Plage Regions Research Articles

Observation and modelling of main-sequence star chromospheres - XII. Two-component model chromospheres for five active dM1e stars

We aim to constrain the Hα, Ca ii H and Ca ii K profiles from quiescent and active regions on active dM1e stars. A preliminary analysis of all the data available for dM1e stars shows that the Hα/Ca ii equivalent width (EW) ratio varies by up to a factor of 7 for different stars in our sample. We find that spectroscopic binaries have a significantly smaller ratio than single dM1e stars. We also find that the pre-main-sequence stars Gl 616.2, GJ 1264 and Gl 803 have a ratio lower than main-sequence single dM1e stars. These differences imply that different chromospheric structures are present on different stars, notably the temperature minimum must decrease with an increasing Hα/Ca ii EW ratio. For these reasons, it is impossible to reproduce all observations with only one grid of model chromospheres. We show that the grid of model chromospheres of Paper VI is adequate to describe the physical conditions that prevail in the chromospheres of spectroscopic binaries and pre-main-sequence M1e stars, but not for the conditions in single dM1e stars. One or more additional grids of model chromospheres will be necessary to reproduce all observations. We use the method developed in Paper XI in this series, in order to build two-component model chromospheres for five M1e field stars: FF And A, FF And B, GJ 1264, AU Mic and Gl 815A. Our solutions provide an exact match of the Hα and the mean Ca ii H & K EWs within measurement uncertainties. We compare the theoretical profiles and the observed profiles of Hα and the Ca ii H & K resonance lines. On the one hand, our fits to the Ca ii lines are reasonably good. On the other hand, our models tend to produce Hα profiles with a central absorption that is too deep. This suggests that the column mass at the transition region for plages is underestimated, but this would imply that the contrast factor between quiescent and active regions in the Ca ii lines is larger than 5. We find that, except in the cases of FF And A and AU Mic, the total Hα profile is dominated by the contributions from plage regions. The Hα profiles in quiescent regions are typically filled in or slightly in emission. We also find that, as in the case of the less active dM1 stars, Ca ii emission in dM1e stars is generally dominated by the contribution from plage regions. We find plage filling factors typically in the range 10 to 60 per cent. These are on average (40 per cent) slightly larger than those derived for the less active dM1 stars (30 per cent). Our modelling shows that the grid of model chromospheres of Paper VI is adequate for only a subset of dM1e stars. Physical conditions that prevail in the chromospheres of dM1e stars are very different from one star to another. For this reason, this study demonstrates the critical need for several improved grids of model chromospheres as well as higher signal-to-noise ratio data.

Solar transition region above sunspots

We study the TR properties above sunspots and the surrounding plage regions, by analyzing several sunspot spectra obtained by SUMER in March 1999 and November 2006. We compare the SUMER spectra observed in the umbra, penumbra, plage, and sunspot plume regions. The Lyman line profiles averaged in each region are presented. For the sunspot observed in 2006, the electron densities, DEM, and filling factors of the TR plasma in the four regions are also investigated. The self-reversals of the Lyman line profiles are almost absent in umbral regions at different locations (heliocentric angle up to $49^\circ$) on the solar disk. In the sunspot plume, the Lyman lines are also not reversed, whilst the lower Lyman line profiles observed in the plage region are obviously reversed. The TR densities of the umbra and plume are similar and one order of magnitude lower than those of the plage and penumbra. The DEM curve of the sunspot plume exhibits a peak centered around $\log(T/\rm{K})\sim5.45$, which exceeds the DEM of other regions by one to two orders of magnitude at these temperatures. We also find that more than 100 lines, which are very weak or not observed anywhere else on the Sun, are well observed by SUMER in the sunspot, especially in the sunspot plume. We suggest that the TR above sunspots is higher and probably more extended, and that the opacity of the hydrogen lines is much smaller above sunspots, as compared to the TR above plage regions. Our result indicates that the enhanced TR emission of the sunspot plume is very likely to be the result of a large filling factor. The strongly enhanced emission at TR temperatures and the reduced continuum make many normally weak TR lines stick out clearly in the spectra of sunspot plumes.

