INTERACTION BETWEEN NETWORK AND INTRANETWORK MAGNETIC FIELDS JUN ZHANG1,2, JINGXIU WANG1,2 , CHIK-YIN LEE3 and HAIMIN WANG3 1 Beijing Astronomical Observatory, Chinese Academy of Sciences, Beijing 100012, China; e-mail: [email protected] 2 National Astronomical Observatories, Chinese Academy of Sciences, Beijing 100012, China 3 Big Bear Solar Observatory, New Jersey Institute of Technology, U.S.A.

(Received 13 October 1999; accepted 6 December 1999)

Abstract. Using high-resolution observations of deep magnetograms and Hα filtergrams obtained at Big Bear Solar Observatory during 17 – 24 October 1997, we have studied the interaction of intranetwork and network elements. The relationship between small-scale magnetic fields and active phenomena is investigated. Most of the small-scale active phenomena are triggered by the interaction either between intranetwork and network magnetic elements or among several network elements. The energy released due to the interaction of intranetwork – network elements and network – network elements is large enough to heat the corona.

1. Introduction The quiet Sun is never quiet. There is a dynamic sea of mixed-polarity magnetic fields on the quiet solar surface. Flux emergence, migration, cancellation, coalescence and fragmentation are observed everywhere at the photosphere. Network fields are found at the network boundaries and in the vertices of supergranule cells (Simon and Leighton, 1964). The intranetwork fields are mixed-polarity magnetic elements inside the network (Livingston and Harvey, 1975). The interaction among ephemeral regions, network and intranetwork magnetic elements continuously drives disturbances from the base of the magnetized solar atmosphere. Network (NT) and intranetwork (IN) magnetic elements have peaks in their magnetic flux distributions at 2 × 1018 Mx and 6 × 1016 Mx (Wang et al., 1995), respectively. IN elements have a lifetime of approximate 2 hours (Zhang et al., 1998a), they contribute a total flux of 1024 Mx per day to the Sun. Important physical processes for the creation and destruction of magnetic elements of the quiet Sun are emergence of ephemeral regions (ERs) (Harvey and Martin, 1973) and flux cancellation (Livi, Wang, and Martin, 1985; Martin, Livi, and Wang, 1985), respectively. Recently, we have obtained a number of sequences of the best quiet-Sun magnetograms and confirmed that intranetwork fields follow supergranular flow and are swept into the network boundaries (Zhang et al., 1998b). They continuously merge into or cancel with network elements, resulting in the endless disturbance of the so-called quiet Sun. Solar Physics 192: 415–426, 2000. © 2000 Kluwer Academic Publishers. Printed in the Netherlands.

416

JUN ZHANG ET AL.

2. Data and Analysis Simultaneous Hα images and longitudinal magnetograms of the quiet Sun near disk center were obtained at Big Bear Solar Observatory (BBSO) from 17–24 October 1997. The observations cover approximately 9 hours every day. The highresolution Hα images were taken with the 65 cm telescope and have a field of view (FOV) of approximately 25600 square with temporal and spatial resolutions of 40 s (20 s between Hα filtergrams) and 100 (0.500 spatial pixel size), respectively. The magnetograms were obtained with the videomagnetograph (VMG) system on the 25 cm telescope and have a FOV of approximately 32000 E–W and 25000 N–S. The mean temporal resolution is six minutes. The spatial resolution of the best magnetograms is about 1.500 (0.6200 spatial pixel size), and the noise level is below 2 G. The working wavelength of the magnetograph is 6103 Å, and the bandpass of the filter is 0.25 Å FWHM. The calibration of the magnetograms is such that 1% polarization equals 115 G in flux density. The two days (17 and 18 October) with the best image quality are selected for this analysis. An important step in the analysis is to co-align the Hα and VMG images. For the images used in the present analysis this was accomplished by identifying strong network magnetic flux elements, seen in the VMG images, with Hα network bright points and rosette structures (the best way to align magnetograms and Hα images is to use the Hα off-band observations. Unfortunately, we do not have the Hα offband data). We estimate the accuracy of the mapping to be of order 2–3 arc sec. Figure 1 shows an example of quasi-cotemporal and quasi-cospatial Hα filtergram and magnetogram, taken from one field of view, centered at (12800 S, 1100 E) on 17 October. The upper part of the figure is an Hα filtergram, and the lower part is a magnetogram, which was acquired from integrating 4096 video frames. The field of view of the Hα filtergram and magnetogram in this figure is 25600 × 21500 . The six windows framed in both the Hα filtergram and magnetogram represent six conjunctions of network, and each arrow presents a network element. Similar to Figure 1, Figure 2 presents another example of an Hα filtergram and magnetogram, centered at 2700 S and 19000 W, observed on 18 October. The field of view of the Hα filtergram and magnetogram is 25600 × 22000 . We should mention that the seeing quality changes continuously during the observational time, and this causes the quantitative results presented in this paper to be uncertain. It is a pity that we have not found a way to estimate the influence of seeing accurately. However, according to our experience, it was noticed that, on the quiet Sun, the total flux and standard deviation of the magnetic elements in the whole FOV are unchanged at a one day interval, when the seeing conditions are unchanged. By measuring the total flux and standard deviation, we can estimate the seeing effect. It was found that, between 16:30 and 20:30 UT on 17 and 24 October (the time interval used in this paper), there are 3–15% fluctuations of total flux and standard deviation.

NETWORK AND INTRANETWORK MAGNETIC FIELDS

417

Figure 1. Quasi-cotemporal and quasi-cospatial Hα filtergram (upper panel) and magnetogram (lower panel) showing one field of view on 17 October, centered at 12800 S and 1100 E. Windows 1 – 6 in the Hα filtergram and magnetogram mark six conjunctions of networks. Each arrow indicates a network element. In the magnetogram, positive flux elements are shown by brighter color, and negative by darker color.

418

JUN ZHANG ET AL.

Figure 2. Similar to Figure 1. Another example of quasi-cotemporal and quasi-cospatial Hα filtergram (upper panel) and magnetogram (lower panel) observed on 18 October.

NETWORK AND INTRANETWORK MAGNETIC FIELDS

419

Figure 3. Time series of magnetograms and Hα filtergrams (labeled by Window 1 in Figure 1) showing the evolution of network elements and the related small-scale active phenomena. The upper four rows are magnetograms, the last row, Hα filtergrams.

420

JUN ZHANG ET AL.

Figure 4. Plot showing an example of convergence of several network elements (Window 6 in Figure 1). The shrinking circle encompasses the same magnetic elements.

3. Interaction of IN and NT Magnetic Fields At network cell boundaries, there often appear strong NT elements and converging flow of many IN elements. Lots of IN flux either merges or cancels with NT elements. The consequence of this interaction is manifested by the continued network brightening, spicules and/or macrospicules ejected from the strong NT elements. In Figure 3, we present a series of magnetograms and Hα filtergrams showing the evolution of network elements and the related small-scale active phenomena. The

NETWORK AND INTRANETWORK MAGNETIC FIELDS

421

Figure 5. A typical example of the evolution of magnetic fields and Hα images (Window 1 in Figure 2).

upper four rows are magnetograms, the last row, Hα filtergrams. Three arrows at 17:27 UT indicate three network magnetic elements. Seen from the time series of magnetograms, elements ‘2’ and ‘3’ at 17:27 UT were cancelling one another (see the two arrows at 18:30 UT). The flux of element ‘2’ decreased from 8.4×1017 Mx (17:27 UT) to 5.1 × 1017 Mx (18:30 UT), while that of element ‘3’ decreased from 2.7 × 1018 Mx to 1.1 × 1018 Mx. Comparing Hα filtergrams with magnetograms we found a macrospicule (shown by an arrow at 17:26 UT) rooted at the large network element ‘1’ of positive polarity (see the arrow at 17:27 UT). It seems that

422

JUN ZHANG ET AL.

