Calculation method of carbon dioxide emissions from fossil fuels based on satellite observation of pollutant gas concentrations

Abstract

The present invention discloses a method for calculating carbon dioxide emissions from fossil fuels based on satellite observation of pollutant gas concentrations. First, the column concentrations of nitrogen dioxide and carbon monoxide observed by satellites and meteorological data are used to calculate the divergence of the two pollutants; then, the prior data of the "bottom-up" emission inventory are used to respectively construct inverse models for calculating nitrogen oxide and carbon monoxide emissions based on satellite remote sensing column concentrations; the coefficients of the inverse model are solved by multivariate linear regression to calculate the emissions of the two; finally, according to the obtained proportional relationship between nitrogen oxide and carbon monoxide emissions and carbon dioxide emissions, two carbon dioxide emissions are respectively calculated, and the best is obtained by complementing each other to obtain the final carbon dioxide emissions. The present invention solves the problem that it is difficult to directly calculate fossil fuel carbon emissions based on satellite observation of carbon dioxide concentrations due to the long survival life of carbon dioxide in the atmosphere and the limited spatial coverage of existing carbon satellite data.

Description

Technical Field

[0001] The present invention belongs to the technical field of atmospheric environment remote sensing application, and in particular relates to a method for calculating fossil fuel carbon dioxide emissions based on satellite observation of polluted gas concentrations. Background Art

[0002] Carbon dioxide (CO2) is recognized as the main greenhouse gas and the driving force of recent and future climate change. Its atmospheric concentration has increased significantly since the Industrial Revolution. Its growth rate has accelerated in the past decade, mainly attributed to anthropogenic sources such as fossil fuel combustion. Therefore, the calculation of CO2 emissions is crucial. At present, CO2 emission estimation methods can be divided into two categories. One is the emission factor method, also known as the "bottom-up" emission calculation method. It is the first carbon emission estimation method proposed by the IPCC and is also the most widely used method. Its basic idea is to construct its activity data and emission factor for each emission source according to the carbon emission inventory list, and use the product of activity data and emission factor as the carbon emission estimate of the emission project. The other category is the observation-based measurement method, also known as the "top-down" emission calculation method, which calculates emissions based on the CO2 concentration in the atmosphere observed on the ground or by satellite. The 49th IPCC Plenary Session in 2019 clarified the use of satellite remote sensing observations to calculate carbon flux through the "top-down" method to support and verify the emission inventory.

[0003] The "bottom-up" method of calculating CO2 emissions from fossil fuel combustion is mainly based on statistical information on economic activities and corresponding technologies, and has the disadvantages of huge workload, low reliability, and large lag. Its uncertainty is mainly due to the inaccurate available data on fuel consumption and fuel chemical composition. Even for power plants in the United States, which are considered to have the most accurate information on fuel usage in countries around the world, the difference between emissions estimated based on fuel usage and emissions reported as part of the Continuous Emissions Monitoring System (CEMS) program is usually about 20%. Therefore, the calculation of CO2 emissions based on satellite observations proposed in the present invention can be used as an important means to verify and improve the current CO2 emission inventory.

[0004] CO2 emissions can be estimated through satellite observations of CO2 concentrations, but existing CO2 satellite sensors (SCIAMACHY, GOSAT, OCO-2, etc.) have limited spatial coverage and cannot distinguish between carbon emissions from fossil fuels and biomass. In addition, CO2 has a long lifespan in the atmosphere, so it is very difficult to estimate emissions directly using satellite remote sensing observations of CO2 concentrations. Since fossil fuel combustion emits both CO2 and nitrogen oxides (NO x ), carbon monoxide (CO) and other pollutants. Compared with CO2, NO xThe atmospheric survival time of CO2 and CO2 is short, which is more representative of local anthropogenic emissions. The corresponding satellite payload has a wide observation range and has the ability to cover the world every day. At present, the Tropospheric Monitoring Instrument (TROPOMI), as a single payload of the European Space Agency's Sentinel-5Precursor satellite mission, is the current satellite remote sensing payload of pollutant gases with the highest spatial resolution, which can identify polluted air masses from large power plants. Therefore, the present invention proposes to invert the pollutant gases (NO2) from fossil fuel combustion through satellite observation. x CO2 emissions can be calculated by using CO and CO) emissions, which can quantify CO2 emissions from fossil fuels and exclude carbon emissions from biogenic sources.

