4. Temporal development: light curves
To study the temporal development of the NBPs, we computed the light curves for the bright points at each wavelength or signature. The curves for each NBP were obtained by selecting an area that contained the bright point throughout the whole observing period (even if it moved spatially), and then averaging, for each time, over all the pixels whose intensity exceeded the threshold value described in Sect. 3.1. A threshold equal to the nominal data noise was used for the magnetic curves. As said in Sect. 3.1, the NBPs are not directly visible in some photospheric signatures, such as white light, NiI pseudo-continuum and velocity images, hence we couldn't use an intensity (velocity) threshold to obtain the corresponding light curves. To guarantee the comparability with the other light curves, in these cases we computed, at each time, the average value over the spatially corresponding magnetic areas.
We also computed the light curves of 11 areas randomly selected in the quiet regions of our FOV, and of size comparable to that of the NBPs (about 3"3"). No threshold was applied for their computation. These quiet areas should represent the so-called internetwork regions, which appear field-free at MDI sensitivity. Using the same number of internetwork areas as of network bright points, and performing the analysis in the same way, will give us confidence in the comparison in a statistical sense. We remark that this is not always the case, especially for spectrographical observations where the slit samples a number of internetwork points that is usually much larger than that of network structures. In Fig. 3, as an example, we show the light curves of different signatures for a NBP. Fluctuations are evident at each wavelength. The red and blue wings of H show a very similar temporal evolution, with simultaneous intensity variations of the same magnitude for most NBPs. For other signatures a direct comparison does not show any simple relation between the variations at different atmospheric heights.
We must comment here on the fact that most authors, when studying the temporal properties of network points, use a different method to determine the "light curves". Basically, the network points are spatially identified using the temporal average of some suitable signature, and then the temporal development of each image pixel belonging to the average structures is considered and analyzed. We believe that the method we adopted has several advantages: - Since the network points might move over the course of time, they do not always correspond spatially to the structures identified on the average maps. Our method guarantees that the structure is properly followed in time, avoiding the loss of relevant pixels, or the inclusion of spurious ones; - If the magnetic structure giving rise to the network point inclines with height, a set of pixels identifying the NBP at a given wavelength might not well represent the same structure at a different wavelength. This is especially relevant when performing comparisons between different atmospheric layers, for example in the phase difference analysis of Sect. 6. We overcame this problem by selecting areas large enough to include the NBPs at each wavelength and each time, as explained earlier.
Computing the light curves as an average over a given area implies the assumption of a spatial coherence over the whole area (of about 3"3", equivalent to a spatial frequency of about 3 Mm-1) competing to each NBP. This seems a reasonable assumption because we do not see any significant inhomogeneities within the single structures.
© European Southern Observatory (ESO) 2000
Online publication: June 5, 2000