Fig. 1: 3-D visualization of a new atmosphere model of a M-dwarf representative of AD Leonis. The magnetic field lines (red) are rooted in the footpoints at the visible surface (grey) and funnel out in the atmosphere above. (Click on the image for a larger version.)
Red dwarf stars of spectral type M, also called M-dwarfs, constitute 75% of all stars in the solar neighbourhood and in our galaxy. They have masses of less than half a solar mass, effective (surface) temperatures of 2000 - 4000 K and low luminosities (L < 0.02 Lsun ). From observations, we know that they can have strong magnetic fields and can thus be magnetically very active, which includes strong flares. Their large number makes them interesting for our understanding of stars in general and their role as host stars of extra-solar planets. And yet many details about the nature of these stars are unclear. A grant by the Research Council of Norway enabled me to extend my research towards M-type dwarf stars (since 2011). A first set of numerical models, which clearly exceed the state-of-the-art in the field, has been produced and is currently analysed [50]. It will be valuable for the interpretation of observations, which so far had to rely on static 1-D models. Furthermore, it will shed light on the physical processes at work in the atmospheres of cool stars.
Click here for the official ITA project page.
Fig. 2: 3-D visualization of a numerical model of the solar atmosphere showing a close-up region with a magnetic tornado. (Click on the image for a larger version.)
Magnetic tornadoes are thought to be abundant on our Sun. They are generated by vortex flows, which form due to the bathtub effect at the solar surface and force the footpoints of magnetic field concentrations to rotate. The magnetic fields extend through the atmospheric layers and thus mediate the rotation upwards, resulting in a net energy transport into the upper layers. There, the energy is dissipated by yet unknown physical processes and may contribute to the heating of the solar corona to temperatures in excess of a million degree Kelvin. Based on the results that I published in Nature in 2012 (Vol. 486, 505-508), I now lead a project funded by the Research Council of Norway, which addresses many fundamental and yet unknown aspects of this novel phenomenon through a combination of high-resolution observations with world-leading facilities like the ground-based Swedish 1-m Solar Telescope (SST) and the space-borne observatories Solar Dynamics Observatory and advanced numerical simulations with state-of-the-art 3-D radiative magnetohydrodynamics computer codes. Similar 3-D simulations for red dwarf stars (which constitute about 75% of all stars in our galaxy) will reveal if the tornado phenomenon is of general importance for cool stars.
Click here for the official ITA project page.
Fig. 3: Revised atmospheric structure of quiet Sun regions. (Click on the image for a larger version.)
Based on 3-D radiation magnetohydrodynamic simulations and high-resolution observations, I reviewed the atmospheric structure of quiet Sun regions, which is more complex and dynamic than previously anticipated. See the review article [ 19] for details. The most important ingredients are illustrated in Fig. 3. This picture includes the newly discovered magnetic "small-scale canopies" and the pronounced dynamics of the chromospheric layer, which have far reaching implications for the structure and heating of stellar atmospheres. Other important aspects concern the generation of m agnetic field (possibly via local small-scale dynamo action) and the interaction of shock waves with the atmospheric magnetic field.
Fig. 4: Illustration of a local 3-D model of the Sun extending from the upper convection zone into the chromosphere/fluctosphere.
Numerical simulations, which were produced with the radiation magnetohydrodynamics code CO5BOLD, were studied with respect to a large number of topics, for instance surface convection and wave phenomena [e.g., 11, 27, 38, 48]. An important finding was the discovery of a complex dynamic small-scale pattern at chromospheric heights [ 2, 3, 4, 27] This pattern is produced by interaction of propagating shock waves, which are self-consistently excited by the convective motions in the lower part of the models. The resulting rapidly changing, mesh-like shock pattern in the atmosphere consists of hot threads and enclosed cool post-shock regions. I introduced the term "fluctosphere" for this shock-dominated domain (see Fig. 3) in order to avoid the common confusion concerning the term chromosphere, which should be reserved for the magnetic field dominated "canopy" domain above [39].
The dynamic and intermittent nature of the fluctosphere turned out to be the key to solve the controversy about the temperature stratification of the solar atmosphere, confirming the 1D finding by Carlsson & Stein (1994). While the shocks can explain the measured chromospheric UV emission, the cool post-shock regions allow for the observed existence of molecules (see below).
Fig. 5: VAPOR visualisation of the top layers of a 3-D MHD model with granulation and magnetic field lines.
In the photosphere, the magnetic field gets almost completely expelled from the granule interiors due to the convective flows, resulting in a horizontally directed but continuously changing "small-scale canopy" field, which overlays these magnetic "voids". The resulting "horizontal internetwork fields" (HIFs) have been observed recently and are currently debated [ 15, 40, 42].
Fig. 6: Simulation of the hydrogen ionization fraction in the solar atmosphere with equilibrium (LTE, top) and time-dependent non-equilibrium treatment (NLTE, bottom).
A number of important processes deviate from equilibrium conditions in stellar chromospheres, which makes a detailed numerical treatment mandatory. Implementing such a realistic numerical description as part of complex atmosphere models is a very challenging task.
An important example concerns the ionization of hydrogen, which is the major constituent of the atmospheric gas. Numerical simulations with a time-dependent non-equilibrium treatment of hydrogen ionization (Fig. 6) proofed that the deviations from the equilibrium state have fundamental influence on the plasma properties (e.g., the density of free electrons) [8, 13, 34, 37]. The implementation of hydrogen ionisation was only a first step and will be followed by a similar description for other atomic species. I also investigated the ionization of calcium (Ca II/III) [21].
Fig. 7: Numerical model of the carbon monoxide concentration in the solar atmosphere.
Fig. 8: The turret of the Swedish Solar Telescope and an intensity map, exhibiting a chromospheric swirl event (dark ring).
During my various projects, I analysed data from different observatories, e.g. the Transition Region and Coronal Explorer (TRACE) spacecraft, the Solar Optical Telescope onboard the Hinode satellite, and the Dutch Open Telescope (La Palma, Spain). I was PI of an observation campaign at the German Vacuum Tower Telescope (Tenerife, Spain), Co-I at the Dunn Solar Telescope (Sunspot, USA), and participated repeatedly in observations with the Swedish Solar Telescope (La Palma, Spain). The obtained high-resolution data provided insight in the small-scale structure and dynamics of the lower to middle solar atmosphere [e.g., 12, 14, 32], including observational support for the predicted fluctospheric shock pattern [7, 9].
Attempts were made to directly measure the line-of-sight component of the magnetic field in the chromosphere from an infrared triplet line of singly ionized calcium [43, 44].
Just recently, I discovered small-scale swirl events in the solar chromosphere. These swirls seem to comprise compact regions (diameters of ~1500 km) of rapidly rotating gas in connection with magnetic flux structures. This process may have fundamental implications for the heating of the upper solar atmosphere [20, 47].
Fig. 9: Synthetic spectra near 4.7 microns with carbon monoxide lines in comparison to ATMOS3 data.
Fig. 10: Artist's impression of the Atacama Large Millimeter Array (Credit ESO) and a synthesized intensity map at a wavelength of 1mm.
Synthetic spectra and intensity images are needed for a direct comparison with observations. In the coursof various projects, I calculated image sequences from magnetohydrodynamical models at different wavelengths from the ultraviolet to the millimetre regime, using different radiative transfer codes. Applications for the Sun include, e.g., spectral lines of iron [31, 41], carbon monoxide (Fig. 9) [10] and singly ionized calcium, and the continua at (sub)-millimetre wavelengths.
The latter were used to predict what the Atacama Large Millimeter Array (ALMA) could observe in the near future. For that purpose, I calculated intensity image sequences for the accessible wavelength range at different positions on the Sun from its disk-centre to its limb. The analysis produced many results that will be valuable for the future planning and interpretation of solar observations with ALMA, defining constraints on required temporal and spatial resolution and scientific objectives. It was found that the formation height range increases with wavelength and also varies with inclination angle so that a combination of instantaneous observations could serve as tomography, revealing the three-dimensional thermal structure of the solar chromosphere [13, 28].
Fig. 11: Emergent continuum intensity synthesized for a MHD model of the Sun at a wavelength of 500 nm.
Fig. 12: Artist's impression of the Hinode space observatory (Credit JAXA) and an illustration of the point spread function for the Solar Optical Telescope onboard (red surface).
Comparisons of solar observations with numerical models are important for: (i) Testing the reliability of numerical models and the applied methods. (ii) In-depth analysis of the relevant physical processes and determination of physical quantities. (iii) Pointing out promising observational targets and yet undiscovered phenomena for future missions. (iv) Optimising observational techniques. For instance, synthetic spectra were used to determine the chemical abundance of silicon in the Sun and other stars [1]. Synthetic intensity images based on magnetohydrodynamic simulations have been compared to solar observations in different wavelength regimes for both continua and various spectral lines. The comparison of observations in the wing of the Ca II H spectral line, carried out with the Dutch Open Telescope, confirmed that the reversed granulation pattern in the middle photosphere is modelled already realistically [5]. Synthetic maps for the line core of the Ca II infrared line at 854.2 nm, which is formed at chromospheric heights, were compared to corresponding observations with the Dunn Solar Telescope [18]. It confirmed that a fluctospheric pattern exists in quiet Sun regions as part of a compound of atmospheric regions, which are dynamically coupled [39].
A particularly successful example concerns the contrast and centre-to-limb variation of the continuum intensity at visible wavelengths. For many years, there were large discrepancies between values derived from models and observations. This fundamental problem was resolved by properly taking into account the detailed instrumental properties of the employed telescope. For this purpose, the instrumental image degradation of the Solar Optical Telescope onboard the Hinode satellite was determined in form of a point spread function with scattered light contributions (see Fig. 12) [16]. With this, the models now reproduce the observations very well, which demonstrates that state-of-the-art numerical simulations provide a realistic description of solar surface convection [17, 45].
CO5BOLD
