Method and system for identifying and diagnosing ozone control area based on multi-source satellite data

Abstract

The invention discloses an ozone control area identification and diagnosis method and system based on multi-source satellite data. The method comprises the following steps: constructing a standardized long sequence data set; representing the regional ozone weekend effect by adopting an ozone weekend-weekend absolute difference; establishing a linear regression model of the HCHO / NO2 ratio and the regional ozone weekend effect according to the standardized long sequence data set and the regional ozone weekend effect, and obtaining a regional adaptive HCHO / NO2 key threshold interval through the linear regression model; and according to the standardized long sequence data set, identifying the trend characteristics of the HCHO / NO2 ratio, and according to the trend characteristics of the HCHO / NO2 ratio and the regional HCHO / NO2 key threshold interval, carrying out diagnostic analysis on the evolution type of the ozone control region. By establishing a regional dependent HCHO / NO2 key threshold system and combining with analysis of HCHO / NO2 trend turning points, accurate identification and diagnosis of global 30-year ozone control region space-time evolution conditions can be realized, and a reliable basis is provided for monitoring, prevention and treatment of global ozone pollution.

Description

Technical Field

[0001] This invention belongs to the field of atmospheric environment remote sensing monitoring technology, specifically relating to a method and system for identifying and diagnosing ozone control zones based on multi-source satellite data. Background Technology

[0002] The key to controlling near-surface ozone (O3) pollution lies in accurately identifying its formation control zones (VOC control, NOx control, or transition zones). Traditional methods rely on single satellite data or short-term ground observations, which suffer from insufficient spatiotemporal coverage, inconsistent regional thresholds, and limited verification methods (only model or weekend effect analysis). Furthermore, the nonlinear relationship between O3 and its precursors, as well as regional emission differences, lead to significant errors in the long-term evolution diagnosis of existing methods.

[0003] Traditional monitoring methods suffer from the following shortcomings: 1) Rigid thresholds: Traditional fixed thresholds for HCHO / NO2 (such as the Duncan threshold) cannot adapt to regional emission differences and the complexity of chemical mechanisms. 2) Spatiotemporal limitations: Sparse ground observation stations and fragmented satellite data periods (e.g., OMI only covers data after 2005) make it difficult to achieve long-term global-scale evolution analysis; traditional studies often rely on single satellites (such as OMI) or short-term ground observations, making it difficult to capture the long-term evolution patterns of O3 control areas. 3) Lack of validation: Relying solely on model simulations or single indicators (such as the O3 weekend effect) and lacking multi-indicator collaborative validation leads to misjudgment of control area types (e.g., misclassifying transition zones as NOx control areas). Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for identifying and diagnosing ozone control zones based on multi-source satellite data, so as to solve the problems of poor adaptability and poor prediction accuracy of traditional monitoring methods.

[0005] This invention is achieved through the following technical solution: Ozone control zone identification and diagnostic methods based on multi-source satellite data include: Acquire satellite HCHO and NO2 data, and construct a standardized long-sequence dataset; O3 and NO2 observation data from ground stations were obtained, and the regional ozone weekend effect was characterized by the ozone weekend-midweek absolute difference, which is the difference between the weekend and midweek average ozone concentrations. Based on standardized long-sequence datasets and regional ozone weekend effects, a linear regression model of the HCHO / NO2 ratio and regional ozone weekend effects was established, and the key threshold range of regional HCHO / NO2 was obtained through the linear regression model. Based on the standardized long-sequence dataset, the trend characteristics of the HCHO / NO2 ratio are identified. Based on the trend characteristics of the HCHO / NO2 ratio and the key threshold range of HCHO / NO2 in the region, the evolution type of the ozone control area is diagnosed and analyzed.

[0006] In some embodiments, the step of constructing a standardized long sequence dataset includes: S011. Acquire HCHO and NO2 column concentration data from fourth-generation satellites GOME / ERS-2, SCIAMACHY / ENVISAT, OMI / Aura, and TROPOMI / Sentinel-5P. S012. Based on OMI satellite data, the NO2 column concentration data of GOME and SCIAMACHY are corrected.

[0007] In some embodiments, step S012 includes: S0121. Use a transit time correction factor to correct the transit time of SCIAMACHY satellite data and GOME satellite data; S0122. Spatial resolution correction is performed on SCIAMACHY satellite data and GOME satellite data using a resolution correction factor.

[0008] In some embodiments, the step of establishing a linear regression model of the HCHO / NO2 ratio and the regional ozone weekend effect based on a standardized long-sequence dataset and the regional ozone weekend effect, and obtaining the key threshold range of regional HCHO / NO2 through the linear regression model, includes: S031. Obtain monitoring data from ground stations in areas affected by anthropogenic pollution emissions exceeding the baseline value, and match the HCHO / NO2 ratio with the hourly average O3 value during the OMI satellite transit period; S032. Establish a linear regression model of the HCHO / NO2 ratio and the ozone weekend effect in the corresponding area of ​​the selected ground station. Define the intersection of the regression line and WE-WD O3=0 as the key threshold. WE-WD O3 represents the absolute difference in ozone between the weekend and midweek. S033. Divide the world into multiple regions, conduct statistical analysis on the key thresholds of ground stations in each region, and construct the HCHO / NO2 key threshold intervals for each region.

