Astron. Astrophys. 348, L37-L40 (1999)

3. Standard siphon flow models

In traditional siphon flow models, such as that of Montesinos & Thomas (1997), the magnetic field strength of the inner footpoint is slightly lower than that of the outer footpoint at equal gravitational potential. However, all observations, with the exception of the 2.2 µm Ti lines, show that B decreases steadily outwards in a sunspot, which is incompatible with the requirements of the standard siphon flow model if both legs of the loop lie within the penumbra (as is suggested by the observations of Westendorp Plaza et al. 1997). It is usually argued that the different heights of formation of the spectral line at both footpoints may lead to a match with the observations. For this explanation to work the background field strength at a given geometrical height would have to be roughly constant in the penumbra (as is indeed assumed by Montesinos & Thomas 1997). This can be tested by considering, e.g., the model of Jahn (1989) or of Jahn & Schmidt (1994). These models reproduce a number of observations, including the radial dependence of B and , as well as the distribution of the continuum intensity, which is very sensitive to the magnetic field structure (Pizzo 1986). In these models the magnetic field strength at a given geometrical height definitely decreases outward. This suggests that the problem associated with siphon flows is not solved just by invoking a different height of formation of spectral lines in different parts of the penumbra.

The 2.2 µm Ti observations reveal a nearly constant field strength along the cool horizontal component of the field harbouring (at least a part of) the Evershed flow. This is consistent with siphon-flow models, since our observations are not accurate enough to detect the small horizontal gradients of the field strength required by such models. Hence the Ti lines resolve the problem posed by the observations of Westendorp Plaza et al. (1997). The height-of-formation explanation, whose adequacy has never been demonstrated for a realistic background penumbral field, need no longer be invoked in the light of our observations.

These may also help to solve another problem faced by siphon flow models. Degenhardt (1991) mentions that his model photospheric loops are normally too short to reproduce penumbral filaments. However he points out that longer filaments are obtained when reducing the magnetic field strength of the arch to approximately 400 G in the upstream penumbral footpoint (for a flux-tube embedded in a field-free atmosphere). Such a low field strength is plausible in view of the weak horizontal magnetic component we observe. According to Thomas & Montesinos (1993) the addition of a surrounding magnetic field also produces an increase of the length of the filaments.

Our observations lead us to propose that siphon-flow calculations be carried out along loops with field strengths in both footpoints of around 500 G. They should be embedded in a background magnetic field which corresponds roughly to the model of Jahn (1989).

© European Southern Observatory (ESO) 1999

Online publication: July 26, 1999
helpdesk.link@springer.de