NL2030716A9

NL2030716A9 - Method for monitoring nighttime pm2.5 concentration based on nighttime light remote sensing data

Info

Publication number: NL2030716A9
Application number: NL2030716A
Authority: NL
Inventors: Zhu Shanyou; Chen Huijuan; Mo Yaping; Xu Yongming; Wang Guojie
Original assignee: Nanjing Univ Of Information Science And Technology
Priority date: 2021-08-18
Filing date: 2022-01-26
Publication date: 2023-04-14
Also published as: NL2030716B1; CN113390769A; NL2030716A; CN113390769B

Abstract

The present disclosure discloses a method for monitoring nighttime pollutant concentration based on nighttime light remote sensing data, which includes: extracting a total of five datasets respectively about DNB radiance, cloud cover, moon phase angles, satellite zenith angles, and satellite overpass time from NPP/VIIRS nighttime light remote sensing data; preprocessing data in the extracted datasets; acquiring PM2_5 station observation data at the former and latter two 10 integral time points adjacent to the satellite overpass time, and conducting linear interpolation, to obtain the PM2_5 concentration at the satellite overpass time; performing humidity correction for the PM2.5 concentration at the satellite overpass time; and analyzing the relationship between nighttime light radiation received by the remote sensing sensor and the PM2.5 concentration based on the radiative transfer theory, and developing a semi-empirical model which comprehensively 15 considers pixel direct light radiation and background scattered light radiation to estimate nighttime PM2.5 concentration from nighttime light remote sensing data. The present disclosure can monitor the nighttime atmospheric environment, reflect a fine spatial distribution pattern of nighttime air pollution, and provide technical support for atmospheric environment monitoring and regulation at night. 20

Description

METHOD FOR MONITORING NIGHTTIME PM2.5 CONCENTRATION BASED ON

NIGHTTIME LIGHT REMOTE SENSING DATA

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of atmospheric environment monitoring at night, and more particularly relates to a method for monitoring nighttime PM>s concentration based on nighttime light remote sensing data.

BACKGROUND

[0002] PM;.s is one of the main causes of atmospheric pollution, which has an important impact on the atmospheric environment, leads to reduced visibility, and induces respiratory diseases, cardiovascular diseases, and neurodegenerative diseases, seriously endangering human health. Acquisition of an accurate spatial distribution of PM. s concentrations can provide effective data support for air pollution control and abatement. Ground air quality stations are a common means of monitoring PM2s, but a limited number and uneven distribution of the observation stations make it difficult to describe a fine spatial distribution of the PM; 5 concentrations. Spatially continuous and wide-area atmospheric environment information can be obtained by means of satellite remote sensing, which is an effective supplement to ground observation data. Currently, many studies have been conducted to monitor daytime PM; 5s based on the aerosol optical depth {AOD} derived from daytime satellite remote sensing data. However, PMs has significant daily-cycle variation characteristics. The nighttime PM>.s concentration presents different spatial distribution characteristics from those during the day with changes in human activities and meteorological factors. Therefore, the remotely sensed daytime PM:s5 data cannot accurately reflect the spatial pattern of air pollution at night.

[0003] Development of nighttime light remote sensing provides the possibility of monitoring the nighttime PM: concentration. The particulate matter in the atmosphere affects the radiative transfer process of surface light in the atmosphere at night. The nighttime light radiation signal received by a satellite sensor contains the influence of the atmosphere, and therefore, the nighttime PM2s concentration can be deduced from the nighttime light remote sensing data. A few studies have explored the monitoring of the nighttime PM;.s concentration based on the nighttime light data. The main idea thereof is to construct a nighttime PM. s concentration empirical model by using the PMs concentration observed at the station as the dependent variable and the nighttime light remote sensing radiation as the main independent variable, in combination with water vapor and other meteorological environmental variables, and by means of statistical methods such as multiple linear regression, a support vector machine, etc. However, the estimation of the PM. 5s concentration from the nighttime light remote sensing data in the existing studies only considers direct attenuation of nighttime light radiation of current pixels of remote sensing images caused by the atmosphere, but does not consider scattering compensation for nighttime light radiation of surrounding background pixels, leading to large errors in estimation of the nighttime PM, s concentration, especially PM. in dark regions. In addition, the existing studies directly use statistical methods to fit empirical relationships between variables such as the PM2s concentration and the nighttime light remote sensing data, which lacks physical mechanism support and has poor universality.

