CN116679356A - Method for predicting PM2.5 and ozone mixed pollution under high space-time resolution - Google Patents
Method for predicting PM2.5 and ozone mixed pollution under high space-time resolution Download PDFInfo
- Publication number
- CN116679356A CN116679356A CN202310580664.9A CN202310580664A CN116679356A CN 116679356 A CN116679356 A CN 116679356A CN 202310580664 A CN202310580664 A CN 202310580664A CN 116679356 A CN116679356 A CN 116679356A
- Authority
- CN
- China
- Prior art keywords
- ozone
- pollution
- concentration
- predicted
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- CBENFWSGALASAD-UHFFFAOYSA-N Ozone Chemical compound [O-][O+]=O CBENFWSGALASAD-UHFFFAOYSA-N 0.000 title claims abstract description 159
- 238000000034 method Methods 0.000 title claims abstract description 37
- 239000013618 particulate matter Substances 0.000 claims description 194
- 230000000875 corresponding effect Effects 0.000 claims description 51
- 238000007637 random forest analysis Methods 0.000 claims description 24
- 238000012544 monitoring process Methods 0.000 claims description 16
- 238000010276 construction Methods 0.000 claims description 13
- 238000012216 screening Methods 0.000 claims description 12
- 238000013507 mapping Methods 0.000 claims description 8
- 230000002596 correlated effect Effects 0.000 claims description 6
- 238000010219 correlation analysis Methods 0.000 claims description 3
- 238000012417 linear regression Methods 0.000 claims description 3
- 238000010586 diagram Methods 0.000 abstract description 10
- 230000001419 dependent effect Effects 0.000 abstract description 4
- 230000002265 prevention Effects 0.000 abstract description 3
- 238000004094 preconcentration Methods 0.000 abstract 1
- 230000009286 beneficial effect Effects 0.000 description 6
- 238000011109 contamination Methods 0.000 description 4
- 238000005070 sampling Methods 0.000 description 4
- 239000004576 sand Substances 0.000 description 4
- 230000007547 defect Effects 0.000 description 3
- 239000003344 environmental pollutant Substances 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 231100000719 pollutant Toxicity 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 230000003044 adaptive effect Effects 0.000 description 2
- 230000001413 cellular effect Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 238000003915 air pollution Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000003090 exacerbative effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000004907 flux Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002195 synergetic effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000003442 weekly effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01W—METEOROLOGY
- G01W1/00—Meteorology
- G01W1/10—Devices for predicting weather conditions
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N15/00—Investigating characteristics of particles; Investigating permeability, pore-volume, or surface-area of porous materials
- G01N15/06—Investigating concentration of particle suspensions
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/0004—Gaseous mixtures, e.g. polluted air
- G01N33/0009—General constructional details of gas analysers, e.g. portable test equipment
- G01N33/0027—General constructional details of gas analysers, e.g. portable test equipment concerning the detector
- G01N33/0036—Specially adapted to detect a particular component
- G01N33/0039—Specially adapted to detect a particular component for O3
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01W—METEOROLOGY
- G01W1/00—Meteorology
- G01W1/02—Instruments for indicating weather conditions by measuring two or more variables, e.g. humidity, pressure, temperature, cloud cover or wind speed
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- Environmental & Geological Engineering (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Environmental Sciences (AREA)
- Pathology (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Ecology (AREA)
- Immunology (AREA)
- Mathematical Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Biodiversity & Conservation Biology (AREA)
- Databases & Information Systems (AREA)
- Mathematical Physics (AREA)
- Pure & Applied Mathematics (AREA)
- Computational Mathematics (AREA)
- Atmospheric Sciences (AREA)
- Mathematical Optimization (AREA)
- Food Science & Technology (AREA)
- Evolutionary Biology (AREA)
- Operations Research (AREA)
- Probability & Statistics with Applications (AREA)
- Bioinformatics & Computational Biology (AREA)
- Algebra (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Medicinal Chemistry (AREA)
- Software Systems (AREA)
- Combustion & Propulsion (AREA)
- Dispersion Chemistry (AREA)
- Computational Linguistics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method for predicting PM2.5 and ozone mixed pollution under high space-time resolution, which comprises the following steps: firstly, taking historical air quality data as a dependent variable, and driving factor data corresponding to the historical air quality data as an independent variable to obtain the optimal PM 2.5 And a concentration prediction model of ozone, then extracting corresponding significantly relevant driving factors by taking a hexagonal grid as a prediction unit, and inputting the driving factors into PM 2.5 And concentration pre-concentration of ozoneObtaining PM at high spatial-temporal resolution in a survey model 2.5 And ozone concentration spatial distribution predictive diagram, finally intersecting them in space to obtain PM 2.5 And a mixed pollution prediction graph of ozone. The prediction method has the advantages of good applicability, good reliability and the like, and can accurately predict PM 2.5 And ozone mixed pollution, for PM in the atmosphere 2.5 And the ozone prevention and control has important guiding significance, high use value and good application prospect.
