Radiative heating and the buoyant rise of magnetic flux tubes in the solar interior

We study the effect of radiative heating on the evolution of thin magnetic flux tubes in the solar interior and on the eruption of magnetic flux loops to the surface. Magnetic flux tubes experience radiative heating because (1) the mean temperature gradient in the lower convection zone and the overshoot region deviates substantially from that of radiative equilibrium, and hence there is a non-zero divergence of radiative heat flux; and (2) the magnetic pressure of the flux tube causes a small change of the thermodynamic properties within the tube relative to the surrounding field-free fluid, resulting in an additional divergence of radiative heat flux. Our calculations show that the former constitutes the dominant source of radiative heating experienced by the flux tube.In the overshoot region, the radiative heating is found to cause a quasi-static rising of the toroidal flux tubes with an upward drift velocity 10^-3|| cm s^-1, where _e – _ad < 0 describes the subadiabaticity in the overshoot layer. The upward drift velocity does not depend sensitively on the field strength of the flux tubes. Thus in order to store toroidal flux tubes in the overshoot region for a period comparable to the length of the solar cycle, the magnitude of the subadiabaticity (< 0) in the overshoot region must be as large as 3 × 10^–4. We discuss the possibilities for increasing the magnitude of and for reducing the rate of radiative heating of the flux tubes in the overshoot region.Using numerical simulations we study the formation of -shaped emerging loops from toroidal flux tubes in the overshoot region as a result of radiative heating. The initial toroidal tube is assumed to be non-uniform in its thermodynamic properties along the tube and lies at varying depths beneath the base of the convection zone. The tube is initially in a state of neutral buoyancy with the internal density of the tube plasma equal to the local external density. We find from our numerical simulations that such a toroidal tube rises quasi-statically due to radiative heating. The top portion of the nonuniform tube first enters the convection zone and may be brought to an unstable configuration which eventually leads to the eruption of an anchored flux loop to the surface. Assuming reasonable initial parameters, our numerical calculations yield fairly short rise times (2–4 months) for the development of the emerging flux loops. This suggests that radiative heating is an effective way of causing the eruption of magnetic flux loops, leading to the formation of active regions at the surface.The National Solar Observatory is one of the National Optical Astronomy Observatories by the Association of Universities for Research in Astronomy, Inc., under cooperative agreement with the National Science Foundation.     

Polar coronal plumes

Polar coronal plumes are modeled using concentrations of magnetic flux at 1.01R , and assuming the field is current-free, or a potential field. Identifying the density enhancement of plumes with magnetic flux concentration produces good agreement between 1.01R and 1.10R , for model conditions of a large background magnetic field and a plume separation of 50 000 to 70 000 km at the base. Beyond 1.10R , both plumes and the potential field diverge very nearly as r ².Also Department of Astrogeophysics, University of Colorado, Boulder, Colo. 80309, U.S.A. Presently visiting Stanford University Institute for Plasma Research, Via Crespi, Stanford, Calif. 94303, U.S.A.     

On a Babcock-Leighton dynamo model with a deep-seated generating layer for the toroidal magnetic field, II

In a previous paper (Paper I), we studied a dynamo model of the Babcock-Leighton type (i.e., the surface eruptions of toroidal magnetic field are the source for the poloidal field) that included a thin, deep seated, generating layer (GL) for the toroidal field, B. Meridional motions (of the order of 12 m s^–1 at the surface), rising at the equator and sinking at the poles were essential for the dynamo action. The induction equation was solved by approximating the latitudinal dependence of the fields by Legendre polynomials. No solutions were found with _p = _f where _p and _f are the fluxes for the preceding and following spot, respectively. The solutions presented in Paper I, had _p = –0.5 _f , were oscillatory in time, and large radial fields, B, were present at the surface.Here, we resume the study of Paper I with a different numerical approach allowing for a much higher resolution in , the polar angle. The time dependent partial differential equations for the toroidal and poloidal field are solved with the help of a second order, time and space centered, finite difference scheme. Oscillatory solutions with _p = _f are found for various values of the meridional motions and diffusivity coefficients. The surface values of B, while considerably smaller than those of Paper I, are still unacceptably large, specially at the poles. The reason can be traced to the eruption of toroidal field at high latitudes. It appears that in order to obtain small values for the radial field in the polar regions, high latitude sources ( smaller than /4, say), must reach their maximum below the surface. Weaker meridional motions near the poles than in the equatorial region are also suggested.     

On The Torsional Oscillations In Babcock–Leighton Solar Dynamo Models

The torsional oscillations at the solar surface have been interpreted by Schüssler and Yoshimura as being generated by the Lorentz force associated with the solar dynamo. It has been shown recently that they are also present in the upper half of the solar convection zone (SCZ). With the help of a solar dynamo model of the Babcock–Leighton type studied earlier, the longitudinal component of the Lorentz force, L , is calculated, and its sign or isocontours, are plotted vs. time, t, and polar angle, (the horizontal and vertical axis respectively). Two cases are considered, (1) differential rotation differs from zero only in the tachocline, (2) differential rotation as in (1) in the tachocline, and purely latitudinal and independent of depth in the bulk of the SCZ. In the first case the sign of L is roughly independent of latitude (corresponding to vertical bands in the t, plot), whereas in the second case the bands show a pole–equator slope of the correct sign. The pattern of the bands still differs, however, considerably from that of the helioseismic observations, and the values of the Lorentz force are too small at low latitudes. It is all but certain that the toroidal field that lies at the origin of the large bipolar magnetic regions observed at the surface, must be generated in the tachocline by differential rotation; the regeneration of the corresponding poloidal field, B _p has not yet been fully clarified. B _p could be regenerated, for example, at the surface (as in Babcock–Leighton models), or slightly above the tachocline, (as in interface dynamos). In the framework of the Babcock-Leighton models, the following scenario is suggested: the dynamo processes that give rise to the large bipolar magnetic regions are only part of the cyclic solar dynamo (to distinguish it from the turbulent dynamo). The toroidal field generated locally by differential rotation must contribute significantly to the torsional oscillations patterns. As this field becomes buoyant, it should give rise, at the surface, to the smaller bipolar magnetic regions as, e.g., to the ephemeral bipolar magnetic regions. These have a weak non-random orientation of magnetic axis, and must therefore also contribute to the source term for the poloidal field. Not only the ephemeral bipolar regions could be generated in the bulk of the SCZ, but many of the smaller bipolar regions as well (at depths that increase with their flux), all contributing to the source term for the poloidal field. In contrast to the butterfly diagram that provides only a very weak test of dynamo theories, the pattern of torsional oscillations has the potential of critically discriminating between different dynamo models.     

Observational constraints on a ‘hidden’, turbulent magnetic field of the Sun

A search for linear polarization due to the transverse Zeeman effect in quiet regions near the heliographic north pole has been carried out. The aim is to determine new constraints on the properties of the hidden or turbulent magnetic flux of the Sun. As more than 90% of the total flux seen in magnetograms has its source in kG fluxtubes with an average filling factor of less than 1%, the term hidden magnetic flux refers to the field in the remaining 99% of the photospheric volume, which remains undetected in ordinary magnetograms (at available levels of spatial resolution and sensitivity).Simultaneous recordings of the Stokes I, Q, and V profiles of the Fei 5250.22 and 5247.06 Å lines with 5 × 5 sec of arc spatial resolution have been made with the NSO McMath solar telescope. The analysis shows how the observed Stokes Q amplitudes, as well as the Q/V ratio in combination with the 5250/5247 Stokes V line ratio, provide constraints on the field strength and the angular distribution of the field vectors of the hidden magnetic flux. The field has to be tangled with opposite polarities mixed on a subarcsec scale, and the field vectors have to have large inclinations with respect to the vertical direction, with an angular distribution not far from being isotropic in the photosphere. Constraints on the strength of this tangled or turbulent magnetic field have been obtained by previous methods, which are reconsidered in view of their dependence on the assumed angular distribution. An upper limit of 100 G comes from determinations of magnetic line broadening, a lower limit of 10 G from observed Hanle-effect depolarization.In our observations the linear polarization has been recorded with a precision of 10^-4 with good spectral resolution. Further improvements are impeded by the lack of telescopes with large photon collecting areas and small instrumental polarization.     

The magnetic field in a protostellar cloud

General forms of theB-p relation are investigated in both the isothermal and the non-isothermal regions. The magnetic flux dissipation either by ambipolar diffusion or by Ohmic dissipation has been studied. The rates of heating due to the magnetic dissipation processes have been calculated in comparison with the rate of compressional heating.The magnetic field strength is derived as a function of flux/mass ratio, mass, density, and geometry of the isothermal cloud. In the non-isothermal region, the temperature is added to the above-mentioned variables.It has been found that the magnetic flux starts to dissipate via ambipolar diffusion at neutral density ofn>3×10⁹ cm^–3. Ambipolar diffusion continues effective until reaching densities ofn>10¹¹ cm^–3, where Ohmic dissipation dominates. Under some conditions, the electrons evaporate from the grain surface atn>10¹³ cm^–3, while the ions are still adsorbed on the grain surfce. In this case, the magnetic flux loss returns to be influenced by ambipolar diffusion.The rates of heating by both Ohmic dissipation _OD and ambipolar diffusion _AD are found to be smaller than the rate of compressional heating _C in case of magnetic dissipation. Assuming that the magnetic field is frozen in the medium, then _C is smaller than both _OD and _AD . The above results of heating were found in the non-isothermal region.     

Sunspot geometry and pressure balance

Measurements on magnetic canopies extending from sunspots show that, at the outer penumbral edge, heights of the bases are independent of sunspot diameter and average 180 km. This places a lower limit on the outer penumbral base; with an assumed thickness of 250 km, the top is 430 km above z = 0 ( _c = 1) in the photosphere.Chistyakov's (1962) observations require the penumbral surface to be convex in radial section. The Wilson depression, able thus to be found only from limb-side penumbras, is 1360 km from his selected measurements. Averaged over all regular sunspots without special selection, this drops to 1040 km. Thus ^* = 1 in umbras lies around z = -610 km.Magnetic field-strength measurements relate probably to ^* 0.02, some 160 km higher, where z -450 km. The magnetic pressure of the typical 3250 G sunspot field would support the external-axial gas-pressure difference at z = -330 km, the difference of 120 km lying well within the uncertainties. Tension forces, commonly invoked to achieve pressure balance, do not exceed the uncertainties of measurement.Beyond the sunspot, the base of the sunspot field rises only slowly over at least 16 000 km horizontally, whereas Beckers (1963) found the inclination of H superpenumbral fibrils to be some 13°. These results are nicely compatible since the field angle is typically of this magnitude at the minimum heights where H fibrils will be observed, say 1400 km.Operated by the Association of Universities for Research in Astronomy, Inc., under contract with the National Science Foundation.     

Implications of a strongly peaked polar magnetic field

Using the flux-transport equation in the absence of sources, we study the relation between a highly peaked polar magnetic field and the poleward meridional flow that concentrates it. If the maximum flow speed _m greatly exceeds the effective diffusion speed /R, then the field has a quasi-equilibrium configuration in which the poleward convection of flux via meridional flow approximately balances the equatorward spreading via supergranular diffusion. In this case, the flow speed () and the magnetic field B() are related by the steady-state approximation () (/R)B()/B() over a wide range of colatitudes from the poles to midlatitudes. In particular, a general flow profile of the form sin ^p cos ^q which peaks near the equator (q p) will correspond to a cos ⁿ magnetic field at high latitudes only if p = 1 and _m = n /R. Recent measurements of n 8 and 600 km² s^–1 would then give _m 7 m s^–1.     

Oscillations of a polytrope with a toroidal magnetic field

The radial and the non-radial (l=2) modes of oscillation of a gaseous polytrope with a toroidal magnetic field are examined using a variational principle. It is found that the frequencies of oscillation of the radial mode and the Kelvin mode (l=2) decrease due to the presence of the magnetic field. The shift in the frequency of the Kelvin mode may be split up into two parts, viz. the shift in frequency due to the magnetic field on the unperturbed sphere [(1²)m, say] and the shift in frequency due to the distortion of the structure by the magnetic field [(1²)s, say]. In the first order calculations using one parameter trial function, it is found that (1²)m is indeed positive but is overweighed by a negative (1²)s. In the second order calculations using a trial function with two variational parameters, we find that the general behaviour of (1²)m and (1²)s is unchanged but that (1²)m becomes negative for polytropic indicesn1.5.In Appendix I we study the effect of a small rotation and toroidal magnetic field on the structure of a polytrope. It is found that the resulting configuration is a prolate spheroid, a sphere or an oblate spheroid according as respectively. Here denotes the magnetic energy andT the kinetic energy due to rotation andq is a constant which depends on the polytropic indexn. The values ofq are given in Table I.     

On the relationship between polar faculae and large-scale magnetic field

We investigate a latitude–time distribution of polar faculae observed at Ussuriysk Observatory in years 1966–1986. The distribution is compared with the longitude-averaged (zonal) magnetic field of the Sun calculated from the data obtained at Mount Wilson Observatory in the years 1966–1976, and at Kitt Peak National Observatory during the period from 1976 to 1985. We found that slow, poleward-directed migration of the polar faculae zones occurring during the course of the solar cycle is not a continuous process, but it contains several episodes of appearance and fast poleward drift of new zones of polar faculae. At the rising phase of the solar cycle, new zones of polar faculae appear at latitudes as low as 40°, but the ones observed during the declining phase of the solar cycle originate at higher latitudes of 50–55°. Such episodes of appearance and fast migration of the polar faculae zones are associated with the poleward-directed streams of magnetic field originated at low latitudes. Moreover, we found some evidence for existence of an additional component of the polar faculae activity that reveals an equatorward migration during the course of the solar cycle. We also investigated a relationship between the number of polar faculae, n, and absolute magnetic flux _z of the zonal mode of the solar magnetic field. We found that within the polar zones of the Sun, substantial correlation between temporal variations of n and _z takes place both on the time scale of the solar cycle and on a shorter time scale of 2–4 years. The relationship between the number of polar faculae and magnetic flux may be approximated by a linear dependence n=0.12_z (where _z is expressed in 10²¹ Mx), except for time interval 1977 through 1980 for which the factor of proportionality is found to have a systematically larger value of 0.20.     

The 3-D boundary element formulation of linear force-free magnetic fields with finite energy content in semi-infinite space

In this paper, a formulation describing a force-free magnetic field with constant, either 0 or = 0, with finite energy content in semi-infinite space,z > 0, above the Sun is proposed and solved uniquely by a numerical method: the boundary element method (BEM). The method is effective and convenient in extrapolating the magnetic field above the solar surface, , directly from the measured magnetogram. Meanwhile, no additional data treatment is needed. Based on the existence of such a field, the uniqueness of the solution in the present formulation is proved and some useful properties are obtained. The validity is also demonstrated by its application to the magnetic field computation in the chromosphere from observed magnetogram data. The practical feasibility is thus discussed and further confirmed.     

Magnetically distorted polytropes

The fundamental frequencies of the non-radial mode of oscillation belonging to the second harmonic (l=2) of magnetically distorted polytropic gas spheres are evaluated in the second approximation by a variational method. The magnetic field is assumed to have both the toroidal and the poloidal components. We find that the frequencies of oscillation are increased due to the presence of the magnetic field and that these depend only slightly on the value of , the ratio of the specific heats. We have also determined the value of <1+1/n for the mode of oscillation which exhibits convective instability. This value is lower than the one which is obtained in the absence of a magnetic field.     

Flare-like energy release by flux pile-up reconnection

Flux pile-up magnetic merging solutions are discussed using the simple robust arguments of traditional steady-state reconnection theory. These arguments determine a unique scaling for the field strength and thickness of the current layer, namely B _s^–1/3, l^2/3, which are consistent with a variety of plasma inflow conditions. Next we demonstrate that flux pile-up merging can also be understood in terms of exact magnetic annihilation solutions. Although simple annihilation models cannot provide unique reconnection scalings, we show that the previous current sheet scalings derive from an optimized solution in which the peak dynamic and magnetic pressures balance in the reconnection region. The build-up of magnetic field in the current sheet implicit in flux pile-up solutions naturally leads to the idea of saturation. Hydromagnetic pressure effects limit the magnetic field in the sheet, yielding an upper limit on the reconnection rate for such solutions. This rate is still far superior to the Sweet–Parker merging rate, which can be derived by seeking solutions that avoid all forms of saturation. Finally we compare time dependent numerical simulations of the coalescence instability with the optimized flux pile-up models. This comparison suggests that merging driven by the relatively slow approach of large flux systems may be favored in practice.     

Simulations of jet formation from a magnetized accretion disk

Axisymmetric magnetohydrodynamic (MHD) simulations have been made of the formation of jets from a Keplerian disk threaded by a magnetic field. The disk is treated as a boundary condition, where matter with high specific entropy is ejected with a Keplerian azimuthal speed and a poloidal speed less than the slow magnetosonic velocity, and where boundary conditions on the magnetic fields correspond to a highly conducting disk. Initially, the space above the disk, the corona, is filled with high specific entropy plasma in the thermal equilibrium in the gravitational field of the central object. The initial magnetic field is poloidal and is represented by the superposition of the fields of monopoles located below the plane of the disk.The rotation of the disk twists the initial poloidal magnetic field lines, and this twist propagates into the corona pushing matter into jet-like outflow in a cylindrical region. After the first switch-on wave, which originates during the first rotation period of the inner radius of the disk, the matter outflowing from the disk starts to flow and accelerate in thez-direction owing to both the magnetic and pressure gradient forces. The flow accelerates through the slow magnetosonic and Alfvén surfaces and at larger distances through the fast magnetosonic surface. The flow velocity of the jet is approximately parallel to thez-axis, with the collimation mainly a result of the pinching force of the toroidal magnetic field. The energy flux of the flow increases with increasing magnetic field strength on the disk. Some of the cases studied have been run for long times, 60 rotation periods of the inner radius of the disk, and show indications of approaching a stationary state.     

Latitude and Magnetic Flux Dependence of the Tilt Angle of Bipolar Regions

Magnetogram data of 517 bipolar active regions are analyzed to study latitude, magnetic flux, polarity separation dependence of tilt angle of the active regions with well-defined bipolar magnetic configurations. The data were obtained at Huairou Solar Observing Station in Beijing during 1988 to October 2001. By statistical analysis, it is found: (1) The tilt angle () is a function of the latitude (). Our observed result, sin=0.5 sin, is in good agreement with that obtained by Wang and Sheeley (1991). (2) The tilt angle is a function of the magnetic flux. The tilt angle increases (decreases) with flux increasing when the flux is smaller (larger) than 5×10²¹ Mx. (3) The tilt angle is a function of the magnetic polarity separation. The tilt angle increases (decreases) with the separation increasing when the separation is smaller (larger) than 8×10⁹ cm. (4) The magnetic flux ( in 10²⁰ Mx) is correlated to the magnetic polarity separation (d in Mm), following ₂₀d ^1.15. The result is close to the observed result of Wang and Sheeley (1989), ₂₀d ^1.3. (5) The tilt fluctuations are independent of the latitude, but depend slightly on the separation, which is similar to the result obtained by Fisher, Fan, and Howard (1995). (6) The distribution function of the ratio of net magnetic flux to total magnetic flux is almost centered around zero net flux. The imbalance of magnetic flux is lower than 10% for 47% of our samples; 31% of active regions are in imbalance of the magnetic flux between 10% and 20%.     

Estimates for the effective electrical conductivity of the core in the interior of Jupiter and Saturn

In the deep interior of the giant planets Jupiter and Saturn, ordinary hydrogen and helium are transformed into a conducting metallic liquid at extremely high pressure. It is likely that the giant planets' observed magnetic field is constantly generated in the metallic fluid core by magnetohydrodynamic processes, converting mechanic energy in the form of convection into magnetic energy. The maximum strength of their magnetic fields is likely to be limited by magnetic field instabilities which convert the magnetic energy back into convection. The parameter which governs the occurrence of magnetic instabilities is the Elsasser number, = B ²/2, where B is the field strength, is the electrical conductivity, is the rotation rate and is the density. Since magnetic instability will be very active when exceeds a critical value _c 10 (the precise value depending on the magnetic field distribution), this imposes an upper bound on the effective electrical conductivity of the metallic fluid which comprises the bulk of Jupiter's interior and much of Saturn's.Stability calculations including both toroidal (model) and poloidal (observed) components of the magnetic field in a rapidly rotating spherical shell, have been performed. The most stable configuration of the field is when the poloidal component of field is strong and the toroidal field is weak; in this case we obtain an upper bound for electrical conductivity of 3 × 10⁶ S/m; while the most unstable configuration of the field is when the toroidal and poloidal fields are comparable, giving rise to _m 3 × 10⁵ S/m. The implications of the results for general dynamo theory are also discussed.     

Thermosolutal instability of compressible hall plasma in porous medium

The thermosolutal instability of a plasma in porous medium in the presence of a vertical magnetic field is considered to include the effects of compressibility and Hall currents. The effects of stable solute gradient and compressibility are found to be stabilizing and the Hall currents have a destabilizing effect. The system is stable for (C _p/g)<1;C _p, , andg denoting specific heat at constant pressure, uniform temperature gradient, and acceleration due to gravity, respectively. In contrast to the non-oscillatory modes in the absence of magnetic field and stable solute gradient, the presence of magnetic field (and, hence, Hall currents) and stable solute gradient introduce oscillatory modes for (C _p/g)>1. The case of overstability is also studied wherein the necessary conditions for the existence of overstability are obtained.     

Stokes parameters for thomson scattering in a cold magnetized plasma

The effect of a strong magnetic field on neutron stars or white dwarfs is calculated for Thomson scattering in a fully ionized collisionless plasma. The Stokes parameters for the scattered radiation are computed explicitly in terms of the state of polarization of the incident wave, the electron-cyclotron frequency, the plasma frequency, the angle of incidence, and the angle of scattering. The effects of the plasma are very insensitive to specific values of ( = ₂ ^p /²,_p denotes the electron plasma frequency) so long as 1, whereas the criterion for the magnetic field to substantially affect the Stokes parameters is that the photon frequency be less than the electron-cyclotron frequency. The effects of classical radiation damping and natural line broadening are briefly mentioned.     

The Energy Lost by Differential Rotation in the Generation of the Solar Toroidal Magnetic Field

The integrals, I_i(t) = _GL u_i j × B _i dv over the volume GL are calculated in a dynamo model of the Babcock–Leighton type studied earlier. Here, GL is the generating layer for the solar toroidal magnetic field, located at the base of the solar convection zone (SCZ); i=r, , , stands for the radial, latitudinal, and azimuthal coordinates respectively; j = (4)^-1 × B, where B is the magnetic field; u_r,u are the components of the meridional motion, and u is the differential rotation. During a ten-year cycle the energy _cycle I(t)dt needs to be supplied to the azimuthal flow in the GL to compensate for the energy losses due to the Lorentz force. The calculations proceed as follows: for every time step, the maximum value of |B| in the GL is computed. If this value exceeds B_cr (a prescribed field) then there is eruption of a flux tube that rises radially, and reaches the surface at a latitude corresponding to the maximum of |B| (the time of rise is neglected). This flux tube generates a bipolar magnetic region, which is replaced by its equivalent axisymmetric configuration, a magnetic ring doublet. The erupted flux can be multiplied by a factor F_t, i.e., by the number of eruptions per time step. The model is marginally stable and the ensemble of eruptions acts as the source for the poloidal field. The arbitrary parameters B_cr and F_t are determined by matching the flux of a typical solar active region, and of the total erupted flux in a cycle, respectively. If E(B) is the energy, in the GL, of the toroidal magnetic field B = B sin cos , B (constant), then the numerical calculations show that the energy that needs to be supplied to the differential rotation during a ten-year cycle is of the order of E(B_cr), which is considerably smaller than the kinetic energy of differential rotation in the GL. Assuming that these results can be extrapolated to larger values of B_cr, magnetic fields 10⁴ G, could be generated in the upper section of the tachocline that lies below the SCZ (designated by UT). The energy required to generate these 10⁴ G fields during a cycle is of the order of the kinetic energy in the UT.     

Active-Region Magnetic Structure Observed in the Photosphere and Chromosphere

The full magnetic vector has been measured in both the photosphere and chromosphere across sunspots and plage in NOAA Active Region 8299. We investigate the vertical magnetic structure above the umbral, penumbral and plage regions using quantitative statistical comparisons of the photospheric and chromospheric magnetic data. The results include: (1) a general decrease in average magnetic flux density with height; (2) the direct detection of the superpenumbral canopy in the chromosphere; (3) values for dB/dz which are consistent with earlier investigations when derived from a straight difference between the two measurements, but which are somewhat small when derived from the B=0 condition, (4) a monolithic structure in the umbrae which extends well into the upper chromosphere, with a very complex and varied structure in penumbrae and plage, as evidenced by (5) a uniform magnetic scale height in the umbrae with an abrupt jump to widely varying scale heights in penumbral and plage regions. Further, we find (6) evidence that field extrapolations using the photospheric flux as the boundary may not agree with expectations or with observed coronal structures as well as those which use the chromospheric magnetic flux as the extrapolation starting point.