[0005] The satellite data involved in this invention mainly include TROPOMI tropospheric NO2 column concentration, CO column total concentration and China Multiscale Emission Inventory Model (MEIC) data. The following are the production technologies of the three main data.

[0006] TROPOMI tropospheric NO2 column concentration: The TROPOMI NO2 inversion system is developed based on the algorithm of DOMINO-2 products and EU QA4ECVNO2 reprocessed OMI datasets and has been improved accordingly. The inversion process includes three steps: (1) The total NO2 oblique column density is inverted from the Level-1b radiation and irradiance spectra measured by TROPOMI using the DOAS method. (2) Based on the information of a data assimilation system, the total NO2 oblique column density is divided into two parts: stratosphere and troposphere. (3) The tropospheric oblique column density is converted to the tropospheric vertical column density.

[0007] TROPOMICO column total concentration data: generated using an improved solar spectral shortwave infrared CO inversion method, the SCIOR algorithm. This algorithm is based on the SCIAMACHY CO inversion algorithm and radiative transfer experience over the past three decades, with additional new features added to the inversion to fully exploit the potential of the TROPOMI SWIR measurements in data processing. This algorithm improves on the current CO inversion methods used for cloudy and aerosol-laden atmospheres. Accurate treatment of clouds and aerosols in the inversion requires simulation of multiple light scattering, which is numerically demanding, so the SICOR algorithm uses a two-stream radiative transfer solver to account for atmospheric scattering in a simplified way.

[0008] China Multiscale Emission Inventory Model (MEIC) data: provides a "bottom-up" emission inventory of CO2, CO and NO2. MEIC has built a unified source classification and grading system and emission factor database, developed emission source characterization methods based on technology and dynamic processes, developed a multi-scale high-resolution emission source model that is seamlessly linked to mainstream atmospheric chemistry models, and further integrated cloud computing and big data technologies to achieve real-time dynamic calculation of emission inventories. Currently, the MEIC model can provide online multi-scale emission inventory data for 10 major air pollutants and carbon dioxide from more than 700 anthropogenic sources in China from 1990 to the present. The emission inventory data developed based on the MEIC model has been widely used by many research institutions and business units at home and abroad, and has supported many large-scale research programs internationally, such as MICS-Asia, HTAP, and CEDS. Summary of the invention

[0009] In order to solve the difficulty of estimating CO2 emissions "top-down" based on satellite observations due to the long survival time of CO2 in the atmosphere and the limited spatial coverage of existing carbon satellite data, the purpose of the present invention is to provide a method for calculating carbon dioxide emissions from fossil fuel combustion based on satellite observations of pollutant gas concentrations.

[0010] To achieve the above object, the technical solution adopted by the present invention is:

[0011] A method for calculating carbon dioxide emissions from fossil fuels based on satellite observation of pollutant gas concentrations comprises the following steps:

[0012] (1) Using TROPOMI NO2 tropospheric column concentration observation data and European Centre for Medium-Range Weather Forecasts (ECMWF) wind field reanalysis data, nitrogen oxides (NO x ) and NO2 proportional factor to obtain NO x Column concentration average level flux;

[0013] (2) Based on the Continuity Equation for Steady State, calculate NO x Column concentration horizontal flux divergence;

[0014] (3) Using the Multi-resolution Emission Inventory for China (MEIC) data with a resolution of 0.1°, grid conversion and resampling were performed, invalid grids in the data were deleted, and “bottom-up” prior emission data with a resolution of 0.05° were obtained;

[0015] (4) Construct calculation NO xThe inverse emission model is used, and the TROPOMI nitrogen dioxide (NO2) tropospheric column concentration observation data in step (1) is used as the first and second data items, the horizontal flux divergence obtained in step (2) is used as the third data item, and the Multi-resolution Emission Inventory for China (MEIC) data in step (3) is used as the prior emissions;

