S P A C E M E A S U R E M E N T S O F V E R T I C A L P R O F I L E S O F T R O P O S P H E R I C M I N O R C O N S T I T U E N T S
T n d a o A o k i
Meteorological Research Institute Tsukuba, lbaraki 305, Japan
A B S T R A C T
A method for remote measurement of tropospheric constituents from satellit.e is describe~, "".h̲ere the̲ yisibl<: and near infrared solar radiation in the sunglint region is measured with tunable etalons and very narrow band pass filter which contains only one absorption line of trace gases concerned. T h e mathematical procedure is describ叫
to retrieve the vertical profiles of constituents. T o study the feasibility of this method a ground based instrument ha.5 been developed and some results of remote mea.511rements of water vapor are presented.
I. I N T R O D U C T I O N
J.i'.or nearly̲tw,? decades ~he r~mote ~o~ndi'!g of t~e tropospl̲1eric gas̲es from ~atellite have been m a d e by measuring the emission from the atmosphere and ground or sea surface. Since 1978 ̲the T J R O S ‑ N satellite s̲eries has operntionally been obtaining the w?‑t:er v~por vertical. profile by the infr:ared_~mission from t~e atmos~her~
with a filter type radiometer. T h e error is more than 20 percent near the ground and worse in the higher altitude (Timchak,1986). T h e remote measurement of C O ha..c;
been m a d e by MAPS(Measurement of Air Pollution from Satellites) on board the space shuttle, also by measuring the infrared emission through g邸 filter. T h e accuracy has been estimated to be an order of 20 percent (Reiche et al., 1982).
In the future satellite program some new type of instruments such as a Fourier transform spectrometer and a grating spectrometer are proposed for remote sounding the tropospheric trace gases. These are again based on the measurement of thermal emission from the atmosphere. T E R S E (Tunable Etalon Remote Sounder for the Earth) is a n e w vertical sounder for the tropospheric trace gases proposed by Aoki et al.(1989, 1993), where the near infrared solar radiation in the sun glint region is observed with tunable eta.Ions where the separation of two etalon plates a.re changed with piezoelectric elements. In this paper our recent work is summarized .
II. R E T R I E V A L A L G O R I T H M
According to Aoki T a. et al. (1993) the signal output of the sensor at the i‑th sampling during a scan of the etalon plate is written in the form:
vj = e/V2 砂
p(U,h)蛉
(v)[Ir,JJ(v)+
lpath(v)],v 1
(1) :,vh~re C is th̲e ̲is. the trans~issi~n func~i_on of the t:mable
is the transmission function of the narrow band pass filter, v1 and v2 a.re the lower and upper limits of the function
¢B,
respectively. Ire/I is the radiance reflected at the surface and Ipath. is that scattered by the atmosphere.In the near Infrared region the path radiance Ipath. is one order of magnitude lcss than Ire/I over the La.mbertian surface whose albedo is about 0.05 w h e n there is no absorption . However, to minimize the effect of the path radiance it is better to use the sun glint. T h e intensity of sun glitter radiance that is reflected into the view angle of about 0.5 degree is the order of 10‑3 10‑5 of the incident radiat.ion(Plass et al., 1977;
Guinn, et al., 1979). This is 104 102 times larger than that over the Lambertian
surface. Thus in c邸e of sun glint I,,01h can be neglect叫 compared to Ire/I・ Otherwise it can be taken into account b邸 ed on, e.g., climatological values of aerosol distribution.
(14) 海 洋 化 学 研 究 第8巻第1号 平 成7年4月
For an ideal etalon ¢F P is given by the Airy function
奸
p(v,t) = (1 ‑ rFP ‑ AFp)21 ‑ 2rFP cos(21rvt)
+
r}p' (2)t = 2 n F p b cos'Y, (3)
where A F P is the absorptance of the etalon, r F P is the reflectance of the surface of the etalon plate and T1F p is the refractive index of the m e d i u m between the etalon plates.
6 is ̲the spacin̲g be~ween _t~e two etalon plates and
‑y
is the incident angle of the light.In this study
6
is changed by changing the voltage to be applied to the piezo elements.T h e transmission function for the absorption due to molecules in a slant path is given by
r(v,0)=exp[‑μ
o l 。
O O苓
k.,,g(z)pg(z)dz],μ = ―
cos0 1,
(4)
(5)
~hE;re 0̲ is the̲ zenith ~n~le of t~e s)an~ pa̲th, fl stands for the species of the molecule, k is the absorption coefficient and p is the density.
