Ozone space source analysis method and device based on multi-dimensional feature fusion and readable storage medium thereof

Abstract

The application provides an ozone spatial source analysis method and device based on multi-dimensional feature fusion and a readable storage medium thereof. The method aims to solve the problems that the prior art cannot analyze the pollution source category attribution, the precursor chemical path and the spatial transport distribution in a unified framework, and the analysis result is difficult to adapt to the business evaluation index. The method determines a target period corresponding to the ozone daily maximum continuous specified time length sliding average value by acquiring source response, precursor sensitivity and spatial weight field data, splits the contribution of each source category into chemical path contributions of different precursor paths according to the sensitivity, performs secondary normalization processing on the spatial weight field in the target region by using a static regional mask, projects each chemical path contribution to the normalized spatial weight field and superimposes them to generate a high spatiotemporal resolution ozone spatial source distribution field with chemical and physical meanings. The application can be used for quantitative spatial source analysis of regional ozone pollution joint prevention and control.

Description

Technical Field

[0001] This invention relates to the field of atmospheric environmental numerical modeling and pollution source analysis technology, and in particular to a method, apparatus and readable storage medium for ozone spatial source analysis using multi-dimensional feature fusion. Background Technology

[0002] Ozone pollution has significant secondary formation characteristics. Its formation process is influenced by precursor emissions, regional transport, photochemical reactions, and meteorological conditions. Accurate ozone source apportionment is an important prerequisite for formulating regional joint prevention and control policies.

[0003] Existing ozone source apportionment technologies mainly fall into the following three categories:

[0004] The first category is chemical labeling methods based on numerical models, such as CMAQ-ISAM (Comprehensive Source Apportionment Module). This method can track the overall contribution of different pollution source categories to ozone concentration, but the output is a scalar response value at the site scale, which cannot reflect the precursor driving type corresponding to the contribution.

[0005] The second category is precursor attribution methods based on sensitivity analysis, such as DDM (Direct Decoupling Method) or chemical box model. These methods can quantitatively describe the nonlinear response of ozone generation to disturbances of nitrogen oxides and volatile organic compounds, but their output results are usually limited to the scalar sensitivity of specific sites and are difficult to directly convert into continuous spatial grid products suitable for regional collaborative management.

[0006] The third category is source tracing methods based on spatial trajectories, such as spatial weighting techniques like BackSA (Backward Sensitivity Analysis). These methods can describe the spatial transport paths of pollutants, but due to the lack of explicit constraints on secondary photochemical reactions, if directly applied to ozone source apportionment, they will be difficult to reflect the differentiated chemical generation mechanisms of nitrogen oxides and volatile organic compounds during transport.

[0007] In recent years, some improved technologies have attempted to combine the above methods, but most of them are simply sequential calls and have failed to establish an inherent coupling relationship between pollution source classification, precursor chemical pathways and spatial transport distribution within a unified framework.

[0008] Therefore, there is an urgent need for a method, apparatus, and readable storage medium for ozone spatial source analysis that integrates multi-dimensional features, in order to solve the problems existing in the prior art. Summary of the Invention

[0009] This invention provides a method, apparatus, and readable storage medium for ozone spatial source analysis based on multi-dimensional feature fusion. It addresses the problems of existing technologies, which struggle to achieve comprehensive analysis of the total contribution of pollution source categories, precursor chemical driving pathways, and high-resolution continuous spatial transport distribution within a unified framework, and which easily lead to the dilution of peak pollution period causal analysis by background period data.

[0010] The core technology of this invention is to extract the source category response contribution by dynamically locking the ozone peak pollution time window, and to decouple it into different precursor path contributions by utilizing the precursor sensitivity nonlinear characteristics. Then, combined with the secondary normalized spatial weight field under the target area mask constraint, the high-precision spatial projection and superposition synthesis of multiple precursor driving paths on the meteorological transport grid is realized.

[0011] In a first aspect, the present invention provides a method for analyzing the spatial sources of ozone through multi-dimensional feature fusion, the method comprising the following steps: Acquire ozone source response data, precursor sensitivity data, spatial weight field data corresponding to different precursors, and static region mask data of stations within the target area; The target time period is determined, and the source response contribution of each preset pollution source category is extracted from the ozone source response data based on the target time period. Based on the precursor sensitivity data, the source response contribution of each source is broken down into the chemical pathway contribution of different precursor pathways. The spatial weight field data of the corresponding precursor is normalized twice within the region using static region mask data to obtain the normalized spatial weight field. The contributions of each chemical pathway are projected onto the corresponding normalized spatial weight field, and the spatial source distribution field of ozone at the site during the target time period is generated by spatial superposition.

[0012] Furthermore, the steps for determining the target time period include: The daily maximum continuous specified duration moving average of the ozone concentration sequence of the station is calculated based on the sliding window, and the continuous time window corresponding to the daily maximum continuous specified duration moving average is determined as the target time period.

[0013] Furthermore, the different precursor pathways include a first precursor pathway and a second precursor pathway; the step of decomposing the response contributions from each source into chemical pathway contributions from different precursor pathways includes: Extract the sensitivity coefficients of the first and second precursors for each preset pollution source category within the target time period; When the sum of the sensitivity coefficients of the first precursor and the second precursor is greater than the set minimum threshold, the source response contribution is proportionally divided into the first precursor path contribution and the second precursor path contribution according to the proportion of each sensitivity coefficient in the sum of sensitivity coefficients. When only a single precursor has a sensitivity coefficient greater than the minimum threshold, the source response contribution is fully assigned to the chemical pathway contribution corresponding to that single precursor's sensitivity coefficient.

[0014] Furthermore, the step of performing a second normalization process within the region on the spatial weight field data of the corresponding precursor using static region mask data includes: Based on static region mask data, the grid weights in the spatial weight field data located outside the target region indicated by the static region mask data are forcibly set to zero; Using the sum of the weights of the remaining grids within the target area as the denominator, the quadratic normalized weight ratio of each grid within the target area is recalculated.

[0015] Furthermore, after the step of generating the ozone spatial source distribution field of the site during the target time period, the method further includes: The ozone spatial source distribution field is generated by dividing the ozone spatial source distribution field by the peak ozone concentration during the target period.

[0016] Furthermore, following the step of generating the spatial source distribution field of ozone, the following steps are also included: Based on the pre-defined multi-scale geographic mapping relationship, the spatial distribution field of ozone sources is spatially aggregated according to administrative boundaries to generate regional-level ozone contribution summary data for each pre-defined pollution source category.

[0017] Furthermore, it also includes: The computation status of multi-site parsing tasks is monitored in real time. When the result file of any target site is missing or the underlying data is abnormal, a secondary recalculation process for that site is automatically triggered.

[0018] Secondly, the present invention provides a multi-dimensional feature fusion ozone spatial source analysis device, comprising: The data acquisition module is used to acquire ozone source response data, precursor sensitivity data, spatial weight field data corresponding to different precursors, and static area mask data of stations within the target area. The chemical decomposition module is used to determine the target time period, extract the source response contribution of each preset pollution source category from the ozone source response data based on the target time period, and decompose each source response contribution into chemical path contribution of different precursor pathways based on precursor sensitivity data. The spatial fusion module is used to perform secondary normalization processing on the spatial weight field data of the corresponding precursors using static regional mask data to obtain a normalized spatial weight field; and to project the contributions of each chemical pathway onto the corresponding normalized spatial weight field, thereby generating the ozone spatial source distribution field of the site in the target time period through spatial overlay.

[0019] Thirdly, the present invention provides an electronic device including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to execute the above-described method for ozone spatial source analysis by multidimensional feature fusion.

[0020] Fourthly, the present invention provides a readable storage medium storing a computer program, the computer program including program code for controlling a process to execute the process, the process including the ozone spatial source analysis method based on the above-described multidimensional feature fusion.

[0021] The main contributions and innovations of this invention are as follows: 1. Complete and unified analytical dimensions: This invention constructs a three-layer coupled framework of source category response, precursor decomposition and spatial mapping, which integrates pollution source category attributes, chemical precursor attributes and spatial transport attributes into a unified spatiotemporal analytical result, effectively solving the problem that traditional single models can only provide scalar contributions or lack precursor driving information.

[0022] 2. Significantly improved attribution accuracy for peak pollution processes: This invention uses the maximum continuous specified duration (e.g., 8 hours) of the day based on dynamic identification of sliding windows as the core evaluation time window, replacing the traditional fixed time period or daily average calculation method. This effectively eliminates the numerical dilution interference of background concentration at night or during inactive periods, and truly restores the driving causes of peak photochemical pollution processes.

[0023] 3. Spatial source resolution and local mapping accuracy are greatly improved: By introducing a static region mask to perform secondary normalization processing on the spatial transport weight field within the region, the numerical dissipation of the high background grid outside the target area is forcibly cut off, avoiding the local transport contribution being masked by the wide-area background. The generated percentage spatial contribution field can more sensitively and accurately reflect the actual spatial transport and contribution distribution of fine-grained grids within the region.

[0024] 4. Strong automation and adaptability of business systems: This invention adopts a standard multi-dimensional grid format to output results, supports the unified aggregation of spatial features at multiple scales such as stations, cities and regions, and has a built-in automatic scheduling and supplementary calculation mechanism for missing stations and data anomalies. It can be seamlessly adapted to and stably run in the routine air quality forecasting and consultation business system at the provincial and large-scale regional levels.

[0025] Details of one or more embodiments of the present invention are set forth in the following drawings and description, so that other features, objects and advantages of the invention will be more readily understood. Attached Figure Description

[0026] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a flowchart illustrating the overall process of the ozone spatial source apportionment method based on source labeling-precursor sensitivity-spatial weight fusion according to an embodiment of this application. Figure 2 This is a flowchart of ozone peak period identification and precursor pathway decomposition according to an embodiment of this application, corresponding to steps S2 to S4; Figure 3 This is a flowchart of spatial weighted region normalization and contribution projection according to an embodiment of this application, corresponding to steps S5 to S6; Figure 4 This is a schematic diagram of the ozone spatial source apportionment results in Chengdu on September 30, 2023, according to an embodiment of this application. Figure 5 This is a block diagram of the ozone spatial source apportionment system according to an embodiment of this application; Figure 6 This is a schematic diagram illustrating the dimensional transformation logic for converting scalar contributions to spatial contribution fields according to an embodiment of this application. Figure 7 This is a timing diagram of multi-source data interaction and task scheduling during the system execution process according to an embodiment of this application; Figure 8 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0027] 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 numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with one or more embodiments of this specification. Rather, they are merely examples of apparatuses and methods consistent with some aspects of one or more embodiments of this specification as detailed in the appended claims.

[0028] It should be noted that the terms "first" and "second" in the specification and claims of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. Unless otherwise specified, the numerical model output data used in this embodiment (such as CMAQ-ISAM source labeling results, DDM sensitivity results, and BackSA spatial weight field) are all output products of atmospheric chemical transport models or adjoint models commonly used in air quality forecasting operational systems. Their specific calculation principles are well known to those skilled in the art and will not be elaborated here.

[0029] Example 1 This embodiment provides a spatial source apportionment method for ozone (O3) based on multi-dimensional feature fusion. Logically, this method constructs a three-layer coupled analytical framework of "source response-precursor sensitivity-spatial weight". The discrete pollution source category scalar contribution is decomposed through precursor chemical pathways and finally transformed into a high-resolution gridded ozone source distribution product that has both chemical mechanism attributes and spatial geographic attributes.

[0030] See Figure 1 The specific execution process of this method includes the following steps: S1. Data Acquisition and Initialization (corresponding to data construction and initialization steps) The system is based on the business reporting start date. Determine the set of target dates for future forecasts For each target station within the target region (e.g., the Southwest region), the following data is read from the multi-source numerical model output results through the data adaptation module: Source-labeled result sets containing O3 scalar responses (e.g., result files using CMAQ-ISAM), sensitivity analysis result sets containing precursor sensitivities (e.g., result files using DDM), and result sets containing nitrogen oxides (NO3) x Spatial weight field data for the transport weights of two types of precursors, namely volatile organic compounds (VOCs) (e.g., spatial weight field output by BackSA), and static area mask data M(x, y) that defines the mapping relationship between grid attributes and cities.

[0031] S2. Target time period determination (corresponding peak time period identification steps) In this embodiment, the aforementioned target time period specifically refers to the peak pollution time window corresponding to the maximum daily 8-hour moving average concentration of ozone.

[0032] See Figure 2 For each target site s and target day d, the system calculates the hourly ozone concentration sequence for site s based on a sliding time window algorithm, and extracts the maximum 8-hour ozone concentration for the day. Its definition is:

[0033] in, This represents the rolling 8-hour average ozone concentration sequence value of site s at time t.

[0034] Obtain the termination time corresponding to the maximum value. This timeframe forms an 8-hour peak pollution window, which is then traced back from that moment. :

[0035] This time window will serve as a unified time reference for all extraction and splitting operations in subsequent steps S3 and S4. This peak pollution time window typically corresponds to the afternoon period of active photochemical reactions, effectively eliminating the dilution interference of background concentrations during inactive periods such as nighttime on the analysis results.

[0036] S3. Source Category Response Extraction (corresponding source category response extraction steps) See also Figure 2 Based on the peak pollution time window determined in step S2 Extract the average ozone response contribution of each preset pollution source category k to the target site s within the time window from the source labeling results set. The preset pollution source categories include, but are not limited to, multiple sources such as transportation, industry, power generation, agriculture, residential, solvent use, dust, and biomass combustion.

[0037] Specifically, the average response contribution of pollution source category k within a time window Calculated using the following formula:

[0038] in, This refers to the number of hours within the peak pollution time window (i.e., 8). The contribution of source category k to ozone concentration at station s at time t in the source labeling results.

[0039] S4. Chemical pathway decoupling (corresponding to the multi-path precursor separation step) In this step, the different precursor pathways mentioned above specifically include the first precursor pathway (nitrogen oxide pathway) and the second precursor pathway (volatile organic compound pathway).

[0040] See also Figure 2 The system extracts NO from each preset pollution source category within the target time period. x Sensitivity mean VOC sensitivity mean Then, the total response contribution is calculated according to the following logical rules. Split into NO x Path contribution and VOC pathway contribution : Rule 1: When the mean NOx sensitivity VOC sensitivity mean The sum of the two is greater than a pre-set minimum threshold. When (the minimum threshold) Desirable (or smaller, to avoid division by zero errors in floating-point operations), indicating that both sensitivities are simultaneously effective. In this case, the total response contribution is calculated according to the proportion of each sensitivity coefficient in the sum of the two. Proportionally split into NO x Path contribution and VOC pathway contribution :

[0041]

[0042] Rule 2: When only one sensitivity coefficient is greater than the minimum threshold (i.e., another sensitivity coefficient less than or equal to) When this occurs, it indicates that only a single precursor exists within that time window, indicating a sensitive response. In this case, the contribution of the total response is... The full amount of contribution of the chemical pathway corresponding to the single effective sensitivity coefficient is assigned, while the contribution of the other pathway is set to zero.

[0043] Rule 3: When both sensitivity coefficients are less than or equal to the minimum threshold If the result indicates that neither of the two sensitivity responses is significant, then the contributions of both chemical pathways are set to zero.

[0044] This decomposition process endows the total contribution of each pollution source category with a clear chemical mechanism attribute, laying the foundation for accurate matching with the spatial weight field of the corresponding precursors.

[0045] S5. Quadratic normalization and projection within the region (corresponding to spatial weight normalization and projection steps) The core of this step is to use static region mask data to perform secondary normalization of the spatial weight field within the target region to obtain a normalized spatial weight field, and then project the path contribution obtained in S4 onto the normalized mesh.

[0046] See Figure 3 and Figure 6 The purpose of this operation is to transform the original spatial weight field, which covers the entire simulation domain, into a localized distribution that is effective only within the target area of ​​interest through target area constraints. This prevents the high background weight values ​​outside the target area (e.g., areas with no actual transport significance, such as the Qinghai-Tibet Plateau, for the analysis of the southwest region) from having a numerical dilution effect on the local spatial contribution distribution within the target area.

[0047] First, the system reads the NOx pathspace weight field of preset key business time periods (e.g., time periods 12Z and 18Z), takes the average value, and obtains the merged NOx... x Weight field Similarly, the merged VOC weight field is obtained. .

[0048] Then, a second normalization process is performed within the target region. Specifically, a predefined static region mask matrix M(x,y) is obtained, which takes a value of 1 when the grid coordinates (x,y) are within the defined target region, and 0 otherwise. The weight field after the second normalization process is then applied. Defined by the following formula:

[0049] The denominator is the summation of all (x,y) grids within the simulation domain.

[0050] This formula ensures that the processed weight field is non-zero only within the target region, and that the sum of the weights of all grid cells within the target region is always equal to 1. For the VOC weight field... Perform the exact same operation to obtain the normalized VOC weight field. .

[0051] Next, the NO obtained from step S4 is split... x Path contribution and VOC pathway contribution As amplification factors, they are respectively related to the corresponding normalized spatial weight fields. and Multiplying these results in the spatial contribution distribution of the source category k across each grid within the target region.

[0052] S6. Spatial superposition and synthesis (corresponding to the spatial contribution synthesis step) See also Figure 3 For source category k, set its NO x The spatial contribution distribution of the path and the spatial contribution distribution of the VOC path are superimposed on each grid coordinate (x, y) to generate the absolute spatial source contribution field of ozone for source category k on target day d to site s. :

[0053] The contribution fields of all preset pollution source categories k are summed to generate the total ozone spatial source distribution field of site s on target day d. :

[0054] Furthermore, the spatial source distribution field of total ozone Each grid contribution value is divided by the peak daily 8-hour ozone concentration determined in step S2. The spatial distribution field of ozone generation percentage :

[0055] in, The sum of the absolute spatial contribution fields of all pollution source categories satisfies This percentage field visually illustrates the relative contribution of emissions from different spatial locations to the ozone peak at each site.

[0056] S7. Multi-scale summarization and standardized output (corresponding steps for multi-scale summarization and standardized output) To meet the business needs of different management levels, this step performs multi-scale data aggregation and output in a standard format.

[0057] On the one hand, based on pre-set multi-scale spatial mapping relationships such as urban mapping tables, the gridded spatial source distribution field of ozone percentage is... Spatial aggregation is performed based on administrative boundaries such as cities. For a given city's geographical boundary c, all grid cells falling within that boundary are counted. The sum of the percentage contribution values ​​is used to generate a city-level summary of ozone contribution data for each preset pollution source category. :

[0058] This aggregated data can be output in structured tables (such as CSV format) to directly serve the diagnosis of pollution causes at the city scale.

[0059] On the other hand, the spatial source distribution of ozone percentage at each site will be analyzed. Write the standard multidimensional region grid file conforming to the NetCDF specification using a four-dimensional array structure (dimensions may include time, station, longitude, and latitude). This standardized format file facilitates subsequent analysis and product mapping on common computing and visualization platforms such as GrADS, NCL, and Python.

[0060] S8. Anomaly detection and automatic recalculation (corresponding to anomaly detection and product packaging steps) This step aims to ensure the integrity of tasks and the robustness of products for large-scale, multi-site business operations.

[0061] See Figure 7 The system's task scheduling and stability management layer monitors the entire computational process of multi-site parsing tasks in real time. When a missing standardized grid file is detected for any target site, or when anomalies in underlying data reading or intermediate variable calculations are captured in the computation log, the system automatically intercepts the anomaly signal and triggers a secondary recalculation process for the faulty site. This recalculation task can be executed through an independent computation thread without affecting the normal processing of other sites, thereby ensuring the completeness of site results in the final business distribution product.

[0062] Specific application scenario analysis: Taking Chengdu business simulation as an example To verify the operational feasibility and computational rigor of the method of this invention, a specific quantitative calculation verification process is given, taking the actual air quality forecast scenario of Chengdu, Southwest China on September 30, 2023 as an example.

[0063] Scenario setting: The target receptor is a set of stations in the main urban area of ​​Chengdu, the target business date is September 30, 2023, and the forecast and analysis indicator is the daily maximum 8-hour moving average O3 concentration.

[0064] After data loading is completed in step S1, the system proceeds to step S2. Using a sliding window algorithm, the system iterates through the hourly forecast concentration sequence for the day to identify the maximum 8-hour ozone concentration in Chengdu for that day. 100.9 μg / m 3 The corresponding peak pollution time window is from 10:00 to 17:00 on the same day (8 consecutive hours), which happens to cover the afternoon period when photochemical reactions are most active.

[0065] Proceeding to step S3, taking traffic sources (TR) and industrial sources (IN) as examples, extract the total response contribution of traffic sources within the time window from 10:00 to 17:00. 15.0 μg / m 3 Total industrial source response contribution 12.4 μg / m 3 The contribution values ​​of the remaining source categories are extracted in the same manner.

[0066] Proceed to step S4 to perform precursor analysis. Continuing with the example of traffic sources, extract their NOx sensitivity within the time window. The VOC sensitivity is 0.40. The value is 0.10. Since the sum of the two, 0.50, is much larger than the minimum threshold (e.g., ...), ... Rule 1 applies, and the result is obtained by proportional division:

[0067]

[0068] For industrial sources and other source categories, the same logical rules are applied to complete the splitting, resulting in their respective dual-path contribution values.

[0069] Proceed to step S5, first obtain the NO for the southwest region. xThe VOC spatial weight field is then applied, followed by the application of pre-configured static mask data for the southwest region (this mask has removed transport-irrelevant areas such as the Qinghai-Tibet Plateau). Based on the static regional mask, the weight values ​​of the grids outside the target region are forcibly set to zero. A second normalization is then performed within the target region using the sum of the weights of the effective grids as the denominator, resulting in a locally normalized weight field. and Subsequently, the NOx path contributions and VOC path contributions of each source category obtained from the previous step are projected onto the corresponding normalized weight fields.

[0070] Proceeding to step S6, the two-path spatial contributions of each source category are superimposed on the grid to generate the absolute spatial source distribution field of ozone for the corresponding day at each site, G(x,y). This field is then divided by 100.9 μg / m³. 3 The peak concentration is converted into the percentage spatial source distribution field P(x,y).

[0071] Figure 4 This diagram illustrates the spatial source distribution of ozone percentage in Chengdu on September 30, 2023, generated according to this embodiment. Clear spatial source analysis conclusions can be drawn from this diagram: high-value grids contributing to the ozone peak in Chengdu on that day are highly concentrated in the main urban area and the northeastern border region, with the largest single grid contribution rate being approximately 2.4%; medium-contribution grids extend northeastward into Deyang and Mianyang, with a small distribution also in Meishan in the south. The direction of contribution distribution is consistent with the transport direction of the northerly wind field on that day, providing a refined quantitative spatial basis for joint prevention and control of air pollution in Chengdu, Deyang, and Mianyang. Furthermore, the sum of the resolvable portion (approximately 37.1%) of the spatial field of all source categories after deducting the background conforms to the mass conservation constraint, verifying the numerical stability of the computational framework.

[0072] The complete calculation process and output of this embodiment demonstrate that the method of the present invention can successfully transform discrete source category scalar responses into a continuous spatiotemporal contribution grid product containing chemically driven paths and highly visualized. Its output not only removes the background concentration dilution effect but also accurately identifies key transport channels and key contribution grids.

[0073] Example 2 See Figure 5 This embodiment provides an ozone spatial source apportionment system based on multi-dimensional feature fusion to implement the above method. The system is logically divided into multiple functional modules, the functions of which are described below: The data acquisition module is used to perform the aforementioned step S1, and is responsible for loading CMAQ-ISAM, DDM, BackSA and static region mask data, and completing format conversion and coordinate correction.

[0074] The chemical separation module is used to perform the aforementioned steps S2 to S4, dynamically calculating and determining the target time period (peak pollution time window) corresponding to the maximum daily 8-hour ozone concentration. Based on this time period, it extracts the source response contribution of each preset pollution source category, and separates the source response contribution into NO based on the precursor sensitivity coefficient ratio. x Dual-path contribution with VOC.

[0075] The spatial fusion module is used to execute the aforementioned steps S5 to S6, including region masking, spatial weight normalization, and meshed spatial projection calculation, to achieve dimensional transformation from scalar response to spatial field (see [link to module name]). Figure 6 (Logic), and overlay the generated ozone spatial source distribution field of the site within the target time period.

[0076] The multi-scale output and scheduling module is used to execute the aforementioned steps S7 to S8, to perform multi-scale data aggregation and standardized file writing (such as NetCDF / CSV / PNG), and is responsible for monitoring the status of the entire task process and automatically supplementing missing sites.

[0077] Example 3 This embodiment also provides an electronic device, see reference. Figure 8 It includes a memory 404 and a processor 402, wherein the memory 404 stores a computer program and the processor 402 is configured to run the computer program to perform the steps in any of the above method embodiments.

[0078] Specifically, the processor 402 may include a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement embodiments of the present invention.

[0079] Memory 404 may include a mass storage device for data or instructions. For example, and not limitingly, memory 404 may include a hard disk drive (HDD), a floppy disk drive, a solid-state drive (SSD), flash memory, an optical disk drive, a magneto-optical disk drive, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 404 may include removable or non-removable (or fixed) media. Where appropriate, memory 404 may be located internally or externally to a data processing device. In a particular embodiment, memory 404 is non-volatile memory. In a particular embodiment, memory 404 may include read-only memory (ROM) and random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable read-only memory (PROM), an erasable read-only memory (EPROM), an electrically erasable read-only memory (EEPROM), an electrically alterable read-only memory (EAROM), or flash memory, or a combination of two or more of these. Where appropriate, the RAM can be Static Random-Access Memory (SRAM) or Dynamic Random-Access Memory (DRAM). DRAM can be Fast Page Mode Dynamic Random Access Memory (FPMDRAM), Extended Data Out Dynamic Random Access Memory (EDODRAM), Synchronous Dynamic Random-Access Memory (SDRAM), etc.

[0080] The memory 404 can be used to store or cache various data files that need to be processed and / or communicated, as well as possible computer program instructions executed by the processor 402.

[0081] The processor 402 reads and executes computer program instructions stored in the memory 404 to implement any of the multi-dimensional feature fusion ozone spatial source analysis methods in the above embodiments.

[0082] Optionally, the electronic device may further include a transmission device 406 and an input / output device 408, wherein the transmission device 406 is connected to the processor 402, and the input / output device 408 is connected to the processor 402.

[0083] Transmission device 406 can be used to receive or send data over a network. Specific examples of the network described above may include wired or wireless networks provided by the communication provider of the electronic device. In one example, the transmission device may include a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In one example, transmission device 406 may be a Radio Frequency (RF) module used for wireless communication with the Internet.

[0084] Input / output device 408 is used to input or output information.

[0085] Example 4 This embodiment also provides a readable storage medium storing a computer program, the computer program including program code for controlling a process to execute the process, the process including an ozone spatial source analysis method based on multidimensional feature fusion according to Embodiment 1.

[0086] It should be noted that the specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementations, and will not be repeated here.

[0087] Generally, various embodiments can be implemented in hardware or dedicated circuitry, software, logic, or any combination thereof. Some aspects of the invention can be implemented in hardware, while others can be implemented by firmware or software executed by a controller, microprocessor, or other computing device, but the invention is not limited thereto. Although various aspects of the invention may be shown and described as block diagrams, flowcharts, or using some other graphical representation, it should be understood that, by way of non-limiting example, these blocks, apparatuses, systems, techniques, or methods described herein can be implemented in hardware, software, firmware, dedicated circuitry or logic, general-purpose hardware or controllers or other computing devices, or some combination thereof.

[0088] Embodiments of the present invention can be implemented by computer software that can execute in a mobile device's data processor, such as a processor entity, or by hardware, or by a combination of software and hardware. Computer software or programs (also referred to as program products), including software routines, applets, and / or macros, can be stored in any device-readable data storage medium and include program instructions for performing specific tasks. The computer program product may include one or more computer-executable components configured to perform the embodiments when the program is run. The one or more computer-executable components may be at least one piece of software code or a portion thereof. Additionally, it should be noted in this respect that, as Figure 1 Any block in the logical flow can represent a program step, or interconnected logic circuits, blocks, and functions, or a combination of program steps and logic circuits, blocks, and functions. Software can be stored on physical media such as memory chips or blocks of storage implemented within a processor, magnetic media such as hard disks or floppy disks, and optical media such as DVDs and their data variants, CDs, etc. The physical media is non-transient.

[0089] Those skilled in the art should understand that the technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments have been described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0090] The above embodiments are merely illustrative of several implementations of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of the present invention should be determined by the appended claims.

Claims

1. A method for spatial source analysis of ozone based on multi-dimensional feature fusion, characterized in that, Includes the following steps: Acquire ozone source response data, precursor sensitivity data, spatial weight field data corresponding to different precursors, and static region mask data indicating the grid range within the target area for stations within the target area; A target time period is determined, and the source response contribution of each preset pollution source category is extracted from the ozone source response data based on the target time period. Based on the precursor sensitivity data, each source response contribution is split into chemical pathway contributions of different precursor pathways. Using the static region mask data, the spatial weight field data of the corresponding precursor is subjected to secondary normalization within the region to obtain a normalized spatial weight field. This includes: based on the static region mask data, resetting the weights of the grids located outside the target region in the spatial weight field data to zero, and using the sum of the weights of the remaining grids within the target region as the denominator to recalculate the secondary normalized weight ratio of each grid within the target region. The contributions of each chemical pathway are projected onto the corresponding normalized spatial weight field, and the ozone spatial source distribution field of the site in the target time period is generated by spatial superposition.

2. The method as described in claim 1, characterized in that, The steps for determining the target time period include: The daily maximum continuous specified duration sliding average of the ozone concentration sequence of the site is calculated based on a sliding window, and the continuous time window corresponding to the daily maximum continuous specified duration sliding average is determined as the target time period.

3. The method as described in claim 1, characterized in that, The different precursor pathways include a first precursor pathway and a second precursor pathway; The step of decomposing the contribution of each source response into chemical pathway contributions of different precursor pathways includes: Extract the first precursor sensitivity coefficient and the second precursor sensitivity coefficient for each of the preset pollution source categories within the target time period; When the sum of the sensitivity coefficients of the first precursor and the second precursor is greater than a set minimum threshold, the source response contribution is proportionally divided into the first precursor path contribution and the second precursor path contribution according to the proportion of each sensitivity coefficient in the sum of the sensitivity coefficients. When only a single precursor has a sensitivity coefficient greater than the minimum threshold, the source response contribution is fully assigned to the chemical pathway contribution corresponding to that single precursor sensitivity coefficient.

4. The method as described in claim 1, characterized in that, Following the step of generating the ozone spatial source distribution field of the site during the target time period, the method further includes: The ozone spatial source distribution field is divided by the peak ozone concentration during the target time period to generate the ozone percentage spatial source distribution field.

5. The method as described in claim 1, characterized in that, After the step of generating the ozone spatial source distribution field, the method further includes: Based on the preset multi-scale geographic mapping relationship, the ozone spatial source distribution field is spatially aggregated according to administrative boundaries to generate regional-level ozone contribution summary data for each preset pollution source category.

6. The method according to any one of claims 1 to 5, characterized in that, Also includes: The computation status of multi-site parsing tasks is monitored in real time. When the result file of any target site is missing or the underlying data is abnormal, a secondary recalculation process for that site is automatically triggered.

7. An ozone spatial source apportionment device based on multi-dimensional feature fusion, characterized in that, include: The data acquisition module is used to acquire ozone source response data, precursor sensitivity data, spatial weight field data corresponding to different precursors, and static area mask data indicating the grid range within the target area. The chemical decomposition module is used to determine the target time period, extract the source response contribution of each preset pollution source category from the ozone source response data based on the target time period, and decompose each source response contribution into chemical path contribution of different precursor pathways based on the precursor sensitivity data. The spatial fusion module is used to perform secondary normalization processing on the spatial weight field data of the corresponding precursor using the static region mask data to obtain a normalized spatial weight field. The module includes: based on the static region mask data, resetting the weights of the grids located outside the target region in the spatial weight field data to zero, and recalculating the secondary normalized weight ratio of each grid in the target region using the sum of the weights of the remaining grids in the target region as the denominator. The contributions of each chemical pathway are then projected onto the corresponding normalized spatial weight field, and the ozone spatial source distribution field of the site during the target time period is generated by spatial superposition.

8. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to run the computer program to perform the method according to any one of claims 1 to 6.

9. A readable storage medium, characterized in that, The readable storage medium stores a computer program, the computer program including program code for controlling a process to execute the process, the process including the method according to any one of claims 1 to 6.

Images