EDP Sciences logo
Subscriber Authentication Point
Search
Menu
Home All issues Volume 656 (December 2021) A&A, 656 (2021) L4 Full HTML
Issue
A&A
Volume 656, December 2021
Solar Orbiter First Results (Cruise Phase)
Article Number L4
Number of page(s) 11
Section Letters to the Editor
DOI https://doi.org/10.1051/0004-6361/202140380
Published online 14 December 2021

© ESO 2021

1. Introduction

The Extreme Ultraviolet Imager (EUI, Rochus et al. 2020) on board the Solar Orbiter (Müller et al. 2020) consists of three telescopes: the Full Sun Imager (FSI), which operates in two extreme UV bandpasses (17.4 nm, and 30.4 nm), and two High Resolution Imagers (HRIs).

The telescope labelled ‘HRIEUV’ has a bandpass centred at 17.4 nm and is dominated by Fe IX and Fe X emission at 1 MK. The telescope labelled ‘HRILya’ has a bandpass centred at 121.6 nm and is dominated by the Lyman-α line of hydrogen. The scientific goal of the two HRIs is to observe the fine structure and dynamics, respectively, in the upper chromosphere (HRILya) and in the low corona (HRIEUV). The transition region between these two layers is where the temperature abruptly increases by two orders of magnitude and where the replenishment of energy and mass that sustains the higher corona and, hence, the whole heliosphere, takes place. Its physics is very complex and it is clear that the crucial processes responsible for the formation of the corona must be operating at very small spatial and temporal scales at all times (Parker 1988; Klimchuk 2015).

The classical picture of the transition region is that of a thermal interface between the coronal and chromospheric plasma (Gabriel 1976). However, it cannot fully describe the high radiance of the transition region lines, so the concept of the “unresolved fine structure” (UFS) of the transition region in the quiet Sun has been proposed (see Feldman 1983; Dowdy et al. 1986): a plethora of small-scale low-lying loops that are magnetically isolated from the corona. Such structures were later detected with the improvement of the spatial resolution of the observations (e.g. Feldman et al. 1999; Warren & Winebarger 2000). More recently, Hansteen et al. (2014) used the data from the Interface Region Imaging Spectrograph (IRIS, De Pontieu et al. 2014, with a spatial resolution of 0.4″, 290 km), and found that low-lying loops (projected full lengths of 4–12 Mm) at low transition-region temperatures (log T around 4.3–4.8) to be a major constituent of the UFS. The High Resolution Coronal Imager (Hi-C) sounding rocket (Kobayashi et al. 2014; Rachmeler et al. 2019) reached the highest resolution in extreme UV coronal observations so far (0.3″, 218 km) and detected fine scale features of less than 200 km wide (Peter et al. 2013; Barczynski et al. 2017). However, thus far, Hi-C has only observed active regions, so the fine structure of the upper transition region of the quiet Sun remains unexplored.

Earlier studies of the quiet Sun revealed a wide range of extreme UV brightenings at transition region and low coronal temperatures, often interpreted in terms of nanoflare heating. We use the term ‘nanoflare’ to refer to the energy range, regardless of the driving mechanism. For coronal heating purposes, an increasing number of miniature heating events are required at scales that are progressively smaller in length, characterised by a power law distribution with an index steeper than −2 (Hudson 1991; Berghmans 2002; Joulin et al. 2016). A power-law distribution was found for the number of events as a function of their energy content (Benz & Krucker 2002; Aschwanden 2004). A power-law index close to −2 was found by Berghmans et al. (1998), whereas a power-law index of −2.5 was found by Harra et al. (2000). However, Aschwanden & Parnell (2002) found a power law of index −1.54, which matches the distribution of active region transient brightenings (Shimizu et al. 1994) and hard X-ray flares (Crosby et al. 1993). The authors explained the difference in terms of a temperature bias imposed by narrow band extreme UV imagers. Recently, Chitta et al. (2021) analysed ‘EUV bursts’ and reached the conclusion that at the resolution of the images of the AIA instrument on board the Solar Dynamics Observatory (SDO, Pesnell et al. 2012), there was, by a factor of 100, an insufficient number of events to account for the energy that is required to heat the corona. At transition region temperatures, explosive events are seen in spectroscopic data with lifetimes of order of minutes and sizes below 1 mm (Brueckner & Bartoe 1983; Dere et al. 1989; Teriaca et al. 2004).

This work uses the first release of properly calibrated and formatted EUI data in December 20201. Before that date, only uncalibrated, quick-look EUI data were publicly available for outreach events, such as the press release of the European Space Agency in July 2020 featuring the EUI team’s first presentation of ‘campfires’. The objective of this Letter is to present novel observations of the EUI/HRIs (Sect. 2) and to discuss what they may contribute to our understanding of coronal structuring and heating (Sect. 3).

2. Observations

As part of the instrument commissioning, the two EUI HRI telescopes were operated on 30 May 2020, for the first time in parallel, at a fast imaging cadence of 5 s and acquired 50 images each. Solar Orbiter was located at 31.5° west in solar longitude from the Earth-Sun line (see Fig. 1). In this study, we also use the data acquired by the AIA instrument (Lemen et al. 2012) on board the SDO (Pesnell et al. 2012) in Earth orbit. To compare EUI and AIA across the whole field of view (FOV), we remapped the images from both instruments (see Appendix A) to a regular 2400 × 2400 grid in Carrington coordinates with a 0.01625 heliographic degrees (197 km) pitch, which corresponds to the original HRIEUV resolution. Appendix A contains the details of the acquired data set.

thumbnail Fig. 1.

Overview of the solar corona of 30 May 2020 as seen by EUI on board Solar Orbiter. Left: location of the HRI field of view is shown as a white square on an EUI/FSI 17.4 nm image. Stonyhurst coordinates are shown, with the crossing of the two yellow great circles indicating the Earth/SDO sub-solar point. We note that the spacecraft roll angle is aligned with the ‘north’ direction of the orbital plane of the Solar Orbiter – and not with the solar meridional plane. The image visualisation is from JHelioviewer (Müller et al. 2017). Middle: first image of the HRIEUV sequence. Right: sub-field corresponding to the red square indicated in the left two panels. The brightest features correspond to the “campfires” studied in this paper.

In Fig. 1, we show the HRIEUV field of view. The subject of the present study are the small-scale brightenings appearing in HRIEUV images with sizes down to the two-pixel resolution limit of the telescope. These brightenings have become known as ‘campfires’, a term which we use for ‘small brightenings observed by HRIEUV’ without necessarily implying that these are physically different from previously observed small extreme UV brightenings in the corona.

We used an automated detection method to determine the statistical significance of small features spatially and track them in time through the sequence. The detection scheme (see Appendix B) applied to the stack of 50 Carrington projected HRIEUV images revealed 1467 campfires. Dividing this number by the area of the field of view and by the duration of our sequence, we arrive at a birthrate of 3.7 × 10−17 m−2 s−1, which is an order of magnitude lower than the birthrate of explosive events (Teriaca et al. 2004 found 4.1 × 10−16 m−2 s−1).

Figure 2 shows the histograms of the projected area, duration, and total intensity of these events. We note that we did not remove the left-most bins of the histograms in Fig. 2 that correspond to events that appear only in one image or only in one pixel. Since these bins are in agreement with the trend seen at larger scales, we kept them in the figure as an illustration that the campfires population extends consistently down to the limitations of our data set in terms of spatial and temporal resolution. Individual events of one pixel in a single frame cannot be confirmed, but some one-pixel events can last several frames and some one-frame events consist of several pixels. Out of the 1467 events, there are 558 which have both their projected area larger than one pixel and their duration larger than five sec (the imaging cadence).

thumbnail Fig. 2.

Observed event distribution as a function of projected area, duration, and total intensity. Projected area refers to all the pixel area that the events occupy at some stage over their duration. The probability density is scaled such that when integrated over all bin widths, the total event count of 1467 is obtained. The smallest areas and durations (only one pixel or one exposure) are displayed as bars without filling.

The detected population is limited by the total duration of our HRIEUV data set, which is only 245 s. The longest event durations in our data set approaches 200 s, but among these, many events are truncated at the end or the start of the sequence. There is no such artificial truncation in the histogram of the projected area, where the largest events are on the order of a few Mm2, which corresponds to roughly one hundred pixels. Strikingly, both the histogram of projected area and the histogram of campfire duration show a power law-like behaviour stretching across at least two orders of magnitude.

Many events have an elongated nature, reminiscent of small loop-like structure. Linear (projected) lengths of these elongated structures range from 1 to 20 pixels (0.2 mm to 4 mm) with aspect ratios distributed mostly between 1 and 5. The smallest observed structures are similar to the smallest active region loops observed by Hi-C (Peter et al. 2013).

As expected from the histograms in Fig. 2, most campfires are very close to the HRIEUV resolution limit and are therefore either dot-like (Fig. 3) or short loop-like (Fig. 4). In most cases, these small events can be recognised in the SDO/AIA bandpasses at 21.1 nm, 19.3 nm, 17.1 nm, and 30.4 nm, although they typically appear weaker and fuzzier, which can be expected from the difference in spatial resolution. No clear event signatures were identified in any of the other SDO/AIA bandpasses. In Appendix C, we show a few more examples, including bigger events with internal structure, including possibly interacting loops.

thumbnail Fig. 3.

Example of a dot-like event. The white square in the two top-left panels corresponds to the field of view of the other, zoomed-in panels.

thumbnail Fig. 4.

Example of a loop-like event. The white square in the two top-left panels corresponds to the field of view of the other, zoomed-in panels.

Remarkably, the observed campfires do not unambiguously reveal localised brightenings in HRILya at the same location and time as the campfires detected in HRIEUV. It is possible that the HRILya signature is a more fuzzy, distributed brightening of a bigger neighbourhood. This will be studied in a future work. In any case, the HRIEUV campfires have a clear tendency to outline the chromospheric network as shown by HRILya. In the top panel of Fig. 5, we show the location of the campfires detected in the HRIEUV data, over-plotted as yellow dots on a reference HRILya image, which was obtained by averaging over the 50 Carrington projected HRILya images.

thumbnail Fig. 5.

Location of the HRIEUV campfires in the HRILya images. Top: averaged intensity map of 50 HRILya images. Detected ‘campfire’ events are displayed in yellow. The contours correspond to signals of 500 DN/s (brown) and 850 DN/s (blue). Bottom: intensity distribution of the HRILya averaged map (solid histogram). The number of detected ‘campfire’ events present in each intensity bin is represented with crosses. Vertical dashed lines embed signals between 500 DN/s (brown) and 850 DN/s (blue), where the majority of the detected events are present. Both y-axes are displayed in logarithmic scale.

The intensity distribution of the averaged HRILya map is shown in the bottom panel of Fig. 5 (solid histogram). The number of detection events falling within each of the 50 equal intensity bins is displayed with crosses. The vertical dashed lines are arbitrarily chosen intensity levels encompassing the peak of the event distribution (crosses) and are used as contours on the intensity map (top panel). The brown and blue colours correspond in both panels to the same intensity levels. The detected events are generally not in the darker cell-centres and there are also very few of them at the brightest locations. The majority of the campfires are located at the network boundaries, which is the same location as where explosive events are found (Porter & Dere 1991). We note that in contrast to the SDO/AIA images, the HRILya images are taken from exactly the same perspective as HRIEUV. Projection effects are therefore not likely to have a major influence on the apparent location of HRIEUV events in the HRILya images.

In order to determine the height of campfires above the photospheric surface, we performed triangulation on sixteen small campfires that could be identified on both HRIEUV 17.4 nm and AIA 17.1 nm images. Figure 6 shows the height above the photosphere versus the projected length for 16 campfires selected for an easy identification. The projected length of each campfire is calculated as the distance between the farthest two pixels of the campfire in the image plane. The photospheric radius value of 695 700 km (Prša et al. 2016) is subtracted to obtain the height above the photosphere. The error bars are derived in Appendix D. From Fig. 6, we can see that the campfires are located between 1000 km and 5000 km above the photospheric surface.

thumbnail Fig. 6.

Height above the photosphere of the 16 campfires versus their projected length (red triangles with their error bars, h ± δh). The red line is a linear fit to the data points. The black line represents the height as half of the projected length as one would expect for a semi-circular loop, for comparison.

The 3D geometry of campfires is unclear, as many of them are barely resolved. Various complex geometries could be envisaged, but here we start with the simplest assumption, namely, that campfires are semi-circular loops. A semi-circular loop has an apex at height H which corresponds to half the length between the loop foot points L. The black H = L/2 line in Fig. 6 is, by the definition of the apex, the upper limit for the height of any point on a semi-circular loop. For our 16 campfires, we triangulated the brightest pixel or the mid-point of a few bright pixels if there are several adjacent pixels of comparable brightness. Figure 6 shows that the resulting campfire heights are systematically higher than the H = L/2 line. Figure 7 shows possible interpretations. A semi-circular loop (a) clearly does not explain the height observations. If campfires represent small-scale loops, then they are either (b) elongated (tall) loops or (c) only their apexes are visible. The latter possibility seems more likely as we do not expect the 17.4 nm emission to extend all the way down to the photosphere. Such a scenario is conceivable when, for instance, the observed campfire corresponds to emission near a contact point for nearby loops (d).

thumbnail Fig. 7.

Scenarios to interpret the campfire heights shown in Fig. 6. The yellow solid lines are the observed campfires, the dotted lines are the invisible loop segments, and the white cross is the triangulation point.

The temperatures of the events were estimated through differential emission measure (DEM) analysis, using the methods of Hannah & Kontar (2012) and Cheung et al. (2015). Here, we show only the result obtained by the former method. The temperatures inferred this way are essentially the same as for the other method. In Fig. 8, we plot the 2D histogram of the total emission measure (EM) for each pixel (integrated over temperature) and the DEM-weighted temperature, as defined in Parenti et al. (2017). We exclude campfires with only one pixel or lasting only one frame. We also correct for the shift in location caused by the different viewpoints by interpolating each event to the location of its intensity barycentre. The temperature distribution peaks at log T ≈ 6.1, with a full width half maximum (FWHM) of 5.43 (log K) and with a relatively long tail extending to 2 MK. While this strongly indicates that campfires are coronal events, namely, that the plasma in these brightenings reaches temperatures of 1 MK, it should be noted that campfires may not always dominate the EM along the line of sight (LOS). Possibly, the DEM-weighted temperature is not always a good representative of the brightening temperature. Nonetheless, we note that both the emission measure and the temperature distribution of the events (but in particular the former) are clearly shifted towards higher values with respect to the background.

thumbnail Fig. 8.

2D histogram of the total emission measure and DEM-weighted temperature for campfire pixels (red) and all pixels (black). The colour is given according to the logarithm of the number of pixels. The number of campfire pixels and all pixels are roughly 9800 and 1.1 × 108, respectively.

3. Summary and discussion

In a first high-cadence data set from the high-resolution telescopes of EUI, we identified a large number of extreme UV brightenings in the quiet Sun that are also referred to as ‘campfires’. To our knowledge, this is the highest resolution image sequence of the quiet Sun corona ever achieved. Given the favourable 31.5° separation in heliocentric longitude with SDO/AIA in Earth orbit, we were also able to apply triangulation techniques at the highest resolution ever to determine the height of campfires. The characteristics of these campfires are summarised below.

An increasing number of campfires are found at progressively smaller scales, down to the resolution of the present HRIEUV data set. We find an apparent power law distributions for the campfire areas, durations, and total intensity. The area distribution is limited at the lower end by spatial resolution (a 2-pixel area is 0.08 mm2). At the higher end, the area distribution ends around 5 mm2, which is a consequence of our wavelet based detection scheme. The duration distribution is limited at the lower end by the temporal resolution of the present data set (twice the image cadence is 10 s). At the higher end, the duration distribution ends around 200 s, which is close to the total duration of our data set, namely 245 s. The appearance of the campfires at progressively smaller scales ranges from small systems of possible interacting loops to individual loops and, finally, to dots at the resolution limit of the data set. The aspect ratios of length versus width range from 1 to 5.

The observed campfires observed in HRIEUV do not, obviously, have corresponding localised brightenings in HRILya at the same time and place. However, the events detected in HRIEUV do line up along the chromospheric network as observed by HRILya. A comparison with simultaneous AIA images shows that most events can be found back in 21.1 nm, 19.3 nm, 17.1 nm, and 30.4 nm as well, however, as compared to HRIEUV, they appear weaker and fuzzier, which is as expected given the difference in resolution. In addition, they are not seen in ‘hotter’ AIA bandpasses.

Using triangulation techniques, we determined for a sub-sample of events that the campfires are located between 1000 km and 5000 km above the photosphere. Campfires are events at coronal temperatures: the distribution of their DEM-weighted temperatures peaks at log T ≈ 6.12, with a FWHM of 0.074 (log K).

Using Hi-C data, Tiwari et al. (2019) reported energy-release events that resemble some of the bigger campfires observed by HRIEUV, but they only analysed active region observations. In the quiet Sun, the smallest extreme UV brightenings have been observed with SDO/AIA (Joulin et al. 2016; Chitta et al. 2021). The superior spatial and temporal resolution of the studied HRIEUV quiet Sun data set allows us to extend the family of extreme UV quiet Sun brightenings down to spatial scales less than half a Mm and down to temporal scales of 10 s.

Campfires are located above the chromospheric network, where the strongest quiet Sun magnetic fields reside. This suggests that campfires are magnetically defined features, as is further supported by their often isolated, distinct loop-like appearance. While the campfires do reach coronal temperatures (1–1.6 MK), they are located at remarkably low heights, between 1000 km and 5000 km from the photosphere. We believe it is the first time such low heights of quiet sun structures emitting around 17.4 nm have been determined. Compared to their lengths, the campfires are systematically higher than what could be expected for semi-circular loops. It is either the case that campfires are vertically elongated loops or that campfires correspond only to loop tops that are brightening. We propose that campfires are the apexes of low-lying small-scale network loops that are heated to coronal temperatures.

The observed heights of campfires at coronal temperatures coincide with the heights of the cold transition region loops reported by Hansteen et al. (2014). This co-existence is coherent with the picture of a plethora of small-scale low lying loops existing at different temperatures (Feldman 1983; Dowdy et al. 1986). Future research will address the reasons for which the campfire loops stand out as particularly hot when compared to neighbouring loops that do not surpass transition region temperatures.

The bigger campfires have durations and sizes that are comparable to explosive events (Dere et al. 1989), even if the latter ones are rather observed at transition region temperatures; see also Peter et al. (2014). In addition, they both appear at network boundaries. Campfires are possibly an order of magnitude less frequent than explosive events, although this might be a detection bias as explosive events are spectroscopically identified. Some of the AIA 17.1 nm quiet Sun small brightenings studied by Innes & Teriaca (2013) are clearly associated with spectral profiles that are typical of explosive events. Due to the lack of spectroscopic data, this aspect, unfortunately, could not be checked for the campfire population.

The current observations were taken at a distance of 0.556 AU from the Sun. Future perihelia of Solar Orbiter will go down to below 0.3 AU and increase the effective spatial resolution of HRIEUV by another factor of 2. Joint campaigns with other remote sensing instruments on board the Solar Orbiter, such as the Polarimetric and Helioseismic Imager (PHI, Solanki et al. 2020) and the SPICE spectrograph (Spice Consortium 2020), will help to confirm or discard our interpretation. This is particularly important for determining the driver of these heating events, their link to explosive events, and their role in coronal heating.

Acknowledgments

The building of EUI was the work of more than 150 people during more than 10 years, we gratefully acknowledge all the efforts that have led to a successfully operating instrument. The authors thank the Belgian Federal Science Policy Office (BELSPO) for the provision of financial support in the framework of the PRODEX Programme of the European Space Agency (ESA) under contract number 4000112292. The French contribution to the EUI instrument was funded by the French Centre National d’Etudes Spatiales (CNES); the UK Space Agency (UKSA); the Deutsche Zentrum für Luft- und Raumfahrt e.V. (DLR); and the Swiss Space Office (SSO). PA and DML acknowledge funding from STFC Ernest Rutherford Fellowships No. ST/R004285/2 and ST/R003246/1, respectively. SP acknowledges the funding by CNES through the MEDOC data and operations center.

References

  1. Aschwanden, M. J. 2004, Physics of the Solar Corona An Introduction (Praxis Publishing Ltd.) [Google Scholar]
  2. Aschwanden, M. J., & Parnell, C. E. 2002, ApJ, 572, 1048 [NASA ADS] [CrossRef] [Google Scholar]
  3. Barczynski, K., Peter, H., & Savage, S. L. 2017, A&A, 599, A137 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  4. Benz, A. O., & Krucker, S. 2002, ApJ, 568, 413 [Google Scholar]
  5. Berghmans, D. 2002, in Solar Variability: From Core to Outer Frontiers, ed. A. Wilson, ESA Spec. Publ., 2, 501 [NASA ADS] [Google Scholar]
  6. Berghmans, D., Clette, F., & Moses, D. 1998, A&A, 336, 1039 [NASA ADS] [Google Scholar]
  7. Brueckner, G. E., & Bartoe, J. D. F. 1983, ApJ, 272, 329 [Google Scholar]
  8. Cheung, M. C. M., Boerner, P., Schrijver, C. J., et al. 2015, ApJ, 807, 143 [Google Scholar]
  9. Chitta, L. P., Peter, H., & Young, P. R. 2021, A&A, 647, A159 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  10. Crosby, N. B., Aschwanden, M. J., & Dennis, B. R. 1993, Sol. Phys., 143, 275 [NASA ADS] [CrossRef] [Google Scholar]
  11. De Pontieu, B., Title, A. M., Lemen, J. R., et al. 2014, Sol. Phys., 289, 2733 [Google Scholar]
  12. Dere, K. P., Bartoe, J. D. F., & Brueckner, G. E. 1989, Sol. Phys., 123, 41 [Google Scholar]
  13. Dowdy, J. F., Rabin, D., & Moore, R. L. 1986, Sol. Phys., 105, 35 [NASA ADS] [CrossRef] [Google Scholar]
  14. Feldman, U. 1983, ApJ, 275, 367 [CrossRef] [Google Scholar]
  15. Feldman, U., Widing, K. G., & Warren, H. P. 1999, ApJ, 522, 1133 [NASA ADS] [CrossRef] [Google Scholar]
  16. Gabriel, A. H. 1976, Philos. Trans. R. Soc. London Ser. A, 281, 339 [NASA ADS] [CrossRef] [Google Scholar]
  17. Hannah, I. G., & Kontar, E. P. 2012, A&A, 539, A146 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  18. Hansteen, V., De Pontieu, B., Carlsson, M., et al. 2014, Science, 346, 1255757 [Google Scholar]
  19. Harra, L. K., Gallagher, P. T., & Phillips, K. J. H. 2000, A&A, 362, 371 [NASA ADS] [Google Scholar]
  20. Hudson, H. S. 1991, Sol. Phys., 133, 357 [Google Scholar]
  21. Inhester, B. 2006, ArXiv e-prints [arXiv: astro-ph/0612649] [Google Scholar]
  22. Innes, D. E., & Teriaca, L. 2013, Sol. Phys., 282, 453 [CrossRef] [Google Scholar]
  23. Joulin, V., Buchlin, E., Solomon, J., & Guennou, C. 2016, A&A, 591, A148 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  24. Klimchuk, J. A. 2015, Philos. Trans. R. Soc. London Ser. A, 373, 20140256 [NASA ADS] [Google Scholar]
  25. Kobayashi, K., Cirtain, J., Winebarger, A. R., et al. 2014, Sol. Phys., 289, 4393 [Google Scholar]
  26. Lemen, J. R., Title, A. M., Akin, D. J., et al. 2012, Sol. Phys., 275, 17 [Google Scholar]
  27. Müller, D., Nicula, B., Felix, S., et al. 2017, A&A, 606, A10 [Google Scholar]
  28. Müller, D., St. Cyr, O. C., Zouganelis, I., et al. 2020, A&A, 642, A1 [Google Scholar]
  29. Murtagh, F., Starck, J. L., & Bijaoui, A. 1995, A&AS, 112, 179 [Google Scholar]
  30. Parenti, S., del Zanna, G., Petralia, A., et al. 2017, ApJ, 846, 25 [CrossRef] [Google Scholar]
  31. Parker, E. N. 1988, ApJ, 330, 474 [Google Scholar]
  32. Pesnell, W. D., Thompson, B. J., & Chamberlin, P. C. 2012, Sol. Phys., 275, 3 [Google Scholar]
  33. Peter, H., Bingert, S., Klimchuk, J. A., et al. 2013, A&A, 556, A104 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  34. Peter, H., Tian, H., Curdt, W., et al. 2014, Science, 346, 1255726 [Google Scholar]
  35. Porter, J. G., & Dere, K. P. 1991, ApJ, 370, 775 [CrossRef] [Google Scholar]
  36. Prša, A., Harmanec, P., Torres, G., et al. 2016, AJ, 152, 41 [Google Scholar]
  37. Rachmeler, L. A., Winebarger, A. R., Savage, S. L., et al. 2019, Sol. Phys., 294, 174 [Google Scholar]
  38. Rochus, P., Auchère, F., Berghmans, D., et al. 2020, A&A, 642, A8 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  39. Shimizu, T., Tsuneta, S., Acton, L. W., et al. 1994, ApJ, 422, 906 [NASA ADS] [CrossRef] [Google Scholar]
  40. Solanki, S. K., del Toro Iniesta, J. C., Woch, J., et al. 2020, A&A, 642, A11 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  41. Spice Consortium (Anderson, M., et al.) 2020, A&A, 642, A14 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  42. Starck, J.-L., & Murtagh, F. 1994, A&A, 288, 342 [NASA ADS] [Google Scholar]
  43. Starck, J. L., & Murtagh, F. 2002, Astronomical Image and Data Analysis (Springer-Verlag) [CrossRef] [Google Scholar]
  44. Teriaca, L., Banerjee, D., Falchi, A., Doyle, J. G., & Madjarska, M. S. 2004, A&A, 427, 1065 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  45. Thompson, W. T. 2006, A&A, 449, 791 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  46. Tiwari, S. K., Panesar, N. K., Moore, R. L., et al. 2019, ApJ, 887, 56 [NASA ADS] [CrossRef] [Google Scholar]
  47. Warren, H. P., & Winebarger, A. R. 2000, ApJ, 535, L63 [NASA ADS] [CrossRef] [Google Scholar]

Appendix A: Data set description

As part of the instrument’s commissioning, the two EUI HRI telescopes were operated, for the first time, in parallel at a fast imaging cadence on 30 May 2020. The Solar Orbiter was located at 31.5° west in solar longitude from the Earth-Sun line (see Fig. 1) and at a distance of 0.556 AU from the Sun. From this vantage point, the HRIEUV angular pixel size of 0.492″ corresponds to 198 km on the solar surface. The HRILya angular pixel size of 0.514″ corresponds to 207 km. HRIEUV effectively achieves a two-pixel resolution of 0.984″, while for HRILya, diffraction from the 30 mm aperture and the type of detector limited the goal to 2″, with an effective resolution estimated now of about 3″. Both HRI cameras produce 2048 × 2048 pixel images that correspond to a 16.8 × 16.8′ FOV.

The data analysed for this study were acquired between 14:54:00 UT and 14:58:20 UT and consist of two sequences of 50 images each, taken simultaneously by the two HRIs at 5 s cadence. The sequence of HRIEUV images started 15 s earlier. HRIEUV images were acquired with 3 s exposure while the HRILya images were exposed for 1 s. The data were acquired with lossless compression and processed to Level 2 FITS files as distributed in the SolO/EUI Data Release 1.02.

We also used the data acquired by the AIA instrument (Lemen et al. 2012) on board the SDO (Pesnell et al. 2012) in Earth orbit. The AIA bandpass at 17.1 nm is similar but not identical to the HRIEUV and FSI bandpasses at 17.4 nm. The detailed differences in bandpasses are not considered here but these will be the subject of future investigations (Gissot & Auchère priv. comm.). The angular pixel sizes of AIA and HRIEUV are comparable, but given the difference in heliocentric distance, EUI provides a factor of 2 improvement in spatial resolution. The difference in heliocentric distance results also in difference in light travel time of 220 s, meaning that EUI images acquired at time T are to be compared with AIA images acquired at T+220 s. The AIA imaging cadence is 12 s, hence, the closest in time AIA and EUI images can be up to 5 s off. We used despiked AIA images. In order to subtract residual thermal signal in faint regions, we subtracted for each quadrant the mean of a 100 × 100 pixels box in the corresponding corner.

To compare EUI and AIA across the whole FOV, we use images from both instruments that are remapped to Carrington coordinates. In this process, the HRI images were re-sampled with bi-cubic interpolation on a regular 2400 × 2400 grid with a 0.0163 heliographic degrees (198 km) pitch, thus preserving the HRIEUV resolution.

No extra data points are created in the conversion from the original 2048 × 2048 to the 2400 × 2400 frame as the extra pixels correspond to a dark border around the actual FOV (see top frame of Fig. 5). Following an optimisation process, the radius of the Carrington projection sphere was chosen at 1.004 R⊙, or 2.8 mm above the photosphere, in order to take out the average parallax over the field of view between the campfires seen in HRIEUV and AIA.

Individual events can still display a parallax between their HRIEUV and AIA images when their height deviates from this average. This can be an explanation for the mismatch seen in Fig. C.5 if this campfire is located exceptionally high above the photosphere.

Appendix B: Automated detection

We use an automated detection method to separate these space-time events from the random intensity fluctuations caused by the various sources of noise in the acquisition chain. In order to determine the statistical significance of a space-time blob of intensity, one must assume a model of the noise present in the data. In the following, we considered only the photon shot noise, which is expected to dominate the other sources of noise in HRIEUV (Rochus et al. 2020). While a detection in space-time directly would be desirable, in the present work we determined the statistical significance of small features spatially and tracked them in time through the sequence. The detection scheme applied to the stack of 50 Carrington projected images was as follows:

  1. Small features are separated from the photon shot noise using the first two scales of a dyadic ‘à trous’ wavelet transform of the images using a B3 spline scaling function (e.g., Starck & Murtagh 2002). In Starck & Murtagh (1994), the wavelet coefficients at each scale are considered significant when they are greater than n times the root-mean-square amplitude expected from Gaussian white noise. In this study, we use the extension to Poisson statistics proposed by Murtagh et al. (1995), with n = 5 sigma thresholding. The thresholding of the wavelet coefficients results in a binary cube.

  2. The 6-connected voxels of the binary cube are clustered into numbered regions. Each region defines an event.

  3. For each event, physically relevant parameters are computed, such as projected area, duration, and the parameters of the ellipse of the same second moments.

The above procedure with 2 wavelet scales and n = 5 sigma thresholding, applied on the sequence of 50 HRIEUV images resulted in 1467 campfires.

Appendix C: Example campfires

We compare the morphology of a few individual campfires as observed in HRIEUV, HRILya and in the SDO/AIA bandpasses at 21.1 nm, 19.3 nm, 17.1 nm, and 30.4 nm. No clear event signatures were identified in any of the other SDO/AIA bandpasses.

The following figures show examples of individual campfires. In each figure, the top row corresponds to HRIEUV and HRILya subfield images, while the panels on the second row show SDO/AIA subfield images, from left to right at respectively 21.1 nm, 19.3 nm, 17.1 nm, and 30.4 nm. Each panel corresponds to nearly (10 mm)2 on the Sun (50 × 50 pixels, each with a footprint on the Sun of (198 km)2, except the two panels top left in each figure that show the (60 mm)2 neighbourhood in which the campfire takes place. The black rectangle shows the maximal extension in the x and y direction over which each campfire is detected over its duration.

As expected from the histograms in Fig. 2, most campfires are very close to the HRIEUV resolution limit. In this section we do not consider events detected as a single pixel or appearing in a single frame. Even then, most campfires appear as essentially dot-like of only a few pixels but their existence is beyond doubt as we can see their evolution over several images. In Figs. C.1, C.2, and 3, we show examples of such dot-like campfires. In each of the cases, there is no corresponding localised brightening in HRILya. A comparison with simultaneous AIA images shows that these smallest events can be identified in the displayed AIA wavelengths as well, although they appear weaker, fuzzier, and typically bigger, which is most likely an effect of the lower spatial resolution. Without the indication from HRIEUV, most of these events would remain unnoticed among noise fluctuations.

thumbnail Fig. C.1.

Example of a dot-like event. The white square in the two top-left panels corresponds to the field of view of the other, zoomed-in panels.

thumbnail Fig. C.2.

Example of a dot-like event. The white square in the two top-left panels corresponds to the field of view of the other, zoomed-in panels.

Amongst the slightly bigger campfires (e.g. Figs. C.3 and 4), we see loop-like events with an aspect ratio of 2 to 4 and a length of up to 10 pixels (2 Mm). Again, there is no corresponding localised brightening in HRILya but the events can be identified in simultaneous AIA images, though they appear weaker and fuzzier, thereby often hiding their elongated structure.

thumbnail Fig. C.3.

Example of a loop-like event. The white square in the two top-left panels corresponds to the field of view of the other, zoomed-in panels.

For the biggest campfires, HRIEUV often shows complex internal structure, which is missing from the corresponding AIA images due to the difference in resolution (Fig. C.4). This internal structure can possibly be interpreted as a group of nearby loops brightening up, either because of some local source in their neighbourhood, or because of mutual interactions. A suggestive example of the latter is the event shown in Fig. C.5. In the evolution of this event (Fig. C.6), we can see that the source location of the event (the black rectangle) later becomes the forking point of what could be interpreted as two interacting loops.

thumbnail Fig. C.4.

Example of complex event. The white square in the two top-left panels corresponds to the field of view of the other, zoomed-in panels.

thumbnail Fig. C.5.

Example of complex event. The white square in the two top-left panels corresponds to the field of view of the other, zoomed-in panels.

thumbnail Fig. C.6.

Time evolution of the event shown in Fig. C.5. The latter corresponds to t = 75 s. Each frame corresponds to an area of 10 × 10 mm and each pixel corresponds to an area of 197 km × 197 km.

Appendix D: Height of campfires

In order to determine the height of campfires above the photospheric surface, we performed triangulation on sixteen small-size campfires that could be identified on both HRIEUV (14:54:10 UT) 17.4 nm and AIA 17.1 nm (14:57:57 UT) images. The difference in time of the two images is chosen to accommodate for the light travel time to the two instruments, so that they image the same solar scene nearly simultaneously (to within 2 s).

The method consists of identifying the same feature (pixel) in the two images (tie-pointing) and calculating the lines of sight that pass through the corresponding pixels in the two images. This allows calculating the position of the intersection point in 3D space (e.g. Inhester 2006). The triangulation was performed using the scc_measure.pro SolarSoft routine, and the location of the campfires in Stonyhurst coordinates (longitude, latitude, and height, h, from the Sun’s centre; see Thompson 2006) was obtained. The half-width triangulation error δh is related to the linear pixel sizes δs1 and δs2 of HRIEUV and AIA, respectively, and can be calculated as follows:

(D.1)

where γ is the separation angle between Solar Orbiter and SDO. On 30 May 2020, γ = 31.5°, one AIA pixel is 441 km and one HRIEUV pixel is 198 km. The error is then δh = 600 km if one can localise a campfire to within one pixel in both HRIEUV and AIA images. This is sometimes difficult to do, so that the error may be up to four HRIEUV and two AIA pixels, which gives an error of 1500 km. This is the random error linked to the limited resolution of the telescopes.

On top of the random error for the height, there is a systematic error due to an uncertainty of a few arcsec on the pointing of the HRIEUV telescope with respect to the centre of the Sun. We estimated that this uncertainty would lead to a systematic error of ±900 km on the derived heights.

All Figures

thumbnail Fig. 1.

Overview of the solar corona of 30 May 2020 as seen by EUI on board Solar Orbiter. Left: location of the HRI field of view is shown as a white square on an EUI/FSI 17.4 nm image. Stonyhurst coordinates are shown, with the crossing of the two yellow great circles indicating the Earth/SDO sub-solar point. We note that the spacecraft roll angle is aligned with the ‘north’ direction of the orbital plane of the Solar Orbiter – and not with the solar meridional plane. The image visualisation is from JHelioviewer (Müller et al. 2017). Middle: first image of the HRIEUV sequence. Right: sub-field corresponding to the red square indicated in the left two panels. The brightest features correspond to the “campfires” studied in this paper.
In the text
thumbnail Fig. 2.

Observed event distribution as a function of projected area, duration, and total intensity. Projected area refers to all the pixel area that the events occupy at some stage over their duration. The probability density is scaled such that when integrated over all bin widths, the total event count of 1467 is obtained. The smallest areas and durations (only one pixel or one exposure) are displayed as bars without filling.
In the text
thumbnail Fig. 3.

Example of a dot-like event. The white square in the two top-left panels corresponds to the field of view of the other, zoomed-in panels.
In the text
thumbnail Fig. 4.

Example of a loop-like event. The white square in the two top-left panels corresponds to the field of view of the other, zoomed-in panels.
In the text
thumbnail Fig. 5.

Location of the HRIEUV campfires in the HRILya images. Top: averaged intensity map of 50 HRILya images. Detected ‘campfire’ events are displayed in yellow. The contours correspond to signals of 500 DN/s (brown) and 850 DN/s (blue). Bottom: intensity distribution of the HRILya averaged map (solid histogram). The number of detected ‘campfire’ events present in each intensity bin is represented with crosses. Vertical dashed lines embed signals between 500 DN/s (brown) and 850 DN/s (blue), where the majority of the detected events are present. Both y-axes are displayed in logarithmic scale.
In the text
thumbnail Fig. 6.

Height above the photosphere of the 16 campfires versus their projected length (red triangles with their error bars, h ± δh). The red line is a linear fit to the data points. The black line represents the height as half of the projected length as one would expect for a semi-circular loop, for comparison.
In the text
thumbnail Fig. 7.

Scenarios to interpret the campfire heights shown in Fig. 6. The yellow solid lines are the observed campfires, the dotted lines are the invisible loop segments, and the white cross is the triangulation point.
In the text
thumbnail Fig. 8.

2D histogram of the total emission measure and DEM-weighted temperature for campfire pixels (red) and all pixels (black). The colour is given according to the logarithm of the number of pixels. The number of campfire pixels and all pixels are roughly 9800 and 1.1 × 108, respectively.
In the text
thumbnail Fig. C.1.

Example of a dot-like event. The white square in the two top-left panels corresponds to the field of view of the other, zoomed-in panels.
In the text
thumbnail Fig. C.2.

Example of a dot-like event. The white square in the two top-left panels corresponds to the field of view of the other, zoomed-in panels.
In the text
thumbnail Fig. C.3.

Example of a loop-like event. The white square in the two top-left panels corresponds to the field of view of the other, zoomed-in panels.
In the text
thumbnail Fig. C.4.

Example of complex event. The white square in the two top-left panels corresponds to the field of view of the other, zoomed-in panels.
In the text
thumbnail Fig. C.5.

Example of complex event. The white square in the two top-left panels corresponds to the field of view of the other, zoomed-in panels.
In the text
thumbnail Fig. C.6.

Time evolution of the event shown in Fig. C.5. The latter corresponds to t = 75 s. Each frame corresponds to an area of 10 × 10 mm and each pixel corresponds to an area of 197 km × 197 km.
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.