S P A C E AND T I M E V A R I A T I O N S OF K 1 7 6 9 9

S O L A R L I N E P R O F I L E

T. R O C A - C O R T E S a n d M . V A Z Q U E Z

lnstituto de Astrofisica de Canarias, CSIC, Universidad de La Laguna, Tenerife, Spain

and

H. W ~ ) H L

Kiepenheuer lnstitut ~ r Sonnenphysik, Sch6neckstr. 6, D-7800 Freiburg, F.R.G.

(Received 8 July, 1982; in final form 24 March, 1983)

Abstract. The temporal and spatial behaviour of the K 17699 line profile is investigated. In particular we have measured the asymmetries of the line profile at several residual intensities using the bisector method. We find, in the bisector of the mean line profile, similar shapes as those obtained before for different positions on the solar disk. However the strong variations of the bisector found with time and geometry of input aperture, warns us against the use of the mean or integrated profiles (either in time or space). Moreover, we find an asaticorrelation between the asymmetry in the line profile at different residual intensities and the shift, found as the distance to a terrestrial line, for any position observed on the solar disk. No limb effect for this line is found, within errors.

I. Introduction

Non-resolved velocity fields produce asymmetries in absorption line profiles. In t he solar photosphere it is believed that they are the consequence of the space average over the granulation velocity field (Voigt, 1956; Schr6ter, 1957) although alternative or complementary mechanisms have been proposed (see Dravins, 1982, for a recent review).

The line K 17699 has been used recently in whole disk measurements of the solar oscillations with a resonant scattering technique (e.g. Brookes et al., 1978). Their asymmetries have been studied by Snider (1972) and Gasanalisade (1979) although with a poor time resolution. The purpose of this paper is to study the behaviour of the line prone looking for space and time variations.

2. Observations and Reduction

The observations were made during two periods: May-June 1978 and Septem- ber- October 1980, in the solar station of the G6ttingen University at Locarno using the 45 cm Gregory-Coud6 reflector. Details of the equipment are given by Wiehr et al.

(1980). A spectral region containing the KI7699 solar line and the terrestrial 02 7697 was

photoelectrically scanned using a stepping motor (1 step = 0.72 m~). At the same time another photomultiplier was recording the nearby continuum. In order to avoid trans- parency fluctuations the ratio between the two detector signals was taken. Every scan consisted of two measurements over the region (first towards the blue then back)

Solar Physics 88 (1983) 1-8. 0038-0938/83/0881-0001501.20. �9 1983 by D. Reidel Publishing Company.

2 T. ROCA-CORTES ET AL.

recording their sum on magnetic tape. The duration of this procedure was 53 s per scan. The observational sets were done at the disk center with different apertures consisting of holes in a mirror placed in front of the spectrograph slit and at several positions on the north part of the solar rotation axis. The census of the observations is given in Table I.

TABLE I

Census of the observations

Observation Position Aperture Date Number of scans series (/~ = cos 0) of observation measured

0 Integral light 1 Oct., 1980 26 1 1.0 40 "a 26 Sept., 1980 53 2 1.0 40 "a 5 Oct., 1980 61 3 1.0 8" 6 June, 1978 50 4 1.0 8" 16 Sept., 1980 63 5 1.0 4" 28 May, 1978 60 6 1.0 4" 5 Oct., 1980 125 7 0.8 4" 29 May, 1978 59 8 0.8 4" 24 Sept., 1980 61 9 0.7 4" 28 Sept., 1980 57

10 0.6 4" 5 Oct., 1980 62 11 0.5 4" 29 May, 1978 19 12 0.5 4" 25 Sept., 1980 30 13 0.3 4" 1 June, 1978 59 14 0.3 4" 1 Oct., 1980 61 15 0,2 4" 1 June, 1978 59

a In these scans the effective aperture was the exit slit of the spectrograph.

The solar region under investigation was displayed on a TV monitor via a Ca + K filter. We carefully avoided observations in or close to active regions.

For the full disk measurements the light was collected by a small coelostat system with 5 mirrors onto the spectrograph slit. Since the diameter of the mirrors was small and this system did not have appropriate tracking we believe that vignetting could happen while taking these spectra.

To determine the wavelength distance between the terrestrial and solar lines we have searched for the apparent minimum of each line and then we fitted different vertical parabola. We tried different numbers of points around the apparent minimum (80-200 for the solar line and 60-130 for the terrestrial one), and the minimum of the parabola which best fitted the data (in the least-squares sense) was taken.

The continuum intensity was measured in 3 segments free of weak lines giving a r.m.s. fluctuation of 1.5~o on average. After a parabolic smoothing (over 5points) we measured the asymmetry of the lines using the bisector method as the positions of the midpoints which have equal residual intensities referred to the minimum of the fitted parabola as zero-reference.

Measurements of the spectrograph instrumental profile were done using a H e - N e laser. The profile was symmetric within 0.2 m*. Using this profile we proceeded to

90

SO

70

--

60

>-

2:50

uJ

z

30

90

80

?0

�9

-..,

60

5o

uJ

40

30

FU

LL

D

ISC

e(1-

10-6

0)

I I

-2~

)0-1

00

0

~10

0 +

20

A

k (M

ET

ER

S

PE

R

SE

CO

ND

)

,u.

o.7

; ~

.u'

e(28

-9-8

0)

I -

2~0

-,~0

0 '

+1

00

+

20

0

A

~ (M

ET

ER

S

PE

R

SE

CO

ND

)

Fig.

1.

}.~=

1.0

; L.

I,O

"

�9

1

-20

0

-10

0

0 I

+10

0 +

20

0

}.J.=

1.0

;

,~=

8"

e(1

6-9

-80

) X

(6-6

-78

)

; X

ex

�9 X

�9 X

�9 X

e

Xo

X

�9

• �9

Xe

X

�9

X�9

I I

I I

-20

0

-I0

0

0 +

10

0

+2

00

,~L=

1.0

;

J~=

/."

el

5-1

0-8

0)

X(2

8-5

-78

)

�9 X

�9 X

eX

Xe X

e x�

9

x�9

xm

X�9

xe

Xe

Xe

-20

0

-1

0 +

10

0

+2

00

}/.

0.6

; ~'

= t,,

,

e(5

-10

-80

] ~

U,O

.3

; ~

', I

...

e(1

-10

-80

)

x(1

-6-7

8)

�9

x

�9

X

�9

x

�9

X

�9

x

�9

x

�9

x

�9

x

�9

X

�9

x

�9 x

)4

1

- ]-

- 1 I .-

i

�9 t

I 0 i

+10

0 +

20

0

)/=

O.S

;

p%

4"

e(2

5-9

-80

) ;<

(29

-5-7

8)

. v

�9

x �9

x �9

x �9

x

�9

x

�9

x

�9

x

�9

x �9

x

�9

x �9

x

xe

, ;

I -2

00

-1

/0

+2

00

+

10

0

-20

0

-10

0

-2

0 -1

0

+1

00

+

20

0

Mea

n b

isec

tors

cor

resp

ondi

ng t

o th

e di

ffer

ent

obse

rvat

iona

l se

ts.

Typ

ical

err

ors

are

of 3

0 m

s-

~.

}.1

.0.8

,;

9'=

L"

e(2

4-9

-f1

0)

X(2

9-5

-78

)

X

�9

X

�9

X

�9

X

�9

X

�9

X

�9

X

�9

>r

�9

X

�9

X

�9

X

�9

X

�9

eK

-20

0

-10

0

~ +

10

0

+2

00

~LI.

0.2

; ~

',/.

"

X(1

-6-7

8)

X

X

X

X

X X

X

X

X

X

X

X

I

-20

0

-1

0 +

10

0

+2

00

�9

�9

4 T. ROCA-CORTES ET AL.

restore the spectra by the optimum filter method, and then recalculated the minima and the asymmetries. The results found were equal, within errors, to the ones found without the restauration process.

3. Variations of the Asymmetries with the Position on the Disk

In Figure 1, the bisectors corresponding to an average of all scans at different positions on the solar disk are displayed. Typical errors are of 30 m s- 1. The terrestrial line shows asymmetries less than 40 m s-~ (-~ 1 m.~) except near the continuum, probably due to a blend.

In general there are blueward shifts except for the 1978 observations near the limb. In the positions, on the solar disc, where we have measurements in the two observing periods, some differences are detected but an interpretation as real long-term variation as reported by Livingston and Hollweger (1982) for whole disk measurements is still speculative.

The bisectors of the disk center have a great similarity with the ones observed by Gasanalisade (1979) except in that we do not observe the kink toward the red at 7 0 ~ residual intensity. This kink was predicted by Dravins et al. (1981, see their Figures 6a and 7) for strong resonance lines in this wavelength region taking into account the granular velocity field only. Near the limb the agreement with Gasanalisade (1979) is good. All qualitative effects seen at disk center remain in integrated light with the amplitude somewhat reduced.

4. The Limb Effect

The variation of the mean distance between the lines, corrected by rotation and orbital motion of the Earth, along the position on the solar disk is the so-called limb effect. From our data we do not find any significant effect, within errors, as can be expected taking into account the spectral region we are dealing with and the height of formation of the line. This behaviour is the same as the one reported by Snider (1972). A bump of 200 m s- 1 was detected between positions # = 0.8 and # = 0.7 which we interpret as the supergranulation velocity field which we cannot average. In accordance with these results we do not find any significant difference between integral light and disk center values.

5. Time Variation

The variation of the bisectors with time for different positions on the solar disk is shown in Figure 2. As one can see their shape and position change greatly confirming the apparent erratic behaviour reported by Snider (1970).

There can be seen that the bisector looks very different for distinct apertures and positions; for instance the scans measured at # = 1.0, with the exit slit of the spectro- graph as the effective aperture (observation series I and 2) show a very much less erratic behaviour than the others, probably due to the fact that they have been measured without a hole-mirror and so have a different distribution of the intensities.

BIS

EC

TOR

S

OF

THE

Kf

SO

LAR

LI

NE

A

T "~

69

9

ANG

STR

OM

S

"['~

1

61

~)1

w

>.-

1,o

~1

cr

mL

o,~

0511

01~

>,..

%0

.t

m

�9

i ii

0.8

�9

6-

0o~

O

0,6

i~1

0.4

0.3

>-

1.0

~s'lO

slO

E::

,~

o~

i~1

o,l

�9

o.t

0.5

0.2

Fig.

2.

Som

e ex

ampl

es o

f ti

me

vari

atio

ns o

f th

e li

ne b

isec

tors

sho

win

g di

ffer

ent b

ehav

iour

. T

ime

inte

rval

53

s.

0 >

6 T. ROCA-CORTES ET AL.

To check this time variation we have used a direct FFT power spectrum analysis of the asymetries at different residual intensities. As most of them showed a significant peak around 5 min (see Figure 3a), we have tried to correlate this temporal behaviour with the 5 min velocity oscillation measured by the distance between the solar and terrestrial lines.

ZSO0

2.000

~.500

~.000

500 _ #

I i

2 6 7 8 9

FRECUENCY (m Hz) (a)

,;/ / "A," ,;-" 1' ~) i . , "~ / 2)

#

v "V

/ /

Lvj ",xx

,4 / | \\\

(b)

Fig. 3. These plots refer to the scans made at # = 1.0 and L = 40" without hole mirror, (observation series number 1). (a) Power spectra of the distance between solar and terrestrial lines, (corrected by Earth rotation in every scan) showing the 5 rain peak. (b) Cross-correlation function of the distance between lines and asymmetries of the solar one at different intensities: (1) at 40 ~o; (2) at 60 ~o; (3) at 80 ~o; (4) at 90 ~o of the

continuum intensity. Dotted lines correspond to the significance levels of 99 and 95 ~o of the null test.

K17699 LINE PROFILE VARIATIONS 7

The cross-correlation function was found between the different series of asymmetries and the velocity determinations. An example is shown in Figure 3b. We set the levels of significance of 95~o and 99~o of a null test and the result is that there is significant anticorrelation, at 0 lag, in most of the cases; only in the observation series numbers 2, 11, and 13, (see Table I) the correlation failed at those significance levels, but in those cases the 5 min oscillation was buried in noise. Similar calculations were done using the distance between the apparent minima of the lines to check if the correlation was an artifact of the parabola fits, obtaining the same result.

This anticorrelation means that when the solar line is shifted to the red due to material motions the profile of the line shows a blueward asymmetry. A similar effect was found by Roddier (1967) in the resonance line Sri4607 showing a change from asymmetric to symmetric profiles.

Slaughter and Wilson (1972) have reported a correlation between the residual intensity at the bottom of the line profile and the 5 min oscillation for the Na D lines. To check that, we have done power spectra of such intensity and found no significant power at any frequency.

6. Conclusions

The asymmetries of the K17699 A solar line varies with the position on the disk, time and field of view. Looking at the time behaviour of the bisectors of the spectral fine, we conclude that the shape of the mean profile can change considerably depending on the time interval used to measure (Higgs, 1962). The geometrical configuration of the aperture used to observe does affect the observed profile as well.

The most important finding is the anticorrelation found between the asymmetries measured at different residual intensities and the photospheric oscillations (mainly by 5 min). Such effects were proposed by Erikssen and Maltby (1967), Kostik and Orlova (1972), Barylko (1978), and Cram et al. (1979). This relation was not detected by Koch et al. (1979) for 4 fines and Kostik and Orlova (1974, 1977) for 11 iron lines. However we think that our fine is more adequate to look for this correlation, for 3 main reasons:

(a) the line is formed higher in the photosphere and is therefore freer of convective motions;

(b) the granulation contrast is smaler in the red part of the spectrum (Wittman, 1978); (c) the amplitude of the acoustic waves increases with height. It is also interesting to note that Altrock and Keil (1977) found in some regions

correlations between velocity and intensity fluctuations in the Mg4571 line formed near the temperature minimum. Staiger (1982) has measured the phase difference between intensity and velocity oscillation for this Mg b line. His results differ from the normally accepted value of 90 ~ for the 5 min oscillation. As our samples are 53 s time averages we cannot exclude this influence in our measurements.

The full disk observations taken were not long enough to draw any conclusion. Although the profile seems to be asymmetric we cannot find any significant time variations. Observations of this kind will be of interest in the light of the use of this line in resonance scattering spectrometry.

8 T. ROCA-CORTES ET AL.

Acknowledgements

W e t h a n k P r o f E . H. SchrOter, Drs D. Drav ins , a n d J. A. Bonet , and M rs In6s M a r q u e z

for in te res t ing d i scuss ions a n d suggest ions. W e express also our gra t i tude to M r

R. G o m e z for helping in the compu ta t i ons .

References

Altrock, R. C. and Keil, S. L.: 1977, Astron. Astrophys. 57, 159. Barylko, M.: 1979, Acta Astron. 29, 43. Brookes, J. R., Isaak, G. R., and Van der Raay, H. B.: 1978, Monthly Notices Roy. Astron. Soc. 185, 1. Cram, L. E., Keil, S. L., and Ulmschneider, P.: 1979, Astrophys. J. 234, 768. Dravins, D., Lindergren, L., and Nordlund, A.: 1981, Astron. Astrophys. 96, 345. Dravins, D.: 1982, Ann. Rev. Astron. Astrophys. 20, 61. Erikssen, G. and Maltby, P.: 1967, Astrophys. J. 148, 833. Gasanalisade, A. G.: 1979, Soln. Dann. No. 7, 85. Higgs, L. A.: 1962, Monthly Notices Roy. Astron. Soc. 124, 51. Koch, A., Kt~veler, G., and Schr6ter, E. H.: 1979, Solar Phys. 64, 13. Kostik, R. I. and Orlova, T. V.: 1972, Solar Phys. 26, 42. Kostik, R. I. and Orlova, T. V.: 1974, Solar Phys. 36, 279. Kostik, R. I. and Orlova, T. V.: 1977, Solar Phys, 53, 353. Livingston, W. and Hollweger, H.: 1982, Astrophys. J. 252, 375. Roddier, F.: 1967, Astrophys. J. 147, 1113. Slaughter, C. D. and Wilson, A. M.: 1972, Solar Phys. 24, 43. SchrOter, E. H.: 1957, Z. Astrophys. 41, 141. Snider, J. L.: 1970, Solar Phys. 12, 352. Snider, J. L. : 1972, Phys. Rev. Letters 28, 853. Staiger, J.: 1982, personal communication. Voigt, H. H.: 1956, Z. Astrophys. 40, 157. Wiehr, E., Wittmann, A., and W6hl, H.: 1980, Solar Phys. 68, 207. Wittmann, A.: 1978, 'Small Scale Motions on the Sun', Mitt. Kiepenheuer-Inst. Nr. 179, p. 29.