Figure 6. An example of ephemeral regions (Window 8 in Figure 2).

the network element is always interacted by intranetwork elements which are at a level of very low flux density. Figure 4 shows an example of convergence of several network elements. The shrinking circles at 16:31 and 18:20 UT indicated regions that included the same network elements. It is obvious that these network elements had converged toward one another. Figure 5 presents another example of the evolution of magnetic fields and their manifestation in the chromosphere. This figure shows that the interaction between network and intranetwork elements is in the form of continuous impacting of very weak intranetwork elements onto network ones. From the sequence of magnetograms, we found that two network elements (shown by two arrows at 16:43 UT) of positive polarity converged first (from the 16:43 UT to 17:46 UT), then they moved apart. Although we have not found an obvious interaction between elements of opposite polarity, we found that, from the time series of Hα filtergrams, many kinds of small-scale active phenomena appeared, for example, the arrow at 18:02 UT shows a macrospicule, and the arrow at 18:41 UT indicates a minifilament. We would further suggest that weak magnetic flux whose flux density is

NETWORK AND INTRANETWORK MAGNETIC FIELDS

423

Figure 7. Upper panel: magnetic fluxes of the ephemeral region shown in Figure 6 as functions of time. The solid and dotted curves indicate the positive and negative flux, respectively. Lower panel: total magnetic flux (solid curve) and standard deviation (dotted curve) against time in the observational FOV on 18 October.

below the detection limit of the magnetograph plays a role in creating quiet-Sun activity. Ephemeral regions (ER) are another kind of magnetic field in the quiet Sun (Zwaan, 1987). In Figure 6 we present an example of an ephemeral region. A bipole of magnetic elements (indicated by two brackets at 17:39 UT) appeared before 17:25 UT. From 17:25 UT to 20:02 UT, the two elements of the ephemeral region moved about 1.0 × 104 km with a relative velocity of about 1.0 km s−1 . The upper panel in Figure 7 shows the magnetic fluxes of the ephemeral region as functions of time. The solid and dotted curves indicate the positive and negative flux, respectively. The rate of flux emergence reaches 5.1×1017 Mx hr−1 , measured from the negative pole of the ER. As we have mentioned in Section 2, the seeing quality changes in time, and this leads to a change in the measured flux. In order to estimate the seeing effect, the total magnetic flux (solid curve) and standard deviation (dotted curve) against time, in the observational FOV on 18 October, are

424

JUN ZHANG ET AL.

Figure 8. Flux of network elements against time for Windows 1 – 6 shown in Figure 1.

shown in the lower panel of Figure 7. It is found that the influence of seeing causes about 14% error in the measured flux.

4. Statistical Results From the series of magnetograms, we found that there were some regions where magnetic elements converged. These regions are seen at network cell boundaries, and there often appear strong network elements and converging flow of many IN elements. These regions have been named convergence centers of magnetic

NETWORK AND INTRANETWORK MAGNETIC FIELDS

425

flux (Zhang et al., 1998b). The interaction between intranetwork flux and network elements is manifested by the continued network brightening, spicules and macrospicules ejected from the strong network elements. From the two-day observations, we identified 78 convergence centers in the areas present in Figures 1 and 2. Six such conjunctions were shown in Figure 1, and another eight conjunctions were indicated in Figure 2. On average, there are four network elements in every conjunction. To assess how much magnetic energy can be released at the network conjunction, we have tried to estimate the flux disappearance at the conjunctions. In order to reduce the error of flux measurement, we measured the flux of 55 network elements within the 14 conjunctions from 17:00 UT to 19:00 UT (in this two-hour interval, there is about a 12% error of the measured flux), when the seeing of the observations is very good and every network elements can be clearly distinguished. Figure 8 presents the flux of network elements against time for Windows 1–6 shown in Figure 1. The decrease of network flux in the interval between 17:00 UT and 18:00 UT (in this one hour interval, the influence of seeing causes about 6% error in the measured flux) is determined as follows: if a magnetic field decreases monotonically, the loss of network flux is the difference of the flux between 17:00 UT and 18:00 UT; if a network flux element increases monotonically, we define the loss of the network flux as zero; if a network flux element decreases first, and then increases, we define the flux loss as the decrease of magnetic flux in the interval. The mean loss of a network flux at the one hour interval is about 4.2 × 1017 Mx. If the decrease of flux is caused by slow magnetic reconnection, some magnetic energy will be released due to the reconnection. Over the whole Sun, there are 8100 conjunctions, and 3.2 × 104 network elements in these conjunctions. This energy dissipation rate can be estimated from (Zhang et al., 1998a) dE B2 N 1 F = V = B NH , (1) dt 8π 1T 3600 8π where B is the intrinsic strength of network element, N is the number of network elements (3.2 × 104 ) on the whole Sun, 1T is the one hour interval (3600 s), V is the volume of a magnetic element, F is the mean loss of network flux in 1 hour interval, and H is the height of a network element. For a conservative reckoning, we adopt H as 3000 km (spicules are observed at the edge of the solar disk, starting from a height of about 3000 km above the photosphere and below this they overlap, forming a continuous chromosphere). B = 1500 G (Stenflo, 1973) and F = 4.2 × 1017 Mx. We then obtain dE/dt = 6.7 × 1028 ergs s−1 . 5. Conclusions and Discussion Sometimes we find coalescence of NT elements of the same polarity and vigorous macrospicule ejection takes place co-spatially and co-temporally with the coalescence. We assume that there might be IN flux below the detection limit which

426

JUN ZHANG ET AL.

continuously interacts with strong network elements. We find in many cases that the magnetic interaction which causes activity is at the level of very low flux density. The nature of the coronal heating mechanism is still a strongly debated topic. The power of 1.8 × 1028 ergs s−1 is required to heat the corona (Withbroe and Noyes, 1977). Although it has been accepted that the Sun’s magnetic field plays a vital role, neither theory nor observations have yet been able to unambiguously settle this argument. Parker (1988) proposed that the corona is heated by the cumulative effect of many small, localized bursts of energy corresponding to magnetic reconnection of the magnetic field. There has been heightened interest in the search for observational evidence of small-scale energy release events in the corona. Schmieder et al. (1994) concluded from observations that the simplest microflare may be composed of a large number of smaller events and that it is likely that the events are triggered successively. Allowing for a large uncertainty of our estimate, we found that the energy due to the interaction between intranetwork network elements and between network–network elements is large enough for coronal heating.

Acknowledgements We thank the anonymous referee for many suggestions to improve the science of this paper. The authors are indebted to Wei Li for taking the observations. This work is supported by the Major Project 19791090, funded by National Natural Science Foundation of China (NSFC), US NSF CAREER Award ATM-9628862 to Haimin Wang and NSF China-US collaboration grant INT-9603534.

References Harvey, K. L. and Martin, S. F.: 1973, Solar Phys. 28, 61. Livi, S. H. B., Wang, J., and Martin, S. F.: 1985, Australian J. Phys. 38, 855. Livingston, W. C. and Harvey, J.: 1975, Bull. Amer. Astron. Soc. 7, 346. Martin, S. F., Livi, S. H. B., and Wang, J.: 1985, Australian J. Phys., 38, 929. Parker, E. N.: 1988, Astrophys. J. 330, 474. Schmieder, B., Fontenla, J., Tandberg-Hanssen, E., and Simnett, G. M.: 1994, Proc. of Kofu Symposium, NRO Report No. 360, p. 339. Simon, G. W. and Leighton, R. B.: 1964, Astrophys. J. 140, 1120. Stenflo, J. O.: 1973, Solar Phys. 32, 41. Wang, J., Wang, H., Tang F., Lee, J., and Zirin, H.: 1995, Solar Phys. 160, 277. Withbroe, G. L. and Noyes, R. W.: 1977, Ann. Rev. Astron. Astrophys. 15, 363. Zhang, J., Lin, G., Wang, J., Wang, H., and Zirin, H.: 1998a, Solar Phys. 178, 245. Zhang, J., Wang, J., Wang, H., and Zirin, H.: 1998b, Astron. Astrophys. 335, 341. Zwaan, C.: 1987, Ann. Rev. Astron. Astrophys. 25, 83.