Method for inverting regional methane emission based on satellite column concentration observation

Abstract

The application discloses a method for inverting regional methane emission based on satellite column concentration observation, and belongs to the technical field of atmospheric remote sensing monitoring, and comprises the following steps: step 1, modifying a CMAQ regional atmospheric chemical transmission model, adding CH4 species as a trace gas to simulate an atmospheric transmission process, and modifying an initial condition module and a boundary condition module in the CMAQ, so that the CMAQ can perform ensemble simulation on CH4; and step 2, based on the CMAQ regional atmospheric chemical transmission model and an EnKF inversion algorithm, a high-resolution nested emission inversion system is constructed; the high-resolution regional methane emission inversion system constructed through the innovative algorithm and program can directly assimilate satellite column concentration data, significantly reduces the uncertainty of CH4 emission estimation, and improves the precision of CH4 simulation. Meanwhile, the system has significant advantages in CH4 emission monitoring and evaluation.

Description

Technical Field

[0001] This invention relates to the field of atmospheric remote sensing monitoring technology, specifically a method for retrieving regional methane emissions based on satellite column concentration observation. Background Technology

[0002] Methane (CH4) is the world's second-largest greenhouse gas, contributing 25% to global warming. Fossil fuel extraction and combustion, as well as land-use change, are the main causes of its increase.

[0003] Currently, multiple satellites have been launched globally to monitor changes in CH4 concentration. However, on the one hand, satellites cannot directly quantify CH4 emission intensity. On the other hand, quantifying emissions at a regional scale faces challenges in improving resolution and is susceptible to errors in CH4 concentration at regional boundaries. Therefore, a method for retrieving regional methane emissions based on satellite column concentration observations is proposed.

[0004] The information disclosed above in this background section is only for enhancing the understanding of the background technology of this invention, and therefore may include prior art that is not known to those skilled in the art. Summary of the Invention

[0005] This invention aims to address at least one of the technical problems existing in the prior art, such as the inability to directly quantify emission intensity and insufficient resolution and boundary errors at the regional scale, despite the deployment of multiple satellites globally to monitor CH4 concentrations. Therefore, one objective of this invention is to propose a method for inverting regional methane emissions based on satellite column concentration observations, which, while considering the influence of regional boundary CH4 concentration errors, enables satellite monitoring of changes in ground-based methane emissions at the regional scale.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A method for retrieving regional methane emissions based on satellite column concentration observations includes the following steps:

[0008] Step 1: Modify the CMAQ regional atmospheric chemical transport model by adding CH4 as a trace gas to simulate atmospheric transport processes. Modify the initial conditions and boundary conditions modules in CMAQ to enable ensemble simulation of CH4.

[0009] Step 2: Based on the CMAQ regional atmospheric chemical transport model and the EnKF inversion algorithm, construct a high-resolution nested emission inversion system;

[0010] Step 3: Using satellite CH4 column concentration data, correct the global CH4 concentration field simulated by the WACCM model, providing an optimized initial field and boundary field for CMAQ regional simulation, and reducing the impact of boundary errors on emission inversion;

[0011] Step 4: Obtain a prior set sample of CH4 emissions through Gaussian random perturbation, input it into the CMAQ atmospheric model to simulate CH4 concentration set, further match and assimilate satellite column concentration data at corresponding locations and times to obtain simulated CH4 column concentration set, combine the above data and their respective error information, input it into the inversion system to obtain optimized CH4 emissions;

[0012] Step 5: Using a two-step optimization strategy, the optimized emissions are re-input into the CMAQ model for forward simulation to generate the boundary field of the inner high-resolution region of the current assimilation window and the initial field for the optimization of the next assimilation window. At the same time, the optimized emissions are used as the prior emissions for the next window.

[0013] Step 6: Based on the optimized boundary field in Step 5, adopt a nested inversion strategy and repeat Step 5 to perform CH4 emission inversion in the inner high-resolution region. Finally, repeat Steps 5 and 6 to perform daily cyclic assimilation.

[0014] As a further optimization of the present invention, in step 2, the construction of the EnKF flux inversion system includes:

[0015] Step 201: EnKF uses perturbation prior emissions to represent its uncertainty and generates a set of prior emissions samples using the Monte Carlo method;

[0016] Step 202: Calculate the background error covariance. The formula is:

[0017]

[0018] in, Represents the average of the set of samples;

[0019] Step 203: Perform quality control on satellite observation data and calculate observation errors;

[0020] Step 204: Assimilation window and localization scheme. The assimilation window is set to 1 day, the set size is set to 50, the Gaspari-Cohn function is used as the localization function, the localization scale is set to 300 km, and optimization is performed only on emission grids that are significantly correlated with concentration (p>0.05).

[0021] Step 205: After obtaining the set of state vectors, CMAQ performs ensemble simulation so that these errors are propagated in the model through each ensemble sample of the state vectors.

[0022] As a further optimization of the present invention, in step 201, the formula used to generate the prior emission set sample is:

[0023] , i = 1, 2, ...,

[0024] Where b represents the background field, It is the identifier of the perturbation sample. It is the size of the set. This represents a randomly perturbed sample, which is added to the prior emissions. In order to generate a set of samples that are input into CMAQ. ; The perturbation of prior emissions in each grid is extracted based on a Gaussian distribution with a mean of zero and a standard deviation equal to the emission uncertainty.

[0025] As a further optimization of the present invention, in step 203, the satellite observation data is processed as follows:

[0026] Step 2030: Only data with qa_value > 0.5 is assimilated;

[0027] Step 2031: Resample the data to the regional scale model grid resolution, and perform weighted processing on multiple satellite observations under the same grid;

[0028] Step 2032: The observation error is estimated based on the accuracy given in the satellite data file.

[0029] As a further optimization of the present invention, in step 205, the observation vector is combined with... Optimized emissions Variance update by minimizing the analysis:

[0030]

[0031] in, As an observation operator, it transforms the model state variables from the model space to the observation space, enabling direct assimilation of satellite data. Represents the average value of the emission set sample. It is the background error covariance matrix and the observation error covariance estimated based on the emission set estimation. Determine the relative contribution to the final analysis field. The response of the uncertainty in simulated concentration to the uncertainty in emissions, within each assimilation window, As the optimal estimate of CH4 emissions, the observation operator serves as the interface for conversion between the assimilation system and satellite data;

[0032] The simulated column concentration data was calculated by weighting the CH4 concentration from the CMAQ model stratified simulation with pressure and interpolating it to the satellite's vertical stratification. :

[0033]

[0034]

[0035] in, For satellite search layer number, For satellite layers The average core of the column, This represents the prior CH4 column concentration. The satellite corresponding to the representative model simulation Layer mixing rate, and These represent the total dry air column and the dry air column of each layer in the simulation, respectively.

[0036] As a further optimization of the present invention, the specific steps for optimizing the initial field and boundary field in step 3 are as follows:

[0037] Step 301: First, convert the WACCM stratified simulation concentration into a column concentration comparable to satellite data according to the above formula;

[0038] Step 302: The quality-controlled satellite data often contains missing values ​​in space and time. To calculate the simulation bias of the grid where the missing values ​​are located, the simulated column concentration and the observed column concentration calculated in step 301 are smoothed radially, latitudinally, and temporally. The latitudinal average value and the average bias between simulation and observation are further calculated. Then, linear interpolation is used to fill in the missing values ​​in the radial direction. For the grid where the missing value is located in the same latitudinal direction, the bias is assumed to be consistent with the average bias of the non-missing value in that latitudinal direction. Finally, the original CH4 concentration simulated by WACCM is corrected grid by grid to provide optimized initial and boundary fields for regional scale inversion.

[0039] Compared with the prior art, the beneficial effects of the present invention are:

[0040] This invention presents a high-resolution regional methane emission retrieval system constructed through innovative algorithms and programs. This system can directly assimilate satellite column concentration data, significantly reducing the uncertainty in CH4 emission estimation and improving the accuracy of CH4 simulation. Furthermore, this system offers significant advantages in CH4 emission monitoring and assessment.

[0041] The above overview is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of the invention will become readily apparent from the accompanying drawings and the following detailed description. Attached Figure Description

[0042] Figure 1 This is a flowchart of the CH4 emission inversion system of the present invention;

[0043] Figure 2 This is a graph showing the changes in CH4 emissions and column concentrations before and after optimization in the nested inversion of the Shanxi region (9 km resolution) in the experimental example of this invention.

[0044] Figure 3 This is a comparison chart of the prior and posterior simulations and the time series observations from five ground stations in the experimental examples of this invention. Detailed Implementation

[0045] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] Example

[0047] This invention provides a technical solution: such as Figure 1 As shown, a method for retrieving regional methane emissions based on satellite column concentration observation includes the following steps:

[0048] First, a CH4 emission retrieval system was constructed based on the regional atmospheric chemical transport models CMAQ and EnKF. The CMAQ model adopts a highly modular design, which can be set with different resolutions as needed, and can be driven by different meteorological datasets and emission inventories for simulation.

[0049] During system operation, a nested inversion strategy was adopted. The resolution for the Chinese region was set to 27 km with a grid size of 219×159. The inner nested region, covering Shanxi and surrounding provinces, had a resolution of 9 km and a grid size of 195×174. The vertical direction was set to 50 layers, from the ground to the top of the atmosphere at 100 Pa. The initial and boundary meteorological fields used NCEP FNL global meteorological reanalysis data.

[0050] The CH4 boundary field outside the Chinese region adopts the global WACCM boundary field optimized by this technique, while the boundary field of the inner region is directly derived from the CH4 concentration field optimized for the Chinese region, thereby reducing the impact of boundary condition errors on the high-resolution emission inversion of the inner region. Except for the first window, the initial CH4 fields of other windows are all derived from the concentration fields generated by the posterior emission drive of the previous window.

[0051] In this embodiment, the prior flux set sample is obtained by perturbing the formula in step 201. The prior anthropogenic CH4 emissions are represented by the EDGAR v8 inventory with a resolution of 0.1°, wetland emissions by the WetCHARTs v1.3 dataset with a resolution of 0.5°, wildfire emissions by the GFED v4.1s dataset with a resolution of 0.25°, and soil uptake by simulation using the MeMo model with a resolution of 1°. All emission data are interpolated to a pattern-matching resolution.

[0052] After obtaining the prior emission set, it is input into the CMAQ model for ensemble simulation to obtain the response relationship between the simulated ensemble concentration and the ensemble emissions. Based on the calculated background error covariance matrix, simulated and observed CH4 column concentrations, and according to the formula in step 205, a "two-step optimization strategy" is adopted to assimilate and invert the CH4 emissions of China (outer layer) and the area surrounding Shanxi (inner layer) respectively.

[0053] Experimental Example

[0054] The assimilation system operated from December 2021 to January 2022, performing nested inversions of CH4 emissions from China and the surrounding Shanxi region daily within the same window, with December serving as the model initialization period. Prior uncertainty was set at 40%. The assimilated satellite observation data was the Tropomi CH4 column concentration product. After emission optimization, both prior and posterior emissions were re-input into the CMAQ model for CH4 concentration simulation. The assimilation effect was evaluated by comparing the results with satellite and ground observations.

[0055] Changes in CH4 emissions and column concentrations before and after optimization in the Chinese region (27 km resolution): Compared with prior emissions, posterior emissions increased in Northeast and Northwest China, while emissions decreased significantly in most parts of Central and Eastern China, especially the North China Plain, the Yangtze River Delta and the Pearl River Delta.

[0056] In January 2022, total CH4 emissions were 3.7 Tg, a decrease of 24.6% compared to the EDGAR inventory. Compared with Tropomi satellite observations, the CH4 column concentrations simulated using prior emissions were significantly overestimated by more than 40 ppb in some areas of central and eastern China. However, the column concentrations simulated using posterior emissions were more consistent with the satellite observations, significantly reducing the overestimation in central and eastern China.

[0057] Figure 2 The figure shows the changes in CH4 emissions and column concentrations before and after optimization in the nested inner Shanxi region (9 km resolution). In the figure, (a) prior emissions, (b) posterior emissions and (c) the difference between the two, (d) concentration observed by the TROPOMI satellite, (e) concentration simulated using prior emissions and (f) concentration simulated using posterior emissions.

[0058] Compared to prior emissions, emissions across almost the entire Shanxi province have decreased significantly. However, emissions increased in Jincheng and Yangquan, major coal-producing areas in Shanxi, indicating potential unrecorded sources in the EDGAR emissions inventory. Statistically, CH4 emissions in Shanxi's EDGAR inventory in January were 756 thousand tons, which decreased to 506 thousand tons after optimization, a reduction of 33.1%. After optimization, the simulated CH4 column concentration decreased from 1924.4 ppb to 1896.4 ppb, and the posterior simulation is more consistent with the TROPOMI observation results (1894.7 ppb), with an average deviation reduction of 94%.

[0059] Figure 3 The data compares the ground-level CH4 concentrations simulated using prior and posterior emission simulations with the observed concentrations at five stations over time. Except for the RYO station, located in the simulation boundary region, the posterior simulation performance at the other four stations is improved. Compared to the AMY, ULD, GSN, and YON stations, the average bias of the posterior simulation is reduced by 63.1%, 81.4%, 94.4%, and 66.4%, respectively. This indicates that the emission inventory uncertainty optimized using this system is significantly reduced and can effectively improve CH4 simulation.

[0060] All parts not described in this invention are the same as or can be implemented using existing technology. Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for inverting regional methane emissions based on satellite column concentration observations, characterized in that, The method comprises the following steps: Step 1, modifying the CMAQ regional atmospheric chemical transport model, adding CH4 species as a trace gas to simulate atmospheric transport process, modifying the initial condition module and the boundary condition module in CMAQ to enable the ensemble simulation of CH4; Step 2, based on the CMAQ regional atmospheric chemical transport model and the EnKF inversion algorithm, a high-resolution nested emission inversion system is constructed; Step 3, using satellite CH4 column concentration data, correcting the global CH4 concentration field simulated by the WACCM model, providing an optimized initial field and boundary field for the CMAQ regional simulation, and reducing the influence of boundary error on emission inversion; The specific steps for providing an optimized initial field and boundary field are as follows: Step 301, first, the WACCM layered simulation concentration is converted into a column concentration comparable to satellite data; wherein the simulated column concentration data is calculated by pressure-weighted weighting and interpolation to the vertical layering of the satellite according to the CH4 concentration simulated by the CMAQ model : ； ； where, retrieving the number of layers for the satellite, retrieving the number of layers for the satellite the column average kernel, representing the prior CH4 column concentration, representing the number of layers for the satellite corresponding to the model simulation the mixing ratio of the layer, and representing the total and the individual layer dry air column of the simulation, respectively; Step 302, after quality control, the satellite data has missing values in space and time, in order to calculate the simulation bias of the grid where the missing value is located, the simulated column concentration calculated in step 301 and the observed column concentration are radially, latitudinally and temporally smoothed, the latitudinal average value and the average bias of simulation and observation are further calculated, then linear interpolation is used to fill in the radial bias missing value, for the same latitudinal missing value grid, the bias assumption is consistent with the average bias of the non-missing value of the same latitudinal, and finally the original CH4 concentration simulated by WACCM is corrected grid by grid, so as to provide an optimized initial and boundary field for regional scale inversion; Step 4, by means of Gaussian random perturbation of prior CH4 emission, prior ensemble samples are obtained, which are input into the CMAQ atmospheric model for CH4 concentration ensemble simulation, further matching and assimilating satellite column concentration data at corresponding positions and times to obtain simulated CH4 column concentration ensemble, combining the above data and respective error information, inputting into the inversion system to obtain optimized CH4 emission; Step 5, using a two-step optimization strategy, the optimized emission is re-input into the CMAQ model for forward simulation to generate the boundary field of the inner high-resolution region in the current assimilation window and the optimized initial field of the next assimilation window, and at the same time, the optimized emission is used as the prior emission of the next window; Step 6, based on the optimized boundary field in step 5, a nested inversion strategy is adopted to repeat step 5 to perform CH4 emission inversion in the inner high-resolution region, and finally, steps 5 and 6 are repeated for day-by-day cycle assimilation.

2. The method of claim 1, wherein the method is based on satellite column concentration observations to invert regional methane emissions. In step 2, the construction of the high-resolution nested emission inversion system comprises: Step 201, EnKF represents the uncertainty by perturbing the prior emission, and generates ensemble samples of the prior emission by using the Monte Carlo method; Step 202, calculate the background error covariance, the formula is: ； wherein represents the average of the set of samples; Step 203, quality control of satellite observation data, calculate the observation error; Step 204, assimilation window and localization scheme, the assimilation window is set to 1 day, the ensemble size is set to 50, the Gaspari-Cohn function is used as the localization function, the localization scale is set to 300 km, and only the emission grids with significant correlation between the ensemble emission and the concentration are optimized; Step 205, after obtaining the ensemble of state vectors, CMAQ performs ensemble simulation to pass these errors in the model through each ensemble sample of the state vector.

3. The method of claim 2, wherein the method is based on satellite column concentration observations to invert regional methane emissions. In step 201, the formula used to generate the ensemble sample of prior emissions is: , i = 1, 2,..., M where b represents the background field, is the identifier of the perturbed sample, is the ensemble size, represents a randomly perturbed sample, the sample is added to the prior emissions to produce an ensemble sample input to CMAQ ; is the perturbation of the prior emissions in each grid according to a Gaussian distribution with mean zero and standard deviation of the emissions uncertainty.

4. The method of inverting regional methane emission based on satellite column concentration observations according to claim 2, characterized in that: In step 203, the satellite observation data is processed as follows: Step 2030, only data with qa_value > 0.5 is assimilated; Step 2031, the data is resampled to the regional scale model grid resolution, and multiple satellite observations under the same grid are weighted; Step 2032, the observation error is estimated according to the given accuracy of the satellite data file.

5. The method of claim 2, wherein: In step 205, the observation vector , optimized emissions Update by minimizing the analysis variance: ； where, The observation operator transforms the model state variables from the model space to the observation space, enabling the direct assimilation of satellite data, representing the average of the emission ensemble samples, is the background error covariance matrix estimated based on the emission ensemble, and the observation error covariance determines the relative contribution to the final analysis field, contains the response of the simulated concentration uncertainty to the emission uncertainty in each assimilation window, as the optimal estimate of CH4 emissions, the observation operator serves as the assimilation system's interface to satellite data; The simulated column concentration data is calculated by pressure-weighting and interpolating the CH4 concentration simulated by the CMAQ model in each layer to the vertical layers of the satellite : ； ； where, the number of layers for the satellite, the number of layers for the satellite the column average kernel, represents the prior CH4 column concentration, represents the number of layers for the satellite corresponding to the model simulation the mixing ratio of the layer, and represent the total and the individual layer dry air column of the simulation, respectively.

Images