Description
Technical Field
The invention belongs to the technical field of atmospheric pollution treatment, and relates to a method for treating atmospheric pollutionPM at high spatial-temporal resolution 2.5 And a method for predicting ozone mixed pollution.
Background
PM 2.5 Is one of the most common atmospheric pollutants, and is being widely studied for its significant impact on human health and the environment. In recent years, researchers have proposed a number of control measures to effectively reduce PM 2.5 Concentration. However, although PM 2.5 The falling speed of (2) is very fast, but O 3 But the level of (c) is abnormally increased. Various studies have shown that PM 2.5 The decrease in (a) results in an increase in radiation flux, exacerbating O 3 Pollution, visible, to control PM 2.5 The change of the emission structure of the pollution source caused by the important emission reduction measures can be O 3 The main cause of the continuous rise of pollution. At present, although monitoring is performed by the arranged detection equipment, PM in the area can be acquired 2.5 And O 3 Pollution conditions but lack of PM 2.5 And O 3 Air quality prediction research of mixed action, which leads people to be unable to intuitively judge PM in a certain area 2.5 And O 3 Is a pollution condition of the above-mentioned equipment. Thus, a PM with high spatial-temporal resolution can be accurately predicted 2.5 And O 3 Method for mixing pollution conditions, for studying PM 2.5 And O 3 The synergy and trade-off between these is very necessary.
Machine learning modeling is an emerging technology for atmospheric prediction and management applications, with Land Use Regression (LUR) models and Random Forest (RF) models being the most widely used. For example, in the previous study, the influence relationship between pollutants and driving factors can be obtained by using the LUR model, however, since the number of monitoring point data is less than 40, the requirement of self-variable data for constructing the LUR model cannot be met, and the sampling period of the monitoring points is short, which results in that the monitoring time also has no complete year-by-hour data, the construction of the LUR model is irregular and error-prone, and in particular, the driving factors are selected to be too single and do not completely contain characteristic factors (such as AOI, POI, landscape index and the like) capable of reflecting the urbanization, the model is too one-sided and single, the wide applicability is lacking, the influence of the urbanization on the air quality is difficult to comprehensively evaluate, and therefore, the reliable LUR model with high space-time resolution is difficult to obtain. In addition, the RF model has a black box effect, the influence relation between pollutants and driving factors cannot be accurately seen, and the defect that the single prediction model is poor in reliability and the like still exists. In addition, the prediction grids adopted in the existing prediction method are generally quadrilateral and circular, so that folding and crossing influences can be generated among the prediction grids, defects of inconsistent neighborhood, isotropy, compactness, low sampling rate and the like are caused, and finally, a high-resolution spatial distribution diagram is difficult to construct.
Therefore, how to construct a highly accurate prediction model for accurately predicting PM 2.5 And O 3 Mixed pollution conditions of (2) and development of PM pairs 2.5 And O 3 The intensive research of mixed pollution conditions is of great significance.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide the PM with good applicability and reliability under high space-time resolution 2.5 And a method for predicting ozone mixed pollution.
In order to solve the technical problems, the invention adopts the following technical scheme:
PM (particulate matter) under high space-time resolution 2.5 And a method for predicting ozone mixed pollution, comprising the following steps:
s1, acquiring historical air quality data of a region to be predicted and driving factor data corresponding to the historical air quality data; the historical air quality data comprises PM measured in each hour in the history of monitoring stations in the area to be predicted 2.5 And ozone concentration data; the driving factor data comprise coordinates, meteorological factors, remote sensing factors, population density, land coverage, road density and landscape indexes;
s2, historically measuring PM (particulate matter) per hour of monitoring stations of areas to be predicted 2.5 And ozone concentration data conversion adult average, quarter average, weekend average, working day average and special day average, respectively constructing different from the correlation factors in the driving factor dataPM corresponding to time period 2.5 And LUR model, PM of ozone 2.5 And an improved RF model of ozone;
s3, corresponding PM (particulate matters) in different time periods obtained in the step S2 2.5 And LUR model, PM of ozone 2.5 Respectively verifying with an improved RF model of ozone, and screening a LUR model or an RF model with the maximum determination coefficient R2 and the lowest Root Mean Square Error (RMSE) as PM 2.5 And a concentration prediction model of ozone;
s4, dividing the area to be predicted into a plurality of hexagonal grids, and taking the hexagonal grids as prediction units;
s5, extracting the PM from each hexagonal grid region 2.5 A driving factor that is significantly correlated with a corresponding one of the concentration predictive models of ozone;
s6, connecting each hexagonal grid region with the PM 2.5 Driving factors significantly related to ozone concentration are respectively input to the PM 2.5 Obtaining PM corresponding to each hexagonal grid area in the ozone concentration prediction model 2.5 And the concentration of ozone to obtain PM at high space-time resolution 2.5 A concentration spatial distribution prediction graph and an ozone concentration spatial distribution prediction graph;
s7, PM at high space-time resolution 2.5 The concentration space distribution prediction graph and the ozone concentration space distribution prediction graph are intersected in space to obtain PM in the region to be predicted 2.5 And a mixed pollution prediction graph of ozone.
