity of the meteorological driving field and provide multi-dimensional cross-validation data to support the auxiliary judgment of emission sources. It should be noted that data is collected within the target area according to a pre-defined spatial distribution plan. 14C observation medium, 14 The observation medium (C) is selected from at least one of plant samples, atmospheric samples, or tree-ring samples. The layout of each sampling point follows the grid design principle of the target area, spatially covering urban functional zones, including residential areas, industrial areas, commercial areas, and ecological protection areas. Sampling points preferentially avoid areas directly disturbed by local point source emissions. The collected plant samples are selected from the aboveground tissues. Leaves from multiple adjacent plants are collected at each sampling point and mixed into one sample. The samples are washed with hydrochloric acid solution to remove surface adhering substances, rinsed with deionized water to neutral pH, dried and ground, and then converted into graphite targets by an automated graphitization system. Accelerator mass spectrometry is used for analysis. 14 C / 12 C ratio, based on the sample 14 C / 12 C ratio and standard substance 14 C / 12 Δ calculation of the deviation of the C ratio 14 C value, obtain Δ at each sampling point 14 After the C value, based on the background Δ of the target region 14 C-value and background CO2 concentration data were extracted from each sampling point. Concentration observation is performed, and parallel sample measurements are conducted at each sampling point during the extraction process. The standard deviation of the parallel samples is calculated to monitor the extraction accuracy. The emission inventory data interface module receives at least two types of bottom-up emission inventory data, including a non-uniform emission inventory based on nighttime light remote sensing inversion and an emission inventory based on sectoral activity level statistics. The module performs format standardization processing on each inventory, converting different data formats to the system's internal standard format. Spatial reprojection is performed to unify the geographic coordinate system of each inventory. Bilinear interpolation is used to resample the spatial resolution of each inventory to the same grid resolution. After processing, at least two sets of prior emission flux fields are output. Each set of prior emission flux fields contains the emission values ​​of each grid cell and the corresponding spatial coordinate index, providing prior emission flux fields for the inversion calculation layer. The meteorological and environmental data interface module receives atmospheric reanalysis meteorological field data and regional air quality monitoring data. It performs data quality checks on the meteorological field data, removing outliers and missing values, and then performs time matching according to the observation periods to output a meteorological driving field. This driving field is used to drive the backward trajectory simulation of the Lagrange particle diffusion model and extract PM2.5 from the regional air quality monitoring data. 2.5 PM 10 The concentration fields of at least one co-polluting element among NO2, CO, and SO2 are collected. The data for each co-polluting element are spatially processed according to sampling points and observation periods to construct a co-polluting element validation dataset. After preprocessing at the data input layer, the output data from each module are correlated and integrated according to a unified spatial grid and time period index. 14 Each sampling point output by the C observation data module The observed concentrations are correlated with the spatial coordinates of the sampling points to form an observed concentration dataset with spatial attributes. The two sets of prior emission flux fields output by the emission inventory data interface module and the meteorological driving field output by the meteorological and environmental data interface module are aligned by grid index and time period to ensure the consistency of matching of each input data in spatial and temporal dimensions. The integrated dataset is encapsulated according to the system's preset data transmission format and transmitted to the inversion calculation layer in sequence. The data format includes grid spatial attributes, time attributes, and numerical fields of each data dimension.

[0022] The inversion calculation layer, connected to the data input layer, is used to perform Bayesian inversion optimization of emission fluxes, achieving optimal estimation of fossil fuel emission fluxes and improving the accuracy of emission estimation. The inversion calculation layer includes: The prior emission field construction module is configured to construct a prior emission flux field and its error covariance matrix B based on emission inventory data, quantify the spatial uncertainty of prior emissions, and provide reasonable initial constraints for Bayesian inversion. The atmospheric transport operator construction module is configured to use the Lagrange particle diffusion model to release virtual particles at each observation point and backtrack them to generate a source-receptor relationship sensitivity field, construct the Jacobian matrix H, establish a quantitative response relationship between the emission source and the observation point, and realize the effective mapping of the observation signal to the emission space. The Bayesian assimilation engine module is configured to solve for the posterior emission flux x by minimizing the cost function, and integrate prior information and observation constraints to obtain the optimal emission flux estimate that takes into account the advantages of both. In this embodiment, the inversion calculation layer specifically performs the following steps: constructing a priori emission flux field based on at least two emission inventory data at a uniform grid resolution using the priori emission field construction module. Calculate the absolute value of the emission difference between the two inventories in each grid cell, and multiply this absolute value by the scaling factor SF. B The prior error covariance matrix B is constructed by squaring the squares, i.e., B = (SF B ×|List 1-List 2|)², objectively reflecting prior uncertainty through differences between multiple lists, avoiding bias caused by subjectively set error distributions. Using an atmospheric transport operator construction module and a Lagrange particle diffusion model, virtual particles are released at preset time intervals at each observation point and tracked backward for a preset number of days. The residence time of virtual particles in the near-surface footprint layer is statistically analyzed to generate a source-receptor sensitivity field. Based on this field, a Jacobian matrix H is constructed. The near-surface footprint layer is the range of 0 to 100 meters above ground level, precisely characterizing the near-surface atmospheric transport process and ensuring accurate and reliable response relationships between emission sources and observation points. A Bayesian assimilation engine module is used to construct an observation error covariance matrix R, a diagonal matrix whose diagonal elements are the sum of squares of measurement uncertainty and transport error. This is achieved by minimizing the cost function. Solving for posterior emission fluxes The total regional emissions are obtained by multiplying the posterior emission flux by the grid area and summing the results, where... For the a priori emission flux field, Based on 14 C observation value estimation concentration, The observation error covariance matrix, The prior error covariance matrix is ​​used to coordinate the trade-off between observation constraints and prior constraints, thereby obtaining a posterior emission flux estimate that combines accuracy and stability. It should be noted that the system receives at least two bottom-up emission inventory data from the emission inventory data interface module. This data has undergone spatial reprojection and bilinear interpolation processing in the data input layer, unifying it to the same spatial resolution grid. The prior emission field construction module compares the emission values ​​of each inventory in each grid cell grid by grid, calculates the absolute value of the emission difference between the two inventories in the same grid cell, and multiplies this absolute value of the difference by a preset scaling factor SF to construct the prior error covariance matrix B. B The product result is squared and assigned to the diagonal position of the corresponding grid cell, while off-diagonal elements are set to zero, forming a diagonal prior error covariance matrix B. The atmospheric transport operator construction module uses the Lagrange particle diffusion model to conduct backward trajectory simulation. Virtual particles are released at hourly intervals at each observation point, with a preset number of virtual particles released each time. All virtual particles trace their atmospheric motion trajectory backward for a preset number of days from the release time. During the integration operation, the model records the three-dimensional position and residence time of each virtual particle within the spatial grid, focusing on the cumulative residence time of virtual particles in the near-surface footprint layer from above the ground to a height of 100 meters. Based on the statistical results of particle residence time in each grid cell at each observation point, a source-receptor sensitivity field is generated. The dimension of this source-receptor sensitivity field is the response coefficient of concentration to emission flux. The sensitivity fields corresponding to all observation points are arranged by grid index to construct a Jacobian matrix H, where the row index corresponds to the observation point and the column index corresponds to the grid cell. The matrix elements characterize the effect of emission flux change in any grid cell on any observation point. Marginal contribution of concentration change; Bayesian assimilation engine module acquires prior emission flux field Jacobian matrix H, based on 14 C observation value estimation After obtaining the concentration vector y and the error covariance matrices B and R, a cost function is constructed. This cost function consists of two terms: the first term is the observation constraint term, which maps the prior emission flux to the observation space through the Jacobian matrix and compares it with the actual observed values, and is weighted by the inverse of the observation error covariance matrix R; the second term is the prior constraint term, which measures the degree of deviation between the posterior emission flux to be solved and the prior emission flux, and is weighted by the inverse of the prior error covariance matrix B. The Bayesian assimilation engine module solves for the posterior emission flux vector x that makes the observation constraint term and the prior constraint term optimal overall by minimizing this cost function. After the solution is completed, the emission flux of each grid cell in the posterior emission flux vector is multiplied by the corresponding grid area and accumulated, and then multiplied by the total duration of the calculation period to obtain the total fossil fuel CO2 emissions of the target area during the calculation period.

[0023] The audit verification layer, connected to the inversion calculation layer, is used to perform independent verification and attribution analysis of emission results, providing objective verification results that are independent of the inventory, and enhancing the credibility of emission data.

[0024] The uncertainty assessment layer, connected to the inversion calculation layer, is used to systematically quantify the uncertainty of the inversion results, clarify the reliability of emission estimates, and avoid biases caused by single parameter assumptions.

[0025] The visualization output layer, connected to the audit verification layer and the uncertainty assessment layer, is used to generate audit result products, intuitively presenting the spatial distribution of emissions and audit conclusions, and supporting carbon governance decision-making needs.

[0026] The uncertainty assessment layer includes: The observation uncertainty assessment module is configured to be based on 14 C. Measurement error and atmospheric transport model error are used to construct the observation error covariance matrix R, which quantifies the random errors in the observation and transport process and ensures the accuracy of the inversion input. The prior sensitivity testing module is configured to perform multiple inversion comparisons by changing the source of the prior list and adjusting the scaling factor of the error covariance matrix B, thereby quantifying the impact of prior hypotheses on posterior results, identifying the sensitivity of prior hypotheses to posterior results, and improving the interpretability of inversion results. The spatial correlation sensitivity test module is configured to set different spatial correlation length parameters for the observation error covariance matrix R and the prior error covariance matrix B, execute multiple sets of inversion experiments, evaluate the impact of the spatial correlation assumption on the inversion stability, clarify the impact of the spatial correlation assumption on the inversion stability, and optimize the selection of spatial smoothness. The aggregate inversion statistics module is configured to summarize the posterior emission results of all sensitivity tests, calculate their mean and standard deviation, output the probability distribution and confidence interval of emissions, provide traceable statistical confidence intervals, and enhance the scientific rigor and credibility of audit reports.

[0027] Example 2, as Figures 1 to 7 As shown, based on Embodiment 1, the present invention provides a technical solution: the audit verification layer includes a multi-list cross-comparison module and a spatial collaborative verification module; Among them, the multi-list cross-comparison module is configured to compare the total posterior emissions with the total of each prior list grid by grid, calculate the total deviation percentage and spatial deviation distribution, output the difference analysis results from top to bottom and bottom to top, reveal the systematic deviation characteristics between the posterior emissions and each prior list, and identify the areas where the lists are overestimated or underestimated. The spatial collaborative verification module is configured to use the posterior... Spatial correlation analysis is performed on the spatial field and the density distribution of industrial facilities in the target area and the concentration field of at least one synergistic pollutant. The spatial consistency between emission hotspots and industrial clusters is quantified by kernel density estimation. Pearson correlation coefficient and inverse variance weighted correlation coefficient are calculated, and a consistency score is output to provide multidimensional independent evidence of the rationality of emission spatial distribution and enhance the credibility of audit conclusions. In this embodiment, the audit verification layer specifically performs the following steps: The multi-inventory cross-comparison module compares the posterior total emissions with the total emissions of each prior inventory grid by grid, calculating the difference between the posterior emissions and the emissions of each prior inventory in each grid cell, as well as the percentage deviation of the total emissions. This generates a spatial deviation distribution map and outputs the difference analysis results between the top-down inversion results and the bottom-up inventory, visually presenting the spatial offset pattern of posterior emissions relative to the prior inventory. This provides a directional reference for inventory improvement. The spatial collaborative verification module uses kernel density estimation to verify the posterior emissions. The spatial fit between the spatial field and the density distribution of industrial facilities in the target area is quantified to identify the degree of spatial overlap between emission hotspots and industrial clusters. A first validation consistency score based on the distribution of industrial facilities is output. Using the spatial distribution of industrial facilities as a benchmark, the spatial matching degree between the inverted emission hotspots and industrial clusters is quantitatively evaluated. The posterior result is calculated through a spatial collaborative validation module. Pearson correlation coefficient and inverse variance-weighted correlation coefficient between the spatial field and the concentration field of co-pollutants, where co-pollutants are selected from PM2.5. 2.5 PM 10 At least one of NO2, CO, and SO2 is used to output a second verification consistency score based on the correlation coefficient, complete the emission source attribution judgment, use the pollutant homology characteristics to assist in emission source identification, and realize multi-dimensional cross-validation attribution judgment; It should be noted that, during the audit verification process, the multi-list cross-comparison module multiplies the emission flux of each grid cell in the posterior emission flux vector obtained by the Bayesian assimilation engine module by the corresponding grid area and sums them up to obtain the total posterior emission amount of the target area. Simultaneously, it retrieves the emission data of each prior list that has undergone spatial reprojection and bilinear interpolation processing from the data input layer. It extracts the prior emission amount of each prior list in each grid cell according to the same grid index, calculates the difference between the posterior emission amount and the emission amount of each prior list for each grid cell, and divides this difference by the corresponding prior list emission amount to obtain the relative deviation percentage of each grid cell. Based on the absolute difference and relative deviation percentage of each grid cell, it renders the generated data. A spatial deviation distribution map is generated, which distinguishes between areas of overestimation and underestimation of posterior emissions relative to the prior inventory by positive and negative differences. The proportion of deviation grids to the total number of grids is then statistically analyzed, outputting the systematic difference analysis results in spatial distribution between the top-down inversion results and the bottom-up inventory. The spatial collaborative verification module employs kernel density estimation to perform the first level of spatial conformity quantification verification for industrial facility distribution. First, a spatial location database of registered industrial polluting enterprises in the target area is acquired. A spatial density surface is generated with each industrial facility location as the center and a preset bandwidth as the radius. This density surface exhibits high density values ​​in areas with concentrated industrial facilities and low density values ​​in areas with sparse facilities. Simultaneously, the posterior... Emission hotspots in the spatial field are extracted as sets of grid cells whose emissions exceed a preset threshold. Spatial overlap analysis is performed between this set and the high-value regions of the industrial density surface. The ratio of emission hotspot grid cells falling into the high-value regions of industrial density is calculated as the first validation consistency score. This score reflects the posterior consistency. The degree of agreement between the spatial distribution of emissions and the spatial clustering of industrial activities; after completing the kernel density estimation verification, the spatial collaborative verification module further performs a second spatial collaborative verification, which will be carried out by the posterior. In the collaborative pollutant validation dataset constructed by the spatial field and data input layer, the concentration fields of each pollutant are paired one by one according to the grid index, and the posterior is calculated separately. With PM 2.5 PM 10 The Pearson correlation coefficient between at least one co-polluting pollutant among NO2, CO, and SO2 is used to measure the correlation coefficient between them. The linear correlation strength of each co-polluting pollutant at the grid scale is calculated, and the weighted correlation coefficient is calculated using the inverse variance weighting method. The reciprocal of the observation variance of each grid is used as a weighting factor to assign the contribution of the corresponding grid to the correlation coefficient calculation, thereby reducing the impact of grids with greater observation uncertainty on the overall correlation coefficient. Based on the calculated correlation coefficient and its significance test results, a second validation consistency score is output. The values ​​of the first validation consistency score and the second validation consistency score are combined to complete the attribution judgment of the emission source.

[0028] In this embodiment, the uncertainty assessment layer specifically performs the following steps: Based on the observation uncertainty assessment module... 14 The measurement uncertainty is determined by the standard deviation of parallel samples measured by C-accelerator mass spectrometry. The transmission error is determined based on the atmospheric transmission model error. The sum of the squares of the measurement uncertainty and the transmission error is used as the diagonal element of the observation error covariance matrix R. The transmission error is taken as the observed... The preset ratio of values ​​comprehensively quantifies random errors in the observation and transmission process, ensuring that the weight configuration of observation constraints is reasonable and reliable. A multi-level scaling factor SF is set through the prior sensitivity test module. B The system takes values ​​from each prior inventory source as the prior emission field and performs multiple sets of inversion comparisons. Through the spatial correlation sensitivity test module, the observation error covariance matrix R and the prior error covariance matrix B are set to three spatial correlation lengths: 0, 1, and 2 times the spatial grid resolution (or set according to specific circumstances). Multiple sets of sensitivity experiment matrices are formed to perform inversion experiments. For example, if the spatial resolution of the inversion is 5km grid, it is recommended to set the spatial correlation length to 0, 5km, and 10km. The system evaluates the impact of prior assumptions and spatial correlation settings on the inversion results, identifies the optimal configuration range of key parameters, and summarizes the posterior emission results of all sensitivity tests through the ensemble inversion statistics module. The posterior emission mean and standard deviation of each experiment are statistically analyzed. Based on the mean and standard deviation, a probability distribution function of emissions is constructed, and emission estimation results with a preset confidence interval are output. The preset confidence interval is a 95% confidence interval, which comprehensively reflects the dispersion of emission estimation caused by parameter changes and provides a traceable statistical confidence interval. It should be noted that the observation uncertainty assessment module is activated after the inversion calculation is completed, based on... 14 The standard deviation of the measured values ​​of parallel samples at each sampling point during C-accelerator mass spectrometry measurements is calculated. This standard deviation serves as a quantitative indicator of measurement uncertainty, reflecting the level of random error throughout the entire process of sample pretreatment, graphitization conversion, and AMS instrument analysis. Simultaneously, the atmospheric transport model error is determined based on the backward trajectory simulation deviation of the Lagrange particle diffusion model. This error is related to the meteorological field resolution, turbulence parameterization scheme, and particle integration time; its specific value depends on the observations. The concentration values ​​are correlated with a preset ratio. The squares of measurement uncertainty and transmission error are summed, and the result is used as the diagonal element of the observation error covariance matrix R. The R matrix is ​​maintained in diagonal form, and the off-diagonal elements are set to zero, indicating that the errors between each observation point are independent. The prior sensitivity test module targets the scaling factor SF in the prior error covariance matrix B. B Perform multi-level parameter configuration, SF B The values ​​are set to three levels: low, medium, and high, to cover different scenarios from tight to loose prior constraints. In each SF... BAt each level, the prior list sources provided by the data input layer are sequentially used as prior emission fields, and the complete inversion process is executed independently. The impact of different prior sources on the offset of the posterior emission results is compared. The spatial correlation sensitivity test module runs independently of the prior sensitivity test. Three spatial correlation lengths (low, medium, and high) are set for the observation error covariance matrix R and the prior error covariance matrix B, respectively (for example, low, medium, and high correspond to 0km, 5km, and 10km, respectively; the specific values ​​can be adjusted according to the actual grid resolution). The multi-level SF B The values ​​are combined with three spatial correlation lengths to form a complete sensitivity experiment matrix. All experiments are performed under the same set of observation data and transmission model conditions. The ensemble inversion statistics module is started after all sensitivity experiments are completed. It reads the posterior emission flux vector of each inversion experiment, summarizes the emission results of all experiments by grid cell, calculates the mean and standard deviation of the posterior emission in each grid cell, and multiplies the mean emission flux of each grid by the total duration of the corresponding grid surface accumulation and accounting period to obtain the mean of the total posterior emission in the target area. Based on the mean and standard deviation, the module constructs a probability distribution function of the emission for each grid cell and the total emission in the area. This distribution function reflects the dispersion of emission estimation caused by changes in the inversion parameter assumptions. The module determines the lower and upper limits of the 95% confidence interval by adding or subtracting 1.96 times the standard deviation from the mean. The confidence interval and the mean of the total emission are output synchronously as the final emission estimation result for use by the audit verification layer and the visualization output layer.

[0029] The visualization output layer includes a spatial interpolation mapping module, an emission spatial distribution map generation module, an automatic verification report generation module, and an uncertainty visualization module; Among them, the spatial interpolation mapping module is configured to... The observed concentrations are interpolated using Kriging space to generate a high-resolution spatial distribution map of the concentration, transforming discrete sampling point concentration observations into continuous spatial distribution information, revealing... Regional distribution characteristics of concentration; The emission spatial distribution map generation module is configured to render a hierarchical color map based on the posterior emission field, supporting gridded display and hotspot annotation, intuitively presenting the spatial differences and clustering characteristics of emission fluxes, and highlighting the spatial location of high emission areas; The automatic verification report generation module is configured to automatically generate carbon emission audit reports based on multi-list comparison, spatial collaborative verification, and uncertainty assessment results, according to a preset template. It presents the conclusions of the entire audit process in a standardized manner, improving the consistency and readability of audit results. The uncertainty visualization module is configured to output the probability distribution of emissions as an error bar chart or a confidence interval rendering chart, clearly conveying the reliability range of emission estimates and assisting decision-makers in scientifically assessing the confidence level of the results. In this embodiment, the visualization output layer specifically performs the following steps: The spatial interpolation mapping module is used to process each sampling point... The observed concentration is interpolated using Kriging space to generate a high-resolution spatial distribution map of the concentration covering the target area based on spatial autocorrelation features. It outputs the observed concentration contour map and gridded concentration distribution data, and displays the concentration gradient change by overlaying contour lines and pseudo-color layers to enhance the visualization and interpretation of spatial distribution. The emission spatial distribution map generation module renders a graded color map based on the posterior emission flux field according to the preset color level. In the graded color map, grid cells with emissions exceeding the preset threshold are marked as emission hotspots. It outputs an interactive gridded emission distribution map that supports spatial overlay analysis and zooming operations, enabling users to conduct detailed analysis of local areas and explore the spatial relationship of multiple layers in an interactive manner. The automatic verification report generation module integrates the spatial deviation distribution of cross-comparison of multiple lists, the consistency score of spatial co-validation, and the confidence interval of ensemble inversion statistics. It automatically generates a structured carbon emission verification audit report according to the preset template. The uncertainty visualization module renders the probability distribution of emissions as an error bar chart or confidence interval strip chart, and displays the dispersion of emission estimates in a graphical way to help users intuitively grasp the fluctuation range and statistical significance of emissions. It should be noted that the spatial interpolation mapping module performs sampling at each sampling point. The concentration was observed using Kriging spatial interpolation to calculate the concentration at each sampling point. The spatial variability function among concentration observations is obtained by fitting a theoretical variability function model to acquire spatial autocorrelation characteristic parameters. Based on the spatial autocorrelation characteristics among sampling points, ordinary kriging is used to calculate the concentration estimate and its variance for each grid node within the target area. During the gridded estimation process, neighboring sampling points are selected for interpolation calculations for each grid node to be estimated. The weighting coefficients of each participating sampling point are determined with the goal of minimizing the variance. The estimated concentration values ​​at each grid node are rendered by the system to generate a high-resolution spatial distribution map of concentration covering the entire target area. This spatial distribution map is output in the form of contour lines overlaid with a pseudo-color layer. The contour line intervals are automatically set according to the distribution characteristics of the concentration data. Gridded concentration distribution data is simultaneously... The data is stored in a standard raster format and output to downstream modules. The emission spatial distribution map generation module performs spatial mapping operations based on the posterior emission flux field. After obtaining the posterior emission flux vector solved by the Bayesian assimilation engine module, it maps the emission flux values ​​of each grid cell to a preset continuous color sequence according to their magnitude. Low-emission grids are rendered with cool colors, and high-emission grids are rendered with warm colors, forming a graded colored emission distribution map. Further statistical distribution analysis is performed on the posterior emission flux field. The upper quartile of the emission flux statistical distribution is used as a preset threshold for emission hotspot determination. Grid cells with emissions exceeding this threshold are identified as emission hotspots and marked with highlighted marks or overlaid special symbols in the graded colored map. After completing the mapping, the data is output. An interactive gridded emission distribution map with layer management capabilities is generated. This interactive map allows operators to perform spatial overlay analysis of emission layers and supports zooming to view emission details in any local area. The emission values ​​and spatial coordinate indices of each grid unit are presented in the interactive interface as floating prompts or attribute queries. The automatic verification report generation module completes the integration and encapsulation of all audit results in the visualization output stage. It obtains spatial deviation distribution data of multi-list cross-comparison and consistency scores of spatial collaborative verification from the audit verification layer, and obtains confidence intervals of ensemble inversion statistics and probability distribution functions of emissions from the uncertainty assessment layer. It executes the report generation process according to the system's preset structured audit report template. The audit report template package... The report includes sections on audit overview, data sources and preprocessing instructions, inversion parameter configuration records, comparative analysis of posterior total emissions and prior inventory, spatial deviation distribution statistics, dual spatial verification scoring of industrial facilities and pollutants, uncertainty quantification results, and 95% confidence intervals. The content of each section is automatically filled in by the corresponding calculated data and graphical results. After the report is automatically formatted, it is output in a standardized format. The uncertainty visualization module simultaneously renders the probability distribution function of emissions as an error bar chart. The center of the error bar chart marks the emission mean, and the upper and lower limits correspond to the upper and lower limits of the confidence interval, respectively. Alternatively, it can be rendered as a confidence interval band chart, with the band width representing the dispersion of emission estimates. The final audit report and visualization charts are output together and delivered to the user.

[0030] Example 3, as Figures 2 to 7 As shown, based on Embodiments 1 and 2, the present invention also provides a method based on... 14 C's urban carbon emission verification and inventory validation method, based on the above-mentioned method... 14 This is achieved through the city's carbon emission verification and inventory validation system, including the following steps: S1: Collect data within the target area. 14 C was observed in the medium and determined by accelerator mass spectrometry. 14 C / 12 Δ is calculated after the C ratio. 14 C value, extracted from each sampling point in combination with baseline data. By observing concentrations, fossil-derived and biogenic CO2 can be effectively separated to obtain high-precision results. Concentration observation data; S2: Receive at least two bottom-up emission inventory data, perform format unification, spatial reprojection and bilinear interpolation processing, unify to the same grid resolution, construct a priori emission flux field, fuse multi-source inventory information to construct a priori emission field, and provide a spatially consistent initial estimate for inversion; S3: Using the Lagrange particle diffusion model, virtual particles are released at each observation point and tracked backward. The residence time of the particles in the near-surface footprint layer is statistically analyzed to generate a source-receptor sensitivity field, construct a Jacobian matrix, and finely characterize the response relationship between emission sources and observation points, thereby improving the accuracy of emission spatial allocation. S4: Based on 14 The observation error covariance matrix is ​​constructed from the measurement error and atmospheric transport model error. By minimizing the cost function composed of observation constraint terms and prior constraint terms, the posterior emission flux field is solved. The observation constraint and prior constraint are weighed in a coordinated manner to obtain a posterior emission estimate that has both accuracy and stability. S5: Compare the total emissions from the post-hoc inventory with the total emissions from each prior inventory grid by grid to generate a spatial deviation distribution map. Then, complete the dual spatial verification of industrial facilities and synergistic pollutants through kernel density estimation and correlation coefficient calculation, providing multi-dimensional independent evidence of the spatial distribution of emissions and enhancing the comprehensiveness and reliability of the audit conclusions. S6: By adjusting the scaling factor and spatial correlation length, multiple sets of sensitivity inversion experiments are performed. All experimental results are summarized to statistically calculate the mean and standard deviation of posterior emissions, and output emission estimation results with confidence intervals. The system quantifies the impact of changes in key assumptions on emission estimation and provides traceable statistical confidence intervals. S7: Based on the post-emission field, render a graded and colored emission distribution map and mark emission hotspots. Integrate the results of multi-inventory comparison, spatial collaborative verification and uncertainty assessment to generate a structured carbon emission verification audit report. The report presents the spatial differences in emissions intuitively with graded coloring and hotspot marking, and outputs the audit conclusions of the entire process in a standardized manner.

[0031] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. The scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method based on 14 C's urban carbon emission verification and inventory validation system is characterized by... It includes a data input layer, an inversion calculation layer, an audit and verification layer, an uncertainty assessment layer, and a visualization output layer; The data input layer is used to receive and preprocess multi-source input data from the target region. The data input layer includes... 14 C. Observation data module, emission inventory data interface module, and meteorological and environmental data interface module; Among them, the 14 The C observation data module is configured to receive data from the target area. 14 C observation data, execute Δ 14 C calculation, baseline deduction and Signal extraction and output target area The specific steps for observing concentration are as follows: through 14 The C observation data module receives at least one of the following: plant samples and atmospheric samples from the target area. 14 C was observed in the medium and determined by accelerator mass spectrometry. 14 C / 12 Δ is calculated after the C ratio. 14 C value, based on the background Δ 14 The C value and background CO2 concentration are calculated using the following formula: Extract each sampling point Observed concentrations, among which This represents the background CO2 concentration. The atmospheric Δ reflected by the observation medium 14 C value, The base Δ 14 C value, fossil fuel Δ 14 C value; The emission inventory data interface module is configured to receive at least two types of bottom-up emission inventory data, perform format unification, spatial reprojection and gridded interpolation, and output a priori emission field with a unified grid resolution. The meteorological and environmental data interface module is configured to receive atmospheric reanalysis meteorological field data and regional air quality monitoring data, perform data quality checks and time matching, and output a meteorological driving field and a co-pollutant validation dataset, wherein the co-pollutants are selected from PM2.

5. 2.5 PM 10 At least one of NO2, CO, and SO2; The inversion calculation layer, connected to the data input layer, is used to perform Bayesian inversion optimization of emission fluxes. The inversion calculation layer includes: The prior emission field construction module is configured to construct a prior emission flux field and its error covariance matrix based on emission inventory data. The atmospheric transport operator construction module is configured to use the Lagrange particle diffusion model to release virtual particles at each observation point and backtrack them to generate a source-acceptor relationship sensitivity field and construct a Jacobian matrix. The Bayesian assimilation engine module is configured to solve for the posterior emission flux by minimizing a cost function. The specific steps are as follows: The Bayesian assimilation engine module constructs a diagonal matrix of observation error covariance, where the diagonal elements are the sum of the squares of measurement uncertainty and transmission error. Then, it minimizes the cost function... Solve for the posterior emission flux, where, For post-emission flux, For the a priori emission flux field, For based on 14 C observation value estimation concentration, The observation error covariance matrix, Let be the prior error covariance matrix. The Jacobian matrix is ​​used, and the total regional emissions are obtained by multiplying the posterior emission flux by the grid area and summing the computation time period length. The audit verification layer, connected to the inversion calculation layer, is used to perform independent verification and attribution analysis of emission results. The audit verification layer includes a multi-list cross-comparison module and a spatial collaborative verification module. The multi-list cross-comparison module is configured to compare the total posterior emissions with the total emissions of each prior list grid by grid, calculate the total deviation percentage and spatial deviation distribution, and output the difference analysis results from top to bottom and bottom to top. The spatial collaborative verification module is configured to perform a post-verification... Spatial correlation analysis was performed on the spatial field, the density distribution of industrial facilities in the target area, and the concentration field of at least one synergistic pollutant. The spatial fit between emission hotspots and industrial clusters was quantified by kernel density estimation. Pearson correlation coefficient and inverse variance weighted correlation coefficient were calculated, and a consistency score was output. The uncertainty assessment layer is connected to the inversion calculation layer and is used to quantify the uncertainty of the inversion results in a systematic manner. The visualization output layer is connected to the audit verification layer and the uncertainty assessment layer to generate audit result products; The uncertainty assessment layer includes: The observation uncertainty assessment module is configured to be based on 14 Construct the observation error covariance matrix using measurement errors and atmospheric transport model errors; The prior sensitivity test module is configured to perform multiple inversion comparisons by changing the source of the prior list and adjusting the scaling factor of the error covariance matrix, thereby quantifying the impact of prior assumptions on posterior results. The spatial correlation sensitivity test module is configured to set different spatial correlation length parameters for the observation error covariance matrix and the prior error covariance matrix, execute multiple sets of inversion experiments, and evaluate the impact of the spatial correlation assumption on the inversion stability. The ensemble inversion statistics module is configured to summarize the posterior emission results of all sensitivity tests, calculate their mean and standard deviation, and output the probability distribution and confidence interval of emissions.

2. A method based on claim 1 14 C's city carbon emission verification and inventory validation system is characterized by: The data input layer further performs the following steps: The system receives at least two bottom-up emission inventory data through the emission inventory data interface module, performs format unification, spatial reprojection and bilinear interpolation on each inventory, unifies each inventory to the same spatial resolution grid, outputs at least two sets of prior emission flux fields, and provides emission values ​​and spatial coordinate indexes for each grid cell. The system receives atmospheric reanalysis meteorological field data and regional air quality monitoring data through the meteorological and environmental data interface module. It performs data quality checks and time matching on the meteorological field data to output the meteorological driving field, extracts the co-pollutant concentration field from the regional air quality monitoring data, and constructs a co-pollutant validation dataset.

3. A method based on claim 1 14 C's city carbon emission verification and inventory validation system is characterized by: The inversion calculation layer further performs the following steps: The prior emission flux field is constructed using the prior emission field construction module based on at least two emission inventory data at a uniform grid resolution. The absolute value of the emission difference between the two inventories in each grid cell is calculated. The absolute value of the difference is multiplied by a scaling factor and then squared to construct the prior error covariance matrix. Using the Lagrange particle diffusion model, the atmospheric transport operator construction module releases virtual particles at preset time intervals at each observation point and tracks them backward for a preset number of days. The residence time of the virtual particles in the near-surface footprint layer is counted to generate a source-receptor sensitivity field. Based on the source-receptor sensitivity field, a Jacobian matrix is ​​constructed. The near-surface footprint layer is the height range of 0 to 100 meters above the ground.

4. A method based on claim 1 14 C's city carbon emission verification and inventory validation system is characterized by: The audit verification layer specifically performs the following steps: The multi-list cross-comparison module compares the total posterior emissions with the total emissions of each prior list grid by grid, calculates the difference between the posterior emissions of each grid cell and the emissions of each prior list, as well as the percentage of total deviation, generates a spatial deviation distribution map, and outputs the difference analysis results between the top-down inversion results and the bottom-up list. The spatial collaborative verification module employs kernel density estimation to perform posterior verification. The spatial fit between the spatial field and the density distribution of industrial facilities in the target area is quantified to identify the degree of spatial overlap between emission hotspots and industrial clusters, and outputs a first verification consistency score based on the distribution of industrial facilities. The posterior is calculated using the spatial collaborative verification module. The Pearson correlation coefficient and inverse variance-weighted correlation coefficient between the spatial field and the co-pollutant concentration field are used to output a second validation consistency score based on the correlation coefficient, thus completing the emission source attribution judgment.

5. A method based on claim 1 14 C's city carbon emission verification and inventory validation system is characterized by: The uncertainty assessment layer specifically performs the following steps: Based on the observation uncertainty assessment module 14 The measurement uncertainty is determined by the standard deviation of parallel samples measured by C-accelerator mass spectrometry. The transmission error is determined based on the atmospheric transmission model error. The sum of the squares of the measurement uncertainty and the transmission error is used as the diagonal element of the observation error covariance matrix. The transmission error is taken as the observed... The preset ratio of the value; By setting multi-level scaling factor values ​​through the prior sensitivity test module, multiple sets of inversion comparisons were performed with each prior inventory source as a prior emission field. By setting the observation error covariance matrix and the prior error covariance matrix to three spatial correlation lengths of 0, 1, and 2 times the spatial grid resolution, respectively, through the spatial correlation sensitivity test module, multiple sets of sensitivity experiment matrices were constructed to perform inversion experiments. The ensemble inversion statistics module summarizes the posterior emission results of all sensitivity tests, calculates the mean and standard deviation of the posterior emissions for each experiment, constructs a probability distribution function of emissions based on the mean and standard deviation, and outputs emission estimation results with a pre-set confidence interval, which is a 95% confidence interval.

6. A method based on claim 1 14 C's city carbon emission verification and inventory validation system is characterized by: The visualization output layer includes a spatial interpolation mapping module, an emission spatial distribution map generation module, an automatic verification report generation module, and an uncertainty visualization module. The spatial interpolation mapping module is configured to perform... The observed concentrations are interpolated using Kriging space to generate a high-resolution spatial distribution map of the concentrations. The emission spatial distribution map generation module is configured to render a graded color map based on the posterior emission field, and supports gridded display and hotspot annotation. The automatic verification report generation module is configured to automatically generate a carbon emission audit report based on a preset template, using the results of multi-list comparison, spatial collaborative verification, and uncertainty assessment. The uncertainty visualization module is configured to output the probability distribution of emissions in the form of an error bar chart or a confidence interval rendering chart.

7. A method based on 14 C's urban carbon emission verification and inventory validation method, based on any one of claims 1-6. 14 This is achieved through the city carbon emission verification and inventory validation system of C, characterized by... Includes the following steps: S1: Collect data within the target area. 14 C was observed in the medium and determined by accelerator mass spectrometry. 14 C / 12 Δ is calculated after the C ratio. 14 C value, extracted from each sampling point in combination with baseline data. Observed concentration; S2: Receive at least two types of bottom-up emission inventory data, perform format unification, spatial reprojection and bilinear interpolation processing to unify to the same grid resolution, and construct a priori emission flux field; S3: Using the Lagrange particle diffusion model, virtual particles are released at each observation point and tracked backward. The residence time of particles in the near-surface footprint layer is statistically analyzed to generate a source-receptor sensitivity field and construct the Jacobian matrix. S4: Based on 14 The observation error covariance matrix is ​​constructed from the measurement error and atmospheric transport model error. The posterior emission flux is solved by minimizing the cost function composed of observation constraint terms and prior constraint terms. S5: Compare the total post-hoc emissions with the totals of each prior inventory grid by grid to generate a spatial deviation distribution map, and complete the dual spatial verification of industrial facilities and synergistic pollutants through kernel density estimation and correlation coefficient calculation. S6: Perform multiple sets of sensitivity inversion experiments by adjusting the scaling factor and spatial correlation length, summarize all experimental results, statistically calculate the mean and standard deviation of posterior emissions, and output emission estimation results with confidence intervals; S7: Based on the post-emission field, render a graded and colored emission distribution map and mark emission hotspots. Integrate the results of multi-inventory comparison, spatial collaborative verification and uncertainty assessment to generate a structured carbon emission verification and audit report.

Images