SUMMARY

[0004] In view of the shortcomings in the prior art, the present disclosure provides a method for monitoring nighttime PM, s concentration based on nighttime light remote sensing data, which gives in-depth analysis of the relationship between the nighttime PM25 concentration and the light radiation received by the remote sensing sensor based on a radiative transfer equation; and constructs, on the basis of simplifying the radiative transfer equation, a nighttime PM: remote sensing monitoring model which comprehensively considers pixel direct light radiation and background scattered light radiation, thus being more reasonable than a method only considering the pixel direct light radiation. Experimental results of the present disclosure manifest that the method proposed by the present disclosure is obviously improved in precision as compared with the current method.

[0005] To achieve the foregoing objective, the present disclosure adopts the following technical solution:

An embodiment of the present disclosure provides a method for monitoring nighttime PMs concentration based on nighttime light remote sensing data, where the monitoring method includes the following steps:

S1. extracting a total of five datasets respectively from NPP/VIIRS nighttime light image data, including DNB radiance, cloud cover, moon phase angles, satellite zenith angles, and satellite overpass time;

S2. preprocessing the extracted datasets, and screening out nighttime light remote sensing data for which the degree of interference of moonlight and cloud cover with surface light is less than a preset interference threshold; performing mask processing for the selected nighttime light remote sensing data by using the cloud cover data, only retaining clear sky pixel values; $3. acquiring PMzs station observation data at the former and latter two integral time points adjacent to the satellite overpass time, and conducting linear interpolation, to obtain the PM>s concentration at the satellite overpass time;

S4. performing humidity correction for the PM. concentration at the satellite overpass time;

S5. analyzing the relationship between nighttime light radiation received by the remote sensing sensor and the PMs concentration based on the radiative transfer theory, where a radiative transfer equation for nighttime surface light is expressed as follows: i gino w fir £8 = Sams pie, fj Z { Jen 7 ESL 0 where T is the vertical optical thickness measured downwards from the upper boundary of the atmosphere, w is the single-scattering albedo, Qo denotes an initial direction of incidence, 0 denotes a single-scattering propagation direction, Q' denotes a propagation direction after multiple scattering events, Ho denotes the cosine of the emission zenith angle of surface light, Fo denotes the radiance scattered by the surrounding surface light sources, namely, the background scattered fr radiation, and P denotes a phase function; Ur denotes a total change in the radiance after the radiance dl propagates in a certain direction through an optical thickness of dt, Hz, Q denotes direct attenuation of surface light, 32 = © denotes single scattering of the surface light, and 33” denotes multiple scattering events of the surface light;

S6. solving the radiative transfer equation for the nighttime surface light, to obtain a theoretical derivation model of the nighttime PMs concentration: 3 where / denotes a sum of radiance after the radiation intensity at the surface passes through the entire atmosphere to the sensor and is attenuated, namely, the nighttime light radiation observed by the satellite; lo denotes the near-surface radiance, namely, the upward radiation of surface light sources; Lo denotes the cosine of the zenith angle; and P(cos9} is the scattering phase function, the aerosol scattering phase function P{cosô} being as follows: where the asymmetry factor g is 0.75, and the angle 0 between the emitted beam and the scattered beam is the satellite zenith angle plus 90°;

S7. based on the theoretical derivation model of the nighttime PM; concentration, constructing a nighttime PM; concentration estimation semi-empirical mode} as follows: where PM: 5 _f(RH)} is the humidity-corrected PM2.5 concentration; po is the cosine of the satellite zenith angle; Ip is the surface upward radiance; I is the DNB radiation value received by the satellite;

P{cosd} is the scattering phase function; Fo is the background scattered radiance; and a, b, and ¢ are empirical coefficients, which are obtained by means of least squares fitting;

S8. applying the constructed nighttime PMs concentration estimation semi-empirical model in spatial independent variables, to obtain a spatial distribution of the humidity-corrected PM2s concentrations; and

S9. inversely transforming the humidity-corrected PMs concentration into the PM>s concentration, to obtain the following PM:2.s concentration spatial distribution:

where f(RH} is the humidity correction factor.

[0006] Alternatively, in step S2, the data in the extracted datasets is preprocessed as follows:

S21. performing preprocessing, such as projection conversion, tessellation, and cropping, for the 5 extracted datasets; and

S22. screening out data with the moon phase angle less than 120° and low cloud cover.

