1: \documentclass[12pt,preprint]{aastex}
2: %\documentclass[manuscript]{aastex}
3:
4: %%%%%%%%%%%%%%
5: \tighten
6: \newcommand\be{\begin{equation}}
7: \newcommand\ee{\end{equation}}
8: \newcommand\jcd{Christensen-Dalsgaard}
9: %%%%%%%%%%%%%%%%%%%%%%%%%%%
10:
11: \usepackage{epsfig}
12: \begin{document}
13: \shortauthors{Antia, Basu \& Chitre}
14: \shorttitle{Solar rotation and its gradient}
15:
16: \title{Solar rotation rate and its gradients during cycle 23}
17: \author{H. M. Antia}
18: \affil{Tata Institute of Fundamental Research,
19: Homi Bhabha Road, Mumbai 400005, India}
20: \email{antia@tifr.res.in}
21: \author{Sarbani Basu}
22: \affil{Astronomy Department, Yale University, P. O. Box 208101,
23: New Haven CT 06520-8101, U.S.A.}
24: \email{basu@astro.yale.edu}
25: \and
26: \author{S. M. Chitre}
27: \affil{Centre for Basic Sciences, University of Mumbai, Mumbai 400098, India}
28: \email{kumarchitre@gmail.com}
29:
30: \begin{abstract}
31: Available helioseismic data now span almost the entire solar activity cycle 23
32: making it possible to study solar-cycle related changes of the
33: solar rotation rate in detail. In this paper we study how the solar
34: rotation rate, in particular, the zonal flows change with time.
35: In addition to the zonal flows that show a well known pattern
36: in the solar convection zone, we also study changes in the radial and
37: latitudinal gradients of the rotation rate, particularly
38: in the shear layer that is present in the immediate sub-surface layers of the Sun.
39: In the case of the zonal-flow pattern,
40: we find that the band indicating fast rotating region close
41: to the equator seems to have bifurcated around 2005.
42: Our investigation of the rotation-rate gradients show that the relative
43: variation in the rotation-rate gradients
44: is about 20\% or more
45: of their average values, which is much larger than the relative variation in
46: the rotation rate itself.
47: These results can be used to test predictions of various solar dynamo models.
48:
49: \end{abstract}
50:
51: \keywords{Sun: oscillations --- Sun: rotation --- Sun: interior}
52:
53: \section{INTRODUCTION}
54: \label{sec:intro}
55:
56: Helioseismology enables us to study dynamics of the solar interior
57: (e.g., Thompson et al.~1996; Schou et al.~1998, etc.). Early results have
58: shown that the differential rotation seen at the solar surface persists
59: till the base of the convection zone and becomes nearly uniform in the radiative
60: interior.
61: Also seen were two shear layers, one near the surface and the other near the
62: base of the convection zone. The latter, which is referred to as the tachocline
63: (Spiegel \& Zahn 1992), is generally believed to be the seat of the solar dynamo.
64: Superposed on the relatively smooth latitudinal variation are alternating
65: bands of slightly faster and slower rotation (Kosovichev \& Schou~1997).
66:
67: Data for the
68: last 12 years covering most of the solar cycle 23 are now
69: available from the Global Oscillation Network Group (GONG) project and
70: the Michelson Doppler Imager (MDI) instrument on board SOHO. As a result, it is now possible
71: to study temporal variations of solar dynamics in detail.
72: Early works that used subsets of the data (e.g., Kosovichev \& Schou 1997;
73: Schou 1999; Howe et al.~2000a; Antia \& Basu 2000)
74: had already established that there is significant variation in the rotation rate of the solar
75: interior. The temporal variation
76: manifests as a system of bands with faster or slower than average rotation
77: rate at fixed depths in the upper layers of the convection zone.
78: And particularly at low latitudes, the zonal flow bands appear to migrate towards the
79: equator with time, and appear to be similar in character to the torsional oscillations
80: discovered in the solar surface rotation rate (Howard \& LaBonte 1980;
81: Ulrich et al.~1988; Snodgrass 1992).
82: A subsequent work (Antia \& Basu 2001) found
83: that in addition to this pattern at low latitudes, there is another system
84: of zonal-flow bands at high latitudes ($> 45^\circ$) that migrate towards
85: the poles as the solar cycle progresses.
86: With accumulation of more seismic data it has now become
87: clear that these zonal flows penetrate through most of the convection zone
88: (Vorontsov et al.~2002; Basu \& Antia 2003, 2006; Howe et al.~2006a).
89: Since the errors in the results obtained by
90: inversions increase with depth, it is not yet clear whether there is any significant
91: temporal variation in the tachocline region near the base of the convection
92: zone --- a region where the solar dynamo is believed to be operating.
93: Howe et al.~(2000b, 2007) reported a 1.3 year oscillation in the equatorial region
94: around $r=0.72R_\odot$ during the rising phase of the solar cycle, but
95: that has not been confirmed by other studies (e.g., Antia \& Basu 2001;
96: Basu \& Antia 2003).
97:
98: The origin of torsional oscillations or zonal flows is not fully
99: understood, but since
100: the zonal flow banded structure and the magnetic activity pattern
101: seem to be closely related,
102: it is generally believed that zonal flows arise from a nonlinear
103: interaction between magnetic fields and differential rotation.
104: Thus the properties of solar zonal flows
105: should provide a constraint on theories of the solar dynamo.
106: Most dynamo models are kinematic, where the velocity field is
107: specified; such models do not yield any information about temporal variations
108: in the velocity field. However, there are some non-kinematic dynamo models that include the
109: Lorentz-force feedback on differential rotation, and these have been used to explain
110: the zonal flow patterns.
111: Covas et al.~(2000, 2004), for example, considered an axisymmetric mean-field dynamo model
112: to study temporal variations of the rotation rate and of magnetic fields in
113: the solar interior.
114: They found
115: temporal variations in the rotation rate that are qualitatively similar
116: to the zonal-flow pattern inferred from the seismic data.
117: Spruit (2003) suggested that thermal fluctuations due to enhanced
118: emission of radiation by small-scale magnetic fields
119: can drive the zonal-flow pattern. His model produces zonal flows
120: with amplitudes that decrease with increasing depth. Lanza (2007)
121: attempted to find the source of the zonal-flow pattern by using a model for
122: angular momentum transport in the convection zone and by comparing
123: the results with observed patterns. His preliminary results led to
124: constraints on the amplitude,
125: phase and location of the perturbations responsible for the observed torsional oscillations.
126: Rempel (2007) found that the poleward
127: moving branch of zonal flows at high latitude may be explained as a
128: response of the coupled differential rotation/meridional flow system to
129: periodic forcing in mid-latitudes by either Lorentz force or thermal
130: perturbations, with thermal fluctuations very likely playing a role in
131: the low-latitude equator-ward moving branch of the zonal flow.
132:
133:
134: Most seismic studies hitherto have investigated
135: temporal variations in the rotation rate itself. For the solar dynamo
136: models, however,
137: spatial gradients of this angular velocity are more important.
138: Although the radial gradient of the rotation rate in the outer shear layer
139: has been studied by
140: Schou et al.~(1998), Basu \& Antia (2001), Corbard \& Thompson (2002), etc.,
141: a detailed study of the temporal variations of both the radial and
142: latitudinal gradients has not been undertaken so far.
143: This motivated us to further investigate changes in both the radial
144: and the latitudinal gradients of the solar rotation rate with a view to
145: studying temporal variations of the shear pattern.
146: We also study the time-derivative of the rotation velocity, which
147: should relate to the driving forces that are possibly responsible for the zonal flow
148: pattern.
149:
150:
151: The rest of this paper is organized as follows: we summarize the data and the
152: analysis technique in \S~\ref{sec:data}; our results are described in
153: \S~\ref{sec:results}, and the discussion of our results along with conclusions
154: are given in \S~\ref{sec:disc}.
155:
156: \section{Data and technique}
157: \label{sec:data}
158:
159: We use data obtained by the GONG (Hill et al.~1996) and MDI
160: (Schou 1999) projects in the present work.
161: These data sets consist of the mean frequency and the splitting coefficients
162: for different $(n,\ell)$ multiplets.
163: Only the odd-order splitting coefficients are required for determining
164: the rotation rate in the solar interior (e.g., Ritzwoller \& Lavely 1991).
165: These splitting coefficients are sensitive only to the north-south symmetric
166: component of the rotation rate and hence, that is the only component that can
167: be determined with these data. Consequently, all results shown in this work are
168: symmetric about the equator. There could
169: be an antisymmetric component of rotation present in the Sun, but that cannot be analyzed using
170: the global-mode data used in this work.
