CN103744069A - Methane profile orthogonal inversion method aiming at AIRS (atmospheric infrared sounder) hyper-spectrum satellite data - Google Patents

Abstract

The invention discloses a methane profile orthogonal inversion method aiming at AIRS (atmospheric infrared sounder) hyper-spectrum satellite data. The method concretely comprises the following steps that a secondary standard product of an AIRS hyper-spectrum satellite is utilized for obtaining atmospheric composition profile data under the conditions of different regions and different seasons; a group of sets which contain a great number of observation results and can represent the atmospheric composition profile data under the conditions of different regions and different seasons are built; a radiation transmission equation is utilized for simulating an atmospheric top layer emergent radiation brightness value on a target passage for each profile, in addition, the observation noise of an instrument is added for obtaining the simulated satellite sensor observation radiation brightness value; the sensitivity of the atmospheric composition and the temperature is analyzed according to the simulated radiation brightness value for selecting the inversion passage; training sample sets are subjected to empirical orthogonal function expansion, and an empirical orthogonal function regression coefficient is calculated; cloud-clearing radiation brightness standard products of the AIRS hyper-spectrum satellite data are obtained, and the obtained regression coefficient is used for inversion.

Description

A kind of methane profile quadrature inversion method for AIRS ultraphotic spectrum satellite data

Technical field

The present invention relates to Atmospheric components inversion technique field, satellite remote sensing field, relate to a kind ofly for AIRS ultraphotic spectrum satellite data methane profile inversion method, relate in particular to a kind of AIRS ultraphotic spectrum satellite data methane profile fast inversion method based on Empirical Orthogonal Function homing method.

Background technology

Methane is a kind of important greenhouse gases, and the greenhouse effect of its unit mass are than CO ₂large 20 times.Up to the present, during greenhouse effect strengthen, methane role accounts for 20%, is only second to CO ₂.The Radiative Forcing of the methane of mankind's discharge is 0.48W/m ², the chances are CO ₂(1.66W/m ²) 1/3rd.Methane has also played vital role in the chemical process of atmosphere convection layer peace flow process, and for example the OH free radical in methane and troposphere forms formaldehyde (CH ₂o), carbon monoxide (CO) and ozone (O ₃), thereby affect oxidability and the tropospheric ozone concentration of atmosphere.In atmosphere, methane concentration has increased by 2.5 times since the industrial revolution, and the methane especially in recent years discharging from permafrost increases gradually, in the world to atmosphere CH ₄the variation of content is especially paid close attention to, and adopts one after another various ways to measure CH ₄content.But due to CH ₄discharge there is very large space and time difference, observation is very limited again, therefore global CH ₄still there is very large uncertainty in the quantification of discharge.Utilize remote sensing monitoring Atmospheric components have fast, economical, can repeatedly obtain atmospheric trace gas Information Superiority in macro-scale.

Current methane profile inversion technique mainly contains empirical algorithms and the large class of physics inversion algorithm two.Empirical algorithms adopts neural network algorithm to carry out inverting more, and because cloud is very complicated on the impact of thermal infrared emergent radiation, so neural network can only be for the inverting under clear sky condition.Physics inversion algorithm is to launch on the basis of optimum estimation, directly relatively do not calculate the difference between radiation and observation radiation, but between ambient field and observation field, obtain a kind of compromise, by priori conditions, separating constraint within the specific limits, then with Newton iteration, progressively approach the maximal possibility estimation of true solution.The general solution formula of atmosphere inverse problem optimal estimation method is:

X=X _a+(K ^TS _e ^-1K+γS _a ^-1) ^-1K ^TS _e ^-1ΔY，

In formula, X _afor initial profile, the first order derivative that K changes atmospheric parameter for observation spoke brightness, Jacobin matrix (Jacobian) namely, S _afor the covariance matrix of initial profile error, S _efor the covariance matrix of observation, model error, Δ Y is the brightness of observation spoke and the difference of calculating spoke brightness.In physics inversion algorithm, there are two very unmanageable problems: first, initial profile X _atruth can not be excessively departed from, otherwise the result of convergence can not be obtained; Next prior uncertainty covariance matrix S _aneed in statistical significance, can represent the time of day for the treatment of inverted parameters, otherwise be difficult to retrain final solution in rational scope.And for physics inversion algorithm, each iteration all needs methane to solve Jacobin matrix K, calculated amount is very large.