[0007] Alternatively, in step 54, humidity correction is performed for the PM.s concentration at the satellite overpass time by using the following formula: where PM s_f(RH} is the PM25 concentration after humidity correction, PMys is the station PMzs concentration, and RH is the ground relative humidity (34).

[0008] Alternatively, in step S6, a process of solving the radiative transfer equation for the nighttime surface light includes the following steps:

S61. expressing the radiative transfer equation for the nighttime surface light according to the relationship between the surface radiance Io and the radiance | received by the sensor:

S62. letting the surface light perpendicularly strike and enter the sensor, where the zenith angle of the surface light is denoted by the satellite zenith angle, and the azimuth angle of the surface light approximates 0; neglecting the angle at which the surface light enters the sensor after scattering, and simplifying the radiative transfer equation for the nighttime surface light in step S61: $63. performing a logarithm operation at both sides of the formula simplified in step 562, to obtain: where the optical thickness T is expressed as a function regarding an effective height of the boundary layer, mass extinction efficiency, the PM2s concentration, and relative humidity: r= PV, , SRY. A 2.8 Sent where H is the effective height of the boundary layer, and Ques is the mass extinction efficiency; and

S64. neglecting the effective height of the boundary layer and the mass extinction efficiency, and setting the aerosol single scattering reflectivity w to 0.95, to obtain the following theoretical derivation model of the nighttime PM; 5 concentration: ay 5 Â GN JA SE A iNAj Ne

[0009] The present disclosure has the following beneficial effects:

For the shortcoming that the existing nighttime PM:s concentration remote-sensing monitoring method only considers direct attenuation of nighttime light radiation of remote sensing pixels caused by the atmosphere but does not consider a physical mechanism, the present disclosure provides a method for monitoring the nighttime pollutant concentration by using NPP/VIIRS nighttime light remote sensing images, which analyzes the relationship between the nighttime light radiation and the PM:25 concentration based on the radiative transfer theory, and based on the analysis, develops a semi-empirical model which considers the impacts of the pixel direct radiation and the background scattered radiation on the nighttime light radiation received by the satellite to estimate nighttime PM2s5 concentration. The present disclosure can effectively monitor the nighttime PM;s5 concentration from the NPP/VIIRS nighttime light remote sensing images, and make up for the deficiency that the existing method is inapplicable to dark regions, thus being applicable to monitoring of atmospheric environment at night, reflecting a fine spatial distribution pattern of nighttime air pollution, and providing technical support for atmospheric environment monitoring and regulation at night.

BRIEF DESCRIPTION OF THE DRAWINGS

[0010] FIG. la+b is a flowchart of a method for monitoring nighttime PM2s5 concentration based on nighttime light remote sensing data in an embodiment of the present disclosure;

[0011] FIGs. 2a to 2d are scatter plots between the humidity-corrected PM; 5s concentration and nighttime light radiation at ground monitoring stations in an embodiment of the present disclosure, where FIG. Za shows an industrial park station, FIG. 2b shows a Huaihe Bridge station, FIG. 2c shows a monitoring station in Huaiyin District, and FIG. 2d shows a city monitoring station;

[0012] FIG. 3 is a scatter plot between PM:2.s concentration estimated by the method of the present disclosure and actually measured PM: s concentration in an embodiment of the present disclosure;

[0013] FIG. 4 is a spatial distribution map of nighttime PM: concentration in Huai'an City in an embodiment of the present disclosure; and

[0014] FIG. 5 is a scatter plot between PM; concentration estimated by the current method and actually measured PMas concentration in an embodiment of the present disclosure.

DETAILED DESCRIPTION OF THE EMBODIMENTS

[0015] The present disclosure is further described in detail below with reference to the accompanying drawings.

[0016] It should be noted that the terms such as "upper”, "lower", "left", "right", "front", "rear", etc., cited in the present disclosure are only used for ease of clear description, and not intended to limit the implementable scope of the present disclosure. The change or adjustment of their relative relationships shall be regarded as the implementable scope of the present disclosure without substantial changes to the technical content.

[0017] FIG. 1a+b is a flowchart of a method for monitoring nighttime PM2s concentration based on nighttime light remote sensing data in an embodiment of the present disclosure. The monitoring method is a nighttime PM.s concentration remote sensing estimation semi-empirical method constructed by comprehensively considers pixel direct light radiation and background scattered light radiation on the basis of simplifying a radiative transfer equation for nighttime light.

