© ESO, 2016

1. Introduction

Coronal holes are low density regions caused by the presence of locally open magnetic field lines, which are often observed at the solar poles. Owing to the low density they are observed as dark regions in coronal extreme (EUV) lines and X-rays. This characteristic is also observed at low radio frequencies formed in the corona, however, at frequencies higher than 15 GHz bright regions inside the the coronal holes are often observed (see for example: Gopalswamy et al. 1999; Brajša et al. 2007).

The presence of bright polar regions is clear in 17 GHz maps, obtained by the Nobeyama Radioheliograph (NoRH, Nakajima et al. 1994), and shows a cycle mimicking the polar faculae cycle (Selhorst et al. 2011; Gopalswamy et al. 2012; Nitta et al. 2014). Selhorst et al. (2010) proposed that these bright patches could be generated by small scale magnetic regions around the positions of intense magnetic patches (kG), that coincide with the polar faculae (Tsuneta et al. 2008). The presence of 17 GHz bright patches were also observed at lower latitudes and have been associated with the enhanced unipolar magnetic regions underlying the coronal holes (Gopalswamy et al. 1999; Maksimov et al. 2006). On the contrary, small bright structures observed in EUV filter images do not have a one-to-one correspondence with the bright 17 GHz patches (Nindos et al. 1999; Riehokainen et al. 2001; Nitta et al. 2014); for this reason, Nitta et al. (2014) suggest that most of the small bright structures are artifacts created by the image synthesis and deconvolution used in NoRH maps.

Selhorst et al. (2010) compared the variation of the polar mean emission at radio (17 GHz) and EUV (304 and 171 Å), obtained with the Extreme ultraviolet Imager Telescope (EIT) on board the Solar and Heliospheric Observatory (SOHO). The 304 and 171 Å polar brightenings have distinct behaviors during the solar cycle. While 171 Å is correlated with the solar cycle, 304 Å is anticorrelated and in agreement with 17 GHz. periods. The authors addressed the 171 Å variations to the presence of coronal holes, whereas the 304 Å behavior was attributed to polar structures in the lower atmosphere, which also determine the polar emission at the 17 GHz radio continuum.

The formation of the He ii 304 Å emission depends on the conditions of the lower atmosphere (transition region and chromosphere), but can also be affected by direct photoionization produced by coronal EUV photons (Hirayama 1971; Zirin 1975; Jordan 1975; Andretta & Jones 1997). Worden et al. (1999) compared the He ii 304 Å emission from different atmospheric structures, namely plage, enhanced network, active networks, and quiet chromosphere, using EIT/SOHO images, over a period of two years during the solar minimum. They found that the intensity contrast of those structures (averaged over their areas) remains constant with time, when compared to the quiet chromosphere. The quiet chromosphere emission also does not vary during the period, however, their definition of quiet chromosphere also includes coronal holes. Worden et al. also suggest that because of the constancy of the quiet chromosphere irradiance in the analyzed period, the He ii 304 Å long-term variations could be traced to the magnetic activity. This idea is supported by Pauluhn & Solanki (2003), who investigated magnetograms and chromospheric EUV irradiance during the period 1996–2000. They found that in quiet regions of the solar disk there was an increase of 10 to 20% in the magnetic flux density, and a corresponding increase in irradiance in He i 584 Å, with similar characteristics to He ii 304 Å emission in the different atmospheric structures (e.g. coronal holes, network regions Glackin et al. 1978). Glackin et al. (1978) indicated that both He i 584 Å and He ii 304 Å show limb brightening in quiet Sun regions and limb darkening in coronal holes. Finally, Del Zanna & Andretta (2011) showed that both He i 584 Å and He ii 304 Å are correlated well with the 10.7 cm flux over the period 1998–2010.

In this paper, we analyze the relationship between the limb emission at radio (17 GHz) and the EUV lines (304 and 171 Å) using synoptic limb charts covering the period between 2010 and 2015. The EUV images were taken by the Atmospheric Imaging Assembly (AIA; Lemen et al. 2012) aboard the Solar Dynamics Observatory (SDO), launched in 2010, whereas the 17 GHz maps were obtained by the Nobeyama Radioheliograph. These data are described in the following section, with the results presented in Sect. 3. The discussion and main conclusions are presented in Sect. 4.

2. Data analysis

For the analysis presented here we use daily AIA EUV images taken with narrowband filters centered on He ii 304 Å and Fe ix 171 Å. While the 304 Å line is an emission line formed in the transition region (TR), with temperatures ranging between 6 and 8 × 10⁴ K, the 171 Å comes from coronal heights, with temperatures of T ~ 10⁶ K. Images are 1024 × 1024 pixel size, and present an angular resolution of around 2.̋40.The prereduced images were obtained at the Joint Science Operations Center¹. To read the images, the standard AIA routines were used (Boerner et al. 2012), which are part of the Solar SoftWare (SSW) package (Freeland & Handy 1998). While the 304 Å images presented a constant exposure time of 2.90 s, at 171 Å the first 68 days of the mission used exposure time of 2.90 s and 2.00 s after that. To make the 171 Ådata base uniform, the digital number (DN) of all maps were normalized to exposure time of 2.00 s.

At 17 GHz (1.76 cm), solar maps have been obtained daily by the NoRH interferometer since 1992 with the same quality resolution (10−18″).We analyzed the online daily presynthesized images², which are taken during local noon using a 10 s integration time. These maps use the Steer CLEAN algorithm (Steer et al. 1984), which has the advantage of restoring diffuse features with a ratio better than 1:300 between the rms noise level and the peak brightness in the images (Koshiishi 2003), which is very good for our work. The 17 GHz maps have 512 × 512 pixels and a spatial resolution of 4.91 arcsecs. Emission at 17 GHz mostly comes from the chromosphere (opacity τ = 1), where the local temperature and electron density are estimated to be ~9.6 × 10³ K and ~9.3 × 10⁹ cm^-3, respectively (Selhorst et al. 2005).

To investigate the relationship between the limb emission at radio and EUV wavelengths, we chose to construct synoptic limb charts. To produce these charts, we calculated the mean emission of delimited limb areas 100′′ wide and angular separation of 5°. Figure 1 shows examples of solar images obtained on May 14, 2010, at the three wavelengths studied here at 17 GHz, 304 Å and 171 Å, respectively, with the delimited limb region used to construct the synoptic limb maps. The angular positions were defined clockwise with 0° at the west limb, 90° at the south pole, 180° at the east and 270° at the north pole.

This procedure results in mean intensity profiles along the solar limb. To compare the variation of these profiles across the years,we normalized the maps by the quiet Sun temperature (or data number (DN) for AIA images). The mean relative intensity (I_R) profiles for each map are shown in the bottom panels of Fig. 1. Distinct structures can be observed at the poles in the images:

• i)

  The south pole presents a coronal hole at171 Å (Fig. 1c),which results in a noticeable reduction of the mean limb intensitybetween 60° and 120° (Fig. 1f). The coronal hole and its associated mean intensity reduction are also observed at 304 Å (Figs. 1b and e). At 17 GHz we observe brights patches inside the EUV coronal hole area (Figs. 1a and d) in agreement with previous observations (Gopalswamy et al. 1999; Maksimov et al. 2006).

• ii)

  The north pole also presents a dark area in the delimited region at 171 Å (Fig. 1c); however, it is a filament instead of a coronal hole, as can be seen in the 17 GHz map (Fig. 1a). In this case, intensities at 17 GHz drop to values lower than the quiet Sun level (Fig. 1d). In the EUV images, the reduction in intensity is more prominent at 171 Å (Fig. 1f) than at 304 Å, whose emission remains similar to the surrounding regions (Fig. 1e).

