At disk-center, individual line profiles corresponding to different atmospheric columns are largely symmetrical, although the line bisectors in upflows typically have a slight -shape while downflows show slightly more pronounced -bisectors due to the increasing vertical velocities with depth (Figure 26). Closer to the limb the spatially resolved line profiles become quite distorted. In the line-forming region, the horizontal velocities are typically somewhat larger than the vertical velocities and indeed approach and occasionally exceed the sound speed. The total velocity span of profiles at inclined viewing angles is therefore substantially larger than at disk-center and the lines can develop multiple components as the line-of-sight passes inhomogeneities with distinctly different velocity shifts as well as temperatures. As a consequence the spatially resolved bisectors become very distorted and depend critically on the exact region of the surface that is being observed.
Before a proper comparison can be made between predictions and observations for spatially resolved lines the theoretical line profiles must be convolved appropriately to account for the finite instrumental resolution and atmospheric seeing (e.g., Nordlund, 1984), which unfortunately somewhat limits the information gained from the observations. Nevertheless, such exercises clearly demonstrate the remarkably good agreement between observations and predictions in general (e.g., Dravins et al., 1981; Dravins and Nordlund, 1990a; Kiselman, 1994; Asplund et al., 2000b; Kiselman and Asplund, 2001; Cauzzi et al., 2006; Pereira et al., 2008). Each line has unique fingerprints in correlation maps between for example continuum intensity and line strength, depth, shift, width, and asymmetry across the granulation pattern depending on their height of formation and sensitivity to the atmospheric conditions, as illustrated in Figure 27. Obviously, the agreement between the predicted and observed behavior in this respect is very satisfactory. Indeed, for most lines studied to date this is achieved even within the assumption of LTE in the 3D line formation. Notable exceptions are for weak low-excitation lines of minority species, which strongly suggests that the problem lies with the LTE line formation rather than a shortcoming of the 3D model atmosphere. In fact, when allowing for departures from LTE in the 3D formation of the Li i 670.8 nm line the good agreement is restored (Kiselman, 1998). The unavoidable conclusion is that the statistical properties of photospheric velocity, temperature, and pressure in the 3D simulations closely resemble the real Sun.
This work is licensed under a Creative Commons License.