171: We use 120 data sets from GONG, each set
172: covering a period of 108 days. The first set starts on
173: 1995, May 7 and the last set ends on 2007, May 15, with a
174: spacing of 36 days between consecutive data sets.
175: Thus there is a considerable overlap between neighboring data sets.
176: The GONG data sets include only p-modes up to $\ell=150$ and we use all modes
177: with frequencies between 1 and 3.5 mHz.
178: The MDI data sets consist of 56 non-overlapping
179: sets each obtained from observations taken over a period of
180: 72 days. The first set begins on 1996, May 1 and the last
181: set ends on 2007, October 6.
182: Contact with SOHO satellite was lost during June 1998
183: and the satellite was finally recovered during February 1999. During this
184: period there is only one data set and that covered a period of 60 days.
185: The MDI data sets include f-modes up to $\ell=300$ and p-modes up to
186: $\ell\approx200$. We use all modes
187: with frequencies between 1 and 3 mHz. Since the higher frequency modes often
188: cause problems, they have not been included in the analysis. The higher degree modes from
189: MDI are also known to give misleading temporal variations in some cases
190: (e.g., Antia et al.~2003). This is indeed found to be the case for
191: the radial gradient of the rotation rate and is discussed later.
192:
193: We use a two dimensional Regularized Least
194: Squares (2D RLS) inversion technique as described by
195: Antia et al.~(1998) to infer the rotation rate in
196: the solar interior from each of the available data sets.
197: The first 8 odd-order splitting coefficients, $a_1,a_3,\ldots,a_{15}$, were
198: used in the inversions.
199: In order to study the time-variation of the rotation rate we first
200: determine the Sun's internal rotation rate using each data set,
201: and then take the time average of the results obtained using all data sets at each latitude
202: and depth. Since we treat the GONG and MDI sets separately, we obtain
203: one average result using GONG sets and another using MDI sets.
204: To obtain the time-varying component of rotation, we subtract the mean from the
205: rotation rate at any given
206: epoch. Thus,
207: \begin{equation}
208: \delta\Omega(r,\theta,t)=\Omega(r,\theta,t)-
209: \langle \Omega(r,\theta,t)\rangle,
210: \label{eq:zonal}
211: \end{equation}
212: where $\Omega(r,\theta,t)$ is the rotation rate as a function of
213: radial distance $r$, latitude $\theta$ and time $t$.
214: Here the angular
215: brackets denote the average over the time duration for which data are available.
216: As is the case for the time-averaged rotation, the residual rotation rate is
217: calculated separately for GONG and MDI data sets.
218:
219: For most parts of this investigation we use the residual rotation velocity, $\delta v_\phi=
220: \delta\Omega r\cos\theta$, rather than the rotation rate $\delta\Omega$.
221: This time varying component, $\delta v_\phi$, is generally called the zonal flow.
222: The definition of zonal flows is somewhat ambiguous: some authors (e.g., Howard \&
223: LaBonte 1980)
224: have defined the zonal flows by subtracting the temporal average of
225: a smooth component of the rotation rate. The smooth component is generally
226: defined using three terms (constant, $\sin^2\theta$ and $\sin^4\theta$)
227: to calculate the latitudinal variation of rotation rate.
228: There is some difference in the resulting pattern in the two cases,
229: as has been described by Antia \& Basu
230: (2000). In this work we adopt the definition given by Eq.~(\ref{eq:zonal})
231: where the temporal mean of the full rotation rate is subtracted to obtain
232: the residuals. This definition has been used previously by other
233: investigators too (e.g., Howe et al.~2000a, 2005, 2006a; Antia \& Basu 2001;
234: Basu \& Antia 2003, 2006, etc.).
235:
236: For the purpose of investigating the spatial gradients of the rotation rate, we use
237: $\Omega$ rather than $v_\phi$ and calculate the radial gradient,
238: ($\Omega_{,r}\equiv\partial \Omega/\partial r$), as well as the latitudinal
239: gradient
240: ($\Omega_{,\theta}\equiv(1/r)\partial \Omega/\partial|\theta|$).
241: These derivatives are computed numerically.
242: Since the inferred
243: zonal flow pattern is symmetric about the equator, the radial gradient
244: is also symmetric. The latitudinal gradient on the other hand is antisymmetric
245: about the equator, and hence it is 0 at the equator. To restore the
246: symmetry of the latitudinal gradient we, have defined the gradient with respect
247: to the absolute value of the latitude. With this definition, both $\Omega_{,r}$ and $\Omega_{,\theta}$
248: are symmetric about the equator.
249: Since the 2D~RLS technique used for inversion gives results that are
250: smooth in $(r,\theta)$, there is no difficulty in computing these derivatives numerically.
251: The errors in these derivatives are, of course,
252: larger than those in $\Omega$ itself.
253: The time-varying component of these derivatives is calculated by
254: differentiating the residual in rotation rate defined by Eq.~(\ref{eq:zonal}),
255: i.e.,$\delta\Omega_{,r}\equiv\partial \delta\Omega/\partial r$ and
256: $\delta\Omega_{,\theta}\equiv(1/r)\partial \delta\Omega/\partial|\theta|$.
257:
258: We also calculate the derivative of the rotation velocity $v_\phi$ with respect to time
259: (note that $\partial v_\phi /\partial t=\partial\delta v_\phi/\partial t$),
260: since it is likely to give an indication of the forcing term that drives
261: the zonal-flow pattern.
262: For calculating time-derivatives numerically, we need $v_\phi$
263: at different times, which we obtain from independent data sets. It should be noted that
264: since the magnitude of the zonal-flow velocity is small, the
265: temporal derivative has significant errors. To get more reliable
266: estimates, we smooth $\delta v_\phi$ at each latitude and radius by assuming that
267: the temporal variation is periodic with a period of 11 years.
268: Since the variation is not strictly sinusoidal, we also include a few higher
269: harmonics and fit the following functional form to $\delta v_\phi$:
270: \be
271: \delta v_\phi(r,\theta,t)=\sum_{k=1}^{k_{\rm max}} a_k(r,\theta)\sin(k\omega_0 t)
272: +b_k(r,\theta)\cos(k\omega_0 t)=\sum_{k=1}^{k_{\rm max}} A_k\sin(k\omega_0 t+
273: \phi_k)\;.
274: \label{eq:harm}
275: \ee
276: In the above equation, $\omega_0$ is the basic frequency for the fundamental component,
277: which is assumed to have a period of 11 years.
278: The coefficients $a_k$ and $b_k$ are determined by a least squares
279: fit at fixed values of $(r,\theta)$. These coefficients determine the
280: amplitude, $A_k$ and phase, $\phi_k$ of the variation at any
281: point defined by the values of $(r,\theta)$ for each harmonic
282: of the 11 year period. We choose time $t=0$ to be the beginning
283: of GONG data sets and the phase is measured with respect to that epoch.
284: This zero-point is about a year before the minimum phase of solar activity.
285: Three terms are found to be sufficient to fit the
286: data. Similar fits were also used by Vorontsov et al.~(2002), Basu \&
287: Antia (2003, 2006) and Howe et al.~(2005, 2006a).
288: These fits are then used to calculate the time-derivative. These fits also enable us
289: to extend the pattern beyond the epochs for which observations exist.
290: These extended patterns are often useful in comparing the observed pattern with those obtained
291: on the basis of dynamo models since dynamo-model predictions are traditionally shown over a period of 22 years.
292:
293:
294: \section{RESULTS}
295: \label{sec:results}
296:
297: The rotation rate in the solar convection zone exhibits temporal variations.
298: Near the solar surface the
299: pattern is observed to be similar to the well known torsional oscillations
300: detected at the surface. In what follows
301: we describe our results in detail. We include results on the
302: temporal variation as well as results
303: on the radial, latitudinal and time derivatives of the rotation rate.
304:
305: Before discussing temporal variations we examine the time-averaged
306: quantities. Since the mean rotation rate inferred by the inversion of
307: seismic data is well-known (e.g., Thompson et al.~1996; Schou et al.~1998),
308: we do not show the results here, but instead we only show the results for
309: the gradients, $\Omega_{,r}$ and $\Omega_{,\theta}$.
310: The upper panels of Fig.~\ref{fig:shearav}
311: show the cuts at fixed latitude, while the lower panels show the
312: color-coded results as a function of both $r$ and $\theta$.
313:
314: The radial gradient of the solar rotation rate
315: is large in the outer shear layer as well as in the tachocline, and
316: is small in other regions. Below the outer shear layer, the radial
317: gradient is positive in the low latitude regions, and is negative in
318: all other regions wherever it can be determined reliably.
319: In the outer shear layer, we find an average negative gradient of
320: the order of a few 100 nHz $R_\odot^{-1}$ i.e, $10^{-15}$ m$^{-1}$ s$^{-1}$.
321: The value of $\Omega_{,r}$ is largely independent of latitude in the outer
322: shear layer, except perhaps at very high latitudes.
323: It should be noted that the regularization used in carrying out
324: the inversions smooths
325: out the radial variations near the tachocline region and hence the actual
326: radial gradient in the tachocline region is expected to be much larger.
327: If the tachocline model obtained by Antia et al.~(1998) is used to define
328: the rotation rate in the tachocline region, the gradient can be calculated
329: more precisely, and the results obtained using MDI data are shown in
330: Fig.~\ref{fig:sheartach}.
331: It is clear that the gradient is much larger and more localized in the
332: tachocline region than is seen in Fig.~\ref{fig:shearav}.
333: The exact shape of the curves, of course, depend on the
334: tachocline model used and may not have much significance.
335: The gradient in the latitudinal direction is significant only within the
336: convection zone, this is of course expected since this is where we
337: find differential rotation. The maximum value of the latitudinal
338: gradient is comparable to the radial gradient
339: in the outer shear layer. This gradient is negative almost everywhere.
340: When studying the temporal evolution of the two gradients, we compare
341: the magnitude of the time-variation with the corresponding time-averaged values.
342:
343: Although the pattern of variation of these gradients is similar for
344: results obtained with GONG and MDI data, the actual values show significant differences
345: in some regions. The results obtained from the two data sets agree reasonably well as
346: far as the
347: latitudinal gradient is concerned, though there are differences at high latitudes.
348: There are
349: noticeable differences in the radial gradients obtained from the two data sets, particularly in the
350: outer shear layer. Such differences between
351: results from GONG and MDI data have been
352: seen even for the rotation rate itself (e.g., Schou
353: et al.~2002), and the reasons for this disagreement are not fully
354: understood. This issue is discussed further in connection with the
355: temporal variations of the radial gradient.
356:
357:
358: \subsection{The time-variation of the rotational velocity: the zonal flows}
359: \label{subsec:zonal}
360:
361: Fig.~\ref{fig:cont8} shows the residual rotational velocity
362: $\delta v_\phi=\delta\Omega r\cos\theta$ obtained
363: using GONG data. Results are shown at a few representative radii in the convection zone.
364: The figure shows the expected, distinct,
365: bands of faster and slower-than-average rotation
366: velocity that, at low latitudes, move towards the equator with time.
367: At high latitudes, the bands seem to move towards the poles.
368: The slow band at high latitudes reaches its end near the poles around
369: 2001, the time when the polar field reversed, while the fast band at high latitude, on the other hand,
370: appears to terminate around 1996 (and will probably do so again in 2008) near the minimum of solar activity.
371: The transition between the equator-ward and pole-ward movements takes
372: place at a latitude of around $45^\circ$.
373: At high latitudes it appears that the
374: 12 years of GONG data have covered a little more than one period, but
375: at low latitudes the periodicity of the zonal-flow pattern is not obvious.
376: The zonal-flow pattern in the outer shear layer appears to be more complex at low
377: latitudes than at high latitudes.
378: In this layer, the low-latitude fast-moving bands from the two hemispheres
379: appear to have merged around 2000, but seem to
380: split around 2005, and then merge again in 2007.
381: A band of fast rotation appears near a latitude of $35^\circ$ in both
382: hemispheres around
383: 2005 and we believe that these bands are a precursor to the next solar cycle.
384: In fact, a closer examination of the figure suggests that these bands have
385: a weak connection to the bands that started at $50^\circ$ latitude around 1997
386: and then bifurcated around 2004 into one branch that moved polewards and another
387: that moved towards the equator.
388: These bands should reach the equator and merge during the next
389: maximum phase of solar activity. Thus the band which started
390: near the $50^\circ$ latitude zone in 1997, i.e., just after the minimum phase of the activity,
391: is expected to reach
392: the equator during the maximum of the following cycle, and
393: will thus span a period of 15--18 years. The band of slow rotation
394: that started a few years after the last solar-activity maximum around a latitude of
395: $50^\circ$ would reach the equator some time after the next minimum, again
396: spanning a period of about one and a half times the fiducial solar cycle.
397:
398: Figure~\ref{fig:cont8} also shows the zonal flow-pattern obtained using GONG data
399: as cuts at a few representative latitudes. These can be used
400: to determine how the zonal-flow pattern changes with radial distance.
401: At low latitudes there is a clear trend
402: of the pattern moving upwards with time and it is also clear that the
403: pattern penetrates through most of the convection zone, probably reaching
404: down to the base of the convection zone. At these low latitudes the average rate
405: of upward movement appears to be about $0.05R_\odot$ per year or about 1 m s$^{-1}$.
406: This follows from the fact that Fig.~\ref{fig:cont8} shows the band moving from
407: close to the base of the convection zone to surface in about 6 years.
408: At a latitude of $60^\circ$ too the band of faster rotation
409: appears to penetrate nearly to the base of the convection zone near
410: the maximum of solar activity.
411:
412: The changing pattern of the zonal flows implies that the
413: maximum and minimum velocities of the flow occur at
414: different times for different latitudes and depths.
415: This has also been found by Schou (1999) and Howe et al.~(2000a).
416: Fig.~\ref{fig:lat}
417: shows the zonal-flow velocity, $\delta v_\phi$, at different latitudes
418: as a function of time at $r=0.98R_\odot$. Both GONG and MDI results are
419: shown and as can be seen, there is good agreement between the results obtained using data
420: from the two projects.
421: It is clear that the time-dependence of $\delta v_\phi$ changes with
422: latitude. While the high latitudes show a nearly sinusoidal variation
423: with a possible period of close to 11 years, the variation at low latitudes
424: is more complicated.
425: The Fourier transform of the results at all latitudes show a peak at the lowest
426: frequency bin, which corresponds to a period of about 11 years.
427: The pattern at low latitudes, however, is clearly not sinusoidal. Thus if it is
428: periodic, higher harmonics must be present.
429: Vorontsov et al.~(2002) and Basu \& Antia (2003) had found the third harmonic
430: to be important. But those studies were based on data covering only about
431: half the solar cycle. With more data Howe et al.~(2005) and Basu \&
432: Antia (2006) found the second harmonic to be more significant than the third harmonic.
433: Clearly data from cycle 24 will be needed to establish the periodicity properly.
434: At a latitude of around $50^\circ$, the amplitude
435: of the flows is distinctly smaller than that at other latitudes.
436: This is close to the latitude where the switch from the equator-ward
437: movement to pole-ward movement of the flows occurs. Thus regions around this latitude
438: seem to separate the two branches of the zonal-flow pattern.
439:
440: While it is not possible to confirm the periodicity of the flows with current data sets,
441: it is tempting to assume a period
442: comparable to the average solar cycle. Thus we assume that the pattern is periodic with
443: a period of 11 years and fit it using Eq.~(\ref{eq:harm}) with
444: 3 terms. The amplitudes of the first two terms are shown in Fig.~\ref{fig:amp}
445: for results obtained from both GONG and MDI data. Both these data sets give similar results
446: and it is clear that although the amplitude of the dominant $k=1$ component
447: decreases with depth, the pattern appears to penetrate to the base
448: of the convection zone.
449: The highest amplitude, of up to 10 m s$^{-1}$, is seen in the outer $0.1R_\odot$ at high latitudes.
450: The amplitude is small in the mid-latitude region around $45^\circ$.
451: It is clear that the $k=2$ harmonic is significant, however, its
452: magnitude is much smaller than that for the first term. The amplitude of the third term is found to
453: be even smaller and it is difficult to estimate the amplitude reliably.
454: The relative magnitude of the various harmonics may depend on the origin
455: of the zonal flow pattern.
456: In the very deep layers close to the tachocline, the errors
457: in the zonal-flow velocity estimates are so large that it is difficult to
458: assign any significance to the fluctuating results obtained there.
459: Below the convection zone the amplitude
460: of both harmonics is very small, indicating that solar cycle-related
461: variations are not clearly seen in these layers.
462:
463:
464: Fig.~\ref{fig:amp} also shows the phase of the $k=1$ term
465: obtained from both GONG and MDI data.
466: It is clear that at low
467: latitudes the phase changes rather steeply around $r=0.9R_\odot$.
468: At high latitudes while MDI results show very little phase variation with depth,
469: GONG results do show some variations. It is not clear if these differences are
470: significant. In the outer half of the convection zone, both GONG and MDI results
471: appear to show a sharp transition around a latitude
472: of $40^\circ$.
473: This may mark the transition between the low and high latitude patterns.
474: There is good agreement between the GONG and MDI results.
475: %though there are
476: %some differences, e.g., at low latitudes the phase change around $0.9R_\odot$
477: %is more clear in MDI results.
478:
479: We can extend the zonal flow pattern over longer time
480: intervals if we assume that the fitted parameters do not change from cycle to cycle,
481: and the resulting pattern can be compared with predictions
482: from different dynamo models. We shall be happy to make such figures available for
483: investigators who wish to make use of them.
484:
485:
486: \subsection{Temporal variations of the rotation-rate gradients}
487: \label{subsec:shear}
488:
489:
490: We calculated the time variation of both radial and latitudinal
491: gradients by numerical differentiation
492: of the rotation-rate residuals that were obtained
493: using Eq.~(\ref{eq:zonal}).
494: Figures~\ref{fig:sheardr} and \ref{fig:sheardth} show the two components of the
495: gradient at different
496: latitudes as a function of time at $r=0.95R_\odot$.
497: There is a reasonable agreement between results obtained using GONG and MDI
498: data at this depth.
499: It is clear that there
500: is significant time-variation in these gradients and that the amplitude
501: of the variation is a sizable fraction of its average value.
502: At low latitudes, the temporal variation of the radial gradient turns
503: out to be about 20\% of its
504: average value.
505: The relative variation in the latitudinal gradient is similar.
506: These changes are much larger than the temporal variation in the rotation rate, which is
507: only about 0.1--1\% of its average value. As a result, we expect this
508: time-variation in the angular velocity gradients to
509: play an important role in workings of the solar dynamo.
510:
511: In order to establish a possible relation between the zonal-flow bands and
512: the magnetic cycle, we compare the pattern of temporal variations of
513: $\delta\Omega_{,r}$ and $\delta\Omega_{,\theta}$
514: at $r=0.95R_\odot$ with the butterfly diagram seen in the distribution
515: of sunspots. The results are shown in Fig.~\ref{fig:butter}.
516: We have chosen this depth because GONG and MDI results for the radial gradient do
517: not agree well in shallower layers.
518: We have multiplied the gradients by $\cos\theta$ in this and subsequent
519: figures in order to compensate for the increase in $\delta\Omega$
520: with latitude. This factor is already present in the residual rotation velocity $\delta v_\phi$
521: since it is part of the conversion from $\Omega$ to $v_\phi$.
522: As a result of this factor residuals in $v_\phi$ show little variation
523: in amplitude with latitude, and with
524: this factor, residuals in $\delta\Omega_{,r}$ or
525: $\delta\Omega_{,\theta}$ also show a similar behavior.
526: It can be seen from Fig.~\ref{fig:butter} that
527: the sunspots are mainly concentrated in regions where the radial gradient
528: is larger than its average value and where the latitudinal gradient
529: is smaller than the average gradient.
530: Note, the mean values of both these gradients are negative at this depth,
531: and hence their magnitude will be larger in regions where $\delta\Omega_{,r}$ or
532: $\delta\Omega_{,\theta}$ are negative. Thus the sunspots are found to
533: occur in regions where the magnitude of radial gradient is smaller than
534: average and that of the latitudinal gradient is larger than average.
535: Clearly, the butterfly pattern is related to
536: zonal flows, as was, indeed, noted earlier by Snodgrass (1987).
537: Around 2005, when the equatorial zonal flow band of faster rotation splits,
538: the latitudinal gradient of rotation rate flips sign in low latitude
539: region.
540:
541: The errors in our results on the gradients are naturally expected to be
542: larger than those in
543: $\delta\Omega$. These large errors give rise to some of the fluctuations seen in Figs~\ref{fig:sheardr}
544: and \ref{fig:sheardth}.
545: As in the case of $\delta v_\phi$, these fluctuations can be smoothed by fitting
546: a periodic signal with a period of 11 years to the results using Eq.~(\ref{eq:harm}).
547: The fitted results, extended to a period of 22 years, are shown in
548: Figs.~\ref{fig:domdrfit},\ref{fig:domdthfit}.
549: The pattern of bands of higher and lower than average gradients can be seen
550: in these figures too, with low latitude bands moving towards
551: the equator and the pattern once again changing at high latitudes.
552: In general there is reasonable agreement between GONG and MDI results, except
553: in the case of the radial gradient at $r=0.98R_\odot$ where there are significant differences
554: between the two results.
555: %The errors in this component are however, large and it is difficult to
556: %identify the actual variation in the radial gradient in the outer shear layer.
557: The agreement
558: between GONG and MDI results is better in deeper layers, despite the fact that
559: the gradient itself is much smaller. As far as the latitudinal
560: gradient is concerned, the agreement between GONG and MDI results is good at all depths.
561: It has been known for some time that there are differences between GONG and MDI rotational-splitting data (Schou et al. 2002).
562: These differences result in the differences between results obtained with GONG and MDI data
563: about the outermost layers at all latitudes.
564: The most prominent difference is however, in the radial-gradient of the rotation rate at
565: high latitudes.
566: MDI data shows the presence of a jet like feature at high latitudes while GONG data do not
567: (e.g., Howe et al.~1998, Schou et al.~2002).
568: Interestingly, the differences between the
569: GONG and MDI data do not affect the time variations of the
570: zonal-flow velocities and the results obtained from the two different
571: data sets agree reasonably well. This may indicate that the systematic
572: differences between the two data sets are largely independent of time.
573: Since we calculate the residual $\delta\Omega$ independently for GONG
574: and MDI the systematic errors cancel when the mean value is subtracted
575: in Eq.~(\ref{eq:zonal}).
576:
577: It is tempting to speculate that the difference in radial gradient in outer
578: layers between the MDI and GONG results is because of the presence of higher
579: degree modes, in particular the f-modes, in MDI data sets. To test this, we
580: repeated the calculations by removing these modes and find that the results remain
581: largely unaffected. Thus the difference between GONG and MDI results is unlikely
582: to be due to difference in resolution due to presence of extra modes and
583: possibly represents the difference in splitting coefficients themselves as
584: found by Schou et al.~(2002). While the zonal velocity, $\delta v_\phi$,
585: computed using the two data sets agree reasonably well,
586: the time variation of the
587: radial gradient of the inversions show significant differences. Thus the
588: process of determining the radial gradient and its time variation appears
589: to amplify the differences. In order to identify the modes that lead to this
590: difference, we did one more set of inversions for MDI data
591: using only modes with $\ell<120$; the resulting radial gradient is closer
592: to the result obtained using GONG data as can be seen from Fig.~\ref{fig:mdil120}.
593: Thus it is clear that the differences are caused by the high-degree modes in the MDI sets.
594: Similar problems have been noticed in even-order splitting coefficients
595: from MDI (Antia et al.~2003). A closer look at Fig.~\ref{fig:domdrfit}
596: suggests that at $r=0.98R_\odot$ there was a rather sharp change around
597: 1999 which is around the period when contact with SOHO was re-established.
598: This is similar to the results on asphericity found by Antia et al.~(2003) and
599: results on solar radius variations found by Antia (2003). All these
600: could be due to some systematic error introduced during recovery
601: of SOHO. High-degree modes which are trapped in the outer layers
602: do not seriously affect the inversion
603: results in the deeper layers and hence, these layers are not affected by the possible
604: problems with the MDI high-degree modes. The re-analysis of MDI data has shown that
605: there are some removable problems with the currently available data sets (Larson \& Schou 2007).
606:
607: The pattern at high latitudes is complicated and it is
608: not possible to say with any certainty whether the pattern is moving towards the poles
609: or towards the equator. From Fig.~\ref{fig:domdrfit} it can be seen that
610: at $r=0.95R_\odot$, the band with lower than average radial gradient that started around a
611: latitude of $50^\circ$ in about 1995 ends at the equator around 2012, thus
612: spanning a period of about one and a half fiducial solar cycles. This is similar to the corresponding band in the
613: zonal-flow pattern. However, unlike the bands in the zonal-flow pattern, these bands appear to
614: last for only about a solar cycle in the latitudinal-gradient pattern.
615: From the panels in Fig.~\ref{fig:domdrfit} that show the radial dependence of
616: $\Omega_{,r}$ it also appears that
617: the pattern rises upwards with time at low latitudes.
618: At the latitude of $30^\circ$
619: the amplitude of the variation in the radial gradient is rather small.
620: This is the latitude where the radial gradient in the lower convection
621: zone changes sign (cf., Fig.~\ref{fig:shearav}) and the magnitude of
622: temporally averaged $\Omega_{,r}$ is small too.
623: At high latitudes ($>45^\circ$) the bands showing the changes in the
624: radial gradient penetrate to the base of the convection zone, while at
625: low latitudes the amplitude decreases below the outer
626: shear layer and there is a distinct phase shift around $r=0.9R_\odot$.
627: For the latitudinal gradient, the temporal variations
628: are generally significant only in the outer shear layer and below a depth
629: of $0.1R_\odot$ this component is generally small (Fig.~\ref{fig:domdthfit}).
630:
631: \subsection{The time derivative of $v_\phi$}
632: \label{subsec:acc}
633:
634: We evaluate the time-derivative of $\delta v_\phi$, the zonal-flow velocity,
635: by differentiating the fits in Eq.~(\ref{eq:harm}). The results are
636: shown in Fig.~\ref{fig:dvrot}. As expected, the magnitude of the derivative
637: is of the order of $\omega_0|\delta v_\phi|$, i.e., about $10^{-7}$ m s$^{-2}$.
638: Since this derivative is calculated by differentiating Eq.~(\ref{eq:harm}),
639: the pattern of the derivative is essentially similar to that of $\delta v_\phi$, except for a phase shift
640: of $\pi/2$. However,
641: the higher harmonics make a larger contribution
642: to the zonal-flow derivative than to the flow itself and as a result,
643: there are additional differences between
644: the patterns that can be seen when the pattern of the zonal-flow
645: derivative is compared to the pattern of the zonal flow.
646: The second and third harmonics appear to dominate
647: in the deeper parts of the convection zone near the tachocline, but
648: since the signal
649: in this region is fairly weak even in $\delta v_\phi$, it is not clear if this
650: is significant.
651:
652:
653:
654: \section{DISCUSSION AND CONCLUSIONS}
655: \label{sec:disc}
656:
657:
658: We have studied the time-variation of the solar-internal rotation rate and its gradients
659: over almost the entire period covering solar cycle 23 using
660: data from the GONG and MDI projects. We used the
661: rotation-rate residuals obtained by subtracting the time-averaged
662: rotation rate from that at each epoch at a given latitude and depth
663: to study the zonal-flow pattern and to determine
664: how it changes with time. These residuals have also been used to study the
665: radial and latitudinal gradients of the
666: zonal flows as well as the time-derivative of solar rotation velocity. Our main results are as follows:
667:
668: \begin{enumerate}
669:
670: \item The rotation-velocity residuals show the well known pattern similar to the observed pattern
671: of
672: torsional oscillations at the solar surface. At low latitudes the results show bands of faster and
673: slower-than-average rotation speed moving towards the equator as the solar
674: cycle progresses. At high latitudes on the other hand, these bands move towards the
675: poles. The transition between the equator-ward and pole-ward movement
676: occurs around a latitude of $45^\circ$.
677:
678: \item We find a precursor of the next solar cycle in the form of a band of fast
679: rotation that appeared at about the $35^\circ$ latitude near the surface around 2005 when
680: the Sun was approaching the activity minimum.
681: We also find that the band
682: of faster rotation found in the region near the equator appears
683: to have bifurcated around the same time.
684:
685: \item The zonal flow pattern at low latitudes moves upwards, towards shallower depths, with an average speed of about 1 m/s.
686:
687: \item The relative time variation of the radial gradient of the
688: rotation-rate is about 20\%
689: of its average value, which is much larger than the relative variation in
690: the rotation rate itself. The time variation of the latitudinal gradient is large
691: only in the outer $0.1R_\odot$ of the Sun, the maximum change being about
692: 20\% of the mean value.
693:
694: \item The magnetic butterfly diagram coincides with band of larger than
695: average radial gradient and with smaller than average
696: latitudinal gradient at $r=0.95R_\odot$. Noting that both gradients are
697: negative, the sunspot activity bands occurs in regions of enhanced
698: latitudinal shear and diminished radial shear, as compared to the
699: magnitude of the average shear.
700:
701: \item The time derivative of the rotation velocity is about $10^{-7}$ m s$^{-2}$
702:
703: \end{enumerate}
704:
705:
706: The equator-ward movement of the zonal-flow pattern at low latitudes is well known
707: (Schou 1999; Howe et al.~2000a; Antia \& Basu 2000; Vorontsov et al.~2002).
708: The pole-ward migrating high latitude branch has also been noted earlier
709: by Antia \& Basu (2001) and Ulrich (2001).
710: The transition between the equator-ward and pole-ward movements
711: occurs around a latitude of $45^\circ$. The differences between the low-latitude
712: and the high-latitude pattern also manifests in integrated quantities
713: like the rotational kinetic energy and angular momentum (Antia et al.~2008).
714: The time-variations of the rotational
715: kinetic energy integrated separately over low and high latitude regions also
716: differ.
717: The high-latitude regions show a kinetic-energy variation that is correlated with
718: the solar activity throughout the convection zone. At the low latitudes,
719: on the other hand, the positive correlation between kinetic energy and solar activity
720: exists only in the outer $0.1R_\odot$, while
721: the variations are anticorrelated with solar activity in the rest of the
722: convection zone. Observations of magnetic
723: features at the solar surface also show
724: an equator-ward movement at low latitudes (manifesting in the well-known
725: butterfly diagram) and a pole-ward movement
726: at high latitudes (e.g., Leroy \& Noens 1983; Makarov \& Sivaraman 1989;
727: Erofeev \& Erofeeva 2000; Benevolenskaya et al.~2001).
728: Theoretical models of the solar dynamo, such as those of
729: Covas et al.~(2000, 2001), Bushby (2005) show these features, though they do differ in
730: detail, such as in the phase. The
731: predicted butterfly diagrams of many models (Jiang et al.~2007; Rempel 2006, etc.)
732: also show a low-latitude equator-ward moving branch and a high-latitude
733: pole-ward drifting branch.
734:
735: The splitting of the band
736: marking fast rotating region near the equator appears to be a new feature
737: and we are not aware of
738: any other investigations that have noted this.
739: Although we have shown only GONG results, MDI data also show the splitting
740: of the bands.
741: The observations during the last solar minimum, however, does not show
742: this feature clearly. Although, a closer look at the Fig.~\ref{fig:cont8}
743: shows a faint signature of two bands merging at the equator around 1996,
744: albeit with a magnitude of $\delta v_\phi$ around 0.5 m s$^{1}$, which is
745: comparable to the error estimates and
746: about a factor of 5 lower than the corresponding value around 2006.
747: This difference could indeed be owing to cycle to cycle variation.
748: We clearly need more data to see the development of this pattern.
749: The near-surface behavior of zonal flows can be seen in the surface
750: rotation rate data from Mt.~Wilson (Ulrich 2001; Howe et al.~2006b)
751: that also covers the previous solar cycle. Since the errors in these
752: observations are larger, it is difficult to say whether the splitting of
753: the near equator band was seen in these results during the last cycle.
754:
755: The rise of the zonal flow pattern with time has been seen in sub-surface layers from
756: other measurements too.
757: Conclusions similar to ours were drawn
758: by Komm, Howard \& Harvey~(1993) who compared the zonal flow pattern
759: obtained using the Doppler
760: measurements at the surface with that from magnetic features that
761: are believed to be anchored underneath the surface.
762: Our result that the zonal flows penetrate
763: through a good fraction of the convection zone, and probably reach
764: the base of the convection zone appears to
765: contradict early inferences (Howe et al.~2000a; Antia \& Basu 2000)
766: that these flows penetrate only to a radius of $0.9R_\odot$.
767: The early results could have been due to the fact that the phase of the
768: zonal flow pattern at low latitudes changes around this depth.
769: With limited data this could have lead to the conclusion that the pattern does
770: not penetrate below this depth.
771: This phase change is seen in
772: variation of rotational kinetic energy in low latitudes too (Antia et al.~2008).
773: Our results are, however, similar to those of Vorontsov et al.~(2002),
774: Basu \& Antia (2003) and Howe et al.~(2005).
775: Vorontsov et al.~(2002) also found that
776: a band of faster rotating elements appears to penetrate almost up to the
777: base of the convection zone at latitude of around $60^\circ$.
778: Dynamo models that assume that the tachocline is the seat of the
779: dynamo (e.g., Covas et al.~2000, 2001, 2004; Bushby 2005, etc.) also predict that the zonal
780: flow pattern should persist through the convection zone.
781:
782:
783: The possible link between time-dependent shear oscillation pattern and the
784: solar activity cycle was first emphasized by Snodgrass (1987). This becomes
785: evident from a striking resemblance of shear zones with the magnetic activity
786: bands revealed through the butterfly diagram (Fig.~\ref{fig:butter}).
787: Making the
788: assumption that seat of the solar dynamo is likely to be located
789: within the shear layers, the observational
790: information that we have gained from the time-variation of radial
791: and latitudinal gradients of rotation rate should provide a valuable input
792: for understanding the dynamo mechanism.
793:
794: We have hitherto been concerned largely with the time-variations of the global
795: rotation rate which is about 0.1--1\% of its mean value. The relative
796: temporal variations in the radial and latitudinal gradients of the rotation rate, however,
797: turn out to be larger, 20\% or more of their mean values. Clearly, the time-varying
798: shear pattern in the activity bands must play an important role in driving
799: the magnetic cycle. In fact, the time-derivatives of angular velocity
800: may be used either as valuable input for dynamo models or at any rate as
801: a constraint on them. Indeed, the time derivative of the differential rotation
802: has been effectively used in dynamo models such as those of Covas et al.~(2000)
803: and Rempel (2006).
804:
805: It is illustrative to consider the azimuthal components of the induction and momentum
806: equations and neglect the presence of other velocity field such as
807: meridional circulation and dissipation, and write the equations
808: \begin{eqnarray}
809: {\partial B_\phi\over\partial t}&=& r\cos\theta{\bf B_p}.\nabla \Omega,
810: \\
811: \rho{\partial v_\phi\over\partial t}&=&{1\over 4\pi r\cos\theta}\nabla\cdot
812: \left({\bf B_p}(r\cos\theta B_\phi)\right),
813: \end{eqnarray}
814: where $\bf B_p$ and $B_\phi$ are respectively the poloidal and toroidal
815: components of the magnetic field, and $\rho$ is the density which is
816: assumed to be independent of time (e.g., Basu \& Antia 2000; Basu 2002). With the
817: available knowledge of the time-dependent radial and
818: latitudinal gradients of the angular velocity, it should be feasible to
819: deduce the temporal variation of the toroidal magnetic field for an
820: assumed configuration of the poloidal field. Equally, the force term
821: can also be estimated using the measured acceleration
822: $\partial v_\phi/\partial t$ provided we assume that the solar torsional
823: oscillations are driven mainly by the Lorentz force, although there will
824: be additional contributions from the temporally varying meridional
825: flow (e.g., Haber et al.~2002; Basu \& Antia 2003) which we neglect.
826: To an order of magnitude we can write
827: \be
828: {\partial v_\phi\over\partial t}\approx {|{\bf B_p}|B_\phi\over 4\pi r\rho}\;,
829: \label{eq:mag}
830: \ee
831: If the magnitude of the poloidal field is estimated through independent means,
832: it should be possible to infer the strength of the azimuthal field using
833: our knowledge of $\partial v_\phi/\partial t$. Thus, if the magnitude
834: of both poloidal and toroidal components are comparable, then the
835: magnetic field would be of the order of 1 G near the surface
836: increasing to 1000 G near the base of the convection zone.
837: The surface value is consistent with observations of the average
838: magnetic field at the surface (e.g., Ulrich
839: \& Boyden 2005).
840: Since the
841: magnetic field is generally concentrated in flux tubes with low filling factors, the
842: magnetic field in these flux tubes is expected to be much higher.
843: On the other hand, if the poloidal field is 100 times weaker than the toroidal field (e.g., Rempel
844: 2007) then the toroidal field will be about 10 times the above estimate,
845: i.e., of the order of 10 kG near the base of the convection zone.
846:
847:
848: In conclusion, we emphasize that observations of the zonal flows of the
849: Sun now cover almost an entire solar cycle, and
850: there are some features
851: of the flows seen at this solar minimum phase that had not been noticed in the earlier solar
852: minimum for which the helioseismic data were sparse. The further evolution of these features
853: should shed light on the differences between different solar cycles.
854: The current results are also precise enough to provide
855: important inputs to or constraints for solar dynamo models.
856:
857: \acknowledgements
858:
859: This work utilizes data obtained by the Global Oscillation
860: Network Group (GONG) project, managed by the National Solar Observatory,
861: which is
862: operated by AURA, Inc. under a cooperative agreement with the
863: National Science Foundation. The data were acquired by instruments
864: operated by the Big Bear Solar Observatory, High Altitude Observatory,
865: Learmonth Solar Observatory, Udaipur Solar Observatory, Instituto de
866: Astrofisico de Canarias, and Cerro Tololo Inter-American Observatory.
867: This work also utilizes data from the Solar Oscillations
868: Investigation/ Michelson Doppler Imager (SOI/MDI) on the Solar
869: and Heliospheric Observatory (SOHO). SOHO is a project of
870: international cooperation between ESA and NASA.
871: MDI is supported by NASA grants NAG5-8878 and NAG5-10483
872: to Stanford University. SB acknowledges partial support from
873: NSF grant ATM 0348837.
874: SMC thanks the Indian National Science Academy for support under the
875: INSA Honorary Scientist programme.
876:
877:
878:
879:
880:
881: \begin{thebibliography}{}
882:
883: \bibitem[Antia(2003)]{ant03}
884: Antia, H. M. 2003, ApJ, 590, 567
885:
886: \bibitem[Antia \& Basu~(2000)]{ant00}
887: Antia, H. M., \& Basu, S. 2000, ApJ, 541, 442
888:
889: \bibitem[Antia \& Basu~(2001)]{an00}%
890: Antia, H. M., \& Basu, S. 2001, ApJ, 559, L67
891:
892: \bibitem[Antia, Basu, \& Chitre~(1998)]{ant98}
893: Antia, H. M., Basu, S., \& Chitre, S. M. 1998, MNRAS, 298, 543
894:
895: \bibitem[Antia et al.~(2008)]{ant08}
896: Antia, H. M., Chitre, S. M., Gough, D. O. 2008, A\&A, 477, 657
897:
898: \bibitem[Antia et al.~(2003)]{antetal03}
899: Antia, H. M., Chitre, S. M., Thompson, M. J. 2003, A\&A, 399, 329
900:
901: \bibitem[Basu(2002)]{bas02}
902: Basu, S. 2000, in proc.\ SOHO 11, From Solar Min to Max: Half a Solar
903: Cycle with SOHO, ed., A. Wilson, ESA SP-508 p.~7
904:
905: \bibitem[Basu \& Antia~(2000)]{ba00a}
906: Basu, S., \& Antia, H. M. 2000, Sol.\ Phys., 192, 449
907:
908: %\bibitem[Basu \& Antia~(2000b)]{ba00b}
909: %Basu, S., \& Antia, H. M. 2000b, Sol.\ Phys., 192, 469
910:
911: \bibitem[Basu \& Antia~(2001)]{ba01} Basu, S., \& Antia, H. M.
912: 2001, in Proc. SOHO 10/GONG 2000 Workshop: Helio- and asteroseismology at
913: the dawn of the millennium, Eds., A. Wilson, P. L. Pall\'e, ESA SP-464,
914: p.~179
915:
916: \bibitem[Basu \& Antia~(2003)]{ba03} Basu, S., \& Antia, H. M.
917: 2003, ApJ, 585, 553
918:
919: \bibitem[Basu \& Antia~(2006)]{ba06} Basu, S., \& Antia, H. M.
920: 2006, in proc.\ SOHO 18/GONG 2006/HELAS I, Beyond the spherical Sun
921: (ESA SP-624) eds., K. Fletcher, M. J. Thompson, p.~128
922:
923: \bibitem[Benevolenskaya, Kosovichev \& Scherrer~(2001)]{ben01}
924: Benevolenskaya, E. E., Kosovichev, A. G., \& Scherrer, P. H.
925: 2001, ApJ, 554, L107
926:
927:
928: %\bibitem[Brun \& Toomre~(2001)]{bru01}
929: %Brun, A. S., \& Toomre, J. 2001, in Helio- and Astero-seismology at the Dawn of the Millennium,
930: %ed. A. Wilson, (ESA SP-464; Noordwijk: ESA), 619
931:
932: \bibitem[Bushby(2005)]{bush05}
933: Bushby, P. J. 2005, AN, 326, 218
934:
935: %\bibitem[Charbonneau et al~(1999)]{cha99}
936: %Charbonneau, P., \jcd, J., Henning, R., Larsen, R. M., Schou, J., Thompson, M. J.,
937: %\& Tomczyk, S. 1999, ApJ, 527, 445
938:
939: \bibitem[Corbard \& Thompson(2002)]{cor02}
940: Corbard, T., \& Thompson, M. J. 2002, Solar Phys., 205, 211
941:
942: \bibitem[Covas et al.~(2000)]{cov00}
943: Covas, E., Tavakol, R., Moss, D., \& Tworkowski, A. 2000, A\&A, 360, L21
944:
945: \bibitem[Covas et al.~(2001)]{cov01}
946: Covas, E., Tavakol, R., \& Moss, D. 2001, A\&A, 371, 718
947:
948: \bibitem[Covas et al.~(2004)]{cov04}
949: Covas, E., Moss, D., \& Tavakol, R. 2004, A\&A, 416, 775
950:
951: %\bibitem[Dikpati \& Charbonneau~(1999)]{dik99}
952: %Dikpati, M., \& Charbonneau, P. 1999, ApJ, 518, 508
953:
954: %\bibitem[Dikpati \& Gilman~(2001)]{dik01}
955: %Dikpati, M., \& Gilman, P. 2001, ApJ, 559, 428
956:
957: \bibitem[Erofeev \& Erofeeva~(2000)]{ero00}
958: Erofeev, D. V., \& Erofeeva, A. V. 2000, Sol.\ Phys., 191, 281
959:
960: \bibitem[Haber et al.(2002)]{hab04}
961: Haber, D. A., Hindman, B. W., Toomre, J., Bogart, R. S., Larsen, R. M.,
962: Hill, F. 2002, ApJ, 570, 855
963:
964: \bibitem[Hill et al.~(1996)]{hi96}
965: %Hill, F., Stark, P. B., Stebbins, R., et al.~1996, Sci., 272, 1292
966: Hill, F., et al.~1996, Sci., 272, 1292
967:
968: \bibitem[Howard \& LaBonte~(1980)]{how80}
969: Howard, R., \& LaBonte, B. J. 1980, ApJ, 239, L33
970:
971: \bibitem[Howe et al.~(1998)]{how98}
972: Howe, R., Antia, H., Basu, S., Christensen-Dalsgaard, J., Korzennik, S. G., Schou, J. \&
973: Thompson, M. J. 1998, in Proc. SOHO6/GONG98, ``Structure and Dynamics of the Interior
974: of the Sun and Sun-like Stars,'' eds. S. Korzennik, A. Wilson, ESA SP 418, 803
975:
976: \bibitem[Howe et al.~(2000a)]{how00}
977: Howe, R., Christensen-Dalsgaard, J., Hill, F., Komm, R. W.,
978: Larsen, R. M., Schou, J., Thompson, M. J., \& Toomre, J. 2000a, ApJ, 533, L163
979:
980: \bibitem[Howe et al.~(2000b)]{ho00}%
981: Howe, R., \jcd, J., Hill, F., Komm, R. W.,
982: Larsen, R. M., Schou, J., Thompson, M. J., \& Toomre, J. 2000b, Sci, 287, 2456
983:
984: %\bibitem[Howe et al.(2001)]{how01}
985: %Howe, R., Hill, F., Basu, S. et al. 2001, in Proc. SOHO 10/GONG2000 workshop
986: %``Helio- and asteroseismology at the dawn of the millennium,''
987: %ed., A Wilson, ESA SP 464, p.~137
988:
989:
990: %\bibitem[2001a]{howe0} Howe, R., Christensen-Dalsgaard, J., Hill, F.,
991: %Komm, R. W., Larsen, R. M., Schou, J., Thompson, M. J., \& Toomre, J.
992: %2001a, in Helio- and Asteroseismology at the Dawn of the Millennium,
993: %ed. A. Wilson (ESA SP-464; Noordwijk: ESA), 19
994:
995: %\bibitem[Howe et al.~(2001b)]{how01}
996: %Howe, R., Hill, F., Basu, S., Christensen-Dalsgaard, J., Komm, R. W.,
997: %Larsen, R. M., Roth, M., Schou, J., Thompson, M. J., \& Toomre, J.
998: %2001b, in Helio- and Astero-seismology at the Dawn of the Millennium,
999: %ed. A. Wilson (ESA SP-464; Noordwijk: ESA), 137
1000:
1001: \bibitem[Howe et al.~(2005)]{how05}
1002: Howe, R., Christensen-Dalsgaard, J., Hill, F., Komm, R. W.,
1003: Schou, J., Thompson, M. J. 2005, ApJ, 634, 1405
1004:
1005: \bibitem[Howe et al.~(2006a)]{how06a}
1006: Howe, R., Rempel, M., Christensen-Dalsgaard, J., Hill, F., Komm, R. W.,
1007: Larsen, R. M., Schou, J., Thompson, M. J. 2006a, ApJ, 649, 1155
1008:
1009: \bibitem[Howe et al.~(2006b)]{how06b}
1010: Howe, R., Komm, R. W., Hill, F., Ulrich, R., Haber, D. A., Hindman, B. W.,
1011: Schou, J., Thompson, M. J. 2006b, Solar Phys., 235, 1
1012:
1013: \bibitem[Howe et al.~(2007)]{how07}
1014: Howe, R., Christensen-Dalsgaard, J., Hill, F., Komm, R. W.,
1015: Schou, J., Thompson, M. J., Toomre, J. 2007, Adv.\ Space Res., 40, 915
1016:
1017: \bibitem[Jiang et al.~(2007)]{jian07}
1018: Jiang, J., Chatterjee, P., \& Choudhury, A. R. 2007, MNRAS, 381, 1527
1019:
1020: \bibitem[Komm, Howard \& Harvey~(1993)]{kom93}
1021: Komm, R. W., Howard, R. F., \& Harvey, J. W. 1993, Sol.\ Phys., 143, 19
1022:
1023: %\bibitem[Komm et al.(2004)]{kom04}
1024: %Komm, R. W., Corbard, T., Durney, B. R., Conzalez Hernandez, I., Hill, F.,
1025: %Howe, R., Toner, C. 2004, ApJ 605, 554
1026:
1027: %\bibitem[Kosovichev et al.~(1997)]{kos97}
1028: %Kosovichev, A. G., Schou, J., Scherrer, P. H., et al.~1997, Sol.\ Phys., 170, 43
1029: %Kosovichev, A. G., et al.~1997, Sol.\ Phys., 170, 43
1030: \bibitem[Kosovichev \& Schou(1997)]{kos97}
1031: Kosovichev, A. G., \& Schou, J. 1997, ApJ, 482, L207
1032:
1033: %\bibitem[LaBonte \& Howard~(1982)]{lab82}
1034: %LaBonte, B. J., \& Howard, R. 1982, Sol.\ Phys., 75, 161
1035:
1036: \bibitem[Lanza(2007)]{lan07}
1037: Lanza, A. F. 2007, A\&A, 471, 1011
1038:
1039: \bibitem[Larson(2007)]{lar07}
1040: Larson. T., Schou, J. 2007, in Proc. Helioseismology, Asteroseismology and MHD Connections,
1041: J. Phys. Conf. Ser., in press
1042:
1043: \bibitem[LeRoy \& Noens~(1983)]{le83}
1044: Leroy, J. -L., \& Noens, J. -C. 1983, A\&A, 120, L1
1045:
1046: \bibitem[Makarov \& Sivaraman~(1989)]{mak89}
1047: Makarov, V. I., \& Sivaraman, K. R. 1989, Sol.\ Phys., 123, 367
1048:
1049: %\bibitem[Miesch~(2000)]{mie00}
1050: %Miesch, M. 2000, Sol.\ Phys., 192, 59
1051:
1052: \bibitem[Rempel(2006)]{rem06}
1053: Rempel, M. 2006, ApJ, 647, 662
1054:
1055: \bibitem[Rempel(2007)]{rem07}
1056: Rempel, M. 2007, ApJ, 655, 651
1057:
1058: \bibitem[Ritzwoller \& Lavely~(1991)]{rit91}
1059: Ritzwoller, M. H., \& Lavely, E. M. 1991, ApJ, 369, 557
1060:
1061: %\bibitem[Schou, Christensen-Dalsgaard \& Thompson~(1994)]{sch94}
1062: %Schou, J., Christensen-Dalsgaard, J., \& Thompson, M. J. 1994, ApJ, 433, 389
1063:
1064: \bibitem[Schou~(1999)]{sch99}
1065: Schou, J. 1999, ApJ, 523, L181
1066:
1067: \bibitem[Schou et al.~(1998)]{sch98}
1068: %Schou, J., Antia, H. M., Basu, S., et al.~1998, ApJ, 505, 390
1069: Schou, J., et al.~1998, ApJ, 505, 390
1070:
1071: \bibitem[Schou et al.~(2002)]{sch02}
1072: %Schou, J., Howe, R., Basu, S., et al.~2002, ApJ, 567, 1234
1073: Schou, J., et al.~2002, ApJ, 567, 1234
1074:
1075: %\bibitem[Snodgrass~(1984)]{sno84} Snodgrass, H. B. 1984, Solar Phys., 94, 13
1076:
1077: \bibitem[Snodgrass~(1987)]{sno87} Snodgrass, H. B. 1987, Solar Phys., 110, 35
1078:
1079: \bibitem[Snodgrass~(1992)]{sno92} Snodgrass, H. B. 1992, in The Solar Cycle,
1080: Proc. NSO 12th Summer Workshop, ed., K. L. Harvey, ASPS 27, 205
1081:
1082: \bibitem[Spiegel \& Zahn~(1992)]{spi92}%
1083: Spiegel, E. A., \& Zahn, J.-P. 1992, A\&A, 265, 106
1084:
1085: \bibitem[Spruit(2003)]{spr03}
1086: Spruit, H. C. 2003, Solar Phys., 213, 1
1087:
1088: \bibitem[Thompson et al.~(1996)]{tho96}
1089: %Thompson, M. J., Toomre, J., Anderson, E., et al.~1996, Sci., 272, 1300
1090: Thompson, M. J., et al.~1996, Sci., 272, 1300
1091:
1092:
1093: %\bibitem[Toomreet al.(2000)]{too00}
1094: %Toomre, J., \jcd, J., Howe, R., Larsen, R. M., Schou, J., \& Thompson, M. J.