[0009] In some embodiments, the step of identifying trend characteristics of the HCHO / NO2 ratio based on a standardized long-sequence dataset includes: S0411. Define the turning point potential function S(t), expressed as: ; Among them, B1, B r and B1+r Let p(B1) and p(B2) represent the linear trends over 5 years to the left, 5 years to the right, and 10 years spanning year t, respectively. r ) represent the p-values ​​for trend significance. This represents the trend fitting error. S0412. For each latitude and longitude grid point, its annual HCHO / NO2 ratio is used as time series data, and a linear model is used for fitting to determine the parameters B1 and B in S(t). r and B 1+r The linear model is represented as: ; Among them, Y t X is the average value in year t. t Let A represent the time variable, A be the average value of the first year, and B be the estimated slope of the trend line, corresponding to parameters B1 and B in S(t). r and B 1+r N t Represents residuals; S0413. If a certain year has the highest S(t) value, and B1, B r If the signs are opposite and the trends on both sides are significant, then the year is identified as the turning point in the evolution of the ozone control area, and it is taken as the trend inflection point; the latitude and longitude grid points for which no turning point year is identified are fitted as linear trends, and the significance of the trend is judged.

[0010] In some embodiments, the step of diagnostic analysis of the evolution type between the NOx control area and the VOC control area based on the trend characteristics of the HCHO / NO2 ratio and the key threshold range of regional HCHO / NO2 includes: When the HCHO / NO2 ratio is consistently outside the critical threshold range, it indicates that the ozone control zone remains within the NOx control zone or VOC control zone, and the evolution type is determined to be stable. When the HCHO / NO2 ratio is within the critical threshold range, but no change in the control mechanism occurs in the ozone control area, the evolution type is determined to be quasi-stable. The method to determine whether a change in the control mechanism has occurred is: if the HCHO / NO2 ratio continuously and stably crosses the critical threshold range for many consecutive years, it is determined that a change in the control mechanism has occurred. When the HCHO / NO2 ratio shows a linear trend and a control mechanism shift occurs in the ozone control area at a certain stage, the evolution type is determined to be a single shift type. When the HCHO / NO2 ratio exhibits a non-linear trend with an inflection point, and the ozone control area has undergone multiple changes in control mechanisms, the evolution type is determined to be a multiple-change type.

[0011] In some embodiments, for grid points of single-transformation and multiple-transformation types, the year of intersection between the HCHO / NO2 trend fitting line and the key threshold is calculated to form an evolution map of the ozone generation mechanism.

[0012] In some embodiments, based on the difference between ozone weekend and midweek absolute difference and the difference in exceedance probability, Theil-Sen slope estimation combined with the Mann-Kendall trend test is used to analyze the long-term trend of the difference between ozone weekend and midweek absolute difference and the difference in exceedance probability based on a 5-year sliding window, classifying the evolution pattern of regional ozone weekend effect and identifying turning point time points. By comparing the evolution types of ozone control areas with the evolution patterns of regional ozone weekend effects, the diagnostic results of the ozone control area evolution types are verified.

[0013] In some embodiments, the evolution patterns of regional ozone weekend effects are classified into non-trend type, weekend effect decline type, weekend effect enhancement type, and transition type; In each 5-year sliding window, if the Mann-Kendall test shows a significant trend, the sliding window period is determined to be either a weekend effect enhancement type or a weekend effect decline type based on the sign of the Theil-Sen slope β; if the trend is not significant, it is determined to be a no-trend type. Starting from the beginning of the time series, the short-term trend within a sliding window of 5 consecutive years is calculated sequentially. The window is then moved forward by one year and the calculation is repeated. The calculation results of each window are aligned with their corresponding median year to obtain a short-term trend sequence arranged in chronological order. The persistence and directional changes that occur in the short-term trend sequence are used to determine whether it is a turning point.

[0014] On the other hand, the present invention also provides an ozone control area identification and diagnosis system based on multi-source satellite data, for performing the ozone control area identification and diagnosis method based on multi-source satellite data, including: The satellite data acquisition and processing unit is used to acquire satellite HCHO and NO2 data and construct them into a standardized long-sequence dataset. The ground station data acquisition and processing unit is used to acquire O3 and NO2 observation data from ground stations and calculate the absolute difference and the probability of exceeding the ozone standard between the weekend and midweek. The key threshold interval acquisition unit is used to establish a linear regression model of the HCHO / NO2 ratio and the regional ozone weekend effect based on the standardized long-sequence dataset and the regional ozone weekend effect, and to obtain the key threshold interval of regional HCHO / NO2 based on the linear regression model. The ozone control zone diagnostic unit is used to identify the trend characteristics of the HCHO / NO2 ratio and, based on the trend characteristics of the HCHO / NO2 ratio and the key threshold range of HCHO / NO2 in the region, to perform diagnostic analysis on the evolution type of the ozone control zone.

[0015] Compared with the prior art, the present invention has the following advantages and beneficial effects: This invention integrates HCHO / NO2 data from four generations of satellites from 1996 to 2024, combines it with data from the global ground-based O3 monitoring network, and establishes a region-dependent key threshold system for HCHO / NO2. By analyzing the inflection points of HCHO / NO2 trends, it can accurately identify and diagnose the spatiotemporal evolution of ozone control areas globally over 30 years, providing a reliable basis for the monitoring and prevention of global ozone pollution. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart of the ozone control zone identification and diagnosis method according to an embodiment of the present invention.

[0018] Figure 2 This is a scatter plot of the monthly average HCHO / NO2 ratio and the weekend-weekend O3 difference (WE-WD O3) for representative city sites (Washington and Los Angeles, USA); the black line represents the linear regression line of the two, and the red triangle represents the intersection of the regression line with the baseline WE-WD O3=0. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, specific embodiments of this application will be described in further detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely for explaining this application and not for limiting it. It should also be noted that, for ease of description, only the parts relevant to this application are shown in the drawings, not all of them. Before discussing exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe operations (or steps) as sequential processes, many of these operations can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but may also have additional steps not included in the drawings. The process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0020] This invention integrates column concentration observation data of formaldehyde (HCHO) and nitrogen dioxide (NO2) from four generations of satellites (GOME / ERS-2, SCIAMACHY / ENVISAT, OMI / Aura, TROPOMI / Sentinel-5P) from 1996 to 2024, and combines it with data from the global ground-based O3 monitoring network to establish a region-dependent HCHO / NO2 key threshold system, thereby achieving accurate identification of O3 generation control areas (VOC control areas, NOx control areas, or transition areas).

[0021] By employing a dual verification mechanism of O3 weekend effect and satellite indicators, combined with trend inflection point detection, control area type classification, and spatiotemporal evolution analysis, this method solves the problems of insufficient single-satellite time period coverage and inconsistent regional thresholds in traditional monitoring methods. It can achieve accurate identification and diagnosis of the spatiotemporal evolution of global O3 control areas over 30 years, and can provide a reliable basis for global ozone pollution monitoring and prevention.

[0022] In some embodiments of the present invention, reference is made to Figure 1 A method for identifying and diagnosing ozone control zones based on multi-source satellite data includes the following steps: S01. Satellite data fusion processing: Integrate HCHO / NO2 data from four generations of satellites from 1996 to 2024 to construct a standardized long-sequence dataset; Includes the following steps: S011, Data Source Selection Table 1 integrates HCHO and NO2 column concentration data from four generations of satellites (GOME / ERS-2, SCIAMACHY / ENVISAT, OMI / Aura, and TROPOMI / Sentinel-5P), covering the period from 1996 to 2024. Table 1. HCHO and NO2 column concentration data from fourth-generation satellites.

[0023] S012, Data Preprocessing Because the NO2 data from multiple satellites are inconsistent in terms of spatiotemporal resolution (e.g., GOME: 320×40 km²; OMI: 24×13 km²) and transit time (e.g., SCIAMACHY: 10:00 LT; OMI: 13:45 LT), the data needs to be preprocessed.

[0024] Based on OMI satellite data, the NO2 column concentration data of GOME and SCIAMACHY were corrected; the OMI satellite has the best spatial resolution (24×13 km²), and its transit time (13:45 local time) can capture the period when ozone photochemical reactions are most active.

[0025] Specifically, the following steps are included: S0121, Transit Time Correction To address the systemic bias caused by differences in satellite transit times (SCIAMACHY: 10:00; GOME: 10:30), a transit time correction factor was added to unify all data to the equivalent observation value at 13:30 local time.

[0026] For SCIAMACHY data, it is first assumed that the bias is independent of instrument resolution. Then, based on the coarse resolution (2°×0.5°) data of OMI and SCIAMACHY during the overlapping observation period (2004-2012), the average difference between the two is calculated using Formula 1.

[0027] ...(1) S0122, Spatial resolution correction To eliminate the differences in horizontal resolution between different instruments, a resolution correction factor is introduced (refer to Formula 5).

[0028] The impact of resolution on observations was quantified by calculating the ratio of NO2 column concentrations at high resolution (0.125°×0.125°) to coarse resolution (2°×0.5°) of the OMI and SCIAMACHY satellites (refer to Formulas 2-4).

[0029] ...(2) ...(3) ... (4) ... (5) The GOME data uses the same transit time correction factor as SCIAMACHY. The temporal variability of the resolution correction factor is driven by the variation characteristics of GOME NO2 (refer to Equations 6-8), where Fine represents fine grid cells and Coarse represents coarse grid cells.

[0030] Finally, the overall correction of the GOME and SCIAMACHY data was completed using Formula 9.

[0031] ... (6) ... (7) ... (8) ... (9) The symbols and meanings of the parameters in Formulas 1-9 are explained in Tables 2 and 3.

[0032] Table 2 Explanation of Parameter Symbols

[0033] Table 3 Explanation of Formula 1 and its parameters

[0034] Formula 1 is used to calculate the transit time correction factor (coarse resolution comparison). On coarse grid cells, the difference in NO2 column concentration between OMI and SCIAMACHY over the overlapping period (2004-2012) is calculated. This difference is mainly used to correct for systematic bias caused by different transit times.

[0035] Table 4. Explanation of Formula 2 and its parameters

[0036] Formula 2 is used to calculate the static resolution correction factor (baseline ratio) of OMI. Based on OMI data, the multi-year average ratio of NO2 concentration in fine grid to coarse grid is calculated, representing the average impact of resolution.

[0037] Table 5 Explanation of Formulas 3 and 4 and their parameters

[0038] Formulas 3 and 4 are used to calculate the final annual resolution correction factor for SCIAMACHY. Using SCIAMACHY's own data, the average ratio of high-resolution data (OMI high resolution) to coarse-resolution data over several years (2002-2012) is calculated; this factor serves as the benchmark for correction.

[0039] Table 6 Explanation of Formula 5 and its parameters

[0040] Formula 5 is used to calculate the final annual resolution correction factor for SCIAMACHY, which combines the OMI baseline ratio and the SCIAMACHY ratio over time; its meaning is: baseline ratio × (annual ratio / average ratio), used to eliminate resolution differences and retain the characteristics of time variation.

[0041] Table 7 Explanation of Formulas 6 and 7 and their parameters

[0042] Formulas 6 and 7 are used to calculate the static resolution correction factor of GOME. Using GOME's own data, the average ratio of high-resolution data (OMI high resolution) to coarse-resolution data over many years (1995-2003) is calculated; this factor serves as the benchmark for correction.

[0043] Table 8. Explanation of Formula 8 and its parameters

[0044] Formula 8 is used to calculate the final annual resolution correction factor for GOME, which combines the baseline ratio of OMI and the change of the GOME ratio over time; its meaning is: baseline ratio × (annual ratio / average ratio), used to eliminate resolution differences and retain the characteristics of time variation.

[0045] Table 9. Explanation of Formula 9 and its parameters.

[0046] Formula 9 is used to calculate the final corrected NO2 column concentration, which involves first standardizing the transit time (systematic bias) and then improving the spatial resolution.

[0047] S013, Data Verification Cross-validation was performed using the results of the ground-based Pandora observation network and the GEOS-Chem model to ensure the accuracy of satellite inversion (error <15%).

[0048] SO2, Quantitative analysis and trend analysis of the weekend effect of ozone; S021, Data Matching and Filtering O3 and NO2 observation data from 23,000 ground stations worldwide were extracted from the TOAR-II database and matched with ERA5 meteorological parameters (temperature, humidity, boundary layer height, etc.) and CAMS aerosol data (PM2.5, PM10).

[0049] Select data from stations with continuous observation for at least 5 years; screen summer data (daily average temperature > 20℃) and exclude stations with significant differences in meteorological conditions between weekend and midweek (p > 0.05); ensure that the NO2 change is significant over the weekend (p < 0.05) to exclude interference from non-emission factors.

[0050] S022 and O3 Weekend Effect Calculation The weekend-midweek difference in O3 was quantified using two indicators: The absolute difference index is expressed as: ... (10) WE-WD O3 represents the difference between the average ozone concentration observed at ground stations over the weekend and the average ozone concentration observed on weekdays; O 3，WE For "weekend ozone concentration", O 3，WD This refers to the ozone concentration on weekdays.

[0051] The weekend-weekday difference in O3 was quantified using a standardized scheme: Sunday was designated as the weekend, and Tuesday through Thursday as weekdays, while Monday and Friday were excluded to minimize the transitional effect between adjacent dates. For each site and the weekly intervals throughout the entire observation period, the average difference in ozone concentration between weekends and weekdays was calculated.

[0052] The probability difference index for exceeding the standard is expressed as: ... (11) Formula 11 represents the statistics of O3 exceedances observed at ground stations; where △Exceedance is the difference in exceedance probability between weekends and weekdays. ，WE>70 ppb This refers to the number of days in a year when the daily O3 concentration exceeds 70 ppb. WE This represents the total number of weekend days in a year. ，WD>70 ppb This indicates the number of days in a year when the daily O3 concentration exceeds 70 ppb. WD This indicates the total number of working days in a year. Similarly, Sunday is designated as the weekend, and Tuesday through Thursday are designated as working days.

[0053] Using Theil-Sen slope estimation combined with the Mann-Kendall trend test, and based on a 5-year sliding window (to reduce the impact of interannual meteorological variations), the long-term trends of WE-WD O3 and ΔExceedance were analyzed. The regional weekend effect evolution patterns were divided into: no trend type, weekend effect decline type, weekend effect enhancement type, and turning point type, and the turning point time point was manually identified.

[0054] Instead of directly calculating a global long-term trend, a 5-year sliding window is used here, aiming to reveal and characterize the evolution of the long-term trend in high resolution through a local dynamic analysis.

[0055] Specifically, this method starts from the beginning of the time series and calculates the short-term trends of 'WE-WDO3' and 'ΔExceedance' sequentially within a five-year sliding window. The window is then moved forward one year and the calculation is repeated until the entire study period is covered. Finally, the calculation results for each window are aligned with its corresponding median year, resulting in a short-term trend sequence arranged chronologically. This sequence clearly demonstrates the changes in the strength and direction of short-term trends over time, and its overall form constitutes the long-term evolution pattern we are interested in.

[0056] Choosing a 5-year window strikes the best balance between reducing the interference of interannual weather fluctuations and maintaining a sufficiently high resolution of trend reversals, thus enabling reliable identification of complex patterns such as 'turning point' and precise location of their occurrence.

[0057] After moving average, Theil-Sen slope estimation (non-parametric trend slope calculation) is performed for time series data ( , (i=1, ..., n), where yi is the annual average O3 value for year i, and xi is the year-time variable for year i (e.g., 2020). The slope estimator β is the median of the slope for all data pairs, expressed as: i <j ……(12) Theil-Sen slope estimation has good noise resistance and is not sensitive to outliers, making it suitable for processing meteorological / pollution data (such as O3 outliers caused by extreme weather).

[0058] Combined with the Mann-Kendall trend test (a nonparametric significance test), the expression for its test statistic S is: ... (13) Where z is the difference between the two data points, i.e. (y j - y i ), is a "placeholder" in Formula 13, used to calculate the S statistic.

[0059] The variance (when there is no repeating data) is calculated as follows: ... (14) The standardized test statistic Z is calculated as follows: ... (15) Significance judgment: like (like =0.05, Z 0.975 If the value is 1.96, then the null hypothesis of no trend is rejected.

[0060] The Mann-Kendall trend test does not require assumptions about the data distribution and is suitable for processing non-normally distributed O3 concentration data.

[0061] Here, Theil-Sen slope estimation and Mann-Kendall trend test are combined to form a nonparametric statistical combination with clear division of labor and complementary advantages, which can robustly quantify trends and determine their statistical significance.

[0062] The Theil-Sen slope estimation is responsible for calculating the magnitude of the trend. It obtains the trend slope by calculating the median of the slopes for all data points. Its core advantage is that it is insensitive to outliers and extreme values ​​in the data, and can reliably quantify the strength of trends. A positive value indicates improvement, while a negative value indicates decline.

[0063] The Mann-Kendall test is responsible for determining the statistical significance of a trend. It does not require the data to follow a specific distribution. It uses hypothesis testing to calculate the p-value to determine whether the trend calculated by Theil-Sen is real or caused by random fluctuations.

[0064] Both methods are used simultaneously within each 5-year sliding window; if the Mann-Kendall test shows a significant trend (p<0.05), then the Theil-Sen slope is used. The sign of the value determines whether the window period is 'enhancing' or 'declining'; if it is not significant, it is determined to be 'no trend'.

[0065] Finally, by observing the slopes of these trends with significance determination ( The continuous changes on the time axis are used to achieve dynamic analysis and classification of long-term trend patterns.

[0066] Specifically, the 'turning point' pattern is defined by identifying persistent and directional changes in the aforementioned short-term trend sequence. For example, when the trend of multiple consecutive windows changes from a significant positive value (enhancing) to a significant negative value (declining), the time period corresponding to the point of change is identified as a trend turning point, and the entire sequence is classified as 'turning point', thus fully depicting the dynamic evolution of the weekend effect from first enhancing to declining (or first declining to enhancing).

[0067] S023, O3 Weekend Effect Cause Assessment Using the same method, the absolute differences between the weekend and midweek of meteorological and pollutant data were calculated, and Theil-Sen slope estimation was performed on them. Combined with the Mann-Kendall trend test, the results were compared with those in step S022 to analyze the influence of meteorological and pollutant data on the evolution of the O3 weekend effect.

[0068] S03. Generate dynamic thresholds; Statistical modeling is performed based on the constructed standardized long-sequence dataset and the weekend effect of ground O3 to establish regional adaptive HCHO / NO2 key threshold intervals; S031, Data Coupling Ground-based stations in regions of the world most affected by anthropogenic pollution emissions were selected (stations with average summer NO2 concentrations > 1.5 × 10⁻⁶). 15molecules / cm²), matching the HCHO / NO2 ratio with the ground hourly average O3 during the OMI satellite transit period (13:00-14:00 LT).

[0069] S032, Statistical Modeling A linear regression model of the HCHO / NO2 ratio and the weekend effect of O3 (WE-WD O3) was established; the key threshold was defined as the intersection of the regression line and WE-WD O3=0, and the key thresholds corresponding to each ground station were obtained.

[0070] Research has revealed a significant negative correlation between satellite-based HCHO / NO2 ratios and the absolute difference in O3 levels between the weekend and midweek. This correlation is qualitatively similar across different cities but varies quantitatively. Figure 2 As shown.

[0071] Assuming the difference in O3 production between the weekend and the weekday is solely attributable to NO X Given the changes in emissions, the crossover point of the HCHO / NO2 ratio at WE-WD O3 = 0 can be considered as the distinction between VOC-controlled emissions and NO2-controlled emissions. X Key thresholds of control mechanisms, such as Figure 2 As shown by the red triangle in the middle.