In order to reduce prior uncertainty covariance matrix S _athe inaccurate methane inversion error of introducing of estimation has developed singular value decomposition algorithm, by γ S on the basis of optimization algorithm _a ^-1item is ignored, and by K ^ts _e ^-1k item carries out svd SVD(Singular value decomposition), for the eigenwert that is less than certain threshold value in diagonal matrix, carry out assignment again, make K ^ts _e ^-1the inverse matrix of K is more stable.Although this algorithm makes the result of inverting reduce to greatest extent the dependence for initial background field and initial background error, but at AIRS, atmospheric methane is changed insensitive troposphere bottom and more than stratosphere carried out strict constraint, inversion result depends on initial profile.

Summary of the invention

The problem existing for prior art, the object of the present invention is to provide a kind of can be stably, quickly and easily for the methane profile quadrature inversion method of AIRS ultraphotic spectrum satellite data.

For achieving the above object, a kind of methane profile quadrature inversion method for AIRS ultraphotic spectrum satellite data of the present invention, is specially:

1) utilize the secondary standard product of AIRS ultraphotic spectrum satellite to obtain the Atmospheric components profile data under zones of different, Various Seasonal condition;

2) set up one group comprise analytic results can represent the Atmospheric components profile data set under zones of different, Various Seasonal condition;

3) utilize radiation transfer equation all to simulate atmospheric envelope on destination channel to each profile and eject and penetrate radiance value, and the observation noise that adds instrument is to obtain the satellite sensor observation radiance value of simulation;

4) according to simulation radiance value, the sensitivity analysis of temperature and Atmospheric components is selected to inverting passage;

5) training sample set is carried out to empirical orthogonal function, calculate Empirical Orthogonal Function regression coefficient;

6) clear sky that obtains AIRS ultraphotic spectrum satellite data is corrected spoke luminance standard product, utilizes the regression coefficient of trying to achieve to carry out inverting.

Further, described step 2) be specially:

A) collect one group of atmosphere profile that represents zones of different, Various Seasonal, each atmosphere profile all comprises atmospheric condition parameter and earth's surface state parameter; Each atmosphere profile has comprised atmospheric condition parameter and earth's surface state parameter, for example atmospheric condition parameter: temperature, moisture content, CO ₂, O ₃, CO, N ₂o, CH ₄profile; , earth's surface state parameter for example: surface temperature, earth's surface moisture content, earth's surface pressure, earth's surface sea level elevation, earth's surface emissivity;

B) all Atmospheric components profile samples are all treated to 101 layers according to the standard vertical barosphere of AIRS, and uniform units:

$P\left(x\right)={({\mathrm{ax}}^2+\mathrm{bx}+c)}^{\frac{7}{2}}---\left(1\right)$

Wherein, P represents air pressure (hundred handkerchiefs); X represents barosphere ordinal number, from 1 to 101; Tri-parameters of a, b and c are by substitution a=-1.55 * 10 ^-4, b=-5.59 * 10 ^-2, c=7.45.

Further, described step 3) is specially:

C) utilize RTTOV11.1 to calculate atmospheric envelope on all passages and eject and penetrate radiance value, be designated as R ₀;

D) for the simulation outgoing radiance value on each passage, add and take the random noise that Instrument observation noise is the upper limit, as the satellite sensor observation radiance value R of simulation.

