The solar network, defined as the boundaries of supergranular flows, contains oscillation with periods longward of the acoustic cut-off and the Brunt-Väisällä cut-off (the low-frequency cut-off for gravity waves). These are not correlated with underlying fluctuations (Lites et al. 1993; Cauzzi, Falchi, & Falciani 2000), and their origins have been attributed to magnetic stochastic processes (Kneer & von Uexküll 1985; 1986; von Uexküll et al. 1989), or events spatially removed from the network, e.g., granular overshoot. Imaged in the light of Ca II K3 (3933.7Å), the network is only partially defined, as brightenings at vertices of several supergranules. Due to the coincidence of these network bright points (NBPs) with high magnetic field concentrations, magnetohydrodynamic (MHD) waves (Kalkofen 1997; Hasan & Kalkofen 1999) and magneto-gravity waves (Damé, Gouttebroze, & Malherbe 1984; Deubner & Fleck 1990; Kneer & von Uexküll 1993) seem to be the modes most likely present. The mechanism responsible for heating the network is still the topic of much debate, with a recent publication by Rosenthal et al. (2002) beginning to readdress the problem.
| Reference | Wavelength | Frequencies | Dur. | Freq. Res. | Cadence | Notes |
| (mHz) | (mins) | (mHz) | (secs) | |||
| This study | Ca II K3 | 1-4 | 150 | 0.11 | 45 | 0.3Å FWHM bandpass |
| Damé 1984 | Ca II K3 | 1.3, 2.0, 2.7, 3.2 | 52 | 0.32 | 10 | 1.2Å FWHM bandpass |
| Kneer & von Uexküll 1986 | Hα,Hα±0.45Å | <2.4, ~3.3 | 54 | 0.31 | 20 | I only |
| von Uexküll 1989 | Hα | 2.4, (3.3(a)) | 64 | 0.26 | 30.4 | (a) V only |
| Kneer & von Uexküll 1993 | Hα | 3.5, <2.4 | 128 | 0.13 | 15 | 0.25Å FWHM bandpass |
| Ca II K | 3.2, <2.4 | 128 | 0.13 | 15 | 0.6Å FWHM bandpass | |
| Mg I b2 | <2.4 | 128 | 0.13 | 15 | 0.3Å FWHM bandpass | |
| Deubner & Fleck 1990 | Ca II 8498Å& 8542Å | 3.3, <2 | 275 | 0.06 | 10 | V&I |
| Kulaczewski 1992 | Mg I | 3.3-5 | 80 | 0.21 | 15 | V&I |
| Ca II K | <2 | 80 | 0.21 | 15 | V&I | |
| Ca II 8542Å | <3 | 75 | 0.22 | 15 | V&I | |
| Lites 1993 | Ca II H | 0.9, 1.8, 2.5 | 62 | 0.27 | 5 | V only, 1 NBP |
| Kariyappa 1994 | Ca II H2v | 2.4 | 35 | 0.48 | 12 | I only, 3 NBPs |
| Bocchialini 1994 | He I, | ~3.3 | 83 | 0.20 | 5 | V&I, some low freq peaks |
| Bocchialini & Baudin 1995 | He I,Ca II | ~3.3 | 83 | 0.20 | 5 | V only, wavelet study |
| Baudin 1996 | He I,Si I | ~3.3 | 83 | 0.20 | 5 | V only, wavelet study |
| Cauzzi 2000 | Hα | 0.6, 1.3, 2.2 | 50 | 0.33 | 12 | 0.25Å FWHM bandpass, 11 NBPs |
| Curdt & Heinzel 1998 | Lyα | 2.2-2.4 | 33 | 0.50 | 33.5 | SUMER |
| Krijger 2001 | 1700,1600,1500 | 3.5 | 90 | 0.19 | 15.0523 | TRACE, all network |
| 1700,1600,1500 | 3.5 | 228 | 0.07 | 21.8625 | TRACE, all network | |
| Judge 2001 | JOP 72 data | <5 | multiple datasets | TRACE, SUMER, MDI | ||
| Krijger 2003 | 1216, 1600 | <2, 3-4 | 48 | 0.35 | 40 | TRACE |
| Lyman series (5,9,15) | 2-3 | 48 | 0.35 | 29 | SUMER | |
| Ca II K2v | 4-5 | 48 | 0.35 | 6 | 0.6Å FWHM bandpass | |
The existence of low-frequency (1-4 mHz) oscillations in the solar network of the quiet Sun is well known. Table 4.1 presents a summary of results from published literature (the frequency resolution term in Table 4.1 will be explained in Section 4.2.3). There are several points concerning these results which need to be addressed. Firstly, most studies simply integrate over all network pixels (normally defined as areas which have a higher intensity than some arbitrary value). This means that the lightcurve will be that of the entire network in the FOV, and any small-scale, localised oscillations may be washed out by stronger, global oscillations. There are some exceptions to this: Cauzzi et al. (2000) identify eleven separate NBPs, but then average over them; Kariyappa (1994) identifies three individual NBPs; Lites et al. (1993) and Curdt & Heinzel (1998) both study only one. Other studies take a different approach, by creating a lightcurve for each and every pixel inside a network `mask'. However this method is very susceptible to errors due to misalignment and changes in atmospheric seeing, and normally any resulting Fourier power spectra will once again be averaged over the entire network.
Secondly, few results identify specific Fourier peaks, instead preferring to refer to either `low-frequency power' below a specific value (e.g., Kneer & von Uexküll 1993) or around the `acoustic band' (e.g., Bocchialini, Vial, & Koutchmy 1994). Of course it must be recognised that this tendency to refer to a frequency band may simply be because of the requirements of the study (e.g., comparing oscillatory power between the network and internetwork). Other studies also prefer to use phase and coherence analysis (e.g., Deubner & Fleck 1990), and do not concentrate on identifying peaks in Fourier power. In cases where specific peaks are identified, their validity may be compromised as they correspond to periods which may be due to insufficiently long observing windows. Through this argument, Kalkofen (1997) discounts all but the 2.7 mHz frequency in Damé et al. (1984), and that at 2.5 mHz in Lites et al. (1993). A similar argument can be used to discount the 0.6 mHz period discovered by Cauzzi et al. (2000). Furthermore, the smoothing window used by Cauzzi et al. (2000) directly affects the 1.3 mHz oscillation found by these authors.
Thirdly, the importance of the frequency resolution and other erroneous Fourier effects must be fully recognised. A poor frequency resolution may result in some peaks being omitted (Section 4.2.4). The bottom half of Table 4.1 are results from SUMER (Solar Ultraviolet Measurements of Emitted Radiation, Wilhelm et al. 1995) and TRACE data. Although the long duration observations, free of atmospheric distortions, makes these data very useful, it should be noted that most results from SUMER studies have been omitted. The reason for this is that in the vast majority of SUMER studies of the quiet Sun, solar rotation is not followed. Hence a new area of the Sun rotates into the SUMER slit every 380 s, introducing a low-frequency cutoff at 2.65 mHz. With only taking specific peaks into account, and addressing the frequency resolution problem, all non-discounted frequencies can be reduced to two broad peaks, one at 2.4 ± 0.3 mHz, and a second at 3.0 ± 0.3 mHz.
In this chapter I use a `contour & contrast' technique to discover further peaks in NBP power spectra. Furthermore these peaks are identified with particular spatial positions in each NBP. The important improvements over previous work are,
- The careful identification of seven individual NBPs.
- The narrow-band Ca II K3 filter.
- Long duration observations leading to good frequency resolution.
- Excellent seeing and high spatial resolution.
- Creation and implementation of a new filtering routine.