((LOOK at the file Abstract_March18.doc))
((INCLUDE THE ABSTRACT that is in Abstarct_March18.doc))
A Coronal Mass Ejection (CME) is now understood as an eruption of magnetized plasma, usually in the form of a flux rope, from the sun with energy up to 10
32
ergs.
Generally a CME has a speed higher than the Alfven speed in the background solar wind, and a shock is driven in front of it (Hudson et al. 2006). The CME is of great importance to space weather for at least two reasons, 1. Some CMEs come across Earth and have great impact on Earth’s magnetosphere (see review, Webb and
Gopalswamy, 2006 ), 2. the
CME-driven shocks are able to generate very-high-energy (GeV) particles which are hazardous to life and instruments on board spacecrafts in the out space.
Remote sensing on CME near the sun demonstrated CMEs have a bright front and a dark cavity (Hundhausen 1987). The bright front is now known as the plasma of higher density and the dark cavity corresponds to the ejected flux rope. In situ observations on
ICME, the corresponding part of CME at 1AU, revealed detailed features on the transition from the sheath to the ICMEs (see review Zurbuchen and Richardson, 2006).
The magnetic field lines drape around ICMEs in the sheath; the transition from the

interplanetary magnetic field to be aligned with the magnetic could usually end up with planar magnetic structures (PMSs) in the sheath region downstream of a quasiperpendicular shock (Farrugia, et al, 1990, Jones et al, 2002). However, these detailed features for magnetic field in the CMEs near the sun can not be investigated with techniques currently available.
In this paper, we investigate, for the first time, the transition from the sheath to the flux rope for a simulated CME at lower solar corona in this letter. The CME is simulated using Space Weather Modeling Framework (SWMF) developed by University of
Michigan (Toth, et al. 2005). For the features characterizing the transition from the sheath to the flux rope, our results show very good agreements with the ICME observations except for the temperature. We also investigate the rotation of magnetic field in the sheath region. The paper is arranged in the following. Section 2, we present our simulation. Section 3, we discuss the transition features like density change. Section 4, we discuss the transition of magnetic field. Conclusions and discussions are presented in
Section 5.
2. Simulation
The space weather modeling framerowork is based on components, each component represents one aspect in the space weather (Toth, et al. 2005). For example, component SC stands for solar corona extending from the sun until 24 Rs (Rs is solar radii) from the sun, component IH stands for inner heliosphere, extending until 250 Rs.
Each component can be coupled with the other, and have capability of adaptive meshes

which allow the user to refine the region dominated by physical processes of small spatial scales, eg, regions around a shock or current sheet.
Since we are interested only in the lower corona, only component SC is used in our simulation. First we simulate a steady solar wind using a semi-empirical model
(Roussev, et al, 2003, Cohen et al, 2007). The solar wind speed at 1AU determined using
Wang-Sheeley-Arge model and the sun’s magnetic field is based on the MDI data during
Carrington Rotation 1922 (Arge and Pizzo, 2000; Arge et al, 2002, Arge et al, 2004).
This Carrington Rotation is chosen because it is a solar minimum. Next, the speed is used to determine the polytropic index which is incorporated into the self-consistent MHD equations to parameterize the process of heating in the solar wind. After the NHD equations and the boundary condition on the sun is set up, the code iterate to 13000 steps to reach a steady state.
After we generate a steady state for the solar corona, a modified TD flux rope is placed near an active region near the equator. To reveal the shock, the grids near the center of shock, which is defined on the radial line going through (X=1.08Rs, Y=0.27Rs,
Z=0.11Rs), are refined to a size of 0.03Rs. After initial adaptation, a CME propagate at a speed ~400 km/s and accelerate to 600km/s in about 50 minutes. The steady state solar wind, the initiation of the CME, and the propagation and acceleration of the CME, the evolution of the CME-driven shock and the post shock compression have been presented in Liu et al, 2008.
3. The transition region

The transition from the sheath to the ejecta has been investigated extensively using in situ observations on ICME, the interplanetary manifestations of CMEs (e.g., see review Zurbuchen and Richardson, 2006)). The observed features for the transition include density change, magnetic field rotation (Klein and Burlaga, 1982) and proton temperature decrease (Gosling et al, 1973; Richardson and Cane, 1995). This observed features are defined at 1AU. In this paper we look at the characteristics near the Sun. To quantify these features, we sample the data on a radial line going through the middle of the flux rope (X=1.08 Rs, Y=0.27 Rs, Z=0.11 Rs). These coordinate is under the HGI coordinate system which is the coordinates we use for the simulation. At t=20, 30 and 40 minutes, the density, the temperature and the magnetic field angle vs. position are plotted in Figure 1. The two dashed lines mark the transition from sheath to the flux rope.
Between these two dashed lines, the density decreases to about one fourth of the maximum density sheath, the zenith angle of magnetic field increases over 60 degrees.
These features are consistent with the observations. The temperature changes in the region between the two dashed lines as well; however, the temperature is higher in the flux rope than the sheath to as opposite to the observations. The temperatures were obtained from the pressure and mass density in the simulated MHD data by assuming that the plasma is comprised of protons only. Probably because we modify the polytropic index

so that the temperature inside the flux rope was higher than it should actually be.
Another feature in the observed ICMEs is that the magnetic field lines drape around the ICME (Gosling and McComas 1987; McComas et al, 1989). Figure 2 shows the flux rope, several magnetic field lines and part of the sun’s surface at t=20 minutes after the CME initiation. The light green surface in Figure 2 is an isosurface of density

10
-17 g/cm
-3
, which represent the flux rope. The red lines are the magnetic field lines; the pink line goes through the center of the flux rope is the line we sample the data. The colored spherical surface is part of the sun; the colors on the surface represent the photosphere magnetic field; the blue and red area is the active region where we insert the flux rope to initiate a CME. All these magnetic field lines in our simulation are draping around the flux rope just as those observed magnetic field line in front of an ICME
4. The rotation of the magnetic field
There is a rotational discontinuity in sheath transition region where the direction of the magnetic field lines drape around the flux rope. Observation on ICMEs have found planar magnetic structures (PMS), characterized by ordering of magnetic fields into laminar sheets in the sheath (Nakagawa et al, 1989). These PMSs were suggested to form due to the processes of magnetic field draping around the magnetic cloud (Farrugia et al,
1990, Neugebauer et al, 1993). A technique was developed to estimate the relative position of a spacecraft to an ICME (Jones, et al, 2002). Recent studies by Kataoka et al
2005 revealed some criterions for the generation of PMSs. These investigations are done to ICMEs near Earth provide some clues to study the magnetic configuration for CMEs near the sun.
To investigate how the magnetic fields change their configuration to evolve from the interplanetary magnetic field to aligned with the magnetic field in the flux rope, the magnetic field lines in the sheath at t=30 minutes are plotted in Figure 3. The yellow surface (density isosurface of 5

10
-18
g/cm
-3
) represents the flux rope. The dark brown magnetic field lines are ahead of the shock and the light green lines are in the behind the

shock. Both these two groups of magnetic field lines are in coplanarity plane (the plane determined by the shock normal and the magnetic field at the shock). The lighter blue lines, which are in the sheath and closer to the flux rope, rotate off the coplanarity plane as they get approach the flux rope. However, although these field lines are rotating, they seem to stay in a plane parallel to the shock plane.
To quantify the field lines rotation from the coplanarity plane, we define coordinate system (x
 y
 z

). The z

axis is along shock normal; the x

axis is along the projection of the upstream magnetic field on the shock plane. The x

-z

plane, therefore, is the coplanarity plane; y

axis is chosen to complete the coordinate. The configuration of the coordinates x
 y
 z

, the zenith angle

, and the azimuth angle

are shown in
Figure 4. The zenith angle

measure the rotation inside the coplanarity plane and the azimuth angle

measure the rotation off it. In the sheath region, the angle

and

in the sheath region are plotted for the CME at t=20, 30 and 40 minutes in Figure 5. We define two layers; Layer 1 close to the shock and Layer 2 close to the sheath. The dashed line in the plots marked Layer 1 in the sheath, the layer close to the flux rope. Another layer, we name as Layer 2 is and the layer close to the flux rope. In the Layer 1, the azimuth angle

is zero and the zenith angle

has small changes and just as we pointed out in the previous section, the magnetic field lines stay in the coplanarity plane. In the
Layer 2, the azimuth angle

increase from zero to an angle larger than 80

, however, the zenith angle

is very close to 90

in this layer, which demonstrates that the rotation is in a plane very close to the shock plane.
Figure 6 shows the evolution of the width of the two layers normalized by the width of the sheath as a function of time after the CME initiation. From at t=20 to 24

minutes, the Layer 2 occupies 90% of the sheath; at t> 30 minutes, each Layer 1 and 2 has about 50 % of the sheath
5. Discussions and Conclusions
We present a CME simulation done with a 3D MHD AMR code (SWMF) and investigate the magnetic structure of the sheath between the flux rope and the CMEdriven shock. Our simulation, focusing on the transition of the magnetic field lines, are consistent with the observed ICMEs for the features that the density decreases, the magnetic field rotation and magnetic field draping around the magnetic cloud. Then we investigate the transition of the magnetic field from the shocked solar wind to be aligned with the flux rope. Here we draw the following conclusions.
1. For the CME near the sun, the magnetic field lines drape around the flux rope, similar to observed ICME which has magnetic field lines drape around it;
2. The sheath between the shock and the flux rope can be divided into two layers, one layer on the shock side and the other on the side of the side of the flux rope. For the magnetic field to transit from the interplanetary magnetic field to the fields draping around the flux rope, the evolution in the two layers of are different. In the shock side, the magnetic field lines stay in the coplanarity layer as if they are not affected by the draping field line. In the sheath side, the field lines rotate off the coplanarity plane to be the shock
Jones et al. 2002 proposed a sketch for the field lines evolution, we furthered his theories into 3D and demonstrate that the field rotation are in two stages and each stages the rotation happens in one plane.

3. The relative width of the two layers to the width of the sheath has also been calculated. The layer one dominates the sheath at the t < 20 minutes, and later after 30 minutes, the sheath is almost evenly divided into Layer 1 and Layer 2.
Although the density and the magnetic field lines we simulated match the observations very well, the temperature change does not. It might be that we modified the polytropic index to parameter the heating in the lower solar corona. However this modification may not change our conclusions since the base of our conclusion is the configuration of the magnetic field lines. The field lines drape around the flux rope as the observed field lines for the ICMEs. A further study on the field configurations using a more realistic heating model will be required in the future.
There are some questions unanswered on the structure evolution of the sheath like what controls the normalized width of the layers and why the widths change with time.
We estimate that the flow very close to the shock stay in the radial direction, and start to go around the flux rope further away from the shock. The going-around flows drag the field lines off the coplanarity plane and form Layer 1 while the field lines stay in the coplanarity plane in Layer 1 ((2?!). Further investigations are required to find more convincing answers for these questions.
The two-layer structure of the sheath behind the CME-driven shock in the solar corona was not detected in the ICME observations at 1AU. The corresponding part of the sheath in front of an ICME at 1AU is very narrow comparing with the width of the sheath. We hypothesis that this structure may be developed only at the center of shock where the flow to the flank is nearly perpendicular to the shock normal; therefore, the flow in the fluid drag the frozen in magnetic field off the coplanarity plane and form the

structure. Most observations on ICMEs did on the flank and don’t have such a structure.
Another possibility is that the shock needs to be strong enough to have a wide Layer 2.
The shock driven by ICMEs. Further investigation with STEREO data might shed some light on magnetic structure in the sheath in front of an ICME.
Acknowledgement
THANKS LWS grant, NSF CAREER and JPL cosmos machines
References
Write explicitely what the acronames are GRL, ApJ etc. Write explicitelly all the names in the references. If there are more than 5 names write at al.
Cohen, O. et al, 2007, ApJ, 654, L163
Farrugia et al., 1990, GRL, 17 (8) 1025
Gosling, J. T.. Pizzo, V. and Bames, S. J., 1973, J. Geophys. Res. 78, 2001
Gosling, J. T and McComas, D. J. 1987, Geophys. Res. Lett. 14, 355
Hundhausen, A. J. 1987 , Sixth International Solar Wind Conference, p181
Hudson, H. S., J.-L. Bougeret and J. Burkepile. 2006, Space Science Review, 123,
13.
Jones G. H. et al., 2002, Geophys. Res. Lett., 29 (11) 1520
Kataoka, R. et al, 2005, Geophys. Res. Lett.,32, L12103
Klein, L. W. and Burlaga, L. F. 1982, J. Geophys. Res. 87, 613

Liu, Y. C.-M., et al, 2008, Apj, in press
McComas, D. J. et al., 1989, J. Geophys. Res. 94, 1465
Nakagawa T. A. et al., , J. Geophys. Res. 94, 11761
Neugebauer, M., 1993, , J. Geophys. Res. 98, 9383
Richardson I. G. and Cane, H. V. 1995, J. Geophys. Res. 100, 23397
Roussev, I. I., et al. 2003, ApJ, 595, L57
Toth, et al., 2005, J. Geophys. Res. 110, 12226
Webb D. F., and N. Gopalswamy, 2006, "Solar Influence on the Heliosphere and
Earth's Environment: Recent Progress and Prospects", ed. N. Gopalswamy and A.
Battacharyya, 71
Zurbuchen, T. H. and I. G. Richardson, 2006, Space Science Review, 123, 31