Referring to FIG. 1a+b, the monitoring method specifically includes the following steps: 1) Processing of remote sensing data

A total of five datasets respectively about DNB radiance, cloud cover, moon phase angles, satellite zenith angles, and satellite overpass time, which are received by a sensor are extracted from

NPP/VIIRS nighttime light image data; and preprocessing, such as projection conversion, tessellation, and cropping, is performed for the extracted datasets. In order to eliminate the interference of moonlight and cloud cover with the surface light, data with the moon phase angle less than 120° and low cloud cover is screened out; and mask processing is performed for the selected nighttime light remote sensing data by using the cloud cover data, only retaining clear sky pixel values.

[0018] 2) Ground data processing

PM. 5 station observation data at the former and latter two integral time points adjacent to the satellite overpass time is acquired, and linear interpolation is conducted, to obtain the PM;s concentration at the satellite overpass time.

[0019] Moisture absorption growth is an important characteristic of PM:s, The particle diameter, spatial distribution, and optical characteristics of fine particle matter change significantly under different relative humidity. In order to reduce the impact of particle hygroscopicity on the relationship between the PM>2s concentration and atmospheric optical properties, humidity correction is performed for PMs as follows: where PM:.s_f(RH) is the PM2s concentration after humidity correction, PM; is the station PM:.s concentration, and RH is the ground relative humidity {%).

[0020] 3} Theoretical derivation based on a radiative transfer equation

The relationship between nighttime light radiation received by the remote sensing sensor and the

PM: concentration is analyzed based on the radiative transfer theory, and a radiative transfer equation for nighttime surface light can be expressed as follows: 8 she = Sir 3 A Sea, £38 ha { Sn TNO ye W

In the foregoing equation, T is the vertical optical thickness measured downwards from the upper boundary of the atmosphere, w is the single-scattering albedo, OQ, denotes an initial direction of incidence, a denotes a single-scattering propagation direction, Q' denotes a propagation direction after multiple scattering events, Ho denotes the cosine of the emission zenith angle of surface light,

Fo denotes the radiance scattered by the surrounding surface light sources, namely, the background scattered radiation, and P denotes a phase function. The left side of the equation indicates a total change in the radiance after the radiance di propagates in a certain direction through an optical thickness of dt. The first term on the right side of the equation indicates the direct attenuation of the surface light, the second term indicates single scattering of the surface light, and the third term indicates multiple scattering events of the surface light.

[0021] FIGs. 2a to 2d are scatter plots between humidity-corrected PM; concentration and remote-sensing nighttime light radiation at four typical monitoring stations in Huai'an City and give a mean nighttime light radiation lmean of each station, which are used for analyzing the relationship between the nighttime light data and the PM:s concentration observed at the stations. The industrial park station and the Huaihe Bridge station are located in suburban regions with very low nighttime light brightness, while the monitoring station in Huaiyin District and the city monitoring station are located in urban regions with high nighttime light brightness. The stations in the urban regions have relatively high brightness values, and then the PMs and the remote-sensing nighttime light radiation show a negative correlation; while the stations in the suburban regions have relatively low brightness values, and then the PM2s and the remote-sensing nighttime light radiation show a positive correlation. The reason is that high-brightness regions have relatively strong nighttime light radiation, and the nighttime light radiation received by the satellite is mainly from the pixels themselves, rather than the background scattered radiation of the surrounding pixels; but low-brightness regions have relatively weak nighttime light radiation, and a large proportion of the nighttime light radiation received by the satellite is from the background scattered radiation of the surrounding pixels. It can be seen from the above that the estimation of the PM;s concentration from the nighttime light remote sensing data cannot only consider the pixel direct radiation, but must also consider the influence of the background scattered radiation.

[0022] Formula 2 can be expressed as follows according to the relationship between the surface radiance Io and the radiance | received by the sensor: 5 ft Ax where / denotes a sum of radiance after the radiation intensity at the surface passes through the entire atmosphere to the sensor and is attenuated, namely, nighttime light radiation observed by the satellite; Ip denotes the near-surface radiance, namely, the upward radiation of surface light sources; Lo denotes the cosine of the zenith angle; and P(cosB)} is the scattering phase function.

