The traditional observational distinction between coronal holes and their surroundings (i.e., active regions and “quiet Sun”) is that coronal holes have the lowest emission in the UV and X-ray. This definition must be amended, however, to exclude filaments (which are often dark when projected against the solar disk) that are cool magnetic structures and not part of the corona (see, e.g., Scholl and Habbal, 2008; Krista and Gallagher, 2009). Coronal hole magnetic fields are known to be more uniform and unipolar than in other regions (see below). The boundaries between coronal holes and surrounding regions are sometimes sharp, sometimes diffuse, and sometimes filled with many small loops (Hudson, 2002).
Coronal holes are more or less indistinguishable from their surroundings in the photosphere and low chromosphere, and usually one cannot see any significant intensity contrast between hole and non-hole regions until the temperature exceeds about 105 K. Spectra, however, can help make the distinction clearer. Teplitskaya et al. (2007) found that central self-reversals in the chromospheric Ca ii H and K lines are noticeably different in coronal holes as opposed to surrounding quiet-Sun regions. A frequently used observational diagnostic of coronal hole boundaries is the He i 10830 Å near-infrared absorption line triplet (Harvey et al., 1975; Harvey and Recely, 2002). At these wavelengths, the absorption is weakest in coronal holes (i.e., the intensity is highest) and spectroheliogram images show the coronal holes quite clearly. It is somewhat counterintuitive that a spectral line of a neutral species (He0) should be sensitive to the properties of the hot corona. However, the atomic level populations that determine the He i 10830 Å source function are unusually responsive to photoionization from UV wavelengths shortward of 504 Å. The overlying solar corona emits these wavelengths in abundance. In coronal hole regions, though, the corona generally has a lower density and temperature, and thus there is less intense UV and X-ray emission to populate the upper levels of the He0 triplet state (see, e.g., Zirin, 1975; Andretta and Jones, 1997; Centeno et al., 2008). This gives rise to reduced absorption. The He i 10830 Å lines are also good probes of supersonic outflow velocities in distant stellar winds (see, e.g., Dupree et al., 2005; Kwan et al., 2007; Sanz-Forcada and Dupree, 2008).
Another observational diagnostic of coronal holes is their elemental abundance signature (e.g., Feldman, 1998; Feldman and Widing, 2003). In the upper chromosphere, transition region, and low corona, holes exhibit abundances very close to those measured in the photosphere. This stands in contrast to other regions (quiet Sun and active regions), which show significant enhancements in the abundances of elements with low first ionization potential (FIP). This selective fractionation is believed to occur in the upper chromosphere, where low-FIP elements become ionized and high-FIP elements remain more neutral. These patterns extend into the heliosphere, where high-speed flows associated with coronal holes are often identifiable from their near-photospheric FIP abundances (von Steiger et al., 1995; Zurbuchen et al., 2002). However, there is still no widespread agreement about the exact physical processes that give rise to this preferential ionization.
The number, sizes, and heliographic locations of coronal holes vary as a function of the solar activity cycle. Large polar holes exist for about 7 years around solar minimum, and are not present for about 3 or 4 years around solar maximum. However, in the declining phase of activity soon after the maximum, it is possible to see the gradual growth of the new-polarity polar coronal holes. This occurs as a number of smaller high-latitude holes eventually collect together at the poles (Timothy et al., 1975; Harvey and Recely, 2002). Taking this growth phase into account, there are only about 1 or 2 years at solar maximum without any distinct high-latitude coronal hole presence. Figure 3 illustrates the growth phase around the peak of solar cycle 23 in 2001. The growth and full “reappearance” of polar coronal holes at this time was described by Miralles et al. (2001a) and McComas et al. (2002). The post-maximum growth of new polar coronal holes lasts about twice as long as their disappearance in the rising phase of the next maximum (Waldmeier, 1981; Fisher and Sime, 1984).
Many low-latitude coronal holes tend to be situated near the edges of magnetically complex active regions. Sometimes active regions emerge within the coronal holes themselves; these have been called “sea anemone” type regions from their spiky, flower-like appearance (Shibata et al., 1994; Asai et al., 2008). Evolving magnetic interactions between active regions and coronal holes have been studied both as a means of enhancing the mass flux of the associated solar wind on nearby open flux tubes (e.g., Habbal et al., 2008; Wang et al., 2009) and as a possible explanation for the nearly rigid rotation of coronal holes (Wang et al., 1996; Antiochos et al., 2007). A burst of emerging magnetic flux in one of these active regions may give rise to new systems of loop connections in the area bordering the coronal hole, and thus cause the coronal hole to decrease in size. This kind of rapid topological evolution of the magnetic field may be relevant in explaining extreme space weather events such as “the day the solar wind disappeared” (i.e., the dramatic drop in the in situ density seen on 11 May 1999; see Janardhan et al., 2008).
Photospheric magnetograms show that coronal holes are more unipolar than other regions; i.e., they have a larger degree of magnetic flux imbalance between the two polarities (Levine, 1982; Wang, 2009). For the large polar coronal holes, this appears to be the long-term outcome of the decay of active-region magnetic fields and their eventual diffusion up to the poles. The unipolar nature of coronal holes is likely to be related to their connection with open-field solar wind streams. As described in Section 2, one reason why coronal holes are dark is that the solar wind carries away both mass and energy, leaving a lower density and pressure. In addition, Abramenko et al. (2006a) and Hagenaar et al. (2008) found that the local rate of emergence of small-scale magnetic flux (mostly in the form of “ephemeral” small-scale bipoles) is substantially lower in unipolar regions than in more mixed or balanced regions of positive and negative magnetic polarity. In most theories of coronal heating of closed loops, the total flux and the overall complexity of the field both drive the total heating rate. Thus, the lower emergence rate of new flux elements in coronal holes may be another factor determining why they have lower densities and pressures (i.e., less coronal heating; see, e.g., Rosner et al., 1978) and thus why they are dark.
Over the past few decades, the magnetic connection between coronal holes and the fast solar wind has been traced down to ever smaller spatial scales. We now can say with some certainty that much of the plasma that eventually becomes the time-steady solar wind seems to originate in thin magnetic flux tubes (with observed sizes of order 50 – 200 km) observed mainly in the dark lanes between the 1000 km size photospheric granulation cells. These strong-field (1 – 2 kG) flux tubes have been called G-band bright points and network bright points, and groups of them have been sometimes termed “solar filigree” (e.g., Dunn and Zirker, 1973; Spruit, 1984; Berger and Title, 2001; Tsuneta et al., 2008). These flux tubes are concentrated most densely in the supergranular network (i.e., the bright lanes between the larger 30,000 km size supergranulation cells). Somewhere in the low chromosphere, the thin flux tubes expand laterally to the point where they merge with one another into a more-or-less homogeneous network field distribution of order 100 G. This merged (mainly vertically oriented) field is associated mainly with the lanes and vertices between supergranular cells. Because this field does not fill the entire coronal volume, it is still susceptible to an additional stage of lateral expansion and weakening. Thus, at a larger height in the chromosphere, these network flux bundles are believed to undergo further broadening into so-called “funnels” (Gabriel, 1976; Dowdy et al., 1986). However, it is still unclear to what extent the closed fields in the supergranular cell centers affect the canopy-like regions between funnels (e.g., Schrijver and Title, 2003). Figure 4 illustrates the successive merging of unipolar field in coronal holes.
Observations of blueward Doppler shifts in supergranular network lanes and vertices, especially in large coronal holes, may be evidence for either the “launching” of the solar wind itself or for upward-going waves that are linked to wind acceleration processes (e.g., Hassler et al., 1999; Peter and Judge, 1999; Aiouaz et al., 2005; Tu et al., 2005). These measurements are consistent with several models of the dynamic, multi-species solar wind in superradially expanding funnels (Byhring et al., 2008; Marsch et al., 2008). However, this interpretation of the data is still not definitive, because there are other observational diagnostics that have shown more of a blueshift in the supergranular cell-centers between funnels (e.g., He i 10830 Å; Dupree et al., 1996; Malanushenko and Jones, 2004). There may be subtle radiative transfer effects that preferentially brighten regions of upflow or downflow (see, e.g., Chae et al., 1997; Avrett, 1999), and these may need to be taken into account in order to understand the meaning of the measured Doppler shifts in these regions.
Lastly, it is important to mention the phenomenon of transient coronal holes (sometimes known as “coronal dimmings”). These are rapid reductions in the UV and X-ray intensity in discrete patches that appear to be spatially and temporally correlated with the liftoff of CME plasma (e.g., Rust, 1983; Thompson et al., 2000; Yang et al., 2008). Spectroscopically, these regions are seen to exhibit similar characteristics as normal coronal holes, including Doppler blueshifts (Harra et al., 2007) and large amplitudes of nonthermal wave broadening (McIntosh, 2009). UV coronal dimmings are beginning to be used as diagnostics for the amount of plasma released – i.e., the total mass – in the CME event (e.g., Aschwanden et al., 2009). It should be noted that transient coronal holes represent just one kind of observed dimming that is associated with time-dependent flare/CME activity; there are other types of dimmings that do not resemble coronal holes (Hudson, 2002).
This work is licensed under a Creative Commons License.