Context. The origin of supergranulation is not understood yet and many scenarios, which range from large-scale deep convection to large-scale instabilities of surface granular flows, are possible. Aims. We characterize the velocities and divergences in supergranulation cells as a function of their size. Methods. Using local correlation tracking, we determine the horizontal flow fields from MDI intensity maps and derive the divergences. The smoothed divergences are used to determine the cells for various spatial smoothings, in particular at the supergranular scale. Results. We find evidence of intermittency in the supergranular range and a correlation between the size of supergranules and the strength of the diverging flow. We also show that the relation between rms velocities and scale (the supergranule radius R) can be represented by a law V rms ∼ R 0.66 . Conclusions. The results issued from our data point towards a scenario where supergranulation is a surface phenomenon of the sun, probably the consequence of a large-scale instability triggered by strong positive correlated rising flows.

Context. The convective origin of supergranulation is highly controversial. Past measurements of intensity variations inside supergranules have often been influenced by the brightness enhancement at the cell boundaries due to the magnetic network. Aims. We conduct a precise determination of intensity variations inside supergranules. Methods. We determine the supergranule cell boundary from smoothed divergence maps derived from horizontal flow maps. We derive these flow maps from intensity maps obtained by MDI/SOHO in high resolution mode. We discuss the different possible approaches to take into account the influence of the magnetic field which can be used to determine the intensity variations inside supergranules.Results. We observe a significant decrease of the intensity from the center to the boundary of supergranules. We also obtain additional clues from the behavior of the maximum intensities and minimum intensities around each pixel, which are related to granules and intergranules: the maximum intensity decreases from center to boundary, while the minimum intensity is constant or increases depending how restrictive the selection is. The difference between intensity profiles versus divergence and relative distance to cell center also provides complementary information. The corresponding temperature differences between cell center and boundary are in the range 0.8-2.8 K. The intensity enhancement (for the magnetic network) or deficit (for intranetwork fields) depends on the localisation inside the cell. Conclusions. It is the first time that such a detailed analysis of intensity variations inside supergranulation is performed. Our results are compatible with a convective origin of supergranulation, as the intensity decreases toward the boundary of the cells. However, new simulations of supergranulation are necessary to verify whether the compared behavior of granule and intergranule intensity variations is in close agreement with convection.

Context. The variation of supergranule cell sizes with the magnetic environment is still controversial. Aims. We study this relation in detail to understand the discrepancies observed between previous results. Methods. We determine the cell size using divergence of horizontal flows derived from local correlation tracking of intensity maps (MDI/SOHO). We study the variation of the cell size as a function of the magnetic field inside the cell. We also consider which component of the magnetic field most influences the cell size. Results. Our main conclusion is that there are no large cells when the magnetic field (in absolute value) averaged over the cell is large. This is mostly due to the magnetic field inside the cell (intranetwork fields), while strong network magnetic fields (at the cell boundary) are associated with larger cells. Further studies of the evolution of the cells and of the flux imbalance suggest that a high level of weak fields may prevent the formation of large cells. This is compatible with the expectation that strong magnetic fields should prevent large-scale flows. Conclusions. The relation between the local activity level determined by the average magnetic field inside the cells and the supergranule size is not linear. Furthermore, it strongly depends on the definition of the activity level (magnetic field inside the cell or magnetic network) and on the magnetic sensitivity of the data. This last point probably explains at least partially the conflicting results obtained up to now.