In the above prediction method, in step S1, the relevant factors in the meteorological factors include temperature, humidity, wind direction and wind speed; the related factors in the remote sensing factors comprise interest point data, interest region data, normalized vegetation indexes, rainfall grid data, lamplight data and building duty ratio data.
In the above prediction method, in step S2, the PM corresponding to the different time periods is further improved 2.5 And the construction method of the LUR model of the ozone comprises the following steps:
(1) PM (particulate matter) 2.5 And annual average, quaternary average, and weekly ozone concentrationsPerforming bivariate correlation analysis on the final average value, the working day average value and the special day average value and the correlation factors in the driving factor data, and screening PM 2.5 A driving factor significantly correlated with ozone concentration;
(2) To be screened and PM 2.5 The driving factors obviously related to the ozone concentration are respectively subjected to stepwise linear regression treatment to construct PM 2.5 Mapping relation between concentration, ozone concentration and obviously related driving factors to obtain corresponding PM in different time periods respectively 2.5 And the LUR model of ozone.
In a further improvement of the above prediction method, in step S2, the PM 2.5 And the construction method of the improved RF model of the ozone comprises the following steps:
(a) PM (particulate matter) 2.5 And annual average, quarternary average, weekend average, workday average and special day average of ozone concentration, and constructing a random forest model with the driving factor data;
(b) Twice screening is carried out through a random forest model, and PM is screened out 2.5 Driving factors significantly related to ozone concentration, construction of PM 2.5 Mapping relation between concentration and ozone concentration and obviously related driving factors to obtain PM respectively 2.5 And an improved RF model of ozone.
In the above prediction method, in step S4, the hexagonal grid is a regular hexagonal grid; the side length of the regular hexagonal grid is 1km.
In step S7, the method is further improved, and the natural breakpoint method is adopted to determine the PM in the corresponding time period 2.5 And O 3 The predicted concentration of (c) is divided into three layers of low, medium and high, and the corresponding pollution conditions are light pollution, medium pollution and high pollution.
In a further improved manner, in step S7, the PM in the area to be predicted 2.5 And the mixed pollution prediction graph of ozone has the following mixed pollution conditions:
first, PM in region to be predicted 2.5 The pollution condition is low pollution and ozone pollutionThe condition is low pollution;
second, PM in the region to be predicted 2.5 The pollution condition of (2) is low pollution, and the pollution condition of ozone is medium pollution;
third, PM in the region to be predicted 2.5 The pollution condition of (2) is low pollution, and the pollution condition of ozone is high pollution;
fourth, PM in the region to be predicted 2.5 The pollution condition of (2) is medium pollution and the pollution condition of ozone is low pollution;
fifth, PM in region to be predicted 2.5 The pollution condition of (2) is medium pollution, and the pollution condition of ozone is medium pollution;
sixth, PM in region to be predicted 2.5 The pollution condition of (2) is medium pollution and the pollution condition of ozone is high pollution;
seventh, PM in region to be predicted 2.5 The pollution condition of (2) is high pollution, and the pollution condition of ozone is low pollution;
eighth, PM in region to be predicted 2.5 The pollution condition of (2) is high pollution, and the pollution condition of ozone is medium pollution;
ninth, PM in region to be predicted 2.5 The pollution condition of (2) is high pollution, and the pollution condition of ozone is high pollution.
Compared with the prior art, the invention has the advantages that:
the invention provides PM with high space-time resolution 2.5 And an ozone mixed pollution prediction method, wherein air quality data obtained by monitoring an air quality micro-station in a complete year time by time are utilized, a prediction model is constructed by combining traditional driving factors and emerging driving factors in seven influencing factors including coordinates, meteorological factors, remote sensing factors, population density, land coverage rate, road density and landscape index, and PM is selected through screening 2.5 Driving factors significantly related to ozone concentration, construction of PM 2.5 Mapping relation between concentration, ozone concentration and obviously related driving factors to obtain corresponding PM in different time periods respectively 2.5 And LUR model, PM of ozone 2.5 And an RF model of ozone, and further, screening by verifying the modelA LUR model or an RF model with the largest determination coefficient R2 and the lowest root mean square error RMSE is obtained as PM 2.5 And a concentration prediction model of ozone, namely a regression model with high space-time resolution; on the basis, the area to be predicted is divided into a plurality of hexagonal grids, the hexagonal grids are used as prediction units, and PM is extracted from the area of each hexagonal grid 2.5 A driving factor which is obviously related to the concentration prediction model of the ozone and is input into PM 2.5 And PM corresponding to each hexagonal grid area is obtained in the ozone concentration prediction model 2.5 And the concentration of ozone to obtain PM (particulate matter) under high space-time resolution corresponding to different hexagonal grid areas 2.5 And finally, PM (particulate matter) in each hexagonal grid area in a corresponding time period is obtained by the concentration space distribution prediction graph and the ozone concentration space distribution prediction graph 2.5 The concentration space distribution diagram and the ozone concentration space distribution diagram are intersected in space to obtain PM in the area to be predicted 2.5 And a mixed pollution prediction graph of ozone. In the invention, the continuous monitoring is adopted to obtain the air quality