Observation and simulation integrated urban CO2 emission plume monitoring method and system

Abstract

The invention belongs to the technical field of greenhouse gas emission monitoring, and discloses an observation and simulation integrated city CO2 emission plume monitoring method and system. The method comprises the following steps: collecting and preprocessing multi-source data; building an atmospheric CO2 column concentration (XCO2) prediction model based on machine learning and multi-source satellite observation; performing artificial CO2 concentration long-term simulation by using an atmospheric transmission model; artificial CO2 concentration data per hour is simulated; and verifying an artificial CO2 emission estimation model based on optimal estimation and a result. The system is high in automation degree, can automatically recognize the emission plume on the basis of multi-source data fusion and model simulation, reduces the demand of manual intervention, improves the monitoring efficiency, and is suitable for monitoring and management of large-scale urban emission sources.

Description

Technical Field

[0001] The present invention belongs to the technical field of greenhouse gas emission monitoring, and in particular relates to an urban CO2 emission plume monitoring method and system integrating observation and simulation. Background Art

[0002] Inversion of anthropogenic CO2 emissions based on atmospheric CO2 concentration observation data has become an effective means to verify and correct greenhouse gas emission inventories. At present, the carbon satellites in orbit include GOSAT / GOSAT-2, OCO-2 / 3 and TanSat. These satellites provide global-scale atmospheric CO2 column concentration (XCO2) inversion data, showing great application potential in monitoring CO2 emissions at urban and point source scales, with the advantages of objective, large-scale and repeated observations. However, due to the influence of satellite observation mode and cloud cover, the number of effective observations over urban areas is often limited. CO2 in the atmosphere is a long-lived gas, and accurately extracting the XCO2 enhancement signal caused by anthropogenic emissions from high background levels is a key step in quantifying CO2 emissions. Affected by many factors such as atmospheric transmission, ecological flux and sensor sensitivity, it is difficult to achieve continuous monitoring and dynamic tracking of urban carbon emissions by relying solely on atmospheric CO2 concentration data observed by carbon satellites.

[0003] Prior art 1: Method for distinguishing point source emission plumes based on carbon satellite data, disclosed:

[0004] The emission plumes are identified using atmospheric CO2 column concentration (XCO2) data obtained using satellite observation technology, mainly using statistical methods such as threshold segmentation method, Gaussian plume simulation method and cross-sectional flux method.

[0005] Specifically, the threshold segmentation method relies on the geographical location and spatial range of the emission source, and uses the downwind area that is less affected by anthropogenic emissions as prior information to accurately divide the anthropogenic emission area and the background area. This method obtains the local XCO2 enhancement value by subtracting the background concentration value from the satellite observation value of the anthropogenic emission area. On this basis, according to the continuity of the spatial distribution of the observation points, the emission plume is edge detected and shape constrained to ensure the accuracy of the plume boundary identification.

[0006] The Gaussian plume model and cross-sectional flux method are the mainstream methods for inverting emission source rates from atmospheric CO2 plume observations. Under the premise of stable wind field conditions, the CO2 concentration distribution near independent emission point sources is taken as a function of wind speed and wind direction. The Gaussian plume assumes that the atmospheric CO2 concentration distribution observed by satellite conforms to the Gaussian shape and fits the satellite observations of the emission source. The cross-sectional flux method derives the source rate as the product of the local wind field and the concentration integrated over the plume cross section.

[0007] Disadvantages of the first prior art: These methods have very strict conditions on the use of research objects and satellite observation data and have certain limitations.

[0008] (1) The location and spatial range of emission sources are predetermined, which makes these methods more suitable for the study of independent large cities or clear emission point sources. When it comes to the study of anthropogenic emissions in multiple cities or complex areas, threshold selection and emission source determination often rely too much on expert knowledge and manual processing, resulting in a lack of objectivity and automation in the identification process, which limits its large-scale promotion and application.

[0009] (2) The core premise of these methods is that satellite observation data must have a spatially continuous pixel distribution. Currently in-orbit carbon satellites, such as OCO-2 / OCO-3 satellites, usually observe in strip or snapshot mode. Due to factors such as clouds, aerosols, and meteorological conditions, the XCO2 data inverted from satellite observations have a large number of gaps and uneven distribution. These data gaps greatly affect the accurate identification of CO2 emission plumes and the accurate inversion of emissions, and increase the uncertainty of the model.

[0010] Prior art 2: Quantification method of anthropogenic carbon dioxide emissions by pollution-carbon collaborative observation, disclosed:

[0011] Anthropogenic CO2 emissions share the same origin as conventional pollutant emissions. NO2 and CO are two major atmospheric pollutants and are also good CO2 tracers. The method of quantifying anthropogenic carbon dioxide emissions through coordinated pollution and carbon observation uses satellite remote sensing technology to jointly observe a variety of pollutants and greenhouse gases in the atmosphere, analyze the correlation and spatiotemporal distribution characteristics of pollutants and CO2 emissions from multiple dimensions and time scales, accurately identify emission sources and indirectly infer anthropogenic CO2 emissions. This method combines CO2 and its accompanying pollutants, such as carbon monoxide (CO), nitrogen oxides (NO x ) and other observation data, and utilizing the spatiotemporal synergy between pollutants and CO2 emissions, the accuracy of emission source identification and emission estimation is optimized, making up for the limitations of single gas observation methods and the low sensitivity of CO2 to emission changes.

[0012] Disadvantages of the second existing technology: Satellite observations of atmospheric CO2 and pollutant data still have problems with insufficient spatial coverage and data discontinuity, especially in cloudy areas or complex terrain, which will affect the spatiotemporal accuracy of collaborative analysis and increase the uncertainty of emission inversion. Although there is a certain correlation between pollutants and CO2 emissions in time and space, this relationship is not always linear or simple. The emission ratio, fuel composition, control technology, etc. of different emission sources will affect the degree of coordination between pollutants and CO2. Therefore, when deriving CO2 emissions based on pollutants, model assumptions may lead to deviations. In areas where multiple sources of emissions overlap, such as large cities or industrial agglomerations, the pollutant characteristics of different emission sources may overlap with each other, making it difficult to accurately identify pollution sources. In addition, the emission intensity and temporal changes of different pollution sources may be similar, further increasing the complexity of emission source classification.

[0013] Prior art three: Reconstruction technology of spatiotemporal continuous data set of atmospheric CO2 using multi-source data and machine learning, disclosed: By constructing a prediction model with atmospheric CO2 column concentration (XCO2) as the target variable, various parameters related to natural cycle processes and anthropogenic emissions are used as prediction variables, and the complex nonlinear relationship of these variables is trained using machine learning algorithms, thereby filling the blank areas of satellite observation of XCO2. Through this technical solution, high spatial resolution, spatiotemporal continuous mapping analysis of atmospheric CO2 concentration can be generated to solve the problem of discontinuous satellite observation data.

[0014] Disadvantages of the third existing technology: Some SCIs use machine learning algorithms to reconstruct atmospheric CO2 datasets, and most studies do not use pollutants from the same source as CO2 emissions as predictive variables. Although some studies have introduced NO2 as a predictive variable, its contribution to the prediction of XCO2 concentration lacks in-depth discussion, and the impact mechanism of pollutants on XCO2 has not been fully analyzed. For example: Guo,K.;Lei,L.;Sheng,M.;Ji,Z.;Song,H.Refining Spatial and Temporal Sens.2024,16,2456.https: / / doi.org / 10.3390 / rs16132456; and He, Z., Fan, G., Li, Journal of Applied Earth Observation andGeoinformation 129,103859.

[0015] https: / / doi.org / https: / / doi.org / 10.1016 / j.jag.2024.103859; The above two papers used NO2 but not CO. In addition, whether the reconstructed XCO2 dataset can accurately reproduce the anthropogenic XCO2 enhancement features identified in satellite observations has not been fully verified, which affects the application effect of using reconstructed data for emission quantification research. Summary of the invention

[0016] In order to overcome the problems existing in the related art, the disclosed embodiments of the present invention provide a method and system for monitoring urban CO2 emission plumes integrating observation and simulation, and specifically relate to a method for monitoring urban CO2 emission plumes integrating observation and simulation. The purpose of the present invention is to fully integrate the atmospheric CO2 and pollutant concentrations, terrestrial ecological remote sensing parameters and meteorological data observed by multi-source satellites, and innovatively develop a CO2 emission plume automatic identification technology that combines machine learning algorithms with atmospheric transmission models, so as to realize long-term dynamic monitoring of urban CO2 emissions.

[0017] The technical solution is as follows: A method for monitoring urban CO2 emission plumes integrating observation and simulation, comprising:

[0018] S1, multi-source data collection and preprocessing;

[0019] S2, building an XCO2 prediction model based on machine learning and multi-source satellite observations;

[0020] S3, using the anthropogenic CO2 concentration simulation results of the atmospheric transport model, simulates the hourly anthropogenic CO2 concentration data;

[0021] S4, verify the anthropogenic CO2 emission estimation model and results based on optimal estimation.

[0022] In step S1, multi-source data collection includes:

[0023] Collect the XCO2 secondary bias correction dataset derived from the observations of the greenhouse gas monitoring satellites OCO-2 and OCO-3;

[0024] Collect multi-source data of atmospheric CO and NO2 pollutant concentration data, vegetation index, and meteorological reanalysis data from the same emission source; process the temporal and spatial resolution into monthly values ​​and 0.1° grid resolution, and perform normalization processing respectively as input data for the XCO2 prediction model;

[0025] Collect ground-based observations of atmospheric CO2 concentrations, anthropogenic emissions inventories, and static geographic data.

[0026] Furthermore, we collected the XCO2 secondary bias correction dataset derived from the observations of the greenhouse gas monitoring satellites OCO-2 and OCO-3, including:

[0027] First, the XCO2 data observed by satellites are quality screened and the observations with a quality flag QF of 0 are retained;

[0028] Then, a curve fitting function is applied to extract the deterministic trends of annual growth and seasonal cycles in the XCO2 time series and decompose the XCO2 time series into the background XCO 2,bg and short-term changes in XCO 2,st, by fitting the XCO2 data in a moving window of every 2° latitude, a spatiotemporal trend model of XCO2 is established, and the expression is:

[0029]

[0030] XCO 2,st =XCO2-XCO 2,bg

[0031] In the formula, a0 and a1 are the intercept and slope of the polynomial function, t is the time unit number in months, i is the number of harmonic functions, ω is the period coefficient, and the calculation formula is 2π / 12, β i and γ i are the variation coefficients of the harmonic functions, δ filter is the interannual variation contained in the residual between the observed and fitted values ​​of XCO2; parameters a0, a1, a2, β i ,γ i All are calculated by least squares fitting;

[0032] Finally, the XCO2 observation data XCO of OCO-2 and OCO-3 were analyzed. 2,st Gridding was performed with a resolution of 0.1°, the monthly average was calculated, and grid points with more than 5 valid points were retained; the observation average was calculated for the same grid in the same month.

[0033] In step S2, an XCO2 prediction model based on machine learning and multi-source satellite observations is constructed, including:

[0034] S201, using XCO 2,st Data is used as the target variable, and the XCO2 prediction model is constructed based on the machine learning algorithm; multi-source data, including CO, NO2, vegetation index NDVI, precipitation PRE, surface solar radiation SSR, surface air pressure SurfP, temperature TEM, east-west wind speed U10 and north-south wind speed V10 at a height of 10 meters, and monthly time labels are used as prediction variables to train the data samples composed of each grid gi and each month ti; the XCO2 prediction model expression is:

[0035]

[0036] Where XCO 2,st[gi,ti] is the XCO corresponding to grid gi and month ti 2,st Value, CO [gi,ti] is the CO concentration value corresponding to grid gi and month ti, NO 2[gi,ti] is the NO2 concentration value corresponding to grid gi and month ti, NDVI [gi,ti is the vegetation index corresponding to grid gi and month ti, PRES [gi,ti]is the precipitation corresponding to grid gi and month ti, SSR [gi,ti] is the surface solar radiation value corresponding to grid gi and month ti, SurfP [gi,ti] is the surface pressure value corresponding to grid gi and month ti, TEM [gi,ti] is the temperature value corresponding to grid gi and month ti, U10 [gi,ti] and V10 [gi,ti They are the east-west wind speed and north-south wind speed at a height of 10 meters corresponding to grid gi and month ti, where ti is the month and gi is the grid;

[0037] S202, the reconstructed XCO2 spatiotemporal continuous dataset was verified and its accuracy evaluated through ten-fold cross validation, feature importance analysis, and ground observation comparison methods.

[0038] In step S3, the anthropogenic CO2 concentration simulation results of the atmospheric transport model are used to simulate the anthropogenic CO2 concentration data per hour, including:

[0039] S301, extracting static geographic data of topography and land use in the study area, and generating initial and boundary meteorological data of the model with meteorological reanalysis data;

[0040] S302, configure the simulation area, time, and physical / chemical schemes, and simulate hourly anthropogenic CO2 concentration data using the annual anthropogenic emission inventory as prior information;

[0041] S303, Test the sensitivity of different physics scenarios or emissions inventories to assess the model’s response to the inputs.

[0042] In step S4, the anthropogenic CO2 emission estimation model and results based on the optimal estimation are verified, including:

[0043] S401, comparative analysis of the spatiotemporal continuous atmospheric CO2 dataset reconstructed by machine learning algorithms and the anthropogenic CO2 concentration data simulated by the atmospheric transport model;

[0044] S402, constructing a cost function and obtaining an emission estimation result when the cost function is the lowest;

[0045] S403, adjusting the initial conditions of the a priori emission inventory based on the preliminary simulation results and combined with the feedback from the optimal estimation model;

[0046] S404, re-simulating atmospheric transport based on the updated emission inventory to obtain optimized hourly anthropogenic CO2 concentration and corresponding emission estimation results;

[0047] S405, monitoring the dynamic changes of the urban CO2 plume through hourly simulation result data during the study period.

[0048] In step S402, the initial conditions of the prior emission inventory are adjusted, including: using an optimization estimation method, taking the prior emission inventory, wind speed, vegetation index, temperature and emissivity parameters in the surface parameters as state variables, and constructing a cost function;

[0049] The iterative method is used to solve the minimum value of the cost function. The error between the observed and simulated values ​​is reduced by gradually adjusting the state variables, and the emission estimation result is obtained when the cost function is the lowest.

[0050] In step S403, the initial conditions of the prior emission inventory are adjusted, including: correcting the CO2 emission intensity and time series distribution.

[0051] In step S405, the dynamic changes of the urban CO2 plume are monitored through the hourly simulation result data during the study period, including:

[0052] Local spatial autocorrelation analysis is used to extract high-value CO2 clusters as potential CO2 emission plume areas; the shape and range of the plume are determined in combination with the city's geographical location and atmospheric CO and NO2 concentrations; and the diffusion direction of emission sources is identified in combination with meteorological data on wind speed and direction.

[0053] Another object of the present invention is to provide an urban CO2 emission plume monitoring system integrating observation and simulation, which is implemented by the urban CO2 emission plume monitoring method integrating observation and simulation, and the system comprises:

[0054] Data collection module, used for multi-source data collection and preprocessing, the multi-source data includes: XCO2 secondary bias correction data set inverted from OCO-2 and OCO-3 observations; atmospheric CO and NO2 pollutant concentration data of the same emission source, vegetation index, meteorological reanalysis data; and ground-observed atmospheric CO2 concentration, anthropogenic emission inventory and static geographic data;

[0055] XCO2 prediction model building module, which builds an XCO2 prediction model based on machine learning and multi-source satellite observations, and verifies and evaluates the accuracy of the reconstructed XCO2 spatiotemporal continuous data set;

[0056] The long-term simulation module of anthropogenic CO2 concentration uses the long-term simulation of anthropogenic CO2 concentration of the atmospheric transport model to simulate the hourly anthropogenic CO2 concentration data;

[0057] Anthropogenic CO2 emission estimation module, anthropogenic CO2 emission estimation model and result verification based on optimal estimation; by constructing a cost function and obtaining the emission estimation result when the cost function is the lowest, and monitoring the dynamic changes of urban CO2 plumes through long-term data.

[0058] In combination with all the above technical solutions, the beneficial effects of the present invention are as follows:

[0059] (1) Multi-source data fusion improves monitoring accuracy: This invention combines multi-source satellite observation data (including CO2 and other pollutant concentrations, terrestrial ecological remote sensing parameters and meteorological data) and comprehensively considers different emission sources and environmental factors to provide more accurate CO2 emission plume monitoring. Compared with the single data source method, this solution can obtain more reliable results in a wider spatial range and under complex environmental conditions.

[0060] (2) Introducing machine learning algorithms and atmospheric transport models to improve emission plume identification and enhance method performance: This technical solution not only applies satellite monitoring technology, but also introduces machine learning algorithms and atmospheric transport models into CO2 emission monitoring. It can handle multi-dimensional, nonlinear and complex input variable relationships, automatically learn the correlation between different data, and track the long-distance transmission process of CO2. This advantage enables this technical solution to provide relatively stable and accurate emission prediction and plume identification when facing missing satellite observation data, uneven data, or complex interactions between pollutants and CO2.

[0061] (3) Long-term dynamic monitoring capability: Compared with traditional static monitoring methods, this invention realizes long-term dynamic monitoring of urban CO2 emission plumes through a combination of satellite observations, machine learning, and atmospheric transport models. This enables the technical solution to continuously track the changing trends of urban CO2 emissions and support more flexible emission control and environmental management decisions.

[0062] (4) Automated processing and identification: This solution has a high degree of automation and can automatically identify emission plumes based on multi-source data fusion and model simulation, reducing the need for manual intervention and improving monitoring efficiency. It is suitable for large-scale urban emission source monitoring and management.

[0063] (5) The present invention accurately monitors urban CO2 emissions, assists in emission plume identification and emission estimation, and provides a scientific basis for urban environmental governance and carbon emission reduction. It can achieve accurate positioning and resource optimization of high-emission areas among different cities and regions, which will help promote the development of the cross-city regional environmental monitoring and governance market and create significant commercial value.

[0064] (6) The present invention is original in terms of dynamic monitoring of emission plumes and precise positioning of emission sources. It solves the problem that there is no mature technical means for high-precision monitoring of urban CO2 emissions at home and abroad, especially the lack of solutions combining multi-source satellite observations and transmission simulations. It also solves the long-standing problem of difficulty in real-time and accurate monitoring of urban CO2 emission plumes. By optimizing data fusion and simulation methods, it realizes the effective integration and error minimization of multi-source data, making dynamic monitoring of CO2 emissions and positioning of emission sources possible, breaking through the monitoring accuracy that traditional technologies cannot achieve. Traditional CO2 emission monitoring relies on ground observations or single satellite data, and often underestimates the value of multi-source observation integration. The present invention overcomes this technical bias, integrates multi-source data and advanced optimization algorithms, fully explores the synergistic potential of satellite and model data, and realizes higher-precision monitoring of urban CO2 emissions, opening up new ideas for the industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] The accompanying drawings herein are incorporated in and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the description, serve to explain the principles of the present disclosure;

[0066] Figure 1 is a flow chart of a method for monitoring urban CO2 emission plume integrating observation and simulation provided by an embodiment of the present invention;

[0067] Figure 2 It is a schematic diagram of an urban CO2 emission plume monitoring system integrating observation and simulation provided by an embodiment of the present invention;

[0068] Figure 3 The results of urban CO2 emission plume monitoring in eastern China at 13:00 local time on January 21, 2019;

[0069] Figure 4 The results of urban CO2 emission plume monitoring in eastern China at 13:00 local time on January 22, 2019;

[0070] Figure 5 The results of urban CO2 emission plume monitoring in eastern China at 13:00 local time on January 23, 2019;

[0071] Figure 6 The results of urban CO2 emission plume monitoring in eastern China at 13:00 local time on January 24, 2019;

[0072] In the figure: 1. Data collection module; 2. XCO2 prediction model construction module; 3. Anthropogenic CO2 concentration long-term simulation module; 4. Anthropogenic CO2 emission estimation module. DETAILED DESCRIPTION

[0073] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without violating the connotation of the present invention, so the present invention is not limited by the specific implementation disclosed below.

[0074] The innovation of the present invention is that the present invention improves the identification accuracy of CO2 emission sources by integrating multi-source satellite data and atmospheric transmission models to construct an optimal estimation method. Combined with spatial autocorrelation analysis, high-concentration clusters are identified, the shape and diffusion path of emission plumes are determined, and high-precision and dynamic monitoring of urban CO2 emissions is achieved. This method provides important technical support for the positioning of urban emission plumes and the formulation of emission reduction policies.

[0075] The present invention provides a method to integrate multi-source satellite observation data, integrate machine learning and atmospheric transmission models, and develop an automated identification technology for urban CO2 emission plumes. Specifically:

[0076] (1) Integration and processing methods of multi-source satellite observation data: CO2 emission monitoring is carried out by combining the concentration data of pollutants such as CO2 and NO2 with meteorological and ecological parameters, filling the gap in satellite observation data;

[0077] (2) Emission plume identification algorithm based on machine learning and atmospheric transport model: monitoring capability at high resolution and dynamic time scale, automated processing to avoid human intervention, including continuous acquisition of satellite data and long-term operation and update mechanism of the transmission model, to ensure accurate capture of emission trends, especially the improvement of the atmospheric transport model in the problem of underestimation of emission simulation, to ensure that the emission simulation results are closer to actual observations.

[0078] Embodiment 1, as Figure 1 As shown, the urban CO2 emission plume monitoring method integrating observation and simulation provided by the embodiment of the present invention includes:

[0079] S1, multi-source data collection and preprocessing;

[0080] S101, collect the secondary bias-corrected XCO2 data set inverted from OCO-2 and OCO-3 observations; first, perform quality screening on the XCO2 data, and only retain high-quality observations with a quality flag (QF) of 0; then, apply a curve fitting function to extract the deterministic trends of the annual growth and seasonal cycle in the XCO2 time series, and decompose the XCO2 time series into two parts: background XCO2 (XCO 2,bg ) and short-term XCO2 changes (XCO 2,st), the temporal and spatial trend models of XCO2 were established by fitting XCO2 data within a moving window of every 2° latitude;

[0081]

[0082] XCO 2,st =XCO2-XCO 2,bg

[0083] In the formula, a0 and a1 are the intercept and slope of the polynomial function, t is the time unit number in months, i is the number of harmonic functions, ω is the period coefficient, and the calculation formula is 2π / 12, β i and γ i are the variation coefficients of the harmonic functions, δ filter is the interannual variation contained in the residual between the observed and fitted values ​​of XCO2; parameters a0, a1, a2, β i ,γ i All the results were calculated by the least squares fitting method.

[0084] Finally, the XCO2 observation data (XCO 2,st ) Gridded with a resolution of 0.1°, the monthly average was calculated, and only grid points with more than 5 valid points were retained. In order to avoid statistical errors caused by different sampling times of different satellites at the same location, OCO-2 and OCO-3 observations were gridded separately, and then the observation average was calculated for the same grid in the same month to ensure the consistency of spatial sampling of different satellites;

[0085] S102, collect multi-source data such as atmospheric CO and NO2 pollutant concentration data, vegetation index, meteorological reanalysis data, etc., and unify their temporal and spatial resolutions into monthly values ​​and 0.1° grid resolution; normalize these data sets respectively and use them as input data for the XCO2 prediction model;

[0086] S103, collect ground-based observations of atmospheric CO2 concentration, anthropogenic emission inventories, and static geographic data.

[0087] S2, building an XCO2 prediction model based on machine learning and multi-source satellite observations;

[0088] S201, using XCO 2,stData is used as the target variable, and the XCO2 prediction model is constructed based on the machine learning algorithm; multi-source data, including CO, NO2, NDVI, precipitation PRE, surface solar radiation SSR, surface air pressure SurfP, temperature TEM, east-west wind speed U10 and north-south wind speed V10 at a height of 10 meters, and monthly time labels are used as prediction variables to train the data samples composed of each grid gi and each month ti; the XCO2 prediction model expression is:

[0089]

[0090] Where XCO 2,st[gi,ti] is the XCO corresponding to grid gi and month ti 2,st Value, CO [gi,ti] is the CO concentration value corresponding to grid gi and month ti, NO 2[gi,ti] is the NO2 concentration value corresponding to grid gi and month ti, NDVI [gi,ti] is the vegetation index corresponding to grid gi and month ti, PRES [gi,ti] is the precipitation corresponding to grid gi and month ti, SSR [gi,ti] is the surface solar radiation value corresponding to grid gi and month ti, SurfP [gi,ti] is the surface pressure value corresponding to grid gi and month ti, TEM [gi,ti] is the temperature value corresponding to grid gi and month ti, U10 [gi,ti] and V10 [gi,ti] They are the east-west wind speed and north-south wind speed at a height of 10 meters corresponding to grid gi and month ti, where ti is the month and gi is the grid;

[0091] It can be understood that the present invention introduces CO as a prediction variable. For the XCO2 prediction model, the introduction of CO as a prediction variable can assist in the identification of CO2 emission sources and emission plumes. Technically: it adds input variables that can characterize carbon emissions and short-term atmospheric transport changes, improves the accuracy of the prediction model; makes up for the problem of insufficient single observation data, enhances the robustness and stability of the model, which is particularly important for areas with sparse CO2 data.

[0092] S202, the reconstructed XCO2 spatiotemporal continuous dataset was verified and its accuracy evaluated through ten-fold cross validation, feature importance analysis, and ground observation comparison methods.

[0093] S3, using the anthropogenic CO2 concentration simulation results of the atmospheric transport model, simulates the hourly anthropogenic CO2 concentration data;

[0094] The atmospheric transport model used in this technical solution is the WRF-Chem model, which is based on real-time online coupled simulation of meteorology and chemistry, combined with atmospheric dynamics and chemical reactions, to track and predict the spatial distribution and temporal changes of greenhouse gases and various pollutants. Operations include preparing input data such as meteorological data, emission data, and static geographic data; compiling and configuring models, setting operating parameters, running simulations, and outputting and analyzing results. The model has the ability to couple meteorological and chemical processes in real time, and can accurately simulate dynamic processes such as CO2 emissions, transmission, and diffusion, and can provide detailed information on the spatiotemporal distribution, source contribution, and diffusion paths of CO2. Its advantages lie in its high spatiotemporal resolution, strong coupling with meteorological conditions, and flexible emission source configuration, making it suitable for multi-scale CO2 research.

[0095] S301, extracting static geographic data of topography and land use in the study area, and generating initial and boundary meteorological data of the model with meteorological reanalysis data;

[0096] S302, configure the simulation area, time, and physical / chemical schemes, and simulate hourly anthropogenic CO2 concentration data using the annual anthropogenic emission inventory as prior information;

[0097] S303, Test the sensitivity of different physics scenarios or emissions inventories to assess the model’s response to the inputs.

[0098] Select different physical schemes and emission inventories, set up multiple experimental groups in the same region and time range, and run the WRF-Chem model group by group; compare the output data of each group of experiments, including the characteristics of spatiotemporal distribution changes, calculation deviations (such as root mean square error, absolute deviation), analyze the sensitivity of different physical schemes and emission inventories, determine the optimal configuration of the model, and explore the impact of input conditions on CO2 simulation results. This sensitivity test can help optimize the model settings and improve prediction accuracy.

[0099] S4, verify the anthropogenic CO2 emission estimation model and results based on optimal estimation;

[0100] S401, comparative analysis of the spatiotemporal continuous atmospheric CO2 dataset reconstructed by machine learning algorithms and the anthropogenic CO2 concentration data simulated by the atmospheric transport model;

[0101] Compare and analyze the results obtained by two different technical means, and compare the two sets of results through methods such as time-space matching, error statistics, and time-space consistency test;

[0102] S402, constructing a cost function and obtaining an emission estimation result when the cost function is the lowest;

[0103] Using the optimization estimation method, the cost function is constructed with the prior emission inventory, wind speed, vegetation index, and surface parameters (temperature and emissivity) as state variables. Cost function construction: Construct the cost function, taking the prior emission inventory, wind speed, vegetation index, surface parameters (such as temperature, emissivity) and other factors as state variables, and minimizing the cost function as the goal to make the simulated data as close to the observed data as possible. The cost function usually includes two parts: the observation term and the prior term. The observation term measures the difference between the simulated value and the observed value, and the prior term reflects the deviation of the state variable from the known emission inventory.

[0104] Gradient descent and iterative solution: Use gradient descent or other iterative optimization algorithms to optimize the cost function by gradually adjusting the state variables. At each iteration, the gradient of the cost function is calculated and the state variables are adjusted in the direction of reducing the error to minimize the difference between the simulated value and the observed value.

[0105] The minimum value of the cost function is solved by iterative method, and the error between the observed and simulated values ​​is reduced by gradually adjusting the state variables to obtain the emission estimation result when the cost function is the lowest.

[0106]

[0107] Where J(x) is the calculation result of the cost function, n is the number of state variables, T is the matrix transpose, y is the spatiotemporal continuous data of atmospheric CO2 reconstructed by the machine learning algorithm, H is the anthropogenic CO2 concentration data simulated by the atmospheric transport model, R is the observation error covariance matrix, which is used to measure the observation error; λ i is a weight parameter used to adjust the relative importance of different state variables in the cost function, x a,i is the prior estimate of the i-th state variable; B i is the prior error covariance matrix of the i-th state variable, which is used to measure the credibility of the prior information; x i is the i-th state variable to be estimated;

[0108] The first term in the formula is the observation deviation term, which represents the difference between the observed data and the simulated value, and measures the deviation between the simulation result and the actual observed value. By weighting with the error covariance matrix R, the influence of data points that are greatly affected by observations on the cost function is reduced. The second term is the prior constraint term, which represents the deviation between different state variables (such as emission inventory, wind speed, vegetation index, etc.) and their prior information. It is weighted using the prior error covariance matrix B to ensure that high-credibility prior information has a greater impact on the cost function. The importance of different state variables is adjusted through the weight parameter λ, making the model more flexible to adapt to specific application requirements.

[0109] Using the gradient descent optimization algorithm, the state variables are iteratively adjusted to find the minimum value of J(x). The process is as follows:

[0110] (1) Calculate the gradient of the cost function J(x) with respect to x

[0111]

[0112] (2) Update the state variable x by gradient descent:

[0113]

[0114] In the formula, x k+1 is the state variable after the k+1th gradient descent update, x k is the state variable at the kth iteration, α is the learning rate, which controls the step size of the update in each iteration. By choosing an appropriate α and the number of iterations k, the value of J(x) can be gradually reduced.

[0115] (3) During the iteration process, when the change of the cost function approaches zero or is less than the preset threshold, the iteration is stopped and it is considered that J(x) has reached the minimum value, and the optimal state variable x is obtained. * .

[0116] In the process of gradually adjusting the state variables to reduce the error, the difference between the observed value and the simulated value gradually decreases. In each iteration, the state variables are gradually adjusted to reduce the error between the observed and simulated values ​​through the following steps: (1) Initialization: First, use the prior information x a Initialize the state variable x and calculate the initial value J(x) of the cost function; (2) Error feedback: At each iteration, adjust the state variable according to the error yH between the observed value y and the simulated value H; (3) Weight adjustment: In the cost function, the weight λ reflects the importance of each state variable. Different state variables are dynamically adjusted through error feedback to make the model match the observed value more accurately. When the cost function reaches the minimum value, the optimal solution x of the state variable x is replaced by * Use the modified model running initial conditions to re-run the model. (Corresponding to the operation steps of S403 and S404)

[0117] S403, based on the preliminary simulation results and combined with the feedback from the optimal estimation model, adjust the initial conditions of the prior emission inventory, including correcting the CO2 emission intensity and time series distribution;

[0118] S404, re-simulating atmospheric transport based on the updated emission inventory to obtain optimized hourly anthropogenic CO2 concentration and corresponding emission estimation results;

[0119] The principle of the model is to input meteorological data, emission data and corresponding static geographic data, and run it to obtain anthropogenic CO2 concentration data. The updated emission inventory data is reprocessed by the model in terms of time and spatial resolution to obtain emission estimation results.

[0120] S405, monitoring the dynamic changes of the urban CO2 plume through hourly simulation result data during the study period.

[0121] Local spatial autocorrelation analysis is used to extract high-value CO2 clusters as potential CO2 emission plume areas; the shape and range of the plume are further determined in combination with the city's geographical location and atmospheric CO and NO2 concentrations; the diffusion direction of emission sources is identified in combination with meteorological data such as wind speed and direction; and the dynamic changes of urban CO2 plumes are monitored through long-term data.

[0122] Local spatial autocorrelation: The k-neighbor weight method is used to construct a row-standardized spatial weight matrix to describe the adjacency relationship of each spatial unit. The local autocorrelation index is calculated based on the attribute value of each spatial unit (such as CO2 concentration) and the values ​​of the surrounding units. The local autocorrelation index calculated by z-score or p-value is tested for significance. The statistical results of the Moran index will be divided into five types: high-high, low-low, high-low, low-high, and non-significant statistics. The spatial units with significant high-high values ​​(i.e., the spatial autocorrelation is positive and the high values ​​are significantly clustered) are screened out, and these units represent potential local high-value clustering areas.

[0123] Plume shape and extent: After obtaining preliminary high-high units, spatial connectivity analysis is performed to further determine local high-high clusters.

[0124] (1) The eight-neighborhood (up, down, left, right, and four diagonal neighbors) rule is used to define the connectivity between cells;

[0125] (2) Based on the adjacency rule, the connected regions of high-high cells are identified. That is, adjacent high-high cells form a continuous cluster, while isolated high-high cells are not considered clusters;

[0126] Identify the diffusion direction of emission sources: For obvious elliptical or rectangular high-high aggregation areas, use the ellipse fitting method to describe their major and minor axes; conduct joint analysis with meteorological wind field data to determine the diffusion direction of emission sources.

[0127] (1) Obtain the coordinate points of each spatial unit in the high-high clustering area, expressed as a set of points (xi, yi);

[0128] (2) Set the main axis direction to be along the wind direction, assuming an angle of θ, that is, the major axis (main axis) of the ellipse will be parallel to the wind direction: θ = arctan(v y / v x), where v x and v y are the components of the wind field in the x and y directions respectively;

[0129] (3) Construct the covariance matrix of the spatial unit coordinate points:

[0130]

[0131] in, and are the variances in the x and y directions, σ xy is the covariance in the x and y directions.

[0132] (4) Calculate the major and minor axes of the ellipse: Determine the lengths of the major and minor axes of the ellipse through the eigenvalues ​​of the covariance matrix. Assume that the eigenvalues ​​of the covariance matrix are λ1 and λ2 (λ1>λ2), then the lengths of the semi-major axis and semi-minor axis are

[0133] Where k is a scaling factor used to control the range covered by the ellipse, and k is usually set to 2 or 3 to cover 95% or 99% of the data points.

[0134] (5) Based on the calculated a, b, and θ, draw an ellipse to represent the diffusion shape and range of the high-concentration concentration area. Combined with the direction of the ellipse and the direction of the wind field, the diffusion trend and range of the high-concentration area are evaluated to identify possible pollution sources and diffusion paths.

[0135] By building a model that integrates multi-source data and multiple technical means, the present invention can more accurately capture and analyze the spatiotemporal variation characteristics of CO2 concentration in the atmosphere, and accurately identify pollution sources and diffusion paths. By combining multi-source information such as satellite observation data, meteorological data (such as wind speed and wind direction), and ground monitoring data, the present invention can perform refined dynamic monitoring of CO2 emissions in different regions.

[0136] At the same time, with the integration of machine learning algorithms and atmospheric transport models, the present invention improves the resolution and inversion accuracy of CO2 emission sources, reduces the errors caused by a single data source or model, and significantly enhances the ability to capture the spatiotemporal distribution of CO2 under complex meteorological conditions. The final effect is reflected in providing more accurate and timely CO2 concentration monitoring and emission source identification, providing scientific support for the formulation of emission reduction strategies and environmental management.

[0137] Embodiment 2, as Figure 2 As shown, an embodiment of the present invention provides an urban CO2 emission plume monitoring system integrating observation and simulation, the system comprising:

[0138] Data collection module 1 is used for multi-source data collection and preprocessing, wherein the multi-source data include: XCO2 secondary bias correction data set inverted from OCO-2 and OCO-3 observations; atmospheric CO and NO2 pollutant concentration data of the same emission source, vegetation index, meteorological reanalysis data; and atmospheric CO2 concentration observed on the ground, anthropogenic emission inventory and static geographic data;

[0139] XCO2 prediction model construction module 2, builds an XCO2 prediction model based on machine learning and multi-source satellite observations, and verifies and evaluates the accuracy of the reconstructed XCO2 spatiotemporal continuous data set;

[0140] Module 3 of long-term simulation of anthropogenic CO2 concentration uses the atmospheric transport model to simulate the hourly anthropogenic CO2 concentration data;

[0141] Anthropogenic CO2 emission estimation module 4, anthropogenic CO2 emission estimation model based on optimal estimation and result verification; by constructing a cost function and obtaining the emission estimation result when the cost function is the lowest, and monitoring the dynamic changes of urban CO2 plumes through long-term data.

[0142] It can be seen from the above embodiments that the present invention provides an integrated multi-source satellite observation data, fusion machine learning and atmospheric transmission model, and develops an automatic identification technology for urban CO2 emission plumes. Specifically:

[0143] Integration and processing methods of multi-source satellite observation data: CO2, NO2 and other pollutant concentration data are combined with meteorological and ecological parameters to monitor CO2 emissions, filling the gap in satellite observation data;

[0144] Emission plume identification algorithm based on machine learning and atmospheric transport model: monitoring capability at high resolution and dynamic time scale, automated processing process to avoid human intervention, including continuous acquisition of satellite data and long-term operation and update mechanism of transmission model, to ensure accurate capture of emission trends, especially the improvement of atmospheric transport model in the problem of underestimated emission simulation, to ensure that emission simulation results are closer to actual observations.

[0145] Furthermore, the present invention focuses on CO2 and pollutant concentration data observed by multi-source satellites, and by combining machine learning algorithms and atmospheric transport models, it can achieve large-scale monitoring on global and regional scales. Its advantage is that it has a wide coverage and can provide CO2 emission trends across large regions, and is particularly suitable for monitoring the long-term emission evolution of multiple large cities. CO2 emission monitoring based on ground-based observation networks and near-ground observations using drones or aircraft platforms can more accurately obtain emission data for specific cities or industrial areas, which is suitable for detailed investigations of short-term or local emission sources. These alternatives can be selected and combined according to specific monitoring needs, regional characteristics, and cost constraints to form a more diverse and flexible urban CO2 emission monitoring system.

[0146] The results of urban CO2 emission plume monitoring in eastern China at 13:00 local time from January 21 to 25, 2019 are as follows: Figure 3-Figure 6 shown.

[0147] (1) High emission areas: Large cities in eastern China usually have higher traffic density and industrial activities during the noon period, and emission plumes form a high CO2 concentration zone in the city and surrounding industrial areas.

[0148] (2) Plume diffusion: By comparing the diffusion paths of emission plumes on different dates, we can see that the areas with higher concentrations extend along the wind direction. In winter, the temperature in eastern China is lower and the wind speed is relatively low, so the diffusion speed of CO2 is slower and the emission plume is more concentrated.

[0149] (3) Topographic influence: The geographical environment in eastern China is relatively complex, and there are differences in the diffusion of CO2 emission plumes between coastal cities and inland cities. For example, coastal cities such as Shanghai are affected by sea breezes, and the emission plumes will diffuse in a certain direction; while inland cities such as Taiyuan and Chifeng are closed in topography, CO2 is difficult to diffuse effectively, and local accumulation in the atmosphere forms high-concentration CO2 concentration areas.

[0150] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principles of the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for monitoring urban CO2 emission plumes integrating observation and simulation, characterized in that: The method includes: S1, multi-source data collection and preprocessing; S2, building an XCO2 prediction model based on machine learning and multi-source satellite observations; S3, using the anthropogenic CO2 concentration simulation results of the atmospheric transport model, simulates the hourly anthropogenic CO2 concentration data; S4, verify the anthropogenic CO2 emission estimation model and results based on optimal estimation.

2. The urban CO2 emission plume monitoring method integrating observation and simulation according to claim 1 is characterized in that: In step S1, multi-source data collection includes: Collect the XCO2 secondary bias correction dataset derived from the observations of the greenhouse gas monitoring satellites OCO-2 and OCO-3; Collect multi-source data of atmospheric CO and NO2 pollutant concentration data, vegetation index, and meteorological reanalysis data from the same emission source; process the temporal and spatial resolution into monthly values ​​and 0.1° grid resolution, and perform normalization processing respectively as input data for the XCO2 prediction model; Collect ground-based observations of atmospheric CO2 concentrations, anthropogenic emissions inventories, and static geographic data.

3. The urban CO2 emission plume monitoring method integrating observation and simulation according to claim 2 is characterized in that: The XCO2 secondary bias correction dataset is collected from the greenhouse gas monitoring satellites OCO-2 and OCO-3 observations, including: First, the XCO2 data observed by satellites are quality screened and the observations with a quality flag QF of 0 are retained; Then, a curve fitting function is applied to extract the deterministic trends of annual growth and seasonal cycles in the XCO2 time series and decompose the XCO2 time series into the background XCO 2,bg and short-term changes in XCO 2,st , by fitting the XCO2 data in a moving window of every 2° latitude, a spatiotemporal trend model of XCO2 is established, and the expression is: In the formula, a0 and a1 are the intercept and slope of the polynomial function, t is the time unit number in months, i is the number of harmonic functions, ω is the period coefficient, and the calculation formula is 2π / 12, β i and γ i are the variation coefficients of the harmonic function, δ filter is the interannual variation contained in the residual between the observed and fitted values ​​of XCO2; parameters a0, a1, a2, β i ,γ i All are calculated by least squares fitting; Finally, the XCO2 observation data XCO of OCO-2 and OCO-3 were analyzed. 2,st Gridding was performed with a resolution of 0.1°, the monthly average was calculated, and grid points with more than 5 valid points were retained; the observation average was calculated for the same grid in the same month.

4. The urban CO2 emission plume monitoring method integrating observation and simulation according to claim 1 is characterized in that: In step S2, an XCO2 prediction model based on machine learning and multi-source satellite observations is constructed, including: S201, using XCO 2,st Data is used as the target variable, and the XCO2 prediction model is constructed based on the machine learning algorithm; multi-source data, including CO, NO2, vegetation index NDVI, precipitation PRE, surface solar radiation SSR, surface air pressure SurfP, temperature TEM, east-west wind speed U10 and north-south wind speed V10 at a height of 10 meters, and monthly time labels are used as prediction variables to train the data samples composed of each grid gi and each month ti; the XCO2 prediction model expression is: Where XCO 2,st[gi,ti] is the XCO corresponding to grid gi and month ti 2,st Value, CO [gi,ti] is the CO concentration value corresponding to grid gi and month ti, NO 2[gi,ti] is the NO2 concentration value corresponding to grid gi and month ti, NDVI [gi,ti] is the vegetation index corresponding to grid gi and month ti, PRES [gi,ti] is the precipitation corresponding to grid gi and month ti, SSR [gi,ti] is the surface solar radiation value corresponding to grid gi and month ti, SUrfP [gi,ti] is the surface pressure value corresponding to grid gi and month ti, TEM [gi,ti] is the temperature value corresponding to grid gi and month ti, U10 [gi,ti] and V10 [gi,ti] They are the east-west wind speed and north-south wind speed at a height of 10 meters corresponding to grid gi and month ti, where ti is the month and gi is the grid; S202, the reconstructed XCO2 spatiotemporal continuous dataset was verified and its accuracy evaluated through ten-fold cross validation, feature importance analysis, and ground observation comparison methods.

5. The urban CO2 emission plume monitoring method integrating observation and simulation according to claim 1 is characterized in that: In step S3, the anthropogenic CO2 concentration simulation results of the atmospheric transport model are used to simulate the anthropogenic CO2 concentration data per hour, including: S301, extracting static geographic data of topography and land use in the study area, and generating initial and boundary meteorological data of the model with meteorological reanalysis data; S302, configure the simulation area, time, and physical / chemical schemes, and simulate hourly anthropogenic CO2 concentration data using the annual anthropogenic emission inventory as prior information; S303, Test the sensitivity of different physics scenarios or emissions inventories to assess the model’s response to the inputs.

6. The urban CO2 emission plume monitoring method integrating observation and simulation according to claim 1 is characterized in that: In step S4, the anthropogenic CO2 emission estimation model and results based on the optimal estimation are verified, including: S401, comparative analysis of the spatiotemporal continuous atmospheric CO2 dataset reconstructed by machine learning algorithms and the anthropogenic CO2 concentration data simulated by the atmospheric transport model; S402, constructing a cost function and obtaining an emission estimation result when the cost function is the lowest; S403, adjusting the initial conditions of the a priori emission inventory based on the preliminary simulation results and combined with the feedback from the optimal estimation model; S404, re-simulating atmospheric transport based on the updated emission inventory to obtain optimized hourly anthropogenic CO2 concentration and corresponding emission estimation results; S405, monitoring the dynamic changes of the urban CO2 plume through hourly simulation result data during the study period.

7. The urban CO2 emission plume monitoring method integrating observation and simulation according to claim 5 is characterized in that: In step S402, the initial conditions of the prior emission inventory are adjusted, including: using an optimization estimation method, taking the prior emission inventory, wind speed, vegetation index, temperature and emissivity parameters in the surface parameters as state variables, and constructing a cost function; The iterative method is used to solve the minimum value of the cost function. The error between the observed and simulated values ​​is reduced by gradually adjusting the state variables, and the emission estimation result is obtained when the cost function is the lowest.

8. The method for monitoring urban CO2 emission plume integrating observation and simulation according to claim 5, characterized in that: In step S403, the initial conditions of the prior emission inventory are adjusted, including: correcting the CO2 emission intensity and time series distribution.

9. The urban CO2 emission plume monitoring method integrating observation and simulation according to claim 5 is characterized in that: In step S405, the dynamic changes of the urban CO2 plume are monitored through the hourly simulation result data during the study period, including: Local spatial autocorrelation analysis is used to extract high-value CO2 clusters as potential CO2 emission plume areas; the shape and range of the plume are determined in combination with the city's geographical location and atmospheric CO and NO2 concentrations; and the diffusion direction of emission sources is identified in combination with meteorological data on wind speed and direction.

10. An urban CO2 emission plume monitoring system integrating observation and simulation, characterized in that: The system is implemented by the urban CO2 emission plume monitoring method integrating observation and simulation as described in any one of claims 1 to 9, and the system comprises: A data collection module (1) is used for multi-source data collection and preprocessing, wherein the multi-source data includes: XCO2 secondary bias correction data set inverted from OCO-2 and OCO-3 observations; atmospheric CO and NO2 pollutant concentration data of the same emission source, vegetation index, meteorological reanalysis data; and ground-observed atmospheric CO2 concentration, anthropogenic emission inventory, and static geographic data; XCO2 prediction model building module (2) builds an XCO2 prediction model based on machine learning and multi-source satellite observations, and verifies and evaluates the accuracy of the reconstructed XCO2 spatiotemporal continuous data set; The long-term simulation module of anthropogenic CO2 concentration (3) uses the long-term simulation of anthropogenic CO2 concentration of the atmospheric transport model to simulate the hourly anthropogenic CO2 concentration data; The anthropogenic CO2 emission estimation module (4) is based on the optimal estimation anthropogenic CO2 emission estimation model and result verification; by constructing a cost function and obtaining the emission estimation result when the cost function is the lowest, and monitoring the dynamic changes of the urban CO2 plume through long-term data.

Images