[0023] When the surface light perpendicularly strikes and enters the sensor, the zenith angle of the surface light can be denoted by the satellite zenith angle, and the azimuth angle of the surface light approximates 0. The angle at which the surface light enters the sensor after scattering has little impact, and can be neglected. Thus, formula 3 can be simplified as follows:

OF mein 23% . ha ee #8

ST

A logarithm operation is performed at both sides of formula 4 at the same time, to obtain:

EE EB I A} (5)

The optical thickness T is expressed as a function regarding an effective height of the boundary layer, mass extinction efficiency, the PM; concentration, and relative humidity: oo DHE FREY EF (8)

LO £8 9 g LAGE J . + LE where H is the effective height of the boundary layer, and Qmest is the mass extinction efficiency.

[0024] The numeric values of the effective height of the boundary layer and the mass extinction efficiency have little changes and can be neglected. According to the previous study, the aerosol single scattering reflectivity w is set to 0.95, Formulas 5 and 6 are combined to obtain: = x

A calculation formula of the aerosol scattering phase function P(cos8) is as follows:

Picos 8} = mmm ss iS) where the asymmetry factor g is 0.75, and the angle 8 between the emitted beam and the scattered beam is the satellite zenith angle plus 90°.

[0025] According to the radiative transfer derivation process, the nighttime PM; 5s concentration can be expressed as a function regarding the nighttime light radiance received by the satellite, the cosine of the satellite zenith angle, the surface upward radiance, the scattering phase function, and the background scattered radiance.

[0026] 4) Model construction

Because the nighttime surface light radiates upwards and the background scattered radiation cannot be accurately calculated, the theoretical model cannot be directly used to calculate the

PMs concentration. Based on the previous theoretical derivation model, a nighttime PMs concentration estimation semi-empirical model is constructed as follows: where PM, s_f(RH) is the humidity-corrected PM, s concentration; po is the cosine of the satellite zenith angle; lg is the surface upward radiance; | is the DNB radiation value received by the satellite;

P{cos8} is the scattering phase function; Fo is the background scattered radiance; and a, b, and c are empirical coefficients, which are obtained necessarily by means of least squares fitting.

[0027] In calculation of the nighttime PMs concentration according to the foregoing semi-empirical model, values of independent variables such as the cosine of the satellite zenith angle, the surface upward radiance, the DNB radiation value received by the satellite, the scattering phase function, and the background scattered radiance need to be acquired. The nighttime light radiance received by the satellite, the cosine of the satellite zenith angle, and the scattering phase function are deduced and calculated according to cloudless and moonless NPP/VIIRS data obtained by screening; the surface light upward radiance is approximately represented by a DNB radiation value on a cloudless and moonless night with the lowest PM:2s concentration in a monitored region; and the background scattered radiance is approximately represented by an average radiance value in a certain spatial range with the current pixel as the center. In order to determine the optimal spatial range of background pixels, mean background pixel radiation values in different spatial ranges are separately calculated as the background radiation values. The change of model precision with the background pixel range is analyzed by means of 10-fold cross-validation, and a mean radiation value in a spatial range corresponding to the highest precision is calculated as the background scattered radiation.

[0028] A remote-sensing nighttime light radiation value on the cloudless and moonless night with the lowest PMs concentration in a study area approximately represents the surface light upward radiance. Based on the semi-empirical model {formula 9) deduced through the radiative transfer theory, by using the humidity-corrected PMs as the dependent variable, and the nighttime light radiance of the station's corresponding pixels, the cosine of the satellite zenith angle, the surface upward radiance, the scattering phase function, and the background scattered radiance as the independent variables, values of the coefficients a, b, and c are calculated by means of least squares fitting, to construct a remote sensing estimation model of the nighttime humidity-corrected PMs concentration. In order to determine the optimal spatial range of background pixels, statistics on the mean spatial independent variables in different spatial ranges with the current pixel as the center are made, the change in estimation accuracy of the humidity-corrected PM: by the semi-empirical model in different background spatial ranges is analyzed, and the background pixel spatial range corresponding to the highest precision is determined as the optimal spatial range. In this instance, the optimal spatial range is within pixels of 75x75, and an average radiation value in this range is calculated as the background scattered radiation.

[0029] 5) Calculation of PM2.s concentration

By using PM;s after humidity correction as the dependent variable, and the nighttime light radiance of the station's corresponding pixels, the cosine of the satellite zenith angle, the surface upward radiance, the scattering phase function, and the background scattered radiance as the independent variables, values of the coefficients a, b, and c are calculated by means of least squares fitting, to construct a nighttime PM; 5s concentration remote sensing monitoring model. By applying the constructed model in the spatial independent variables, a spatial distribution of the humidity-corrected PM 5 concentrations is obtained.

[0030] Finally, the humidity-corrected PM2s5 concentration is inversely transformed into the

PM: concentration, to obtain the following PM. concentration spatial distribution: og PM} (RH) 10 where f{RH} is the humidity correction factor.

[0031] The model constructed based on the optimal spatial range is used as a final remote sensing estimation model of the nighttime humidity-corrected PM: s concentration. By applying the model in the spatial independent variables such as the nighttime light radiance, the cosine of the satellite zenith angle, the surface upward radiance, the scattering phase function, and the background scattered radiation, the humidity-corrected PMas concentration in the study area is calculated. Then, the estimated humidity-corrected PM25 concentration is inversely transformed into the PM; 5 concentration according to formula 10, to obtain the spatial distribution of the PM, 5 concentrations, FIG, 3 is a scatter plot of the cross validation accuracy obtained by applying the method of the present disclosure in remote sensing monitoring of nighttime PMas in Huai'an City.

The samples are mainly distributed around the 1:1 line, where R? is 0.69, and the RMSE and the

MAE are respectively 36.31 ug/m? and 25.27 ug/m?, which indicates that the estimation precision is high. FIG. 4 shows a spatial distribution of the nighttime PM; 5 concentrations from September to

December in 2019 in Huai'an City that is obtained by applying the method of the present disclosure.

The PMs; concentrations mainly range from 60 ug/m? to 170 ug/m3, which shows a significant spatial difference. The PM:5 concentrations in the main urban region are generally above 140 ug/m?3, which are obviously higher than the surrounding regions; and the area of the high-value region is relatively large. The urban regions of county districts such as Lianshui County, Hongze

District, and Jinhu County also show relatively high PMs concentrations, generally above 120 ug/m?, but the areas are far smaller than that of the main urban region. In addition to the main urban region and the urban regions of county districts, some regions with high PM;s concentrations are distributed in a strip mainly along the highway. The PM:25 concentrations in suburbs are generally low and below 100 pg/m?3.

[0032] In order to make comparison between the method of the present disclosure and the current method without considering the impact from the background scattering and the radiative transfer mechanism, the current method is also applied in the estimation of the nighttime PMas concentration in Huai'an City. FIG. 5 shows a scatter plot of the cross validation accuracy of the current method. It can be seen from the graph that, the samples are scattered, with many samples deviating from the 1:1 line, especially in the regions with high and low PM:s values. R? is 0.48, and the RMSE and the MAE are respectively 51.23 ug/m? and 31.67 ug/m3, which indicates that the estimation precision is obviously lower than that of the method of the present disclosure.

[0033] The estimation result of the present disclosure shows that, the semi-empirical model constructed based on the radiative transfer equation can estimate the nighttime PM2s5 concentration from the nighttime light remote sensing data and reflect a fine spatial distribution of nighttime air pollution, and has a monitoring precision obviously superior to that of the current method, thus providing technical support for atmospheric environment monitoring and regulation at night.

[0034] The above merely describes the preferred embodiments of the present disclosure, but the protection scope of the present disclosure is not limited to the above-described embodiments. All technical solutions that fall under the idea of the present disclosure belong to the protection scope of the present disclosure. It should be noted that, several improvements and modifications may be made by those of ordinary skill in the art without departing from the principle of the present disclosure, and these improvements and modifications should also be construed as falling within the protection scope of the present disclosure.

Claims

CONCLUSIONS

1. Method to monitor PM:s concentration at night based on remote night light detection data, where the monitoring method includes the following steps:

S1. extract from NPP/VIIRS nightlight image data five sets of data on DNB radiation, cloud cover, lunar phase angles, satellite highest position angles, and time passing satellite respectively, which are received by a sensor;

S2. process data in extracted datasets and remotely screen nightlight detection data for which the degree of disturbance of moonlight and cloud cover with surface light is less than a preset threshold; process masks for the selected remote night light detection data using the cloud cover data and keep only open sky pixel values;

53. Acquire PM>s observation data at the station at the first and last two integral time points adjacent to the satellite overtime and perform linear interpolation to obtain the PM:.s concentration during the satellite overtime ;

S4. perform humidity correction for PM;.5 concentration during satellite transit time;

SS. analyze the relationship between night light radiation received by the remote sensing sensor and the PM:5 concentration based on radiative transfer theory, expressing an equation of radiative transfer for surface light at night as follows: where t is the vertical optical thickness measured down from the is upper limit of the atmosphere, w is the one-time scattering albedo, Qp denotes an initial direction of incidence, 0 denotes a one-time scattered motion direction, 0' denotes a direction of motion after multiple scattering events, po is the cosine of the angle of the crest position of the surface light emission indicates, Fo indicates the radiation scattered by the surrounding surface light sources, namely, the scattered background radiation, and P a

Sr, Qu indicates phase function; 4 indicates a complete change in the radiation after the radiation dl-moves in a certain direction through an optical thickness of dt, * *“* *“ indicates direct attenuation of surface light, *% one-time scattering of the FF i od I = TC. J NE indicates surface light and TR indicates multiple scattering events of the surface light;

56. solve the equation of radiative transfer for the surface light at night to obtain a theoretical derivation model of the PM:5 concentration at night: {SD se © NE 5 Sip SURES OD Jd SUNG S fy BIA sss} where / is a sum of the radiation indicates after the radiation intensity at the surface passes through the entire atmosphere to the sensor and is attenuated, namely the light radiation observed by the satellite at night; Ip denotes the radiation at the surface, namely the radiation upwards from surface light sources; uo denotes the cosine of the angle of the highest position; and P{cos3} is the scattering phase function, where the aerosol scattering phase function P{cosô) is: where the asymmetry factor g is 0.75 and the angle between the emitted beam and the scattered beam is the angle of the satellite's highest position plus 90°;

S7. based on the theoretical derivation model of the PM25 concentration at night, construct a semi-empirical estimation model for the PM as follows: 5 concentration at night: Ln DDS Dies 838 = where PM25_f(RH)} is the humidity-corrected PM; s concentration is; po is the cosine of the angle of the satellite's highest position; lp is the surface radiation upwards; / de

DNB radiation value received by the satellite is; P{cosd) is the scattering phase function; Fo is the background scattering radiation; and a, b and c are empirical coefficients, which. are obtained from least squares that fit;

S8. apply the composite semi-empirical estimation model for the PM25 concentration at night in the space-independent variables to obtain a spatial distribution of the humidity-corrected PM25 concentrations; and

S9. reverse the humidity-corrected PM>.5 concentration to the PM;5 concentration to obtain the following spatial distribution of PM,s concentration: ee TREY BA is TT “FRED where f{RH} is the humidity correction factor.

A method to monitor PMs concentration at night based on remote night light detection data according to claim 1, wherein in step S2, the data in the extracted data sets is pre-processed as follows:

S21. do pre-processing, such as projection conversion, tessellation, and cropping, for the extracted datasets; and

522. screen data with moon phase angle less than 120° and low cloud cover.

The method of monitoring PM:2.s concentration at night based on remote night light detection data according to claim 1, wherein in step S4, a humidity correction is performed for the PM:s concentration during the time of passing from a satellite using the following formula: where PM2.5_f(RH} is the PM;s concentration after humidity correction, PM25 is the PM25 concentration at the station, and RH is the ground relative humidity {9%}.

The method of monitoring PM:5 concentration at night based on remote night light detection data according to claim 1, wherein in step S6, a process to solve the equation of radiative transfer for the surface light at night has the following steps includes:

S61. express the equation of radiative transfer for the surface light at night according to the relationship between the surface irradiance Io and the radiance | received by the sensor: Js V,

$62. beam the surface light perpendicularly and enter the sensor, with the angle of the top position of the surface light indicated by the angle of the top position of the satellite and the azimuth angle of the surface light being around 0; Ignore the angle the surface light enters the sensor after scattering and simplify the equation for radiative transfer for the surface light at night in step S61: Ne RES FIT misses § aR

$63. perform a logarithmic operation on both sides of the formula simplified in step S62, to obtain: * ° zE where the optical thickness T is expressed as a function of an effective height of the boundary layer, mass extinction efficiency, the PM2.5 concentration and relative humidity: Poe DI AND BN) i LEER g + YER SY sort where H is the effective height of the boundary layer and Qmes: is the mass extinction efficiency; and

S64. ignoring the effective boundary layer height and mass extinction efficiency and setting the aerosol single scatter reflection w to 0.95 to obtain the following theoretical derivation model of PM25 concentration at night: