0709.3731/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: 
3: 
4: \shorttitle{The Nature of RP Waves Revealed}
5: \shortauthors{D. S. Bloomfield et~al.}
6: 
7: 
8: \begin{document}
9: 
10: 
11: \title{The Nature of Running Penumbral Waves Revealed}
12: 
13: 
14: \author{D. Shaun Bloomfield, Andreas Lagg, and Sami K. Solanki}
15: \email{bloomfield@mps.mpg.de}
16: \affil{Max-Planck-Institut f\"{u}r Sonnensystemforschung, 
17: 	Max-Planck-Str. 2, 37191 Katlenburg-Lindau, Germany}
18: 
19: 
20: \begin{abstract}
21: We seek to clarify the nature of running penumbral (RP) waves: are they 
22: chromospheric trans-sunspot waves or a visual pattern of upward-propagating 
23: waves? Full Stokes spectropolarimetric time series of the photospheric 
24: Si\,{\sc{i}}\,10827~\AA\ line and the chromospheric He\,{\sc{i}}\,10830~\AA\ 
25: multiplet were inverted using a Milne-Eddington atmosphere. Spatial pixels 
26: were paired between the outer umbral/inner penumbral photosphere and the 
27: penumbral chromosphere using inclinations retrieved by the inversion and the 
28: dual-height pairings of line-of-sight velocity time series were studied for 
29: signatures of wave propagation using a Fourier phase difference analysis. The 
30: dispersion relation for radiatively cooling acoustic waves, modified to 
31: incorporate an inclined propagation direction, fits well the observed phase 
32: differences between the pairs of photospheric and chromospheric pixels. We 
33: have thus demonstrated that RP waves are in effect low-$\beta$ slow-mode waves 
34: propagating along the magnetic field.
35: \end{abstract}
36: 
37: 
38: \keywords{Sun: infrared --
39: 	  Sun: magnetic fields --
40: 	  Sun: sunspots --
41: 	  Techniques: polarimetric -- 
42: 	  Waves}
43: 
44: 
45: \section{INTRODUCTION}
46: \label{sec:int}
47: The term running penumbral (RP) wave was created by 
48: \citet{1972ApJ...178L..85Z} to describe chromospheric H$\alpha$ velocity and 
49: intensity fronts that were observed moving out through sunspot 
50: penumbrae. Since then, a host of work has been carried out on reporting their 
51: properties (see, e.g., the series of papers by \citealp{2000A&A...354..305C, 
52: 2001A&A...375..617C} and \citealp{2000A&A...363..306G}), while their exact 
53: nature has remained unidentified. Currently, the two most likely possibilities 
54: for the form of these oscillatory disturbances are:
55: \begin{enumerate}
56: \item trans-sunspot waves generated in the umbra (e.g., by umbral flashes) and 
57: limited to the chromospheric layer, 
58: \item a ``visual pattern'' resulting from field-aligned waves propagating up 
59: from the photosphere.
60: \end{enumerate}
61: To date, many findings point toward RP waves being due to the ``visual 
62: pattern'' scenario \citep[for an extensive discussion of this topic see the 
63: recent review by][]{2006RSPTA.364..313B}. This seems especially likely now 
64: that \citet{2006ApJ...640.1153C} have successfully identified chromospheric 
65: 3-min umbral oscillations as propagating, field-aligned, acoustic waves. 
66: However, recent work by \citet{2006A&A...456..689T, 2007A&A...463.1153T} has 
67: not been able to decide between either of the two possible RP wave scenarios. 
68: Thus, the sum of the evidence is still not conclusive.
69: 
70: In this paper we use velocity time series observations that possess a two-fold 
71: advantage over those of previous studies. The first is the simultaneous 
72: recording of photospheric and chromospheric lines, allowing the connection 
73: between velocities in the lower and upper atmosphere to be accurately 
74: investigated. The second is the retrieval of the magnetic vector at both 
75: heights in the atmosphere through the use of full-Stokes spectropolarimetry, 
76: thus circumventing the need for any assumptions about the possible orientation 
77: of the magnetic field.
78: 
79: 
80: \section{OBSERVATIONAL DATA}
81: \label{sec:data}
82: The dataset used here was obtained from the main spot of active region NOAA 
83: 9451 on 2001 May 9 with the Tenerife Infrared Polarimeter 
84: \citep[{\sc{tip}};][]{1999ASPC..183..264M} attached to the German Vacuum Tower 
85: Telescope in Tenerife, Canary Islands. The 0.5\arcsec$\times$40\arcsec\ 
86: spectrograph slit was positioned across NOAA 9451 for approximately 70~min 
87: with no spatial scanning of the slit, while the solar image was kept 
88: stationary beneath the slit via a correlation tracking device 
89: \citep{1996A&AS..115..353B}. The main umbra from this dataset was previously 
90: analyzed and presented as dataset 2 in the work of 
91: \citet{2006ApJ...640.1153C}, so only a brief description of the observational 
92: setup and format of the data is supplied here. Note the correct heliographic 
93: coordinates of the observed sunspot are S22\degr E20\degr. 
94: 
95: The spectral region obtained in these observations was recorded with a 
96: wavelength sampling of 31~m\AA~pixel$^{-1}$ and includes the photospheric 
97: Si\,{\sc{i}}\,10827.09~\AA\ line, the upper-chromospheric 
98: He\,{\sc{i}}\,10830~\AA\ multiplet (with triplet components at 10829.09~\AA, 
99: 10830.25~\AA, and 10830.34~\AA), and the telluric H$_2$O line at 10832.11~\AA. 
100: The {\sc{tip}} instrument was used to record simultaneous spectral images of 
101: the four Stokes parameters ($I$, $Q$, $U$, $V$) for each of the 0.4\arcsec\ 
102: spatial pixels along the slit. Multiple images were coadded online to increase 
103: the signal-to-noise ratio, resulting in a final cadence of 2.1~s.
104: 
105: 
106: After dark current subtraction, flat-field correction, polarization 
107: calibration, and removal of polarization cross talk 
108: \citep{2003SPIE.4843...55C}, Stokes ($I$, $Q$, $U$, $V$) profiles were 
109: inverted separately for each line using the Milne-Eddington inversion code of 
110: \citet{2004A&A...414.1109L}. For the He\,{\sc{i}} inversions an atmospheric 
111: model with one magnetic component was used, while a non-magnetic component was 
112: included in the Si\,{\sc{i}} inversions to account for stray light. A 
113: continuum intensity space-time plot of the time series is given in 
114: Fig.~\ref{fig:con_B_vel}, alongside absolute magnetic field strengths, 
115: line-of-sight (LOS) velocities, and magnetic field inclinations in solar 
116: coordinates (see \S~\ref{subsubsec:atm_cou} for more details) retrieved by 
117: the Si\,{\sc{i}} and He\,{\sc{i}} inversions.
118: 
119: The Si\,{\sc{i}} and He\,{\sc{i}} LOS velocities retrieved from the inversions 
120: both show systematically decreasing velocity (i.e., blueshift) with time. This 
121: trend arises from the relative motion along the observer's LOS caused by the 
122: Earth's rotation, as these observations were obtained during local morning. 
123: Simple linear fits proved a suitable approximation to the trends and LOS 
124: velocity time series from each spatial pixel had these linear background 
125: trends removed prior to any form of temporal analysis.
126: 
127: 
128: \section{ANALYSIS METHOD}
129: \label{sec:ana_met}
130: Fourier phase difference analysis is a useful tool through which the 
131: propagation characteristics of waves may be determined. The form of analysis 
132: used here is based on the standard Fourier equations that are discussed in 
133: depth by \citet{2001A&A...379.1052K}. This technique has been used extensively 
134: in the past for various solar studies (e.g., \citealp{1963ApJ...138..252J}, 
135: \citealp{1973SoPh...33..333S}, \citealp{1984ApJ...277..874L}, 
136: \citealp{2000ApJ...531.1150W}) and it remains one of the most robust 
137: methodologies in use.
138: 
139: The phase difference spectrum, $\Delta \phi (\nu)$, between two temporal 
140: signals measures the phase lag at discrete frequencies. For signals separated 
141: by some spatial distance, this lag is the cycle-time that it takes different 
142: frequency components to travel from the first location to that of the second. 
143: When waves propagate between two locations in a normally dispersive medium, 
144: phase difference spectra will show zero phase difference at low frequencies 
145: (where waves are evanescent and not propagating) followed by phase differences 
146: which increase in magnitude at higher frequencies (as high-frequency 
147: components travel more slowly than low-frequency components).
148: 
149: The Fourier phase coherence spectrum, $C^2 (\nu)$, between two signals is a 
150: measure of the quality of phase difference variation. However, unless 
151: averaging in frequency is performed, the coherence between two Fourier 
152: components will be unity irrespective of the phase differences. The Fourier 
153: squared coherence of randomly distributed phase differences then approaches 
154: 1/$n$ for averaging over $n$ points in frequency. In this work coherence 
155: values are calculated using an average over 5 frequency intervals, providing a 
156: ``noise'' level of 0.2 for randomly distributed (i.e., uncorrelated) phase 
157: differences. In the following phase difference diagrams, Fourier squared 
158: coherence is represented by the degree of symbol shading (white for 0; black 
159: for 1) while the symbol size represents the cross-spectral power -- a measure 
160: of the co-variance between the Si\,{\sc{i}} and He\,{\sc{i}} LOS velocity 
161: signals.
162: 
163: 
164: \subsection{Trans-sunspot Wave}
165: \label{subsec:tra_sun_wav}
166: If RP waves are due to waves propagating across the sunspot chromosphere, 
167: spectra calculated between the chromospheric LOS velocities from the umbra and 
168: those from pixels at sequentially greater distances into the penumbra should 
169: show phase difference values increasing linearly with frequency, becoming 
170: steeper with greater spatial separation of the signals. In this analysis, 
171: Fourier phase differences and coherences were calculated between the 
172: He\,{\sc{i}} LOS velocity from pixel 55 (located in the umbra) and the 
173: He\,{\sc{i}} LOS velocities from pixels at increasing distances into the 
174: penumbra.
175: 
176: Figure~\ref{fig:fft_tra_sun} shows the output from such an analysis, where 
177: phase difference spectra between the He\,{\sc{i}} LOS velocity from pixel 55 
178: and those pixel numbers listed in each panel are overplotted. For example, 
179: panel {\emph{a}} contains $\Delta \phi_{\mathrm{He}(t, 
180: 55)\rightarrow\mathrm{He}(t, 55)}$ and $\Delta \phi_{\mathrm{He}(t, 
181: 55)\rightarrow\mathrm{He}(t, 54)}$, while panel {\emph{b}} contains $\Delta 
182: \phi_{\mathrm{He}(t, 55)\rightarrow\mathrm{He}(t, 53)}$ and 
183: $\Delta \phi_{\mathrm{He}(t, 55)\rightarrow\mathrm{He}(t, 52)}$, where 
184: $\mathrm{He}(t, Y)$ denotes the temporal He\,{\sc{i}} LOS velocity signal from 
185: spatial pixel $Y$. Although groups of phase difference spectra are overplotted 
186: to increase the clarity of any relations, no clear form of propagating wave 
187: behaviour is seen. Note that around 4~mHz (i.e., 4-min period) values of 
188: increasing phase difference are observed when moving into the penumbra. 
189: However, the finding is marginal since the values of Fourier squared coherence 
190: in the spectra rapidly approach the ``noise'' level of 0.2 for randomized 
191: phase differences.
192: 
193: 
194: \subsection{Upward-propagating Waves}
195: \label{subsec:upw_pro_wav}
196: Although a number of differing forms of propagating wave can exist in the 
197: outer atmosphere of a sunspot (e.g., fast/slow magneto-acoustic and Alfv\'{e}n 
198: waves) we shall restrict our analysis of upward-propagating waves to that of 
199: field-aligned acoustic waves, as these were shown by 
200: \citet{2006ApJ...640.1153C} to describe the phase behaviour of 3-min waves in 
201: sunspot umbrae. The extension of these waves from travelling along 
202: near-vertical field lines in the umbra to travelling along inclined field 
203: lines which expand out over the penumbra is not unexpected since:
204: \begin{enumerate}
205: \item the magnetic field inclination increases smoothly from the umbral centre 
206: out through the penumbra (Figs.~\ref{fig:con_B_vel}{\emph{d}} and 
207: \ref{fig:con_B_vel}{\emph{g}}),
208: \item the photospheric LOS velocity signals are fairly coherent across the 
209: umbra/penumbra boundary (Fig.~\ref{fig:con_B_vel}{\emph{c}}).
