Retrieval of total ozone quantity from high resolution infrared spectra : influence of spectroscopic and physical parameters

Retrieval of

total

ozone

quantity from high resolution infrared spectra

:

influence

of

spectroscopic

and

physical parameters

A.

Barbe,

A.

Hamdouni,

J.J.

Plateaux, (1)

Ph. Demoulin (2)

1

G.S.M.A., U.R.A.

D

1434 CNRS,

UFR

Sciences

BP

347

st06z

RErMS

France

2

Université

de Liège

-

Institut

_{d'Astrophysique B}

4000

COINTE

Liège Betgique

In

this

work, we

present

the

results obtained

for

total

ozone

quantity

(TOQ) from

Jungfraujoch observatory spectrometer using 3 different spectral regions.

The spectrometer is a University

of

Liège FTS

with

a resolution power

of

0.0025

cm-l

and typical SÂ'l ratio larger than 1500. The spectra cover the region

from

1900

to

3500 cm-1.

In

principle the use

of

isolated lines is perfect, as all the spectroscopic parameters are known, as well as their temperature dependence, that is

to

say position, intensity, halwidth and ground energy level. The three band are given in the next table

I,

as

well

as their band centers and the band strength given in cm-1/molecule.cm2. We note that the ratio of intensities is roughly a

factor

10

or

50,

Vn

vtl

vttv'?

vl v?v?

Sband

2705.239r

ur

_{000 0.243}

_x

_l0-le

2110.7843 101 000 0.124

x

10-l

3046 0882 003 000 0.131

x

10-Iô

permitting

to

cover different zenith angles,

but which

makes simultaneous

retrieval

not

very

easy.

The line positions and intensities

for vl+v2+v3,

v1+v3, 3v3

are given respectively

in

references

l,

2,3,

as

well

as ground energy levels. The

air

broadening coefiicients are given

in

referenc

e

4

and pressure

shifts

in

reference

5.

We

derived

TOQ

by

least

square

fitting

between observed and

calculated

transmittances.

The

calculated transmittance

uses

Voigt

function, using

Whiting

expression modified by Oliveiro and Longbothum and the atmosphere is divided in 30 layeis in whicir Pressure

(P)

and Temperature

(T)

are considered as constants (reference

6). P

and

T

have been obtained by soundings during the same day as the recording spectra.

We

show

in figure

1

to

5

the

examples

of fits

which

have been obtained

in

the different

spectral regions. For each of these fits the TOQ is

given

in Dobson units.

The influence

of

different spectroscopic parameters of the retrieved TOQ has been tested :

*

_{Line strength S :}

_retrieved

_roQ

_{is of course inversely}

_{proportionalto}

s.

*

_Halfrryidth

_:

_we

_have

_plotted

_in

_figure

₆

_{the value}

_of

T

O

_Q.

with

respect

to

cr

for two

different spectral regions.

t

_{n coefficient, defined by}

_a(T):cr(To)

f

+)"

The results are the

following, for

the 3040

cm-

I

and

\.r/

with

cr:8410-1 cm-l

atm-i.

n:

0.75

n:1.3

-+

_-+

TOQ:323

-

:344

D.U

D.U=f(o"

₎

lfawo Nunbcr (oa-1)

Figure 7

60

70

*For the

pressure shifts

we

have

plotted

in

figure

7 the

differences

(in

%) calculated

with

Âv ₌

_{-10.10-3cm-l.atm-l}

andÂv ₌ 0

All

these results are summarized in table

II,

part A.

A.

Spectroscopic parameters G.1, !g mK/atm

between

spectra 1.u 2.S 3.ct

4.Âv

5.n Total

-+

-+

-)

-)

-)

0eÂ

4%

1%

0%

20Â 1 _L/O0,/ posrtron line strength halfividth pressure shift variation

ofcr

In

table

II,

part

B,

we have estimated the errors coming

from

instrument function, that is

to

say the determination of zero level of transmittance and of the apparatus function.

B.

Instrument

t.

2 0o% transmittance Apparatus function #3% JJ.O / ttL /o

To

estimate

the

influence

of

geophysical parameters

we

have tested

the

influence

of a

systematic

variation

of

3"c

on

T.o.Q.,

for the 3 different spectral regions (part.

_c).

C. Geophysical parameters

l-Temperature.

For

2775

and

3040

cm-l

regions

(lower

state energy

of

17.6

cm-l

and

50.64

cm-l

respectively, we have found .

ÂT:

*

3oC

_->

ÂDU:

+1.8oÂ

For

v1+v3

(E"

:

447

cm-l),we

have found

AT:*3oC-+ÂDU:-3%

2.Shape

of

the

03

vertical distribution Difificult

to

estimate Order : Z oÂ

I 6 e. d J Ê ç a o c c E E

l(avc Nuab* (o-r)

e F <P1211 2Ô3/N _" rcqlçurâl'd)--Figure 3 Figure 5 290 5 6 É

*

)ul wrv- !{uEtr-r (oa-1)

Ir-v€ NuEbGF (ca-a)

Figure 4

But

the final error

of

T.O.Q. is not the sum of the three

d

B

and C parts, as many

of

these errors are correlated

(T,n,a) or

have the same effect (ex : S and \Yo transmittance).

Then we finally estimate the

theoretical

maximum

eror

of absolute

total

ozone quantity as :

4%+

s,00lo

loÂ+ 2oÂ

+

2%o

:9

Yo

c[,n,T, shape

The experimental final results obtained from the different fits are summarized in table

III.

Results

2080

(v1+v3;

2775

(v1+u2+v3;

30a0

(3vr)

wide spectral region

(wsr)

I

line wsr

I

line

wsr

lliné

Afrxed

256

277

287

300

Ovanable 277

282

290

307

312 cr

variable

293

282+

11

290+

5

312+

t2

CONCL.

Z9S

+ t7

DU

6'/"

Results

They are given

for

the 3 different spectral regions, covering a large spectral region

or

the best single

line

of this

region.

We let

the radius

of

aperture

of

FTS free

to

have the best

fit

and then the final

results are 282

_;290

and 312 Dobson

Units

corresponding

to

295

xl7

Dobson

unit.

These results are in the range of theoretical error

.9

oÂ.

Conclusion

. We believe that there are

two

points which can lead

to

a better ozone retrieval : one is concerned

with

absolute determination of intensity where the goal is

to

have

a

_{real lo/o accuracy ; the}

second

is

to

have

a

better

knowledge

of

instrumental apparatus

function

(including

zero

level transmittance). In this case, we may expect a final retrieval of T.O.Q. better than 4 Yo.

References

1 The 3.6 pm region of ozone. Line positions and intensities

M.A.

Smith, C.P. Rinsland,

M.

Devi, J.M. Flaud,

C.

Camy-peyret,

A.

Barbe,

Journ.

of

Mol.

Spectr.,1990, 139

p

171-181

2 Intensities and self broadening coefficients of

03

in the 5 prm region

J.J. Plateaux, S. Bouazza,

A.

Barbe, Journ. of

Mol.

Spectr.,199l,746, p

314-325

3 The 3.3 prm region

of

ozone. Line positions and intensities

C.

Cany-Peyret,

J.M.

Flattd,

M.

Smith, C. Rinsland,

M.

Devi, J.J.

Plateaux,

A.

Barbe,

Journ.

of

Mol.

Spectr.,l99A,

l4l,

p

134-144

4 Nitrogen and oxygen broadening coefficients of ozone

S. Bouazza,

A.

Barbe,

J.J.

Plateaux,

L.

Rosenmen,

J.M.

Hartmann,

R. Gamache, Journ. of

Mol.

Spectr., 1993, 157 _,

p27l-289

5 Pressure shifts

of

03

broadened by

NZ

and C,2,

A. Barbe, S. Bouazza

andJ.J.

Ploteaux, Applied Optics, 1991, Vol.30,

6 P. Marché, Thesis, Reims, 1980

Acknowledsments:

We

tlrank

R

Zancler

_for

useful

discussions snd

L.

Delbouille

ancl G.

Roland

for

their

help

during tlte

obsen'ations at the

Jungfraujoch-C.

Camy-Peyret,

J.M.

Flaud,

nol8, p

2431-2436

s :ar,:a;;:

n'g

, ,t,t';.

I

d

p

çC,IROSCOp

3o,

ffi'\

't1,1,$1:-',j

?&

/

o

7

o

,a:.: ..' -.:. :.,

&

+B

R tt

l

IT .s IT, r.. l i.i 'i _{I .t.}S,