Astron. Astrophys. 363, 279-288 (2000)

2. Observations and reductions

We (Sivaraman, Livingston and Gupta) have been collecting data from new observations every year from 1993 to 1997 with the primary aim of securing high-quality observations under the best seeing conditions possible. These observations consist of a sequence of spectroheliograms in the K-line at the East Auxiliary facility of the McMath-Pierce telescope and temporally simultaneous magnetic area scans using the spectromagnetograph (Jones et al., 1992) at the Vacuum Telescope closeby. A spectroheliograph for use at the focal plane of the spectrograph of the McMath - Pierce telescope was fabricated by Keith Pierce especially for these observations. These scans covered quiet regions around the centre of the disk. Among the observations collected over the years, the ones of April 16, 1994 are the best. The exit slit of the spectroheliograph was set at 400 µm which corresponds to a band pass of 40 mÅ positioned on the K_2V emission peak. The spectroheliograph scanned an area of 560 440 arc sec² on the sun in 30 seconds while the duration of the magnetic scans that covered an area of 700 512 arc sec² over the same region on the sun was about 5 minutes. The spectroheliograms (abbreviated as SHGs henceforth) were obtained on T-Max 400 Kodak films and were developed in D19 for 5-min. The magnetic scans were obtained with 26 integrations, twice the normal value adopted for the synoptic observations and this improved the S/N ratio by a factor of 1.4.

From the sequence of 10 K-line SHG and magnetic scans obtained on April 16, 1994, three K-line SHGs that temporally coincided with the three magnetic scans, constituting three K-line SHG - magnetic scan pairs, (designated pair 1, 2 and 3) were chosen for a detailed analysis (Table 1). Because of the narrow passband centred on the K_2V emission peak, the INBPs and the network stand out strikingly in emission in comparison with broad-band filtergrams. On our scans, there are about 15 INBPs per cell and we also notice that there is diffuse emission less bright than the INBPs almost everywhere within the network (see also Hale & Ellerman 1903, Fig. 3 of Plate IV), which shows the "minute calcium flocculi" and the diffuse emission at the "H₂ level". These diffuse emissions do not show up in the filtergrams with a broad band pass. At the time of minimum of solar activity the network elements, which consist of only the "quiet sun network" component, do not form a continuous boundary in all the cells, but are broken in between, giving an impression that the cells are incomplete. This is because the coarse mottles (De Jager, 1959) that form the additional network component are virtually absent during a deep solar minimum, whereas during other phases of the solar activity, even the regions considered quiet do contain a significant amount of the coarse mottles on the network boundaries along with the quiet sun network component which clearly defines the cell boundaries.

Table 1. Details of the scans of April 16, 1994. The times are in Mountain Standard Time (MST).

The K-line SHGs were digitized at the same resolution as the magnetic scans (1.14 arc sec per pixel). The analysis of the K-line SHG and magnetic scan pairs was done using the IRAF software system. The entire field of the two images forming a SHG - magnetic scan pair was divided into 4 subfields to ensure the best registration of the image pairs. The subfield was registered to 1.2 to 1.4 pixels. The bright points within the network were identified visually. A cursor was placed on a bright point and the brightness (film density) was recorded. We then flipped to the magnetic image and recorded the magnetic field at the same co-ordinates. This was done for all the INBPs in the whole sub-field identifying the INBPs one by one. We show in Fig. 1 the INBPs and the spatially-coincident magnetic elements in one area bounded by the circle as an example. Those elements that resembled the INBPs, but lay very close to the network boundary and so could have a possible membership with the network, were not included in the measurements; only those bright points that lay clearly within the network were measured. The brightness of each INBP is the maximum value over a 2 2 pixel area and the magnetic field for each IN magnetic element is the maximum absolute value of the field over the same 2 2 pixels in Mx cm^-2. We repeated this measurement procedure for the three SHG - magnetic scan pairs.

Fig. 1. A section of the spectroheliogram (left) and the coaligned magnetic area scan (right) pair. An example of the innernetwork bright points (INBPs) in the SHG that are cospatial with the innernetwork magnetic elements is shown within the circle (of diameter 30 arc sec) in the respective frames. There is a striking example of a bright INBP just below the centre of the circle that coincides with a negative polarity field. There is another INBP at the 9 o´clock position that coincides with a positive-negative polarity pair. The third one at the 3 o´clock position coincides with a positive polarity element and the fourth one approximately at the 4 o´clock position coincides with a negative polarity element. Examples of the Network Bright Points (low brightness) can be seen within boxes A and B.

Using a very narrow spectral band to isolate the K_2V emission peak, the number of INBPs per cell ranges from 10 to 20 and on most occasions it is closer to the latter figure. With this number of bright points, the interior of the network looks very crowded. Similarly, in the magnetic scans we see almost the same number of IN magnetic elements within each network as the INBPs. Thus it is a challenge to look for coincidences (or their absence) in crowded fields in the image pairs.

One way to rule out the possibility of chance coincidences is to reduce the population of the INBPs and also the diffuse emission within each cell and then look for coincidences. To do this we applied a threshold in brightness, to the K_2V Spectroheliograms mentioned in Table 1 and derived brightness images with only those INBPs left within the network that have brightness 1.5 (or 50% above ), where is the mean brightness over the network interior. We did the brightness - magnetic field measurements for all the INBPs that were now in the field. In the next step, we increased the threshold to 70% of and derived images with only those INBPs having brightness 1.7 . The measurements of the brightness - magnetic field were again done as before for all those INBPs remaining in the field. The brightness - magnetic field measurements were done for the three brightness threshold levels for the 3 SHG - magnetic scan pairs listed in Table 1. We wish to emphasise that we have made the coincidence measurements only on the INBPs and not on the diffuse emission regions. Another way to rule out the chance coincidence is to compare the INBPs of a cell with the IN magnetic elements of another cell in a different region on the sun.

The noise level in the magnetic scans was determined from a set of simultaneous magnetic scans obtained at the spectromagnetograph as well as at the McMath - Pierce telescope using ZIMPOL I (Keller et al., 1994) over a quiet area around the solar disk centre on another occasion. The noise level from a comparison of these scans (using the ZIMPOL I scans as the reference standard) is about 3 Mx cm^-2. This value is consistent with the increased integration time we used for the magnetograms that improved the S/N ratio by a factor of 1.4. However, we have set the noise level for our present measurements at 4 Mx cm^-2, a more conservative value which is the value normally used for the NSO magnetograms in all studies.

As we have described in our search for coincidences, we have read the maximum absolute value of the magnetic fields over a 2 2 pixel area for every IN magnetic element examined (which is the counterpart of an INBP) to decide whether the coincidence is with fields 4 Mx cm^-2 or 4 Mx cm^-2. Note that the lower detection limit for the flux corresponding to the noise level of 4 Mx cm^-2 would be about 2 10¹⁶ Mx. In addition, to have an idea of the magnetic environment around the 2 2 pixel area, we examined the field over a 10 10 pixel area centered around the 2 2 pixels which provided us the magnetic field of the IN magnetic elements. In all cases where the fields exceeded 4 Mx cm^-2 over the 2 2 pixels it was found that the field values outside the 2 2 pixels were also high and of either polarity. We also noticed that at each brightness level about 25% of the INBPs (as bright as the ones associated with fields 4 Mx cm^-2) coincided with magnetic elements (over the 2 2 pixel) with fields 4 Mx cm^-2. In such cases it was evident from the 10 10 pixel matrix that the 2 2 pixel area lay at the interface of merger of two opposite polarity elements, while the fields outside the 2 2 pixels were far above 4 Mx cm^-2 for both polarities. Thus these represent the meeting places of opposite polarity fields which eventually cancel, and examples of such cancellations have been illustrated well by Zhang et al. (1998b) and the increase in brightness of the INBPs above the sites of bipoles by Nindos & Zirin (1998). We are not able to follow the evolution of cancellation events as our data do not form a time sequence. But from an examination of the pixels in the immediate neighbourhood of the 2 2 pixels we are able to infer the cancellation of the opposite polarity fields and there are several such cases in our large sample. Zhang et al. (1998b) estimate that about 30% of the IN magnetic elements cancel with magnetic elements of opposite polarity at any time, and this is in agreement with the data reported here. The rest 15% of the INBPs lay over areas where the fields were much lower than 4 Mx cm^-2 both over the 2 2 pixels as well as in the neighbouring areas. These probably represent the unresolved emission and unresolved fields within the network.

© European Southern Observatory (ESO) 2000

Online publication: December 5, 2000
helpdesk.link@springer.de