[0072] S033, Constructing key threshold intervals In real atmospheric environments, there may not be a clear transition between the two O3 chemical mechanisms. Factors such as meteorological conditions and satellite detection noise can blur the boundary between the two, resulting in a wider transition region rather than a clear threshold. Existing studies usually assume a certain range of fluctuation.

[0073] To address the aforementioned issues, this invention employs a regional aggregation strategy, dividing the globe into several key regions: North America, Europe, East Asia, Southeast Asia, South Asia (India), South America, Oceania (Australia), and Southern Africa. Within each key region, several smaller economic areas are further subdivided: for countries with large land areas, key economic regions (such as North China and the Yangtze River Delta in China; California and New York in the United States) are selected for research; while for countries with smaller land areas (such as European countries and Japan), they are analyzed as a whole.

[0074] By statistically analyzing the key thresholds at various surface monitoring stations within the region, a spatially representative range of HCHO / NO2 key thresholds was constructed. Specifically, based on the HCHO / NO2 key thresholds at various surface monitoring stations within the region, a normal distribution was fitted, and the spatially representative key threshold range for the region was defined using the mean ± 3 standard deviations. Simultaneously, the causes of differences in key thresholds across different regions were compared and analyzed, including factors such as emission differences and meteorological conditions.

[0075] S04. Spatiotemporal evolution diagnosis of ozone control areas: using the inflection potential function S(t) to identify trend inflection points, classifying four types of control area evolution patterns, and combining meteorological and aerosol data for correction. S041, Trend and Inflection Point Detection S0411. Define a turning point potential function S(t) to identify the trend inflection point of the HCHO / NO2 ratio (for example, the year corresponding to the turning point from a negative trend to a positive trend (or vice versa) is defined as the trend inflection point), expressed as: ... (16) Among them, B1, B r and B 1+r Let p(B1) and p(B2) represent the linear trends centered on year t for the left of the year ([t-4, t]), the right of the year ([t, t+4]), and the 10-year trend spanning that year ([t-4, t+4]); p(B1) and p(B2) represent the linear trends for the left of the year ([t-4, t]), the right of the year ([t, t+4]), and the 10-year trend spanning that year ([t-4, t+4]). r ) represent the p-values ​​for trend significance. This represents the trend fitting error. A 5-year interval was chosen to mitigate the impact of weather fluctuations and extreme events.

[0076] S0412. For each latitude and longitude grid point (i.e., a single pixel in the satellite data), its annual HCHO / NO2 ratio is used as time series data, and a linear model is used for fitting to quantify its long-term trend, i.e., to determine the parameters B1 and B2 corresponding to B in S(t). r and B 1+r The linear model is represented as: ... (17) Among them, Y t X is the average value in year t. t Let A represent the time variable (year t), A be the average value of the first year, B be the estimated slope of the trend line, and N be the value of the time variable (year t). t This represents the residual (the difference between the fitted value and the actual value).

[0077] If a certain year has the highest S(t) value, and B1, B r Opposite signs, significant trends on both sides (p(B1) and p(B)) r If the value is less than 0.05, it is identified as the turning point year of the ozone control area evolution direction (i.e., the trend inflection point year, denoted as key time point 1).

[0078] Choosing the year with the highest S(t) value ensures that the difference in the trend slope between the two sides of that year is maximized (max|). |), and the 10-year trend fitting error passing through this point. As large as possible. Using this method, the turning point year can be accurately identified with an error of 0.5-1%.

[0079] For latitude and longitude grid points where no turning point is identified, a linear trend is directly fitted. When the trend significance p-value is <0.05, the trend is considered significant; otherwise, the point is considered to have no significant trend change.

[0080] Evolution Type Classification and Identification of Control Zones S042 and O3 Based on the trend characteristics of the HCHO / NO2 ratio and the regionally representative key threshold range obtained in step S03, and fully considering the initial conditions and evolution trend of the O3 mechanism, the evolution types of the ozone control areas (corresponding to NOx control areas or VOC control areas) at the grid points are divided into: Stable type: Always in the typical single NOx / VOC control zone (outside the critical threshold range), no change in O3 chemical control mechanism has occurred (i.e., the HCHO / NO2 ratio has not exceeded the critical threshold, and the ozone control zone has always remained in the NOx control zone or VOC control zone). Quasi-stable: It is in the transitional stage between the NOx control region and the VOC control region (within the critical threshold range), but the O3 chemical control mechanism has not yet changed; Single-transformation type: The HCHO / NO2 ratio shows a linear trend, and there is a significant change in the O3 chemical control mechanism at a certain stage; the criterion for judgment is that the HCHO / NO2 ratio continuously and stably crosses the critical threshold range for several consecutive years (such as 3 years or more) (i.e., from below / above the critical threshold to a stable range above / below the critical threshold), which is judged as a significant change in the O3 chemical control mechanism. Multiple-transformation type: The HCHO / NO2 ratio exhibits a non-linear trend with a trend inflection point, and repeatedly crosses the critical threshold range, indicating multiple shifts in the control mechanism. The criterion is the presence of a statistically significant trend inflection point in the time series, with the value repeatedly and alternately crossing the critical threshold range before and after the inflection point (e.g., in at least two different time periods, the HCHO / NO2 ratio data for several consecutive years are consistently located on different sides of the threshold range). This is considered an indication of multiple shifts in the O3 chemical control mechanism. For grid points of type 4 evolution (multiple transformation type), the inflection point of the HCHO / NO2 ratio usually marks the change in the development direction of the O3 control area, i.e., the key time point 1 mentioned in step S041; for grid points of type 3 and 4 evolution, it is necessary to identify the year in which the HCHO / NO2 ratio stably crosses the key threshold, and record the starting year in which the observation value for the first time is stably located on the other side of the key threshold for many consecutive years (such as 3 years or more) as the key time node for the completion of the transformation of the ozone control area (i.e., the year of threshold crossing, recorded as key time point 2); finally, a gridded ozone generation control mechanism evolution map is generated to show the transformation process of the control mechanism and the key time nodes.

[0081] S043. Assess the explanatory power of O3 control area transitions for long-term surface O3 variations. The evolution type of the ozone control area in step S042 is compared with the regional ozone weekend effect evolution pattern identified in step S022. The evaluation is conducted to determine whether the identified regional ozone weekend effect evolution pattern is consistent with the control area transformation trend and key time points identified based on the satellite HCHO / NO2 ratio. The identification and diagnosis results are then double-verified to improve their reliability.

[0082] When inconsistencies are found, analyze the underlying causes, such as complex precursor sources or the randomness of meteorological conditions.

[0083] On the other hand, in some embodiments of the present invention, the ozone control zone identification and diagnostic system based on multi-source satellite data includes: The satellite data acquisition and processing unit is used to acquire satellite HCHO and NO2 data and construct them into a standardized long-sequence dataset. The ground station data acquisition and processing unit is used to acquire O3 and NO2 observation data from ground stations and calculate the absolute difference and the probability of exceeding the ozone standard between the weekend and midweek. The key threshold interval acquisition unit is used to establish a linear regression model of the HCHO / NO2 ratio and the regional ozone weekend effect based on the standardized long-sequence dataset and the regional ozone weekend effect, and to obtain the key threshold interval of regional HCHO / NO2 based on the linear regression model. The ozone control zone diagnostic unit is used to identify the trend characteristics of the HCHO / NO2 ratio and, based on the trend characteristics of the HCHO / NO2 ratio and the key threshold range of HCHO / NO2 in the region, to perform diagnostic analysis on the evolution type of the ozone control zone.

[0084] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Any simple modifications or equivalent changes made to the above embodiments based on the technical essence of the present invention shall fall within the protection scope of the present invention.

Claims

1. A method for identifying and diagnosing ozone control zones based on multi-source satellite data, characterized in that, The application relates to a method for diagnosing evolution types of ozone control regions. The method comprises the following steps: Satellite HCHO and NO2 data are acquired to construct a standard long-sequence data set; O3 and NO2 observation data of ground stations are acquired, and an ozone weekend-weekday absolute difference is used to represent regional ozone weekend effect, wherein the ozone weekend-weekday absolute difference is the difference between average ozone concentrations in the weekend and in the week; A linear regression model of HCHO / NO2 ratio and regional ozone weekend effect is established according to the standard long-sequence data set and the regional ozone weekend effect, and a regional HCHO / NO2 key threshold interval is obtained through the linear regression model; 2.The method of claim 1, wherein, Trend characteristics of the HCHO / NO2 ratio are identified according to the standard long-sequence data set, and evolution types of ozone control regions are diagnosed and analyzed according to the trend characteristics of the HCHO / NO2 ratio and the regional HCHO / NO2 key threshold interval. The step of constructing the standard long-sequence data set comprises the following steps: S011, HCHO and NO2 column concentration data of four generations of satellites GOME / ERS-2, SCIAMACHY / ENVISAT, OMI / Aura and TROPOMI / Sentinel-5P are acquired; 3.The method of claim 2, wherein, S012, NO2 column concentration data of GOME and SCIAMACHY are corrected based on OMI satellite data. Step S012 comprises the following steps: S0121, transit time correction factors are used to perform transit time correction on SCIAMACHY satellite data and GOME satellite data; 4.The method of claim 1, wherein, S0122, resolution correction factors are used to perform spatial resolution correction on SCIAMACHY satellite data and GOME satellite data. The step of establishing a linear regression model of HCHO / NO2 ratio and regional ozone weekend effect according to the standard long-sequence data set and the regional ozone weekend effect and obtaining a regional HCHO / NO2 key threshold interval through the linear regression model comprises the following steps: S031, monitoring data of ground stations in a region affected by human pollution emissions and exceeding a benchmark value are acquired, and HCHO / NO2 ratio and O3 hourly average value in an OMI satellite transit period are matched; S032, a linear regression model of HCHO / NO2 ratio and regional ozone weekend effect of selected ground stations is established, a key threshold is defined as an intersection of a regression line and WE-WD O3=0, and WE-WD O3 represents an ozone weekend-weekday absolute difference; 5.The method of claim 1, wherein, S033, the global is divided into multiple regions, key thresholds of ground stations in the regions are statistically analyzed, and HCHO / NO2 key threshold intervals of the regions are constructed. The step of identifying trend characteristics of the HCHO / NO2 ratio according to the standard long-sequence data set comprises the following steps: ； where B1, B r and B 1+r represent the linear trend for the left 5 years, right 5 years and 10 years spanning year t, respectively, p(B1) and p(B r ) are the trend significance p-values, is the trend fitting error; S0412、For each latitude and longitude grid point, the HCHO / NO2 ratio value per year is taken as time series data, and a linear model is used for fitting to determine the parameters B1, B r and B 1+r in S(t); the linear model is represented as: ； where Y t is the average value for year t, X t represents the time variable, A is the average value for the first year, B is the estimated slope of the trend line, corresponding to the parameters B1, B r and B 1+r in S(t); N t represents the residuals; S0413、If a certain year has the highest S(t) value, and B1, B r If the sign of the slope is opposite, and the trend is significant on both sides, then the year is identified as a turning point of the evolution direction of the ozone control region, and it is taken as the trend turning point. The grid points that are not identified as turning points are fitted as a linear trend, and the significance of the trend is judged. 6.The method of claim 1 or 5, wherein, S0411, a turning potential function S(t) is defined and is expressed as: The step of diagnosing and analyzing evolution types between NOx control regions and VOC control regions according to the trend characteristics of the HCHO / NO2 ratio and the regional HCHO / NO2 key threshold interval comprises the following steps: When the HCHO / NO2 ratio is always outside the critical threshold interval, indicating that the ozone control zone is always maintained in the NOx control zone or the VOC control zone, the evolution type is determined to be stable; When the HCHO / NO2 ratio is within the critical threshold interval, but the control mechanism in the ozone control zone does not change, the evolution type is determined to be quasi-stable; the method for determining whether the control mechanism changes is that if the HCHO / NO2 ratio continuously and stably crosses the critical threshold interval for many years, it is determined that the control mechanism changes; When the HCHO / NO2 ratio shows a linear trend and the control mechanism in the ozone control zone changes once at a certain stage, the evolution type is determined to be single transition; When the HCHO / NO2 ratio shows a nonlinear trend and there is a trend inflection point, and the control mechanism in the ozone control zone changes many times, the evolution type is determined to be multiple transitions.

7. The method of claim 6, wherein the method further comprises: For grid points of single transition and multiple transitions, identify the years when the HCHO / NO2 ratio stably crosses the critical threshold, and take the starting year when the first continuous multi-year observation value stably locates on the other side of the critical threshold as the key time node when the ozone control zone completes the transition, and generate a grid-based ozone generation control mechanism evolution map. 8.The method of claim 6, wherein, According to the absolute difference and the probability difference of the ozone weekend-weekday, the long-term trend of the absolute difference and the probability difference of the ozone weekend-weekday is analyzed based on a 5-year sliding window using Theil-Sen slope estimation combined with Mann-Kendall trend test, the evolution pattern of regional ozone weekend effect is classified and the turning point is identified; The evolution type of the ozone control zone is verified by comparing the evolution type of the ozone control zone with the evolution pattern of the regional ozone weekend effect. 9.The method of claim 8, wherein, The evolution pattern of the regional ozone weekend effect is divided into no trend type, weekend effect decline type, weekend effect enhancement type and turning type; In each 5-year sliding window, if the Mann-Kendall test shows a significant trend, then according to the positive or negative of the Theil-Sen slope β, the sliding window period is determined to be the weekend effect enhancement type or the weekend effect decline type, if the trend is not significant, it is determined to be the no trend type; The short-term trend in the continuous 5-year sliding window is calculated from the starting point of the time series, the window is slid by one year and the calculation is repeated, the calculation results of each window are aligned with the corresponding median year to obtain a short-term trend sequence in chronological order, and whether it is a turning type is determined by identifying the persistence and directionality change in the short-term trend sequence.

10. A system for identification and diagnosis of ozone control zones based on multi-source satellite data, characterized in that, The method for identifying and diagnosing the ozone control zone based on multi-source satellite data according to any one of claims 1-9, comprising: a satellite data acquisition and processing unit for acquiring satellite HCHO and NO2 data and constructing a standardized long sequence data set; a ground station data acquisition and processing unit for acquiring O3 and NO2 observation data of ground stations and calculating ozone weekend-weekday absolute difference and probability difference; The key threshold interval acquisition unit is configured to establish a linear regression model of the HCHO / NO2 ratio and the regional ozone weekend effect according to the standardized long sequence dataset and the regional ozone weekend effect, and to acquire the regional HCHO / NO2 key threshold interval according to the linear regression model. The ozone control area diagnosis unit is configured to identify the trend feature of the HCHO / NO2 ratio and to diagnose and analyze the evolution type of the ozone control area according to the trend feature of the HCHO / NO2 ratio and the regional HCHO / NO2 key threshold interval.

Images