W e define a. nondimensional quantity of the output signal for the i‑th sampling by
Fj = ‑ M
V (6)For the normalization f四 or,
V,
w e take an average of measurements:︐ M
M
▽ 台
1‑ M
= ‑
,: :: ,.. (7)
where M is the total number of the samples. It should be noted that
Fi
does not cont.ain the constants C appeared in (1). This means that the absolute m e邸 urement of the intensity of radiRtion is not necessary in this method.B y expanding
Fi
around an initial guess w e obtainN
。
Fi(& ) = Fi° (& )+こ 8Fi
z J' •j = l
8
巧祐 =
F,,ob, ‑ F;゜=こ
Iくijりj
(8) (9)
k99
{)Fi
" =
OXj' (10)where F° is the F theoretically calculated under the condition of the initial guess. x is the unknown parameter to be determined and the density of the j‑th layer is given by
P;
= 月
(1+
rr;x;), (11)Transactions of The Research Institute of. (15) Oceanochemlstry Vol. 8, No. 1, April 1 9 9 5
where pj is the initial guess value for P; and u; is the standard deviation of the error in initial guess, which was given artificially in the present work. This formula was used for making every x to be ‑the same order. T h e ‑unknown parameters are iteratively det.ermined with the use of the m a x i m u m likelihood solution{Rodgers, 1976):
x ( n + l )
= [
Sご+
Iく(n)1 £S ‑ l [((n)戸
Iく(n)tx
+
Iく(n)X(n)], (12)where X ( n ) is the vector of x at the n‑th step of the iteration,
S,,,
is the covariance matrix of X, f〈(n) is the K appeared in (12) a.t the n‑th step and Si is the covariance matrix of the m e邸 urement error.III. G R O U N D B A S E D M E A S U R E M E N T S
T o examine the performance of the algorithm a ground b邸ed instrument h邸 been developed for water vapor soundings. T h e block diagram of this instrument is shown in Fig.1. Since in this case w e observe the direct solar radiation, I,eJI in (1) becomes (Aoki et al., 1994),
Iairect(v)
=
1°炉
(v)Or, (13)where
J 8 炉
(v) is the extraterrestrial solar radiation, which m a y have a strong wavenum‑her dependence in the range between v1 an.d v2 due to absorption by the solar atmo‑
sphere. W e assume that tlie absorption due to the solar atmosphere is well known so that w e can give the extraterrestrial solar r叫iation with a constant [ 8 and a known function <t>0(v).
n
is the solid angle of the field of view of the sensor. In the present study the view angle is 0.19 degree.T h e transmittance spectra ofthe tunable etalon for some ofthe etalon rlate spacing are shown in the top panel of Fig.2. A n etalon and an interference filter together constitute a very narrow band pass filter as shown in the panel marked as "FILTER"
in Fig.2. This band pass filter only covers a small wavenumber region which essentially contains only one absorption line of water vapor.
A n example of the output signal of measurements is shown in Fig.3 for the ca.c;e of the line at 654lcm‑1, where the abscissa is the electric voltage apllied to the piezo elements and the ordinate is the normalized output Fi, T h e solid line is the measurement and dashed one is the calculated. A deep dip of the curve corresponds to the water vapor absorption line at 654lcm‑1. T w o dips appear in the figure. This is because the continuous application of electric volt昭e from
O
to 1000 volt makes movement of the line of tunable etalon for more than one free spectral range.T h e density profiles have been obtained by iteratively fitting the calculated curve of the normalized output to the measured one. Actual fitting has been m a d e for the curves only within a narrow range of the voltage around one of two dips in Fig.3. T h e froundtrnth of thevertical p晶°tle of water vapor h邸 been obtained within two hours
rom the radiosonde observation at the site far by about 100 meter. A n example of ret.rieved profiles is shown in Fig.4. Here the thin solid line is the retrieval profile, the thick solid line is the rrofile from radiosonde observation, the dashed line is the initial profile and the thin solid line on the lefthand side of the figure is the difference between the retrieved and sonde observed profiles. It can be seen that the profile in the upper layers does not move greatly from the initial value. This is because the region around the absorpt.ion line center, where the information of the upper atmospheric layers is m u c h involved, is completely absorbed and so it does not bring the information of the upper layers. T o improve this w e have to choose a line of weak line intensity.
Fig.5 is another example of measurement m a d e in the Taklimakan Desert in China, where the water vapor amount is small (Aoki Te. et al., 1993). A s seen in the figure, in this case the absorption at line center is not large and the profile of the upper troposphere
(16) 海 洋 化 学 研 究 第8巻第1号 平 成7年4月
...J
ヽ I /
ー●一 /
I~忍
芝
Fig. l ~Joe~ diag̲ra.m of grnund‑ba.sed T E R S E , w here H S is the heliosta.t, E T F is the etalon, B F is the interference filter, T E is the tunable etalon and L is the H e N e laser.
' 1 0 1
WUNv'llIWSNv'~l
A
人 A
lh 1
人 A
A
0
, 0 1
一 \
,l Jj 65 76 51
マ 了 ▽ マ 了 了
マ Y T 戸 戸 □ X
了 ア □ 丁
一
IFP
FRAUN
!FILTER C
02 1 H 2 01 7 s u R F A C E
]FILTERED
6581 6583 6585 6587 6589 6S91 6593 6595 6597 6599
WAVENUMBER
(CH" 1 )Fig.2 Spectra of the wavenumber region for the 1120 sounding. T h e lowest two panels are the transmittance of the column atmosphere(the bottom panel is the filtered), the next two are those of the
出
0 and CO2, respectively. T h e panel signed as "FILTER" is the transmittance of the band p邸s filter. The・ next panel is the relative intensity of the e~t~ate!restrial sol:ir ra9iati?n a~d _ the top p̲anel is that of the tunable etalon for some value of etalon plate separation.Transactions of The Research Institute of. (17) Oceanochemistry Vol. 8, No. I, April 1 9 9 5
AFTER PEAK FITTIHG
I. 2
WU N' v' IG 'v 'c l
I. O',,,
0.8
G3NI~vk-WI
rJ N
0.6 0. 4
0.2
IAN0 [[H9(t : 6S《1, 36 1[A9 . H6HIH. IIT : 92/06/25 9 1 H ( 9 0 9 9 0 2 9 0 2
"' "" "
'9(C 1:9000 I N I(C9A 1 10N"° ・
A 1 nos. N01(l : 6 S O L U I(員11H : SS. 16
" t(fl1cruk[986.oo U9 9TH 6r 9[ 1 9 9. 26 IOIAl •ot•IE·ll• 11 .l《l O l IAIA N O . , I o n DATA MO . "
9 [ 2 r s 9 ,2 . 1S612 T[ 9909 : 0 . 0?00
ー: OBSERVED
‑‑‑:CALCULATED
̀ ` ` ^ ‑
0 . 0
。
100 200 300 400 500 600RAHP VOLT AGE (V l
700 800 900
Fig.3 A n example of the normalized output of measurement (solid line) with T E R S E and calculated one (dashed line). T h e abscissa is the electric voltage given to the piezo element of the tunable etalon.
" ● ● 1191
・ ・ "''・ "
叫
・4.19.i.':1漏沿
荘ttよ
in:
『;I . i n•OO
10.0
(
エ
H︶
1H :l l3 H
5 . 0
R E T R I E V E D P R O F I L E H I T H S O N D E
H2(l PRClFILE FRClH TERSE U S E D B A N D S :
6 5 4 1 . 3 CH―I U S E D FP SPACING:
D . 1 8 1 3 C H H E A S U R E H E N T ERROR:
0 . 0 1 0
0 . 0
‑ 1 0 . 0 0 . 0 10.0 2 0 . 0 3 0 . 0 4 0 . 0 5 0 . 0 6 0 . 0 7 0 . 0 8 0 . 0 9 0 . 0 1 0 0 . 0
N U M B E R D E N S I T Y (102
IH B L / K H )
Fig.4 Comparison between retrieved vertical profile of water vapor (thin solid line) and sonde observation (thick solid line). T h e dashed line is the ini‑
tial profile and the curve on the left hand side is the difference between retrieved and sonde observed profiles.
(18) 海洋化学研究 第8巻 第1号 平成7年4月
1 0 . 0
H
四
U S E D B A N D S : 6 5 8 9 . 3 c w t
(
エH
︶
1H :J I3 H
5 . 0
P R 0 F I L E F R 0 M T E R S E
0 . 0
‑ 5 . 0 0 . 0
l i
︑`
̀ ̀`
︑`
︑¥
\
`
ヽ︑
`
1 ,
`
`
`
¥
\
︐
︑̀
`
'
︐
`
`'
`
,`
ー
︐' ``
︐
U S E D F P S P A C I N G : 0. 1 8 1 3 Ct1
M E A S U R E M E N T E R R O R : 0 . 0 1 0
5 . 0 1 5 . 0 2 0 . 0
N U M B E R D E N S I T Y (1
02 1M C J L / K M )
Fig.5 A s in Fig.4 except for observation at Taklimakan Desert in China.2 4 . 0 (
T O H
20.0t‑
go I)
゜ ゜
I 6. 0
・ p
x‑‑o
。 ‑ • - ‑ ‑ ‑ ‑ ‑ ‑ ‑ ‑ ‑ ‑ ‑ ‑ ‑12. 0 t‑ O
N H
8. 0 i‑‑
T V L O L
H匹 P R O F I L E F R O H T E R S E U S E D BANDS•
6 5 4 1 , 3 C H―I
‑ ‑ 0 ‑ ‑ ‑ ‑ ‑ ‑ ‑ ‑ ‑
0 0 ―
O : T E R S E X:R!,DI(lS(lNDE
〇.
o
8
︐
10 11 12
T I M E
13 I 4 15 16
Fig.6 Diurnal variation of the total column amount of water vapor mea‑
sured by T E R S E .
'!'1°ansact̲lons̲ or The̲ ~e~earch Instlt~te or. (19) Oceanochemlstry Vol. 8, No. 1, April 1 9 9 5
is well retrieved. Fig.6 shows the diurnal variation of the total column amount of water vapor observed by T E R S E . T h e sonde observation is shown by the mark of cross. It cRn be seen that inspite of the complete absorption of the part of the line center the total column amounts of water vapor are retrieved in good accuracy. T h e error is the order of a few percent.
R E F E R E N C E S
Aoki Ta., A remote sounder for the vertical profiles of the atmospheric gases: simulation, Proc. of Spring Meeting of Japan Meteorological Society, M a y 22‑2./, Tokyo, 62‑
62, 1989. {in Japanese)
Aoki Ta., Fukabori M .
,
Aoki Te, Trace gas remote sounding from near I R sun glint observation with tunable etalons, N A T O Advanced Ruearch W o r b h o p "High Spectral Re.,olution Infrared Remote SenJing for Earth•_, Weather and Climate Slttdiu", Springer‑Verlag, Berlin, 309‑322, 1993.Aoki T a., Fukabori M., Aoki Te., Remote measurements of atmospheric constituent.s in the lower layers, Proceeding‑'of the International Sympo.,ium on "Global Cycle., of Atmo.,phcric Greenhotue GaJu", Tohoku Univ., Sendai, Japan, 165‑170, 1994.
Aoki Te., Aoki Ta., Fukabori M., Remote sounding of the atmospheric water vapor at Hotan in Taklimakan Desert, Proceeding‑'of the Japan‑China International Sympo.,ium on the "Study of the Mcchaninn of DeJertification", Science and Technology Agency, Japan, 180‑184, 1993.
Guinn J. A., Jr., Plass G. N., Kattawar G. W., Sunglint glitter on a wind‑ruffied sea:
further studies, Applied Optic‑', 18, 842‑849, 1979.
Plass G. N., Kattawar G. W ., Guinn J. A., Jr., Isophotes ofsunglint glitter on a wind‑
ruffied sea, Applied Optic‑', 16, 643‑653, 1977.
Reichle H. G., Jr., Beck S. M., Haynes IL E., Hesketh W . D., Holland J. A., Hypes W . D., Orr III II. D., Sherrill R. T., Wallio 11. A., Casas J. C ., Saylor M . S., Gormsen B. B., Carbon monoxide measurements in the troposphere, Science, 218, 1024‑1026, 1982.
Rodgers C. D., Retrieval of atmospheric temperature and composition from remol.e measurements of thermal radiation, Review of Geophy.,icJ and Space Phy.,ic.'
,
11, 609‑624, 1976.
Timchak A., Satellite‑derived moisture profiles, N O A A Tech. Rep. NESD!S, 24, 1986.
(20) 海 洋 化 学 研 究 第8巻第1号 平 成7年4月