Further, described step 4) is specially:

E) for each profile of atmosphere profile data centralization, by CH ₄concentration increases by 10%, re-uses RTTOV11.1 and calculates atmosphere on all passages and eject and penetrate radiance value, is designated as R _cH4, calculate corresponding atmosphere and eject the mean value of penetrating radiance value difference value

in like manner, by the whole 1K that increases of temperature profile, re-use RTTOV11.1 and calculate atmosphere on all passages and eject and penetrate radiance value, be designated as R _t, calculate corresponding atmosphere and eject the mean value of penetrating radiance value difference value

vapour concentration is increased to 10%, re-use RTTOV11.1 and calculate atmosphere on all passages and eject and penetrate radiance value, be designated as R _h2O, calculate corresponding atmosphere and eject the mean value of penetrating radiance value difference value

by N ₂o concentration increases by 10%, re-uses RTTOV11.1 and calculates atmosphere on all passages and eject and penetrate radiance value, is designated as R _n2O, calculate corresponding atmosphere and eject the mean value of penetrating radiance value difference value

earth's surface emissivity is reduced to 0.02, re-use RTTOV11.1 and calculate atmosphere on all passages and eject and penetrate radiance value, be designated as R _ε, calculate corresponding atmosphere and eject the mean value of penetrating radiance value difference value

F) select

be greater than 0.2K and be less than 0.8K and

be less than 0.4K and

be less than 0.2K and

be less than the passage of 0.1K.

Further, described step 5) is specially:

G) calculate respectively the covariance matrix of methane profile sample set and simulation outgoing radiance,

C _cov＝ΔC(ΔC) ^T/n (2)，

R _cov＝ΔR(ΔR) ^T/n (3)，

C wherein _cov[l * l] is the covariance matrix of carbon monoxide profile sample set, Δ C[l * n] be the deviation matrix of profile and profile sample set average; Similarly, R _cov[m * m] is the covariance matrix of simulation AIRS observation spoke brightness, Δ R[m * n] be the deviation matrix of analogue observation radiation value and sample set average; Subscript T representing matrix transposition;

H) by C _covand R _covcarry out respectively empirical orthogonal expansion:

C _cov＝ΓΛΓ ^T (4)，

R _cov＝Γ _rΛ _rΓ _r ^T (5)，

Wherein Г [l * l] is the Empirical Orthogonal Function collection of profile sample set, and the proportion that each EOF accounts in population variance according to it is arranged by row, value is maximum come before, the l that Λ is comprised of the eigenwert of covariance matrix ties up diagonal matrix; Equally, Г _r[m * m] is the Empirical Orthogonal Function collection of analogue observation spoke luma samples collection, Λ _rthe m dimension diagonal matrix forming for the eigenwert of covariance matrix;

I) for atmosphere profile sample set and simulation outgoing radiance, select respectively front 50 and 40 proper vectors, carry out empirical orthogonal function:

ΔC＝Γ _1-50×A (6)，

ΔR=Γ _r1-40×B (7)；

J) calculate respectively EOF score A and the B of Δ C and Δ R:

A=Γ _1-50 ^T×ΔC (8)，

$B={{{\mathrm{\Γ}}_r}_{1-50}}^T\×\mathrm{\ΔR}---\left(9\right);$

K) calculate the transition matrix D between two EOF scores, then obtain regression coefficient matrix S,

D=A×B ^T×(B×B ^T) ^-1 (10)，

S=Γ×D×Γ _r ^T (11)；

L) incident angle and view angle are made as respectively to 0 °, 10 °, 20 ° ... 70 °, take 10 ° as interval, for each profile of atmosphere profile data centralization, utilize RTTOV11.1 to calculate corresponding atmosphere and eject and penetrate radiance, then according to step G)-step K) method calculate regression coefficient S corresponding to different angles.

Further, described step 6) is specially:

M) by picture dot, read in AIRS clear sky and correct spoke brightness product, read quality tab radiances_QC and the radiance value radiances of each each wave band of picture dot, then carry out clear sky and correct quality judgement; According to step F) passage selected, check that the clear sky of each passage is corrected quality; If inverting passage has clear sky over half to correct quality for poor, radiances_QC=2, does not carry out inverting; If carry out inverting, but the clear sky of part inverting passage is corrected quality for poor, this passage clear sky is corrected to spoke brightness and is made as sample average

;

N) by picture dot read in solar zenith angle and observation zenith angle, at step L) in find corresponding matrix of coefficients;

O) by picture dot, utilize step M) clear sky that reads in corrects spoke brightness and step N) matrix of coefficients that reads in, according to following relation, obtain corresponding Atmospheric components vertical concentration profile:

$C_{\mathrm{obs}}=\stackrel{\‾}{C}+S\×\mathrm{\Δ}{R}_{\mathrm{obs}}---\left(12\right),$

Wherein, C _obs[l] is the gas concentration Vertical Profile l layer of inverting,

for gas profile sample set average, Δ R _obs[m * n] is observation radiance and simulation spoke luma samples collection mean bias matrix;

P) by picture dot, read in longitude and latitude information, by step O) methane profile inversion result and corresponding latitude and longitude information by picture dot, be written in HDF, Output rusults data layout is HDF.

Method of the present invention has been chosen the atmosphere profile that represents Various Seasonal, region from the atmosphere profile storehouse in European Meso-scale meteorology forecasting centre, has guaranteed the representativeness of training sample; Utilize front several proper vectors of Empirical Orthogonal Function to return, can reduce the error that observation noise is introduced, also can reduce other interference gas (H for example ₂o, N ₂o) interference, can be fast, stable, realize the inverting of methane profile exactly.

Accompanying drawing explanation

Fig. 1 is schematic flow sheet one of the present invention;

Fig. 2 is schematic flow sheet two of the present invention.

Embodiment

Below, the present invention is more fully illustrated, and show exemplary embodiment of the present invention.Yet the present invention can be presented as multiple multi-form, and should not be construed as the exemplary embodiment that is confined to narrate here.But, these embodiment are provided, thereby make the present invention comprehensively with complete, and scope of the present invention is fully conveyed to those of ordinary skill in the art.

Satellite data is in the present invention selected U.S. earth observing system EOS(Earth Observation System) the atmosphere infrared sensor AIRS(Atmospheric Infrared Sounder that carries on-AQUA satellite) clear sky correct spoke brightness secondary standard product (L2-standard-cloud-cleared-radiance-product).This product has the picture dot of cloud by the advanced microwave detector AMSU(Advanced Microwave Sounding Unit of same load to part) carried out spoke brightness and corrected, significantly improve the utilization factor of data.AIRS is sweep type acquisition sensor, and it adopts infrared light grating array light splitting technology, and 2378 passages cover 650-2700cm ^-1infrared spectrum region, passage divides three sections: 3.74-4.61 μ m, 6.20-8.22 μ m and 8.80-15.4 μ m, spectral resolution λ/Δ λ=1200, actinometry absolute precision is better than 0.2K.For the reference target of 250K, instrument effective noise temperature is from 0.14K(4.2 μ m) to 0.35K(15 μ m).The field of View angle of instrument is 1.1 °, and the spatial resolution of corresponding substar is 13.5km, and scanning angle is ± 48.95 °, and sweep length is 1650km.

In the present invention, atmosphere profile storehouse is selected from the observation sample of ECMWF based on radiosonde data, geographic position covers the most of region, the whole world from north latitude 75 degree to south latitude 75 degree, time distributes upper containing having covered Various Seasonal, and atmospheric condition parameter comprises temperature, steam, O ₃, CO ₂, CO, N ₂o, CH ₄profile, earth's surface state parameter has comprised surface temperature, earth's surface moisture content, earth's surface pressure, earth's surface sea level elevation, earth's surface emissivity.The radiative transfer model adopting in the present invention is RTTOV11.1, and this model is to belong to parameterized accelerated model, has merged AIRS instrument spectral response functions, quickly and accurately the radiance value at analog sensor entrance pupil place.

A kind of methane profile quadrature inversion method for AIRS ultraphotic spectrum satellite data of the present invention, first need to utilize the Atmospheric components profile under AIRS ultraphotic spectrum moonscope zones of different, Various Seasonal condition, obtain corresponding Atmospheric components profile data, utilize these data to carry out following methods processing procedure, Fig. 1, Fig. 2 are the process flow diagram of following methods processing procedure, comprise the steps:

S1. utilize the secondary standard product of AIRS ultraphotic spectrum satellite to obtain the Atmospheric components profile data under zones of different, Various Seasonal condition

S2. set up one group comprise analytic results can represent the Atmospheric components profile data set under zones of different, Various Seasonal condition;

S3. utilize radiation transfer equation all to simulate atmospheric envelope on destination channel to each profile and eject and penetrate radiance value, and the observation noise that adds instrument is to obtain the satellite sensor observation radiance value of simulation;

S4. according to simulation radiance value, the sensitivity analysis of temperature and Atmospheric components is selected to inverting passage;

S5. training sample set is carried out to empirical orthogonal function, calculate Empirical Orthogonal Function regression coefficient;

S6. obtain the clear sky of AIRS ultraphotic spectrum satellite data and correct spoke luminance standard product, utilize the regression coefficient of trying to achieve to carry out inverting.

Wherein, step S2 further comprises:

S2.1 collects one group of atmosphere profile that represents zones of different, Various Seasonal, and each atmosphere profile all comprises atmospheric condition parameter and earth's surface state parameter; Each atmosphere profile has comprised atmospheric condition parameter and earth's surface state parameter, for example atmospheric condition parameter: temperature, moisture content, CO ₂, O ₃, CO, N ₂o, CH ₄profile; , earth's surface state parameter for example: surface temperature, earth's surface moisture content, earth's surface pressure, earth's surface sea level elevation, earth's surface emissivity;

The all Atmospheric components profile samples of S2.2 are all treated to 101 layers according to the standard vertical barosphere of AIRS, and uniform units:

$P\left(x\right)={({\mathrm{<figure-callout\; id="2"\; label="ax"\; filenames="131231115936.png"\; state="\{\{state\}\}">ax</figure-callout>}}^2+\mathrm{bx}+c)}^{\frac{7}{2}}---\left(1\right),$

Wherein, P represents air pressure (hundred handkerchiefs); X represents barosphere ordinal number, from 1 to 101; Tri-parameters of a, b and c are by substitution a=-1.55 * 10 ^-4, b=-5.59 * 10 ^-2, c=7.45.

Wherein, step S3 further comprises:

S3.1 utilizes RTTOV11.1 to calculate atmospheric envelope on all passages and ejects and penetrate radiance value, is designated as R ₀;

S3.2 adds and take the random noise that Instrument observation noise is the upper limit for the simulation outgoing radiance value on each passage, as the satellite sensor observation radiance value R of simulation.

Wherein, step S4 further comprises:

S4.1 is for each profile of atmosphere profile data centralization, by CH ₄concentration increases by 10%, re-uses RTTOV11.1 and calculates atmosphere on all passages and eject and penetrate radiance value, is designated as R _cH4, calculate corresponding atmosphere and eject the mean value of penetrating radiance value difference value

in like manner, by the whole 1K that increases of temperature profile, re-use RTTOV11.1 and calculate atmosphere on all passages and eject and penetrate radiance value, be designated as R _t, calculate corresponding atmosphere and eject the mean value of penetrating radiance value difference value

vapour concentration is increased to 10%, re-use RTTOV11.1 and calculate atmosphere on all passages and eject and penetrate radiance value, be designated as R _h2O, calculate corresponding atmosphere and eject the mean value of penetrating radiance value difference value

by N ₂o concentration increases by 10%, re-uses RTTOV11.1 and calculates atmosphere on all passages and eject and penetrate radiance value, is designated as R _n2O, calculate corresponding atmosphere and eject the mean value of penetrating radiance value difference value

earth's surface emissivity is reduced to 0.02, re-use RTTOV11.1 and calculate atmosphere on all passages and eject and penetrate radiance value, be designated as R _ε, calculate corresponding atmosphere and eject the mean value of penetrating radiance value difference value

S4.2 selects

be greater than 0.2K and

be less than 0.8K and be less than 0.4K and

be less than 0.2K and be less than the passage of 0.1K.

Wherein, step S5 further comprises:

S5.1 calculates respectively the covariance matrix of methane profile sample set and simulation outgoing radiance,

C _cov＝ΔC(ΔC) ^T/n (2)，

R _cov＝ΔR(ΔR) ^T/n (3)，

C wherein _cov[l * l] is the covariance matrix of carbon monoxide profile sample set, Δ C[l * n] be the deviation matrix of profile and profile sample set average.Similarly, R _cov[m * m] is the covariance matrix of simulation AIRS observation spoke brightness, Δ R[m * n] be the deviation matrix of analogue observation radiation value and sample set average.Subscript T representing matrix transposition;

S5.2 is by C _covand R _covcarry out respectively empirical orthogonal expansion:

C _cov＝ΓΛΓ ^T (4)，

R _cov＝Γ _rΛ _rΓ _r ^T (5)，

Wherein Г [l * l] is the Empirical Orthogonal Function collection of profile sample set, and the proportion that each EOF accounts in population variance according to it is arranged by row, value is maximum come before, the l that Λ is comprised of the eigenwert of covariance matrix ties up diagonal matrix.Equally, Г _r[m * m] is the Empirical Orthogonal Function collection of analogue observation spoke luma samples collection, Λ _rthe m dimension diagonal matrix forming for the eigenwert of covariance matrix;

S5.3, for atmosphere profile sample set and simulation outgoing radiance, selects respectively front 50 and 40 proper vectors, carries out empirical orthogonal function:

ΔC＝Γ _1-50×A (6)，

ΔR=Γ _r1-40×B (7)；

S5.4 calculates respectively EOF score A and the B of Δ C and Δ R:

A=Γ _1-50 ^T×ΔC (8)，

$B={{{\mathrm{\&Gamma;}}_r}_{1-50}}^T\&times;\mathrm{\&Delta;R}---\left(9\right);$

S5.5 calculates the transition matrix D between two EOF scores, then obtains regression coefficient matrix S:

D=A×B ^T×(B×B ^T) ^-1 (10)，

S=Γ×D×Γ _r ^T (11)，

S5.6 is made as respectively 0 ° by incident angle and view angle, 10 °, 20 ° ... 70 °, take 10 ° as interval, for each profile of atmosphere profile data centralization, utilizing RTTOV11.1 to calculate corresponding atmosphere ejects and penetrates radiance, then according to the method for S4.1-S4.5, calculate regression coefficient S corresponding to different angles.

Wherein, step S6 further comprises:

S6.1 reads in AIRS clear sky by picture dot and corrects spoke brightness product, can from ( http:// mirador.gsfc.nasa.gov/) download AIRS clear sky and correct spoke luminance standard product (keyword is AIRS2CCF), form is HDF4.Read quality tab (radiances_QC) and the radiance value (radiances) of each each wave band of picture dot, then carry out clear sky and correct quality judgement; The passage of selecting according to S3.2, checks that the clear sky of each passage is corrected quality; If inverting passage has clear sky over half to correct quality for poor (radiances_QC=2), do not carry out inverting; If carry out inverting, but the clear sky of part inverting passage is corrected quality for poor, now this passage clear sky is corrected to spoke brightness and is made as sample average

S6.2 reads in solar zenith angle (solzen) and observation zenith angle (satzen) by picture dot, finds corresponding matrix of coefficients in S4.6

S6.3 utilizes clear sky that S5.1 reads in to correct the matrix of coefficients that spoke brightness and S5.2 read in by picture dot, according to following relation, obtains corresponding Atmospheric components vertical concentration profile:

$C_{\mathrm{obs}}=\stackrel{\&OverBar;}{C}+S\&times;\mathrm{\&Delta;}{R}_{\mathrm{obs}}---\left(12\right),$

Wherein, C _obs[l] is the gas concentration Vertical Profile (l layer) of inverting,

for gas profile sample set average, Δ R _obs[m * n] is observation radiance and simulation spoke luma samples collection mean bias matrix.

S6.4 reads in longitude and latitude information by picture dot, and the methane profile inversion result of step S5.3 and corresponding latitude and longitude information are written in HDF by picture dot, and Output rusults data layout is HDF.

Explanation of nouns in instructions:

ECMWF: the European Centre for Medium-range Weather Forecasts of European Study of Meso Scale Weather forecasting centre;

AIRS: atmosphere infrared sensor Atmospheric Infrared Sounder;

AMSU: advanced microwave detector Advanced Microwave Sounding Unit;

EOS: earth observing system Earth Observation System;

SVD: svd SVDSingular value decomposition.

Claims (6)

1. for a methane profile quadrature inversion method for AIRS ultraphotic spectrum satellite data, it is characterized in that, the method is specially:

1) utilize the secondary standard product of AIRS ultraphotic spectrum satellite to obtain the Atmospheric components profile data under zones of different, Various Seasonal condition;

2) set up one group comprise analytic results can represent the Atmospheric components profile data set under zones of different, Various Seasonal condition;

3) utilize radiation transfer equation all to simulate atmospheric envelope on destination channel to each profile and eject and penetrate radiance value, and the observation noise that adds instrument is to obtain the satellite sensor observation radiance value of simulation;

4) according to simulation radiance value, the sensitivity analysis of temperature and Atmospheric components is selected to inverting passage;

5) training sample set is carried out to empirical orthogonal function, calculate Empirical Orthogonal Function regression coefficient;

6) clear sky that obtains AIRS ultraphotic spectrum satellite data is corrected spoke luminance standard product, utilizes the regression coefficient of trying to achieve to carry out inverting.

2. the methane profile quadrature inversion method for AIRS ultraphotic spectrum satellite data as claimed in claim 1, is characterized in that described step 2) be specially:

A) collect one group of atmosphere profile that represents zones of different, Various Seasonal, each atmosphere profile all comprises atmospheric condition parameter and earth's surface state parameter; Each atmosphere profile has comprised atmospheric condition parameter and earth's surface state parameter, and atmospheric condition parameter comprises: temperature, moisture content, CO ₂, O ₃, CO, N ₂o, CH ₄profile; , earth's surface state parameter comprises: surface temperature, earth's surface moisture content, earth's surface pressure, earth's surface sea level elevation, earth's surface emissivity;

B) all Atmospheric components profile samples are all treated to 101 layers according to the standard vertical barosphere of AIRS, and uniform units:

$P\left(x\right)={({\mathrm{ax}}^2+\mathrm{bx}+c)}^{\frac{7}{2}}---\left(1\right),$

Wherein, P represents air pressure, and unit is hundred handkerchiefs; X represents barosphere ordinal number, from 1 to 101; Tri-parameters of a, b and c are by substitution a=-1.55 * 10 ^-4, b=-5.59 * 10 ^-2, c=7.45.

3. the methane profile quadrature inversion method for AIRS ultraphotic spectrum satellite data as claimed in claim 1, is characterized in that, described step 3) is specially:

C) utilize RTTOV11.1 to calculate atmospheric envelope on all passages and eject and penetrate radiance value, be designated as R ₀;

D) for the simulation outgoing radiance value on each passage, add and take the random noise that Instrument observation noise is the upper limit, as the satellite sensor observation radiance value R of simulation.

4. the methane profile quadrature inversion method for AIRS ultraphotic spectrum satellite data as claimed in claim 1, is characterized in that, described step 4) is specially:

E) for each profile of atmosphere profile data centralization, by CH ₄concentration increases by 10%, re-uses RTTOV11.1 and calculates atmosphere on all passages and eject and penetrate radiance value, is designated as R _cH4, calculate corresponding atmosphere and eject the mean value of penetrating radiance value difference value

in like manner, by the whole 1K that increases of temperature profile, re-use RTTOV11.1 and calculate atmosphere on all passages and eject and penetrate radiance value, be designated as R _t, calculate corresponding atmosphere and eject the mean value of penetrating radiance value difference value

vapour concentration is increased to 10%, re-use RTTOV11.1 and calculate atmosphere on all passages and eject and penetrate radiance value, be designated as R _h2O, calculate corresponding atmosphere and eject the mean value of penetrating radiance value difference value

by N ₂o concentration increases by 10%, re-uses RTTOV11.1 and calculates atmosphere on all passages and eject and penetrate radiance value, is designated as R _n2O, calculate corresponding atmosphere and eject the mean value of penetrating radiance value difference value

earth's surface emissivity is reduced to 0.02, re-use RTTOV11.1 and calculate atmosphere on all passages and eject and penetrate radiance value, be designated as R _ε, calculate corresponding atmosphere and eject and penetrate the average of radiance value difference value

Value $\stackrel{\&OverBar;}{|R_{\mathrm{\&epsiv;}}-R_0|};$

F) select

be greater than 0.2K, and

be less than 0.8K, and

be less than 0.4K, and

be less than 0.2K, and

be less than the passage of 0.1K.

5. the methane profile quadrature inversion method for AIRS ultraphotic spectrum satellite data as claimed in claim 1, is characterized in that, described step 5) is specially:

G) calculate respectively the covariance matrix of methane profile sample set and simulation outgoing radiance,

C _cov＝ΔC(ΔC) ^T/n (2)，

R _cov＝ΔR(ΔR) ^T/n (3)，

C wherein _cov[l * l] is the covariance matrix of carbon monoxide profile sample set, Δ C[l * n] be the deviation matrix of profile and profile sample set average; Similarly, R _cov[m * m] is the covariance matrix of simulation AIRS observation spoke brightness, Δ R[m * n] be the deviation matrix of analogue observation radiation value and sample set average; Subscript T representing matrix transposition;

H) by C _covand R _covcarry out respectively empirical orthogonal expansion:

C _cov＝ΓΛΓ ^T (4)，

R _cov＝Γ _rΛ _rΓ _r ^T (5)，

Wherein Г [l * l] is the Empirical Orthogonal Function collection of profile sample set, and the proportion that each EOF accounts in population variance according to it is arranged by row, value is maximum come before, the l that Λ is comprised of the eigenwert of covariance matrix ties up diagonal matrix; Equally, Г _r[m * m] is the Empirical Orthogonal Function collection of analogue observation spoke luma samples collection, Λ _rthe m dimension diagonal matrix forming for the eigenwert of covariance matrix;

I) for atmosphere profile sample set and simulation outgoing radiance, select respectively front 50 and 40 proper vectors, carry out empirical orthogonal function:

ΔC＝Γ _1-50×A (6)，

ΔR=Γ _r1-40×B (7)；

J) calculate respectively EOF score A and the B of Δ C and Δ R:

A=Γ _1-50 ^T×ΔC (8)，

$B={{{\mathrm{\&Gamma;}}_r}_{1-50}}^T\&times;\mathrm{\&Delta;R}---\left(9\right);$

K) calculate the transition matrix D between two EOF scores, then obtain regression coefficient matrix S,

D=A×B ^T×(B×B ^T) ^-1 (10)，

S=Γ×D×Γ _r ^T (11)；

L) incident angle and view angle are made as respectively to 0 °, 10 °, 20 ° ... 70 °, take 10 ° as interval, for each profile of atmosphere profile data centralization, utilize RTTOV11.1 to calculate corresponding atmosphere and eject and penetrate radiance, then according to step G)-step K) method calculate regression coefficient S corresponding to different angles.

6. the methane profile quadrature inversion method for AIRS ultraphotic spectrum satellite data as claimed in claim 1, is characterized in that, described step 6) is specially:

M) by picture dot, read in AIRS clear sky and correct spoke brightness product, read quality tab radiances_QC and the radiance value radiances of each each wave band of picture dot, then carry out clear sky and correct quality judgement; According to step F) passage selected, check that the clear sky of each passage is corrected quality; If inverting passage has clear sky over half to correct quality for poor (radiances_QC=2), do not carry out inverting; If carry out inverting, but the clear sky of part inverting passage is corrected quality for poor, this passage clear sky is corrected to spoke brightness and is made as sample average

;

N) by picture dot read in solar zenith angle and observation zenith angle, at step L) in find corresponding matrix of coefficients;

O) by picture dot, utilize step M) clear sky that reads in corrects spoke brightness and step N) matrix of coefficients that reads in, according to following relation, obtain corresponding Atmospheric components vertical concentration profile:

$C_{\mathrm{obs}}=\stackrel{\&OverBar;}{C}+S\&times;\mathrm{\&Delta;}{R}_{\mathrm{obs}}---\left(12\right),$

Wherein, C _obs[l] is the gas concentration Vertical Profile (l layer) of inverting,

for gas profile sample set average, Δ R _obs[m * n] is observation radiance and simulation spoke luma samples collection mean bias matrix;

P) by picture dot, read in longitude and latitude information, by step O) methane profile inversion result and corresponding latitude and longitude information by picture dot, be written in HDF, Output rusults data layout is HDF.

Images