data as the dependent variable, so that the unnormalization and error of model construction can be effectively avoided, the traditional driving factors and the emerging driving factors in seven corresponding influence factors in the adaptive time are taken as independent variables, the characteristic factors of the area to be predicted can be completely covered, the integrity is better, the adaptability is better, and thus the method is beneficial to constructing the prediction model with higher accuracy, meanwhile, the area to be predicted is divided into a plurality of hexagonal grids, the hexagonal grids are beneficial to constructing the cellular prediction area, the folding and poor folding among the grids are reduced, the sampling rate is improved, and the PM under high resolution is obtained 2.5 And ozone concentration spatial distribution prediction graph, and is simultaneously beneficial to analyzing PM at high resolution 2.5 And ozone, and specifically provides a solution strategy by accurately identifying the mixed pollution conditions in different areas, especially for PM (particulate matter) exposure at the same time 2.5 And O 3 The heavily polluted area can be treated effectively in time, and has important practical significance for effectively treating the atmospheric pollution. PM at high spatial-temporal resolution of the present invention 2.5 Pre-mixing contamination with ozoneThe measurement method has the advantages of good applicability, good reliability and the like, and can accurately predict PM 2.5 And ozone mixed pollution, for PM in the atmosphere 2.5 And the ozone prevention and control has important guiding significance, high use value and good application prospect.
Drawings
FIG. 1 is a PM at high spatial-temporal resolution in example 1 of the present invention 2.5 And a flow chart of a prediction method of ozone mixed pollution.
Fig. 2 is a diagram showing a site distribution diagram for monitoring land utilization and air quality in an area to be predicted in embodiment 1 of the present invention.
FIG. 3 shows ozone and PM corresponding to the time of the annual average and quaternary average in example 1 of the present invention 2.5 A scatter plot of actual concentration and predicted concentration of (c).
FIG. 4 is a graph showing the mean operating day and the mean weekend time of the present invention for ozone and PM 2.5 A scatter plot of actual concentration and predicted concentration of (c).
FIG. 5 shows ozone and PM corresponding to the time of day average in example 1 of the present invention 2.5 A scatter plot of actual concentration and predicted concentration of (c).
FIG. 6 is a PM at high spatial-temporal resolution in example 1 of the present invention 2.5 Annual average and ozone annual average mixed pollution prediction graph.
FIG. 7 is a PM at high spatial-temporal resolution in example 1 of the present invention 2.5 A weekday/weekend mean and an ozone weekday/weekend mean mixed pollution prediction graph.
Detailed Description
The invention is further described below in connection with the drawings and the specific preferred embodiments, but the scope of protection of the invention is not limited thereby. The materials and instruments used in the examples below are all commercially available.
Example 1:
PM (particulate matter) under high space-time resolution 2.5 And the method for predicting the mixed pollution of ozone, the flow diagram of which is shown in figure 1, comprises the following steps:
s1, as shown in FIG. 2, the area to be predicted is the area of the Changsha city.PM (particulate matter) of 144 air quality monitoring stations is continuously and automatically measured per hour by the environment protection agency of Changsha city in 2020 2.5 And O 3 Data (street level), obtaining measured PM per hour 2.5 And ozone concentration data, with them as dependent variables. According to meteorology (northern hemisphere), spring is 3-5 months; summer for 6 months to 8 months; autumn 9 months to 11 months; winter, 12 months to 2 months. The long sand region represents a longer time series level with a mean annual value, a mean seasonal value, and daily mean and hour concentration data representing instantaneous concentration levels.
Meanwhile, taking an air quality monitoring site of a main urban area of a long sand city in 2020 as a circle center, establishing a plurality of buffer areas, and collecting driving factor data corresponding to the main urban area of the long sand city, wherein the driving factor data (self-variable data) corresponding to the long sand city in 2020 can be extracted by establishing four buffer areas (namely 500-2000 meters with an interval of 500 meters), and the driving factor data comprises coordinates, meteorological factors (such as temperature, humidity, wind direction and wind speed), remote sensing factors (such as area of interest data (AOI), point of interest data (POI), normalized vegetation index (NDVI), rainfall grid data, lamplight data and building duty ratio data), population density, land coverage, road density and landscape index.
S2, historically measuring PM (particulate matter) per hour of monitoring stations of areas to be predicted 2.5 And converting the concentration data of ozone into an adult average value, a quarter average value, a weekend average value, a working day average value and a special day average value, and respectively constructing PM corresponding to different time periods with the correlation factors in the extracted driving factor data 2.5 And LUR model, PM of ozone 2.5 And an improved RF model of ozone, specifically:
PM corresponding to different time periods 2.5 And a method for constructing LUR model of ozone, in particular to constructing PM by utilizing R4.1.2 2.5 And O 3 A stepwise multiple regression model of the correlation factor in concentration and driving factor data comprising the steps of:
(1) PM using R4.1.2 version software 2.5 And annual average, quarterly average, weekend average, workday average and special day average of ozone concentration, and drivePerforming bivariate correlation analysis on correlation factors in the dynamic factor data, and filtering PM 2.5 And an insignificant correlation between ozone concentration (P>0.05 Screening out and PM 2.5 A driving factor significantly correlated with ozone concentration.
(2) To be screened and PM 2.5 The driving factors obviously related to the ozone concentration are respectively subjected to stepwise linear regression treatment to construct PM 2.5 Mapping relation between concentration, ozone concentration and obviously related driving factors to obtain corresponding PM in different time periods respectively 2.5 And the LUR model of ozone.
As shown in table 1, the variance expansion factor VIF <10 for the model, which illustrates that none of the models of the present invention have multiple collinearity.
Table 1 corresponding PM over the year period 2.5 And parameters of the LUR model of ozone
TABLE 2 PM corresponding to weekday and weekend time periods 2.5 And parameters of the LUR model of ozone
PM 2.5 And the construction method of the improved RF model of the ozone comprises the following steps:
(a) PM (particulate matter) 2.5 And annual, quarternary, weekend, workday and special day averages of ozone concentration, and constructing a random forest model with the driving factor data, in particularPM for all time scales (year-season-day-weekend-weekday-18 h) using Python 3.10 2.5 And O 3 Concentration) and a corresponding self-variable dataset to construct a random forest model.
(b) Twice screening is carried out through a random forest model, and PM is screened out 2.5 Driving factors significantly related to ozone concentration, construction of PM 2.5 Mapping relation between concentration and ozone concentration and obviously related driving factors to obtain PM respectively 2.5 And an improved RF model of ozone.
S3, corresponding PM (particulate matters) in different time periods obtained in the step S2 2.5 And LUR model, PM of ozone 2.5 Performing ten times cross validation test on the model and the improved RF model of ozone respectively, and screening a LUR model or an RF model with the largest determination coefficient R2 and the lowest Root Mean Square Error (RMSE) as PM 2.5 And a concentration prediction model of ozone.
As shown in FIG. 3, FIG. 3 (a) and FIG. 3 (f) are ozone and PM corresponding to each other in time of average annual value 2.5 A scatter plot of actual concentration and predicted concentration; FIGS. 3 (b) and 3 (g) are ozone and PM corresponding during the spring mean time 2.5 A scatter plot of actual concentration and predicted concentration; FIGS. 3 (c) and 3 (h) are the corresponding ozone and PM during the summer mean time 2.5 A scatter plot of actual concentration and predicted concentration; FIGS. 3 (d) and 3 (i) are ozone and PM corresponding to the autumn mean time 2.5 A scatter plot of actual concentration and predicted concentration; fig. 3 (e) and 3 (j) are scatter plots of actual concentration and predicted concentration corresponding to the winter mean time.
As shown in FIG. 4, FIG. 4 (a) and FIG. 4 (c) are the corresponding ozone and PM for the average time of day 2.5 A scatter plot of actual concentration and predicted concentration; FIGS. 4 (b) and 4 (d) are ozone and PM corresponding during the weekend mean time 2.5 A scatter plot of actual concentration and predicted concentration of (c).
As shown in fig. 5, fig. 5 (a) is a scatter diagram of the actual concentration and the predicted concentration of ozone over the time of day average; FIG. 5 (b) is PM in average time of day 2.5 A scatter plot of actual concentration and predicted concentration of (c).
S4, dividing the area to be predicted (Changsha city) into a plurality of regular hexagonal grids by using corresponding geographic processing software, and taking the regular hexagonal grids as prediction units, wherein the side length of the regular hexagonal grids is 1km.
S5, extracting and PM from each hexagonal grid area 2.5 A driving factor that is significantly correlated with the corresponding concentration prediction model of ozone.
S6, associating PM in each hexagonal grid region 2.5 The driving factors which are obviously related to the concentration of the ozone are respectively input into PM 2.5 Obtaining PM corresponding to each hexagonal grid area in the ozone concentration prediction model 2.5 And the concentration of ozone to obtain PM at high space-time resolution 2.5 An annual average concentration spatial distribution prediction map and an ozone annual average concentration spatial distribution prediction map are shown in fig. 6.
S7, PM at high space-time resolution 2.5 The concentration space distribution prediction graph and the ozone concentration space distribution prediction graph are intersected in space to obtain PM in the region to be predicted 2.5 And a mixed pollution prediction graph of ozone.
In step S7, PM in corresponding time period is obtained by adopting natural breakpoint method 2.5 And O 3 The predicted concentration of (1) is divided into three layers of low, medium and high, the corresponding pollution conditions are light pollution, medium pollution and high pollution, and PM in the same area is identified by means of geographic processing software, space connection and the like 2.5 And O 3 The degree of contamination was further analyzed for PM at high spatial-temporal resolution 2.5 And O 3 Is a synergistic contamination of (a).
FIG. 6 is a PM at high spatial-temporal resolution in example 1 of the present invention 2.5 Annual average and ozone annual average mixed pollution prediction graph. As shown in fig. 6, the black color belongs to a classified high concentration area, that is, a pollution prevention area to be found in the area to be predicted. The white areas belong to areas of no or less contamination. If we go to predict the polluted area of street level from 144 sites alone, the prediction result resolution is low and the prediction result is inaccurate, however, it is obviously impractical to arrange as many sample points as possible to cover the air quality condition of the Changsha city. Because too many arrangements are needed to be moreAnd the operation cost is too high. Therefore, the invention provides a method for obtaining PM with high space-time resolution in long-time market by constructing a model with high space-time resolution 2.5 And O 3 Pollution concentration profile. High concentration PM using geographic processing software 2.5 And O 3 The concentrations overlap in time-space, resulting in PM at high time-space resolution 2.5 And O 3 The concentration distribution area, such as black hexagons of a 1Km hexagonal grid in fig. 7, is what we want to get. Air treatment related personnel can more accurately identify high PM according to the black square 2.5 And high O 3 Meanwhile, the fine area is applied, so that the control requirement of the fine area for air pollution treatment is facilitated. Exploration of PM 2.5 And O 3 The time precision is also particularly important in the mixed pollution exposure condition, so the method researches PM from a plurality of time scales 2.5 And O 3 Is a mixed pollution condition of the above-mentioned materials. FIG. 7 is a PM at high spatial-temporal resolution in example 1 of the present invention 2.5 A weekday/weekend mean and an ozone weekday/weekend mean mixed pollution prediction graph.
From the above results, it can be seen that the present invention provides PM at high spatial-temporal resolution 2.5 And an ozone mixed pollution prediction method, wherein air quality data obtained by monitoring an air quality micro-station in a complete year time by time are utilized, a prediction model is constructed by combining traditional driving factors and emerging driving factors in seven influencing factors including coordinates, meteorological factors, remote sensing factors, population density, land coverage rate, road density and landscape index, and PM is selected through screening 2.5 Driving factors significantly related to ozone concentration, construction of PM 2.5 Mapping relation between concentration, ozone concentration and obviously related driving factors to obtain corresponding PM in different time periods respectively 2.5 And LUR model, PM of ozone 2.5 And an RF model of ozone, and further, by verifying the models, a LUR model or an RF model with the largest determination coefficient R2 and the lowest root mean square error RMSE is selected as PM 2.5 And a concentration prediction model of ozone, namely a regression model with high space-time resolution; on the basis, the region to be predicted is divided into a plurality of regionsHexagonal grids, taking the hexagonal grids as prediction units, and extracting PM (particulate matter) from each hexagonal grid area 2.5 A driving factor which is obviously related to the concentration prediction model of the ozone and is input into PM 2.5 And PM corresponding to each hexagonal grid area is obtained in the ozone concentration prediction model 2.5 And the concentration of ozone to obtain PM (particulate matter) under high space-time resolution corresponding to different hexagonal grid areas 2.5 And finally, PM (particulate matter) in each hexagonal grid area in a corresponding time period is obtained by the concentration space distribution prediction graph and the ozone concentration space distribution prediction graph 2.5 The concentration space distribution diagram and the ozone concentration space distribution diagram are intersected in space to obtain PM in the area to be predicted 2.5 And a mixed pollution prediction graph of ozone. In the invention, the continuous monitoring is adopted to obtain the air quality data as the dependent variable, so that the unnormalization and error of model construction can be effectively avoided, the traditional driving factors and the emerging driving factors in seven corresponding influence factors in the adaptive time are taken as independent variables, the characteristic factors of the area to be predicted can be completely covered, the integrity is better, the adaptability is better, and thus the method is beneficial to constructing the prediction model with higher accuracy, meanwhile, the area to be predicted is divided into a plurality of hexagonal grids, the hexagonal grids are beneficial to constructing the cellular prediction area, the folding and poor folding among the grids are reduced, the sampling rate is improved, and the PM under high resolution is obtained 2.5 And ozone concentration spatial distribution prediction graph, and is simultaneously beneficial to analyzing PM at high resolution 2.5 And ozone, and specifically provides a solution strategy by accurately identifying the mixed pollution conditions in different areas, especially for PM (particulate matter) exposure at the same time 2.5 And O 3 The heavily polluted area can be treated effectively in time, and has important practical significance for effectively treating the atmospheric pollution. PM at high spatial-temporal resolution of the present invention 2.5 The prediction method of ozone mixed pollution has the advantages of good applicability, good reliability and the like, and PM can be accurately predicted 2.5 And ozone mixed pollution, for PM in the atmosphere 2.5 The method has important guiding significance for preventing and controlling ozone, high use value and good application prospect。
The above description is only of the preferred embodiment of the present invention, and is not intended to limit the present invention in any way. While the invention has been described in terms of preferred embodiments, it is not intended to be limiting. Any person skilled in the art can make many possible variations and modifications to the technical solution of the present invention or equivalent embodiments using the method and technical solution disclosed above without departing from the spirit and technical solution of the present invention. Therefore, any simple modification, equivalent substitution, equivalent variation and modification of the above embodiments according to the technical substance of the present invention, which do not depart from the technical solution of the present invention, still fall within the scope of the technical solution of the present invention.
Claims (7)
1. A method for predicting PM2.5 and ozone mixed pollution at high spatial-temporal resolution, comprising the steps of:
s1, acquiring historical air quality data of a region to be predicted and driving factor data corresponding to the historical air quality data; the historical air quality data comprises PM measured in each hour in the history of monitoring stations in the area to be predicted 2.5 And ozone concentration data; the driving factor data comprise coordinates, meteorological factors, remote sensing factors, population density, land coverage, road density and landscape indexes;
s2, historically measuring PM (particulate matter) per hour of monitoring stations of areas to be predicted 2.5 And converting the concentration data of ozone into an adult average value, a quarter average value, a weekend average value, a working day average value and a special day average value, and respectively constructing PM corresponding to different time periods with the related factors in the driving factor data 2.5 And LUR model, PM of ozone 2.5 And an improved RF model of ozone;
s3, corresponding PM (particulate matters) in different time periods obtained in the step S2 2.5 And LUR model, PM of ozone 2.5 Respectively verifying with an improved RF model of ozone, and screening a LUR model or an RF model with the maximum determination coefficient R2 and the lowest Root Mean Square Error (RMSE) as PM 2.5 And a concentration prediction model of ozone;
s4, dividing the area to be predicted into a plurality of hexagonal grids, and taking the hexagonal grids as prediction units;
s5, extracting the PM from each hexagonal grid region 2.5 A driving factor that is significantly correlated with a corresponding one of the concentration predictive models of ozone;
s6, connecting each hexagonal grid region with the PM 2.5 Driving factors significantly related to ozone concentration are respectively input to the PM 2.5 Obtaining PM corresponding to each hexagonal grid area in the ozone concentration prediction model 2.5 And the concentration of ozone to obtain PM at high space-time resolution 2.5 A concentration spatial distribution prediction graph and an ozone concentration spatial distribution prediction graph;
s7, PM at high space-time resolution 2.5 The concentration space distribution prediction graph and the ozone concentration space distribution prediction graph are intersected in space to obtain PM in the region to be predicted 2.5 And a mixed pollution prediction graph of ozone.
2. The method according to claim 1, wherein in step S1, the relevant factors among the meteorological factors include temperature, humidity, wind direction and wind speed; the related factors in the remote sensing factors comprise interest point data, interest region data, normalized vegetation indexes, rainfall grid data, lamplight data and building duty ratio data.
3. The prediction method according to claim 2, wherein in step S2, the PM corresponding to the different time periods 2.5 And the construction method of the LUR model of the ozone comprises the following steps:
(1) PM (particulate matter) 2.5 And annual average value, quarternary average value, weekend average value, working day average value and special day average value of ozone concentration, performing bivariate correlation analysis on the correlation factors in the driving factor data, and screening PM 2.5 A driving factor significantly correlated with ozone concentration;
(2) To be screened and PM 2.5 The driving factors obviously related to the ozone concentration are respectively subjected to stepwise linear regression treatment to construct PM 2.5 Mapping relation between concentration, ozone concentration and obviously related driving factors to obtain corresponding PM in different time periods respectively 2.5 And the LUR model of ozone.
4. The prediction method according to claim 2, wherein in step S2, the PM 2.5 And the construction method of the improved RF model of the ozone comprises the following steps:
(a) PM (particulate matter) 2.5 And annual average, quarternary average, weekend average, workday average and special day average of ozone concentration, and constructing a random forest model with the driving factor data;
(b) Twice screening is carried out through a random forest model, and PM is screened out 2.5 Driving factors significantly related to ozone concentration, construction of PM 2.5 Mapping relation between concentration and ozone concentration and obviously related driving factors to obtain PM respectively 2.5 And an improved RF model of ozone.
5. The prediction method according to any one of claims 1 to 4, wherein in step S4, the hexagonal mesh is a regular hexagonal mesh; the side length of the regular hexagonal grid is 1km.
6. The method according to any one of claims 1 to 4, wherein in step S7, PM under the corresponding time period is used by a natural break point method 2.5 And O 3 The predicted concentration of (c) is divided into three layers of low, medium and high, and the corresponding pollution conditions are light pollution, medium pollution and high pollution.
7. The method according to claim 6, wherein in step S7, the PM in the region to be predicted 2.5 And the mixed pollution prediction graph of ozone has the following mixed pollution conditions:
first, PM in region to be predicted 2.5 The pollution condition of (2) isLow pollution, the pollution condition of ozone is low pollution;
second, PM in the region to be predicted 2.5 The pollution condition of (2) is low pollution, and the pollution condition of ozone is medium pollution;
third, PM in the region to be predicted 2.5 The pollution condition of (2) is low pollution, and the pollution condition of ozone is high pollution;
fourth, PM in the region to be predicted 2.5 The pollution condition of (2) is medium pollution and the pollution condition of ozone is low pollution;
fifth, PM in region to be predicted 2.5 The pollution condition of (2) is medium pollution, and the pollution condition of ozone is medium pollution;
sixth, PM in region to be predicted 2.5 The pollution condition of (2) is medium pollution and the pollution condition of ozone is high pollution;
seventh, PM in region to be predicted 2.5 The pollution condition of (2) is high pollution, and the pollution condition of ozone is low pollution;
eighth, PM in region to be predicted 2.5 The pollution condition of (2) is high pollution, and the pollution condition of ozone is medium pollution;
ninth, PM in region to be predicted 2.5 The pollution condition of (2) is high pollution, and the pollution condition of ozone is high pollution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310580664.9A CN116679356A (en) | 2023-05-22 | 2023-05-22 | Method for predicting PM2.5 and ozone mixed pollution under high space-time resolution |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310580664.9A CN116679356A (en) | 2023-05-22 | 2023-05-22 | Method for predicting PM2.5 and ozone mixed pollution under high space-time resolution |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116679356A true CN116679356A (en) | 2023-09-01 |
Family
ID=87786426
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310580664.9A Pending CN116679356A (en) | 2023-05-22 | 2023-05-22 | Method for predicting PM2.5 and ozone mixed pollution under high space-time resolution |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116679356A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117574061A (en) * | 2024-01-16 | 2024-02-20 | 暨南大学 | PM2.5 and ozone pollution cooperative prevention and control prediction method and system |
-
2023
- 2023-05-22 CN CN202310580664.9A patent/CN116679356A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117574061A (en) * | 2024-01-16 | 2024-02-20 | 暨南大学 | PM2.5 and ozone pollution cooperative prevention and control prediction method and system |
| CN117574061B (en) * | 2024-01-16 | 2024-04-05 | 暨南大学 | PM2.5 and ozone pollution cooperative prevention and control prediction method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Munir et al. | Analysing the performance of low-cost air quality sensors, their drivers, relative benefits and calibration in cities—A case study in Sheffield | |
| Borck et al. | Population density and urban air quality | |
| Huang et al. | Development of land use regression models for PM2. 5, SO2, NO2 and O3 in Nanjing, China | |
| Shi et al. | Investigating the influence of urban land use and landscape pattern on PM2. 5 spatial variation using mobile monitoring and WUDAPT | |
| Donnelly et al. | Real time air quality forecasting using integrated parametric and non-parametric regression techniques | |
| Bravo et al. | Racial isolation and exposure to airborne particulate matter and ozone in understudied US populations: Environmental justice applications of downscaled numerical model output | |
| Data | Guidelines on analysis of extremes in a changing climate in support of informed decisions for adaptation | |
| Chen et al. | Effects of urban green space morphological pattern on variation of PM2. 5 concentration in the neighborhoods of five Chinese megacities | |
| Yang et al. | New method for evaluating winter air quality: PM2. 5 assessment using Community Multi-Scale Air Quality Modeling (CMAQ) in Xi'an | |
| Masiol et al. | Hourly land-use regression models based on low-cost PM monitor data | |
| Stirnberg et al. | Meteorology-driven variability of air pollution (PM 1) revealed with explainable machine learning | |
| Demuzere et al. | A new method to estimate air-quality levels using a synoptic-regression approach. Part I: Present-day O3 and PM10 analysis | |
| CN114371260A (en) | Gridding monitoring, diffusion early warning and tracing method for non-organized VOCs of industrial enterprise | |
| Hua et al. | Improved PM2. 5 concentration estimates from low-cost sensors using calibration models categorized by relative humidity | |
| Lee et al. | An efficient spatiotemporal data calibration approach for the low-cost PM2. 5 sensing network: A case study in Taiwan | |
| CN116679356A (en) | Method for predicting PM2.5 and ozone mixed pollution under high space-time resolution | |
| Liang et al. | Efficient data preprocessing, episode classification, and source apportionment of particle number concentrations | |
| Chen et al. | The influence of neighborhood-level urban morphology on PM2. 5 variation based on random forest regression | |
| CN110503348B (en) | Individual air pollution exposure simulation measurement method based on position matching | |
| Ai et al. | The impact of greenspace on air pollution: Empirical evidence from China | |
| Zhang et al. | Natural and human factors influencing urban particulate matter concentrations in central heating areas with long-term wearable monitoring devices | |
| Hanel et al. | Multi-model analysis of RCM simulated 1-day to 30-day seasonal precipitation extremes in the Czech Republic | |
| Bonilla-Bedoya et al. | Spatiotemporal variation of forest cover and its relation to air quality in urban Andean socio-ecological systems | |
| Achite et al. | Analysis of monthly average precipitation of Wadi Ouahrane basin in Algeria by using the ITRA, ITPAM, and TPS methods | |
| CN110991930B (en) | Method for calculating dust load grade of highway section |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |