CN109709577A - A kind of Three-dimensional Variational Data Assimilation method of the aerosol LIDAR inverting PM2.5 based on WRF-Chem mode - Google Patents
A kind of Three-dimensional Variational Data Assimilation method of the aerosol LIDAR inverting PM2.5 based on WRF-Chem mode Download PDFInfo
- Publication number
- CN109709577A CN109709577A CN201811621885.1A CN201811621885A CN109709577A CN 109709577 A CN109709577 A CN 109709577A CN 201811621885 A CN201811621885 A CN 201811621885A CN 109709577 A CN109709577 A CN 109709577A
- Authority
- CN
- China
- Prior art keywords
- data
- assimilation
- inverting
- wrf
- observation
- 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
- 238000000034 method Methods 0.000 title claims abstract description 58
- 239000000443 aerosol Substances 0.000 title claims abstract description 35
- 239000011159 matrix material Substances 0.000 claims abstract description 14
- 238000012544 monitoring process Methods 0.000 claims abstract description 13
- 230000002123 temporal effect Effects 0.000 claims abstract description 7
- 230000008033 biological extinction Effects 0.000 claims description 14
- 238000007781 pre-processing Methods 0.000 claims description 9
- 238000009826 distribution Methods 0.000 claims description 7
- 239000003344 environmental pollutant Substances 0.000 claims description 7
- 231100000719 pollutant Toxicity 0.000 claims description 7
- 230000006870 function Effects 0.000 claims description 6
- 239000013618 particulate matter Substances 0.000 claims description 6
- 230000003068 static effect Effects 0.000 claims description 6
- 238000012360 testing method Methods 0.000 claims description 6
- 230000006855 networking Effects 0.000 claims description 5
- 238000012549 training Methods 0.000 claims description 5
- 238000002360 preparation method Methods 0.000 claims description 4
- 238000013441 quality evaluation Methods 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 4
- 238000007619 statistical method Methods 0.000 claims description 4
- 238000003860 storage Methods 0.000 claims description 4
- 238000004140 cleaning Methods 0.000 claims description 3
- 238000004088 simulation Methods 0.000 claims description 3
- 238000006467 substitution reaction Methods 0.000 claims description 3
- 230000005856 abnormality Effects 0.000 claims description 2
- 239000003102 growth factor Substances 0.000 claims description 2
- 238000013480 data collection Methods 0.000 claims 1
- 230000003287 optical effect Effects 0.000 claims 1
- 230000001932 seasonal effect Effects 0.000 abstract description 2
- 239000002245 particle Substances 0.000 description 5
- 230000008859 change Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000007547 defect Effects 0.000 description 3
- 230000007812 deficiency Effects 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 239000000809 air pollutant Substances 0.000 description 1
- 231100001243 air pollutant Toxicity 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000009792 diffusion process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000013213 extrapolation Methods 0.000 description 1
- 239000003292 glue Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 230000035699 permeability Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 238000003908 quality control method Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
- 238000005303 weighing Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Optical Radar Systems And Details Thereof (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The Three-dimensional Variational Data Assimilation method of the invention discloses a kind of aerosol LIDAR inverting PM2.5 based on WRF-Chem mode, belong to environment monitoring technical field, after laser radar AOD inverting assimilation, forecast is significantly improved with observation correlation, accuracy, average deviation also declines to a great extent, analog result is closer to observation, and each grade scoring result is also significantly increased after assimilation, forecasts that the class of pollution and fact are closer after showing assimilation.The assimilation method is succinctly direct, using laser radar data, the quality of data is reliable, spatial and temporal resolution is higher, considers simulated domain local features and Seasonal variation, constructs dynamic background field error covariance matrix, improve mode applicability, it is corrected, is reasonably adjusted with humidity by highly correcting, improved laser radar AOD and accuracy is forecast to PM2.5.
Description
Technical field
The present invention relates to a kind of Three-dimensional Variational Data Assimilation methods, molten more particularly to a kind of gas based on WRF-Chem mode
The Three-dimensional Variational Data Assimilation method of glue laser radar inverting PM2.5, belongs to environment monitoring technical field.
Background technique
With industrialization, the rapid development of urbanization, cities in China is in recent years frequently by gray haze pollution effect；
Spring and summer replaces season, causes northern sandstorm, causes to seriously endanger to urban air-quality, visibility.
Gray haze, sandstorm diffusion, caused by Air Pollutants be PM10 and PM2.5, the most for particulate matter at present
Accurate monitoring means be weighing is collected using ground particulate sampler, but the limited amount of particle sampler and
It is mainly distributed on more flourishing urban area, western and northern city is rare, leads to monitor particulate matter transmission key area and big
The variation of range of particle object concentration is restricted.
To gray haze, sandstorm frequency, remote sensing monitoring and electronic skyscreen are on the one hand used, is based on atmospheric circulation background
, power extrapolation is carried out, on the other hand by power-chemical model on-line/off-line coupling, is changed within forecast following 3-7 days, with
The continuous improvement of further investigation and computer performance to mode power, chemism, model predictions accuracy and resolution ratio
It is continuously improved, already becomes the main means of gray haze, sandstorm frequency.
The factors such as the uncertainty due to emission source, meteorological field and mode itself, the forecast result of air quality model
Still there is larger uncertainty, using Data Assimilation, couple the analog result and pollutant concentration prison of air quality numerical model
Survey data, can effectively improve the initial fields of numerical model, and realization observation (various routines, non-conventional observation) has with mode
Effect combines, and finally improves the value of forecasting and forecast accuracy of air quality model.
It is still immature based on laser radar assimilation technique both at home and abroad, also without the assimilation system for laser radar networking foundation
System；The majority pollution Aerosol Extinction (AOD) that is obtained using satellite remote sensing inverting of assimilation or ground monitoring obtain PM10 and
PM2.5 is directly or indirectly assimilated；Even if also only being calculated with the AOD data that it is acquired using best interpolation using laser radar
Method (OI) is assimilated indirectly；Or assimilated using the air quality model of minority, achievement does not have generality and popularization
Property.
The satellite remote sensing AOD of air quality model assimilates.Value of Remote Sensing Data is by observation point longitude and latitude, ground surface type, too
The external informations such as positive azimuth, solar zenith angle, satellite scan angle, wavelength influence, and are based on dark blue algorithm or dark pixel algorithm, instead
It drills to obtain AOD, there are larger system and non-systematic error, needs to carry out stringent quality control and deviation early period and correct, number
According to quality by multifactor impact, there are larger uncertainties；Later period is based on remotely-sensed data and constructs AOD Observation Operators, causes error
Transmitting；Assimilation obtains AOD increment, by linear relationship, distribution to PM2.5, so that change pattern forecasts initial fields；Entire technology
Process is only applicable to GOCART aerosol scheme, does not have generalization, sport technique segment is cumbersome, and there are error propagation phenomenons, and defend
Twice, temporal resolution is lower for observation in sing data one day.
The PM10 and PM2.5 of air quality model directly assimilate.It is mesh for the direct assimilation scheme of PM10 and PM2.5
It is preceding the most generally to assimilate scheme and means with conventional pollutant, but monitoring data all exist respectively on time and spatial resolution
From defect and deficiency, it is difficult to covering and represents most of mode grid, time and spatially with limitation.
Best interpolation assimilates algorithm (OI).OI algorithm belongs to single-node analysis method, and basic thought is to analyze field error
The minimum standard of covariance determines the statistics optimal weights of observation increment；The having some limitations property of algorithm: since OI algorithm limits
In the scale and substantial single-node analysis method of the system of linear equations for solving weight, Observation Operators cannot be too complicated, and exists
The regional choice problem of observational data；Observation Operators use three-dimensional space interpolation, including space interpolation both horizontally and vertically,
It is but simple linear interpolation method (bilinear interpolation of horizontal direction and the linear interpolation of vertical direction)；In addition, the party
Method does not account for the propagation and development of the error in forecasting process, is a kind of local parser of static state, so as to cause observation
Data can not effectively correct mode initial fields, influence to assimilate result.
Laser radar AOD assimilation.Collection, the observation of PM2.5 is in ground progress, and the particle of particle sampler acquisition
The process that object has a high temperature to dry, numerical value represents " dry " particulate matter quality concentration, but AOD is not only influenced by PM2.5 concentration,
Vertical distribution is changeable, uneven, cannot reflect aerosol particle principle condition near the ground；Its backscatter signal is dry by relative humidity
It disturbs, influences the accuracy of echo inverting assimilation.
In conclusion prior art defect and deficiency have:
1, structure is complicated, sport technique segment is cumbersome, causes error propagation；
2, time, the spatial resolution for observing data are low, are not able to satisfy routine work forecast demand and fine forecast lattice
Point requires；
3, related assimilation method belongs to state algorithm, domain of the existence select permeability；
4, AOD assimilation is influenced by vertical distribution and relative humidity, influences inverting assimilation result；
5, technology or Forecast Mode minority forecast networking capability, technology without promoting or establishing provinces and cities' observation+assimilation
Applicability is relatively narrow.
Summary of the invention
The main object of the present invention is to solve defect and deficiency in the prior art, and provide a kind of based on WRF-Chem mould
The Three-dimensional Variational Data Assimilation method of the aerosol LIDAR inverting PM2.5 of formula.
The purpose of the present invention can reach by using following technical solution:
A kind of Three-dimensional Variational Data Assimilation method of the aerosol LIDAR inverting PM2.5 based on WRF-Chem mode, including
Laser radar data preprocessing module and assimilation module, are supervised based on WRF-Chem air quality model and aerosol LIDAR
Networking is surveyed, using inverting PM2.5 concentration, corrects mode initial fields using Three-dimensional Variational Data Assimilation method, problem-solving pattern forecast is initial
Field and forecast accuracy.
The data preprocessing method of laser radar data preprocessing module, includes the following steps:
Step 11: the module input data includes laser radar aerosol related data after echo inverting, pollution
Monitoring data and meteorological measuring carry out data cleansing to various observation data, and carry out quality evaluation, according to time series
It is write as fixed format data, is prepared for next step standard data；
Step 12: carrying out vertical direction extinction coefficient integral, obtain flood AOD；
Step 13: by laser radar Vertical Profile data and meteorological measuring, carry out AOD height correct with it is wet
Degree is corrected, and the AOD data by arrangement are obtained；
Step 14: merging with PM2.5 and other meteorological, pollutant concentration data, bring machine algorithm into, successive ignition obtains
Time PM2.5 concentration when specified；
Step 15: according to assimilation time window, carrying out data screening and format storage；Finally PM2.5 after inverting is switched to
BUFR format provides several preparations for assimilation module.
In step 11, carrying out data cleansing to various observation data includes that repetition values are rejected, missing values substitute and exceptional value
It rejects.
The assimilation method for assimilating module, includes the following steps:
Step 21: the module input data includes gfs weather forecast data, elevation terrain data, land use data, dirt
Area side data boundary and emission inventory data are contaminated, duration and forecast initial time are more newly downloaded according to weather report and make；
Step 22: initial fields are not assimilated in initial start-up, generation；
Step 23: while NMC statistical method is called, each Season background error co-variance matrix is constructed, judges to simulate season,
Call corresponding Season background field error covariance matrix；
Step 24: followed by Three-dimensional Variational Data Assimilation, initial fields assimilation being carried out to PM2.5, including both horizontally and vertically
Assimilation obtains updating initial fields；
Step 25: updating lateral boundaries using initial fields after assimilation, carry out thermal starting；
Step 26: finally running WRF-Chem mode.
In step 21, elevation terrain data and land use data are static data, and depending on simulated domain, gfs is meteorological
Forecast data, Polluted area lateral boundaries data and emission inventory data are dynamic data.
Using inverting assimilation technique, using machine autonomous learning, collection, arrangement and cleaning data set, and data set is divided
For training group, test group and forecast group, judgment threshold and the number of iterations are set, AOD and PM2.5 quantitative model, exploitation right are established
Weight coefficient, set of computations forecast result.
Vertical Observation height using microwave laser radar data is 15km, vertical resolution 7.5m, temporal resolution
For 5min, assimilation radius is 200km, and assimilation time window is ± 3h.
Using Three-dimensional Variational Data Assimilation algorithm, which is based on Bayes' theorem, in priori ambient field error and observation error
It is to export the posterior probability density function of mode state variable x under the conditions of given observation yo under incoherent hypothesis:
Pa(x)=P (x | yo)=P (yo| x) P (x)=Po(H(x)-yo)Pb(x-xb)
Posterior probability density Pa(x) it is proportional to the probability density P of observationo(H(x)-yo) and priori probability density Pb(H(x)-
yo) product, it is unbiased and under meeting the hypothesis of normal state in background field error and observation error, pass through variation, solve posterior probability
The maximal solution of density just converts the minimal solution for solving cost functional J (x), and formula is as follows:
J (x)=(x-xb)TB-1(x-xb)+(H(x)-y)TR-1(H(x)-y)
Wherein: x is assimilation field, xbIt is ambient field, B is background error covariance, and H is Observation Operators, and y is observation
Vector, R are observation error covariances；
The control variable used in formula include stream function, non-equilibrium velocity potential, non-equilibrium temperature, non-equilibrium earth's surface air pressure,
The data such as false relative humidity；
It is calculated generally for simplifying, converts incremental form for above formula, assimilated by increment, forecast amendment initial fields and side
Boundary field:
Wherein, δ x=x-xb, δ y=x-Hxb
By solving the minimal solution of cost functional J (x), complicated non-linear Observation Operators are introduced, while according to simulation region
Domain constructs the background error covariance matrix of every month, obtains the Dynamical statistic property of each month ambient field, including sports ground
Geostrophic equilibrium between quality field, the constraint relationships such as standing balance between the air pressure and potential of quality field.
Using laser radar aerosol vertical distribution profile, extinction coefficient near the ground is calculated using absolute altitude method；According to
Local aerosol scattering absorbs growth factor, in conjunction with local relative humidity, carries out humidity to extinction coefficient near the ground and corrects, establish
Extinction coefficient near the ground and particulate matter quality concentration relationship near the ground.
Use WRF-Chem air quality model.
Advantageous effects of the invention:
The three-dimensional variation of aerosol LIDAR inverting PM2.5 based on WRF-Chem mode provided by the invention a kind of is same
Change method advantage is as follows:
1. the assimilation method is succinctly direct, not needing to construct complicated inversion algorithm using tripartite's module, impact factor is limited,
The error and uncertainty of introducing can be quantitatively evaluated well.
2. using laser radar data, the quality of data is reliable, and spatial and temporal resolution is higher, and it is pre- to better meet present mode
Report demand.
3. three-dimensional assimilation algorithm can construct dynamic background field error more effectively using observation data revision mode initial fields
Covariance matrix considers simulated domain local features and Seasonal variation, improves mode applicability.
It corrects, reasonably adjusts with humidity 4. being corrected by height, improve laser radar AOD and accuracy is forecast to PM2.5.
5. the air quality model used is with a wide range of applications, assimilation method is not limited by aerosol scheme, tool
There are very strong applicability and transplantability.
Detailed description of the invention
Fig. 1 is the laser radar assimilation system of this patent of the present invention design, includes two submodules: laser radar number in figure
Data preprocess module and assimilation module.
Specific embodiment
To make the more clear and clear technical solution of the present invention of those skilled in the art, the present invention is made below further
Detailed description, embodiments of the present invention are not limited thereto.
As shown in Figure 1, the three of the aerosol LIDAR inverting PM2.5 provided in this embodiment based on WRF-Chem mode
Variational Assimilation method is tieed up, including is based on WRF-Chem air quality model and aerosol Monitoring by Lidar networking, inverting PM2.5
Concentration corrects mode initial fields using Three-dimensional Variational Data Assimilation method, and problem-solving pattern forecasts initial fields and forecast accuracy, for existing
Having assimilation technique, structure is complicated, sport technique segment is cumbersome, using inverting assimilation technique, using machine autonomous learning, collection, arrangement and
Data set is cleaned, and data set is divided into training group, test group and forecast group, judgment threshold and the number of iterations is set, establishes AOD
With PM2.5 quantitative model, weight coefficient, set of computations forecast result, to improve training pattern accuracy, for observation number are utilized
According to time, spatial resolution it is low, using microwave laser radar data, which is 15km, vertically
Resolution ratio is 7.5m, and temporal resolution is 5min or so, and assimilation radius reaches 200km, and assimilation time window is ± 3h, sufficiently,
Reasonably using concern period laser radar data, forecast accuracy is improved.
In the present embodiment, the data preprocessing method of laser radar data preprocessing module, includes the following steps:
Step 11: the module input data includes laser radar aerosol related data after echo inverting, pollution
Monitoring data and meteorological measuring carry out data cleansing to various observation data, and carry out quality evaluation, according to time series
It is write as fixed format data, is prepared for next step standard data；
Step 12: carrying out vertical direction extinction coefficient integral, obtain flood AOD；
Step 13: by laser radar Vertical Profile data and meteorological measuring, carry out AOD height correct with it is wet
Degree is corrected, and the AOD data by arrangement are obtained；
Step 14: merging with PM2.5 and other meteorological, pollutant concentration data, bring machine algorithm into, successive ignition obtains
Time PM2.5 concentration when specified；
Step 15: according to assimilation time window, carrying out data screening and format storage；Finally PM2.5 after inverting is switched to
BUFR format provides several preparations for assimilation module.
In step 11, carrying out data cleansing to various observation data includes that repetition values are rejected, missing values substitute and exceptional value
It rejects.
In the present embodiment, the assimilation method for assimilating module, includes the following steps:
Step 21: the module input data includes gfs weather forecast data, elevation terrain data, land use data, dirt
Area side data boundary and emission inventory data are contaminated, duration and forecast initial time are more newly downloaded according to weather report and make；
Step 22: initial fields are not assimilated in initial start-up, generation；
Step 23: while NMC statistical method is called, each Season background error co-variance matrix is constructed, judges to simulate season,
Call corresponding Season background field error covariance matrix；
Step 24: followed by Three-dimensional Variational Data Assimilation, initial fields assimilation being carried out to PM2.5, including both horizontally and vertically
Assimilation obtains updating initial fields；
Step 25: updating lateral boundaries using initial fields after assimilation, carry out thermal starting；
Step 26: finally running WRF-Chem mode.
In step 21, elevation terrain data and land use data are static data, and depending on simulated domain, gfs is meteorological
Forecast data, Polluted area lateral boundaries data and emission inventory data are dynamic data.
Using inverting assimilation technique, using machine autonomous learning, collection, arrangement and cleaning data set, and data set is divided
For training group, test group and forecast group, judgment threshold and the number of iterations are set, AOD and PM2.5 quantitative model, exploitation right are established
Weight coefficient, set of computations forecast result.
Vertical Observation height using microwave laser radar data is 15km, vertical resolution 7.5m, temporal resolution
For 5min, assimilation radius is 200km, and assimilation time window is ± 3h.
In the present embodiment, for state algorithm and regional choice the problems such as, using Three-dimensional Variational Data Assimilation algorithm, the algorithm
Based on Bayes' theorem, in the case where priori ambient field error and observation error are incoherent hypothesis, given observation y is exportedoCondition
The posterior probability density function of lower mode state variable x:
Pa(x)=P (x | yo)=P (yo| x) P (x)=Po(H(x)-yo)Pb(x-xb)
That is posterior probability density Pa(x) it is proportional to the probability density P of observationo(H(x)-yo) and priori probability density Pb(H
(x)-yo) product, it is unbiased and under meeting the hypothesis of normal state in background field error and observation error, pass through variation, solve posteriority
The maximal solution of probability density just converts the minimal solution for solving cost functional J (x), and formula is as follows:
J (x)=(x-xb)TB-1(x-xb)+(H(x)-y)TR-1(H(x)-y)
Wherein: x is assimilation field, xbIt is ambient field, B is background error covariance, and H is Observation Operators, and y is observation
Vector, R are observation error covariances.The control variable used in formula include stream function, non-equilibrium velocity potential, non-equilibrium temperature,
The data such as non-equilibrium earth's surface air pressure, false relative humidity.It is calculated generally for simplifying, converts incremental form for above formula, pass through increasing
Amount assimilation, forecast amendment initial fields and lateral boundaries field:
Wherein, δ x=x-xb, δ y=x-Hxb
By solving the minimal solution of cost functional J (x), complicated non-linear Observation Operators can be introduced, and execute the overall situation
The regional choice problem solve, used because observational data may be not present；Simultaneously according to simulated domain, National Climate method is utilized
(NMC), the background error covariance matrix for constructing every month obtains the Dynamical statistic property (sports ground of each month ambient field
Geostrophic equilibrium between quality field, the constraint relationships such as standing balance between the air pressure and potential of quality field), solve static state not
The background error covariance matrix problem of change.
In the present embodiment, assimilating for laser radar AOD is influenced by vertical distribution and relative humidity, and laser radar is utilized
Extinction coefficient near the ground is calculated using absolute altitude method in aerosol vertical distribution profile；It is absorbed and is increased according to local aerosol scattering
The long factor carries out humidity to extinction coefficient near the ground and corrects in conjunction with local relative humidity, establish " dry " extinction coefficient near the ground and
" dry " particulate matter quality concentration relationship near the ground.
In the present embodiment, for technical method minority, WRF-Chem air quality model is used.
In the present embodiment, laser radar data preprocessing module: the module input data includes after echo inverting
Laser radar aerosol related data, pollution monitoring data and meteorological measuring, to various observation data carry out data it is clear
(including repetition values are rejected, missing values substitution, abnormality value removing) is washed, and carries out quality evaluation, is write as fixation according to time series
Formatted data is prepared for next step standard data；Vertical direction extinction coefficient integral is carried out, flood AOD is obtained；Pass through
Laser radar Vertical Profile data and meteorological measuring, the height for carrying out AOD is corrected to be corrected with humidity, is obtained by arranging
AOD data；Merge with PM2.5 and other meteorological, pollutant concentration data, brings machine algorithm into, successive ignition is specified
When time PM2.5 concentration；According to assimilation time window, data screening and format storage are carried out；PM2.5 after inverting is finally switched into BUFR
Format provides several preparations for assimilation module.
In the present embodiment, assimilate module: the module input data include gfs weather forecast data (demand according to weather report,
3-7 day data can be inputted), elevation terrain data and land use data, Polluted area lateral boundaries data and emission inventory
Data, wherein elevation terrain data and land use data are static data, depending on simulated domain, gfs weather forecast data,
Polluted area lateral boundaries data and emission inventory data are dynamic data, under duration and forecast initial time update according to weather report
It carries and makes；Initial fields are not assimilated in initial start-up, generation；NMC statistical method is called simultaneously, constructs each Season background error association side
Poor matrix judges to simulate season, calls corresponding Season background field error covariance matrix；It is right followed by Three-dimensional Variational Data Assimilation
PM2.5 carries out initial fields assimilation (including both horizontally and vertically assimilating), obtains updating initial fields；More using initial fields after assimilation
New lateral boundaries carry out thermal starting；Finally run WRF-Chem mode.
It in the present embodiment, is further verifying this method feasibility, we choose the dirty three times of Jiangsu Province's different months
Dye process, using environmental monitoring station, province laser radar and pollutant monitoring data, this patent assimilation side is not assimilated and uses in design
Two groups of testing programs of method, compare simulation.Tables 1 and 2 is respectively that PM2.5, PM10, AQI forecast of forecast and observation are accurate
Property and according to air quality concentration scale carry out risk score table.
The comparison of 1 model test outcome evaluation of table
2 forecast result of table is classified risk score table
The three-dimensional variation of aerosol LIDAR inverting PM2.5 based on WRF-Chem mode manufactured in the present embodiment a kind of
Assimilation method, after laser radar AOD inverting assimilation, forecast is significantly improved with observation correlation, accuracy, average deviation
It declines to a great extent, analog result is closer to observation, and each grade scoring result is also significantly increased after assimilation, forecasts after showing assimilation
The class of pollution is closer with fact.
The above, further embodiment only of the present invention, but scope of protection of the present invention is not limited thereto, and it is any
Within the scope of the present disclosure, according to the technique and scheme of the present invention and its design adds those familiar with the art
With equivalent substitution or change, protection scope of the present invention is belonged to.
Claims (10)
1. a kind of Three-dimensional Variational Data Assimilation method of the aerosol LIDAR inverting PM2.5 based on WRF-Chem mode, feature
It is, including laser radar data preprocessing module and assimilation module, is swashed based on WRF-Chem air quality model and aerosol
Optical radar monitors networking and corrects mode initial fields, problem-solving pattern using Three-dimensional Variational Data Assimilation method using inverting PM2.5 concentration
Forecast initial fields and forecast accuracy.
2. the three-dimensional of aerosol LIDAR inverting PM2.5 based on WRF-Chem mode according to claim 1 a kind of
Variational Assimilation method, which is characterized in that the data preprocessing method of laser radar data preprocessing module includes the following steps:
Step 11: the module input data includes laser radar aerosol related data after echo inverting, pollution monitoring
Data and meteorological measuring carry out data cleansing to various observation data, and carry out quality evaluation, are write as according to time series
Fixed format data are prepared for next step standard data；
Step 12: carrying out vertical direction extinction coefficient integral, obtain flood AOD；
Step 13: by laser radar Vertical Profile data and meteorological measuring, the height for carrying out AOD is corrected to be ordered with humidity
Just, the AOD data by arrangement are obtained；
Step 14: merging with PM2.5 and other meteorological, pollutant concentration data, bring machine algorithm into, successive ignition is specified
When time PM2.5 concentration；
Step 15: according to assimilation time window, carrying out data screening and format storage；PM2.5 after inverting is finally switched into BUFR lattice
Formula provides several preparations for assimilation module.
3. the three-dimensional of aerosol LIDAR inverting PM2.5 based on WRF-Chem mode according to claim 2 a kind of
Variational Assimilation method, which is characterized in that in step 11, data cleansing is carried out to various observation data and includes repetition values rejecting, lack
Mistake value substitution and abnormality value removing.
4. the three-dimensional of aerosol LIDAR inverting PM2.5 based on WRF-Chem mode according to claim 1 a kind of
Variational Assimilation method, which is characterized in that the assimilation method for assimilating module includes the following steps:
Step 21: the module input data includes gfs weather forecast data, elevation terrain data, land use data, contaminated area
Domain lateral boundaries data and emission inventory data, duration and forecast initial time are more newly downloaded according to weather report and make；
Step 22: initial fields are not assimilated in initial start-up, generation；
Step 23: while NMC statistical method is called, each Season background error co-variance matrix is constructed, judges to simulate season, call
Corresponding Season background field error covariance matrix；
Step 24: followed by Three-dimensional Variational Data Assimilation, initial fields assimilation is carried out to PM2.5, including both horizontally and vertically assimilate,
It obtains updating initial fields；
Step 25: updating lateral boundaries using initial fields after assimilation, carry out thermal starting；
Step 26: finally running WRF-Chem mode.
5. the three-dimensional of aerosol LIDAR inverting PM2.5 based on WRF-Chem mode according to claim 4 a kind of
Variational Assimilation method, which is characterized in that in step 21, elevation terrain data and land use data are static data, with simulation
Depending on region, gfs weather forecast data, Polluted area lateral boundaries data and emission inventory data are dynamic data.
6. the three-dimensional of aerosol LIDAR inverting PM2.5 based on WRF-Chem mode according to claim 1 a kind of
Variational Assimilation method, which is characterized in that use inverting assimilation technique, use machine autonomous learning, collection, arrangement and cleaning data
Collection, and data set is divided into training group, test group and forecast group, judgment threshold and the number of iterations are set, AOD and PM2.5 are established
Quantitative model utilizes weight coefficient, set of computations forecast result.
7. the three-dimensional of aerosol LIDAR inverting PM2.5 based on WRF-Chem mode according to claim 1 a kind of
Variational Assimilation method, which is characterized in that the use of the Vertical Observation height of microwave laser radar data is 15km, vertical resolution is
7.5m, temporal resolution 5min, assimilation radius are 200km, and assimilation time window is ± 3h.
8. the three-dimensional of aerosol LIDAR inverting PM2.5 based on WRF-Chem mode according to claim 1 a kind of
Variational Assimilation method, which is characterized in that use Three-dimensional Variational Data Assimilation algorithm, which is based on Bayes' theorem, in priori background
Field error and observation error are to export given observation y under incoherent hypothesisoUnder the conditions of mode state variable x posterior probability
Density function:
Pa(x)=P (x | yo)=P (yo| x) P (x)=Po(H(x)-yo)Pb(x-xb)
Posterior probability density Pa(x) it is proportional to the probability density P of observationo(H(x)-yo) and priori probability density Pb(H(x)-yo)
Product, it is unbiased and under meeting the hypothesis of normal state in background field error and observation error, by variation, solve posterior probability density
Maximal solution just converts the minimal solution for solving cost functional J (x), and formula is as follows:
J (x)=(x-xb)TB-1(x-xb)+(H(x)-y)TR-1(H(x)-y)
Wherein: x is assimilation field, xbIt is ambient field, B is background error covariance, and H is Observation Operators, and y is observation vector, R
It is observation error covariance；
The control variable used in formula includes stream function, non-equilibrium velocity potential, non-equilibrium temperature, non-equilibrium earth's surface air pressure, false appearance
To data such as humidity；
It is calculated generally for simplifying, converts incremental form for above formula, assimilated by increment, forecast amendment initial fields and lateral boundaries
:
Wherein, δ x=x-xb, δ y=x-Hxb
By solving the minimal solution of cost functional J (x), complicated non-linear Observation Operators are introduced, while according to simulated domain, structure
The background error covariance matrix for building every month obtains the Dynamical statistic property of each month ambient field, including sports ground and matter
Measure the geostrophic equilibrium between field, the constraint relationships such as standing balance between the air pressure and potential of quality field.
9. the three-dimensional of aerosol LIDAR inverting PM2.5 based on WRF-Chem mode according to claim 1 a kind of
Variational Assimilation method, which is characterized in that utilize laser radar aerosol vertical distribution profile, near-earth is calculated using absolute altitude method
Face extinction coefficient；Absorb growth factor according to local aerosol scattering, in conjunction with local relative humidity, to extinction coefficient near the ground into
Row humidity is corrected, and extinction coefficient near the ground and particulate matter quality concentration relationship near the ground are established.
10. the three-dimensional of aerosol LIDAR inverting PM2.5 based on WRF-Chem mode according to claim 1 a kind of
Variational Assimilation method, which is characterized in that use WRF-Chem air quality model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811621885.1A CN109709577A (en) | 2018-12-28 | 2018-12-28 | A kind of Three-dimensional Variational Data Assimilation method of the aerosol LIDAR inverting PM2.5 based on WRF-Chem mode |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811621885.1A CN109709577A (en) | 2018-12-28 | 2018-12-28 | A kind of Three-dimensional Variational Data Assimilation method of the aerosol LIDAR inverting PM2.5 based on WRF-Chem mode |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109709577A true CN109709577A (en) | 2019-05-03 |
Family
ID=66259082
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811621885.1A Pending CN109709577A (en) | 2018-12-28 | 2018-12-28 | A kind of Three-dimensional Variational Data Assimilation method of the aerosol LIDAR inverting PM2.5 based on WRF-Chem mode |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109709577A (en) |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110161532A (en) * | 2019-05-30 | 2019-08-23 | 浙江大学 | A method of based on multi-wavelength laser radar inverting microfluidic aerosol physical characteristic |
| CN110910963A (en) * | 2019-10-29 | 2020-03-24 | 暨南大学 | Three-dimensional variation assimilation method and system for optical thickness of aerosol and storage medium |
| CN110987741A (en) * | 2019-11-06 | 2020-04-10 | 中国辐射防护研究院 | Method for estimating content of secondary aerosol in atmosphere |
| CN111048161A (en) * | 2019-10-22 | 2020-04-21 | 中国人民解放军国防科技大学 | Aerosol extinction coefficient three-dimensional variation assimilation method based on IMPROVE equation |
| CN111221004A (en) * | 2020-01-18 | 2020-06-02 | 北京环拓科技有限公司 | Method for detecting VOC distribution by utilizing laser radar 3D scanning |
| CN111382506A (en) * | 2020-03-02 | 2020-07-07 | 苏州工业园区洛加大先进技术研究院 | Method for evaluating influence of aerosol and radiation interaction on atomization effect |
| CN111581897A (en) * | 2020-06-02 | 2020-08-25 | 孙溦 | Data processing method, device and equipment |
| CN111651875A (en) * | 2020-05-25 | 2020-09-11 | 北京大学 | Air quality forecast improvement method and system based on error transmission |
| CN111738600A (en) * | 2020-06-23 | 2020-10-02 | 南通大学 | Urban road air quality evaluation method based on high-precision PM2.5 inversion result |
| CN111881590A (en) * | 2020-07-30 | 2020-11-03 | 中国科学院空天信息创新研究院 | Spatial analysis method for concentration of atmospheric particulate matter |
| CN111965666A (en) * | 2020-07-16 | 2020-11-20 | 中国矿业大学 | Aerosol three-dimensional distribution mapping method |
| CN112033870A (en) * | 2020-07-20 | 2020-12-04 | 宁波市气象台 | Regional boundary layer PM2.5 stereo distribution and total amount calculation method |
| CN112307602A (en) * | 2020-10-13 | 2021-02-02 | 河海大学 | Method for joint inversion of underground water pollution source information and hydraulic permeability coefficient field |
| CN112800603A (en) * | 2021-01-26 | 2021-05-14 | 北京航空航天大学 | Atmospheric environment data assimilation method based on set optimal interpolation algorithm |
| CN113111565A (en) * | 2021-02-23 | 2021-07-13 | 中国人民解放军国防科技大学 | Aerosol data assimilation method based on MIE scattering equation observation operator |
| CN113127530A (en) * | 2021-03-05 | 2021-07-16 | 中国气象科学研究院 | Dynamic atmospheric three-dimensional structure detection information group assimilation system based on railway network |
| CN113156395A (en) * | 2021-04-07 | 2021-07-23 | 中国科学院大气物理研究所 | Aerosol laser radar data fusion method and system |
| CN113435068A (en) * | 2021-08-27 | 2021-09-24 | 中国人民解放军国防科技大学 | Radionuclide assimilation prediction method based on logarithmic variational assimilation |
| CN113740220A (en) * | 2021-09-07 | 2021-12-03 | 中国人民解放军国防科技大学 | Multi-scale three-dimensional variational assimilation method based on high-resolution aerosol data |
| CN113834902A (en) * | 2021-08-16 | 2021-12-24 | 中国人民解放军国防科技大学 | Sulfur dioxide emission source inversion method based on four-dimensional variational assimilation |
| CN114112995A (en) * | 2021-12-01 | 2022-03-01 | 中国人民解放军国防科技大学 | Aerosol optical characteristic data assimilation method and device based on three-dimensional variational technology |
| CN115293190A (en) * | 2022-05-27 | 2022-11-04 | 河海大学 | Sand storm dynamic inversion method based on microwave signal blind source separation and SCA |
| CN116381719A (en) * | 2023-01-13 | 2023-07-04 | 中国科学技术大学 | Laser radar-based polluted area enhanced identification method |
| CN112132336B (en) * | 2020-09-22 | 2024-02-20 | 南京创蓝科技有限公司 | Quarterly prediction method for PM2.5 concentration |
| CN112800603B (en) * | 2021-01-26 | 2024-05-24 | 北京航空航天大学 | Atmospheric environment data assimilation method based on set optimal interpolation algorithm |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107273995A (en) * | 2016-04-08 | 2017-10-20 | 株式会社日立制作所 | Urban Air Pollution Methods |
-
2018
- 2018-12-28 CN CN201811621885.1A patent/CN109709577A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107273995A (en) * | 2016-04-08 | 2017-10-20 | 株式会社日立制作所 | Urban Air Pollution Methods |
Non-Patent Citations (5)
| Title |
|---|
| 扬.克莱斯尔: "《太阳能资源预报与评估》", 31 January 2018 * |
| 王桥 等: "《PM2.5卫星遥感技术及其应用》", 31 October 2017 * |
| 范水勇 等: "MM5三维变分系统在北京地区冷暖季背景场误差的对比分析", 《高原气象》 * |
| 蒋维东: "AOD卫星观测资料三维变分同化试验研究", 《中国知网》 * |
| 项衍: "激光雷达区域组网数据的三维同化与综合分析方法研究", 《中国知网》 * |
Cited By (34)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110161532A (en) * | 2019-05-30 | 2019-08-23 | 浙江大学 | A method of based on multi-wavelength laser radar inverting microfluidic aerosol physical characteristic |
| CN111048161A (en) * | 2019-10-22 | 2020-04-21 | 中国人民解放军国防科技大学 | Aerosol extinction coefficient three-dimensional variation assimilation method based on IMPROVE equation |
| CN111048161B (en) * | 2019-10-22 | 2022-09-13 | 中国人民解放军国防科技大学 | Aerosol extinction coefficient three-dimensional variation assimilation method based on IMPROVE equation |
| CN110910963A (en) * | 2019-10-29 | 2020-03-24 | 暨南大学 | Three-dimensional variation assimilation method and system for optical thickness of aerosol and storage medium |
| CN110910963B (en) * | 2019-10-29 | 2022-03-29 | 暨南大学 | Three-dimensional variation assimilation method and system for optical thickness of aerosol and storage medium |
| CN110987741B (en) * | 2019-11-06 | 2022-05-17 | 中国辐射防护研究院 | Method for estimating content of secondary aerosol in atmosphere |
| CN110987741A (en) * | 2019-11-06 | 2020-04-10 | 中国辐射防护研究院 | Method for estimating content of secondary aerosol in atmosphere |
| CN111221004A (en) * | 2020-01-18 | 2020-06-02 | 北京环拓科技有限公司 | Method for detecting VOC distribution by utilizing laser radar 3D scanning |
| CN111221004B (en) * | 2020-01-18 | 2022-04-19 | 北京环拓科技有限公司 | Method for detecting VOC distribution by utilizing laser radar 3D scanning |
| CN111382506A (en) * | 2020-03-02 | 2020-07-07 | 苏州工业园区洛加大先进技术研究院 | Method for evaluating influence of aerosol and radiation interaction on atomization effect |
| CN111651875A (en) * | 2020-05-25 | 2020-09-11 | 北京大学 | Air quality forecast improvement method and system based on error transmission |
| CN111581897A (en) * | 2020-06-02 | 2020-08-25 | 孙溦 | Data processing method, device and equipment |
| CN111581897B (en) * | 2020-06-02 | 2023-11-03 | 孙溦 | Data assimilation, device and equipment for ground observation of atmospheric pollution |
| CN111738600A (en) * | 2020-06-23 | 2020-10-02 | 南通大学 | Urban road air quality evaluation method based on high-precision PM2.5 inversion result |
| CN111965666A (en) * | 2020-07-16 | 2020-11-20 | 中国矿业大学 | Aerosol three-dimensional distribution mapping method |
| CN112033870B (en) * | 2020-07-20 | 2021-03-09 | 宁波市气象台 | Regional boundary layer PM2.5 stereo distribution and total amount calculation method |
| CN112033870A (en) * | 2020-07-20 | 2020-12-04 | 宁波市气象台 | Regional boundary layer PM2.5 stereo distribution and total amount calculation method |
| CN111881590A (en) * | 2020-07-30 | 2020-11-03 | 中国科学院空天信息创新研究院 | Spatial analysis method for concentration of atmospheric particulate matter |
| CN112132336B (en) * | 2020-09-22 | 2024-02-20 | 南京创蓝科技有限公司 | Quarterly prediction method for PM2.5 concentration |
| CN112307602A (en) * | 2020-10-13 | 2021-02-02 | 河海大学 | Method for joint inversion of underground water pollution source information and hydraulic permeability coefficient field |
| CN112800603A (en) * | 2021-01-26 | 2021-05-14 | 北京航空航天大学 | Atmospheric environment data assimilation method based on set optimal interpolation algorithm |
| CN112800603B (en) * | 2021-01-26 | 2024-05-24 | 北京航空航天大学 | Atmospheric environment data assimilation method based on set optimal interpolation algorithm |
| CN113111565A (en) * | 2021-02-23 | 2021-07-13 | 中国人民解放军国防科技大学 | Aerosol data assimilation method based on MIE scattering equation observation operator |
| CN113127530A (en) * | 2021-03-05 | 2021-07-16 | 中国气象科学研究院 | Dynamic atmospheric three-dimensional structure detection information group assimilation system based on railway network |
| CN113156395A (en) * | 2021-04-07 | 2021-07-23 | 中国科学院大气物理研究所 | Aerosol laser radar data fusion method and system |
| CN113834902A (en) * | 2021-08-16 | 2021-12-24 | 中国人民解放军国防科技大学 | Sulfur dioxide emission source inversion method based on four-dimensional variational assimilation |
| CN113435068A (en) * | 2021-08-27 | 2021-09-24 | 中国人民解放军国防科技大学 | Radionuclide assimilation prediction method based on logarithmic variational assimilation |
| CN113740220A (en) * | 2021-09-07 | 2021-12-03 | 中国人民解放军国防科技大学 | Multi-scale three-dimensional variational assimilation method based on high-resolution aerosol data |
| CN114112995B (en) * | 2021-12-01 | 2024-01-30 | 中国人民解放军国防科技大学 | Aerosol optical characteristic data assimilation method and device based on three-dimensional variation technology |
| CN114112995A (en) * | 2021-12-01 | 2022-03-01 | 中国人民解放军国防科技大学 | Aerosol optical characteristic data assimilation method and device based on three-dimensional variational technology |
| CN114662127B (en) * | 2022-03-09 | 2024-05-31 | 中金金融认证中心有限公司 | Method, apparatus and readable storage medium for data assimilation |
| CN115293190A (en) * | 2022-05-27 | 2022-11-04 | 河海大学 | Sand storm dynamic inversion method based on microwave signal blind source separation and SCA |
| CN116381719B (en) * | 2023-01-13 | 2023-08-29 | 中国科学技术大学 | Laser radar-based polluted area enhanced identification method |
| CN116381719A (en) * | 2023-01-13 | 2023-07-04 | 中国科学技术大学 | Laser radar-based polluted area enhanced identification method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109709577A (en) | A kind of Three-dimensional Variational Data Assimilation method of the aerosol LIDAR inverting PM2.5 based on WRF-Chem mode | |
| CN109697323B (en) | Rainfall observation method integrating satellite remote sensing and mobile communication base station signals | |
| Li et al. | Modeling land surface processes and heavy rainfall in urban environments: Sensitivity to urban surface representations | |
| Beyrich et al. | Evaporation over a heterogeneous land surface: EVA_GRIPS and the LITFASS-2003 experiment—an overview | |
| CN105243435B (en) | A kind of soil moisture content prediction technique based on deep learning cellular Automation Model | |
| CN108536908B (en) | Method for evaluating watershed water environment safety based on non-point source nitrogen and phosphorus loss risk | |
| CN107316095A (en) | A kind of region meteorological drought grade prediction technique for coupling multi-source data | |
| CN107064937A (en) | A kind of measuring method of Dual-linear polarization radar system and strong rain | |
| Zhang | Application of a convection‐permitting ensemble prediction system to quantitative precipitation forecasts over southern China: Preliminary results during SCMREX | |
| CN114265836A (en) | All-weather assimilation method and device of satellite microwave hygrothermograph based on cloud area temperature and humidity profile inversion | |
| CN105069295B (en) | Satellite and surface precipitation measured value assimilation method based on Kalman filtering | |
| CN110674467A (en) | Method for monitoring response of hydrological process to climate change based on SWAT model | |
| CN112785024A (en) | Runoff calculation and prediction method based on watershed hydrological model | |
| CN107480781A (en) | The nuclear accident Source Term Inversion method of neutral net adaptive Kalman filter | |
| CN114254802B (en) | Prediction method for vegetation coverage space-time change under climate change drive | |
| CN108871586B (en) | Inversion method of ground-based infrared remote sensing ground surface temperature | |
| Liu et al. | Wind dynamics over a highly heterogeneous oasis area: An experimental and numerical study | |
| Zou et al. | Impact of eddy characteristics on turbulent heat and momentum fluxes in the urban roughness sublayer | |
| Petersen et al. | Modelling the atmospheric transport of trace metals from Europe to the North Sea and the Baltic Sea | |
| CN113516418A (en) | Apple planting area drought disaster risk assessment method | |
| CN115544706A (en) | Wavelet and XGboost model integrated atmospheric fine particle concentration estimation method | |
| CN109948175B (en) | Satellite remote sensing albedo missing value inversion method based on meteorological data | |
| CN114878748B (en) | CO (carbon monoxide) 2 Emission monitoring method and emission monitoring system | |
| CN115840908B (en) | Method for constructing PM2.5 three-dimensional dynamic monitoring field based on microwave link of LSTM model | |
| Wang et al. | Evaluation of two observation operator schemes for wind profiler radar data assimilation and its impacts on short-term forecasting |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190503 |
|
| RJ01 | Rejection of invention patent application after publication |