Method and system for population migration dynamic attribution assessment of health risks of air pollution

Abstract

The application discloses a population migration dynamic attribution assessment method and system for air pollution health risk, and belongs to the technical field of environmental health risk assessment and spatio-temporal big data analysis. The method firstly fuses multi-source spatio-temporal data and carries out standardization preprocessing; secondly, based on night light and built-up area statistical data, a provincial adaptive threshold algorithm is used to dynamically identify three types of spatial units, i.e. urban core area, expansion area and rural area; then, a "no migration" counterfactual population distribution scenario is constructed based on the population pattern in the base year; finally, the health risk in the real and counterfactual scenarios is calculated in parallel, and the net health risk caused by population migration is quantified by difference. The application effectively solves the exposure misclassification and attribution capability loss problems caused by the use of static population and fixed boundaries in the traditional assessment method, improves the spatio-temporal accuracy of the assessment and the calculability of the attribution result, and provides a standardized analysis tool for environmental health risk management.

Description

Technical Field

[0001] This invention relates to the field of environmental health risk assessment and spatiotemporal big data analysis technology. Specifically, it relates to an assessment method, system, and electronic device that integrates dynamic boundary identification and counterfactual attribution frameworks for quantitatively assessing the impact of population migration on the health risks of air pollution exposure. Background Technology

[0002] With the continued acceleration of global urbanization, large-scale and frequent population migration has become a core driving force reshaping regional population distribution patterns. This process has not only changed socioeconomic structures but has also directly led to a dramatic spatial restructuring of population air pollution exposure levels and health risks. However, faced with this dynamic and complex reality, existing air pollution health risk assessment technologies are encountering a series of systemic bottlenecks that urgently need to be overcome in achieving accurate and quantitative attribution analysis.

[0003] Current mainstream assessment methods are mostly based on static or periodically updated population distribution data, implicitly assuming that the spatial pattern of population remains stable during the assessment period. This static perspective fundamentally fails to capture the "spatial reconstruction of exposed populations" effect caused by rural-to-urban migration and cross-regional flows. Significant discrepancies may arise between the actual exposure levels and the estimates of traditional models in areas of population inflow and outflow, greatly reducing the reliability of risk assessments in rapidly urbanizing frontier areas. One root cause of this limitation lies in the rigid division of assessment spatial units. Many studies rely on fixed administrative divisions (such as provincial, municipal, and county boundaries) to mechanically distinguish between urban and rural areas, while the actual scope of urban built-up areas expands year by year. This outdated division leads to a large number of people already within urban expansion areas and exposed to high levels of pollution being incorrectly classified as "rural exposed populations," and vice versa, resulting in serious spatial misclassification of exposed populations and an inherent deficiency in the "spatial accuracy" of the assessment.

[0004] Therefore, there is an urgent need in this field for an evaluation scheme that can overcome the limitations of static perspective, achieve dynamic boundary recognition, and have accurate attribution capabilities. Summary of the Invention

[0005] This invention aims to overcome the technical shortcomings of existing air pollution health risk assessment technologies, such as misclassification of exposed populations due to the use of static population data and fixed administrative boundaries, and the inability to quantitatively separate the independent health effects of population migration. It provides a systematic solution capable of dynamically identifying urban-rural boundaries, constructing counterfactual control scenarios, and ultimately achieving quantitative attribution of net health risks from population migration. The solution is detailed below:

[0006] In a first aspect, the present invention provides a method for dynamic attribution assessment of health risks from air pollution due to population migration, comprising the following steps:

[0007] S1. Acquire and preprocess multi-source spatiotemporal data of the target area, including grid data of air pollutant concentration, remote sensing data of nighttime light, grid data of population distribution and statistical yearbook data, and resample all data to a unified geographic coordinate system and spatial resolution;

[0008] S2. Dynamic Urban-Rural Boundary Identification: Based on nighttime light remote sensing data and urban built-up area data in statistical yearbooks, the adaptive thresholds of each provincial unit within the target area in different years are calculated; based on the adaptive thresholds and the urban-rural classification results of the base year, the target area is divided into three types of dynamic spatial units each year: urban core area, urban expansion area, and rural area.

[0009] S3. Counterfactual scenario construction: Using the spatial distribution of population in the base year as a fixed pattern, and only adding the natural population growth factor, construct a counterfactual population distribution grid under the assumption that no inter-regional population migration occurs;

[0010] S4. Health risk calculation and migration attribution: Based on the real population distribution grid and the counterfactual population distribution grid respectively, combined with the air pollutant concentration grid data, the real scenario health risk value and the counterfactual scenario health risk value are calculated in parallel.

[0011] The difference between the real-world health risk value and the counterfactual health risk value is calculated, and the net health risk value caused by population migration is output.

[0012] As an optional implementation, calculating the adaptive threshold in step S2 specifically includes:

[0013] S21. For the target year and provincial unit, calculate the proportion of urban built-up area of ​​the provincial unit based on statistical yearbook data;

[0014] S22. Obtain the set of all nighttime light pixel values ​​within the provincial unit and calculate its cumulative area ratio function;

[0015] S23. The minimum nighttime light pixel value that satisfies the cumulative area ratio function value being greater than or equal to the urban built-up area ratio is determined as the adaptive threshold for the provincial unit in the target year.

[0016] As an optional implementation, the division rules for the three types of dynamic spatial units in step S2 are as follows:

[0017] The area that is classified as a city in the base year is defined as the urban core area;

[0018] The area that was classified as a city at the end of the baseline year but is classified as a city in the target year is defined as the urban expansion area.

[0019] Regions where both the base year and the target year are classified as rural are defined as rural areas.

[0020] As an optional implementation, the construction of the counterfactual population distribution grid in step S3 is specifically achieved through the following formula:

[0021] );

[0022] in, Indicates the grid position of the target year (yr) and province (prov). The counterfactual population figures on the )

[0023] This indicates the actual population of the corresponding location, year, and province.

[0024] This represents the province's actual total population for that year;

[0025] This represents the total cumulative net migration population of the province from the base year (2011) to the target year yr.

[0026] As an optional implementation, the health risk value is calculated in step S4 using the following formula:

[0027] ;

[0028] in, This represents the health risk value calculated based on a real population distribution grid.

[0029] Indicates year (yr) and age group Baseline mortality rate;

[0030] The attribution score is calculated from the pollutant exposure-response relationship coefficient.

[0031] Indicates the total number of exposed individuals;

[0032] For age group The proportion of the population.

[0033] As an optional implementation, the difference calculation in step S4 is performed using the following formula:

[0034] ;

[0035] in, Indicates the grid position ( Net health risk value caused by population migration in year yr;

[0036] The health risk value is calculated based on a counterfactual population distribution grid.

[0037] In a second aspect, the present invention provides a population migration attribution and spatiotemporal dynamic quantitative assessment system for air pollution health risks, used to implement the method described in the first aspect above, specifically including:

[0038] The data preprocessing module is used to acquire and fuse multi-source spatiotemporal data, complete spatial registration and resampling, and form a basic dataset with a unified spatiotemporal reference system.

[0039] The dynamic zoning module, connected to the data preprocessing module, is used to generate a dynamic urban-rural zoning map that divides the target area into urban core area, urban expansion area and rural area each year based on nighttime light data and urban built-up area statistics, by calculating the adaptive thresholds of each provincial unit in the target area in different years, and based on the adaptive thresholds and the urban-rural classification results of the base year.

[0040] The scenario building module, connected to the data preprocessing module and the dynamic partitioning module, is used to construct a counterfactual population distribution grid under the assumption that no inter-regional population migration occurs, with the population spatial distribution of the base year as a fixed pattern and only superimposed with the natural population growth factor.

[0041] The risk attribution module, connected to the data preprocessing module and the scenario construction module, is used to calculate health risk values ​​in parallel based on the real population distribution grid and the counterfactual population distribution grid, combined with air pollutant concentration data, and to perform differential calculation on the health risk values ​​under the two scenarios, outputting the net health risk value caused by population migration.

[0042] As an optional implementation, the dynamic partitioning module includes:

[0043] The threshold calculation unit is used to determine the corresponding minimum pixel value in the cumulative area ratio function of nighttime light pixels according to the proportion of urban built-up area by provincial unit and year, so as to serve as an adaptive threshold.

[0044] The spatial classification unit, connected to the threshold calculation unit, is used to classify each grid unit into the urban core area, urban expansion area or rural area year by year according to the adaptive threshold and the classification result of the base year.

[0045] As an optional implementation, the scenario construction module includes:

[0046] The migration scale calculation unit is used to calculate the annual net migration scale of each provincial unit based on the total population and natural population growth rate data in the statistical yearbook.

[0047] The population reconstruction unit, connected to the migration scale calculation unit, is used to reconstruct counterfactual population data at the grid level based on the migration scale, according to the rule of scaling the target year's real population grid proportionally to the total number after deducting the cumulative migration population.

[0048] Thirdly, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the method as described in the first aspect.

[0049] Compared to existing technologies, the benefits of this invention are significant and multifaceted. First, by introducing a counterfactual scenario construction rule that "locks in the baseline year's population spatial pattern and considers only natural growth," this invention successfully creates a computable and comparable analytical framework. This allows for the clear separation and quantitative calculation of the "population migration" factor from the complex compound effects of air pollution on health, fundamentally solving the core problems of effect confusion and unclear attribution in traditional assessments. Second, utilizing a provincial dynamic threshold algorithm based on the coupling of nighttime light remote sensing and built-up area statistical data, this invention achieves automatic and accurate identification of the annually changing urban expansion boundaries, thereby dividing the city into three dynamic spatial units: urban core area, expansion area, and rural areas. This innovation overcomes the spatial misclassification problem of exposed populations caused by relying on fixed administrative boundaries, thus elevating the spatiotemporal accuracy of health risk assessment to a new level. Furthermore, this invention integrates multi-source data fusion, dynamic boundary identification, counterfactual scenario construction, and differential attribution calculation into a logically rigorous and clearly defined standardized technical process. This process not only ensures the reproducibility and comparability of the analysis results, but also, due to its engineered design, possesses good portability and promotion potential. It can provide a stable and reliable analysis tool for environmental health risk management in different regions and at different times, ultimately achieving a leap from methodological innovation to improved decision support capabilities. Attached Figure Description

[0050] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0051] Figure 1 This is a schematic diagram of the overall technical route of the method and system for quantitative assessment of population migration attribution and spatiotemporal dynamics of air pollution health risks provided in the embodiments of the present invention. Detailed Implementation

[0052] This invention provides a method and system for dynamic attribution assessment of health risks from air pollution due to population migration. It aims to address the problems of misclassification of exposed populations caused by reliance on static population data and fixed administrative boundaries in existing technologies, as well as the lack of technical solutions for independent quantitative attribution of the health effects of population migration. The following is in conjunction with… Figure 1 The technical solution of the present invention will be described in detail and completely through specific embodiments.

[0053] Example 1:

[0054] This embodiment provides a method for quantitatively assessing the spatiotemporal dynamics of population migration attribution to air pollution health risks. The overall execution flow of this method is as follows: Figure 1 As shown, it mainly covers four closely connected stages: data preprocessing stage, urban-rural dynamic zoning stage, scenario construction stage, and attribution calculation stage.

[0055] The primary prerequisite for implementing this method is to construct a high-precision, multi-source fused spatiotemporal data foundation. This involves acquiring and preprocessing multi-source spatiotemporal data of the target region, a process that standardizes data from different sources, formats, and resolutions. In this embodiment, the required data mainly includes grid data of air pollutant concentrations, remote sensing data of nighttime light, grid data of population distribution, and statistical yearbook data. Taking seven provinces in East China—Shandong, Jiangsu, Anhui, Zhejiang, Jiangxi, Fujian, and Shanghai—as an example, the evaluation time span is selected from 2011 to 2020. This period represents the most active stage of urbanization and population movement in the region, making it highly representative.

[0056] For the air pollutant concentration grid data, this embodiment preferably uses a publicly available high-precision dataset, such as the ChinaHighNO2 dataset, to obtain daily nitrogen dioxide concentration data from 2011 to 2020. The spatial resolution of the original data can be 10 km × 10 km, or other specifications. To meet the needs of refined assessment, it needs to be processed into a unified standard.

[0057] For nighttime light remote sensing data, publicly available VIIRS annual nighttime light data can be used. The raw spatial resolution of this data is usually high, about 15 arcseconds (about 500 meters at the equator). Its high sensitivity can effectively characterize the intensity of human activities on the Earth's surface and is a key basis for identifying the spatial extent of urbanization.

[0058] For population distribution data, the annual population count grid data provided by the World Population Grid Dataset can be used. Its original resolution is 30 arcseconds, which can reflect the subtle differences in the spatial distribution of the population.

[0059] For supplementary statistical data, one can consult the officially published "China Statistical Yearbook" as needed, which systematically compiles data on the urban built-up area, natural population growth rate, and total population of each province for each year.

[0060] In addition, to calculate health risks, basic health data is required. Specifically, baseline mortality rates for each age group within the study area can be extracted from the Global Burden of Disease Study database, and the exposure-response coefficient between long-term nitrogen dioxide exposure and all-cause mortality can be determined. For example, in this embodiment, this coefficient is set as the percentage increase of 10 μg / m². 3 The concentration increases the relative risk by 2%, which translates to a relative risk RR of 1.02 with a confidence interval of 1.01 to 1.04.

[0061] After collecting the massive amounts of data, rigorous data preprocessing and spatiotemporal alignment are essential. This is because different data sources often employ different projection methods and grid division standards. This step utilizes geographic information technology to resample and crop all air pollutant concentration data, nighttime light data, and population distribution data to a unified geographical area, namely the aforementioned seven provinces in East China, and uniformly set the spatial resolution to 0.1° × 0.1°. Simultaneously, a spatiotemporal data cube based on the same geographic coordinate system, such as the WGS 1984 coordinate system, is established. This operation ensures that each subsequent grid cell has a precise correspondence across different data layers, thus forming a standardized basic dataset suitable for pixel-by-pixel computation, thereby completing the process. Figure 1 The first phase of the task, "data preprocessing," achieved spatiotemporal alignment and resampling of air pollutant concentration grids, nighttime light remote sensing data, population distribution grids, and statistical yearbook data.

[0062] Immediately following data preprocessing, the method proceeds to... Figure 1 The second stage, "Dynamic Urban-Rural Zoning," involves implementing dynamic urban-rural boundary identification. The core innovation of this step lies in overcoming the spatial misclassification defects caused by the use of fixed administrative boundaries in traditional methods. It introduces a dynamic threshold algorithm to accurately capture the changing boundaries of urban entities year by year. Specifically, this step calculates adaptive thresholds for each provincial unit within the target area in different years based on nighttime light remote sensing data and urban built-up area data from statistical yearbooks. Based on this, the target area is divided into three dynamic spatial units each year: urban core area, urban expansion area, and rural area. To achieve this goal, the proportion of urban built-up areas in each province needs to be calculated first. For each target year (yr) and each provincial unit (prov), using data obtained from the statistical yearbook, the proportion of the urban built-up area of ​​that provincial unit to its total administrative area is calculated. This proportion is denoted as . The calculation is based on the following formula:

[0063] ;

[0064] In the above formula, the molecule This represents the built-up area of ​​the province (prov) in a specific year (yr) as recorded in the statistical yearbook.

[0065] denominator This represents the total administrative area of ​​the province (prov).

[0066] After calculating this ratio, the system further determines the dynamic urban nighttime light index threshold. This process requires traversing the set of all nighttime light pixel values ​​within the provincial unit and calculating their cumulative area ratio function. Specifically, the operation logic is to find a minimum nighttime light pixel value x such that the cumulative area ratio represented by all pixels within the province that are greater than or equal to this value x is exactly greater than or equal to the previously calculated urban built-up area ratio. The smallest pixel value found is determined as the adaptive threshold for that provincial unit in the target year, denoted as... The mathematical logic for determining this threshold is shown in the following formula:

[0067] ;

[0068] in, A function representing the cumulative proportion of the total area to which the nighttime light intensity is greater than or equal to x;

[0069] Indicates all A set of pixel values.

[0070] Through this adaptive algorithm, the nighttime light intensity values ​​can ensure that the total amount of the designated urban area is consistent with the statistical data, while the spatial form closely matches the actual range of human activity reflected by the nighttime lights.

[0071] After determining the adaptive threshold for each year, the dynamic attribute division of each grid unit can be performed year by year based on this threshold and the urban-rural classification results of the baseline year. The specific division rules are as follows:

[0072] First, compare the grid position of the target year yr ( Nighttime light index Based on the province's adaptive threshold for that year, if the grid's nighttime light index is greater than or equal to the threshold, it is initially determined to be a city; if it is less than the threshold, it is determined to be a township, as shown in the following formula:

[0073] ;

[0074] Based on this, the classification status of the baseline year, 2011, is used to refine the definition:

[0075] If a grid was already classified as a city in the base year 2011, it is defined as an "urban core area," which represents mature existing urban space.

[0076] If a grid was not classified as urban in the base year 2011 (i.e., it was rural at that time), but is classified as urban in the target year yr, it is defined as an "urban expansion area". This type of area accurately describes the newly expanded space in the process of urbanization.

[0077] If a grid is classified as rural in both the base year and the target year, it is defined as either a "rural area" or a "township area".

[0078] Starting with 2011, using its nighttime light data and the proportion of built-up areas in that year to determine the initial urban-rural classification, and then executing the above judgment logic year by year, a dynamic urban-rural boundary zoning map can be generated from 2011 to 2020, thus completing the process. Figure 1 The diagram illustrates the process of dividing a region into an urban core area, an urban expansion area, and rural areas.

[0079] After completing the fine-grained dynamic partitioning, the method proceeds... Figure 1 The third stage, "scenario construction," involves building a counterfactual "no migration" scenario. This is a crucial step in enabling the quantitative separation of the independent health effects of migration. The core idea of ​​this step is to construct a counterfactual population distribution grid assuming no inter-regional migration, using the spatial distribution of the base year as a fixed pattern and only adding natural population growth factors. This grid serves as a reference for comparison with the actual population distribution. To construct this scenario, the scale of migration must first be calculated. This is based on the total population figures from the statistical yearbooks of the target year (yr) and the province (prov). and the natural population growth rate Based on the previous year's population figures, the net migration number compared to the previous year is calculated. The calculation logic is as follows: subtract the theoretically expected population from the previous year's total population under the influence of the natural growth rate from the actual total population of the current year; the difference is the mechanical population growth, or net migration. This calculation is achieved through the following formula:

[0080] ;

[0081] In the above formula, Indicates population number, This represents the population figure for the previous year. Indicates the natural growth rate. This represents the migration volume, i.e., the number of people migrating compared to the previous year. Using this formula, this example calculates the scale of population migration in seven provinces and municipalities in East China from 2011 to 2020. Through the summary and analysis of the data, the population flow characteristics of different provinces and regions can be clearly revealed. According to the actual calculation results, Shandong Province absorbed 739,000 people in its urban core area and 224,000 people in its urban expansion area between 2011 and 2020, but its rural areas saw an outflow of 1.761 million people, resulting in a net outflow of 798,000 people for the entire province. In contrast, Zhejiang Province showed extremely strong population attraction, with 1.866 million people flowing into its urban core area, 693,000 people into its urban expansion area, and even its rural areas attracting 3.747 million people due to its unique economic model, resulting in a total net inflow of 6.306 million people for the entire province. These detailed data provide a solid foundation for subsequent attribution calculations. The specific population migration scale data for each province by region are shown in the table below:

[0082] Table 1. Population Migration Scale in East China, 2011-2020

[0083]

[0084] After calculating the annual net migration scale of each province, we can begin to construct a population distribution scenario with "no migration". The specific operation is as follows: using 2011 as the base year, the actual population grid of each province in the target year yr is scaled proportionally. The scaling target is "the theoretical total after deducting the cumulative migration population". Specifically, for the target year yr, the province prov is located at ( The counterfactual population figure is denoted as . Its value is equal to the proportion of the actual population at that location to the total actual population of the province in that year, multiplied by the total actual population of the province in that year minus the total cumulative net migration from 2011 to the target year. This reconstruction process is achieved through the following formula:

[0085] );

[0086] in, Indicates the grid position of the target year (yr) and province (prov). The counterfactual population figures on the )

[0087] This indicates the actual population of the corresponding location, year, and province.

[0088] This represents the province's actual total population for that year;

[0089] This represents the total cumulative net migration population of the province from the base year (2011) to the target year yr.

[0090] This formula, while maintaining the weights of the actual spatial distribution of the population in that year, eliminates changes caused by migration in the total population, thus simulating a virtual population distribution state where the population only grows naturally in its place and there is no mechanical movement across regions. This state is... Figure 1 The third phase output is a "counterfactual scenario: locking in the base year structure".

[0091] After constructing the population distribution data for both real-world and counterfactual scenarios, the method proceeds... Figure 1 The fourth stage, "attribution calculation," is the calculation of health risks from air pollution exposure and migration attribution. This stage uses parallel computing and differential analysis to ultimately quantify the net impact of population migration on health risks.

[0092] First, health risk values ​​need to be calculated based on both the real population distribution grid and the aforementioned counterfactual population distribution grid, combined with air pollutant concentration grid data. In this embodiment, the health risk indicator is selected as the premature mortality burden caused by long-term nitrogen dioxide exposure. The calculation logic is the same for both real and counterfactual scenarios, and is calculated by accumulating data using the following formula:

[0093] ;

[0094] In the above formula, This represents the calculated health risk value (number of premature deaths).

[0095] Indicates a specific year Specific age groups Baseline mortality rate;

[0096] This indicates the total exposed population. Real-world grid population data is used in real-world scenarios, while counterfactual grid population data is used in counterfactual scenarios.

[0097] This indicates the population proportion of the corresponding age group;

[0098] This represents the attribution score, which is calculated from the pollutant exposure-response relationship coefficient, specifically: ,

[0099] In this embodiment, the relative risk (RR) is set to 1.02. By summing the values ​​for all grids and all age groups, the real-world health risk values ​​can be obtained. Health risk values ​​under a scenario without population migration .

[0100] Finally, to quantify the net health risk of migration, the health risk values ​​under the two scenarios above are calculated using a difference calculation. That is, the net health risk value caused by migration. The health risk value is equal to the real-world health risk value minus the counterfactual health risk value. The calculation formula is as follows:

[0101] ;

[0102] The determination is made based on the calculation result of this formula: if A value greater than 0 indicates that actual population migration increases the overall nitrogen dioxide health risk burden compared to the scenario assuming no migration; conversely, if... A value less than 0 indicates that population migration has reduced the overall health risk burden. Based on detailed calculations of data from seven provinces and cities in East China from 2011 to 2020 in this embodiment, the net impact of nitrogen dioxide exposure on health risks in each province due to population migration was obtained. The specific results are summarized in the table below:

[0103] Table 2. Net Health Risks of Nitrogen Dioxide Exposure Due to Population Migration in Seven Provinces and Cities in East China, 2011-2020

[0104]

[0105] A deeper analysis of the data in Table 2 reveals significant spatial heterogeneity in the impact of population migration on health risks. For example, Zhejiang Province, a major net inflow province, attracts a large influx of people due to its vibrant economy, resulting in a net increase in health risks in urban core areas, urban expansion areas, and rural areas, with a total increase of 2038 premature deaths. This indicates a significant agglomeration effect of population towards high-pollution or high-exposure risk areas. Jiangsu Province shows a similar pattern, with a total increase of 1312 people. In stark contrast, Anhui and Jiangxi Provinces, typical population-exporting regions, experienced a large outflow of people, especially from rural areas (as shown in Table 1), reducing the base of people exposed to air pollution in these areas and resulting in a net decrease in health risks of -262 and -38 people, respectively. Shandong Province exhibits a balanced state; although the risk increases in urban areas, the significant outflow from rural areas offsets this increase, resulting in a net impact of only -3 people. This assessment accurately reveals how population mobility reshapes the risk pattern of regional environmental health, validating the effectiveness and sensitivity of the counterfactual attribution framework proposed in this invention, and achieving… Figure 1 The ultimate goal is to output the "net health risks caused by population migration".

[0106] In summary, the method provided by the embodiments of the present invention, through Figure 1 The complete technical approach presented successfully combines multi-source spatiotemporal data fusion, dynamic urban-rural boundary identification technology, and counterfactual scenario construction methods. It not only solves the problem of spatial misclassification of exposed populations caused by the use of static population data and fixed administrative boundaries in traditional assessments, but more importantly, it innovatively establishes a computable control system, making the independent impact of the complex sociodemographic process of "population migration" on environmental health quantifiable and attributable. This method provides a strong scientific basis and quantitative tool for governments to formulate differentiated population guidance policies, optimize urban spatial layout, and implement precise environmental health risk management.

[0107] Example 2:

[0108] Based on Embodiment 1 above, this embodiment provides a dynamic attribution assessment system for air pollution health risks from population migration. This system, in its logical structure, highly corresponds to the aforementioned method flow and aims to automate execution through the collaborative work of computer hardware and software. Figure 1 The steps shown.

[0109] The system mainly includes a data preprocessing module, a dynamic partitioning module, a scenario construction module, and a risk attribution module. The data preprocessing module is primarily used for execution... Figure 1 The first phase of the task involves acquiring and integrating multi-source spatiotemporal data, including air pollutant concentration grids, nighttime light remote sensing data, population distribution grids, and statistical yearbooks. Through internally integrated resampling and coordinate transformation algorithms, spatial registration and resolution unification are achieved to form a standardized basic dataset.

[0110] The dynamic partitioning module connects to the data preprocessing module for execution. Figure 1 The second phase of the task comprises a threshold calculation unit and a spatial classification unit. The threshold calculation unit is responsible for calculating the adaptive threshold year by year based on the proportion of urban built-up area and the cumulative distribution function of nighttime light. The spatial classification unit, based on the threshold and the baseline year's status, dynamically divides each grid cell into urban core area, urban expansion area, or rural area, generating a dynamic urban-rural zoning map year by year.

[0111] The scenario building module connects to the data preprocessing module and the dynamic partitioning module to perform... Figure 1 The third phase of the task comprises a migration scale calculation unit and a population reconstruction unit. The migration scale calculation unit calculates net migration based on census and natural growth data; the population reconstruction unit uses the aforementioned counterfactual reconstruction formula to generate a virtual population distribution grid assuming no migration has occurred.

[0112] Finally, the risk attribution module connects to the aforementioned modules to perform... Figure 1 The fourth phase of the task involves calculating health risk values ​​in both real-world and counterfactual scenarios in parallel, performing differential operations, and ultimately outputting the net health risk value caused by population migration.

[0113] Through its modular design, the system automates and standardizes the assessment process, greatly improving the efficiency and accuracy of environmental health risk assessment.

[0114] Example 3:

[0115] This embodiment provides an electronic device including a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it can implement all the steps of the population migration attribution and spatiotemporal dynamic quantitative assessment method for air pollution health risks described in the above embodiments. This electronic device can be a high-performance server, workstation, or cloud computing node. In actual operation, the processor first reads instructions from the memory, controls the communication interface to acquire external multi-source data, and then allocates a buffer in memory for data resampling and alignment. Next, the processor calls a mathematical operation library to traverse massive amounts of grid data, calculates adaptive thresholds, and performs dynamic partitioning and labeling. Subsequently, the processor constructs a virtual population matrix based on counterfactual logic and uses parallel computing technology to simultaneously solve health risk equations under two scenarios. Finally, the processor presents the calculated difference results to the user through an output interface or stores them in a database for further analysis. Through this combination of hardware and software, this invention can rapidly process large-scale, long-term spatiotemporal big data, providing real-time intelligent support for regional environmental governance and public health decision-making.

Claims

1. A method for assessing the dynamic attribution of health risks from air pollution due to population migration, characterized in that, Includes the following steps: S1. Acquire and preprocess multi-source spatiotemporal data of the target area, including grid data of air pollutant concentration, remote sensing data of nighttime light, grid data of population distribution and statistical yearbook data, and resample all data to a unified geographic coordinate system and spatial resolution; S2. Dynamic Urban-Rural Boundary Identification: Based on nighttime light remote sensing data and urban built-up area data in statistical yearbooks, adaptive thresholds for each provincial unit within the target area in different years are calculated. Based on the adaptive threshold and the urban-rural classification results of the baseline year, the target area is divided into three dynamic spatial units each year: urban core area, urban expansion area and rural area. S3. Counterfactual scenario construction: Using the spatial distribution of population in the base year as a fixed pattern, and only adding the natural population growth factor, construct a counterfactual population distribution grid under the assumption that no inter-regional population migration occurs; S4. Health risk calculation and migration attribution: Based on the real population distribution grid and the counterfactual population distribution grid respectively, combined with the air pollutant concentration grid data, the real scenario health risk value and the counterfactual scenario health risk value are calculated in parallel. The difference between the real-world health risk value and the counterfactual health risk value is calculated, and the net health risk value caused by population migration is output.

2. The method for assessing the dynamic attribution of air pollution health risks based on population migration according to claim 1, characterized in that, Step S2, calculating the adaptive threshold, specifically includes: S21. For the target year and provincial unit, calculate the proportion of urban built-up area of ​​the provincial unit based on statistical yearbook data; S22. Obtain the set of all nighttime light pixel values ​​within the provincial unit and calculate its cumulative area ratio function; S23. The minimum nighttime light pixel value that satisfies the cumulative area ratio function value being greater than or equal to the urban built-up area ratio is determined as the adaptive threshold for the provincial unit in the target year.

3. The method for assessing the dynamic attribution of air pollution health risks based on population migration according to claim 1, characterized in that, The division rules for the three types of dynamic spatial units mentioned in step S2 are as follows: The area that is classified as a city in the base year is defined as the urban core area; The area that was classified as a city at the end of the baseline year but is classified as a city in the target year is defined as the urban expansion area. Regions where both the base year and the target year are classified as rural are defined as rural areas.

4. The method for assessing the dynamic attribution of air pollution health risks based on population migration according to claim 1, characterized in that, The construction of the counterfactual population distribution grid in step S3 is specifically achieved through the following formula: )； in, Indicates the grid position of the target year (yr) and province (prov). The counterfactual population figures on the ) This indicates the actual population of the corresponding location, year, and province. This represents the province's actual total population for that year; This represents the total cumulative net migration population of the province from the base year to the target year yr.

5. The method for assessing the dynamic attribution of air pollution health risks based on population migration according to claim 1, characterized in that, The health risk value is calculated in step S4 using the following formula: ； in, This represents the health risk value calculated based on a real population distribution grid. Indicates year (yr) and age group Baseline mortality rate; The attribution score is calculated from the pollutant exposure-response relationship coefficient. Indicates the total number of exposed individuals; For age group The proportion of the population.

6. The method for assessing the dynamic attribution of air pollution health risks based on population migration according to claim 1, characterized in that, The difference calculation in step S4 is achieved through the following formula: ； in, Indicates the grid position ( Net health risk value caused by population migration in year yr; The health risk value is calculated based on a counterfactual population distribution grid.

7. A population migration dynamic attribution assessment system for health risks from air pollution, characterized in that, include: The data preprocessing module is used to acquire and fuse multi-source spatiotemporal data, complete spatial registration and resampling, and form a basic dataset with a unified spatiotemporal reference system. The dynamic zoning module, connected to the data preprocessing module, is used to generate a dynamic urban-rural zoning map that divides the target area into urban core area, urban expansion area and rural area each year based on nighttime light data and urban built-up area statistics, by calculating the adaptive thresholds of each provincial unit in the target area in different years, and based on the adaptive thresholds and the urban-rural classification results of the base year. The scenario building module, connected to the data preprocessing module and the dynamic partitioning module, is used to construct a counterfactual population distribution grid under the assumption that no inter-regional population migration occurs, with the population spatial distribution of the base year as a fixed pattern and only superimposed with the natural population growth factor. The risk attribution module, connected to the data preprocessing module and the scenario construction module, is used to calculate health risk values ​​in parallel based on the real population distribution grid and the counterfactual population distribution grid, combined with air pollutant concentration data, and to perform differential calculation on the health risk values ​​under the two scenarios, outputting the net health risk value caused by population migration.

8. The population migration dynamic attribution assessment system for air pollution health risks according to claim 7, characterized in that, The dynamic partitioning module includes: The threshold calculation unit is used to determine the corresponding minimum pixel value in the cumulative area ratio function of nighttime light pixels according to the proportion of urban built-up area by provincial unit and year, so as to serve as an adaptive threshold. The spatial classification unit, connected to the threshold calculation unit, is used to classify each grid unit into the urban core area, urban expansion area or rural area year by year according to the adaptive threshold and the classification result of the base year.

9. The population migration dynamic attribution assessment system for air pollution health risks according to claim 7, characterized in that, The scenario construction module includes: The migration scale calculation unit is used to calculate the annual net migration scale of each provincial unit based on the total population and natural population growth rate data in the statistical yearbook. The population reconstruction unit, connected to the migration scale calculation unit, is used to reconstruct counterfactual population data at the grid level based on the migration scale, according to the rule of scaling the target year's real population grid proportionally to the total number after deducting the cumulative migration population.

10. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1 to 6.

Images