COMPARISON AMONG Ca II K SPECTROHELIOGRAM TIME SERIES WITH AN APPLICATION TO SOLAR ACTIVITY STUDIES

Various observatories around the globe started regular full-disk imaging of the solar atmosphere in the Ca II K line in the early decades of the 20th century. The archives made by these observations have the potential of providing far more detailed information on solar magnetism than just the sunspot number and area records to which most studies of solar activity and irradiance changes are restricted. We evaluate the image quality and contents of three Ca II K spectroheliogram time series, specifically those obtained by the digitization of the Arcetri, Kodaikanal, and Mt Wilson photographic archives, in order to estimate their value for studies focusing on timescales longer than the solar cycle. We analyze the quality of these data and compare the results obtained with those achieved for similar present-day observations taken with the Meudon spectroheliograph and with the Rome-PSPT. We also investigate whether image-segmentation techniques, such as those developed for identification of plage regions on present-day Ca II K observations, can be used to process historic series. We show that historic data suffer from stronger geometrical distortions and photometric uncertainties than similar present-day observations. The latter uncertainties mostly originate from the photographic calibration of the original data and from stray-light effects. We also show that the image contents of the three analyzed series vary in time. These variations are probably due to instrument changes and aging of the spectrographs used, as well as changes of the observing programs. The segmentation technique tested in this study gives reasonably consistent results for the three analyzed series after application of a simple photographic calibration. Although the plage areas measured from the three analyzed series differ somewhat, the difference to previously published results is larger.

HELIOSEISMIC SIGNATURE OF CHROMOSPHERIC DOWNFLOWS IN ACOUSTIC TRAVEL-TIME MEASUREMENTS FROM HINODE

We report on a signature of chromospheric downflows in two emerging flux regions detected by time-distance helioseismology analysis. We use both chromospheric intensity oscillation data in the Ca II H line and photospheric Dopplergrams in the Fe I 557.6 nm line obtained by Hinode/SOT for our analyses. By cross-correlating the Ca II oscillation signals, we have detected a travel-time anomaly in the plage regions; outward travel times are shorter than inward travel times by 0.5-1 minute. However, such an anomaly is absent in the Fe I data. These results can be interpreted as evidence of downflows in the lower chromosphere. The downflow speed is estimated to be below 10 km s-1. This result demonstrates a new possibility of studying chromospheric flows by time-distance analysis.

Comparison of transient horizontal magnetic fields in a plage region and in the quiet Sun

<i>Aims. <i/>The properties of <i>transient horizontal magnetic fields<i/> (THMFs) in both plage and quiet Sun regions are obtained and compared.<i>Methods. <i/>Spectro-polarimetric observations with the Solar Optical Telescope (SOT) on the <i>Hinode<i/> satellite were carried out with a cadence of about 30 s for both plage and quiet regions located near the disk center. We selected THMFs that have net linear polarization (<i>LP<i/>) higher than 0.22%, and an area larger than or equal to 3 pixels, and compared their occurrence rates and distribution of magnetic field azimuth. We obtained probability density functions (PDFs) of magnetic field strength and inclination for both regions.<i>Results. <i/>The occurrence rate in the plage region is the same as for the quiet Sun. The vertical magnetic flux in the plage region is ~8 times more than in the quiet Sun. There is essentially no preferred orientation for the THMFs in either region; however, THMFs in the plage region with higher <i>LP<i/> have a preferred direction consistent with that of the plage-region's large-scale vertical field pattern. PDFs show that there is no difference in the distribution of field strength of horizontal fields between the quiet Sun and the plage regions when we avoid the persistent vertical flux concentrations for the plage region.<i>Conclusions. <i/>The similarity between the PDFs and the occurrence rates in plage and quiet regions suggests that a local dynamo process due to the granular motion may generate THMFs all over the Sun. The preferred orientation for higher <i>LP<i/> in the plage indicates that the THMFs may be somewhat influenced by the larger-scale magnetic field pattern of the plage.

