POLARIZARION OF FeI LINES IN SOLAR
MAGNETOCONVECTION SIMULATIONS
.
gotzon
18-1-95
.
E. V. Khomenko(1, 2), S. Shelyag(3), S. K. Solanki(3) and A. Vögler(3)
(1) Main Astronomical Observatory, Kyiv, Ukraine;
(2) Instituto de Astrofísica de Canarias, Tenerife, Spain, [email protected];
(3) Max-Planck-Institut für Sonnensystemforschung, Katlenburg-Lindau, Germany.
RADIATIVE MHD SIMULATIONS
SUMMARY
We have used a 3D MHD code called MURAM
(MPAe/University of Chicago RAdiative MHD). The code
includes non-grey radiative transfer, full compressibility, and
the effects of partial ionization for the 11 most abundant
chemical elements (see Vogler, 2003). The size of the
computational domain is 6000 by 6000 by 1400 km3. The
horizontal resolution is 20 km. The atmosphere extends 600 km
above the level t5000=1. The snapshots used here have a bi-polar
structure of the magnetic field with an average (unsigned)
strength <B> = 30, 80 and 140 G (as representative of quiet
solar regions).
Realistic solar magneto-convection simulations are used to study
the polarization of the following Zeeman- sensitive photospheric
lines: Fe I 6301, 6302, 15648, 15652 A. The Stokes spectra
are synthesised in a series of snapshots with a bi-polar
structure of the magnetic field and average unsigned flux
varying from <B> = 30 G to 140 G. The profiles are smeared to
simulate the effects of the telescope and atmospheric seeing. A
Principle Component Analysis approach is applied to classify the
profiles. The synthetic Stokes spectra have many properties in
common with those observed in the solar quiet network and
inter-network regions. In particular:
SPECTRAL SYNTHESIS
(1) The simulations reproduce width and depth of the Stokes I
profile (Fig. 1).
The spectra of FeI 15648, 15652, 6301 and 6302 A lines
where calculated under LTE in 1.5D approach for every vertical
column of the simulation box.
ASYMMETRIES
(2) The amplitudes of the polarization signals of the Fe I 15648
line in the <B>=30 G snapshot are similar those observed in the
inter-network (Fig. 2).
Fig. 3.Histograms of the amplitude (top) and area (bottom)
(3) The Stokes V profiles show, on average, positive asymmetry.
Under reduced resolution it becomes close to the observed in the
quiet Sun (Fig. 3 ,4, 5).
(4) The average asymmetry of Stokes V depends on the
magnetic flux in the simulations (Fig. 3). The amplitude
asymmetry decreases and the area asymmetry increases with
decreasing <B>.
(5) The profiles belonging to the same class are clustered
together and form patches on the solar surface, close to the
spatial resolution in size (Fig. 4, 5).
Fig. 1. Comparison between the synthetic (averaged over the
snapshot, red line) and observed (circles) Stokes I profiles
from the Liège atlas. The line widths and depths are
reproduced reasonably well by the simulations (no macroneither microturbulent velocity was used for the synthesis).
This justifies that the convective velocity field and
temperature in the simulations are sufficiently realistic.
(6) The Stokes V profiles show positive asymmetry in upflows
and negative asymmetry in downflows (Fig. 4).
(7) Under reduced spatial resolution, the irregular-shaped
Stokes V profiles appear in between the opposite polarity areas
(Fig. 5). The abundance of the irregular profiles in the
simulations increases with decreasing the magnetic flux from
13-15 % for <B> = 140 G up to 35% for <B> = 30 G.
asymmetries of the Stokes V profiles of the Fe I 6302 line with
the resolution ~1". The mean values together with standard
deviations corresponding to each curve are indicated in the
figure. Red line: <B> = 140 G; green line: <B> = 80 G; blue line:
<B> = 30 G.
The observations of the quiet Sun (mainly network fields) at
6300 A give positive values of da = 15 % and dA = 6% (Sigwarth
1999). The range of the area asymmetries in the considered MHD
model agrees with the observational constrains, while are
amplitude asymmetry is larger. Note, that the amplitude
asymmetry increases by 30 % with the flux decreasing from 140
to 30 G. The area asymmetry behaves in the opposite way and
decreases almost twice over the same flux interval.
PRINCIPAL COMPONENT ANALYSIS
STOKES V AMPLITUDES
Fig. 4. Left panel: map of the classes of Stokes V profiles of
Fig. 2.Histograms of the LOG of the amplitudes of the Fe I
6302 and 15648 A Stokes V profiles for <B> = 30 G. Black
line: original spatial resolution of 20 km. Blue line: spatial
-4
resolution of ~1 arcsec and noise level of 2·10 Ic. Red line:
observations with the Tenerife Infrared Polarimeter
(Khomenko et al, 2003). Only profiles with the amplitudes
-3
exceeding 10 were used for the histograms in the last two
cases.
The spatial smearing causes the average polarization signal to
become 3 times lower. Among the considered snapshots, the
amplitudes in the <B> = 30 G case are the closest the those
observed by TIP. Still, the polarization in the simulations is
slightly larger. In observations, about 60% of the total area is
-3
occupied by the signals above 10 , while the simulations give a
larger fraction of 80%. Note that the Stokes V amplitudes of
the infrared and the visible line in the <B>=30 G case are similar.
Fig. 5. PCA classification of the Fe I 15648 Stokes V profiles
with the resolution ~1 arcsec for <B> = 30 G case. The format
of the figure is the same as Fig. 4.
All the shown classes of the profiles can be found in observations
of the quiet Sun (Khomenko et al. 2003, Sanchez Almeida & Lites
2000). The mixed-polarity profiles make about 30% of all the
classified profiles. They are located in between the opposite
polarity areas. The average size of patches formed by the profiles
of the same class has increased to 600-700 km, which is close to
the spatial resolution.
the FeI 15648 A line for <B> = 30 G (original resolution). Right
panel: examples of the class profiles. Green colors mark the
locations of the blue-shifted profiles having mostly positive
asymmetry (classes d and e). Blue and white colors mark the
red-shifted profiles having negative or no asymmetry (class a
and b). The yellow and red colors mark the location of the
irregular-shaped profiles (classes c and f).
The distribution of the class profiles shows a clear correlation
with the granulation structure. Note, that the profiles in granular
regions show positive asymmetry, while the profiles in
intergranules often show negative asymmetry. These behavior
was indeed observed in the quiet Sun inter-network by SocasNavarro et al. (2004). The irregular-shaped Stokes V profiles are
located in the canopy regions at granular-intergranular borders.
They make 20% of all the classified profiles.
REFERENCES
E. Khomenko, M. Collados, S. K. Solanki, A. Lagg, J. Trujillo Bueno, 2003, A&A, 408, 1115
D. Rees, A. Lopez Ariste, J. Thatcher, M. Semel 2000, A&A, 355, 759
J. Sanchez Almeida and B. Lites, 2000, ApJ, 532, 1215
M. Sigwarth, K. S. Balasubramaniam, M. Knolker, W. Schmidt, 1999, A&A, 349, 941
H. Socas-Navarro, V. Martinez Pillet, B. Lites, 2004, ApJ
A. Vogler, 2003, Three-dimentional simulations of magneto-convection in the solar photosphere,PhD thesis, Univ. Gottingen