3. The quiet Sun

The 17 GHz quiet Sun was determined for each individual map using a histogram (bin = 100), in which the most common temperature was identified. Then we performed a weighed mean including two values immediately before the histogram peak and two values immediately after it. The result was assumed as the quiet Sun. Since the NoRH are normalized to 10⁴ K, the quiet Sun oscillates around it.

As reported by Boerner et al. (2014), the AIA responses have degraded over the years, most notably in the 304 Å channel. Using the SolarSoft routine aia_get_response.pro, one can see the degradation of the AIA responses over time with respect to the EUV Variability Experiment (EVE) on board the SDO. At the end of 2015, the responses of the 171 Å and 304 Å channels decreased to ~86% and ~20%, respectively, of the initial values as of May 2010. We corrected the AIA 171 Å and 304 Å images for the time-dependent instrument degradation and used them to calculate the quiet Sun intensity in DN with the same procedure that was used to obtain the 17 GHz quiet Sun brightness temperature from the NoRH images. The bin sizes were set to ~1% of the quiet Sun DN at the beginning of SDO observations, namely, 4 and 1 at 171 Å and 304 Å, respectively.

Figure 2 shows the quiet Sun intensity derived in the above procedure. At 171 Å, the quiet Sun presents an overall decrease over the years (black dots). The quiet Sun at 304 Å (in gray) stayed almost flat until the latter half of 2014, when it started to decrease from ~100 DN to ~50 DN at the end 2015. The decrease of the quiet Sun in 304 Å images since 2014 could represent true solar variations. However, it could also be due to residual calibration issues, as the last calibration of the AIA responses with the EVE took place in 2014, which is called version 6. The EVE calibration rocket scheduled for launch in 2016 will help us identify the origin of the downturn of the 304 Å quiet Sun that started in 2014. We note that the synoptic results as presented in the next section and Fig. 3 are not strongly affected by the apparent decrease of the quiet Sun intensity at 304 Å.

4. Results

In the studied period (2010–2015), around 2000 images were analyzed for each wavelength,after manually checking and discarding images with problems (<2%) . The results at 17 GHz show that I_R in the 5° cells varies between 0.70 and 29.02; the lower values could be addressed as the presence of dark filaments in the delimited region, and the higher values are due to the presence of extreme active regions, which can reach brightness temperatures above 5 × 10⁵ K, i.e. 50 times greater than the quiet Sun values (Selhorst et al. 2014, and references therein). We attribute the stronger emission from active regions to the gyro-resonance contribution that becomes important in the presence of strong magnetic fields (~2000 G) in the transition region or at coronal heights (Shibasaki et al. 2011). Nevertheless, these extreme active regions are rarely observed at the limb in the analyzed period: only 24 days present relative intensities and 6 days have .

In EUV filter images, the presence of coronal holes results in a stronger decrease of limb emission with 304 and 171 Å reaching I_R minimum values of 0.57 and 0.49, respectively. On the other hand, the active regions at the limb increase I_R at a lesser pace than in radio with a maximum around 18.11 at 304 Å and 26.42 at 171 Å, which likely reflects the single free-free emission mechanism that is not affected by strong magnetic field in active regions.

The daily I_R angular distributions are organized in synoptic limb charts, in which the angular position is represented in the ordinates, and the time in the abscissas; intensities are color-coded. Since our main objective is to study the behavior of the polar regions as a whole, the I_R values in Fig. 3 are artificially saturated at 5.50, 3.30, and 1.40 at 171 Å, 304 Å, and 17 GHz, respectively, to enhance the visualization of the tenuous variations at the polar regions. We now discuss the results in terms of the equatorial and polar regions separately.

4.1. Equatorial emission

We calculated the monthly mean emission in the northern and southern hemispheres to quantify the intensity increase in the equatorial region observed in Fig. 3 and verify a possible north-south asymmetry in the solar activity. The monthly mean equatorial emission in the southern hemisphere is here defined asthe monthly average of intensities in the ranges 0 to 45° and 135 and 180° (see Fig. 1b). In the northern hemisphere, it is defined asthe monthly average between 180−225° and 315−360° (see Fig. 1c). The results are shown in Fig. 4. For comparison with the solar cycle, the bottom panel showsthe monthly sunspot number (SSN) for each hemisphere³. The southern and northern hemispheres are shown in red an blue, respectively, and yearly running averages (12 points) were applied to all the values (thick lines).

The equatorial results were are consistent with the hemispheric sunspot number. The three wavelengths have similar mean equatorial intensity (⟨ I_R ⟩) variations, which is evident from the significant correlation coefficients between the radio and the EUV for the monthly averages (Table 1). Following the sunspot number, the synoptic charts showed a northern dominance until 2012 and the increase in the southern activity at the end of 2013. The differences between the equatorial hemispheric emission were more evident at the chromospheric level, at 304 Å and 17 GHz (Figs. 4b and c), which resulted in high correlation indexes, 0.79 and 0.80, for the southern and northern hemispheres, respectively.

The hemispheric differences at 171 Å (Fig. 4a ) are smaller than those observed in the chromosphere, however, the behavior is still the same. Moreover, good correlations were found in the comparison with the 17 GHz monthly values, namely 0.73 for the south and 0.62 for the north.

4.2. Polar emission

While in the equatorial analysis the EUV and radio emission showed a good agreement, at the poles the scenario changes completely and the differences between EUV and radio are clear. At 171 Å the presence of coronal holes, observed as dark regions at the poles (Fig. 3 top), are reflected in the synoptic map as a decrease in the mean intensity. The dark regions remain stable for months and are present in high numbers at the south pole. At the north pole, the last large coronal hole was observed at the end of 2010, but coronal holes can be observed in the southern hemisphere until the first half of 2012and after 2015. Without the coronal holes both poles present an increase at 171 Å. The coronal holes observed at 171 Å were also identified in the 304 Å synoptic chart (Fig. 3 middle). On the other hand, at 17 GHz, instead of coronal holes the poles present a prominent intensity increase (Fig. 3 bottom) until 2012and after 2015. In the absence of the coronal holes (around 2013), I_R decreases.

In order to compare the temporal variation of the polar regions, the mean ⟨ I_R ⟩ of the south (60° and 120°) and north (240° and 300°) poles were calculated for each month. Results are shown in Fig. 5, and present the following characteristics:

• 171 Å: the mean⟨ I_R ⟩ increases at both poles (Fig. 5a) with the north (blue curve) more intense than the south pole (red curve);

• 304 Å: the mean ⟨ I_R ⟩ is almost constant at the north pole with a small reduction in the beginning of 2014 (Fig. 5b, blue curves). In the south the mean ⟨ I_R ⟩ started to increase with the disappearance of coronal holes and reached a maximum in 2013. Then ⟨ I_R ⟩ decreased to a minimum at the end of 2014 and, even with the presence of a coronal hole, it has started to increase again (Fig. 5b, red curves);

• 17 GHz: contrarily to the EUV, the south pole presented greater mean ⟨ I_R ⟩ values during the analyzed period. The north pole showed only small variations in the mean ⟨ I_R ⟩ during the analyzed period (Fig. 5c, blue curves). The south pole varied in the opposite way to the sunspot number, i.e., the mean ⟨ I_R ⟩ decreased in the period in which the sunspot number reached the maximum (red curves in Figs. 5c and d).

The correlation between the poles in radio and EUV resulted in lower values than those obtained at the equator. In the south the indexes comparing 17 GHz with 171 and 304 Å were −0.39 and −0.36, respectively. These negative values can be attributed to the presence of coronal holes, which are dark at EUV, but presented a bright counterpart at 17 GHz. At the north pole, the coronal holes were less frequently observed, which resulted in a correlation index of −0.28, between 17 GHz and 171 Å and a index close to zero (−0.04) in the comparison between 17 GHz and 304 Å.

5. Discussion and conclusions

We analyzed the polar and equatorial emission at 17 GHz, 171 Å and 304 Å through synoptic limb charts constructed from NoRH and AIA maps. The synoptic charts cover the period of 2010−2015 show increasing solar activity in the equatorial region and a reduction in the radio polar brighteningduring the period of maximum activity.

5.1. Equatorial region

At the equatorial region, the results for the three wavelengths clearly show the hemispheric asymmetry in the solar activity (see for example: Hathaway 2010). The dominance of the northern hemisphere coincided with the first SSN peak, whereas the second SSN peak occurs simultaneously with increase in the activity at the south (see Fig. 4). There is an interval between the peaks (around the beginning of 2013), which indicates an inversion in the asymmetry of the hemispheric activity. Moreover, the comparison between the hemispheres shows that the peak of activity in the south is more intense for all three wavelengths.

The stronger southern activity started in the first semester of 2013 and is coincident with the complete polar field reversion of the north pole (Mordvinov et al. 2015). Since solar cycle 20, polar field reversals first occurred in the north (Svalgaard & Kamide 2013), all those cycles showed a more intense first peak. This trend has clearly changed in the present cycle, being the more intense the second peak.

The monthly equatorial radio ⟨ I_R ⟩ in both hemispheres presented an overall positive correlation with the EUV (see Table 1).The trend has clearly changed in the present cycle, in which the second peak is more intense. A stronger activity in the southern hemisphere was more noticeable at 17 GHz and 304 Å, which both formed in the lower atmosphere.

5.2. Polar regions

In the polar regions, the presence of stable and long-lived coronal holes is evident at both EUV wavelengths; this is identified by the darker patches in Fig. 3. We also note an asymmetry of their hemispheric activity, in which the darker patches at the north pole started to disappear earlier than at the south pole, in agreement with the results obtained by Karna et al. (2014).Remarkably, the south pole showed coronal holes in the beginning of 2015, whereas no coronal holes were observed at the north pole in the same period.

The 17 GHz radio emission at the south pole clearlydecreases while the SSN increases, and vice versa (Fig. 5); on the other hand, the radio emission at the north pole is almost constant during the studied period.For the EUV bands, ⟨ I_R ⟩ at the north pole is more intense than the ⟨ I_R ⟩ at the south throughout the entire analyzed period. At 17 GHz, the south pole dominated in the period (cf. Selhorst et al. 2011; Gopalswamy et al. 2012; Nitta et al. 2014).

This work strengthens the association between coronal holes and the 17 GHz polar brightenings as it is evident in the synoptic limb charts shown in Fig. 3, in agreement with previous case study works (Gopalswamy et al. 1999; Maksimov et al. 2006). The enhancement of the radio brightness in coronal holes is explained by the presence of bright patches closely associated with the presence of increased unipolar magnetic regions underlying the coronal holes (Gopalswamy et al. 1999; Brajša et al. 2007; Selhorst et al. 2010). However, the physical mechanisms that link the radio brightenings and coronal holes are not yet fully understood. Observations of these brightenings with better spatial resolution and also at different radio wavelengths, such as solar observations with the Atacama Large Millimetric/Submillimetric Array (ALMA) (Wedemeyer et al. 2016) might be fundamental to investigate this association.

---

Acknowledgments

The authors would like to thank the referee for valuable comments and suggestions that helped to improve the manuscript. We also thank Peter Levens for language corrections. We would like to thank the Nobeyama Radioheliograph, which is operated by the NAOJ/Nobeyama Solar Radio Observatory. A.J.O.S. acknowledge the scholarship form CAPES. C.L.S. acknowledge financial support from the São Paulo Research Foundation (FAPESP), grant number 2014/10489-0.

References

All Tables

All Figures