[0016] (5) Multiple linear regression was used to fit the three types of coefficients in the inverse emission model pixel by pixel, and their probability density functions were calculated. The Bootstrap method was used to obtain NO x The coefficient ranges for the three categories of coefficients in the inverse model;

[0017] (6) Using the observed data of TROPOMI carbon monoxide (CO) column concentration, the same operations as steps (1) to (5) were performed to obtain the three types of coefficients of the CO emission inverse model and compared with the NO x The three types of coefficients of the inverse model are verified and optimized;

[0018] (7) Using the three types of coefficients obtained, the column concentration data observed by satellite are brought into the inverse model to obtain the final NO x and CO emissions;

[0019] (8) Obtaining the “bottom-up” annual average anthropogenic NO based on MEIC x Calculate NO with CO2 emission data x Conversion factors to CO2 emissions;

[0020] (9) Using the emission factor obtained in step (8), substitute the NO obtained in step (7) x Emissions, calculate and obtain CO2 emissions;

[0021] (10) Based on the MEIC, the annual average anthropogenic CO and CO2 emissions data are obtained from the bottom up. The same operations are performed according to steps (8) to (9) to obtain CO2 emissions.

[0022] (11) The two sets of CO2 emission data obtained in steps (9) and (10) are optimized. When there are missing values ​​in the emission data, the two sets are filled in. When the emission data exist at the same time, the average of the two sets is taken to obtain the final CO2 emission data.

[0023] In the step (1), the NO2 tropospheric column concentration, NO x / NO2 proportional factor and wind field data, and calculate the average horizontal flux of NO2 column concentration after local background correction;

[0024] The wind speed of the wind field data is 950hPa, and the satellite transit time is 13:30 local time. Therefore, the wind field data at 13:00 and 14:00 are used to calculate the average value, and the resolution is consistent with the satellite data after resampling.

[0025] Seek NO x When calculating the horizontal flux of column concentration, taking the average horizontal flux value can prevent the ambiguity of NO2 tropospheric column concentration caused by wind field changes. x The equation for flux is expressed as:

[0026] F=L×V×W

[0027] Where V represents the NO2 tropospheric column concentration, and the data is converted to μg / m 2 , W represents the wind field, the unit is m / s, L is NO x / NO2 proportional factor; NO in the local area x The / NO2 ratio factor is affected by ozone (O3) because NO in the form of nitric oxide (NO) x It cannot be observed by satellites. Once NO is mixed with O3 and converted into NO2, the NO2 flux will increase, forming a temporary virtual source. If the conversion occurs within 25 km around the emission source to consider the fitting interval, the total integral when calculating the divergence will be less affected. Therefore, NO is selected. x The / NO2 scaling factor L is 1.32, representing a typical value for the TROPOMI satellite.

[0028] In step (2), according to the steady-state continuity equation, the flux divergence produces NO x Sources and sinks, through NO x The average flux of column concentration was used to determine NO x Horizontal flux divergence of column concentration, calculation of NO x The horizontal flux divergence equation for column concentration is expressed as:

[0029]

[0030] Among them, V represents the NO2 tropospheric column concentration, W represents the wind field, and L represents the NO x / NO2 scaling factor, Indicates NO x The spatial derivative of the average flux represents the flux density outward from the point. The numerical spatial derivative can calculate the pixel value (flux value) using the fourth-order central finite difference method. The pixel distribution is divided into u direction and v direction according to the wind direction. u represents the wind in the longitude direction, the same latitude, and v represents the wind in the latitude direction, the same longitude. Take the pixel y in the two directions respectively. t The values ​​of the four surrounding pixels (yt-2 ,y t-1 ,y t+1 ,y t+2 ), h represents the distance corresponding to the unit pixel, and the haversine formula is used to calculate the distance between the unit pixels in two directions. The specific solution formula of the fourth-order central finite difference method is expressed as (y t-2 -8y t-1 +8y t+1 -y t+2 ) / 12*h, and obtain the divergence in the u direction and the v direction, and finally get the total divergence D; E and S are the source and sink respectively. When D is greater than 0, the point is the source of the radiated flux, and when D is less than 0, the point is the sink of the absorbed flux.

[0031] In the step (3), a mask is generated according to the high-value area using the China Multi-scale Emission Inventory Model data (MEIC) with a resolution of 0.1°, and the high-value area grids with values ​​greater than 97.5% percentile are taken as valid grids, and resampled to obtain the 0.05°*0.05° grid unit data of the prior emission.

[0032] In step (4), according to the principle of conservation of mass, the following equation is constructed to estimate NO x Inverse model of emissions:

[0033]

[0034] Among them, all the terms in the formula are converted into a matrix of uniform spatial range, which consists of each pixel, x, y is the position information of the pixel in the matrix, the subscript i represents the observation date of the satellite, C(x, y) represents the NO2VCDs obtained by satellite observation, V(x, y) represents the wind field, and the first term in the equation represents the rate of change of TVCDs of NO2 to NO x The second term represents the impact of NO x The third term is the horizontal flux divergence obtained in step (2) as the third data, representing the convection term, and the three types of coefficients α1, α2, and α3 are the three coefficient ranges that need to be calculated based on the prior emission data;

[0035] In the first item, It indicates the change in gas column concentration during the satellite transit between two days. Generally, the satellite transit time is one day, and the calculation unit is μg / m 2 *s,

[0036] The third divergence item was screened in time to remove outliers. Days when the pixel value in the high-value area of ​​divergence D was less than 0 were considered invalid, and the 3-times sigma method was used to remove outliers.

[0037] In step (5), the NO2 VCDs data obtained from satellite observations and the MEIC prior data are used to fit the inverse emission model pixel by pixel according to the multivariate linear regression. The regression process is repeated for each month to obtain the three types of coefficients α1, α2, and α3 to draw a probability density function (PDF) graph, and the Bootstrap confidence interval is calculated to obtain the NO x The coefficient range of the three types of coefficients in the inverse model uses the kernel density estimation (KDE) method to estimate the density function of the coefficient sample, and the coefficient value corresponding to the maximum value point of the KDE fitting curve is taken as the estimated value;

[0038] In step (6), the proportional factor L is not required when calculating the average CO flux. By performing the same operations as steps (1) to (5), the ranges of the three types of coefficients in the CO inverse model are obtained, which are consistent with those of NO. x The three types of coefficients in the inverse model are verified and optimized to determine the three types of coefficients in the inverse model.

[0039] In step (7), the three types of coefficients obtained after verification and optimization are substituted into the inverse model equation, and the TROPOMI NO2 and CO column concentration data observed by satellite and the divergence data calculated separately are substituted into the inverse model to obtain the final NO x and CO emissions and

[0040] In step (8), the NO of the whole department is obtained from the bottom up based on MEIC. x With CO2 emission data, calculate CO2 and NO x The conversion factor F (Conversion Factor) of emissions is as follows:

[0041]

[0042] in, and All are “bottom-up” emissions data.

[0043] In step (9), the emission factor F obtained in step (8) and the NO obtained in step (7) are used to calculate the emission factor F. x Emissions The carbon emissions from fossil fuel combustion based on satellite observations are estimated as follows:

[0044]

[0045] In step (10), the annual average anthropogenic CO and CO2 emission data are obtained based on MEIC, and the same operations as steps (8) to (9) are performed to obtain CO2 emissions.

[0046] In step (11), the two sets of CO2 emission data obtained in steps (9) and (10) are optimized. x In case of missing value, it is filled by CO; in case of missing value of CO, it is filled by NO. x To fill; when NO x When it exists with CO, the average of the two is taken to obtain the final CO2 emissions.

[0047] Beneficial effect: The method for calculating carbon dioxide emissions provided by the present invention first calculates the "indicator" emissions in advance and calculates the CO2 emissions based on the relationship between the "indicator" and CO2. Compared with the "bottom-up" emission factor method for calculating CO2 emissions with a long survival life, the method is more accurate; secondly, the present invention adopts NO x The two "indicators" of NO and CO are cross-validated and mutually compensated for and optimized to obtain more reliable results. At the same time, the method for calculating the emission of "indicators" is also based on the principle of mass conservation, building an inverse model, solving the coefficients, and better obtaining NO x and CO emissions. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is the overall flow chart of the present invention;

[0049] Figure 2 for Figure 1 Flowchart of step one in the figure;

[0050] Figure 3 for Figure 1 Flowchart of step 2 in the figure;

[0051] Figure 4 for Figure 1 Flowchart for step three in . DETAILED DESCRIPTION

[0052] The present invention will be further described below in conjunction with the accompanying drawings.

[0053] like Figure 1 The main framework flow of a method for calculating carbon dioxide emissions based on satellite-observed atmospheric component concentrations is shown, including the following steps:

[0054] (1) Using TROPOMI NO2 tropospheric column concentration observation data and European Centre for Medium-Range Weather Forecasts (ECMWF) wind field reanalysis data, nitrogen oxides (NO x) and NO2 proportional factor to obtain NO x Based on the Continuity Equation for Steady State, NO x Column concentration horizontal flux divergence.

[0055] (2) Using the 0.1° resolution Multi-resolution Emission Inventory for China (MEIC) data, we screened high-value grids as valid grids and obtained the prior emission grid cell data with a resolution of 0.05°. We then constructed an estimation of NO x The emission inverse model is constructed by bringing the horizontal flux divergence and other data obtained in step (1) into the inverse model, and using multivariate linear regression to solve the three types of coefficients in the emission inverse model pixel by pixel.

[0056] (3) Using the observed data of TROPOMI carbon monoxide (CO) column concentration, the same operations as steps (1) and (2) were performed to obtain the three types of coefficients of the CO emission inverse model and compared with the NO x The three types of coefficients of the inverse model are verified and optimized.

[0057] (4) Using the three types of coefficients obtained, the column concentration data observed by satellite are brought into the inverse model to obtain the final NO x and CO emissions.

[0058] (5) Obtaining the “bottom-up” annual average anthropogenic NO based on MEIC x Calculate NO with CO2 emission data x The conversion factor of NO obtained in step (4) is used to calculate the conversion factor of CO2 emissions. x Emissions, calculate the individual emission estimates for the combined CO2 emissions estimate, and obtain the NO x The amount of CO2 emitted.

[0059] (6) Based on MEIC, the annual average anthropogenic CO and CO2 emissions data are obtained. The same operations are performed according to steps (4) and (5) to obtain CO2 emissions based on CO emissions.

[0060] (7) The two sets of CO2 emission data obtained in steps (5) and (6) are optimized. x In case of missing value, it is filled by CO; in case of missing value of CO, it is filled by NO. x To fill; when NO x When CO and CO exist at the same time, the average of the two is taken to obtain the final CO2 emissions.

[0061] like Figure 2 The figure shows a method for calculating NO based on the principle of conservation of mass. x The method for calculating emissions, as a detailed description of step 1 of the general flow chart, includes the following steps:

[0062] (1.1) Using TROPOMI nitrogen dioxide (NO2) tropospheric column concentration observations and European Centre for Medium-Range Weather Forecasts (ECMWF) wind field reanalysis data, NO2 tropospheric column concentrations are corrected by local background correction. Without background correction, the sink term will be systematically increased. For estimated point source emissions, background correction has no effect; for "top-down" spatial integration calculated on a large scale, background correction is relevant.

[0063] Using nitrogen oxides (NO x ) and NO2 proportional factor, NO x The / NO2 ratio factor is generally taken as L = 1.32, which represents the typical value of the selected TROPOMI. x The / NO2 ratio factor is different from the L balance value, but as long as the conversion occurs within the 25 km fitting interval around the source, it is not affected and reaches equilibrium. x / NO2 is still suitable.

[0064] Seek NO x When calculating the horizontal flux of column concentration, taking the average horizontal flux value can prevent the NO2 tropospheric column from being blurred due to wind changes, and maintain the point source strong gradient in the time average. x The average column concentration flux.

[0065] (1.2) Based on the Continuity Equation for Steady State, the divergence of flux produces NO x Sources and sinks, through NO x The average flux of column concentration was used to determine NO x The divergence of the horizontal flux of column concentration is used to calculate NO x The divergence equation for the horizontal flux of column concentration can be expressed as:

[0066]

[0067] Where V represents the NO2 tropospheric column concentration, W represents the wind field, L is the proportional factor, and E and S are the source and sink, respectively.

[0068] (1.3) Using the Multi-resolution Emission Inventory for China (MEIC) data with a resolution of 0.1°, grid conversion and resampling were performed, invalid grids in the data were deleted, and grid cells in the high-value area of ​​the MEIC mean map with a value greater than the 95th percentile were selected as valid grids to obtain the prior emission grid cell data with a resolution of 0.05°.

[0069] (1.4) According to the principle of mass conservation, we can estimate NO x The inverse emission model of NO x The tropospheric column can be determined by the balance maintained between emissions E, chemical-physical losses L and transport Q, NO x The velocity formula of the tropospheric column can be expressed as:

[0070]

[0071] The above tropospheric column rate formula can be used to construct an estimate of NO x The emission inverse model is:

[0072]

[0073] The subscript i represents the satellite observation date, C(x, y) represents the NO2 VCDs obtained by satellite observation, and V (x, y) represents the wind field. x The effect of the rate of change of TVCDs of NO2 in emission estimation, the second term represents the NO x The third term is the horizontal flux divergence obtained in step (1.2) as the third data, representing the advection term.

[0074] (1.5) Based on the prior emission data obtained in step (1.3), the horizontal flux divergence obtained in step (1.2) is substituted as the third term, and the NO2 VCDs data obtained from satellite observations are used to solve the emission inverse model pixel by pixel according to multivariate linear regression to obtain three types of coefficients α1, α2, and α3 to draw the PDF diagram, and the NO2 VCDs data obtained in step (1.3) are used to obtain the NO2 VCDs data. x Coefficient ranges for the three categories of coefficients in the inverse model.

[0075] like Figure 3 The figure shows a method for calculating CO emissions based on the principle of conservation of mass. As a detailed explanation of step 2 of the overall flow chart, the observed data of TROPOMI carbon monoxide (CO) column concentration is used to perform the same algorithm operation as step 1 to obtain three types of coefficients of the CO emission inverse model.

[0076] The three types of coefficients of the inverse model of CO emission and NO x The three types of coefficients of the inverse model are verified and optimized, the three types of coefficients in the inverse model are determined, the obtained coefficients are substituted into the inverse model equation, and the satellite observation data are substituted into the inverse model equation to obtain the final NO x and CO emissions.

[0077] like Figure 4 Shown are emission factors based on NO x The calculation method of CO2 emissions, as a detailed explanation of step 3 of the overall flow chart, includes the following steps:

[0078] (3.1) Obtaining the “bottom-up” average annual anthropogenic NOx based on MEIC x With CO2 emission data, NO x Conversion factors to CO2 emissions (Conversion Factor), the equation is as follows:

[0079]

[0080] in and All are annual average anthropogenic sources from “bottom up”.

[0081] (2) Calculate the emission factor obtained according to step (3.1) Using the previously obtained NO x Emissions Calculate individual emission estimates for mixed CO2 emissions estimates Finally, we get the value based on NO x Emissions CO2 emissions, the process is as follows:

[0082]

[0083]

[0084] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for calculating carbon dioxide emissions from fossil fuels based on satellite observation of pollutant gas concentrations, characterized in that: The following steps are involved: (1) Using TROPOMI nitrogen dioxide tropospheric column concentration observation data and European Center for Medium-Range Weather Forecasts wind field reanalysis data, nitrogen oxides NO x NO is calculated with the NO2 proportional factor x Column concentration average level flux; (2) Based on the steady-state continuity equation, calculate NO x Column concentration horizontal flux divergence; (3) Using the Chinese multi-scale emission inventory model data with a resolution of 0.1°, grid conversion and resampling were performed, invalid grids in the data were deleted, and "bottom-up" prior emission data with a resolution of 0.05° were obtained; (4) Construct calculation NO x The inverse emission model is constructed, and the TROPOMI nitrogen dioxide tropospheric column concentration observation data in step (1) is used as the first and second data, the horizontal flux divergence obtained in step (2) is used as the third data, and the China Multi-Scale Emission Inventory Model data in step (3) is used as the prior emissions; In step (4), according to the principle of conservation of mass, the following estimation of NO is constructed: x Inverse model of emissions: Among them, all the terms in the formula are converted into a matrix of uniform spatial range, which consists of each pixel, x, y is the position information of the pixel in the matrix, the subscript i represents the observation date of the satellite, C(x, y) represents the NO2VCDs obtained by satellite observation, V(x, y) represents the wind field, and the first term in the equation represents the rate of change of TVCDs of NO2 to NO x The second term represents the impact of NO x The third term is the horizontal flux divergence obtained in step (2) as the third data, representing the convection term, and the three types of coefficients α1, α2, and α3 are the three coefficient ranges that need to be calculated based on the prior emission data; In the first item, It indicates the change in gas column concentration during the satellite transit between two days. Generally, the satellite transit time is one day, and the calculation unit is μg / m 2 *s, The third divergence item was screened in time to remove outliers. Days when the pixel value in the high-value area of ​​divergence D was less than 0 were considered invalid, and the 3-times sigma method was used to remove outliers. (5) Multiple linear regression was used to fit the three types of coefficients in the inverse emission model pixel by pixel, and their probability density functions were calculated. The Bootstrap method was used to obtain NO x The coefficient ranges for the three categories of coefficients in the inverse model; (6) Using the observed data of TROPOMI carbon monoxide column concentration, the same operations are performed according to steps (1) to (5), and the three types of coefficients of the inverse model of CO emission are obtained and compared with those of NO x The three types of coefficients of the inverse model are verified and optimized; (7) Using the three types of coefficients obtained, the column concentration data observed by satellite are brought into the inverse model to obtain the final NO x and CO emissions; (8) Obtaining the “bottom-up” annual average anthropogenic NO based on MEIC x Calculate NO with CO2 emission data x Conversion factors to CO2 emissions; (9) Using the emission factor obtained in step (8), substitute the NO obtained in step (7) x Emissions, calculate and obtain CO2 emissions; (10) Based on the MEIC, the annual average anthropogenic CO and CO2 emissions data are obtained from the bottom up. The same operations are performed according to steps (8) to (9) to obtain CO2 emissions. (11) The two sets of CO2 emission data obtained in steps (9) and (10) are optimized. When there are missing values ​​in the emission data, the two sets are filled in. When the emission data exist at the same time, the average of the two sets is taken to obtain the final CO2 emission data.

2. The method for calculating fossil fuel carbon dioxide emissions based on satellite observation of pollutant gas concentrations according to claim 1, characterized in that: In step (1), the NO2 tropospheric column concentration, NO x / NO2 proportional factor and wind field data, and calculate the average horizontal flux of NO2 column concentration after local background correction; The wind speed of the wind field data is 950hPa, and the satellite transit time is 13:30 local time. Therefore, the wind field data at 13:00 and 14:00 are used to calculate the average value, and the resolution is consistent with the satellite data after resampling. Seek NO x When calculating the horizontal flux of column concentration, taking the average horizontal flux value can prevent the ambiguity of NO2 tropospheric column concentration caused by wind field changes. x The equation for the flux is expressed as: F=L×V×W Where V represents the NO2 tropospheric column concentration, and the data is converted to μg / m 2 , W represents the wind field, the unit is m / s, L is NO x / NO2 proportional factor; NO in the local area x The / NO2 ratio factor is affected by ozone because NO in the form of nitric oxide x It cannot be observed by satellites. Once NO is mixed with O3 and converted into NO2, the NO2 flux will increase, forming a temporary virtual source. If the conversion occurs within 25 km around the emission source to consider the fitting interval, the total integral when calculating the divergence will be less affected. Therefore, NO is selected. x The / NO2 scaling factor L is 1.32, representing a typical value for the TROPOMI satellite.

3. The method for calculating fossil fuel carbon dioxide emissions based on satellite observation of pollutant gas concentrations according to claim 1, characterized in that: In step (2), according to the steady-state continuity equation, the flux divergence produces NO x Sources and sinks, through NO x The average flux of column concentration was used to determine NO x Horizontal flux divergence of column concentration, calculation of NO x The horizontal flux divergence equation for column concentration is expressed as: Among them, V represents the NO2 tropospheric column concentration, W represents the wind field, and L represents the NO x / NO2 scaling factor, Indicates NO x The spatial derivative of the average flux represents the flux density outward from the point. The numerical spatial derivative can calculate the pixel value using the fourth-order central finite difference method. The pixel distribution is divided into u direction and v direction according to the wind direction. u represents the wind in the longitude direction, the same latitude, and v represents the wind in the latitude direction, the same longitude. Take the pixel y in the two directions respectively. t The values ​​of the four surrounding pixels (y t-2 ,y t-1 ,y t+1 ,y t+2 ), h represents the distance corresponding to the unit pixel, and the haversine formula is used to calculate the distance between the unit pixels in two directions. The specific solution formula of the fourth-order central finite difference method is expressed as (y t-2 -8y t-1 +8y t+1 -y t+2 ) / 12*h, and obtain the divergence in the u direction and the v direction, and finally get the total divergence D; E and S are the source and sink respectively. When D is greater than 0, the point is the source of the radiated flux, and when D is less than 0, the point is the sink of the absorbed flux.

4. The method for calculating fossil fuel carbon dioxide emissions based on satellite observation of pollutant gas concentrations according to claim 1, characterized in that: In step (3), the Chinese multi-scale emission inventory model data with a resolution of 0.1° is used to generate a mask based on the high-value area, and the high-value area grids with values ​​greater than 97.5% percentile are taken as valid grids, and resampled to obtain the 0.05°*0.05° grid unit data of the prior emission.

5. The method for calculating fossil fuel carbon dioxide emissions based on satellite observation of pollutant gas concentrations according to claim 1, characterized in that: In step (5), the NO2 VCDs data obtained from satellite observations and the MEIC prior data are used to fit the inverse emission model pixel by pixel according to the multivariate linear regression. The regression process is repeated for each month to obtain the three types of coefficients α1, α2, and α3 to draw the probability density function diagram, and calculate the Bootstrap confidence interval to obtain NO x The coefficient range of the three types of coefficients in the inverse model uses the kernel density estimation method to estimate the density function of the coefficient sample, and the coefficient value corresponding to the maximum value of the KDE fitting curve is taken as the estimated value; In step (6), the proportional factor L is not required when calculating the average CO flux. By performing the same operations as steps (1) to (5), the ranges of the three types of coefficients in the CO inverse model are obtained, which are consistent with those of NO. x The three types of coefficients in the inverse model are verified and optimized to determine the three types of coefficients in the inverse model.

6. The method for calculating fossil fuel carbon dioxide emissions based on satellite observation of pollutant gas concentrations according to claim 1, characterized in that: In step (7), the three types of coefficients obtained after verification and optimization are substituted into the inverse model equation, and the TROPOMI NO2 and CO column concentration data observed by satellite and the divergence data calculated separately are substituted into the inverse model to obtain the final NO x and CO emissions and 7. The method for calculating fossil fuel carbon dioxide emissions based on satellite observation of pollutant gas concentrations according to claim 1, characterized in that: In step (8), the NO of the whole department is obtained from the bottom up based on MEIC. x With CO2 emission data, calculate CO2 and NO x The conversion factor F for emissions is given by the following equation: in, and All are "bottom-up" emissions data.

8. The method for calculating fossil fuel carbon dioxide emissions based on satellite observation of pollutant gas concentrations according to claim 1, characterized in that: In step (9), the emission factor F obtained in step (8) and the NO obtained in step (7) are used to calculate the emission factor F. x Emissions The carbon emissions from fossil fuel combustion based on satellite observations are estimated as follows:

9. The method for calculating fossil fuel carbon dioxide emissions based on satellite observation of pollutant gas concentrations according to claim 1, characterized in that: In step (10), the annual average anthropogenic CO and CO2 emissions data are obtained based on MEIC, and the same operations as steps (8) to (9) are performed to obtain CO2 emissions.

10. The method for calculating fossil fuel carbon dioxide emissions based on satellite observation of pollutant gas concentrations according to claim 1, characterized in that: In step (11), the two sets of CO2 emission data obtained in steps (9) and (10) are optimized. x In case of missing value, it is filled by CO; in case of missing value of CO, it is filled by NO. x To fill; when NO x When it exists with CO, the average of the two is taken to obtain the final CO2 emissions.

Images