System and method for monitoring greenhouse gas emissions

Abstract

A system and method for monitoring greenhouse gas emissions is disclosed. A plurality of satellite observations associated with emissions of greenhouse gases in a first area of interest are received from a plurality of satellite data sources, respectively. The plurality of satellite observations are fused to produce a fused input dataset. An emissions estimation model is used to generate a first estimate of emissions of greenhouse gases in the first area of interest based on the fused input dataset.

Description

[0001] Cross-references to related applications

[0002] This application claims priority to U.S. Patent Application No. 17 / 540,205, filed December 1, 2021, the entire contents of which are incorporated herein by reference. Technical Field

[0003] This disclosure relates to gas emission monitoring, and more specifically, to a system, method, and storage medium for monitoring greenhouse gas emissions. Background Technology

[0004] The quantification and governance scale of anthropogenic carbon dioxide (CO2) emissions at point sources (also referred to as emission sources in this paper) are fundamental to achieving, monitoring, and evaluating the progress towards carbon neutrality. For example, implementing cap-and-trade programs and carbon taxes requires verifying CO2 emissions at the relevant power plants and factories. In another example, understanding the heterogeneity of CO2 emissions at the city and county levels helps in policy implementation to effectively reduce CO2 emissions. Yet another example, at the international level, disputes over national greenhouse gas emission inventories hinder the implementation of carbon boundary adjustment taxes and the attribution of national responsibility for climate change. These disputes can be resolved through the quantification of CO2 emissions. Summary of the Invention

[0005] On one hand, a method for monitoring greenhouse gas emissions is disclosed. Multiple satellite observations associated with greenhouse gas emissions in a first region of interest are received from multiple satellite data sources. The multiple satellite observations are fused to produce a fused input dataset. An emission estimation model is used to generate a first emission estimate of greenhouse gases in the first region of interest based on the fused input dataset.

[0006] In another aspect, a system for monitoring greenhouse gas emissions is disclosed. The system includes a memory and a processor. The memory is configured to store instructions. The processor is coupled to the memory and configured to execute the instructions to perform a process including: receiving multiple satellite observations associated with greenhouse gas emissions in a first region of interest from multiple satellite data sources; fusing the multiple satellite observations to generate a fused input dataset; and generating a first emission estimate of the greenhouse gas in the first region of interest using an emission estimation model based on the fused input dataset.

[0007] In another aspect, a non-transitory computer-readable storage medium is disclosed. This computer-readable storage medium is configured to store instructions that, in response to execution by a processor, cause the processor to perform a process comprising: receiving, respectively, multiple satellite observations associated with greenhouse gas emissions in a first region of interest from multiple satellite data sources; fusing the multiple satellite observations to generate a fused input dataset; and, based on the fused input dataset, using an emissions estimation model to generate a first emissions estimate of the greenhouse gas in the first region of interest.

[0008] It should be understood that the foregoing general description and the following detailed description are merely exemplary and explanatory, and do not limit the claimed invention. Attached Figure Description

[0009] The accompanying drawings, which are incorporated herein and form part of the specification, illustrate implementations of the present disclosure and, together with the specification, further serve to explain the present disclosure and enable those skilled in the art to make and use it.

[0010] Figure 1A-1C This is a graphical representation illustrating an exemplary implementation of the Gaussian plume model for estimating CO2 emissions disclosed in this invention.

[0011] Figure 2 The illustration shows an exemplary operating environment of a system disclosed in this invention for configuring to monitor greenhouse gas emissions.

[0012] Figure 3A The illustration shows an exemplary process for monitoring greenhouse gas emissions disclosed in this invention.

[0013] Figure 3B The illustration shows another exemplary process for monitoring greenhouse gas emissions disclosed in this invention.

[0014] Figure 3C The illustration shows yet another exemplary process for monitoring greenhouse gas emissions disclosed in this invention.

[0015] Figure 4 The illustration shows an exemplary decomposition process disclosed in this invention for decomposing greenhouse gas emission values ​​associated with a grid into one or more decomposed greenhouse gas emission values ​​from one or more emission sources within the grid.

[0016] Figure 5 This is a flowchart of an exemplary method for monitoring greenhouse gas emissions disclosed in this invention.

[0017] Figure 6 This is a flowchart of another exemplary method for monitoring greenhouse gas emissions disclosed in this invention.

[0018] The implementation of this disclosure will be described with reference to the accompanying drawings. Detailed Implementation

[0019] Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. Where possible, the same reference numerals are used in all the drawings to denote the same or similar parts.

[0020] Traditional methods for quantifying greenhouse gas emissions (e.g., CO2 emissions) involve multiplying emission-generating activities (e.g., cement production) by an emission factor specifying greenhouse gas emissions per unit of activity. While conceptually simple, this approach can encounter various challenges in implementation. For example, emission factors are not only region-specific but also differentiated by technology and operating conditions. The derivation of emission factors is infeasible at many developing countries and at local scales across the country. In another example, reporting of emission-generating activities provided by the emission sources themselves can be subject to delays, errors, and fraud. Yet another example, the accuracy and consistency of greenhouse gas emission estimates can be greatly affected by data sources and model assumptions.

[0021] Alternative methods for monitoring CO2 emissions include satellite observations of the column-mean dry air mole fraction (xCO2) of CO2 in the atmosphere, which can be termed satellite remote sensing-based methods. For example, Gaussian plume models can be applied to estimate CO2 emissions at a point source scale using xCO2 satellite observations. See below for further details. Figure 1A-1C A more detailed description of the Gaussian plume model is provided. It has been demonstrated that a Gaussian plume can be fitted onto xCO2 data obtained from the Orbiting Carbon Observatory 2 (OCO-2) satellite. By correlating the plume with its nearby emission sources (e.g., power plants) and by quantifying the xCO2 difference between the background and the plume, the daily average CO2 emissions from the emission sources can be estimated when xCO2 data from the OCO-2 satellite are available.

[0022] Satellite-based remote sensing methods are promising because they address the problems of traditional approaches. For example, xCO2 data from the OCO-2 satellite covers the globe with a spatial resolution of 2.25 km per pixel. Furthermore, xCO2 data from the Greenhouse Gas Observing Satellites 1 (GOSAT-1) and GOSAT-2 can have a spatial resolution of 10 km per pixel. Therefore, CO2 emission estimates can potentially be achieved globally at a local scale. In another example, xCO2 data from the OCO-2 satellite was publicly available on the day of initial observations, and the gridded spatial estimate of CO2 emissions can be validated using ground-based measurement sensors (e.g., gas analyzers). Therefore, this method is less susceptible to spoofing and delays, and different estimation methods can be systematically evaluated using the same set of field measurements.

[0023] However, satellite-based remote sensing methods may have several technical problems and bottlenecks. For example, the OCO-2 satellite may not provide sufficient xCO2 data for the desired temporal and spatial coverage. Specifically, the OCO-2 satellite can provide xCO2 measurements within a narrow ground field of view (eight 2.25 x 2.25 km² pixels) perpendicular to the satellite bridge. Although the OCO-2 satellite can cover the globe every 16 days, the narrow field of view can result in large spatial gaps between strips. Furthermore, the presence of clouds and smog can invalidate xCO2 data collected by the OCO-2 satellite. For Gaussian plume models with specific imagery standards for estimating emissions, it has been reported that only one image from a two-year time period is suitable at a specific point source of interest.

[0024] In another example, Gaussian plume models in satellite-sensing-based methods can be applied to isolated medium and large point sources, which is not feasible for estimating CO2 emissions from single point sources clustered in small areas. In yet another example, previous studies in satellite-sensing-based methods have only demonstrated the feasibility of estimating CO2 emissions using manually selected data at limited locations. On a daily and global scale, this approach may be labor-intensive.

[0025] To address one or more of the aforementioned problems, this disclosure provides a system and method for monitoring greenhouse gas emissions by utilizing satellite data fusion and surrogate estimation. The greenhouse gases described herein can be any gas capable of trapping heat in the atmosphere and warming the Earth, and may include CO2, methane, nitrous oxide, fluorinated gases, etc. Without loss of generality, CO2 may be used as an example of a greenhouse gas described in the following description of this disclosure.

[0026] In the systems and methods disclosed herein, satellite observations related to greenhouse gas emissions (e.g., including various xCO2 data from OCO-2, OCO-3, and other satellites) can be fused to produce a fused xCO2 input dataset, thereby enabling a better approximation of the plume's extent and the xCO2 values ​​of both the background and the plume. Daily observations of wind and other industrial gases (e.g., methane) can be used to guide the reconstruction of plumes outside the field of view of the OCO-2 or OCO-3 satellites. Gap-filling techniques can be used to increase the number of satellite images that can be applied to Gaussian plume models. Furthermore, when xCO2 data is unavailable, observations from other reference resources (e.g., methane, NO2, or thermal signals) can be used as proxies for estimating CO2 emissions. Additionally, decomposition processing can be applied to integrate coarse-resolution satellite-based emission estimates into fine-scale emission maps by facility object (e.g., plant clusters).

[0027] The systems and methods disclosed in this paper improve the applicability of Gaussian plume models for monitoring greenhouse gas emissions from point sources with short delays. For example, emission estimates can be obtained within 2–3 days of the emission, which is well-suited for monitoring and implementing carbon neutrality policies. The systems and methods disclosed in this paper can significantly increase the frequency of satellite-based emission estimates by using surrogate estimation methods, and emission estimates for small industrial clusters can be achieved using decomposition processes. Implementing surrogate estimation methods over large areas allows for the estimation of industrial emissions at the governance scale.

[0028] Figure 1A-1C This is a graphical representation illustrating exemplary implementations of Gaussian plume models for estimating CO2 emissions, based on several examples. (Reference) Figure 1A Image 100 shows wind direction 104 and point source 102 emitting CO2. Point source 102 can be, for example, a power plant, cement plant, steel plant, or any other power plant or plant emitting greenhouse gases such as CO2. Strip 106 shows the OCO-2 overpass or flyover that correlates xCO2 data from the OCO-2 satellite with a portion of the CO2 emissions from point source 102. Gaussian plume models can be used to fit and model xCO2 data from the OCO-2 satellite to provide an estimate of CO2 emissions associated with point source 102.

[0029] For example, for each overpass or flyover, the magnitude of the vector average of the ERA-intermediate and MERRA2 winds can be used as the wind speed to simulate the plume. The Gaussian plume model equations can be expressed by the following expressions (1) and (2):

[0030]

[0031]

[0032] In the above expressions (1) and (2), V represents the vertical column of CO2 at the point source and downwind of the point source, in g / m². The x-direction is parallel to the wind direction, and x represents the distance (in meters (m)) from the point source to the direction parallel to the wind direction. The y-direction is perpendicular to the wind direction, and y represents the distance (in meters) across the wind direction. V depends on the emission rate F (g / s), the lateral wind distance y (m), the wind speed u (m / s), and the standard deviation in the y-direction (e.g., σ). y (m)). x0 = 1000m can be the characteristic length, making the independent variable of the exponent dimensionless. α represents the atmospheric stability parameter, which can be determined by classifying the source environment according to Gaussian plume stability, and depends on surface wind speed, cloud cover, and time of day. Surface wind speed and cloud cover can be taken from the ERA period.

[0033] A region of the OCO-2 stripe (e.g., upwind and therefore unaffected by point sources) can be selected as the background, and xCO2 from these points in that region can be averaged. The model plume can then be defined as the area downwards from the x-axis (wind vector) in both the positive and negative y-directions at an intensity threshold (e.g., a 5% intensity threshold). The observed plume can be defined based on the points corresponding to the model plume, taking into account the optical path. The extent of the plume can be defined by V(x, y) in the above expressions (1) and (2), where V(x, y) is greater than a predetermined intensity threshold (e.g., 5% greater than the background concentration of CO2 or another emitted gas). For example, the extent of the plume can be the region defined by V(x, y), where V(x, y) is greater than a predetermined intensity threshold (e.g., 5% greater than the background concentration of CO2).

[0034] Figure 1B An exemplary xCO2 plume is shown relative to the background. Figure 1C An exemplary xCO2 plume relative to the background is shown, as would be seen by the OCO-2 satellite. Figure 1B-1C The dashed line in the figure shows the 5% plume density cutoff value from the axial value of the Gaussian plume model.

[0035] Estimating greenhouse gas (e.g., CO2) emissions over large areas at the point source and / or governance scale can be challenging, given the limitations of data availability from one or more satellites. To address this challenge, this paper discloses a system and method for monitoring greenhouse gas emissions, which will be described below with reference to... Figure 2-6 This will be described in more detail. In the systems and methods disclosed herein, the availability of input data to emission estimation models (e.g., Gaussian plume models) can be increased by compiling and fusing xCO2 observations from multiple satellites. Furthermore, the amount of appropriate input data for emission estimation models can be increased by simultaneously using industrial gas data to augment xCO2 data. Additionally, gap-filling of CO2 emission estimates in regions where no xCO2 data is available can be performed by deriving and applying a mapping between CO2 emissions and one or more reference sources. One or more reference sources can include industrial gas emissions, such as methane, land surface temperature, or shortwave infrared thermal signals. Furthermore, the spatial resolution of CO2 emission estimates can be further improved by decomposing the total emissions values ​​over a grid into multiple identifiable emission sources within the grid. Therefore, the systems and methods disclosed herein can provide rigorous, consistent, and efficient CO2 emission estimates for diverse emission sources worldwide.

[0036] Figure 2A block diagram of an exemplary operating environment 200 for a system 201 configured to monitor greenhouse gas emissions according to an embodiment of the present disclosure is illustrated. The operating environment 200 may include the system 201, a user equipment 212, and a plurality of data sources 218A, ..., 218N. Components of the operating environment 200 may be coupled to each other via a network 210.

[0037] In some embodiments, system 201 may be contained within a cloud computing device. Alternatively, system 201 may be contained within a local computing device. The computing device may be, for example, a server, desktop computer, laptop computer, tablet computer, or any other suitable electronic device including a processor and memory. In some embodiments, system 201 may include processor 202, memory 203, storage device 204, and associated database 215. It should be understood that system 201 may also include any other suitable components for performing the functions described herein.

[0038] For example, system 201 may have different components in a single device, such as integrated circuit (IC) chips, or separate devices with dedicated functions. The IC may be implemented as an application-specific integrated circuit (ASIC) or a field-programmable gate array (FPGA). In another example, one or more components of system 201 may be located in a cloud computing environment, or alternatively in a single location or distributed locations, but communicate with each other via network 210.

[0039] Processor 202 may include any suitable type of general-purpose or special-purpose microprocessor, digital signal processor, microcontroller, graphics processing unit (GPU), etc. Processor 202 may include one or more hardware units (e.g., part of an integrated circuit) designed to be used with other components or as part of a program. This program may be stored on a computer-readable medium and, when executed by processor 202, may perform one or more functions.

[0040] Processor 202 can be configured as a separate processor module dedicated to monitoring greenhouse gas emissions. Alternatively, processor 202 can be configured as a shared processor module for performing other functions. Processor 202 may include several modules, such as fusion module 205, enhancement module 206, estimation module 207, mapping module 208, and decomposition module 209. Although Figure 2The fusion module 205, enhancement module 206, estimation module 207, mapping module 208, and decomposition module 209 are shown within a single processor 202, but they may also be implemented on different processors, either close to or far from each other. For example, mapping module 208 may be implemented by a processor (e.g., a GPU) dedicated to offline training of machine learning models to derive mapping relationships, while estimation module 207 may be implemented by another processor for estimating greenhouse gas emissions based on the mapping relationships.

[0041] The fusion module 205, enhancement module 206, estimation module 207, mapping module 208, and decomposition module 209 (and any corresponding submodules or subunits) may be hardware units (e.g., portions of an integrated circuit) of the processor 202, designed to be used in conjunction with other components or software units implemented by the processor 202 by executing at least a portion of a program. The program may be stored on a computer-readable medium such as memory 203 or storage device 204, and when executed by the processor 202, it may perform one or more functions.

[0042] The following is for reference. Figure 3A-6 The fusion module 205, enhancement module 206, estimation module 207, mapping module 208, decomposition module 209, and association database 215 are described in more detail.

[0043] Memory 203 and storage device 204 may include any suitable type of mass storage device provided to store any type of information that processor 202 may need to operate. For example, memory 203 and storage device 204 may be volatile or non-volatile, magnetic, semiconductor-based, magnetic tape-based, optical, removable, non-removable, or other types of storage devices or tangible (i.e., non-transitory) computer-readable media, including but not limited to ROM, flash memory, dynamic RAM, and static RAM. Memory 203 and / or storage device 204 may be configured to store one or more computer programs that can be executed by processor 202 to perform the functions disclosed herein. For example, memory 203 and / or storage device 204 may be configured to store a program that can be executed by processor 202 to estimate greenhouse gas emissions. Memory 203 and / or storage device 204 may also be configured to store information and data used by processor 202.

[0044] User equipment 212 may be a computing device including a processor and memory. For example, user equipment 212 may be a desktop computer, laptop computer, tablet computer, smartphone, game controller, television (TV), music player, wearable electronic device such as a smartwatch, internet device, smart vehicle, or any other suitable electronic device with a processor and memory. User equipment 212 may be operated by a user. In some embodiments, user equipment 212 may receive a request from a user for estimating greenhouse gas emissions in a region of interest, and may forward the request to system 201, causing system 201 to generate an emission estimate of greenhouse gases in the region of interest. System 201 may send the greenhouse gas emission estimate to user equipment 212, allowing user equipment 212 to present the greenhouse gas emission estimate to the user via its screen.

[0045] Multiple data sources 218A, ..., 218N may include multiple satellite data sources configured to store multiple satellite observations associated with greenhouse gas emissions in the region of interest, respectively. For example, the multiple satellite data sources may include one or more of the following: an OCO-2 data source configured to store xCO2 data from the OCO-2 satellite, an OCO-3 data source configured to store xCO2 data from the OCO-3 satellite, a GOSAT-1 data source configured to store xCO2 data from the GOSAT-1 satellite, a GOSAT-2 data source configured to store xCO2 data from the GOSAT-2 satellite, or a TANSAT data source configured to store xCO2 data from the TANSAT satellite.

[0046] Multiple data sources 218A,…,218N may also include multiple reference data sources configured to store observations of multiple reference resources in the region of interest, respectively. For example, the multiple reference resources may include at least one of industrial gas resources, land surface temperature resources, or shortwave infrared heat sources. These multiple reference data sources may include one or more of the following: an industrial gas data source configured to store industrial gas observations obtained from the Sentinel-5P satellite; a land surface temperature data source configured to store land surface temperature observations obtained from an Earth Resources Satellite; or a shortwave infrared data source configured to store shortwave infrared thermal observations obtained from the Sentinel-2 infrared band (e.g., for structures above 1300 K, such as blast furnaces).

[0047] Figure 3AAn exemplary process 300 for monitoring greenhouse gas emissions in a region of interest according to an embodiment of the present disclosure is illustrated. In process 300, the region of interest may include one or more point sources (e.g., power plants) that may emit greenhouse gases. One or more first satellite observations 304 associated with greenhouse gas emissions in the region of interest may be obtained from one or more first satellite data sources. For example, the greenhouse gas may be CO2. The one or more first satellite observations may include one or more first observations of the column-mean dry air mole fraction (xCO2) of CO2 in the atmosphere (referred to herein as “xCO2 observations”). The one or more first xCO2 observations may include, for example, xCO2 data from the OCO-2 satellite, xCO2 data from the OCO-3 satellite, or both. In some cases, the one or more first xCO2 observations may include sufficient xCO2 data such that CO2 emissions in the region of interest can be estimated using an emission estimation model 306 based solely on the one or more first xCO2 observations.

[0048] Initially, estimation module 207 can be configured to receive wind data 302 (such as wind direction and speed) and one or more first satellite observations 304 (e.g., one or more first xCO2 observations). Estimation module 207 can input the wind data 302 and the one or more first satellite observations 304 into an emissions estimation model 306. Emissions estimation model 306 may include, for example, a Gaussian plume model. Estimation module 207 can apply emissions estimation model 306 to generate greenhouse gas emissions estimates 308 for the region of interest based on the wind data 302 and the one or more first satellite observations 304.

[0049] Mapping module 208 can associate greenhouse gas emission estimates 308 with one or more observations from one or more reference resources in the region of interest to generate a training dataset. Mapping module 208 can store the training dataset in an association database 215. In some embodiments, one or more observations from one or more reference resources may include at least one of the following associated with the region of interest: industrial gas observation 310, surface temperature observation 312, or shortwave infrared thermal observation 314. Industrial gas emissions 372 can be estimated using emission estimation model 306 based on wind data 302 and industrial gas observation 310. Mapping module 208 can associate greenhouse gas emission estimates 308 with each of industrial gas observation 310 (or equivalently, industrial gas emissions 372), surface temperature observation 312, and shortwave infrared thermal observation 314. For example, mapping module 208 can pair greenhouse gas emission estimates 308 with each of industrial gas observation 310 (or equivalently, industrial gas emissions 372), surface temperature observation 312, and shortwave infrared thermal observation 314 to generate a training dataset. (Refer to the following...) Figure 3CDescribe the training dataset in more detail.

[0050] Figure 3B Another exemplary process 350 for monitoring greenhouse gas emissions in a region of interest according to an embodiment of the present disclosure is illustrated. In process 350, the region of interest may include one or more point sources that may emit greenhouse gases. One or more first satellite observations 304 associated with greenhouse gas emissions in the region of interest may be obtained from one or more first satellite data sources. One or more second satellite observations 354 associated with greenhouse gas emissions in the region of interest may be obtained from one or more second satellite data sources.

[0051] For example, a greenhouse gas could be CO2. One or more first satellite observations 304 may include one or more first xCO2 observations, including, for example, xCO2 data from the OCO-2 satellite, xCO2 data from the OCO-3 satellite, or both. One or more second satellite observations 354 may include one or more second xCO2 observations, including, for example, at least one of the following: xCO2 data from the GOSAT-1 satellite, xCO2 data from the GOSAT-2 satellite, or xCO2 data from the TANSAT satellite. xCO2 data from various satellites may vary in terms of accuracy (0.5-4 ppm), spatial resolution (2-10 km), field of view, or orbital revisit time (3-16 days). xCO2 data from the OCO-2 or OCO-3 satellites may have a higher spatial resolution than xCO2 data from the GOSAT-1, GOSAT-2, or TANSAT satellites.

[0052] In some embodiments, one or more first xCO2 observations may include sufficient xCO2 data for Gaussian plume fitting, such that CO2 emissions in the region of interest can be estimated based solely on one or more first xCO2 observations. Alternatively, one or more first xCO2 observations may not include sufficient xCO2 data for Gaussian plume fitting, such that one or more first xCO2 observations can be fused with one or more second xCO2 observations to estimate CO2 emissions in the region of interest.

[0053] The fusion module 205 can be configured to perform data fusion 352 based on multiple satellite observations to generate a fused input dataset 356 of the region of interest. For example, the fusion module 205 can fuse one or more first satellite observations 304 with one or more second satellite observations 354 to generate the fused input dataset 356. The region of interest may include a geographic area divided into multiple grids (e.g., multiple 2.25 x 2.25 km² pixels). Through the operation of data fusion 352, the spatial coverage of xCO2 data can be improved as input to the emission estimation model 306. The output grid of the operation of data fusion 352 may have an xCO2 data spatial resolution of approximately 2 km (e.g., 2 km, 2 km ± 5%, 2 km ± 10%, etc.), where the box boundaries enclose the region of interest (e.g., including both emission point sources and CO2 plumes from Gaussian plume models).

[0054] Specifically, fusion module 205 can resample each satellite observation into one or more resampled observations associated with one or more grids in a plurality of grids. For example, each satellite observation may include one or more initial observations (e.g., initial xCO2 values) associated with a region of interest obtained from a corresponding satellite data source. Fusion module 205 may apply a geographic weighting method to generate one or more resampled observations (e.g., resampled xCO2 values) associated with one or more grids based on one or more initial observations. Exemplary geographic weighting methods may include, but are not limited to, the average of neighbors, the nearest Gaussian weighting method, or any other suitable weighting method. Next, for each of the plurality of grids, fusion module 205 may determine the availability of resampled observations associated with that grid. Fusion module 205 may generate fusion input values ​​associated with the grid based on the availability of resampled observations associated with the grid. Fusion module 205 may then generate a fused input dataset 356 to include corresponding fusion input values ​​for each of the plurality of grids.

[0055] For example, one or more first initial observations associated with a region of interest can be extracted from xCO2 data obtained from OCO-2 and / or OCO-3 satellites, and these one or more first initial observations can be used with a geographic weighting method to generate one or more first resampled observations associated with one or more first grids. The one or more first grids may be at least a portion of a plurality of grids, and the remaining grids in the plurality of grids may not have first resampled observations. Similarly, one or more second initial observations associated with a region of interest can be extracted from xCO2 data from GOSAT-1, GOSAT-2, and / or TANSAT satellites, and these second initial observations can be used with a geographic weighting method to resample them across a plurality of grids to generate one or more second resampled observations associated with one or more second grids. The one or more second grids may be at least a portion of a plurality of grids, and the remaining grids in the plurality of grids may not have second resampled observations. The one or more first grids may or may not overlap with the one or more second grids.

[0056] For a grid that simultaneously has both first and second resampled observations (i.e., the grid is an overlapping grid between one or more first grids and one or more second grids), the fusion module 205 can determine the fusion input value associated with that grid (e.g., the fusion xCO2 input value) as the average of the first and second resampled observations. Then, the average resampled observation (denoted as ) in the overlapping region between one or more first grids and one or more second grids can be calculated as the average of the fusion input values ​​of the overlapping grids between one or more first grids and one or more second grids.

[0057] For a grid with a first resampled observation or a second resampled observation (i.e., the grid is only in one or more first grids or one or more second grids, but not in two grids), the fusion module 205 can determine the fusion input value associated with the grid as the first resampled observation or the second resampled observation.

[0058] For a grid without resampled observations (i.e., the grid is not included in one or more first grids and one or more second grids), the fusion module 205 can use the following expression (3) to calculate the fusion input value for that grid:

[0059] xCO 2(filled) =((n) coarse -n oco )×(xcO 2(coarse) )- xCO 2(oco) ) / n coarse (3)

[0060] In the above expression (3), xCO 2(filled) This represents the fusion input value associated with the mesh (e.g., fusion xCO2 input value), n coarse xCO represents the total number of grids associated with each individual satellite observation (e.g., 2.25 x 2.25 km² pixels). 2(coarse) This represents the initial observations (e.g., initial xCO2 values) obtained from GOSAT-1, GOSAT-2, or TANSAT satellite observations, and n oco This represents the total number of overlapping grids between one or more first grids and one or more second grids (e.g., an overlap of 2.25 x 2.25 km² pixels).

[0061] Enhancement module 206 can be configured to enhance the fused input dataset 356 using industrial gas observations 310 associated with the region of interest. Specifically, enhancement module 206 can apply wind data 302 and industrial gas observations 310 to emission estimation model 306 to estimate one or more model parameters 358 associated with emission estimation model 306. Emission estimation model 306 may include a Gaussian plume model, and one or more model parameters 358 may include the extent of the plume in the Gaussian plume model.

[0062] In some embodiments, atmospheric concentrations in certain industries can be observed by satellite, and there are numerous methods to estimate their ground-based emissions. Routine industrial gaseous products (e.g., methane, carbon monoxide, sulfur dioxide, etc.) from the sentinel-5P satellite can then be used to approximate the extent of a CO2 plume in a Gaussian plume model. For example, if a CO2 emission point source also emits sulfur dioxide, a Gaussian plume can be fitted onto sulfur dioxide data to determine the plume's emission rate and extent. The plume's emission rate and extent determined from the sulfur dioxide data can then be used to estimate CO2 emissions. For illustrative purposes, it is assumed that chemical reactions are negligible and that the rates of movement and dispersion in the emitted gases are similar. This assumption allows the plume extent determined from the sulfur dioxide data to be directly used as the plume extent for CO2 emission estimation. If the estimated plume partially overlaps with actual OCO observations, the Gaussian plume model equations (e.g., expressions (1)-(2)) above) and the actual OCO observations can be used to inversely calculate one or more parameters of the equations (e.g., including the CO2 emission rate). In fact, more complex calculations involving chemical transport models can be used to convert the plume associated with one type of emission gas into the range of a plume associated with another type of emission gas.

[0063] For example, when fitting a Gaussian plume model to non-CO2 air pollutant emissions (e.g., methane), the wind direction can be adjusted so that V (the vertical column of pollution across position (x, y)) best matches the shape of the pollution plume on satellite imagery. Then, based on wind speed u and atmospheric stability parameter a, the emission rate of gases such as methane (denoted as F-methane) can be estimated. The adjusted wind direction, methane emission rate (F-methane), wind speed u, atmospheric stability parameter a, etc., can be reused in CO2 emission estimation. For example, by assuming that CO2 emissions are proportional to air pollution emissions (e.g., methane), the CO2 emission rate (denoted as F_CO2) in the Gaussian plume model can be written as F_CO2 = r × Fmethane, where r is optimized such that V(x, y) for the Gaussian plume model of CO2 emissions best fits the limit of xCO2 data availability from the OCO-2 and / or OCO-3 satellites. The plume range from the Gaussian plume model can then be determined based on one or more of the following: the CO2 emission rate (F_CO2) based on the above expressions (1) and (2), the tuned wind direction, the wind speed u, and the atmospheric stability parameter a.

[0064] The estimation module 207 can be configured to apply the fused input dataset 356 to the emission estimation model 306 to generate greenhouse gas emission estimates 308 for the region of interest based on one or more model parameters 358. The mapping module 208 can be configured to correlate the greenhouse gas emission estimates 308 with each of industrial gas emissions 372 (or equivalently, industrial gas observations 310), surface temperature observations 312, and shortwave infrared thermal observations 314 to generate a training dataset. The mapping module 208 can store the training dataset in an association database 215.

[0065] Figure 3C Another exemplary process 370 for monitoring greenhouse gas emissions according to embodiments of the present disclosure is illustrated. Process 370 illustrates a surrogate estimation method for estimating greenhouse gas emissions when no satellite observations associated with greenhouse gas emissions are available in the region of interest. In the surrogate estimation method, a mapping relationship between greenhouse gas emissions and one or more reference resources for each grid in the region of interest can be derived. In some embodiments, the mapping relationship can be established by emission infrastructure category (e.g., coal-fired power plants versus natural gas-fired power plants) or by search distance from the point source of interest. The derived mapping relationship can be applied to estimate CO2 emissions in the absence of xCO2 data, or to identify anomalous estimates using a Gaussian plume model.

[0066] Specifically, mapping module 208 can be configured to perform the mapping relationship derivation 374 operation to determine the mapping relationship 376 between greenhouse gas emissions and one or more reference resources based on the association database 215. The one or more reference resources may include at least one of industrial gas resources, surface temperature resources, or shortwave infrared heat sources.

[0067] Next, estimation module 207 can be configured to obtain one or more observations from one or more reference resources in the region of interest. Estimation module 207 can determine greenhouse gas emission estimates 308 in the region of interest based on mapping relation 376 and the one or more observations from the one or more reference resources. The one or more observations from the one or more reference resources may include at least one of industrial gas observations 310, surface temperature observations 312, or shortwave infrared thermal observations 314. For example, industrial gas emissions 372 can be estimated using emission estimation model 306 based on wind data 302 and industrial gas observations 310. Estimation module 207 can apply mapping relation 376 based on industrial gas emissions 372, surface temperature observations 312, and shortwave infrared thermal observations 314 to generate greenhouse gas emission estimates 308.

[0068] In some embodiments, the mapping relationship 376 can be modeled by a mapping model. The mapping model can be a Gaussian process regression model, a random forest model, a multivariate regression model, a gradient-enhanced decision tree model, a neural network model, or any other suitable model. The mapping model can take at least one of industrial gas observations 310 (or industrial gas emissions 372), surface temperature observations 312, or shortwave infrared thermal observations 314 as input, and can generate a greenhouse gas emission estimate 308 as output. For example, the mapping module 208 can be configured to train the mapping model using a training dataset stored in the association database 215. The estimation module 207 can apply the trained mapping model to generate the greenhouse gas emission estimate 308 based on one or more of the following: industrial gas observations 310 (or industrial gas emissions 372), surface temperature observations 312, or shortwave infrared thermal observations 314.

[0069] Figure 4The illustration illustrates an exemplary decomposition process according to embodiments of the present disclosure for decomposing a greenhouse gas emission value associated with a grid into one or more decomposed greenhouse gas emission values ​​from one or more emission sources within the grid. In some embodiments, the region of interest may be divided into multiple grids, and the greenhouse gas emission estimate in the region of interest may include multiple greenhouse gas emission values ​​for the multiple grids respectively. For each of the multiple grids, the decomposition module 209 may identify one or more emission sources within the grid and may decompose the greenhouse gas emission value 408 associated with the grid into one or more decomposed greenhouse gas emission values ​​412 from one or more emission sources.

[0070] For example, mapping module 208 can perform the mapping derivation 374 operation to determine the mapping relationship 376 based on the association database 215. The mapping relationship 376 between greenhouse gas emissions and each of ground temperature and shortwave infrared thermal signal can be used to decompose the greenhouse gas emission values ​​408 associated with the grid. First, one or more emission sources within the grid can be classified and depicted based on high-resolution satellite imaging, maps, and other auxiliary data. If only the emission source of interest (e.g., a single potential emitter) exists at the upwind location away from the plume, decomposition module 209 can attribute the greenhouse gas emission value 408 of the grid to the emission source of interest, and decomposition of the greenhouse gas emission value 408 is not required. However, if multiple emission sources are identified as potential emitters (e.g., a cluster of factories), decomposition module 209 can perform the decomposition 410 operation to decompose the greenhouse gas emission value 408.

[0071] Specifically, for each of the multiple emission sources, the decomposition module 209 can obtain one or more observations from one or more reference resources associated with that emission source. The one or more observations from the one or more reference resources may include at least one of the following: land surface temperature observation 402 or shortwave infrared thermal observation 404. Land surface temperature observation 402 and shortwave infrared thermal observation 404 may have spatial resolutions of approximately 100 m and 20 m, respectively, which are higher than the spatial resolution of greenhouse gas emission values ​​408 (e.g., 2 km).

[0072] Next, the decomposition module 209 can determine an initial emission estimate 406 for greenhouse gases at emission sources based on mapping relationship 376 using one or more observations from one or more reference resources (e.g., surface temperature observation 402, shortwave infrared thermal observation 404). Therefore, by performing a similar operation, the decomposition module 209 can determine initial emission estimates 406 for greenhouse gases at multiple emission sources separately.

[0073] Then, the decomposition module 209 can perform a decomposition operation 410 to decompose the greenhouse gas emission value 408. For example, the decomposition module 209 can decompose the greenhouse gas emission value 408 to generate a decomposed greenhouse gas emission value 412 based on the initial emission estimate 406 of the greenhouse gas for the identified emission sources. Specifically, the decomposition module 209 can decompose the greenhouse gas emission value 408 using the following expressions (4) and (5) to generate the decomposed greenhouse gas emission value 412:

[0074] CO 2(2km) =CO 2(1) +...+CO 2(n) (4)

[0075]

[0076] In the above expressions (4) and (5), n represents the number of emission sources within the grid (e.g., 2.25 by 2.25 km). 2 (pixels), CO 2(2km) The greenhouse gas emission value at the grid is 408, CO. 2(i) This represents the greenhouse gas emissions of 412 at emission source i, where CO... 2e(i) The initial emission estimate at emission source i is 406. In expression (5), CO... 2e(i) It can be determined based on the mapping relationship 376 as described above.

[0077] In some embodiments, the decomposition module 209 can identify emission sources with potentially falsified emission reports from one or more emission sources based on decomposed greenhouse gas emission values ​​associated with the emission source. The decomposition module 209 can provide notifications to prioritize on-site inspections of emission sources with potentially falsified emission reports.

[0078] Therefore, the decomposition process 400 disclosed herein can be used to help government agencies identify potential emission sources with falsified emission reports. By applying the decomposition process 400, government agencies can prioritize which areas or plants need to be inspected by the field measurement team. For example, if the difference between the greenhouse gas emission value 408 of the grid and the sum of the reported emission values ​​in the grid is less than a predetermined error threshold, then a field trip inspection in the grid may not be necessary. However, if the difference is equal to or greater than the predetermined error threshold, the emission source with the largest difference between its decomposed greenhouse gas emission value 412 and the reported emission value can be selected for priority field trip inspection. If the inspected or audited emission source is not fraudulent, its reported emission value can be verified and used to update the above expressions (4)-(5) and determine the next most suspicious emission source. At the same time, the verified reported emission value can also be used to improve the mapping relationship 376. To further improve the accumulation process 400 and accelerate fraud detection, ground-based measuring devices can also be installed at emission sources where the greenhouse gas emissions 412 they decompose differ significantly from their reported emissions (e.g., the difference between the decomposed greenhouse gas emissions and the reported emissions is greater than an error threshold). As a result, measurements from the ground-based measuring devices can be used to improve the monitoring of greenhouse gas emissions at emission sources.

[0079] In conjunction with the above Figure 2-4 This paper provides an exemplary operational procedure for monitoring greenhouse gas (e.g., CO2) emissions. Initially, xCO2 data from various satellite data sources associated with a region of interest (GIO) (e.g., OCO-2, OCO-3, GOSAT-1, GOSAT-2, and TANSAT satellites) are downloaded and fused to generate a fused xCO2 input dataset. The GIO can be divided into multiple grids (e.g., each grid can be 2 km × 2 km in size). Concurrent grid wind data (e.g., CFS v2), industrial gas data (e.g., Sentinel-5P atmospheric industrial gas data), land surface temperature data from Landsat (with a resolution of 100 m), and infrared thermal data (e.g., Sentinel-2 longwave infrared data with a resolution of 20 m) can also be downloaded. A Gaussian plume model can be used to fit the fused xCO2 input dataset to produce CO2 emission estimates. Alternatively, a Gaussian plume model can be used to fit the industrial gas data to produce industrial gas emission estimates.

[0080] In cases where the fused xCO2 input dataset does not meet the data criteria for the Gaussian plume model, the xCO2 enhancement process can be used to compute one or more model parameters (e.g., including the range of the xCO2 Gaussian plume). The Gaussian plume equations (1) and (2) can then be used to backcalculate CO2 emissions using one or more model parameters.

[0081] If the CO2 emission estimate for the region of interest is successful, the CO2 emission estimate can be paired with concurrent observations of methane, surface temperature, and infrared thermal data, and compiled into a training dataset. The training dataset can be used to develop generalized or location-specific mappings between CO2 emissions and one or more of the following: industrial gas emissions, land surface temperature, and shortwave infrared thermal signals. These mappings can be used to estimate CO2 emissions at locations where xCO2 data is unavailable.

[0082] In some cases, emission sources in the grid are clustered (e.g., closely located). The mapping relationship between CO2 emissions and one or more of surface temperature and shortwave infrared thermal signals can also be used to decompose the CO2 emission values ​​associated with the grid into decomposed CO2 emission values ​​for individual emission sources.

[0083] In addition, the mechanism for using on-site confirmed CO2 emission values ​​to improve mapping relationships and decomposition methods can be referenced as above. Figure 4 This is achieved through measures such as government agencies sending audit teams to inspect CO2 emissions at emission sources and installing more emission sensors at those sources, which allows for continuous improvement in CO2 emission monitoring.

[0084] It should be noted that in some application scenarios, industrial gas emissions 372 can be publicly disclosed and directly available to the public, and the emission estimation model 306 is not required (e.g., Figure 3A (As shown in the dashed box 399 of 3C) Industrial gas emissions 372 are derived from industrial gas observations 310. Therefore, during the establishment of the correlation database 215, the derivation of greenhouse gas emission estimates 308, and / or the decomposition of greenhouse gas emission values ​​408 410, it may not be necessary to... Figure 3A or Figure 3C The derivation process is shown in the dashed box 399. That is, during the establishment of the association database 215, the derivation of the greenhouse gas emission estimate 308 and / or the decomposition of the greenhouse gas emission value 408 410, publicly available and publicly accessible industrial gas emissions 372 can be used directly.

[0085] Figure 5 This is a flowchart of an exemplary method 500 for monitoring greenhouse gas emissions according to embodiments of the present disclosure. Method 500 may be implemented by system 201, particularly fusion module 205 and estimation module 207, and may include steps 502-506 as described below. Some steps may be optional to perform the disclosure provided herein. Furthermore, some steps may be performed simultaneously or in conjunction with... Figure 5 The different sequences shown will be executed.

[0086] In step 502, the fusion module 205 can receive multiple satellite observations associated with greenhouse gas emissions in the first region of interest from multiple satellite data sources.

[0087] In step 504, the fusion module 205 can fuse multiple satellite observations to generate a fused input dataset.

[0088] In step 506, estimation module 207 can use an emission estimation model based on the fused input dataset to generate a first emission estimate of greenhouse gases in the first region of interest.

[0089] Figure 6 This is a flowchart of another exemplary method 600 for monitoring greenhouse gas emissions according to embodiments of the present disclosure. Method 600 may be implemented by system 201, specifically including fusion module 205, enhancement module 206, estimation module 207, mapping module 208, and decomposition module 209, and may include steps 602-624 as described below. Some steps may be optional to perform the disclosure provided herein. Furthermore, some steps may be performed simultaneously or in conjunction with... Figure 6 The different sequences shown will be executed.

[0090] In step 602, the fusion module 205 can receive multiple satellite observations associated with the first region of interest from multiple satellite data sources.

[0091] In step 604, the fusion module 205 can fuse multiple satellite observations to generate a fused input dataset.

[0092] In step 606, enhancement module 206 can use industrial gas observations associated with the first region of interest to enhance the fusion input dataset to be input into the emission estimation model.

[0093] In step 608, estimation module 207 may use an emission estimation model based on the fused input dataset to generate a first emission estimate of greenhouse gases in a first region of interest. In some embodiments, the first region of interest may be divided into multiple grids. The first emission estimate of greenhouse gases in the first region of interest may include multiple greenhouse gas emission values ​​for each of the multiple grids.

[0094] In step 610, the mapping module 208 may associate the first greenhouse gas emission estimate with one or more first observations of one or more reference resources in the first region of interest to generate a training dataset.

[0095] In step 612, the mapping module 208 can store the training dataset in the associated database 215.

[0096] In step 614, the mapping module 208 can derive the mapping relationship between greenhouse gas emissions and one or more reference resources based on the association database 215.

[0097] In step 616, the estimation module 207 may obtain one or more second observations of one or more reference resources in a second region of interest from multiple satellite data sources, in which no satellite observations associated with greenhouse gases are available.

[0098] In step 618, the estimation module 207 may determine a second emission estimate of greenhouse gases in the second region of interest based on the mapping relationship and one or more second observations from one or more reference resources. In some embodiments, the second region of interest may be divided into multiple grids. The second emission estimate of greenhouse gases in the second region of interest may include multiple greenhouse gas emission values ​​for the multiple grids respectively.

[0099] In step 620, the decomposition module 209 can select a grid from either a first region of interest or a second region of interest. Greenhouse gas emission values ​​associated with the grid can be obtained from either a first or second greenhouse gas emission estimate.

[0100] In step 622, the decomposition module 209 can identify one or more emission sources within the grid.

[0101] In step 624, the decomposition module 209 can decompose the greenhouse gas emission values ​​associated with the grid to generate one or more decomposed greenhouse gas emission values ​​for one or more emission sources.

[0102] Another aspect of this disclosure relates to a non-transitory computer-readable medium storing instructions that, when executed, cause one or more processors to perform the methods described above. The computer-readable medium may include volatile or non-volatile, magnetic, semiconductor, magnetic tape, optical, removable, non-removable, or other types of computer-readable media or computer-readable storage devices. For example, as disclosed, the computer-readable medium may be a storage device or memory module storing computer instructions. In some embodiments, the computer-readable medium may be a disk or flash drive on which computer instructions are stored.

[0103] According to one aspect of the present invention, a method for monitoring greenhouse gas emissions is disclosed. Multiple satellite observations associated with greenhouse gas emissions in a first region of interest are received from multiple satellite data sources. The multiple satellite observations are fused to produce a fused input dataset. An emission estimation model is used to generate a first emission estimate of greenhouse gases in the first region of interest based on the fused input dataset.

[0104] In some embodiments, industrial gas observations associated with a first region of interest are used to enhance the input dataset to be fused into the emission estimation model.

[0105] In some embodiments, the input dataset used to enhance the fusion of industrial gas observations into the emission estimation model includes applying industrial gas observations to the emission estimation model to estimate one or more model parameters associated with the emission estimation model.

[0106] In some embodiments, using an emission estimation model to generate a first emission estimate of greenhouse gases in a first region of interest includes applying a fused input dataset to the emission estimation model to generate a first emission estimate of greenhouse gases based on one or more model parameters.

[0107] In some embodiments, the emission estimation model includes a Gaussian plume model. One or more model parameters include the range of the plume in the Gaussian plume model.

[0108] In some embodiments, a first greenhouse gas emission estimate is associated with one or more first observations from one or more reference resources in a first region of interest to generate a training dataset. The training dataset is stored in an associated database.

[0109] In some embodiments, one or more reference resources include at least one of industrial gas resources, surface temperature resources, or shortwave infrared heat sources. The one or more first observations include at least one of the following associated with the first region of interest: industrial gas observations, surface temperature observations, or shortwave infrared thermal observations.

[0110] In some embodiments, the mapping between greenhouse gas emissions and one or more reference resources is derived based on an association database.

[0111] In some embodiments, one or more second observations from one or more reference resources are obtained in a second region of interest, where no satellite observations associated with greenhouse gas emissions are available from multiple satellite data sources. A second estimate of greenhouse gas emissions in the second region of interest is determined based on this mapping and the one or more second observations from the one or more reference resources.

[0112] In some embodiments, the first region of interest includes a geographical area divided into multiple grids. A first greenhouse gas emission estimate within the first region of interest includes multiple greenhouse gas emission values ​​from each of the multiple grids.

[0113] In some embodiments, one or more emission sources are identified within one of a plurality of grids. The greenhouse gas emission values ​​associated with the grid are decomposed into one or more decomposed greenhouse gas emission values ​​for one or more emission sources.

[0114] In some embodiments, decomposing a greenhouse gas emission value associated with a grid into one or more decomposed greenhouse gas emission values ​​for one or more emission sources includes: for each emission source, obtaining one or more observations of one or more reference resources associated with the grid; determining an initial emission estimate of the greenhouse gas for the emission source using the one or more observations of the one or more reference resources based on a mapping relationship between the greenhouse gas emission and the one or more reference resources, such that one or more initial emission estimates of the greenhouse gas are generated for the one or more emission sources; and decomposing the greenhouse gas emission value associated with the grid into the one or more decomposed greenhouse gas emission values ​​based on the one or more initial emission estimates of the greenhouse gas for the one or more emission sources.

[0115] In some embodiments, emission sources with potentially falsified emission reports are identified from one or more emission sources based on decomposed greenhouse gas emission values ​​associated with the emission source. Notification is provided to prioritize on-site inspections of emission sources with potentially falsified emission reports.

[0116] In some embodiments, the first region of interest includes a geographic region divided into multiple grids. Fusing the multiple satellite observations to generate the fused input dataset includes: resampling each satellite observation into one or more resampled observations associated with one or more of the multiple grids; and for each of the multiple grids, determining the availability of resampled observations associated with that grid, and generating fused input values ​​associated with that grid based on the availability of the resampled observations associated with that grid.

[0117] In some embodiments, each satellite observation includes one or more initial observations associated with a first region of interest. Resampling each satellite observation to one or more resampled observations associated with one or more grids includes applying a geographic weighting method to generate one or more resampled observations associated with one or more grids based on one or more initial observations.

[0118] In some embodiments, the greenhouse gas includes CO2. Multiple satellite observations include multiple observations of the column-mean dry air mole fraction (xCO2) of CO2 in the atmosphere. The multiple xCO2 satellite observations are associated with a first region of interest and are obtained from multiple xCO2 data sources, respectively.

[0119] According to another aspect of the present invention, a system for monitoring greenhouse gas emissions is disclosed. The system includes a memory and a processor. The memory is configured to store instructions. The processor is coupled to the memory and configured to execute the instructions to perform a process including: receiving multiple satellite observations associated with greenhouse gas emissions in a first region of interest from multiple satellite data sources; fusing the multiple satellite observations to generate a fused input dataset; and generating a first emission estimate of the greenhouse gas in the first region of interest using an emission estimation model based on the fused input dataset.

[0120] In some embodiments, the method further includes using industrial gas observations associated with a first region of interest to enhance the input dataset to be fused into the emissions estimation model.

[0121] In some embodiments, in order to use industrial gas observations to enhance the fused input dataset to be input into the emission estimation model, the method further includes applying the industrial gas observations to the emission estimation model to estimate one or more model parameters associated with the emission estimation model.

[0122] In some embodiments, in order to use an emission estimation model to generate a first emission estimate of greenhouse gases in a first region of interest, the process further includes applying a fused input dataset to the emission estimation model to generate a first emission estimate of greenhouse gases based on one or more model parameters.

[0123] In some embodiments, the emission estimation model includes a Gaussian plume model. One or more model parameters include the range of the plume in the Gaussian plume model.

[0124] In some embodiments, the process further includes associating a first greenhouse gas emission estimate with one or more first observations from one or more reference resources in a first region of interest to generate a training dataset. The training dataset is stored in an association database.

[0125] In some embodiments, one or more reference resources include at least one of industrial gas resources, surface temperature resources, or shortwave infrared heat sources. The one or more first observations include at least one of the following associated with the first region of interest: industrial gas observations, surface temperature observations, or shortwave infrared thermal observations.

[0126] In some embodiments, the process further includes deriving a mapping relationship between greenhouse gas emissions and one or more reference resources based on an association database.

[0127] In some embodiments, the process further includes: obtaining one or more second observations of the one or more reference resources from the plurality of satellite data sources in a second region of interest, where no satellite observations associated with the greenhouse gas emissions are available in the second region of interest; and determining a second emission estimate of the greenhouse gas in the second region of interest based on the mapping relationship and the one or more second observations of the one or more reference resources.

[0128] In some embodiments, the first region of interest includes a geographical area divided into multiple grids. A first greenhouse gas emission estimate within the first region of interest includes multiple greenhouse gas emission values ​​from each of the multiple grids.

[0129] In some embodiments, the process further includes: identifying one or more emission sources within a grid of the plurality of grids; and decomposing the greenhouse gas emission value associated with the grid into one or more decomposed greenhouse gas emission values ​​for the one or more emission sources.

[0130] In some embodiments, to decompose a greenhouse gas emission value associated with a grid into one or more decomposed greenhouse gas emission values ​​for one or more emission sources, the method further includes: for each emission source from the one or more emission sources, obtaining one or more observations of one or more reference resources associated with the grid; determining an initial emission estimate of the greenhouse gas for the emission source using the one or more observations of the one or more reference resources based on a mapping relationship between the greenhouse gas emission and the one or more reference resources, such that one or more initial emission estimates of the greenhouse gas are generated for the one or more emission sources; and decomposing the greenhouse gas emission value associated with the grid into the one or more decomposed greenhouse gas emission values ​​based on the one or more initial emission estimates of the greenhouse gas for the one or more emission sources.

[0131] In some embodiments, the process further includes: identifying emission sources with potentially falsified emission reports from the one or more emission sources based on the decomposed greenhouse gas emission values ​​associated with the emission source; and providing notification to prioritize on-site inspections of emission sources with potentially falsified emission reports.

[0132] In some embodiments, the first region of interest includes a geographic area divided into multiple grids. The process of fusing the multiple satellite observations to generate the fused input dataset further includes: resampling each satellite observation into one or more resampled observations associated with one or more of the multiple grids; and for each of the multiple grids, determining the availability of the resampled observations associated with that grid, and generating fused input values ​​associated with that grid based on the availability of the resampled observations associated with that grid.

[0133] In some embodiments, each satellite observation includes one or more initial observations associated with a first region of interest. To resample each satellite observation into one or more resampled observations associated with one or more grids, the process further includes applying a geographic weighting method to generate one or more resampled observations associated with one or more grids based on one or more initial observations.

[0134] In some embodiments, the greenhouse gas includes CO2. Multiple satellite observations include multiple observations of the column-mean dry air mole fraction (xCO2) of CO2 in the atmosphere. The multiple xCO2 satellite observations are associated with a first region of interest and are obtained from multiple xCO2 data sources, respectively.

[0135] According to another aspect of this disclosure, a non-transitory computer-readable storage medium is disclosed. The computer-readable storage medium is configured to store instructions that, in response to execution by a processor, cause the processor to perform a process comprising: receiving, respectively, multiple satellite observations associated with greenhouse gas emissions in a first region of interest from multiple satellite data sources; fusing the multiple satellite observations to generate a fused input dataset; and, based on the fused input dataset, using an emissions estimation model to generate a first emissions estimate of the greenhouse gas in the first region of interest.

[0136] The above description of a particular implementation can be readily modified and / or adapted to various applications. Therefore, based on the teachings and instructions presented herein, such adaptations and modifications are intended to be within the meaning and scope of equivalents of the disclosed implementation.

[0137] The breadth and scope of this disclosure should not be limited by any of the exemplary implementations described above, but should be defined solely by the appended claims and their equivalents.

Claims

1. A method for monitoring greenhouse gas emissions, characterized in that, The method includes: Multiple satellite observations associated with greenhouse gas emissions in the first region of interest were received from multiple satellite data sources. The multiple satellite observations are fused to generate a fused input dataset; and The first emission estimate of the greenhouse gas in the first region of interest is generated using the emission estimation model based on the fused input dataset; The first emission estimate of the greenhouse gas is correlated with one or more first observations of one or more reference resources in the first region of interest to generate a training dataset; The training dataset is stored in an associated database; Based on the associated database, the mapping relationship between the greenhouse gas emissions and the one or more reference resources is derived; Obtain one or more second observations of the one or more reference resources from the plurality of satellite data sources in a second region of interest, where no satellite observations associated with the greenhouse gas emissions are available in the second region of interest; and A second emission estimate of the greenhouse gas in the second region of interest is determined based on the mapping relationship and the one or more second observations from the one or more reference resources.

2. The method according to claim 1, characterized in that, The method also includes: The fused input dataset is augmented using industrial gas observations associated with the first region of interest.

3. The method according to claim 2, characterized in that, The step of using industrial gas observations associated with the first region of interest to enhance the fused input dataset to be input into the emission estimation model includes: The industrial gas observations are applied to the emission estimation model to estimate one or more model parameters associated with the emission estimation model.

4. The method according to claim 3, characterized in that, The step of generating the first emission estimate of the greenhouse gas in the first region of interest based on the fused input dataset using an emission estimation model includes: The fused input dataset is applied to the emission estimation model to generate a first emission estimate of the greenhouse gas based on one or more model parameters.

5. The method according to claim 3, characterized in that, The emission estimation model includes a Gaussian plume model, wherein one or more model parameters include the range of the plume in the Gaussian plume model.

6. The method according to claim 1, characterized in that, The one or more reference resources include at least one of industrial gas resources, surface temperature resources, or shortwave infrared heat sources, wherein the one or more first observations include at least one of the following associated with the first region of interest: industrial gas observation, surface temperature observation, or shortwave infrared thermal observation.

7. The method according to any one of claims 1, characterized in that, The first region of interest includes a geographical area divided into multiple grids, and the first emission estimate of the greenhouse gas in the first region of interest includes multiple greenhouse gas emission values ​​for each of the multiple grids.

8. The method according to claim 7, characterized in that, The method also includes: Identify one or more emission sources within a grid of the plurality of grids; and The greenhouse gas emissions associated with the grid are decomposed into one or more decomposed greenhouse gas emissions from the one or more emission sources.

9. The method according to claim 8, characterized in that, The step of decomposing the greenhouse gas emission values ​​associated with the grid into one or more decomposed greenhouse gas emission values ​​from the one or more emission sources includes: For each emission source from the one or more emission sources, obtain one or more observations of one or more reference resources associated with the grid; Based on the mapping relationship between the greenhouse gas emissions and the one or more reference resources, the one or more observations of the one or more reference resources are used to determine an initial emission estimate of the greenhouse gas from the emission source, thereby generating one or more initial emission estimates of the greenhouse gas for the one or more emission sources; and Based on one or more initial emission estimates of the gas from the one or more emission sources, the greenhouse gas emission value associated with the grid is decomposed into the one or more decomposed greenhouse gas emission values.

10. The method according to claim 8, characterized in that, The method also includes: Based on the decomposed greenhouse gas emission values ​​associated with the emission sources, identify emission sources with potentially falsified emission reports from the one or more emission sources; and Provide notification to prioritize on-site inspections of emission sources with potentially falsified emission reports.

11. The method according to claim 1, characterized in that, The first region of interest comprises a geographic area divided into multiple grids, and the multiple satellite observations are fused to generate a fused input dataset, including: Each satellite observation is resampled into one or more resampled observations associated with one or more of the multiple grids; and For each of the plurality of grids Determine the availability of resampled observations associated with the grid; and The fused input value associated with the grid is generated based on the availability of resampled observations associated with the grid.

12. The method according to claim 11, characterized in that, Each satellite observation includes one or more initial observations associated with the first region of interest, and Resampling each satellite observation into one or more resampled observations associated with the one or more grids includes: A geographic weighting method is applied based on the one or more initial observations to generate the one or more resampled observations associated with the one or more grids.

13. The method according to any one of claims 1-12, characterized in that, The greenhouse gases include carbon dioxide (CO2), and The multiple satellite observations include multiple column-averaged dry air mole fraction (xCO2) observations of CO2 in the atmosphere, which are associated with the first region of interest and obtained from multiple xCO2 data sources.

14. A system for monitoring greenhouse gas emissions, characterized in that, The system includes: Memory configured to store instructions; and A processor is connected to the memory and configured to implement the method as described in any one of claims 1 to 13 when executing the instructions.

15. A non-transitory computer-readable storage medium for storing program instructions, characterized in that, When the program instructions are executed by at least one processor, they implement the method as described in any one of claims 1 to 13.

Images