On oscillations found in an active region with EIS on Hinode

Aims. Using temporal series (40 �� slot) data in the Fe xii 195 A line from the extreme-ultraviolet imaging spectrometer (EIS) on board the Hinode satellite, we seek to carry out a statistical study of oscillations in an active region. Methods. Using wavelet techniques, we measure the frequency and duration of oscillations found in the time series data. Stokes I and V data from Hinode’s SOT are used to measure photospheric magnetic fields, while an EIS raster is used to provide estimates of LOS velocity and electron density. The measured frequencies are displayed as maps in different bands to aid their analysis. Results. Oscillations over a broad range of frequencies (2–154 mHz) are found throughout the active region at the temperature of Fe xii (log T = 6.1 K). Oscillations with frequencies between 2–8 mHz are concentrated in bright plage areas, while oscillations of higher frequency preferentially group at the edges of these areas. Evidence is found for harmonics on the boundary of the active region loop, indicating the presence of standing waves. From a measurement of the lengths and electron density of loops appearing in our active region field-of-view (FOV), together with a knowledge of the periods (frequencies) of the oscillations within the area of these loops, we find a magnetic field value of between 12.2 ± 2.08−18.9 ± 3.40 G, assuming standing fast kink waves. Conclusions. We conclude that waves, which the observed oscillations are signatures of, are common in the active region studied, over a broad range of frequencies. The change in the distribution of the oscillations present at frequencies less than 8 mHz and those at all frequencies above this indicates a variation of the wave mode with frequency, e.g., possibly from slow magnetoacoustic waves in the bright plage regions to fast magnetoacoustic waves at the boundaries of these regions. The preference for higher frequency oscillations to cluster on the edges of regions of high magnetic field is considered indicative of resonant absorption. Based on our measurements of magnetic field, we conclude that the oscillations found by us in the active region loop are fast kink waves. Previous studies such as O’Shea et al. (2001 )a ndIreland et al. (1999) carried out statistical studies of the oscillatory characteristics of active regions using transition region and coronal lines observed with CDS on board the SOHO spacecraft. Here, we seek to carry out a similar statistical study but with the superior temporal resolution and imaging capabilities of the EUV imaging spectrometer (EIS) on board the Hinode spacecraft (Culhane et al. 2007). In this work we will present a global statistical analysis of all intensity oscillations, measured in a Fe xii 195 A line, present in an active region on the disk. From this statistical study we will seek to identify different wave modes present in the active region and its structures. We note that O’Shea et al. (2001) found evidence for non-compressive (transverse) waves to be more prevalent in coronal lines (at temperatures between Log T = 6.0−6.3 K) than in lower temperatures lines. We also, therefore, seek to find further evidence for these non-compressive waves. From the measured oscillation frequencies (periods), together with the observed physical quantities of loop length, radius and electron density, we will seek to provide estimates of the magnetic field in the corona. We note that Nakariakov & Ofman (2001), using a similar technique to that which will be used by us, obtained values of 13 ± 9 G for global standing kink waves in an active region.

Determination of Magnetic Diffusivity from High‐Resolution Solar Magnetograms

The magnetic diffusivity in the solar photosphere is determined by applying a new method to the magnetic induction equation. The magnetic field evolution is specified by a time sequence of high-resolution magnetograms of plage regions, taken by Hinode/SOT and SOHO/MDI. The mean value of magnetic diffusivity determined from SOT magnetograms with the smallest pixel size of 116 km is about -->0.87 ? 0.08 km2 s?1. This is the smallest value that has been empirically determined so far. High-resolution and full-disk MDI magnetograms with the pixel sizes of 440 and 1400 km yielded larger values of -->4.4 ? 0.4 and -->18 ? 7.4 km2 s?1, respectively. The measured diffusivity values at different length scales are consistent with a turbulent cascade that ends at a resistive dissipation scale of about 30 km. The results suggest that turbulent magnetic diffusivity should be taken into account in the analysis of the observed rate of magnetic flux cancellation in the photosphere.

Convection-driven Emergence of Small-Scale Magnetic Fields and their Role in Coronal Heating and Solar Wind Acceleration

Recent observations by the Solar Optical Telescope (SOT) on board Hinode have revealed that the surface of the Sun is, on average, covered with small-scale horizontal magnetic fields. Frequent emergence of horizontal magnetic flux on a granular scale is found in the quiet Sun and in plage regions. In this Letter we present the results of magnetohydrodynamic simulations that cover the upper convection zone and the corona. It is found that, even when the initial magnetic field is uniform and vertical, a disordered magnetic field is produced in the convection zone. The photospheric magnetic field is then characterized by strong vertical fields concentrated in the intergranular lanes and relatively weak, horizontal fields both in the granules and in the intergranular lanes. Occasionally, fragments with large magnetic fluxes are driven above the photosphere by the upward convective flows. These characteristics are consistent with the SOT observations. Moreover, the simulated flux emerging on a granular scale undergoes magnetic reconnection with the expanding vertical magnetic concentrations in the chromosphere. These reconnection events heat the local plasma and emit high-frequency waves that propagate into the corona. Such an interplay between the small-scale horizontal fields and the vertical flux may play an important role in coronal heating and wind acceleration in the Sun and stars.

Transient horizontal magnetic fields in solar plage regions

Aims.We report the discovery of isolated, small-scale emerging magnetic fields in a plage region with the Solar Optical Telescope aboard Hinode.

Discovery of a Temperature-Dependent Upflow in the Plage Region during a Gradual Phase of the X-Class Flare

Abstract We present Hinode/EIS raster scan observations of the plage region taken during the gradual phase of the GOES X3.2 flare that occurred on 2006 December 13. The plage region is located 200${}^{\prime\prime}$ east of the flare arcade. The plage region has a small transient coronal hole. The transient coronal hole is strongly affected by an X-class flare, and upflows are observed at its boundary. Multi-wavelength spectral observations allow us to determine velocities from the Doppler shifts at different temperatures. Strong upflows along with stationary plasma have been observed in the Fe XV line 284.2 Å (log$T/$ K $=6.3$) in the plage region. The strong upflows reach almost 150 km s$^{-1}$, which was estimated by a two-component Gaussian fitting. On the other hand, at a lower corona/transition region temperature (He II, 256.3 Å, log$T/$ K $=$ 4.9), very weak upflows, almost stationary, have been observed. We find that these upflow velocities clearly depend on the temperature with the hottest line, Fe XV, showing the fastest upflow velocity and the second-highest line, Fe XIV, showing the second-highest upflow velocity (130 km s$^{-1}$). All velocities are below the sound speed. The trend of the upflow dependence on temperature dramatically changes at 1 MK. These results suggest that heating may have an important role for strong upflow.

Flow Field Evolution of a Decaying Sunspot

We study the evolution of the flows and horizontal proper motions in and around a decaying follower sunspot based on time sequences of two-dimensional spectroscopic observations in the visible and white-light imaging data obtained over 6 days from 2005 June 7 to 12. During this time period the sunspot decayed gradually to a pore. The spectroscopic observations were obtained with the Fabry-Perot-based Visible-Light Imaging Magnetograph (VIM) in conjunction with the high-order adaptive optics (AO) system operated at the 65 cm vacuum reflector of the Big Bear Solar Observatory (BBSO). We apply local correlation tracking (LCT) to the speckle-reconstructed time sequences of white-light images around 600 nm to infer horizontal proper motions, while the Doppler shifts of the scanned Fe I line at 630.15 nm are used to calculate line-of-sight (LOS) velocities with subarcsecond resolution. We find that the dividing line between radial inward and outward proper motions in the inner and outer penumbra, respectively, survives the decay phase. In particular the moat flow is still detectable after the penumbra disappeared. Based on our observations, three major processes removed flux from the sunspot: (1) fragmentation of the umbra, (2) flux cancelation of moving magnetic features (MMFs; of the same polarity as the sunspot) that encounter the leading opposite polarity network and plages areas, and (3) flux transport by MMFs (of the same polarity as the sunspot) to the surrounding network and plage regions that have the same polarity as the sunspot.

Relationships between magnetic foot points and G-band bright structures

Aims. Magnetic elements are thought to be described by flux tube models, and are well reproduced by MHD simulations. However, these simulations are only partially constrained by observations. We observationally investigate the relationship between G-band bright points and magnetic structures to clarify conditions, which make magnetic structures bright in G-band. Methods. The G-band filtergrams together with magnetograms and dopplergrams were taken for a plage region covered by abnormal granules as well as ubiquitous G-band bright points, using the Swedish 1-m Solar Telescope (SST) under very good seeing conditions. Results. High magnetic flux density regions are not necessarily associated with G-band bright points. We refer to the observed extended areas with high magnetic flux density as magnetic islands to separate them from magnetic elements. We discover that Gband bright points tend to be located near the boundary of such magnetic islands. The concentration of G-band bright points decreases with inward distance from the boundary of the magnetic islands. Moreover, G-band bright points are preferentially located where magnetic flux density is higher, given the same distance from the boundary. There are some bright points located far inside the magnetic islands. Such bright points have higher minimum magnetic flux density at the larger inward distance from the boundary. Convective velocity is apparently reduced for such high magnetic flux density regions regardless of whether they are populated by G-band bright points or not. The magnetic islands are surrounded by downflows. Conclusions. These results suggest that high magnetic flux density, as well as efficient heat transport from the sides or beneath, are required to make magnetic elements bright in G-band.

High‐Resolution Observations and Modeling of Dynamic Fibrils

We present unprecedented high-resolution Hα observations, obtained with the Swedish 1 m Solar Telescope, that, for the first time, spatially and temporally resolve dynamic fibrils in active regions on the Sun. These jetlike features are similar to mottles or spicules in quiet Sun. We find that most of these fibrils follow almost perfect parabolic paths in their ascent and descent. We measure the properties of the parabolic paths taken by 257 fibrils and present an overview of the deceleration, maximum velocity, maximum length, and duration, as well as their widths and the thickness of a bright ring that often occurs above dynamic fibrils. We find that the observed deceleration of the projected path is typically only a fraction of solar gravity and incompatible with a ballistic path at solar gravity. We report on significant differences of fibril properties between those occurring above a dense plage region and those above a less dense plage region where the magnetic field seems more inclined from the vertical. We compare these findings to advanced numerical two-dimensional radiative MHD simulations and find that fibrils are most likely formed by chromospheric shock waves that occur when convective flows and global oscillations leak into the chromosphere along the field lines of magnetic flux concentrations. Detailed comparison of observed and simulated fibril properties shows striking similarities of the values for deceleration, maximum velocity, maximum length, and duration. We compare our results with observations of mottles and find that a similar mechanism is most likely at work in the quiet Sun.

Rapid Temporal Variability of Faculae: High‐Resolution Observations and Modeling

We present high-resolution G-band observations (obtained with the Swedish 1 m Solar Telescope) of the rapid temporal variability of faculae, which occurs on granular timescales. By combining these observations with magnetoconvection simulations of a plage region, we show that much of this variability is not intrinsic to the magnetic field concentrations that are associated with faculae, but rather a phenomenon associated with the normal evolution and splitting of granules. We also show examples of facular variability caused by changes in the magnetic field, with movies of dynamic behavior of the striations that dominate much of the facular appearance at 01 resolution. Examples of these dynamics include merging, splitting, rapid motion, apparent fluting, and possibly swaying.

The Contrast of Magnetic Elements in Synthetic CH‐ and CN‐Band Images of Solar Magnetoconvection

We present a comparative study of the intensity contrast in synthetic CH-band and violet CN-band filtergrams computed from a high-resolution simulation of solar magnetoconvection. The underlying simulation has an average vertical magnetic field of 250 G with kG fields concentrated in its intergranular lanes, and is representative of a plage region. To simulate filtergrams typically obtained in CH- and CN-band observations we computed spatially resolved spectra in both bands and integrated these spectra over 1 nm FWHM filter functions centred at 430.5 nm and 388.3 nm, respectively. We find that the average contrast of magnetic bright points in the simulated filtergrams is lower in the CN-band by a factor of 0.96. This result strongly contradicts earlier semi-empirical modeling and recent observations, which both etimated that the bright-point contrast in the CN-band is \emph{higher} by a factor of 1.4. We argue that the near equality of the bright-point contrast in the two bands in the present simulation is a natural consequence of the mechanism that causes magnetic flux elements to be particularly bright in the CN and CH filtergrams, namely the partial evacuation of these elements and the concomitant weakening of molecular spectral lines in the filter passbands. We find that the RMS intensity contrast in the whole field-of-view of the filtergrams is 20.5% in the G band and 22.0% in the CN band and conclude that this slight difference in contrast is caused by the shorter wavelength of the latter. Both the bright-point and RMS intensity contrast in the CN band are sensitive to the precise choice of the central wavelength of the filter.

Localizing plages on BO Mic

We have obtained a densely sampled time series of Ca  H&K line profiles of the ultrafast rotating K-dwarf star BO Mic. Taken at high resolution, the spectra reveal pronounced variations of the emission core profiles. We interpret these variations as signs of concentrated chromospherically active regions, in analogy to solar plages. We further interpret the varia- tions as partly due to the rapid growth and decay of plages, while other variations appear to be caused by plages moved over the visible stellar disk by rotation. The equivalent width of the Ca K core emission changes approximately in anti-phase to the photospheric brightness, suggesting an association of the chromospheric plage regions with pronounced dark photospheric spots. We believe that further analysis of the presented spectral time series will lead to a chromospheric Doppler image of BO mic.

Magnetic Flux Ropes in the Solar Photosphere: The Vector Magnetic Field under Active Region Filaments

This study is an observational search for evidence of prominence-associated magnetic flux ropes at the photospheric level in plage regions away from sunspots. Although there are a few cases of supposed flux rope geometry in the photosphere of complex sunspot regions (notably δ sunspots), such cases appear to be fairly rare and therefore do not contribute significantly to the suspected emergence of magnetic helicity in active regions leading to the common eruption of coronal mass ejections. In the present study, the Advanced Stokes Polarimeter (ASP) at the Dunn Solar Telescope of the National Solar Observatory was used to measure simultaneously the evolution of the photospheric vector magnetic field and the structure of the chromosphere as seen in the Hα line. The high angular resolution of these vector field measurements coupled with the high polarimetric precision of the ASP has allowed the possible identification of flux rope (concave up) geometry in the photosphere below some narrow (and supposedly low-lying) active region filaments. Two cases of such occurrence are presented in detail. These results indicate that flux ropes may be rather common in normal (i.e., non-δ sunspot) active regions. Further observations should shed some light on origin of these flux ropes as either emerging from the solar interior or generated in the solar atmosphere as a consequence of driven magnetic footpoint motion and reconnection.

Connecting solar radiance variability to the solar dynamo with the virial theorem

AbstractThe variability of solar radiance over a solar cycle is thought to result from a delicate balance between the radiative deficit of sunspots and the extra contribution of plage and network regions. Although the net effect is tiny, it must imply structural and thermal changes in the Sun or in partial layers of it as an unavoidable consequence of the virial theorem. Using the virial theorem for continua–including the magnetic field–it can be shown how solar radiance variability might be connected to a deeply seated flux‐tube dynamo and how this connection is established on a hydrodynamical time scale. (© 2005 WILEY‐VCH Verlag GmbH &amp; Co. KGaA, Weinheim)

Investigations of the MnI 539.47 nm and MnI 542.04 nm lines in solar plages

We observed the MnI 539.47 nm and 542.04nm line profiles in the quiet photosphere and plage regions in 1998 and 2001, and measured the full widths at half maximum, equivalent widths and depths of these profiles. The relative changes of the MnI 539.47 nm line parameters, normalized to the values obtained for quiet photosphere, as a function of magnetic field strength in plages is analyzed. We found, for plage region observed in 1998, that both the equivalent width and the depth of the line profile decrease with increasing strength of magnetic filed in plage at a rate of 9x10?4 /Gauss, but the full width at half maximum does not show any significant regular changes. Based on these results, the variations of the MnI 539.47 nm spectral line in solar flux with activity cycle could be explained by the variation of solar surface coverage with plages. For observations in 2001, the equivalent width and the depth of this line profile also decrease with increasing strength of magnetic filed in plage, but there are significant differences in the behavior of line parameters in comparison with the 1998 values. Comparison of changes of the MnI 539.47 nm line parameters with the parameters of the MnI 542.04 nm line in 1998 shows a clear discrepancy between them. On the contrary, in 2001 the full widths at half maximum and the equivalent widths of these two lines behaved in a similar fashion. .

Simulations of magneto-convection in the solar photosphere

We have developed a 3D magnetohydrodynamics simulation code for applications in the solar convection zone and photosphere. The code includes a non-local and non-grey radiative transfer module and takes into account the effects of partial ionization. Its parallel design is based on domain decomposition, which makes it suited for use on parallel computers with distributed memory architecture. We give a description of the equations and numerical methods and present the results of the simulation of a solar plage region. Starting with a uniform vertical field of 200 G, the processes of flux expulsion and convective field amplification lead to a dichotomy of strong, mainly vertical fields embedded in the granular downflow network and weak, randomly oriented fields filling the hot granular upflows. The strong fields form a magnetic network with thin, sheet- like structures extending along downflow lanes and micropores with diameters of up to 1000 km which form occasionally at vertices where several downflow lanes merge. At the visible surface around optical depth unity, the strong field concentrations are in pressure balance with their weakly magnetized surroundings and reach field strengths of up to 2 kG, strongly exceeding the values corresponding to equipartition with the kinetic energy density of the convective motions. As a result of the channelling of radiation, small flux concentrations stand out as bright features, while the larger micropores appear dark in brightness maps owing to the suppression of the convective energy transport. The overall shape of the magnetic network changes slowly on a timescale much larger than the convective turnover time, while the magnetic flux is constantly redistributed within the network leading to continuous formation and dissolution of flux concentrations.