1095: %2000, Sol.\ Phys., 192, 437
1096:
1097: \bibitem[Ulrich~(2001)]{ulr01}
1098: Ulrich, R. K. 2001, ApJ, 560, 466
1099:
1100: \bibitem[Ulrich \& Boyden~(2005)]{ulr05}
1101: Ulrich, R. K. \& Boyden, J. E. 2005, ApJ, 620, L123
1102:
1103: \bibitem[Ulrich et al.~(1988)]{ulr88}
1104: Ulrich, R. K., Boyden, J. E., Webster, L., Snodgrass, H. B., Padilla, S. P.,
1105: Gilman, P., \& Shieber, T. 1988, Sol.\ Phys., 117, 291
1106:
1107:
1108: \bibitem[Vorontsov et al.~(2002)]{vetal02}
1109: Vorontsov, S. V., \jcd, J., Schou, J., Strakhov, V. N.,
1110: \& Thompson, M. J. 2002, Sci., 296, 101
1111:
1112:
1113: \end{thebibliography}
1114: \clearpage
1115:
1116:
1117: \begin{figure}
1118: \plotone{f1.eps}
1119: \caption{The mean radial gradient, $\Omega_{,r}$, and latitudinal gradient,
1120: $\Omega_{,\theta}$, of the solar rotation rate obtained
1121: using GONG and MDI data. Panels in the upper row show the cuts at fixed latitude
1122: as a function of radius, while those in the lower row shows the same results as
1123: color-coded diagrams. All values are in units of nHz $R_\odot^{-1}$.}
1124: \label{fig:shearav}
1125:
1126: \end{figure}
1127:
1128:
1129: \clearpage
1130:
1131: \begin{figure}
1132: \plotone{f2.eps}
1133: \caption{The mean radial gradient, $\Omega_{,r}$, obtained from
1134: MDI data using the tachocline model of Antia et al.~(1998) shown at a
1135: few selected latitudes. The color scheme is same as that in the upper
1136: row of Fig.~\ref{fig:shearav}. Errorbars are shown by dotted lines.}
1137: \label{fig:sheartach}
1138: \end{figure}
1139:
1140:
1141: \clearpage
1142:
1143: \begin{figure}
1144: %\epsscale{1.0}
1145: \plotone{f3.eps}
1146: \caption{Diagrams showing $\delta v_\phi$,
1147: the residuals of the rotation velocity obtained using
1148: GONG data. We show the results
1149: at a few representative depths and latitudes as marked above the respective panels.
1150: The scale is in m s$^{-1}$.}
1151: \label{fig:cont8}
1152: \end{figure}
1153:
1154: \clearpage
1155:
1156:
1157: \begin{figure}
1158: %\epsscale{.60}
1159: \plotone{f4.eps}
1160: \caption{The zonal-flow velocity as a function of time at different
1161: latitudes at $r=0.98R_\odot$. The latitudes are marked in each panel.
1162: The squares show results obtained with GONG data, while triangles
1163: show results with MDI data. The dash-dotted, solid and dashed lines show the fits
1164: using Eq.~(\ref{eq:harm})
1165: to the GONG results with 1,2 and 3 terms respectively.
1166: For clarity, errorbars are not shown on the points, however, a typical errorbar
1167: is shown by a point in the upper right corner of each panel.
1168: }
1169: \label{fig:lat}
1170: \end{figure}
1171:
1172: \clearpage
1173:
1174: \begin{figure}
1175: \plotone{f5.eps}
1176: \caption{Diagram showing amplitudes of the $k=1$ and $k=2$
1177: (Eq.~\ref{eq:harm})
1178: components and the phase of $k=1$ component of zonal-flow expansion as obtained using
1179: GONG and MDI data.
1180: For panels that show the amplitudes, the scale is in units of m s$^{-1}$.}
1181: \label{fig:amp}
1182: \end{figure}
1183:
1184: \clearpage
1185:
1186:
1187:
1188: \begin{figure}
1189: \plotone{f6.eps}
1190: \caption{The radial gradient, $\Omega_{,r}$
1191: at a few selected latitudes is shown for $r=0.95R_\odot$.
1192: The squares and triangles show results obtained using GONG and MDI data
1193: respectively. The solid and dashed lines show fits obtained using equations
1194: similar to Eq.~(\ref{eq:harm}) to these points.
1195: For clarity, errorbars are not shown on the points, but a typical errorbar
1196: is shown by a point in a corner of each panel.}
1197: \label{fig:sheardr}
1198: \end{figure}
1199:
1200: \clearpage
1201:
1202: \begin{figure}
1203: \plotone{f7.eps}
1204: \caption{The latitudinal gradient $\Omega_{,\theta}$
1205: at a few selected latitudes is shown for $r=0.95R_\odot$.
1206: The squares and triangles show the results obtained using GONG and MDI data
1207: respectively. The solid and dashed lines show fits obtained using equations
1208: similar to Eq.~(\ref{eq:harm}) to these points.
1209: For clarity, errorbars are not shown on the points, but a typical errorbar
1210: is shown by a point in a corner of each panel.}
1211: \label{fig:sheardth}
1212: \end{figure}
1213:
1214: \clearpage
1215:
1216: \begin{figure}
1217: \epsscale{0.8}
1218: \plotone{f8.eps}
1219: \caption{The radial and latitudinal gradients, at $r=0.95R_\odot$
1220: are superposed on the butterfly diagram showing the distribution
1221: of sunspots which is obtained from the Greenwich sunspot data.
1222: The scale is marked in nHz $R_\odot^{-1}$.}
1223: \label{fig:butter}
1224: \end{figure}
1225:
1226: \clearpage
1227:
1228: \begin{figure}
1229: \epsscale{1.0}
1230: \plotone{f9.eps}
1231: \caption{Residuals in the radial gradient
1232: obtained using GONG and MDI data at a few selected depths or latitudes. are shown.
1233: We actually show $\delta\Omega_{,r}\cos\theta$ to minimize latitudinal variation.
1234: The scale is marked in units of nHz $R_\odot^{-1}$.}
1235: \label{fig:domdrfit}
1236: \end{figure}
1237:
1238: \clearpage
1239:
1240: \begin{figure}
1241: \plotone{f10.eps}
1242: \caption{Residual in the latitudinal gradient
1243: obtained using GONG and MDI data at a few selected depths or latitudes.
1244: We actually plot $\delta\Omega_{,\theta}\cos\theta$
1245: to minimize latitudinal variation.
1246: The scale is marked in units of nHz $R_\odot^{-1}$.}
1247: \label{fig:domdthfit}
1248: \end{figure}
1249:
1250: \clearpage
1251:
1252: \begin{figure}
1253: \epsscale{0.8}
1254: \plotone{f11.eps}
1255: \caption{Residual in the radial and latitudinal gradient
1256: obtained at $r=0.98R_\odot$ using MDI data with $\ell<120$ modes.
1257: We actually plot $\delta\Omega_{,r}\cos\theta$ and
1258: $\delta\Omega_{,\theta}\cos\theta$
1259: to minimize latitudinal variation.
1260: The scale is marked in units of nHz $R_\odot^{-1}$.}
1261: \label{fig:mdil120}
1262: \end{figure}
1263:
1264: \clearpage
1265:
1266: \begin{figure}
1267: \epsscale{1.0}
1268: \plotone{f12.eps}
1269: \caption{The acceleration $(1/\omega_0)\partial v_\phi\partial t$
1270: obtained using GONG and MDI data at a few selected depths or latitudes.
1271: The scale is marked in units of m s$^{-1}$.}
1272: \label{fig:dvrot}
1273: \end{figure}
1274:
1275:
1276:
1277: \end{document}
1278: