Local Helioseismology

3.3.1 Direct solution in plane-parallel models

3.3.2 Normal-mode summation approximation

Figure 3 compares Green’s functions computed numerically and approximated via normal-mode summation. Notice that the exact numerical result shows a discontinuity at the source depth; this comes from the jump across the delta function on the right hand side of Equation (11). The normal mode approximation, being a finite sum of continuous functions, is everywhere continuous.

Open image in new window

3.3.3 Green’s functions for the observable

4.1.1 Wavefield decomposition

4.1.2 Absorption coefficient

The original motivation for the Hankel analysis was to search for wave absorption by sunspots. Braun et al. (1987) defined the absorption coefficient by

$${\alpha _m}(L,\nu) = {{{P_{{\rm{out}}}} - {P_{{\rm{in}}}}} \over {{P_{{\rm{out}}}}}} = 1 - {{\vert {B_m}(L,\nu){\vert ^2}} \over {\vert {A_m}(L,\nu){\vert ^2}}}.$$

(37)

Braun (1995) remarks that this definition of the absorption coefficient may not necessarily represent wave dissipation as it ignores mode mixing. In practice, some averaging must be done to reduce the noise level. For example, Braun et al. (1987) averaged |A _m |² and |B _m |² over all azimuthal orders and over the frequency range 1.5 mHz < ν < 5 mHz. Braun (1995) obtained a reasonable signal-to-noise level only by averaging in frequency space over the width of a ridge (radial order n) at fixed L:

$${\alpha _m}(L,n) = 1 - {{{{\langle {\vert {B_m}(L,\nu){\vert ^2} - {\lambda _m}(L,\nu)} \rangle}_n}} \over {{{\langle {\vert {A_m}(L,\nu){\vert ^2} - {\lambda _m}(L,\nu)} \rangle}_n}}},$$

(38)

where the angle brackets denote a frequency average over the width of the n-th ridge around the frequency of the mode (L, n). The function λ _m (L, ν) is introduced to remove the background power (see Figure 8). Braun (1995) also made averages over all azimuthal orders, to obtain an absorption coefficient for each ridge and each wavenumber denoted by α(L, n). Hankel analysis revealed that sunspots are strong absorbers of incoming p and f modes. As a check, the same analysis applied to quiet-Sun data does not show significant absorption.

Open image in new window

4.1.3 Phase shifts

Braun et al. (1992) and Braun (1995) were successful at measuring the relative phase shift between outgoing and ingoing waves. Phase shifts measurements require a lot of care. Braun et al. (1992) pointed out that the phases of A _m and B _m represent averages over a resolution element with size (ΔL, Δν). For example, Braun (1995) has ΔL = 20 and Δν = 4 µHz.

Let us now consider actual observations. The p modes are distributed along ridges corresponding to different radial orders n, whose frequency positions in the L-ν diagram are specified by known functions ν_n(L). There may be several unresolved modes in the resolution element (ΔL, Δν). The condition θ_max < uT/(2R_⊙) (Section 4.1.1) is equivalent to ΔL > 2Δν/ν _n ′(L), where the prime denotes a derivative. This means that the spread in L of the ridge across the frequency step Δν is much smaller than the wavenumber resolution ΔL. In this case Braun et al. (1992) estimate the spurious phase shift from Equation (42) by letting L equal to the mean wavenumber L of the unresolved modes. Since \({\nu _j} - {\nu _n}({L_i}) \simeq \nu _n^{\prime}({L_i})(\bar L - {L_i})\), the frequency dependence of the spurious shifts is given by (Braun et al., 1992)

$$\Delta {\phi ^{{\rm{spur}}}}({L_i},{\nu _j}) \simeq - {{{\theta _{\max}} + {\theta _{\min}}} \over {\nu _n^{\prime}({L_i})}}[{\nu _j} - {\nu _n}({L_i})].$$

(43)

Figure 9 shows phase shifts measurements for quiet-Sun data (Braun, 1995). The predicted values of the spurious shifts, given by the above equation, agree with the observations. The corrected average phase shift is obtained from the phase of B _m A _m * exp(−iΔφ^spur) averaged over m and some frequency range. As expected for the quiet Sun, the average phase shift is zero once the spurious shift has been subtracted. This control experiment shows, in particular, that the values of ν _n (L _i ) and (L _i ) interpolated from modern p-mode frequency tables are precise enough for the analysis.

Open image in new window

4.1.4 Mode mixing

4.2.1 Local power spectra

Ring-diagram analysis is a powerful tool to infer the speed and direction of horizontal flows below the solar surface by observing the Doppler shifts of ambient acoustic waves from power spectra of solar oscillations computed over patches of the solar surface (typically 15° × 15°). Thus ring analysis is a generalization of global helioseismology applied to local areas on the Sun (as opposed to half of the Sun).

The ring diagram method is based on the computation and fitting of local k − ω power spectra and was first introduced by Hill (1988). A small patch is tracked as it moves across the disk. In this process, images are remapped onto a projection grid, such as the cylindrical equal-area projection, Postel’s azimuthal equidistant projection, or the transverse cylindrical equidistant projection (see, e.g., Corbard et al., 2003). A series of tracked images form a data cube, i.e., the surface Doppler velocity as a function of the two spatial coordinates and time Φ(x, t). The data cube is Fourier transformed to obtain Φ(k, ω), where k is the horizontal wavevector and ω is the angular frequency (approximately a plane-wave decomposition). The three-dimensional power spectrum of the resulting data cube, P(k, ω) = |Φ(k, ω)|², yields the basic input data for ring diagrams. As in the global power spectrum, the main features are the ridges corresponding to the normal modes of the Sun. As there are two wavenumber directions, the ridges now appear as rings when seen in a cut through the spectrum at constant frequency (Figure 10). Flows in regions, over which the power spectrum is computed, introduce Doppler shifts in the oscillation spectrum, and changes in sound speed alter the locations of the rings. Thus by fitting the positions and shapes of the rings in the power spectrum the subsurface flows and sound-speed can be estimated.

Open image in new window

4.2.2 Measurement procedure

Other fitting techniques are described by Patrón et al. (1997) and Basu et al. (1999). Essentially all ring-diagram fitting has been done assuming symmetric profiles for the ridges in the power spectrum. Basu and Antia (1999) concluded that including line asymmetry in the model power spectrum improves the fits, but does not substantially change the inferred flows.

The most important parameters used in ring-diagram analysis are the flow parameters U _x and U _y , and the central frequency ω₀. For each radial order n and wavenumber k (or spherical harmonical degree l) corresponds a new set of values: It is convenient to use the notations \(U_{x}^{nl},U_{y}^{nl}\) for the flow parameters, and δw _nl for the difference between the fitted mode frequency and the mode frequency calculated from a standard solar model.

Open image in new window

4.2.3 Depth inversions

The goal of inverting ring fit parameters is to determine the sound speed, density, and mass flows in the region underneath the tile over which the local power spectrum was computed. Generally, the assumption is that the sound speed, density, and flows are only functions of depth within the region covered by a particular tile. The forward problem for ring diagrams has traditionally been done by analogy with global mode helioseismology.

The RLS method (see, e.g., Hansen, 1998; Larsen, 1998) selects the model that minimizes a combination of the misfit between the observed data and the data predicted by the model and some function of the model, for example the integral of the square of the second derivative. In particular RLS minimizes a function of the form

$${\chi ^2} = \sum\limits_i {{1 \over {\sigma _i^2}}} {\left[ {{d_i} - \int {dz} \;{K^i}(z)f(z)} \right]^2} + \lambda {\cal R}f.$$

(51)

Here we have assumed a diagonal error covariance matrix with σ _i being the standard deviation of the error on the measurement d _i . The regularization operator ℛ takes the function f(z) and returns a scalar. The regularization parameter which controls the relative importance of minimizing the misfit to the data and smoothing the solution is λ. Example resolution kernels for RLS inversion are plotted in Figure 12.

Open image in new window

4.3.1 Fourier filtering

The first operation in time-distance helioseismology is to track Doppler images at a constant angular velocity to remove the main component of solar rotation, as is done in ring-diagram analysis (see Section 4.2.1). Postel’s azimuthal equidistant projection is often used in time-distance helioseismology. The resulting data cube is Fourier transformed to obtain Φ(k, ω).

A filtering procedure is then applied to the data (Duvall Jr et al., 1997). First, frequencies below 1.5 mHz are filtered out in order to remove granulation and supergranulation noise. The data are further filtered to select parts of the wave propagation diagram.

A separate filtering procedure is applied for surface gravity waves, which are used to probe the near surface layer. In this case the filter function F (k, w) is 1 if \(\vert\omega\pm\sqrt{gk}\vert < kU_{\rm cut}\) and 0 otherwise. The parameter U_cut controls the width of the region around the f-mode dispersion relation ω² = gk. A reasonable choice is U_cut = 1 km s⁻¹. This value allows for large Doppler frequency shifts introduced by flows, and does not let the p₁ ridge through.

There is some freedom in the selection of the Fourier filters. For example one may construct filters that depend explicitly on the direction of the wavevector k, and not simply on the wavenumber (see, e.g., Giles, 1999).

4.3.2 Cross-covariance functions

The cross-correlation is useful as it is a phase-coherent average of inherently random oscillations. It can be seen as a solar seismogram, providing information about travel times, amplitudes, and the shape of the wave packets traveling between any two points on the solar surface. Figure 13 shows a theoretical cross-covariance Δ as a function of distance between x₁ and x₂, and time lag t. This time-distance diagram was computed for a spherically symmetric solar model (see Sekii and Shibahashi, 2003). The first ridge corresponds to acoustic waves propagating between the two points without additional reflection from the solar surface. The next ridge corresponds to waves which arrive after one reflection from the surface, and the ridges at greater time delays result from waves arriving after multiple bounces. The backward branch associated with the second ridge corresponds to waves reflected on the far-side of the Sun. In most applications, only the direct (first-bounce) travel times are measured. Figure 14 shows the p-mode cross-covariance function at short distances, computed with the Fourier filters given in Table 1. In this example, cross-covariance functions were averaged over pairs of points at fixed distance to reduce noise. An example cross-covariance function for the f-mode ridge is plotted in Figure 15.

Open image in new window

Open image in new window

Open image in new window

The travel times of the wave packets are measured from the (first-bounce) cross-covariance function. Local inhomogeneities in the Sun will affect travel times differently depending on the type of perturbation. For example, temperature perturbations and flow perturbations have very different signatures. Given two points x₁ and x₂ on the solar surface the travel time perturbation due to a temperature anomaly is, in general, independent of the direction of propagation between x₁ and x₂. However, a flow with a component directed along the direction x₁ → x₂ will break the symmetry in travel time for waves propagating in opposite directions: Waves move faster along the flow than against it. Magnetic fields introduce a wave speed anisotropy and will have yet another travel time signature (this has not been detected yet).

4.3.3 Travel time measurements

Cross-covariance functions can have a large amount of realization noise due to the schochastic nature of solar oscillations, and it has proved difficult to measure travel times between two individual pixels on the solar surface. Some averaging is often required. The time over which the cross-covariances are computed is usually T ≥ 8 hr (this puts a limit on the temporal scale of the solar phenomena that can be studied). In order to further enhance the signal-to-noise ratio, Duvall Jr et al. (1993) and Duvall Jr et al. (1997) suggested to average C(x₁, x₂, t) over points x₂ that belong to an annulus or quadrants centered at x₁. For instance, to measure flows in the east-west direction in the neighborhood of point x, cross-covariances are averaged over two quadrant arcs \({\cal A}_{\rm w}(x,\Delta)\) and \({\cal A}_{\rm e}(x,\Delta)\) that include points a distance Δ from x:

$${C_{{\rm{ew}}}}(x,t;\Delta) = \sum\limits_{{x^{\prime}} \in {{\cal A}_{\rm{w}}}} {C(x,{x^{\prime}},t)} + \sum\limits_{{x^{\prime}} \in {{\cal A}_{\rm{e}}}} {C({x^{\prime}},x,t)},$$

(57)

Figure 16 shows the geometry of the averaging procedure. An example cross-covariance C_ew is shown in Figure 17 in the f-mode case. Correlations at positive times (t > 0) correspond to waves that propagate westward, and correlations at t < 0 to waves that propagate eastward. As shown in Figure 17, cross-covariances can be further averaged along lines of constant phase within a range of distances. Likewise an average cross-correlation C_sn is constructed from south-north quadrants to measure flows in the meridional direction. Another average C_ann is obtained by averaging over the whole annulus,

$${C_{{\rm{ann}}}}(x,t;\Delta) = \sum\limits_{{x^{\prime}} \in {\cal A}} {C(x,{x^{\prime}},t),} $$

(58)

where \({\cal A}(x,\Delta)\) is an annulus of radius Δ centered at x. This average can be used to measure separately the waves that propagate inward and outward from the central point.

Open image in new window

Open image in new window

At fixed x and Δ, the cross-covariance function oscillates around two characteristic (first-bounce) times t = ±t_g. Center-to-annulus or center-to-quadrants travel times are often measured by fitting Gaussian wavelets (Duvall Jr et al., 1997; Kosovichev and Duvall Jr, 1997). This procedure distinguishes between group and phase travel times by allowing both the envelope and the phase of the wavelet to vary independently. The positive-time part of the cross-correlation is fitted with a function of the form

$${w_+}(t) = A\exp [ - {\gamma ^2}{(t - {t_{\rm{g}}})^2}]\cos [{\omega _0}(t - {\tau _+})],$$

(59)

where all parameters are free, and the negative-time part of the cross-correlation is fitted separately with

$${w_-}(t) = A\exp [ - {\gamma ^2}{(t - {t_{\rm{g}}})^2}]\cos [{\omega _0}(t - {\tau_-})].$$

(60)

The times τ₊ and τ₋ are the so-called phase travel times. The basic observations in time-distance helioseismology are the travel time maps τ₊(x, Δ) and τ₋(x, Δ), measured for each of the three averaged cross-correlations C_ew(x, t; Δ), C_sn(x, t; Δ), and C_ann(x, t; Δ). The travel time differences τ_diff = τ₊ − τ₋ are mostly sensitive to flows while the mean travel times τ_mean = (τ₋ + τ₊)/2 are sensitive to wave-speed perturbations. Maps of measured travel times are shown in Figure 18: Most of the signal in these maps is due to supergranular flows (15–30 Mm length scales). We note that an alternative definition of travel time can be obtained by fitting a model cross-covariance function to the data (Gizon and Birch, 2002). This last definition is often used in geoseismology (see, e.g., Zhao and Jordan, 1998).

Open image in new window

This definition has a number of useful properties. First, it is very robust with respect to noise. The fit reduces to a simple sum that can always be evaluated whatever the level of noise. Second, it is linear in the cross-covariance. As a consequence, averaging various travel time measurements is equivalent to measuring a travel time on the average cross-covariance. This is unlike previous definitions of travel time that involve non-linear fitting procedures. Third, the probability density functions of τ₊ and τ₋ are unimodal Gaussian distributions. This means, in particular, that it makes sense to associate an error to a travel time measurement. The Born sensitivity kernels discussed in Section 4.3.5 were derived according to this definition of travel times.

4.3.4 Noise estimation

In global helioseismology, it is well understood that the precision of the measurement of the pulsation frequencies is affected by realization noise resulting from the stochastic nature of the excitation of solar oscillations (see, e.g., Woodard, 1984; Duvall and Harvey, 1986; Libbrecht, 1992; Schou, 1992). It is important to study these properties since the presence of noise affects the interpretation of travel time data. In particular, the correlations in the travel times must be taken into account in the inversion procedure.

An interesting approach, pioneered by Jensen et al. (2003a), consists of estimating the noise directly from the data by measuring the rms travel time within a quiet Sun region. The underlying assumptions are that the fluctuations in the travel times are dominated by noise, not by ‘real’ solar signals, and that the travel times measured at different locations can be seen as different realizations of the same random process. By ‘real’ solar signals we mean travel time perturbations due to inhomogeneities in the solar interior that are slowly varying over the time of the observations. Jensen et al. (2003a) studied the correlation between the center-to-annulus travel times as a function of the distance between the central points, at fixed annulus radius.

Gizon and Birch (2004) derived a simple model for the full covariance matrix of the travel time measurements. This model depends only on the expectation value of the filtered power spectrum and assumes that solar oscillations are stationary and homogeneous on the solar surface. The validity of the model is confirmed through comparison with MDI measurements in a quiet Sun region. Gizon and Birch (2004) showed that the correlation length of the noise in the travel times is about half the dominant wavelength of the filtered power spectrum. The signal-to-noise ratio in quiet-Sun travel time maps increases roughly like the square root of the observation time and is maximum for a distance near half the length scale of supergranulation.

4.3.5 Travel time sensitivity kernels

Notice that kernels can be computed for quantities other than travel times. For example, in geophysics Dahlen and Baig (2002) computed the first-order sensitivity of the amplitude of the observed waveform to a small local change in sound speed. In particular it might be helpful to compute kernels for the mean frequency and amplitude of the cross-correlations, with the aim of using these quantities to help constrain inversions.

4.3.5.2 Finite-wavelength kernels in the single-source approximation

The three approximations that have been used to treat small scale perturbations for time-distance helioseismology are the Born approximation, the Rytov approximation, and the Fresnel zone approximation. Here we describe some results that have been obtained using these methods in the single-source approximation. The single-source approximation, which was motived by the “Claerbout Conjecture” (Rickett and Claerbout, 1999) and by analogy with previous ray-theory based work, says that the one-way travel time perturbation can be found by looking at the time-shift of a wavepacket created at one observation location and then observed at the other. Gizon and Birch (2002) showed that the single-source approximation ignores a potentially important scattering process; this issue is discussed in detail in the paragraph on distributed source models.

Motivated by the use of the Born approximation to compute travel time sensitivities in the context of geophysics (see, e.g., Marquering et al., 1998), Birch and Kosovichev (2000) used the Born approximation to compute travel time kernels for time-distance helioseismology. Figure 19 shows an example result. The shape of the kernel is the characteristic “banana-doughnut” shape first described by Marquering et al. (1998). The kernel is hollow along the ray path, travel times are not sensitive to small-scale sound-speed perturbations located along the ray path. This has been explained in terms of wavefront healing (Nolet and Dahlen, 2000). The ringing in the kernel with distance from the ray path is a result of the finite bandwidth of the wavefield.

Open image in new window

Another approach to producing finite-wavelength kernels is the Fresnel zone approximation (see, e.g., Sneider and Lomax, 1996; Jensen et al., 2000). The Fresnel zone approach makes a simple parametrized approximation to the kernels. In particular, in the Fresnel zone approximation, the travel time kernels, for fixed x₁ and x₂, are written as (Jensen et al., 2001)

$$K({x_1},{x_2};r) = {A \over {\tau ({x_2}\vert r)\tau (r\vert {x_1})}}\sin ({\omega _0}\;\Delta \tau)\exp \left[ {- {{\left({{{a\;\Delta \tau} \over {{T_0}/2}}} \right)}^2}} \right],$$

(65)

where τ(x|r) is the ray-theoretical travel time from x to r, and Δτ = τ(x₁|r)+ τ(r|x₂) − τ(x₁|x₂). The parameter ω₀ is the central frequency of the wavepacket, T₀ = 2π/ω₀ is the corresponding period, and a is related to the frequency bandwidth of the wavepacket. The amplitude A of the kernel is determined by demanding that the total integral of the kernel be equal to τ(x₁|x₂), which is the ray theory value for the total integral of the kernel. Notice that the value of the kernel goes to infinity as r tends to x₁ or x₂. Jensen et al. (2000) showed, by comparison with direct simulation, that a two dimensional analogue of the Fresnel zone kernel (Equation (65)) was a reasonable approximation to the actual linear sensitivity of travel time to changes in the sound speed.

The Rytov approximation gives a first order correction to the phase of the wavefield, instead of a first order correction to the wavefield itself as one obtains from the Born approximation. Jensen and Pijpers (2003) used the Rytov approximation to write travel time kernels for the effects of sound-speed perturbations. The results were qualitatively similar to the Born approximation results of Birch and Kosovichev (2000). No comparison has been made, in the helioseismology literature, of the ranges of validity of the Rytov and Born approximations.

4.3.5.3 Distributed source models

Gizon and Birch (2002), motived by Woodard (1997), gave the first comprehensive treatment of the linear forward problem for time-distance helioseismology. Such a treatment must include a physical model of wave excitation which takes into account the fact that waves are stochastically excited by sources (granules) that are distributed over the solar surface (see Section 3.2). An example calculation (Figure 20) demonstrates that the single-source approximation is qualitatively incorrect, as it ignores a scattering process that can be dominant for some perturbations. Figure 20 compares an f-mode travel time kernel for local changes in the wave damping rate, computed in the single-source and distributed-source models. The hyperbolic feature in the distributed-source kernel is not present in the single-source model. As discussed by Gizon and Birch (2002), the single-source approximation ignores the scattered wave that is seen at the observation point that is treated as a source in the single-source approximation.

Open image in new window

Birch et al. (2004) used the recipe derived by Gizon and Birch (2002) to obtain the linear sensitivity of p-mode travel times to changes in sound speed. The results showed that the sensitivity depends on the filtering that is done to the wavefield before the travel times are measured. Figure 21 shows sound-speed kernels computed for three different cases: a model with no filters, a model with an MTF (see Section 3.4) roughly like the one for MDI full-disk data, and a model with an MTF and also a phase-speed filter of the type employed in routine time-distance analysis. The three kernels are all strikingly different. In the unfiltered case, the kernels looks much like the traditional “banana-doughnut” kernel (see Figure 19). With the introduction of the MTF, the travel time becomes more sensitive to deeper sound-speed perturbations and less sensitive to the near-surface perturbations. This is because the MTF preferentially removes high wavenumbers, and thus modes with low phase speeds and shallow turning points. With the introduction of the phase-speed filter the kernel again changes drastically. The phase-speed filter removes all of the waves with lower turning point much below the turning point of the target ray. As a result, the sensitivity much below the lower turning point of the ray is greatly reduced. This example emphasizes the importance of including the filtering in models of travel time sensitivity kernels. The strong dependence of the shape of the kernels on the filtering suggests that the kernels could to be tailored by adjusting the filters that are used in the data analysis procedure.

Open image in new window

4.3.6 Inversions of travel times

The inverse problem is to determine the perturbations δq _α (r) to the solar model that are consistent with a particular set of observed travel times δτ _i . For the inverse problem, the kernel functions and the noise covariance are in general assumed to be known.

The minimization of χ², Equation (66), has been done using either the LSQR algorithm (Paige and Saunders, 1982), the multi-channel deconvolution (MCD; Jacobsen et al., 1999), or via singular value decomposition (Hughes and Thompson, 2003). The inputs to either of these methods are the travel time kernels, which result from the linear forward problem (see Section 3.5), the regularization operator, the choice of regularization parameter, and in general also the covariance matrix of the noise.

The central question about inversions is the degree to which any particular inversion method is able to retrieve the true Sun, e.g., the actual sound-speed variations and flows in the interior, from a set of observed travel times. The degree to which any inversion method succeeds will presumably depend on a number of things: (i) the accuracy of the forward model, (ii) the noise level in the travel times, (iii) the depths and spatial scales of the real variations in the Sun, (iv) the number and type of travel times that are available as input to the inversion, and (v) the accuracy of the travel time noise covariance matrix. The two main approaches that have been taken to study inversion methods are either to invert artificial data or to invert real data for perturbations that are relatively well known from global helioseismology or from surface measurements. It is also useful to compare the results of different inversion methods applied to the same set of observed travel times.

The approach of inverting real data is appealing as it is by definition “realistic”. This approach for time-distance was pioneered by Giles (1999) who did ray theory based inversions to measure differential rotation. The results were in approximate agreement with the results of global model inversions, which validated time-distance inversions applied to slowly-varying large scale flows. Gizon et al. (2000) inverted f-mode travel times for horizontal flows in the near-surface. Figure 22 shows a comparison between the projection of the inferred flows onto the line of sight and the directly observed line-of-sight Doppler velocity. The correlation coefficient between the two is 0.7 (Gizon et al., 2000), which shows that their inversion method, an iterative deconvolution, is able to approximately retrieve the horizontal flow velocities from the travel times.

Open image in new window

There have been a number of efforts to validate inversions by the generation and inversion of artificial data. These tests are useful as they provide an intuitive understanding of various regularization schemes, the choice of regularization parameter, the effects of limited sets of input travel times, the effects of incorrect assumptions regarding the noise covariance, the potential resolution of any particular inversion, and the cross-talk between different model parameters. A drawback to testing inversions with artificial data is that one always wonders about the realism of the artificial data. Hopefully, in the not distant future quite realistic data will be available for the purpose of testing inversion schemes.

Couvidat et al. (2004) compared inversions done using ray-approximation kernels with inversions done with Fresnel-zone kernels. The results were quite similar for depths between the surface and the lower turning point of the deepest rays used in the inversion. Inversions done with the ray kernels cannot retrieve sound speed perturbations below the lower turning point of the deepest ray, while the Fresnel-zone kernels extend below the depth of the deepest ray. The results of Couvidat et al. (2004), as well as those of Giles (1999), suggest that even though the ray approximation overestimates travel times for small scale perturbations this effect does not seriously corrupt the large scale features found by inversion.

Another test of inversion methods was done by Zhao and Kosovichev (2003a). The two main results of this study, regarding inversions methods, were that there is cross-talk between upflows and convergent flows, and between downflows and divergent flows, and that the MCD and LSQR methods give essentially the same results when applied to quiet sun MDI data. Figure 23 shows an example of a test of the LSQR method. The initial model is shown in the top panel. Travel times were generated using the ray approximation. The travel times were then inverted using five iterations of LSQR to obtain the flow field shown in the second panel of Figure 23. Notice that there are small vertical flows near the surface at the centers and boundaries of the supergranulations cells. After 100 iterations of LSQR, these small vertical velocity features are removed (bottom panel of Figure 23).

Open image in new window

The first end-to-end test of time distance helioseismology using artificial data was done by Jensen et al. (2003b). In this important study, artificial data were generated by numerical wave propagation through a three-dimensional stratified and horizontally inhomogeneous model. Waves were generated by stochastic sources distributed over a layer 50 km below the upper boundary of the surface of the model. The artificial wave field was “observed” at the spatial resolution and temporal cadence of MDI high-resolution observations. This artificial data set was then subjected to standard time-distance analysis, and the inversion was performed with the Fresnel-zone kernels described by Jensen and Pijpers (2003). Figure 24 shows a comparison of the original sound-speed model and the result of the inversion of the artificial data. Notice that the inversion recovers much of the original structure. The inversion result contains noise, which is to be expected from travel times computed from finite time series. The only noise source in this example is realization noise, noise resulting from the stochastic nature of the wave excitation.

Open image in new window

4.4.1 Ingression and egression

The ingression and egression are attempts to the answer the following question: Given the wavefield observed at the solar surface, what is the best estimate of the wavefield at some point in the solar interior assuming that the observed wavefield resulted entirely from waves diverging from that point (for the egression) or waves converging towards that point (for the ingression)?

The development of holography has, historically, been motived by analogy with optics (e.g., Lindsey and Braun, 2000a) and as a result the holography literature often employs optical terminology. The target point in the solar interior at which we attempt to estimate the wavefield is termed the “focus point”. In practice, only data in a restricted area on the surface above the focus point is used to compute the egression and ingression. This region is called the “pupil”. The choice of pupil geometry depends on the particular application and will be described when we discuss particular applications of holography.

An intuitive way to interpret Equation (70) is to view the Green’s functions as propagators. In this way, the egression can be seen as the result of using the anti-causal Green’s function \(G_{+}^{{\rm holo}}\) to propagate the surface wavefield backwards in time. Likewise, the ingression can be seen as the result of using the causal Green’s function \(G_{-}^{{\rm holo}}\) to propagate the observed surface wavefield forwards in time.

The egression and ingression are the essential quantities in helioseismic holography. There are many particular ways in which these quantities can be combined to learn about the solar interior. The main techniques are control-correlations (Section 4.4.3), acoustic power holography (Section 4.4.4), phase-sensitive holography (Section 4.4.5), and far-side imaging (Section 4.4.6) which is a special case of phase-sensitive holography. Of central importance to all of these methods is the choice of the holography Green’s function \(G_{\pm}^{{\rm holo}}(r,\omega)\), which we now describe.

4.4.2 Holography Green’s functions

The two approaches that have been used to construct the holography Green’s functions, \(G_{\pm}^{{\rm holo}}\), involve ray theory and wave theory.

4.4.2.1 Ray theory

Lindsey and Braun (2000a), motived by ray theory, prescribed the holography Green’s function as

$$G_\pm ^{{\rm{holo}}}(r,t) = \sum\limits_N {{A_N}} (x,z){\delta _{\rm{D}}}(t \pm {\tau _N}(x,z)),$$

(71)

where r = (x, z) and x = ∥x∥ is the horizontal distance. The role of the function \(G_{-}^{{\rm holo}}(G_{+}^{{\rm holo}})\) is to propagate the observed surface wavefield forwards (backwards) in time and into the interior of the solar model. Theses functions can also be seen as the “wavefield” response at a horizontal distance x, depth z, and time t to a surface source located at the origin at time t = 0. The observed surface wavefield can be thought of as a source, in the spirit of Huygen’s principle. The index N refers to the number of skips the ray path has taken off the solar surface. Each function A _N (x, z) is the amplitude of a single skip component of the Green’s function and can be estimated from simple ray theory. In particular, in ray theory the amplitude functions A _N are determined by the conservation of energy and the increase, with distance along the ray path, of the cross-sectional area of a ray bundle emerging from the source location. In Equation (71), the functions τ _n are the ray-theoretical travel times along the N-skip ray paths connecting a point on the surface with a point at depth z and a horizontal distance x away from the first point. The sum over skips N in Equation (71) is convenient as it can easily be truncated to estimate, for example, just the “one-skip” Green’s function, i.e., the Green’s function that contains only contributions from waves that have traveled once into the solar interior. The holography Green’s functions in the time domain (Equation 71) must then be Fourier transformed to obtain G_±(r, ω) referred to in Equation (70).

Notice that there are a great many simplifications involved in writing the Green’s function in Equation (71). The substantial frequency dependence of the travel time (e.g., Jefferies et al., 1994) has been ignored. There is no account taken of damping, which is quite significant for high-degree modes (e.g., Duvall Jr et al., 1998). Furthermore, as the ray approximation requires the wavelength to be much smaller than the length scale of the variation of the background medium, the ray approximation is not expected to be valid near the solar surface. Further issues include the neglect of the buoyancy and cut-off frequencies in the calculation of τ _N , and also in the computation of the ray paths themselves (Lindsey and Braun 2000a). Recent work by Lindsey and Braun (2004) suggests that after empirical dispersion corrections have been applied (see Section 4.4.3), the results of ray theory calculations are quite similar to the results of wave theory calculations (see Figure 25).

Open image in new window

4.4.3 Local control correlations

The local control correlation is the correlation between the observed signal at a particular point on the solar surface and the holographic reconstruction of the signal at that same point computed from the data in a surrounding region. In particular the local control correlations are defined as (see, e.g., Lindsey and Braun, 2004)

$${C_-}(x,\omega) = {\langle {\Phi (x,\omega)H_- ^{\ast}(x,{z_{\rm{o}}},\omega)} \rangle _{\Delta \omega}},$$

(74)

$${C_+}(x,\omega) = {\langle {{H_+}(x,{z_{\rm{o}}},\omega){\Phi ^{\ast}}(x,\omega)} \rangle _{\Delta \omega}}.$$

(75)

Here the depth z₀ denotes the z coordinate of the solar surface. The angle brackets denote the average over a frequency band of width Δω centered at frequency ω. Notice that we have not specified the pupil which goes into the ingression/egression calculations. The pupil that is used depends on the particular situation. Figure 26 shows quiet Sun local control correlations, measured using both ray theory and wave theory Green’s functions.

Open image in new window

If the egression/ingressions were perfect reconstructions of the observed signal, then the phases of the C_± would both, on average, be zero in the quiet Sun. In practice, this is not the case (Lindsey and Braun, 2004). The phases of the holography Green’s functions are typically corrected so that the phase of the average quiet Sun local control correlation is zero (e.g., Lindsey and Braun, 2000a). We refer to this correction as the empirical dispersion correction.

4.4.4 Acoustic power holography

4.4.5 Phase-sensitivity holography

4.4.6 Far-side imaging

Far-side imaging is a special case of phase-sensitive holography (e.g., Lindsey and Braun, 2000b; Braun and Lindsey, 2001). The idea is to use the wavefield on the visible disk to learn about active regions on the far-side of the Sun. Figure 27 shows the geometry that was used by Braun and Lindsey (2001). In order to obtain full coverage of the far-side of the Sun, two different geometries were employed. For regions near the antipode of the center of the visible disk a two-skip geometry and two-skip Green’s functions were employed (panel (a) of Figure 27). For focus positions near the limb, the ingression/egression were computed using a single-skip pupil and corresponding Green’s functions, and then correlated with the egression/ingression computed using a three-skip pupil and Green’s functions (the geometry is shown in panel (b) of Figure 27).

Open image in new window

4.4.7 Acoustic imaging

Acoustic imaging was first introduced by Chang et al. (1997); for a recent review see Chou et al. (2003). We have included acoustic imaging in the section on helioseismic holography as the definitions and philosophical motivation for the two techniques are quite similar.

The central computations in acoustic imaging (see, e.g., Chou et al., 1999, 2003) are wavefield reconstructions in the solar interior, Ψ_in and Ψ_out, defined by

$${\Psi _{{\rm{out}},{\rm{in}}}}(r,t) = \sum\limits_{\tau = {\tau _1}}^{{\tau _2}} {W(\tau, z)} \overline \Phi (\Delta (\tau, z),t \pm \tau).$$

(81)

The focus position is r = (x, z). The function \(\overline{\Phi}(\Delta,t)\) denotes the azimuthal average, around x, of the surface wavefield Φ measured a horizontal distance Δ away from x. The quantity τ represents the ray-theoretical travel times from the subsurface focus point at r to surface points a horizontal distance Δ away from the focus point. For a given τ, the distance Δ satisfies the time-distance relation established between the focus point and a surface point. This relation, Δ = Δ(τ, z), is computed from a standard solar model, using the ray approximation (see Figure 28). The function W is a smooth weight function of τ (or Δ), explicitly given by Chou et al. (1999). The sum in Equation (81) involves the observed wavefield inside an annulus with inner and outer radii specified by the range τ₁ < τ < τ₂.

Open image in new window

In some sense, acoustic imaging is a special case of holography. Here, the pupil is an annulus with inner radius Δ₁ and outer radius Δ₂ given by the time-distance relation Δ _j = Δ(τ _j , z). The equivalent holography Green’s function is given by Wδ_D(t − τ). The signal Ψ_out corresponds to the egression, i.e., is the signal reconstructed from the observations of waves diverging from the focal point, while corresponds to the ingression, the signal estimated from the waves seen converging towards the focus position.

The forward problem that has received the most attention in acoustic imaging is the dependence of travel times on changes in the sound speed. Chou and Sun (2001) used the ray approximation to estimate the sensitivity of acoustic imaging phase travel times to changes in sound speed. The results showed that in general the horizontal resolution is greater than the vertical resolution and that the resolution decreases with increasing focus depth.

The inverse problem of determining the sound speed from a given set of travel times has been studied in the ray approximation as well. Sun and Chou (2002) tested ray-theory RLS inversions of phase-time measurements on artificial data. These tests showed that the RLS inversions were, for good choices of the regularization parameter, capable of reconstructing the model used to generate the artificial data. Inversions of travel times for sound speed were also discussed in detail by Chou and Sun (2001).

4.5.1 Forward problem

4.5.2 An example calculation

Woodard and Fan (2005) make the further approximation that oscillations of different radial orders are excited incoherently. This is certainly not the case in reality, as evidenced by the asymmetry in the power spectrum. It is not clear what effect this approximation will have on the final result, as most of the power is near the resonances, where line-asymmetry is not important.

Notice that we also need to compute the term \(E[\Phi(k,\omega)\delta\Phi^{\ast}(k^{\prime},\omega^{\prime})]\) as the final result that we wish to obtain is the first order change in \(E[\Phi(k,\omega)\Phi^{\ast}(k^{\prime},\omega^{\prime})]\) introduced by a small change in the model. This second term can be computed in exactly the same manner as the \(E[\delta\Phi(k,\omega)\Phi^{\ast}(k^{\prime},\omega^{\prime})]\) term, which we have already done.

4.5.3 Inverse problem

So far we have only addressed the forward problem. The other half of the direct-modeling approach is the inverse problem, that is inferring the flows in the Sun from the Fourier domain correlations seen in the data. The inverse problem in direct modeling has been solved as a linear least-squares problem (Woodard, 2002; Woodard and Fan, 2005). A crucial observation is that a flow with horizontal wavenumber q only couples modes with wavenumbers k and k′ when ∥k − k′∥ = q. This is a result of the assumption that the background model is plane-parallel and translation invariant. As a result the different wavenumber components of the flow can be inferred one at a time, in the spirit of the MCD inversion scheme (Jacobsen et al., 1999). The same argument applies in the time domain, and flows can be inferred one frequency component at a time.

Woodard (2002) tested his direct-modeling approach on a quiet sun region using MDI data. The method was used to invert for near-surface supergranular-scale flow. Although many approximations were made in carrying out the inversion, general agreement (a correlation coefficient of 0.68) was found between the Doppler component of the seismically inferred flow in the photosphere and the directly observed surface Doppler signal (see Figure 29). Thus direct modelling appears to be a very promising technique.

Open image in new window

5.1.1 Rotation and torsional oscillations

Measuring the mean rotation using local techniques is a useful exercise, in particular to compare with estimates from inversions of frequency-splitting coefficients. Giles et al. (1998) and Giles (1999) measured rotation with time-distance helioseismology applied to MDI data (structure and dynamics programs; see Scherrer et al., 1995). Figure 30 shows the angular velocity as a function of latitude for radii r > 0.886R_⊙. There is a good qualitative agreement between the time-distance and the global-mode inversions, although the agreement is not as good at high latitudes. Basu et al. (1998, 1999) used the ring-diagram method to obtain the rotational velocity for r > in the latitude range |λ| < 50°. The north-south symmetric component of rotation compares favorably with the rotation determined from the frequency splittings of the global p modes (Figure 31). Since the inversion results of p-mode splittings were not expected to be particularly reliable near the surface, the ring diagram results complement and support the earlier conclusion that rotation increases with depth, near the surface, at low latitudes. Possible explanations for the existence of this near-surface radial shear have been discussed by Gilman (2000) and Robinson and Chan (2001). The local helioseismology results also show measurable differences in the rotation profile between the northern and southern hemispheres (see, e.g., Basu et al., 1999; Giles, 1999; González Hernández and Patrón, 2000). It would be premature, however, to trust the validity of the north-south asymmetry in the rotation rate, since systematic errors are not fully understood and do vary with latitude.

Open image in new window

Open image in new window

Howard and Labonte (1980) discovered small (±10 m s⁻¹) latitudinal variations in the surface Doppler rotation profile that propagate toward the equator with the sunspot latitudes. These variations, known as torsional oscillations, have since been confirmed by global helioseismology (Kosovichev and Schou, 1997; Schou, 1999). Inversions of global-mode frequency splittings have shown that torsional oscillations persist at a depth over a large fraction of the convection zone (Howe et al., 2000a), especially at high latitudes where they extend down to the bottom of the convection zone (Vorontsov et al., 2002). At low latitudes, the data support the idea that the torsional oscillation pattern originates inside the convection zone and propagates both upward and equatorward (Vorontsov et al., 2002; Basu and Antia, 2003). Above 45° latitude, bands of faster and slower rotation appear to move toward the poles (Antia and Basu, 2001; Schou, 2003a) suggesting a connection with the poleward migration of high-latitude magnetic activity seen at the surface (see, e.g., Callebaut and Makarov, 1992).

Both time-distance (Giles et al., 1998; Beck et al., 2002; Zhao and Kosovichev, 2004) and ring-diagram (Basu et al., 1999; Basu and Antia, 2000; Haber et al., 2000, 2002) analyses have confirmed the global-mode measurements of torsional oscillations. Figure 32 shows a plot of the zonal flows at depths of 0.9 Mm and 7.1 Mm obtained from ring-diagram analysis. At both these depths, propagating zonal fast bands are seen to migrate toward the equator as the solar cycle progresses. As will be later discussed in Section 5.2, local techniques can also provide information about the longitudinal structure of the zonal flows.

Open image in new window

Schüssler (1981) and Yoshimura (1981) suggested that the torsional oscillations may be driven by the Lorentz force due to a migrating dynamo wave. Recent numerical calculations by Covas et al. (2000) seem to be consistent with this hypothesis; their calculations also reproduce the poleward propagating torsional oscillations at high latitudes. Other explanations attribute the torsional oscillations to the feedback of the smaller-scale magnetic fields on the angular momentum transport mechanisms responsible for differential rotation, e.g., changes in the Reynolds/Maxwell stresses (Küker et al., 1996; Kitchatinov et al., 1999) or the local suppression of turbulent viscosity by active regions (Petrovay and Forgács-Dajka, 2002). An alternative explanation by Spruit (2003) suggests that zonal flows may be driven by temperature perturbations near the surface due to the magnetic field: Local flows around regions of enhanced magnetic activity act through the Coriolis force to balance horizontal pressure gradients. The model of Spruit (2003) naturally predicts a solar-cycle variation in the meridional flows. The reader is referred to the review paper of Shibahashi (2004) for other possible explanations.

5.1.2 Meridional flow and its variations

Surface Doppler measurements have shown the existence of a meridional flow from the equator to the poles with an amplitude of about 20 m s⁻¹ (e.g. Hathaway, 1996). Meridional circulation is an important ingredient of solar dynamo models where magnetic flux is transported by a deep equatorward flow (flux-transport dynamo models). In such models, the solar-cycle period is largely determined by the amplitude of the meridional flow near the base of the convection zone (Dikpati and Charbonneau, 1999). Hathaway et al. (2003) claimed that the butterfly diagram is evidence for flux-transport dynamo models, and estimated a 1.2 m s⁻¹ flow at the bottom of the convection zone from the drift of sunspots toward the equator. However, Schüssler and Schmitt (2004) argue that the butterfly diagram is equally well reproduced by a conventional dynamo model with migrating dynamo waves without transport of magnetic flux by a flow. The depth of penetration of the meridional flow is another parameter in some dynamo models (see, e.g., Nandy and Choudhuri, 2002). Knowing meridional circulation is also important to understand its feedback on differential rotation (Rekowski and Rüdiger, 1998; Gilman, 2000).

Meridional circulation was first detected at depth by Giles et al. (1997) with time-distance helioseismology. The inversions, constrained by mass conservation, show that the data are consistent with a meridional circulation which is 20 m s⁻¹ poleward at the solar surface and about 3 m s⁻¹ equatorward at the base of the convection zone, with a turnover point near 0.8R_⊙ (Giles, 1999, see Figure 33). Other estimates of the meridional circulation have been obtained in the near-surface layers with time-distance helioseismology (see, e.g., Duvall Jr and Gizon, 2000; Zhao and Kosovichev, 2004), Hankel analysis (Braun and Fan, 1998), acoustic imaging (e.g. Chou and Dai, 2001), and ring-diagram analysis (see, e.g., Basu et al., 1999; González Hernández et al., 1999, 2000; Haber et al., 2000). Giles (1999) and Haber et al. (2002) found a submerged equatorward flow in the north during the years 1998–2001. This last result, however, may be contaminated in part by errors in the orientation of the MDI Dopplergrams which cause rotation to leak into the meridional flow signal: a P-angle error due to the misalignement of the MDI telescope on the SOHO spacecraft (C. Toner, 2000, private communication) and an error in the Carrington elements that specify the direction of the solar rotation axis (Giles, 1999).

Open image in new window

Meridional circulation (averaged in longitude over several rotation periods) is observed to change from year to year through the solar cycle. To study the time-varying component of meridional circulation, we consider the residuals obtained after subtraction of a temporal average at each latitude. Meridional flow residuals have an amplitude which is less than 10 m s⁻¹. Like torsional oscillations, the north-south flow residuals drift equatorward with active latitudes. Unlike torsional oscillations, the residuals of meridional circulation change sign across the uppermost 70 Mm. Near the surface, the time-varying residuals converge toward active latitudes (Gizon, 2003; Basu and Antia, 2003; Zhao and Kosovichev, 2004, see Figure 34), while at greater depths meridional flow residuals diverge from the dominant latitude of activity (Chou and Dai, 2001; Beck et al., 2002, see Figure 35). The time-varying component of meridional circulation has also been measured near the surface from the advection of supergranulation (Gizon and Duvall Jr, 2004, see Section 5.3.5).

Open image in new window

Open image in new window

These observations, summarized in Figure 36, lead Zhao and Kosovichev (2004) to postulate the existence of extra meridional ciculation rolls on each side of the mean latitude of activity, superimposed on the mean global-scale poleward flows. However, it is suggested in Section 5.2 that the changes in the longitudinal averages of meridional flows can be attributed for a large part to flows that are localized both in longitude and latitude around active regions.

Open image in new window

5.1.3 Vertical flows

Because of their small amplitude, vertical flows are much harder to measure than horizontal flows. The ring-diagram analysis is not, in principle, sensitive to vertical flows. However, Komm et al. (2004) showed how to estimate the vertical flows from ring-diagram inversions by adding a mass conservation constraint (using GONG data). This mass conservation constraint is enforced locally for each 15° × 15° region, under the assumption is that horizontal density fluctuations are negligible because of the large size of the region and the large integration time. Figure 37 shows the longitudinal average of the near-surface vertical velocity over one Carrington rotation (Komm et al., 2004). Weak downflows (less than 1 m s⁻¹) are observed near active latitudes. This result is consistent with the observations of a residual horizontal flow converging toward the mean latitude of activity near the surface (see Figure 36). For depths in the range 0–16 Mm, the amplitude of the downflow increases with depth. A small upflow is seen near the equator.

Open image in new window

5.1.4 Search for variability at the tachocline

Howe et al. (2000b) discovered temporal changes in the rotational velocity near the bottom of the convection zone with a period of approximately 1.3 yr and a peak amplitude at 0.72R_⊙. The angular velocity variations at 0.72R_⊙ and 0.63R_⊙ are anticorrelated (the signal seemed much weaker near the second half of 2001 but it has come back since). These results are somewhat controversial since they have not been confirmed by others (see, e.g., Basu and Antia, 2001; Vorontsov et al., 2002). However, using wavelet analysis over cycles 21–23, Boberg et al. (2002) found signals with a 1–2 yr period in the solar mean magnetic field and Knaack et al. (2005) detected transient 1.3 yr oscillations in the unsigned photospheric magnetic flux, which may be related to the large-scale magnetic surges towards the poles. The 1.3 yr periodicity is also seen in observations of the interplanetary magnetic field and geomagnetic activity (Lockwood, 2001). Gough (2000) suggested that a 500 G radial magnetic field across the tachocline may lead to exchange of angular momentum with the right time scale, while Covas et al. (2001) suggested a highly non-linear mechanism that does not require different physical time scales for torsional and tachocline oscillations.

A deeply challenging task for local helioseismology is to detect the signature of the magnetic field in the tachocline region. The results of Howe et al. (2000b) have not been confirmed by local helioseismology yet, but there have been a few attempts to search for wave-speed perturbations near the base of the convection zone. Chou and Serebryanskiy (2002) selected wave packets that return to the same spatial point after traveling around the Sun with an integral number of bounces N. The ray paths taking N = 8 bounces to go around the Sun have a lower turning point close to the bottom of the convection zone. Chou and Serebryanskiy (2002) observed that the change in one-bounce travel time at solar maximum relative to minimum is approximately the same for all N values studied, except for N = 8 which shows an additional decrease in travel time. Figure 38 shows the relative decrease in the one-bounce travel time at N = 8 relative to the other N values as a function of time (about 15 ms at solar maximum). The correlation with the sunspot number suggests that the additional decrease in travel time at N = 8 might be caused by wave-speed perturbations at the base of the convection zone. With the simplified assumption that this additional decrease is caused only by a magnetic field perturbation at the base of the convection zone, the field strength was estimated to be 0.4 MG at solar maximum. However, Chou and Serebryanskiy (2002) caution that this interpretation is oversimplified since it would also imply an additional decrease in travel time for the wave packets with N ≥ 7, which is not observed. We note here that Ruzmaikin and Lindsey (2003) argue that only field strengths greater than 1 MG should be detectable near the bottom of the convection zone with currently available data.

Open image in new window

A different approach by Birch (2002) and Duvall Jr (2003) was to search for longitudinal structures near the bottom of the convection zone. Raw travel times measured between surface points separated by 45° were observed to be correlated from day to day, indicating a real solar signal (Duvall Jr, 2003). However, it is not known whether this signal is due to surface effects or deep perturbations.

It should be mentioned that global seismology may also be used to detect a toroidal magnetic field at the level of a fraction of a megagauss, if such a field were confined to a thin layer near the base of the convective envelope (e.g., Basu, 1997; Dziembowski and Goode, 2004). For example, Basu (1997) gave an upper limit of 0.3 MG near solar minimum.

It is fair to say that local helioseismological techniques have encountered difficulties in probing the tachocline. In order to reach a depth of 200 Mm, it is necessary to consider ray paths that connect surface points separated by almost 45°. All local methods thus suffer from foreshortening near the limb. Also, the plane-wave approximation is obviously not appropriate for probing the deep convection zone. Certain methods, like ring-diagram analysis, need to be adapted to take this into account (see, e.g., Haber et al., 1995).

5.2.1 Ordered flows near complexes of magnetic activity

Here we consider flows near regions of enhanced magnetic activity. We do not discuss local flows associated with sunspots, but larger-scale flows around complexes of magnetic activity. Using local helioseismology, synoptic maps of local horizontal flows can be constructed by averaging the data in time in a frame of reference that co-rotates with the Sun. Weak 50 m s⁻¹ surface flows that converge toward active regions were detected by Gizon et al. (2001) with time-distance helioseismology (Figure 39). These flows, which exist as far as 30° from the centers of active regions, are also seen in ring diagram analyses with a coarser resolution (Haber et al., 2001). We note that Hindman et al. (2004) have shown that the two methods give remarkably similar results near the surface. Recently, both time-distance (Zhao and Kosovichev, 2004) and ring-diagram (Haber et al., 2004) analyses have provided depth inversions of horizontal flows down to about 15 Mm. Depths inversions indicate that, below 10 Mm, horizontal flows often diverge from active-region centers with velocities on the order of 50 m s⁻¹. Figure 40 shows the flows measured by Haber et al. (2004) around a particular active region. Although vertical flows have not been directly measured yet, motions appear to be organized in the form of a toroidal cell with a surface inflow and a deeper outflow. The picture is not so simple on time scales that are shorter than a week since flows do evolve from day to day.

Open image in new window

Open image in new window

The observed flows around active regions could be caused by the magnetic field. Indeed, the model of Spruit (2003) mentioned earlier predicts a surface inflow toward regions of enhanced magnetic field. An alternative explanation by Yoshimura (1971) suggests that the longitudinal ordering of solar magnetic fields could be due to the existence of large convective patterns that favor the formation of active regions at particular sites on the solar surface. We note that far away from active regions, complex flows like meanders, jets, and vortices are seen in the synoptic maps (Figure 39); Toomre (2002) suggested that they may be related to the largest scales of deep convection.

5.2.2 Effect on longitudinal averages of large-scale flows

An important question is to know whether the local flows that surround active regions contribute significantly to the solar-cycle variations of the longitudinal averages of rotation and meridional circulation in the upper layers of the convection zone (as discussed in Section 5.1.1 and Section 5.1.2). In other words, are the time-varying components of rotation and meridional circulation global phenomena or are they modulated in longitude by the presence of active regions?

In order to help answer this question, we consider synoptic maps of surface flows obtained in 1999 for Carrington rotations 1948 and 1949. By excluding the local areas of magnetic activity from the computation of the longitudional averages of rotation and meridional circulation, we obtain “quiet-sun” estimates that can be compared to the flows averaged over all longitudes (see Gizon, 2003, for details). From Figure 41 it is obvious that the torsional oscillation (rotational shear at active latitudes) is present in between active longitudes. Thus the time-varying zonal flows appear to be mostly independent of longitude as they slowly drift in latitude through the solar cycle. Yet, active regions do rotate a little faster than their surroundings, thus affecting the mean rotation rate at active latitudes (see, e.g., Zhao et al., 2004). The influence of local active-region flows on longitudinal averages of north-south flows is more serious. Indeed, the 50 m s⁻¹ organized flows that surround active regions introduce a kink at active latitudes in the average meridional circulation profile, on the order of ±5 m s⁻¹. Thus the solar-cycle variations of the meridional circulation (Section 5.1.2) would appear to be caused by active regions: convergence toward active latitudes near the surface and, presumably, divergence below 10 Mm.

Open image in new window

5.2.3 Sunspot flows

Here we consider flows in the immediate vicinity of sunspots, which should not be confused with the flows discussed in the previous two sections. Duvall Jr et al. (1996) used the technique of time-distance helioseismology to measure the travel time difference between incoming and outgoing p-mode wavepackets around sunspots. They suggested that the observations are consistent with the existence of a downflow below sunspots with a velocity of about 2 km s⁻¹. It was estimated that the downflow persists down to a depth of 2 Mm below the surface. Kosovichev (1996) performed a 3D inversion of travel time measurements. The inversion results are consistent with 1 km s⁻¹ downflows that can reach depths of about 25 Mm.

Lindsey et al. (1996) employed “knife-edge” diagnostics to estimate horizontal flows around sunspots. In knife-edge diagnostics the Fourier transform of the data is multiplied by some filter F(k, ω), and then transformed back to the space-time domain. The square of the absolute value of the signal is then averaged over time. The filter is constructed to let through the part of the signal that has been Doppler shifted by a flow in a particular direction. An estimate of the x component, for example, of the flow is then formed by subtracting the local amplitudes that correspond to waves that have been Doppler shifted in the \(+\hat{x}\) and \(-\hat{x}\) directions. Using this technique, Lindsey et al. (1996) were able to detect a horizontal outflow from the center of sunspots at depths of about 15 Mm. The amplitude of this flow reaches a maximum of about 180 m s⁻¹ at a distance of 40 Mm from the center of a sunspot.

The Hankel analysis also has the capability of detecting flows below sunspots by interpreting the phase shifts between inward and outward traveling acoustic waves. Braun et al. (1996) applied this technique to south pole data from 1988 and 1991, and reported phase shifts that are consistent with the Doppler effect of a radial outflow from sunspots. The mean horizontal outflow appears to increase with depth and reaches 200 m s⁻¹ at a depth of 20 Mm. A more detailed analysis by Sun et al. (1997) using TON data demonstrates that there is a positive frequency shift between outgoing and ingoing waves which corresponds to a 40–80 m s⁻¹ radial outflow from sunspots. These results exhibit properties similar to those reported by Lindsey et al. (1996).

In order to gain some confidence about flow measurements in sunspots, it is useful to infer surface flows with local helioseismology and check for consistency with direct Doppler measurements. Gizon et al. (2000) used f-mode time-distance helioseismology to study flows very close to the surface (1 Mm deep). Wave-based sensitivity kernels for horizontal flows were used in an iterative deconvolution of the travel times. Figure 42 shows a map of the horizontal flows near the surface for a sunspot observed on 1998 December 6 (the spatial resolution is about 3 Mm). A horizontal outflow around sunspots is observed with an amplitude of about 500 m s⁻¹, which extends to roughly twice the penumbral radius. A high correlation was found between the mean Dopplergram and the projection of the inferred horizontal flows onto the line of sight, confirming the validity of the time-distance inversion. This outflow, often called the moat flow, was seen earlier using surface Dopplergrams (Sheeley, 1972) and magnetic tracers (see, e.g., Brickhouse and Labonte, 1988). The moat flow is believed to be driven by a pressure gradient caused by the blockage of heat by sunspots (see, e.g., Nye et al., 1988).

Open image in new window

Zhao et al. (2001) used a damped least-squares inversion (Tikhonov, 1963) of travel times to infer mass flows around a sunspot below the solar surface. Figure 43 shows the horizontal and vertical components of the flows at three different depths. Converging and downward directed flows were detected at depths of 0–3 Mm, while outflows extending more than 30 Mm from sunspot axis were found below the downward and converging flows. Two vertical cuts through the sunspot, shown in Figure 44, show strong flows across the sunspot at depths of 9–12 Mm, which may provide some evidence in support of the cluster model, as opposed to the monolithic sunspot model. Zhao and Kosovichev (2003b) also applied their time-distance inversion technique to determine the subphotospheric dynamics of an unusually fast-rotating sunspot observed in August 2000, which revealed that the vortical flows can be seen beneath the visible surface of this active region. On the basis of the three-dimensional velocity fields obtained from the time-distance helioseismology inversions, Zhao and Kosovichev (2003b) estimated the subsurface kinetic helicity and concluded that it is comparable to the current helicity estimated from vector magnetograms.

Open image in new window

Open image in new window

Braun and Lindsey (2003) have recently extended their applications of helioseismic holography to include Doppler diagnostics of quiet Sun regions. Phase-correlation holography measures travel time perturbations from the temporal correlations between the egression and ingression. By dividing the pupil into four quadrants is is possible to infer vector flows. Braun and Lindsey (2003) performed the analysis separately for bands of power around 3, 4, and 5 mHz, and the resulting maps were then averaged together. As shown in Figure 45, it is easy to identify the moat outflow at a focus depths of 3 Mm, as well as supergranules (24 hr time average).

Open image in new window

Open image in new window

Almost all local helioseismology results seem to indicate that there exists a radial outflow around sunspots down to at least 10 Mm, which relates to the moat flow seen at the surface. An exception is the inflow measured with p-mode time-distance helioseismology at a depth of 0–3 Mm. It is especially puzzling that the f-mode and p-mode results are inconsistent. Regarding vertical flows, it is difficult to understand how both downflows and horizontal outflows can coexist just below the sunspot. We caution that many complications introduced by the magnetic field are simply ignored by all local techniques. For example, the effect of the magnetic field on excitation and damping mechanisms is known to introduce a small travel time difference between incoming and outgoing waves (Woodard, 1997; Gizon and Birch, 2002).

5.2.4 Sinks and sources of acoustic waves

The very first success of local helioseismology was the detection of acoustic absorption by sunspots. Braun et al. (1987) decomposed the oscillations in the annular region surrounding a sunspot into incoming waves and outgoing waves (Hankel decomposition, Section 4.1) to measure the incoming and outgoing wave power. It was found that sunspots with a typical radius of 20 Mm absorb as much as 50% of the incoming acoustic power (Braun et al., 1987, 1988). The absorption coefficient reaches 70% for the giant sunspot group studied by Braun and Duvall (1990). Braun et al. (1988) found that the absorption coefficient a increases as a function of horizontal wavenumber k. The decrease of a with m/k is consistent with the absorbing region being mostly confined to the sunspot area. Weaker magnetic regions, such as plage, also display a significant level of p-mode absorption. Using 68 hr of south pole data, Braun (1995) was able to measure the absorption coefficient as a function of degree, azimuthal order, and frequency. It was found that the m dependence is weak for |m| < 20. The m-averaged absorption coefficient is plotted in Figure 47 for two different sunspots as function of frequency and for different harmonic degrees. There is a peak in the absorption around 3 mHz.

Open image in new window

Spruit and Bogdan (1992) suggested that the partial conversion of the incoming acoustic waves into slow magnetoacoustic waves that propagate downward, channelled by the magnetic field, may explain the observations. Cally and Bogdan (1993), Bogdan (1997), and Cally et al. (2003) showed using numerical models that this mechanism is indeed capable of producing strong absorption coefficients, as defined by Braun. The best agreement is obtained when the magnetic field is inclined to the vertical, to simulate a spreading magnetic field with height (Crouch and Cally, 2003; Cally et al., 2003). In these models the maximum absorption is near 30° inclination.

The acoustic absorption by magnetic regions was confirmed by application of acoustic imaging (Chang et al., 1997) to data from TON. Acoustic holography, applied to MDI data, is ideal for the detection of sources and sinks of acoustic waves on the Sun. Braun and Fan (1998) discovered a region of lower acoustic emission in the 3–4 mHz frequency band which extends far beyond the sunspots (the ‘acoustic moat’). Acoustic moats extend beyond magnetic regions into the quiet Sun. In addition, Braun and Lindsey (1999) discovered high-frequency emission (‘acoustic glories’) surrounding active regions. Figure 48 shows egression power maps of the active region AR 8179 averaged over 24 hr, at different focal depths. The acoustic glory is seen in panel (b) as a bright halo of excess 5 mHz emission surrounding the entire active region complex. Isolated sunspots usually do not have any acoustic glory. The acoustic glory should not be confused with the ‘acoustic halos’ (Doppler acoustic power maps near 6 mHz) that surround all magnetic features (Braun et al., 1992). Neither acoustic halos nor acoustic glories are understood phenomena.

Open image in new window

We note that acoustic absorption by sunspots may also be studied with local power maps from ring-diagram analysis (e.g., Rajaguru et al., 2001; Howe et al., 2004), and by comparing cross-correlation amplitudes for incoming and outgoing waves in time-distance helioseismology (Duvall, 2002, private communication).

5.2.5 Phase shifts and wave-speed perturbations

The scattering phase shifts induced by sunspots were first measured with Hankel analysis by Braun et al. (1992) and Braun (1995). The m-averaged difference between incoming and outgoing p-mode phases is plotted in Figure 49 for sunspot AR 5254 observed in 1988 at the south pole (Braun, 1995). The f-mode phase shifts could not be determined accurately. The acoustic phase shifts are generaly positive and increase with frequency (or harmonic degree) to reach values in excess of 90°. Notice that the increase with frequency is not linear: The rate of increase is smaller at smaller frequencies. This could indicate that the origin of the scattering phase shifts is due to perturbations that are confined to shallow depths (Braun, 1995). The phase shifts are also observed to decrease with |m|, suggesting that scattering occurs over a small area (Braun, 1995). Fan et al. (1995) estimate that the phase shifts are mostly happening in the region inside the outer edge of the penumbra. Braun (1995) find no significant phase shifts in plage regions. Both the m and frequency dependence of the phase shifts have been modelled successfully by Cally et al. (2003, Figure 49). This provides a strong confirmation of the central physical mechanism that they advocate: partial conversion of incoming waves into downward propagating slow magnetoacoustic waves. We note that organized flows in and around sunspots should also introduce small phase shifts.

Open image in new window

Chen et al. (1998) used acoustic imaging to map phase shifts as well. They measured phase shifts between ingoing waves and outgoing waves and found a difference between quiet and active regions. Their maps correlate sharply with the surface magnetic field. According to Chou (2000), the phase shifts represent the phase perturbation accumulated along the wave path, which can be due to a change in the phase velocity, wave path, but also magnetic field and changes in the mode cavity. Figure 50 shows phase-shift maps focusing at the solar surface obtained with TON data (Chou et al., 1999). The maps confirm the original findings obtained with Hankel analysis. The acoustic imaging technique is very similar to phase-sensitive holography based on the correlation between the egression and the ingression (Lindsey and Braun, 1997; Braun and Lindsey, 2000, and Section 4.4). The two techniques give consistent results. Braun and Lindsey (2000) report that the phase shifts (or reduced travel time perturbations) increase in amplitude like the logarithm of the surface magnetic flux density.

Open image in new window

The travel time perturbations below active regions was studied by Kosovichev et al. (2000, 2001) using time-distance helioseismology. Mean travel times are usually interpreted in terms of wave speed. Kosovichev et al. (2000) found that the absolute difference in wave-speed between a sunspot (AR 8131) and the quiet Sun is up to 1 km s⁻¹. For reference, the quiet Sun sound speed is about 20 km s⁻¹ at a depth of 4 Mm and 35 km s⁻¹ at 10 Mm. The relative change in the squared sound speed is about (δc²/c² ≃ −0.1 at a depth of 4 Mm, which they say would correspond to a 10% temperature decrease. At greater depths (7–15 Mm), the sound-speed perturbation switches sign and becomes positive. This increase in sound-speed could be due to a temperature change or the direct effect of the magnetic field. The perturbations vanish at depths greater than 15 Mm, which may be an indication of the depth extent of active regions or perhaps of the poor resolution of the inversions there. Kosovichev et al. (2000) also detected the presence of narrow sound-speed anomalies (termed ‘fingers’) that connect internally the sunspot with two nearby pores (confirmed by Couvidat et al., 2004). Kosovichev and colleagues studied several other active regions, including AR 9393 which was seen on the disk for more than 3 months in 2001 (with a different name at each disk passage). The sound-speed perturbations under this giant active region are shown in Figure 51 (movie).

Open image in new window

Surface Doppler velocity measurements are not always reliable in highly magnetized regions. Following a suggestion by Duvall Jr (1995), Duvall Jr and Kosovichev (2001) have implemented a travel time averaging scheme to focus below a sunspot without using surface measurements directly inside the sunspot. This is achieved by considering all the rays that intersect at a target location in the solar interior (deep focusing technique). The standard and deep-focusing time-distance techniques seem to give consistent measurements of wave-speed perturbations (Duvall Jr and Kosovichev, 2001). Another way to avoid contamination is to perform sound-speed inversions of selected travel times that involve only data outside sunspots (Zhao and Kosovichev, 2003b). Figure 52 shows a comparison of sound-speed inversions below a sunspot obtained with and without cropping the data in the sunspot (GONG data; Hughes et al., 2005). The results are qualitatively similar in the two cases, i.e., the sign of the sound-speed perturbation as a function of depth is preserved. General agreement is also found between GONG and MDI for both travel times (Rajaguru et al., 2004) and inversion results (Hughes et al., 2005).

Open image in new window

Kosovichev et al. (2000) studied the emergence in time of an active region. Their results show that the wave-speed perturbations rise very fast across the upper 18 Mm of the convection zone. An analysis for 2 hr time steps suggests that the emerging magnetic flux travels the upper 10 Mm in less than 2 hr, implying a minimum speed of 1.3 km s⁻¹. Jensen et al. (2001) inferred the wave-speed structure beneath the same emerging active region using an inversion technique based on Fresnel-zone sensitivity kernels, as opposed to ray-based kernels. Their results are similar to those of Kosovichev et al. (2000).

Using local mode frequency differences between active and quiet Sun regions measured with the ring-diagram technique (see, e.g., Hindman et al., 2000; Rajaguru et al., 2001), Basu et al. (2004) performed structure inversions to estimate the sound speed and adiabatic index in active regions. The results are for regions with horizontal size 15° and 7-day averages. An attempt was made to remove the uncertainty in the modelisation of the surface layers (Section 4.2.3): inversions do not provide information for depths less than 1 Mm. The uncertainties in modeling the surface layers were ‘removed’ in this process (Section 4.2.3), such that no reliable information at depths of less than 1 Mm could be extracted.

The relative difference of the sound speed between active and quiet regions is plotted in Figure 53 for 12 different active regions. For active regions with strong magnetic fields, a reduction in sound-speed is observed in the upper layers (for depths less than, say, 7 Mm). At greater depths, the sound speed is increased relative to the quiet Sun. This is consistent with the work of Kosovichev et al. (2000) based on time-distance helioseismology. The amplitude of the perturbations are much smaller here than in the time-distance case simply because the magnetic region covers only a few percent of the area of ring-diagram analysis. Because the analysis of Basu et al. (2004) is based on the physics of normal modes, it provides an independent check of the results obtained using time-distance helioseismology.

Open image in new window

The wave-speed anomalies below sunspots could be caused by a variety of physical effects, for example thermal and magnetic perturbations. It has not been possible to disentangle these effects yet. The low sound-speed regions just below the surface have been attributed to a smaller temperature (Kosovichev et al., 2000; Basu et al., 2004). On the other hand, the higher wave speeds measured at a depth of 10 Mm below sunspots are unlikely to be due only to the direct effect of the magnetic field (or it would imply very large field strengths of a several tens of kG). The likely cause is a combination of magnetic and structural/thermal effects (Brüggen and Spruit, 2000; Basu et al., 2004). Barnes and Cally (2001) and Cally et al. (2003), however, have questioned the interpretation of travel time anomalies in terms of linear perturbations to the wave speed. Their numerical simulations of wave propagation through a model sunspot have been able to reproduce the phase shifts measured by Hankel analysis without the need for a thermal perturbation. This stresses the need for a proper solution of the forward problem of time-distance helioseismology in sunspots.

Lindsey and Braun (2003) argued that strong magnetic fields near the photosphere introduce large phase shifts in waves passing upwards into the photosphere of active regions and termed this effect the “showerglass” effect. They argue that the showerglass makes measurements of local variations in the subsurface more difficult, and thus attempts should be made to correct for this effect before inverting for subphotospheric structure and flows. Measurements of the showerglass were made by Lindsey and Braun (2003), and in more detail by Lindsey and Braun (2005a). In both cases the local egression and ingression control correlations (Section 4.4.3) were computed in active regions using MDI data. Maps of the phase and amplitude of the control correlations showed a clear relationship with the line-of-sight magnetograms, suggesting that the surface magnetic field was altering the amplitudes and phases of the waves used to compute the control correlation. Lindsey and Braun (2005b) used the measurements of the phase shifts of Lindsey and Braun (2005a) to correct the data before doing phase-sensitive holography (Section 4.4.5). They suggest that with the effect of near-surface magnetic field removed there is no clear evidence for sound-speed perturbations at depths greater than 5 Mm below a sunspot and that the effect of strong photospheric magnetic field on local helioseismic measurements should be studied further.

5.2.6 Far-side imaging

The first observational results from far-side helioseismic holography are due to Lindsey and Braun (2000b). Their results showed that active regions on the far-side of the Sun introduce travel time deficits compared to the quiet Sun of order ten seconds. The demonstration by Lindsey and Braun (2000b) that far-side imaging was practical is important for space weather predictions, as it allows about a week of warning before an active region will be seen at the East limb. Braun and Lindsey (2001) expanded the original idea presented by Lindsey and Braun (2000b) to allow imaging not just of the central region of the far-side of the Sun, but also the regions nearer to the limb. Daily far-side images computed from MDI data are available from the web at http://soi.stanford.edu/data/farside/. Figure 55 shows far-side images computed by Braun and Lindsey (2001) for subsequent days together with magnetograms for the visible part of the disk. Active regions seen on the far-side can be seen in the synoptic magnetogram of the subsequent Carrington rotation. Figure 54 and the companion movie show simultaneously the front-side and farside of the Sun during the period March–June 2001 which saw the evolution of AR 9393.

Open image in new window

Open image in new window

The sign of the travel time perturbation associated with active regions seen by far-side imaging is consistent with increased wave speeds or shorter paths traveled by the waves. This is consistent with the results of phase-sensitive holography of plage regions on the visible disk (see, e.g., Braun and Lindsey, 2000) and may be related to the observed solar cycle variation of normal mode frequencies (see, e.g., Lindsey and Braun, 2000b, and references therein).

5.2.7 Excitation of waves by flares

Kosovichev and Zharkova (1998) used data from the SOHO/MDI instrument (Scherrer et al., 1995) to make the first clear observations of a helioseismic wave produced by a flare. A roughly circular wave front, though with not insignificant quadrupole component, was seen emerging from the location of the now famous Bastille Day flare, an X-class flare of 1996. The wave packet was seen out to a distance of 120 Mm from the source and took about 55 min to travel that distance (see Figure 56). Earlier attempts using ring diagrams (Haber et al., 1987) and Hankel analysis (Braun and Duvall, 1990) reported small, but not significant, variations of the p-mode signal in flaring regions.

Open image in new window

Later, Donea et al. (1999) used acoustic power holography (see Section 4.4.4) of SOHO/MDI data to investigate the same flare as was studied by Kosovichev and Zharkova (1998). The acoustic power holography showed that the wave source was strongest around 3 mHz, though the signal-to-noise ratio was higher around 6 mHz. The signal in the 6 mHz band appeared about 4 min after the signal in the 3 mHz band. Donea and Lindsey (2004) found significant acoustic power signatures associated with two other flares and showed a possible connection between the fine-scale spatial structure of the acoustic power maps and motions of the footpoints of the flaring loops. We note that detection of flare-induced acoustic waves using ring-diagram analysis has been claimed by Ambastha et al. (2003).

The basic mechanism by which a flare excites helioseismic waves is not entirely clear. Wolff (1972) suggested that the energy released in the flare heats the atmosphere, the expansion of which in turn causes a downward propagating wave. Zharkova and Kosovichev (1998) argue, on the basis of the momentum needed to explain the observed Bastille Day flare wave amplitude, that the shock wave created by chromospheric evaporation may not be the complete explanation and suggested that perhaps momentum transfer by electron or proton beams impacting directly on the photosphere may be important. Zharkova and Kosovichev (2001) argued that electron beams can carry sufficient momentum to explain the observed amplitude of the helioseismic wave caused by the Bastille Day flare.

A separate issue is the mechanism that leads to the initiation of solar flares. The contribution of local helioseismology to this problem has been rather poor so far. We note that Dzifcakova et al. (2003) have searched for a relationship between the evolution of subsurface flows and flaring activity.

5.3.1 Paradigms

Granulation, with a typical size of 1.5 Mm, is well understood as a convective phenomenon and can be studied with realistic numerical simulations (see, e.g., Stein and Nordlund, 2000). Supergranules, however, have remained puzzling since their detection by Hart (1954, 1956). In a classic paper, Leighton et al. (1962) reported “large cells of horizontally moving material distributed roughly uniformly over the entire solar surface” with a characteristic scale of 30 Mm that are outlined by the chromospheric network; they suggested that the cells are a “surface manifestation of a supergranulation pattern of convective currents”. Unfortunately, there is no accepted theory that explains why solar convection should favor a 30 Mm scale. The solar convection zone is so highly turbulent and stratified that numerical modeling at supergranular scales has remained elusive (see, e.g., Rincon, 2004, and references therein).

Simon and Leighton (1964) presumed that a convective instability due to the recombination of ionized Helium is at the origin of the distinct supergranular scale. Forty years later, this hypothesis has not been proved to be right or wrong. The depth of the supergranulation layer is not known either. We note that Parker (1973) believes that the effect of stratification on convection may imply that supergranules are a deep phenomenon, with depths in excess of their horizontal diameters. Antia and Chitre (1993) investigated the stability of linear convective modes in the solar convection zone and found that a few convective modes dominate the power spectrum, one of which was identified as supergranulation. Using 1D numerical caculations, Ploner et al. (2000) suggested that the scale of supergranulation may be due to the interaction and merging of individual granular plumes (see also Rast, 2003). A somewhat related model was proposed by Rieutord et al. (2000, 2001) whereby supergranulation is the result of a non-linear large-scale instability of the granular flow, triggered by exploding granules. In the two previous models, supergranulation is not a proper scale of thermal convection. Regarding the influence of supergranular flows on magnetic fields, it is well established that a stationary cellular flow tends to expel the magnetic field from the regions of fluid motion and concentrate the magnetic flux into ropes at the cell boundaries (Parker, 1963; Galloway et al., 1977; Galloway and Weiss, 1981).

On the observational side, the original work of Leighton and coworkers has been refined. A variety of methods have been used to characterize the distribution of the cell sizes. A characteristic scale can be obtained from the spatial autocorrelation function (see, e.g., Hart, 1956; Simon and Leighton, 1964; Duvall Jr, 1980), the spatial Fourier spectrum (see, e.g., Hathaway, 1992; Beck, 1997), and segmentation or tessellation algorithms (Hagenaar et al., 1997). Although definitions vary, average cell sizes are in the range 15–30 Mm. It is unclear whether there is a variation of cell sizes with latitude: Rimmele and Schroeter (1989) and Komm et al. (1993a) report a possible decrease with latitude, Berrilli et al. (1999) an increase, and Beck (1997) no significant variation. The typical lifetime of the supergranular/chromospheric network is found to be in the range 1–2 d (e.g Rogers, 1970; Worden and Simon, 1976; Duvall Jr, 1980; Wang and Zirin, 1989). The rms horizontal velocity of supergranular flows is known to be about 300 m s⁻¹ (see, e.g., Hathaway et al., 2000). The vertical component of the flows, however, has been extremely difficult to measure (see, e.g., Giovanelli, 1980) or infer (November, 1989). Indeed, Miller et al. (1984) caution that Doppler velocity measurements at the cell boundaries may be polluted by the network field. Estimates of the rms vertical flow are provided by Chou et al. (1991) and Hathaway et al. (2002) who find speeds of about 30 m s⁻¹ (the topology of the vertical flows is largely unknown). Even more difficult to measure are the related temperature fluctuations. Observers have searched for the thermal signature of a convective process, i.e., rising hot material at the cell centers and sinking cool material at the cell boundaries. Unfortunately, answers vary too widely (see Lin and Kuhn, 1992, and references therein).

Estimates of the pattern rotation rate of supergranulation can be obtained by correlation tracking techniques. Duvall Jr (1980) and Snodgrass and Ulrich (1990) used Doppler scans separated in time by Δt = 24 hr. The results indicate that the equatorial rotation of the supergranulation pattern is faster than the spectroscopic rate (Snodgrass and Ulrich, 1990) by about 4% and faster than the magnetic features (Komm et al., 1993b) by about 2%. Duvall Jr (1980) also considered same-day scans with Δt = 6 hr and found a slightly smaller pattern rotation rate than for Δt = 24 hr.

It was suggested by Foukal (1972) that the difference between the rotation of magnetic features and the spectroscopic rate may be due to magnetic structures being rooted in deeper, more rapidly rotating layers. Supergranules must also sense increase rotation in the shear layer below the surface. However, Beck (2000) remarks that the pattern rotation of supergranulation measured from correlation tracking with Δt = 24 hr is significantly faster than the rotation of the solar plasma measured by helioseismology at any depth in the interior. It is especialy puzzling that the rotation of the magnetic network is significantly less than that of the supergranular pattern, since small magnetic elements are believed to be advected by supergranular flows. Hathaway (1982) suggested that a faster supergranular rate may be a direct consequence of the interaction of convection and rotation.

As explained in the following sections, local heliososeismology has become a powerful tool to study the structure and evolution of supergranular flows. In fact, helioseismological measurements have transformed our knowledge of the dynamics of supergranulation.

5.3.2 Horizontal flows and vertical structure

In order to resolve structures at supergranular scales, Duvall Jr et al. (1996) measured p-mode travel times at distances shorter than 10 Mm over an 8.5 hr time interval. Directional information was obtained by cross-correlating a point on the surface with surrounding quadrants centered on the four cardinal directions. In a first order approximation, the south-north and east-west travel time differences were converted into an apparent horizontal vector flow field without inversion. Duvall Jr et al. (1997) found that the line-of-sight projection of the inferred flow field is highly correlated with the mean MDI Dopplergram (correlation coefficient 0.74). This comparison was initialy used as a validation of the time-distance technique.

Three dimensional inversions of quiet-sun travel times have been presented by Duvall Jr et al. (1997), Kosovichev and Duvall Jr (1997), and Zhao and Kosovichev (2003a). Tests of the ray-based inversion procedure of Zhao and Kosovichev (2003a) show that the horizontal components of the velocity can be infered with some confidence in the upper 5 Mm (and perhaps down to 10 Mm). The small vertical flows, on the other hand, cannot be inferred reliably near the surface due to significant “cross-talk” with the horizontal divergence signal.

Near-surface horizontal flows have been measured at a depth of 1 Mm with f-mode time-distance helioseismology (Duvall Jr and Gizon, 2000). In this case travel time differences are directly sensitive to horizontal flows because f modes propagate horizontaly. The f-mode time-distance technique gives results that are comparable to correlation tracking of granulation (De Rosa et al., 2000). Shown in Figure 22 is an inversion of f-mode travel times that employs 2D Born kernels (Gizon et al., 2000). The correlation coefficient between the estimated line-of-sight velocity and the surface Doppler image is about 0.7.

The holographic technique also measures flows at supergranular scales. As shown earlier in Figure 45, supergranules are easy to identify as regions of outflows for a 3 Mm focus depth (24 hr time average).

A practical way to detect and display supergranulation is to plot the horizontal divergence of the flow field. Time-distance helioseismology applied to f modes is particularly well adapted to measuring the horizontal divergence of the velocity near the surface. Duvall Jr and Gizon (2000) measured the time it takes for solar f modes to propagate from any given point on the solar surface to a concentric annulus around that point. The difference in travel times between inward and outward propagating waves is a proxy for the local horizontal divergence. An example of the divergence signal (inward minus outward travel time) for one 8-hour interval analyzed is shown in Figure 57 with magnetic field information overlaid (MDI full-disk data). A white, or positive signal, corresponds to a horizontal outflow. From the size of the features present, their lifetime, and the presence of the magnetic field in the dark lanes regions, supergranulation is identified as the main contributor to the signal. This can also be seen by making a spatial power spectrum of the divergence signal: Power peaks near degree l = 120. The movie associated with Figure (57) shows that the magnetic features stay in the regions of horizontal convergence over the time interval of the observations (one week).

Open image in new window

Local heliososeismology opens prospects for mapping the structure of supergranular flows below the surface. A major goal is to answer the long-standing question of how deep supergranular flows persist below the surface.

Duvall Jr et al. (1997) and Kosovichev and Duvall Jr (1997) presented three-dimensional inversions of quiet-sun travel times using ray-theoretical sensitivity kernels. Figure 58 shows a vertical cut through the velocity field; the horizontal sampling is 4.3 Mm and the vertical sampling is about 1 Mm. In the upper 5 Mm the flows are positively correlated with the direct surface observations. To assess the average structure of superganular flows with depth, Duvall Jr (1998) used the inversion results from Duvall Jr et al. (1997) to measure the correlation of horizontal flows between the surface and any given depth. As seen in Figure 59, the correlation drops to zero at a depth of 5 Mm and is negative in the range 5–8 Mm, suggesting the existence of a “return flow”. Since the correlation seems to disappear below 8 Mm, Duvall Jr (1998) concluded that the depth of supergranulation is 8 Mm. The same type of analysis was repeated by Zhao and Kosovichev (2003a) with more recent (and presumably more reliable) inversions of travel times. They found that the correlation of the horizontal divergence falls off away from the surface, changes sign at a depth of about 6 Mm, and is negative at depths in the range 6–14 Mm (see Figure 59). In particular, Zhao and Kosovichev (2003a) find a correlation coefficient of −0.5 at a depth of 10 Mm. Zhao and Kosovichev (2003a) concluded that supergranulation is about 15 Mm deep.

Open image in new window

Open image in new window

We note that inversion results heavily rely on the assumed travel time sensitivity kernels. While most inversions use ray-based kernels, Jensen et al. (2000) and Birch and Kosovichev (2000) showed that finite-wavelength effects must be taken into account. It may be that wave-based kernels could yield depth inversion results that are qualitatively different from the ones presented above.

Braun and Lindsey (2003) used a 24 hr MDI data cube and seismic holography to make maps of convective flows at focus depths from 3 Mm to 37 Mm. The horizontal-divergence maps are shown in the left panels of Figure 60. The behavior of the divergence maps with depth can be studied by making scatter plots of the divergence values observed at a given depth versus the near-surface values. The slope of a line fit to the scatter is observed to change sign at 10 Mm and to remain negative in the range 10–37 Mm. Although it is tempting to interpret the reversal of the horizontal divergence signal at focus depths greater than 10 Mm, Braun and Lindsey (2003) suspect that this reversal could be due to the increasing contribution of oppositely directed surface flows as the pupil increases in size with focus depth. To test this hypothesis, they considered a control experiment in which the observed flows at 3 Mm depth are used to estimate the surface contribution to the seismic measurements at greater focus depths. The divergence maps of the control experiment, which are shown in the right panels of Figure 60, ressemble the observed divergence maps (left panels). The good correspondence between the two suggests that the observations at depths are mostly due to leakage from the surface (the “showerglass effect”). Modelling efforts by Braun et al. (2004) have confirmed that the change of sign of the correlation below 10 Mm is a predominantly surface contamination of the velocity signal from neighboring supergranules. However, the fact that the control data do not reproduce exactly the observations may indicate that there is at least some sensitivity to velocities at depth.

Open image in new window

Open image in new window

It would appear that local helioseismology has not yet provided a definitive answer regarding the depth of supergranulation.

5.3.3 Rotation-induced vorticity

Since the typical lifetime of supergranules is significantly less than the solar rotation period, the influence of rotation on supergranular convection is expected to be small. Solar rotation effects in supergranules were illustrated by Hathaway (1982) in non-linear numerical simulations. The Coriolis force causes divergent and convergent horizontal flows to be associated with vertical components of vorticity of opposite signs (on average). In the northern hemisphere, cells rotate clockwise where the horizontal divergence is positive, while they rotate counterclockwise in the convergent flow towards the sinks. The sense of circulation is reversed in the southern hemisphere.

The effect of rotation on supergranules was first observed by Duvall Jr and Gizon (2000) and Gizon and Duvall Jr (2003)². They computed the vertical vorticity and horizontal divergence from vector flow maps obtained with f-mode time-distance helioseismology (1 Mm deep, 8 hr time invervals). The vertical large-scale vorticity due to differential rotation and meridional circulation was removed to study the vorticity at supergranular scales. Panel (a) of Figure 62 shows the correlation coefficient between div and curl at latitude λ,

$$C(\lambda) = {{\langle {{\rm{div}}\;{\rm{curl}}} \rangle} \over {\sqrt {\langle {{\rm{di}}{{\rm{v}}^2}} \rangle \langle {{\rm{cur}}{{\rm{l}}^2}} \rangle}}},$$

(93)

where the angle brackets denote a spatial average over the area of a 10° longitudinal strip centered around A. In the north, positive (negative) divergence is correlated with clockwise (anticlockwise) vorticity; the correlation changes sign in the south. Thus, away from the equator, the number of right-handed cyclones is not equal to the number of left handed cyclones. The sign and the latitudinal variation of C(λ) are both characteristic of the effect of the Coriolis force on the flows. Panel (b) of Figure 62 shows horizontal averages of the vertical vorticity 〈curl〉_± versus f(λ), where the averages 〈·〉₊ and 〈·〉₋ are restricted to the regions of positive and negative divergence, respectively. Note that there is a small difference between the total area covered by regions of positive and negative divergence (Duvall Jr and Gizon, 2000). A nearly perfect linear relationship between 〈curl〉_± and ∓f(λ) is observed, with 〈curl〉_± = ∓f(λ) Ms⁻¹. This is again consistent with the interpretation as a Coriolis effect.

Open image in new window

5.3.4 Pattern evolution

As mentioned above, the pattern rotation rate of supergranulation derived from correlation tracking of Doppler scans appears to be significantly faster than the spectroscopic rate (for a 24 hr time lag). Correlation tracking algorithms can also be applied to images of supergranular flows derived from local helioseismology, such as the horizontal divergence signal. Unlike raw Doppler images, the divergence signal has uniform sensitivity across the solar disk and is subject to few systematic errors.

Gizon and Duvall Jr (2003) used a 60-day sequence of 90° × 90° divergence maps obtained with f-mode time-distance helioseismology (0.24° spatial sampling). The original MDI Doppler velocity images were tracked at the Carrington rate to remove the main component of rotation. Small regions, apodized by a Gaussian surface with a full width at half maximum of 3.84° and separated by a time-lag Δt, were spatially cross-correlated. For Δt < 1.5 d the peak of maximum correlation is easy to identify, and the spatial displacement Δx at correlation maximum gives an apparent pattern velocity U_LCT = Δx/Δt, where the subscript LCT stands for ‘local correlation tracking’. Figure 63 shows that the apparent motion of the supergranulation pattern depends on Δt. In particular, the pattern rotation rate increases rapidly with Δt. The apparent meridional motion is poleward for Δt < 18 hr with a very small amplitude (< 5 m s⁻¹), while, suprisingly, it is slightly equatorward for greater time lags. The apparent rotation of supergranulation obtained by tracking the divergence maps confirms the original findings of Duvall Jr (1980) and Snodgrass and Ulrich (1990) obtained by tracking Dopplergrams.

Open image in new window

Beck and Duvall Jr (2001) showed that the observed correlation function of the divergence signal switches sign at Δt ∼ 40 hr. The same analysis is repeated here with the 60-day sequence used by Gizon and Duvall Jr (2003). Figure 64 shows a still at Δt = 24 hr from a movie of the temporal evolution of the spatial correlation, with Δt varying from zero to 5.5 d (equatorial region, 6 hr time steps, Carrington tracking rate). The spatial cross-correlation decays and oscillates as a function of time. For Δt < 36 hr, the peak of maximum correlation is well defined and is surrounded by a ring of negative correlation due to adjacent supergranules. For Δt > 48 hr a peak of negative correlation has appeared (east), which is surrounded by a ring of positive correlation. When the analysis is repeated away from the equator, one finds that the peak of negative correlation forms poleward of the east direction. For Δt ≃ 5.5 d the main correlation peak is again positive and is surrounded by a ring of negative correlation.

Open image in new window

Earlier observations of solar convection assumed that supergranulation can be characterized by an autocorrelation function that exhibits a simple exponential decay in time (Harvey, 1985; Kuhn et al., 2000). It is now obvious that supergranulation does not follow such a simple model. The oscillation period of the correlation function, of the order of 6 d, suggests an underlying long-range order. As shown below, these puzzling observations are easier to describe in Fourier space (Section 5.3.5).

Another interesting method of analysis to study the evolution of the supergranulation pattern is described by Lisle et al. (2004). They used horizontal-divergence maps obtained from local correlation tracking of granules (1 min time lag). Lisle et al. (2004) constructed temporal averages (8 d) of the divergence maps for different tracking velocities v _t . The spatial variations of a 15° × 15° section of the time-averaged divergence map were characterized by the rms values in the east-west (σ _x ) and south-north (σ _y ) directions. Lisle et al. (2004) found that at the equator the ratio σ _x /σ _y is maximum at a tracking velocity U_LRT which is 110 m s⁻¹ above the Carrington velocity, i.e., faster than any rotation rate measured so far. They suggested that a ratio σ _x /σ _y greater than 1 is an indication that supergranules are preferentially aligned in a north-south direction. Junwei Zhao has recently repeated their analysis using time-distance divergence maps. Figure 66 and the companion Movie ?? shows a 15° × 15° time-averaged divergence map for various tracking velocities. Figure 66 shows the ratio σ _x /σ _y at the equator as a function of tracking velocity, averaged over all available data. The maximum occurs at a tracking rate 123 m s⁻¹ above the Carrington velocity. It is evident that there is another local maximum 8 m s⁻¹ below the Carrington velocity. This local maximum could only be hinted at from the plots of Lisle et al. (2004) due to lack of averaging. We will come back to this observation near the end of the next section.

Open image in new window

Open image in new window

5.3.5 Traveling-wave convection

The 3D power spectrum of the divergence signal was studied first by Gizon et al. (2003). The same analysis was extended by Gizon and Duvall Jr (2004) to cover the period from 1996 to 2002. They considered series of MDI full-disk Dopplergrams from the Dynamics campaigns (two to three months each year). Dopplergrams were tracked at the Carrington angular velocity to remove the main component of rotation. Every 12 hr, a 120^° × 120^° map of the horizontal divergence of the near-surface flows was obtained using f-mode time-distance helioseismology. At a given target latitude λ a longitudinal section of the data was extracted, 10° wide in latitude. Gizon and Duvall Jr (2004) rearranged the data in a frame of reference with angular velocity Ω_mag(λ) = 14.43° − 1.77° sin² λ − 2.58° sin⁴ λ d⁻¹ (rotation rate of small magnetic features; Komm et al., 1993b). This choice of reference is convenient, although arbitrary. The position vector is denoted by x = (x, y) in the neighborhood of latitude λ, where coordinate x is prograde and y is northward, with a spatial sampling of 2.92 Mm in both coordinates.

Figure 67 shows a cyclindrical cut in the equatorial power spectrum at a constant k typical of the supergranulation, and a model fit to the data according to Equations (95, 96, 97). For each azimuth ψ, the power has two peaks at frequencies ω₊ = ω _r (k) and ω₋ = −ω_r(−k). Observations show that the difference ω₊ − ω₋ = 2ω₀(k) is independent of azimuth (Gizon et al., 2003). No Galilean transformation can cause these peaks to coalesce, at zero frequency or otherwise. This implies that supergranulation undergoes oscillations. In the plane power is distributed along two parallel sinusoids at frequencies ±ω₀ + kU _x cos ψ + kU _y sin ψ. The presumption by Rast et al. (2004) that power is distributed along two intersecting sinusoids must be rejected: It is inconsistent with the observations. As one does in helioseismological ring analysis (Schou and Bogart, 1998), it is natural to interpret the frequency shift k · U as a Doppler shift produced by an advective flow U (some average flow in the supergranulation layer). The weak k-dependence of the measured U in the range 40 < kR_⊙ < 180 is consistent with this interpretation. Furthermore the functional form of the power is preserved for different tracking rates (i.e., different latitudes). As shown in Figure 68, it is remarkable that the inferred rotation U _x and meridional circulation U _y are both similar to that of the small magnetic features (Komm et al., 1993b, c). This property is consistent with the view that magnetic fields are advected by supergranular flows. In addition, the torsional oscillations are seen with the same phase and amplitude as in global helioseismology. The temporal changes in the meridional circulations are also consistent with an inflows toward the mean latitude of activity (see Section 5.3.4).

Open image in new window

Open image in new window

The dynamics of the supergranulation is best studied once the background flow U has been removed. In a co-moving frame, each spatial component oscillates at a characteristic frequency ω₀. There is a clear relationship between ω₀ and the wavenumber k, well described by a power law (see panel (a) of Figure 69):

$${{{\omega _0}} \over {2\pi}} \simeq 1.8{\left({{{k{R_\odot}} \over {120}}} \right)^{0.45}}\mu \;{\rm{Hz}}.$$

(99)

This is a fundamental relationship as it is measured to be essentialy independent of ψ, λ, and the phase of the solar cycle. The data are consistent with a spectrum of traveling waves with a dispersion relation ω = ω₀(k). The azimuthally averaged power spectrum is shown in panel (b) of Figure 69. In the range 60 < kR_⊙ < 160, the peak of power away from zero temporal frequency is well resolved since the quality factor Q = ω₀/γ is larger than 1. Near kR_⊙ = 120 we have Q ≃ 2. The variations of the oscillation frequency with latitude and time are less than 5% during the period 1996–2002 for kR_⊙ = 115. Since ω₀ and the dominant size of supergranules are observed to be essentially independent of latitude, the general dynamics determining the time scale and the spatial scale of supergranulation is not affected by the Coriolis force associated with the large scale vorticity (rotation). However, there is a pronounced anisotropy in the azimuthal distribution of wave power at fixed k. The power anisotropy is measured to be η ≃ 1 for kR_⊙ = 115. Furthermore, power is maximum in a direction equatorward of the prograde direction in both hemispheres: For |λ| < 25° observations show ψ_max ≃ −λ, and for |λ| > 25° ψ_max ≃ −25°signλ. The pattern therefore senses the effect of rotation. A snapshot of the divergence field would not reveal this as the sum of the powers measured in opposite directions is nearly isotropic. We recall that the effect of rotation on supergranulation was detected in the vorticity field.

Open image in new window

The lifetime of supergranules is about 2.3 d at kR_⊙ = 115, i.e., a somewhat larger value than other estimates derived from the decay of the autocorrelation function. This is because the lifetime is not strictly given the decay of the correlation function at short time lags, but by the decay of the envelope of the correlation function, which oscillates. The lifetime decreases by about 20% at active latitudes (Gizon and Duvall Jr, 2004).

As mentioned earlier, estimates of supergranulation rotation obtained by tracking the Doppler pattern are systematically found to be higher than the rotation of the magnetic network (for large time lags, see panel (a) of Figure 63). This apparent superrotation of the pattern can now be understood as the result of the waves being predominantly prograde. The east-west motion of the pattern is effectively a power-weighted average of the true rotation and the non-advective phase speed u_w = ω₀/k ∼ 65 m s⁻¹ for kR_⊙ = 120. Similarly, the excess of wave power toward the equator is reflected in the equatorward meridional motion of the pattern at large time lags (see panel (b) of Figure 63) even though the advective flow is poleward. Correlation tracking measurements must take into account the fact that the pattern propagates like a modulated traveling wave.

It was a source of concern that the wavelike properties of supergranulation had not been seen previously in surface Doppler shifts. Schou (2003b) showed that the same phenomenon can be observed directly using MDI Dopplergrams, thereby confirming the observations of Gizon and Duvall Jr (2003). In addition to confirming those results, Schou (2003b) was able to extend the measurements for the rotation and meridional velocities, U _x and U _y , beyond ±70° latitude. We note that, earlier, Beck and Schou (2000) had also used a spectral method to estimate the equatorial rotation of supergranulation from surface Doppler images; this method, however, was incorrect as it did not take account of the full complexity of the power spectrum.

All the evidence shows that supergranulation displays a high level of organization in space and time. Although no serious explanation has been proposed yet, it would seem that supergranulation is an example of traveling-wave convection. The prograde excess of wave power is perhaps due to the influence of rotation (or rotational shear) that breaks the east-west symmetry, allowing for new instabilities to propagate (see, e.g., Busse, 2003, 2004). We note that convection in oblique magnetic fields also exhibits solutions that take the form of traveling waves (Hurlburt et al., 1996). Future work should focus on measuring the evolution of the pattern at different depth in the interior and the phase relationship between the different Fourier components. More than forty years after its discovery, supergranulation remains a complete mystery.