Method, device, equipment and medium for optimizing greenhouse gas emission inventory data

Abstract

This application provides a method, apparatus, device, and medium for optimizing greenhouse gas emission inventory data. The method acquires emission inventory data, multi-source observation concentration data, and auxiliary data for a target greenhouse gas in a target area. The emission inventory data includes the emission amount of the target greenhouse gas in the spatiotemporal dimensions within the target area. Using an atmospheric transport model, gas concentration simulation is performed based on the emission inventory data to obtain simulated gas concentration data in the spatiotemporal dimensions. The emission inventory data, multi-source observation concentration data, and auxiliary data are spatiotemporally aligned and correlated with the simulated gas concentration data to obtain correlated data. The correlated data is input into a trained physical-guided neural network to obtain an optimization coefficient sequence. The emission inventory data is then optimized based on the optimization coefficient sequence to obtain optimized target emission inventory data. Embodiments of this application can improve the accuracy of the emission inventory.

Description

Technical Field

[0001] This application relates to the field of greenhouse gas emission monitoring technology, and more specifically, to a method, apparatus, equipment, and medium for optimizing greenhouse gas emission inventory data. Background Technology

[0002] Methane is one of the greenhouse gases that has a significant impact on global climate. Its greenhouse effect potential per unit mass far exceeds that of carbon dioxide, and it has a shorter residence time in the atmosphere. Precise emission reduction can quickly curb the trend of global warming. As global climate governance measures continue to upgrade, the accuracy and timeliness of greenhouse gas quantification have become the core bottleneck restricting the effectiveness of emission reduction.

[0003] Currently, methane emissions are characterized by complex sources and strong spatiotemporal heterogeneity. Traditional quantification methods suffer from insufficient observation coverage and large model simulation biases, making it difficult to meet the needs of atmospheric science research, precise environmental monitoring, and refined management in the energy industry. Therefore, developing efficient and accurate quantification technologies for methane and other greenhouse gases to provide scientific support for emission reduction policy formulation and emission source control has become a critical issue that urgently needs to be addressed in related fields. Summary of the Invention

[0004] In view of this, this application provides a method, apparatus, equipment and medium for optimizing greenhouse gas emission inventory data, so as to at least solve the problems existing in the related technologies.

[0005] Specifically, this application is implemented through the following technical solution: This application provides a method for optimizing greenhouse gas emission inventory data, including: Acquire emission inventory data, multi-source observation concentration data, and auxiliary data for the target greenhouse gas in the target area; the emission inventory data includes the emission amount of the target greenhouse gas in the spatiotemporal dimensions within the target area; Using an atmospheric transport model, gas concentration simulation is performed based on the emission inventory data to obtain simulated gas concentration data in the spatiotemporal dimension. The emission inventory data, the multi-source observation concentration data, and the auxiliary data are spatiotemporally aligned and correlated with the simulated gas concentration data to obtain correlated data. The associated data is input into a trained physical guidance neural network to obtain an optimization coefficient sequence, and the emission inventory data is optimized based on the optimization coefficient sequence to obtain optimized target emission inventory data.

[0006] This application also provides an apparatus for optimizing greenhouse gas emission inventory data, including: The data acquisition module is used to acquire emission inventory data, multi-source observation concentration data, and auxiliary data for the target greenhouse gas in the target area; the emission inventory data includes the emission amount of the target greenhouse gas in the spatiotemporal dimensions within the target area; The concentration simulation module is used to simulate gas concentrations based on the emission inventory data using an atmospheric transport model, and to obtain simulated gas concentration data in the spatiotemporal dimension. The data association module is used to perform spatiotemporal alignment and association between the emission inventory data, the multi-source observation concentration data, the auxiliary data and the simulated gas concentration data to obtain associated data. The data optimization module is used to input the associated data into a trained physical guided neural network to obtain an optimization coefficient sequence, and to optimize the emission inventory data based on the optimization coefficient sequence to obtain optimized target emission inventory data.

[0007] This application also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the method for optimizing greenhouse gas emission inventory data as described in any of the foregoing embodiments.

[0008] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method for optimizing greenhouse gas emission inventory data as described in any of the foregoing embodiments.

[0009] This application also provides a computer program product, including a computer program that, when run by a processor, performs the steps of any of the possible methods for optimizing greenhouse gas emission inventory data described above.

[0010] The technical solutions provided by the embodiments of this application may include the following beneficial effects: In this embodiment, emission inventory data, multi-source observation concentration data, and auxiliary data cover dimensions such as basic data, measured data, and environmental data, providing a wide range of data sources. Furthermore, an atmospheric transport model generates spatiotemporally consistent simulated gas concentration data, which is then spatiotemporally aligned and correlated to achieve a precise correspondence between emission inventory data, multi-source observation concentration data, auxiliary data, and simulated gas concentration data, generating correlated data. Finally, a physical-guided neural network is used, leveraging deep learning to uncover data correlation patterns while incorporating physical constraints to ensure emission rationality. Optimization coefficients are generated based on the correlated data, and these coefficients are used to optimize the emission inventory data. This improves the accuracy of the emission inventory data, providing a reliable data foundation for the formulation of emission reduction measures, precise control of emission sources, and climate science research.

[0011] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this specification. Attached Figure Description

[0012] Figure 1 This is a flowchart illustrating an exemplary embodiment of an emission inventory data optimization method. Figure 2 This is a flowchart illustrating the alignment and correlation of multi-source observation concentration data in an exemplary embodiment of this application; Figure 3 This is a flowchart illustrating another method for optimizing emission inventory data, as shown in an exemplary embodiment of this application; Figure 4 This is a flowchart illustrating an emission inventory data optimization process in an exemplary embodiment of this application; Figure 5 This is a schematic diagram of the structure of an emission inventory data optimization device shown in an exemplary embodiment of this application; Figure 6 This is a schematic diagram of the structure of another emission inventory data optimization device illustrated in an exemplary embodiment of this application; Figure 7 This is a hardware structure diagram of a computer device illustrated in an exemplary embodiment of this application. Detailed Implementation

[0013] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0014] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0015] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0016] Methane is one of the greenhouse gases that has a significant impact on global climate. Its greenhouse effect potential per unit mass far exceeds that of carbon dioxide, and it has a shorter residence time in the atmosphere. Precise emission reduction can quickly curb the trend of global warming. As global climate governance measures continue to upgrade, the accuracy and timeliness of greenhouse gas quantification have become the core bottleneck restricting the effectiveness of emission reduction.

[0017] Currently, methane emissions are characterized by complex sources and strong spatiotemporal heterogeneity. Traditional quantification methods suffer from insufficient observation coverage and large model simulation biases, making it difficult to meet the needs of atmospheric science research, precise environmental monitoring, and refined management in the energy industry. Therefore, developing efficient and accurate quantification technologies for methane and other greenhouse gases to provide scientific support for emission reduction policy formulation and emission source control has become a critical issue that urgently needs to be addressed in related fields.

[0018] Regional methane emission estimation primarily includes bottom-up emission inventory methods and top-down atmospheric inversion methods. Emission inventory methods rely heavily on yearbook emission data and empirical emission factors, summarizing estimates across different emission source types. However, they generally suffer from time lags, low spatial resolution, and difficulty in reflecting sudden or anomalous emissions. Furthermore, they often exhibit systematic underestimation uncertainties in complex emission areas related to oil and gas extraction, coal mining, and urban areas. Atmospheric inversion methods, on the other hand, are typically based on atmospheric chemical transport models, such as the global-scale GEOS-Chem model or the regional-scale WRF-CHEM model. They deduce regional methane emission flux by comparing simulated and observed concentrations. However, the accuracy of this method is highly dependent on the accuracy of the meteorological field simulation, the spatiotemporal representativeness of the observed concentrations, and the model parameter settings, leading to significant uncertainties in the inversion results.

[0019] Although progress has been made in methane emission inversion estimation, the following defects and limitations still exist in practical applications, which restrict the accuracy, timeliness and generalizability of regional methane emission inventories.

[0020] (1) The emission estimation results face reliability challenges in terms of accuracy and consistency.

[0021] There are often orders of magnitude differences between existing emission inventories and inversion results based on observational data. For example, existing emission databases generally underestimate methane emissions from open landfills. In oil and gas producing areas, emission estimates based on observations can be more than ten times the reported values ​​in some years. These differences reflect that existing emission inventories and inversion techniques are unable to accurately depict real emissions in complex emission scenarios, resulting in insufficient reliability of emission estimates.

[0022] (2) Insufficient ability to synergize observation data of different scales and types.

[0023] Top-down emission estimation results often use single-source data for inversion, such as relying solely on base station observation data or satellite observation concentration data. A unified and mature multi-source observation concentration data fusion framework has not yet been formed. Due to the significant differences in spatial coverage, temporal resolution, and vertical sensitivity among different observation methods, the lack of effective fusion methods will directly lead to low utilization of observation information and make it difficult to fully leverage the complementary advantages of multi-source data.

[0024] (3) The computational efficiency of the inversion method based on the physical model is low, making it difficult to support high-resolution applications.

[0025] Top-down methane emission estimation methods rely heavily on atmospheric chemical transport models. Under conditions of high spatiotemporal resolution at the regional scale (e.g., hourly or kilometer-scale simulations), they often require a large amount of computational resources and long running time. This shortcoming is particularly prominent when frequent updates to emission estimation results or multi-scenario analysis are needed, which seriously restricts the feasibility of inversion methods in operational and real-time applications.

[0026] (4) Insufficient ability to propagate model errors and quantify uncertainty.

[0027] The accuracy of the relevant inversion method depends to some extent on the simulation accuracy of the atmospheric transport model. However, such models inevitably have systematic errors in meteorological fields, boundary fields, and chemical mechanisms. These errors are propagated and amplified step by step during the inversion process, and are difficult to effectively correct and uniformly characterize. This results in insufficient assessment of the uncertainty of emission results and makes it difficult to provide a reliable reference for the error range.

[0028] Based on the above research, this disclosure provides a method for optimizing greenhouse gas emission inventory data. The method first acquires emission inventory data, multi-source observation concentration data, and auxiliary data for a target greenhouse gas in a target region. The emission inventory data includes the emissions of the target greenhouse gas in the spatiotemporal dimensions within the target region. Then, using an atmospheric transport model, gas concentration simulation is performed based on the emission inventory data to obtain simulated gas concentration data in the spatiotemporal dimensions. The emission inventory data, the multi-source observation concentration data, and the auxiliary data are spatiotemporally aligned and correlated with the simulated gas concentration data to obtain correlated data. Finally, the correlated data is input into a trained physical-guided neural network to obtain an optimization coefficient sequence, and the emission inventory data is optimized based on the optimization coefficient sequence to obtain optimized target emission inventory data.

[0029] In this embodiment, emission inventory data, multi-source observation concentration data, and auxiliary data cover dimensions such as basic data, measured data, and environmental data, providing a wide range of data sources. Furthermore, an atmospheric transport model generates spatiotemporally consistent simulated gas concentration data, which is then spatiotemporally aligned and correlated to achieve a precise correspondence between emission inventory data, multi-source observation concentration data, auxiliary data, and simulated gas concentration data, generating correlated data. Finally, a physical-guided neural network is used, leveraging deep learning to uncover data correlation patterns while incorporating physical constraints to ensure emission rationality. Optimization coefficients are generated based on the correlated data, and these coefficients are used to optimize the emission inventory data. This improves the accuracy of the emission inventory data, providing a reliable data foundation for the formulation of emission reduction measures, precise control of emission sources, and climate science research.

[0030] To facilitate understanding of this embodiment, a detailed description of the method for optimizing greenhouse gas emission inventory data disclosed in this disclosure is provided first. The execution entity of the emission inventory data optimization method provided in this disclosure is generally a computer device. This computer device can be a server, which can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing initial cloud computing services such as cloud services, cloud databases, cloud computing, cloud storage, big data, and artificial intelligence platforms. In other embodiments, the computer device can also be a terminal device, which can be a mobile device, terminal, handheld device, computing device, vehicle-mounted device, etc.

[0031] In other embodiments, the method can also be applied to an implementation environment consisting of computer equipment and servers, or an implementation environment consisting of terminal equipment and servers. Furthermore, the method for optimizing emission inventory data can also be implemented by a processor calling computer-readable instructions stored in memory.

[0032] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0033] Please see the appendix Figure 1 This is a flowchart illustrating an exemplary embodiment of the present application of a method for optimizing greenhouse gas emission inventory data. Figure 1 As shown, the method for optimizing greenhouse gas emission inventory data in this embodiment of the present disclosure may include the following steps S101~S104: S101: Obtain emission inventory data, multi-source observation concentration data, and auxiliary data for the target greenhouse gas in the target area; the emission inventory data includes the emission amount of the target greenhouse gas in the spatiotemporal dimensions within the target area.

[0034] The target area refers to the specific spatial range for quantifying and optimizing target greenhouse gas emissions. It can be set according to actual needs. Specifically, it can be divided according to urban agglomerations (such as provincial-level regions, municipal-level regions, district-level regions, etc.) or river basins (such as the Yangtze River, the Yellow River and its branches, etc.), without any restrictions here.

[0035] In this embodiment, the target greenhouse gas includes methane gas. In other embodiments, the target greenhouse gas may be other greenhouse gases, such as carbon dioxide, nitrous oxide, Freon, etc., and is not limited thereto.

[0036] In this embodiment, the emission inventory data includes the emission amount of the target greenhouse gas in the spatiotemporal dimension within the target area. Specifically, the target area is divided into multiple two-dimensional grids, and the emission inventory data includes the emission amount of the target greenhouse gas corresponding to each of the multiple two-dimensional grids.

[0037] Optionally, the emission inventory data can be obtained by processing the original emission inventory product. The original emission inventory data is obtained from the Emissions Database for Global Atmospheric Research (EDGAR), and optimized based on activity data (such as energy consumption and industrial production intensity) of the target region. The optimized original emission inventory data is then spatially allocated using spatial proxy factors and time-based on hourly intervals.

[0038] The multi-source observation concentration data refers to the actual concentration data of the target greenhouse gas obtained through multiple observation methods, including satellite remote sensing observation, mobile platform (such as UAV, mobile monitoring vehicle) observation, and ground base station fixed-point observation. The multi-source observation concentration data includes satellite observation concentration data (column concentration of the target greenhouse gas corresponding to each satellite observation pixel), mobile observation concentration data, and base station observation concentration data.

[0039] Here, the concentration data observed by satellite can be obtained from one satellite or from multiple satellites; no limitation is made here.

[0040] The auxiliary data refers to the auxiliary data that affects the accuracy of gas concentration observation and transmission and diffusion, including meteorological data (wind speed, temperature, humidity, boundary layer height), surface data (such as land use type, elevation, surface roughness), and human activity data (nighttime light, POI density), etc. Here, meteorological data refers to the data obtained through observation.

[0041] The spatiotemporal dimensions include a time dimension and a spatial dimension. The time dimension can be a time interval, such as an hour. The spatial dimension includes a grid dimension and height. In this embodiment, the target area is divided into multiple two-dimensional grids according to a preset resolution. The preset resolution can be set according to actual needs (such as the area range of the target area). For example, if the target area is a provincial area, it is divided into grids of 3km×3km as the dividing unit. If the target area is a municipal area, it is divided into grids of 1km×1km as the dividing unit.

[0042] In some implementations, multi-source observation concentration data can be generated through the following steps: The raw multi-source observation concentration data is obtained, which includes raw satellite observation concentration data, raw mobile observation concentration data, and raw base station observation concentration data. The raw satellite observation concentration data, raw mobile observation concentration data, and raw base station observation concentration data are filtered and preprocessed respectively to obtain satellite observation concentration data, mobile observation data, and base station observation concentration data.

[0043] The data processing procedures for satellite-observed concentration data, mobile-observed concentration data, and base station-observed concentration data are described below.

[0044] (1) For the original satellite observation concentration data: The original satellite observation concentration data is in the form of satellite cloud image. It can be filtered according to the Quality Assessment (QA) band it carries. The Quality Assessment band is used to indicate the valid data (valid pixels) in the original satellite observation concentration data, and then invalid pixels are deleted to filter the original satellite observation concentration data.

[0045] (2) For the original mobile observation concentration data: the standard deviation of the original mobile concentration data can be determined, and the screening range can be determined based on the standard deviation. The original mobile observation data can be screened based on the screening range to obtain the mobile observation concentration data.

[0046] In this embodiment, the screening range can refer to a multiple of the standard deviation, such as three times the standard deviation. In other embodiments, it can also be other multiples, which are not limited here.

[0047] It is easy to understand that the original moving concentration data includes the concentration data (concentration, latitude and longitude, altitude and time) of each trajectory point of the observation trajectory. In this way, the mean and standard deviation can be determined based on the concentration data of each trajectory point, the screening range can be determined, and the original moving observation data can be filtered based on the screening range to obtain the moving observation concentration data. For example, the concentration data of each trajectory point can be verified. If the concentration data of any trajectory point is not within the screening range, it can be deleted to obtain the moving observation concentration data.

[0048] (3) For base station observation concentration data: using a preset background extraction algorithm, the enhanced concentration value is removed from the original base station observation concentration data to obtain the base station observation concentration data.

[0049] The enhanced concentration value refers to the concentration data of the target greenhouse gas emitted from other areas that is observed by base stations within the target area.

[0050] It is easy to understand that the target greenhouse gas emitted in other areas is affected by factors such as emission intensity and wind direction, which will affect the concentration monitoring of the target area. That is, the original base station observed concentration data refers to the data observed by the base station in the target area. It is affected by gas emissions in the target area and may also be affected by gas emissions in other areas. In this embodiment, a background extraction algorithm is used to separate and remove the enhanced concentration caused by emissions from other areas (upwind areas). For example, if the concentration value of the original base station observed concentration data is 100 ppm, and the enhanced concentration value is 10 ppm, then the base station observed concentration data is 90 ppm, so as to reduce background interference in long-distance transmission.

[0051] In this embodiment, by processing the original multi-source observation concentration data, invalid or abnormal data can be removed, which helps to improve the accuracy of the multi-source observation concentration data.

[0052] S102: Using an atmospheric transport model, gas concentration simulation is performed based on the emission inventory data to obtain simulated gas concentration data in the spatiotemporal dimension.

[0053] Among them, the atmospheric transport model refers to the Weather Research and Forecasting Model coupled with Chemistry (WRF-CHEM) model.

[0054] Specifically, for step S102, meteorological driving data and initial boundary data can be acquired, and the meteorological driving data, emission inventory data and boundary data can be input into the atmospheric transport model to simulate gas concentration and obtain simulated gas concentration data in the spatiotemporal dimensions.

[0055] The meteorological driving data is obtained from the GADS / FNL dataset, including key atmospheric parameters from the ground to the upper atmosphere. Specifically, these include surface and sea level pressure, 2m temperature and relative humidity, 10m wind field, geopotential height, temperature, wind field and vertical height at each standard pressure level, soil temperature and humidity, sea surface temperature, ice cover, and downward shortwave radiation flux.

[0056] The initial boundary data is used to provide background concentrations, which are obtained based on the output of a global chemical transport model (such as CAMS).

[0057] In this step, the diffusion, transport, and chemical reaction processes of the target greenhouse gas in the atmosphere are simulated based on the atmospheric transport model combined with emission inventory data, meteorological driving data, and initial boundary data. Finally, the model outputs simulated gas concentration data with the same spatiotemporal dimensions as the emission inventory data. The simulated gas concentration data includes horizontal simulated concentration data and vertical stratified concentration data corresponding to multiple two-dimensional grids at different times.

[0058] S103: The emission inventory data, the multi-source observation concentration data, and the auxiliary data are spatiotemporally aligned and correlated with the simulated gas concentration data to obtain correlated data.

[0059] This step aims to unify the spatiotemporal benchmarks of various types of data, thereby improving the correlation between different types of data.

[0060] Specifically, multi-source observed concentration data corresponding to the same time interval can be matched and correlated with the horizontal simulated concentration data of each two-dimensional grid, and emission inventory data and auxiliary data corresponding to the same time interval can be matched and correlated with the horizontal simulated concentration data of each two-dimensional grid to obtain correlated data.

[0061] For example, the associated data can be represented as multidimensional data {emissions, multi-source observation concentration data, auxiliary data, and simulated gas concentration data} within the same two-dimensional grid over the same time interval.

[0062] For multi-source observation concentration data: For the same time interval, satellite observation concentration data, mobile observation concentration data, and base station observation concentration data can be mapped to various two-dimensional grids in the target area respectively. If the observation data is single-point or strip data, it can be matched to the corresponding two-dimensional grid based on interpolation.

[0063] For auxiliary data: For the same time interval, grid registration is performed on auxiliary data such as meteorological data (wind speed, temperature) and topographic data. Static environmental data (such as topography) can be directly mapped to a two-dimensional grid, while dynamic environmental data (such as hourly weather) are synchronized according to time nodes and matched to the corresponding two-dimensional grid, so that each two-dimensional grid can be further associated with environmental feature parameters at that time node.

[0064] It should be noted that for observational data with altitude (such as concentration data observed by drones or meteorological data such as wind speed), vertical alignment is also required. The specific details will be introduced later.

[0065] In this way, with "time and two-dimensional grid" as the core index, the simulated gas concentration data, multi-source observation concentration data and auxiliary data of each two-dimensional grid at the corresponding time are integrated to form structured correlation data, ensuring that a single correlation data can completely reflect the three-dimensional information of "simulation-observation-environment" of a certain spatiotemporal grid.

[0066] The following section details the spatiotemporal alignment and correlation of auxiliary data and simulated gas concentration data.

[0067] For auxiliary data, alignment is performed mainly in the time dimension, grid dimension, and height dimension, and then correlated with the simulated gas concentration data.

[0068] Alignment in terms of the time dimension refers to aligning the time resolution of the auxiliary data with that of the simulated gas concentration data. For example, if the auxiliary data is updated hourly and the simulated gas concentration data has a 6-hour resolution, then the auxiliary data is aggregated by averaging over a 6-hour period. If the auxiliary data is updated daily, then interpolation can be used to supplement the data to a 6-hour time scale, thereby ensuring that the data in the same time interval correspond.

[0069] For grid alignment: Alignment can be performed based on the type of auxiliary data. For example, for surface data, which is static data corresponding to the target area, it can be directly mapped to each two-dimensional grid. However, for meteorological driven data, it is necessary to map it to the resolution of the two-dimensional grid based on the resolution of its data sampling.

[0070] Alignment for height dimension: For auxiliary data with height information (or vertical dimension) (such as wind speed, temperature, etc. at different heights), it is necessary to match and associate the auxiliary data with the vertically layered data according to the height of the auxiliary data.

[0071] S104: Input the associated data into the trained physical guidance neural network to obtain an optimization coefficient sequence, and optimize the emission inventory data based on the optimization coefficient sequence to obtain the optimized target emission inventory data.

[0072] Among them, the Physics-Guided Neural Network (PGNN) combines a graph neural network (GNN) with a physical constraint layer.

[0073] In this step, the trained physical guidance neural network integrates physical constraints and deep learning. It uses the physical guidance neural network to mine the correlation patterns between emission inventory data, multi-source observation concentration data, simulated observation concentration data and auxiliary data in the associated data, and outputs an optimization coefficient sequence. This allows the emission inventory data to be optimized based on the optimization coefficient sequence, resulting in optimized target emission inventory data and achieving accurate optimization of the emission inventory data.

[0074] Specifically, the trained physical-guided neural network uses the deviation between simulated gas concentration data and observed concentration data as its learning target. Under the guidance of the model's physical constraints (auxiliary variables), it learns the spatial distribution characteristics of emission errors. That is, the model can learn the pattern under what meteorological and geographical conditions the magnitude of the difference between observed and simulated data means what adjustments need to be made to the emission inventory data, thus reflecting the complex mapping relationship from the associated data to the optimization coefficient sequence. Based on the above method, it can effectively distinguish between concentration differences caused by emission deviations and concentration changes caused by the observation environment (such as meteorology and surface conditions), thereby improving the stability and physical rationality of emission optimization.

[0075] Finally, based on the optimization coefficient sequence, the emission inventory data is optimized. During the optimization process, satellite observation concentration data, due to its wide spatial coverage and strong spatial continuity, can constrain the spatial distribution structure of emissions. Mobile observation concentration data and base station observation concentration data, due to their sensitivity to near-surface concentration changes, are used to correct the total emissions of the target area or local scale. Through the synergistic constraints of multi-source observation concentration data on spatial distribution and emission intensity, the emission inventory data of the target area is refined and scaled.

[0076] In this embodiment of the application, before inputting the associated data into the trained physical guidance neural network, different label weights can be configured for the multi-source observation concentration data in the associated data, wherein the label weights are used to distinguish the data source and credibility.

[0077] Specifically, the identification value of satellite observation concentration data is 99, the identification value of UAV observation concentration data is 9, the identification value of mobile observation concentration data is 2, and the identification value of base station observation concentration data is 1, thereby guiding the physical guidance neural network to learn the relative contribution of different types of multi-source observation concentration data to the optimization coefficient sequence.

[0078] In this embodiment, the optimization coefficient sequence includes the optimization coefficients corresponding to each two-dimensional grid.

[0079] For step S104, the emissions of the corresponding grids in the emission inventory data can be optimized based on the optimization coefficients corresponding to each two-dimensional grid to obtain the optimized target emission inventory data.

[0080] Specifically, when optimizing the emissions of the corresponding two-dimensional grids in the emission inventory data based on the optimization coefficients corresponding to each two-dimensional grid, the target emissions can be determined for each two-dimensional grid based on the product of the optimization coefficient and the emissions. Then, the optimized target emission inventory data can be generated based on the target emissions of each two-dimensional grid.

[0081] In this embodiment, emission inventory data, multi-source observation concentration data, and auxiliary data cover dimensions such as basic data, measured data, and environmental data, providing a wide range of data sources. Furthermore, an atmospheric transport model generates spatiotemporally consistent simulated gas concentration data, which is then spatiotemporally aligned and correlated to achieve a precise correspondence between emission inventory data, multi-source observation concentration data, auxiliary data, and simulated gas concentration data, generating correlated data. Finally, a physical-guided neural network is used, leveraging deep learning to uncover data correlation patterns while incorporating physical constraints to ensure emission rationality. Optimization coefficients are generated based on the correlated data, and these coefficients are used to optimize the emission inventory data. This improves the accuracy of the emission inventory data, providing a reliable data foundation for the formulation of emission reduction measures, precise control of emission sources, and climate science research.

[0082] In some implementations, when aligning and associating the multi-source observed concentration data corresponding to the same time interval with the horizontal simulated concentration data of each of the two-dimensional grids, please refer to [link to relevant documentation]. Figure 2 The flowchart illustrating the alignment and correlation of multi-source observation concentration data in an exemplary embodiment of this application includes the following steps S201-S203: S201: For any time interval, based on the first correspondence between the satellite observation pixels corresponding to the satellite observation concentration data and the two-dimensional grid, the satellite observation concentration data is allocated to the corresponding two-dimensional grid.

[0083] In this context, a satellite observation pixel refers to the smallest unit of observation of the Earth's surface by a satellite. For example, the observation pixel resolution of a certain satellite is 5km × 5km. Of course, the observation pixel resolution may vary between different satellites.

[0084] It is understandable that, since the spatial scale (resolution) of satellite observation pixels and two-dimensional grids is usually inconsistent (the resolution of satellite observation pixels is usually larger than that of two-dimensional grids), satellite observation concentration data is split or aggregated into two-dimensional grids based on the first correspondence, thereby allocating satellite observation concentration data to the corresponding two-dimensional grids.

[0085] The first correspondence includes the ratio between the distribution area of ​​satellite observation pixels in their respective grids and the area of ​​the observation pixels.

[0086] In some implementations, step S201 may include the following steps (1) to (2): (1) For each satellite observation pixel, the distribution weight of the column concentration corresponding to the satellite observation pixel in each corresponding two-dimensional grid is determined according to the proportional relationship between the distribution area of ​​the satellite observation pixel in each corresponding two-dimensional grid and the area of ​​the satellite observation pixel.

[0087] As can be understood from the preceding text, since satellite observation pixels are usually larger than two-dimensional grids, the ratio between the distribution area of ​​each satellite observation pixel in its corresponding two-dimensional grid and the area of ​​the satellite observation pixel can be determined. This ratio is the weight of the column concentration corresponding to the satellite observation pixel in the corresponding two-dimensional grid.

[0088] For example, for satellite observation pixel A1, its area is 25 km². 2 (square kilometers), corresponding to two-dimensional grids B1, B2, B3, and B4. The distribution area of ​​satellite observation pixel A1 within two-dimensional grids B1, B2, B3, and B4 is 2.5 km². 2 5km 2 7.5 km 2 10 km 2 Therefore, the proportional relationships between the area of ​​each distribution and the area of ​​A1 can be determined to be 0.1, 0.2, 0.3, and 0.4, respectively. Then, the assigned weights of each two-dimensional grid B1, B2, B3, and B4 are 0.1, 0.2, 0.3, and 0.4.

[0089] (2) For each two-dimensional grid, determine and allocate the satellite observation concentration of the two-dimensional grid by assigning the column concentration of each satellite observation pixel corresponding to the two-dimensional grid and the column concentration of each satellite observation pixel within the grid.

[0090] In this step, for each two-dimensional grid, the weighted sum of the column concentrations of each satellite observation pixel covering the two-dimensional grid and the corresponding assigned weights is obtained to obtain the satellite observation concentration of the two-dimensional grid.

[0091] For example, if a two-dimensional grid B1 is simultaneously covered by two satellite observation pixels A1 and A2, with pixel A1 having a column density of 100 and a weight of 0.1, and pixel A2 having a column density of 200 and a weight of 0.5, then the satellite observation density of the two-dimensional grid B1 is... .

[0092] In this embodiment, the area ratio is used as the allocation weight to reflect the contribution ratio of satellite observation signals in different grids, avoiding the spatial deviation caused by "directly assigning satellite observation concentration data to a single grid" and realizing the accurate allocation of coarse-resolution satellite observation concentration data to two-dimensional grids.

[0093] S202: For any time interval, based on the second correspondence between the moving trajectory corresponding to the moving observation concentration data and the two-dimensional grid, the moving observation concentration data is matched and associated with the simulated gas concentration of the corresponding two-dimensional grid.

[0094] It is easy to understand that mobile observation concentration data is obtained through mobile observation platforms (such as drones and mobile monitoring vehicles). Therefore, the mobile trajectory refers to the travel path of the mobile observation platform.

[0095] Here, the second correspondence refers to the spatial mapping relationship between the movement trajectory and the two-dimensional grid (for example, which two-dimensional grid each trajectory point falls into). In this way, the movement observation concentration data can be matched and associated with the simulated gas concentration of the corresponding two-dimensional grid according to the second correspondence.

[0096] As mentioned above, mobile observation concentration data includes mobile observation data and UAV observation concentration data. It should be noted that since mobile observation data is usually near the ground, it is only necessary to associate the observation concentration data corresponding to each trajectory point on the mobile trajectory with each two-dimensional grid. However, for UAV observation concentration data, the altitude of the UAV will affect the observation concentration. Therefore, for UAV observation concentration data, the matching of the UAV altitude needs to be considered.

[0097] For mobile observation data, the following steps (i) to (ii) may be included: (i) For each two-dimensional grid, determine at least one target trajectory point covered by the two-dimensional grid based on the two-dimensional coordinates of each trajectory point.

[0098] Here, for each two-dimensional grid, by matching the two-dimensional coordinates of each trajectory point, at least one target trajectory point falling within the two-dimensional grid can be determined. For example, if the longitude range of a certain two-dimensional grid is 116.2°-116.3° and the latitude range is 39.9°-40.0°, then all trajectory points whose coordinates are within this range are target trajectory points.

[0099] (ii) Based on the observed concentration data of the at least one target trajectory point, determine the moving observed concentration, and match and associate the moving observed concentration data with the simulated gas concentration of the two-dimensional grid.

[0100] For the observed concentration data of at least one target trajectory point, a statistical method (such as calculating the mean) is used to determine the moving observed concentration, and then the moving observed concentration data is matched and correlated with the simulated gas concentration of the corresponding two-dimensional grid.

[0101] In this embodiment, mobile observation is a "point-based, continuous" sampling method. By filtering target trajectory points and aggregating concentrations, discrete point data is transformed into grid-scale observations, solving the spatial matching problem of "points not corresponding to grids" in mobile observation. This transforms the fine observations of the mobile observation platform into two-dimensional grid-scale data, which not only adapts to the spatial reference of the model but also retains the high-resolution value of mobile observation.

[0102] For concentration data observed by UAVs, the following steps (a) to (b) may be included: (a) Based on the three-dimensional coordinates corresponding to the concentration data observed by the UAV, determine multiple target two-dimensional grids corresponding to the three-dimensional coordinates.

[0103] It should be understood that the concentration data observed by the UAV includes the observed concentration corresponding to each trajectory point on the movement trajectory of the UAV. Each trajectory point has corresponding three-dimensional coordinates. In this way, multiple target two-dimensional grids associated with the concentration data observed by the UAV can be determined.

[0104] (b) For each target two-dimensional grid, determine the target trajectory point corresponding to the target two-dimensional grid.

[0105] For each target 2D grid, the target trajectory point corresponding to the target 2D grid can be determined based on the horizontal coordinate of the target 2D grid. For example, the trajectory point can be projected in the horizontal direction, so that each trajectory point can be divided into the corresponding target 2D grid.

[0106] (c) Based on the observed concentration of each target trajectory point, the interpolation method is used to determine the target UAV observed concentration data corresponding to the target two-dimensional grid, so as to match and associate the target UAV observed concentration data with the simulated gas concentration of the target two-dimensional grid.

[0107] It is understandable that the observed concentrations corresponding to each trajectory point are discrete point data, and they are characterized by being scattered in horizontal position and having diverse flight altitudes. Therefore, they cannot directly represent the concentration level of the entire two-dimensional grid. Thus, it is necessary to standardize the spatial dimension of the trajectory point data through interpolation.

[0108] Specifically, it may include the following steps (c1) to (c4): (c1) Based on the observed concentration of each target trajectory point, the horizontal interpolation concentration of the center point of the target two-dimensional grid is determined by linear interpolation.

[0109] Here, the center point of the target two-dimensional grid is taken as the interpolation target position. Based on the observed concentration of target trajectory points within and near the target two-dimensional grid, a linear interpolation method is used to determine the horizontal interpolation concentration of the center point of the target two-dimensional grid, thereby eliminating the spatial deviation caused by the uneven horizontal distribution of UAV trajectory points.

[0110] (c2) Obtain the UAV vertical layer data, and determine the height of the model near the ground layer, the height of the UAV upper layer, and the height of the UAV lower layer from the UAV vertical layer data.

[0111] Here, the drone vertical stratification data includes the stratification height of the drone's flight.

[0112] The height of the model's near-surface layer refers to the target height for vertical interpolation in subsequent steps. The height of the upper and lower layers of the UAV determines the height range for vertical interpolation, thereby ensuring that the observed concentration corresponding to the model's near-surface layer is interpolated.

[0113] (c3) From the observation concentrations of each target trajectory point, determine the first observation concentration corresponding to the upper layer height of the UAV and the second observation concentration corresponding to the lower layer height of the UAV. Based on the first observation concentration and the second observation concentration, use the interpolation method to determine the vertical interpolation concentration corresponding to the near-ground layer height of the model.

[0114] In this step, in order to solve the problem of the different heights of UAVs and the mismatch with the near-ground reference height, the observation data is aligned with the vertical reference of the target two-dimensional grid.

[0115] First, from the observed concentrations of the target trajectory points, the first observed concentration corresponding to the upper layer height of the UAV and the second observed concentration corresponding to the lower layer height of the UAV are selected. Using the interpolation method, based on the difference between the upper and lower layer heights and the difference between the corresponding observed concentrations, the vertical interpolated concentration corresponding to the height of the model near the ground layer is determined, thereby eliminating the concentration deviation caused by vertical height.

[0116] (c4) Based on the horizontal interpolation concentration of the center point of the target two-dimensional grid and the vertical interpolation concentration, determine the target UAV observation concentration data corresponding to the target two-dimensional grid.

[0117] By fusing the horizontal and vertical interpolation concentrations of the target two-dimensional grid center point, a standardized target observation concentration that simultaneously satisfies the requirements of "horizontal anchoring of the grid center point and vertical anchoring of the near-surface layer" can be obtained.

[0118] In this way, by standardizing the concentration data observed by UAVs in two dimensions and by using step-by-step interpolation in both horizontal and vertical directions, the spatial deviation between the discrete observations by UAVs and the model grid reference is eliminated, and the accurate alignment of the observed data and the simulated data is achieved, which is conducive to improving the scientificity and reliability of subsequent emission inventory optimization.

[0119] S203: For any time interval, based on the third correspondence between the base station observation location corresponding to the base station observation concentration data and the two-dimensional grid, the base station observation concentration data is allocated to the corresponding two-dimensional grid.

[0120] Among them, the concentration data observed by the base station refers to the data observed by the base station with a fixed observation location.

[0121] The third correspondence refers to the spatial association between the base station's observation location (latitude and longitude coordinates) and the two-dimensional grid, determining the two-dimensional grid where the base station is located (if the base station's coordinates fall within a certain grid, then that grid is the target grid corresponding to the base station). In this way, for any time interval node, the base station's observed concentration data obtained at that time is directly associated with the corresponding two-dimensional grid.

[0122] In this implementation, based on the principle of "time synchronization and spatial correspondence", heterogeneous observation data from satellites, mobile devices, and base stations are transformed into unified data at the grid scale through three different types of spatial mapping relationships, laying the foundation for subsequent matching with simulated gas concentrations and optimization of emission inventories.

[0123] Please see the appendix Figure 3 This is a flowchart illustrating another method for optimizing greenhouse gas emission inventory data, as shown in an exemplary embodiment of this application. Figure 3 As shown, the method for optimizing greenhouse gas emission inventory data in this embodiment may further include the following steps S105-S106: S105: Obtain the uncertainty assessment index values ​​corresponding to the emission inventory data, the multi-source observation concentration data, and the physical guided neural network, respectively.

[0124] Among them, the uncertainty assessment index value is used to quantify the reliability of data or models (such as standard deviation, relative error, confidence interval, etc.). The smaller the value, the higher the credibility of the data or model.

[0125] Uncertainty index values ​​for emission inventory data are used to characterize the inherent biases of emission inventory data, which originate from errors in activity data, emission factor errors, spatiotemporal allocation errors, etc.

[0126] Uncertainty indices for multi-source observation concentration data include indices that quantify the biases of observation data from satellites, mobile devices, and base stations, covering sensor accuracy errors, inversion errors, representativeness errors, etc.

[0127] The uncertainty index of a physically guided neural network is used to reflect the bias of the model itself, which comes from network structure limitations, training data randomness, physical constraint approximation error, etc.

[0128] S106: Based on the uncertainty assessment index values ​​corresponding to the emission inventory data, the multi-source observation concentration data, and the physical guided neural network, respectively, determine the uncertainty assessment index value of the target emission inventory data; the uncertainty assessment index value is used to characterize the confidence level of the target emission inventory data.

[0129] Specifically, we can assume that the three types of uncertainty assessment index values ​​are independent of each other, and use the first-order error propagation formula to fuse the uncertainty index values ​​of emission inventory data, multi-source observation concentration data and physical guided neural network to obtain the comprehensive uncertainty index value of the target emission inventory data (such as the posterior emission standard deviation).

[0130] In this embodiment, the confidence level of the target emission inventory data is quantified: the indicator values ​​(such as confidence interval width and standard deviation) directly reflect the reliability range of the optimized inventory and provide a quality reference for data use.

[0131] Furthermore, by breaking down the contribution ratio of the three types of indicators, the main sources of uncertainty can be identified (e.g., if the proportion of model error is too high, the network structure needs to be optimized).

[0132] Please see Figure 4 This is a flowchart illustrating an emission inventory data optimization process, as shown in an exemplary embodiment of this application. Figure 4 As shown, meteorological driving data, initial boundary data, and emission inventory data are input into the atmospheric transport model. The atmospheric transport model outputs simulated gas concentration data. The emission inventory data, simulated gas concentration data, multi-source observation concentration data, and auxiliary data are matched and correlated, and then input into a trained physical guidance neural network to obtain an optimization coefficient sequence. Based on the optimization coefficient sequence, the gas emissions in the emission inventory data are optimized to obtain the optimized target emission inventory data.

[0133] Corresponding to the embodiments of the aforementioned emission inventory data optimization method, this application also provides embodiments of an emission inventory data optimization apparatus.

[0134] Please see Figure 5 This is a schematic diagram illustrating the structure of a greenhouse gas emission inventory data optimization device, as shown in an exemplary embodiment of this application. Figure 5 As shown, the greenhouse gas emission inventory data optimization device 500 includes: The data acquisition module 510 is used to acquire emission inventory data of the target greenhouse gas for the target area, multi-source observation concentration data, and auxiliary data; the emission inventory data includes the emission amount of the target greenhouse gas in the spatiotemporal dimensions within the target area; The concentration simulation module 520 is used to simulate gas concentration based on the emission inventory data using an atmospheric transport model to obtain simulated gas concentration data in the spatiotemporal dimension. The data association module 530 is used to perform spatiotemporal alignment and association between the emission inventory data, the multi-source observation concentration data, the auxiliary data and the simulated gas concentration data to obtain associated data. The data optimization module 540 is used to input the associated data into a trained physical guided neural network to obtain an optimization coefficient sequence, and optimize the emission inventory data based on the optimization coefficient sequence to obtain optimized target emission inventory data.

[0135] In some embodiments, the target area is divided into multiple two-dimensional grids, and the simulated gas concentration data includes horizontal simulated concentration data corresponding to the multiple two-dimensional grids at different times; the data association module 530 is specifically used for: The multi-source observed concentration data corresponding to the same time interval are matched and associated with the horizontal simulated concentration data of each of the two-dimensional grids, and the emission inventory data and the auxiliary data corresponding to the same time interval are matched and associated with the horizontal simulated concentration data of each of the two-dimensional grids respectively to obtain the associated data.

[0136] In some embodiments, the multi-source observation concentration data includes satellite observation concentration data, mobile observation concentration data, and base station observation concentration data; the data association module 530 is specifically used for: For any given time interval, based on the first correspondence between the satellite observation pixels corresponding to the satellite observation concentration data and the two-dimensional grid, the satellite observation concentration data is allocated to the corresponding two-dimensional grid. For any time interval, based on the second correspondence between the movement trajectory corresponding to the moving observation concentration data and the two-dimensional grid, the moving observation concentration data is matched and associated with the simulated gas concentration of the corresponding two-dimensional grid. For any time interval, based on the third correspondence between the base station observation location corresponding to the base station observation concentration data and the two-dimensional grid, the base station observation concentration data is allocated to the corresponding two-dimensional grid.

[0137] In some embodiments, the satellite observation pixel corresponds to at least one two-dimensional grid, the first correspondence includes the proportional relationship between the distribution area of ​​the satellite observation pixel in each corresponding grid and the area of ​​the observation pixel, and the satellite observation concentration data includes the column concentration corresponding to each satellite observation pixel; the data association module 530 is specifically used for: For each satellite observation pixel, the distribution weight of the column concentration corresponding to the satellite observation pixel in each corresponding two-dimensional grid is determined according to the proportional relationship between the distribution area of ​​the satellite observation pixel in each corresponding two-dimensional grid and the area of ​​the satellite observation pixel. For each two-dimensional grid, the satellite observation concentration of the two-dimensional grid is determined and allocated by assigning the column concentration of each observed pixel corresponding to the two-dimensional grid and the allocation weight of the column concentration of each observed pixel within the grid.

[0138] In some embodiments, the mobile observation concentration data includes UAV observation concentration data, which includes each trajectory point on the mobile trajectory and the corresponding observation concentration; the data association module 530 is specifically used for: Based on the three-dimensional coordinates corresponding to the concentration data observed by the UAV, multiple target two-dimensional grids corresponding to the three-dimensional coordinates are determined; For each target two-dimensional grid, determine the target trajectory point corresponding to the target two-dimensional grid and the corresponding observation concentration; Based on the observed concentrations at each target trajectory point, an interpolation method is used to determine the target UAV observed concentration data corresponding to the target two-dimensional grid, so as to match and associate the target UAV observed concentration data with the simulated gas concentration of the target two-dimensional grid.

[0139] In some embodiments, the simulated gas concentration data has corresponding vertically layered data; the data association module 530 is specifically used for: Based on the observed concentrations of each target trajectory point, the horizontal interpolation concentration of the center point of the target two-dimensional grid is determined using a linear interpolation method. Acquire the vertical layering data of the UAV, and determine the height of the near-ground layer of the model, the upper layer height of the UAV, and the lower layer height of the UAV from the vertical layering data of the UAV; From the observed concentrations of each target trajectory point, a first observed concentration corresponding to the upper layer height of the UAV and a second observed concentration corresponding to the lower layer height of the UAV are determined. Based on the first observed concentration and the second observed concentration, an interpolation method is used to determine the vertical interpolation concentration corresponding to the near-ground layer height of the model. Based on the horizontal interpolation concentration and the vertical interpolation concentration of the center point of the target two-dimensional grid, the target UAV observation concentration data corresponding to the target two-dimensional grid is determined.

[0140] In some embodiments, the optimization coefficient sequence includes optimization coefficients corresponding to each two-dimensional grid; the data optimization module 540 is specifically used for: Based on the optimization coefficients corresponding to each two-dimensional grid, the emissions of target greenhouse gases in the emission inventory data are optimized to obtain the optimized target emission inventory data.

[0141] In some embodiments, the data acquisition module 510 is specifically used for: Acquire raw multi-source observation concentration data; the raw multi-source observation concentration data includes raw satellite observation concentration data, raw mobile observation concentration data, and raw base station observation concentration data. The original satellite observation concentration data, the original mobile observation concentration data, and the original base station observation concentration data are respectively subjected to data filtering processing to obtain the satellite observation concentration data, the mobile observation concentration data, and the base station observation concentration data.

[0142] Please see Figure 6 This is a schematic diagram illustrating the structure of another apparatus for optimizing greenhouse gas emission inventory data, as shown in an exemplary embodiment of this application. Figure 6 As shown, the emission inventory data optimization device 500 also includes a data evaluation module 550; The data acquisition module 510 is also used for: Obtain the uncertainty assessment index values ​​corresponding to the emission inventory data, the multi-source observation concentration data, and the physical guided neural network, respectively; The data evaluation module 550 is used for: Based on the uncertainty assessment index values ​​corresponding to the emission inventory data, the multi-source observation concentration data, and the physical guided neural network, the uncertainty assessment index value of the target emission inventory data is determined; the uncertainty assessment index value is used to characterize the confidence level of the target emission inventory data.

[0143] The specific implementation process of the functions and roles of each unit in the above device can be found in the implementation process of the corresponding steps in the above method, and will not be repeated here.

[0144] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this application according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0145] Corresponding to the above-described method for optimizing greenhouse gas emission inventory data, this disclosure also provides a computer device, such as... Figure 7 The diagram shown is a structural schematic of a computer device provided in an embodiment of this disclosure. Figure 7As shown, the computer device 700 includes a processor 710, an internal bus 720, memory 730, a network interface 740, and non-volatile memory 750, and may also include other hardware required for its functions. One or more embodiments of this specification can be implemented in software, for example, the processor 710 reads the corresponding computer program from the non-volatile memory 750 into the memory 730 and then runs it. Of course, besides software implementation, one or more embodiments of this specification do not exclude other implementation methods, such as logic devices or a combination of hardware and software, etc. That is to say, the execution entity of the following processing flow is not limited to individual logic units, but can also be hardware or logic devices.

[0146] The memory 730, also known as internal memory, is used to temporarily store the computational data in the processor 710, as well as the data exchanged with non-volatile memory 750 such as hard disk. The processor 710 exchanges data with the non-volatile memory 750 through the memory 730.

[0147] In this embodiment, memory 730 is specifically used to store application code that executes the solution of this application, and its execution is controlled by processor 710. That is, when the computer device is running, processor 710 communicates with network interface 740, memory 730 and non-volatile memory 750 through internal bus 720, so that processor 710 executes the application code stored in memory 730 and non-volatile memory 750, thereby executing the emission inventory data optimization method described in the above method embodiment.

[0148] Processor 710 may be an integrated circuit chip with signal processing capabilities. The aforementioned processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware microservices. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor.

[0149] It is understood that the structures illustrated in the embodiments of this application do not constitute a specific limitation on the computer device 700. In other embodiments of this application, the computer device 700 may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0150] This disclosure also provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the steps of the greenhouse gas emission inventory data optimization method described in the above-described method embodiments. The storage medium may be a volatile or non-volatile computer-readable storage medium.

[0151] This disclosure also provides a computer program product carrying program code. The program code includes instructions that can be used to execute the steps of the method for optimizing greenhouse gas emission inventory data in the above method embodiments. For details, please refer to the above method embodiments, which will not be repeated here.

[0152] The aforementioned computer program product can be implemented through hardware, software, or a combination thereof. In one optional embodiment, the computer program product is specifically embodied in a computer storage medium; in another optional embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0153] The embodiments of the subject matter and functional operation described in this specification can be implemented in the following ways: digital electronic circuits, tangibly embodied computer software or firmware, computer hardware including the structures disclosed in this specification and their structural equivalents, or combinations thereof. Embodiments of the subject matter described in this specification can be implemented as one or more computer programs, i.e., one or more modules of computer program instructions encoded on a tangible, non-transitory program carrier for execution by a data processing apparatus or for controlling the operation of a data processing apparatus. Alternatively or additionally, the program instructions may be encoded on artificially generated propagation signals, such as machine-generated electrical, optical, or electromagnetic signals, which are generated to encode information and transmit it to a suitable receiving device for execution by the data processing apparatus. The computer storage medium may be a machine-readable storage device, a machine-readable storage substrate, a random or serial access memory device, or combinations thereof.

[0154] The processing and logic flow described in this specification can be executed by one or more programmable computers that execute one or more computer programs to perform corresponding functions by operating on input data and generating output. The processing and logic flow can also be executed by dedicated logic circuitry—such as FPGAs (Field-Programmable Gate Arrays) or ASICs (Application-Specific Integrated Circuits), and the device can also be implemented as dedicated logic circuitry.

[0155] Computers suitable for executing computer programs include, for example, general-purpose and / or special-purpose microprocessors, or any other type of central processing unit. Typically, the central processing unit receives instructions and data from read-only memory and / or random access memory. Basic computer microservices include a central processing unit for implementing or executing instructions and one or more memory devices for storing instructions and data. Typically, a computer will also include one or more mass storage devices for storing data, such as disks, magneto-optical disks, or optical disks, or the computer will be operatively coupled to such mass storage devices to receive data from or transfer data to them, or both. However, a computer is not required to have such devices. Furthermore, a computer can be embedded in another device, such as a mobile phone, a personal digital assistant (PDA), a mobile audio or video player, a game console, a global positioning system (GPS) receiver, or a portable storage device such as a universal serial bus (USB) flash drive, to name a few.

[0156] Computer-readable media suitable for storing computer program instructions and data include all forms of non-volatile memory, media, and memory devices, such as semiconductor memory devices (e.g., EPROM, EEPROM, and flash memory devices), magnetic disks (e.g., internal hard disks or removable disks), magneto-optical disks, and CD-ROM and DVD-ROM disks. Processors and memory may be supplemented by or incorporated into dedicated logic circuitry.

[0157] While this specification contains numerous specific implementation details, these should not be construed as limiting the scope of any invention or the scope of the claims, but rather are primarily intended to describe features of specific embodiments of a particular invention. Certain features described in the various embodiments herein may also be implemented in combination in a single embodiment. Conversely, various features described in a single embodiment may also be implemented separately in various embodiments or in any suitable sub-combination. Furthermore, while features may function in certain combinations as described above and even initially claimed in this way, one or more features from a claimed combination may be removed from that combination in some cases, and a claimed combination may refer to a sub-combination or a variation thereof.

[0158] Similarly, although the operations are depicted in a specific order in the accompanying drawings, this should not be construed as requiring these operations to be performed in the specific order shown or sequentially, or requiring all illustrated operations to be performed to achieve the desired result. In some cases, multitasking and parallel processing may be advantageous. Furthermore, the separation of various system modules and microservices in the above embodiments should not be construed as requiring such separation in all embodiments, and it should be understood that the described program microservices and systems can generally be integrated together in a single software product or packaged into multiple software products.

[0159] Thus, specific embodiments of the subject matter have been described. Other embodiments are within the scope of the appended claims. In some cases, the actions recited in the claims may be performed in a different order and still achieve the desired result. Furthermore, the processes depicted in the drawings are not necessarily shown in a specific order or sequence to achieve the desired result. In some implementations, multitasking and parallel processing may be advantageous.

[0160] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for optimizing greenhouse gas emission inventory data, characterized in that, include: Acquire emission inventory data, multi-source observation concentration data, and auxiliary data for the target greenhouse gas in the target area; the emission inventory data includes the emission amount of the target greenhouse gas in the spatiotemporal dimensions within the target area; Using an atmospheric transport model, gas concentration simulation is performed based on the emission inventory data to obtain simulated gas concentration data in the spatiotemporal dimension. The emission inventory data, the multi-source observation concentration data, and the auxiliary data are spatiotemporally aligned and correlated with the simulated gas concentration data to obtain correlated data. The associated data is input into a trained physical guidance neural network to obtain an optimization coefficient sequence, and the emission inventory data is optimized based on the optimization coefficient sequence to obtain optimized target emission inventory data.

2. The method according to claim 1, characterized in that, The target area is divided into multiple two-dimensional grids, and the simulated gas concentration data includes horizontal simulated concentration data corresponding to each of the multiple two-dimensional grids at different times; the emission inventory data, the multi-source observation concentration data, and the auxiliary data are spatiotemporally aligned and correlated with the simulated gas concentration data to obtain correlated data, including: The multi-source observed concentration data corresponding to the same time interval are matched and associated with the horizontal simulated concentration data of each of the two-dimensional grids, and the emission inventory data and the auxiliary data corresponding to the same time interval are matched and associated with the horizontal simulated concentration data of each of the two-dimensional grids respectively to obtain the associated data.

3. The method according to claim 2, characterized in that, The multi-source observation concentration data includes satellite observation concentration data, mobile observation concentration data, and base station observation concentration data; the step of matching and associating the multi-source observation concentration data corresponding to the same time interval with the horizontal simulated concentration data of each of the two-dimensional grids includes: For any given time interval, based on the first correspondence between the satellite observation pixels corresponding to the satellite observation concentration data and the two-dimensional grid, the satellite observation concentration data is allocated to the corresponding two-dimensional grid. For any time interval, based on the second correspondence between the movement trajectory corresponding to the moving observation concentration data and the two-dimensional grid, the moving observation concentration data is matched and associated with the simulated gas concentration of the corresponding two-dimensional grid. For any time interval, based on the third correspondence between the base station observation location corresponding to the base station observation concentration data and the two-dimensional grid, the base station observation concentration data is allocated to the corresponding two-dimensional grid.

4. The method according to claim 3, characterized in that, The satellite observation pixel corresponds to at least one two-dimensional grid, and the first correspondence includes the proportional relationship between the distribution area of ​​the satellite observation pixel in each corresponding grid and the area of ​​the observation pixel; the satellite observation concentration data includes the column concentration corresponding to each satellite observation pixel; the allocation of the satellite observation concentration data to the corresponding two-dimensional grid based on the first correspondence between the satellite observation pixel and the two-dimensional grid includes: For each satellite observation pixel, the distribution weight of the column concentration corresponding to the satellite observation pixel in each corresponding two-dimensional grid is determined according to the proportional relationship between the distribution area of ​​the satellite observation pixel in each corresponding two-dimensional grid and the area of ​​the satellite observation pixel. For each two-dimensional grid, the satellite observation concentration of the two-dimensional grid is determined and allocated by assigning the column concentration of each observed pixel corresponding to the two-dimensional grid and the allocation weight of the column concentration of each observed pixel within the grid.

5. The method according to claim 3, characterized in that, The mobile observation concentration data includes UAV observation concentration data, which includes each trajectory point on the mobile trajectory and the corresponding observation concentration; the matching and association of the mobile observation concentration data with the simulated gas concentration of the corresponding two-dimensional grid based on the second correspondence between the mobile trajectory corresponding to the mobile observation concentration data and the two-dimensional grid includes: Based on the three-dimensional coordinates corresponding to the concentration data observed by the UAV, multiple target two-dimensional grids corresponding to the three-dimensional coordinates are determined; For each target two-dimensional grid, determine the target trajectory point corresponding to the target two-dimensional grid and the corresponding observation concentration; Based on the observed concentrations at each target trajectory point, an interpolation method is used to determine the target UAV observed concentration data corresponding to the target two-dimensional grid, so as to match and associate the target UAV observed concentration data with the simulated gas concentration of the target two-dimensional grid.

6. The method according to claim 5, characterized in that, The observed concentration based on each target trajectory point is determined using an interpolation method to determine the target UAV observed concentration data corresponding to the target two-dimensional grid, including: Based on the observed concentrations of each target trajectory point, the horizontal interpolation concentration of the center point of the target two-dimensional grid is determined using a linear interpolation method. Acquire the vertical layering data of the UAV, and determine the height of the near-ground layer of the model, the upper layer height of the UAV, and the lower layer height of the UAV from the vertical layering data of the UAV; From the observed concentrations of each target trajectory point, a first observed concentration corresponding to the upper layer height of the UAV and a second observed concentration corresponding to the lower layer height of the UAV are determined. Based on the first observed concentration and the second observed concentration, an interpolation method is used to determine the vertical interpolation concentration corresponding to the near-ground layer height of the model. Based on the horizontal interpolation concentration and the vertical interpolation concentration of the center point of the target two-dimensional grid, the target UAV observation concentration data corresponding to the target two-dimensional grid is determined.

7. The method according to claim 1, characterized in that, The optimization coefficient sequence includes optimization coefficients corresponding to each two-dimensional grid; the optimization of the emission inventory data based on the optimization coefficient sequence to obtain optimized target emission inventory data includes: Based on the optimization coefficients corresponding to each two-dimensional grid, the emissions of target greenhouse gases in the emission inventory data are optimized to obtain the optimized target emission inventory data.

8. The method according to claim 3, characterized in that, The multi-source observation concentration data is generated through the following steps: Acquire raw multi-source observation concentration data; the raw multi-source observation concentration data includes raw satellite observation concentration data, raw mobile observation concentration data, and raw base station observation concentration data. The original satellite observation concentration data, the original mobile observation concentration data, and the original base station observation concentration data are processed respectively to obtain the satellite observation concentration data, the mobile observation concentration data, and the base station observation concentration data.

9. The method according to claim 1, characterized in that, After obtaining the optimized target emission inventory data, the method further includes: Obtain the uncertainty assessment index values ​​corresponding to the emission inventory data, the multi-source observation concentration data, and the physical guided neural network, respectively; Based on the uncertainty assessment index values ​​corresponding to the emission inventory data, the multi-source observation concentration data, and the physical guided neural network, the uncertainty assessment index value of the target emission inventory data is determined; the uncertainty assessment index value is used to characterize the confidence level of the target emission inventory data.

10. An optimization device for greenhouse gas emission inventory data, characterized in that, include: The data acquisition module is used to acquire emission inventory data, multi-source observation concentration data, and auxiliary data for the target greenhouse gas in the target area; the emission inventory data includes the emission amount of the target greenhouse gas in the spatiotemporal dimensions within the target area; The concentration simulation module is used to simulate gas concentrations based on the emission inventory data using an atmospheric transport model, and to obtain simulated gas concentration data in the spatiotemporal dimension. The data association module is used to perform spatiotemporal alignment and association between the emission inventory data, the multi-source observation concentration data, the auxiliary data and the simulated gas concentration data to obtain associated data. The data optimization module is used to input the associated data into a trained physical guided neural network to obtain an optimization coefficient sequence, and to optimize the emission inventory data based on the optimization coefficient sequence to obtain optimized target emission inventory data.

11. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the method for optimizing greenhouse gas emission inventory data according to any one of claims 1-9.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the method for optimizing greenhouse gas emission inventory data according to any one of claims 1-9.

Images