210: \end{enumerate}
211: 
212: 
213: To study the possible propagation of field-aligned waves between two 
214: atmospheric heights we must first accurately determine the photospheric pixels 
215: that provide the lower atmospheric signal for the chromospheric pixels which 
216: lie above the penumbra. This is necessary because the field is significantly 
217: inclined here (Figs.~\ref{fig:con_B_vel}{\emph{d}} and 
218: \ref{fig:con_B_vel}{\emph{g}}) and velocity signals in the upper atmosphere 
219: will be spatially removed from their originating photospheric pixels. The 
220: expected picture for field-aligned, upward-propagating waves is indicated in 
221: the schematic diagram of Fig.~\ref{fig:cartoon}, where increasingly inclined 
222: field lines at the photosphere reach further into the chromospheric penumbra.
223: 
224: 
225: \subsubsection{Atmospheric Height Coupling}
226: \label{subsubsec:atm_cou}
227: In order to correctly pair spatial pixels between the photosphere and the 
228: chromosphere we require reliable determination of the magnetic field vector 
229: in solar coordinates. This is complicated by the 180\degr\ azimuthal 
230: ambiguity, whereby two equally valid but opposite azimuth orientations exist 
231: in the observer's coordinate frame. This uncertainty in the field 
232: azimuth impacts on the whole magnetic vector; the two differing azimuthal 
233: solutions yield different solar inclinations, $\gamma^{\prime}$.
234: 
235: To overcome this ambiguity we have implemented a ``smoothest magnetic vector'' 
236: form of ambiguity solution. Namely, a pixel region with the most realistic 
237: solution is selected as a trusted starting point (e.g., in the umbra where 
238: the true solution should be closest to vertical). Moving away from this seed 
239: region, either the 0\degr\ or 180\degr\ azimuth solution is chosen on a 
240: pixel-by-pixel basis to minimise the spatial variation in the three 
241: orthogonal components of the solar magnetic vector ($B_x$, $B_y$, $B_z$), 
242: where the $z$-direction is normal to the solar surface. The field inclinations 
243: achieved in solar coordinates are shown in Figs.~\ref{fig:con_B_vel}{\emph{d}} 
244: and \ref{fig:con_B_vel}{\emph{g}} for the inversion results from Si\,{\sc{i}} 
245: and He\,{\sc{i}}, respectively. 
246: 
247: Spatial pixels were then paired between the photosphere and chromosphere using 
248: the temporal averages of Si\,{\sc{i}} inclinations in the solar coordinate 
249: frame\footnote{Inclinations determined by the Si\,{\sc{i}} inversion were 
250: chosen for this task because the greater signal-to-noise achieved in this line 
251: means that its magnetic vector, and hence solar inclination, is more reliably 
252: determined in comparison to that from the He\,{\sc{i}} inversion.}, $\langle 
253: \gamma^{\prime}_{\mathrm{Si}} \rangle$. Coupled with an expected height 
254: separation, $\Delta H$, these inclinations provide pixel offsets between 
255: photospheric and chromospheric pixel pairs in the direction along the slit by 
256: $\Delta S = \vert \tan \langle \gamma^{\prime}_{\mathrm{Si}} \rangle \vert 
257: \cos \alpha \Delta H / s_{\mathrm{pix}}$, where $\Delta H$ was taken as 
258: 1000~km \citep[following the findings of][for this sunspot 
259: umbra]{2006ApJ...640.1153C}, $s_{\mathrm{pix}}$ is the spatial sampling of the 
260: slit ($\approx$$300$~km\,pixel$^{-1}$), and $\alpha$ is the angle between the 
261: field azimuth and the slit direction. Regions of pixels paired together in 
262: this work are outlined in panels {\emph{b}}-{\emph{g}} of 
263: Fig.~\ref{fig:con_B_vel} by dotted lines, while details of the pixel pairs are 
264: provided in Table~\ref{tab:pixel_pairs} with their corresponding field 
265: inclinations. Although this approach uses the simplifying assumption that the 
266: magnetic field remains essentially linear between the two formation heights 
267: (i.e., there is no field curvature), it is somewhat justified by the resulting 
268: pixel pairs in Table~\ref{tab:pixel_pairs} having inclinations that differ by 
269: $\leqslant$7\degr\ -- i.e., $\langle \gamma^{\prime}_{\mathrm{Si}} \rangle 
270: \approx \langle \gamma^{\prime}_{\mathrm{He}} \rangle$. 
271: 
272: 
273: The Fourier phase difference spectra resulting from these dual-height pixel 
274: pairs are presented in Fig.~\ref{fig:fft_upw_pro}, where spectra from groups 
275: of adjacent pixel pairs are again overplotted to enhance any relations. In 
276: contrast to the trans-sunspot case depicted in Fig.~\ref{fig:fft_tra_sun}, the 
277: expected form of phase difference variation due to propagation (i.e., 
278: increasing values of phase difference with frequency) is clearly apparent in 
279: most of the panels. In addition, throughout panels {\emph{a}}-{\emph{g}} 
280: Fourier squared coherence values remain reasonably high.
281: 
282: 
283: \subsubsection{Dispersion Relation Comparison}
284: \label{subsubsec:dis_rel_fit}
285: We make use of the equations provided in \citet{2006ApJ...640.1153C} which 
286: describe the dispersion relation for vertical acoustic waves propagating 
287: in the presence of a vertical magnetic field within a stratified isothermal 
288: atmosphere with radiative cooling. The equations were modified for this work 
289: to simplistically mimic the first order effects that acoustic-like 
290: (low-$\beta$ slow-mode) waves would experience when propagating along inclined 
291: field lines instead of purely vertically in a vertical magnetic field -- i.e., 
292: $\cos \gamma^{\prime}$ reduced gravity and 1/$\cos \gamma^{\prime}$ increased 
293: path length.
294: 
295: The solid curves in Fig.~\ref{fig:fft_upw_pro} were calculated using the 
296: measured values of field inclination in solar coordinates along with the 
297: temperature (4000~K), radiative cooling time (55~s), and vertical height 
298: separation (1000~km) given by \citet{2006ApJ...640.1153C} for this sunspot 
299: umbra. Although only a simple approximation to the expected dispersion 
300: relation for such waves, these curves show an encouraging association with the 
301: measured data points.
302: 
303: 
304: \section{DISCUSSION}
305: \label{sec:dis}
306: In this section we present our findings in the context of results from 
307: previous studies in an attempt to provide answers to a few of the outstanding 
308: issues surrounding the relationship that RP waves share with other forms of 
309: sunspot waves.
310: 
311: 
312: It has been long known that sunspot chromospheres oscillate at differing 
313: periods in different spatial regions \citep{1972SoPh...27...71G}. 
314: Figure~\ref{fig:he_fft_pow} displays the variation of Fourier power from the 
315: Si\,{\sc{i}} and He\,{\sc{i}} LOS velocities in the form of space-frequency 
316: diagrams. Individual power spectra from each spatial pixel have been 
317: normalized to the variance of the respective time series, resulting in white 
318: noise having power of 1 and 18.4 being the 99.99\% significance level of 
319: Poisson noise. Normalization was performed to aid in the comparison of 
320: spectral profiles between spatial regions that exhibit vastly different LOS 
321: velocity amplitudes. Rather than indicating some form of physical 
322: discontinuity \citep[c.f.,][]{2006A&A...456..689T, 2007A&A...463.1153T}, the 
323: change from dominant chromospheric 3-min power to longer periods near the 
324: umbra/penumbra boundary in Fig.~\ref{fig:he_fft_pow}{\emph{b}} may just result 
325: from the magnetic field inclination becoming large enough to allow 
326: photospheric low-frequency (i.e., 5-min $p$-mode) power to tunnel through the 
327: higher-frequency acoustic cutoff (5.2~mHz) at the temperature minimum 
328: \citep{2004Natur.430..536D}. We note that the classical interpretation of an 
329: acoustic cutoff is effectively negated by the inclusion of radiative cooling 
330: in the modeled dispersion relation of \S~\ref{subsubsec:dis_rel_fit}, which 
331: allows wave reflection and transmission at all frequencies. However, the 
332: dominant chromospheric frequency in Fig.~\ref{fig:he_fft_pow}{\emph{b}} is 
333: modified by the magnetic field inclination, closely following the strong-field 
334: limit $\cos \gamma^{\prime}$ relation of \citet{1977A&A....55..239B}. Power 
335: existing below the cutoff may be explained by the gradual transition from 
336: mainly reflected to transmitted waves around the cutoff (i.e., the slow 
337: turn-up in the curves of Fig.~\ref{fig:fft_upw_pro}). Another possibility is 
338: the presence of unresolved structure in the chromospheric penumbra 
339: \citep{1995A&A...293..252R}, consisting of either two spatially-separated 
340: distributions of field inclination along the LOS or an uncombed magnetic field 
341: configuration \citep{2007A&A...462.1147L}. If the more vertical distribution 
342: has the measured field inclinations and the other has values $\sim$20\degr\ 
343: larger, the acoustic cutoff curve in Fig.~\ref{fig:he_fft_pow}{\emph{b}} could 
344: be pulled to even lower frequencies. At larger field inclinations (i.e., 
345: further into the penumbra) little evidence is found of 3-min waves because 
346: power at 5-min period vastly exceeds that at 3-min in the underlying 
347: photosphere (Fig.~\ref{fig:he_fft_pow}{\emph{a}}).
348: 
349: The termination of 3-min wave patterns at the umbra/penumbra boundary noted by 
350: \citet{2004A&A...424..671K} and \citet{2006SoPh..238..231K} and the fact that 
351: not all 3-min wavefronts can be traced out from the umbra into the penumbra 
352: has been used to suggest that RP waves are not associated with similar waves 
353: in the umbra. A simple check for the linkage of either 3-min or 5-min waves 
354: between the chromospheric umbra and penumbra can be made by bandpass filtering 
355: the He\,{\sc{i}} velocity time series. The spatial variation of He\,{\sc{i}} 
356: LOS velocities through the umbra and penumbra is presented in 
357: Fig.~\ref{fig:filt_5min} before and after filtering in the period ranges $2.5 
358: - 3.5$~min and $4.5 - 5.5$~min. It is clear that each of the 3-min umbral 
359: wavefronts has a rapidly diminishing counterpart in the penumbra 
360: (Fig.~\ref{fig:filt_5min}{\emph{b}}), while each of the 5-min (i.e., RP) 
361: wavefronts has an only somewhat weaker counterpart within the umbra 
362: (Fig.~\ref{fig:filt_5min}{\emph{c}}). These findings once again support the 
363: picture of a continuous variation of RP wave behaviour through the entirety of 
364: the sunspot atmosphere.
365: 
366: 
367: If RP waves are indeed the ``visual pattern'' of upward-propagating waves we 
368: expect that the wave velocity amplitude along the field would be essentially 
369: constant through the penumbra, because each wavefront will have experienced 
370: the same degree of wave growth caused by the decrease of density with 
371: altitude. Examination of the unfiltered He\,{\sc{i}} LOS velocity signal in 
372: Fig.~\ref{fig:filt_5min}{\emph{a}} shows that the RMS LOS velocity decreases 
373: throughout the penumbra. However, the correct quantity to consider is the 
374: RMS velocity along the magnetic field vector. This was obtained from the RMS 
375: LOS velocity (solid curve in the right-hand panel of 
376: Fig.~\ref{fig:filt_5min}{\emph{a}}) using the measured field inclinations from 
377: the observer's LOS -- resulting field-aligned RMS velocities are depicted in 
378: Fig.~\ref{fig:filt_5min}{\emph{a}} by a dotted curve. The field-aligned RMS 
379: velocities still show a decrease close to the umbra/penumbra boundary (from 
380: a greatly diminished contribution of transmitted power at 3-min period) but 
381: also a nearly constant value of $\sim$1~km~s$^{-1}$ over pixels 45 to 25 of 
382: the penumbra, lending more credence to the ``visual pattern'' scenario.
383: 
384: Our results support the conclusion of \citet{2006SoPh..238..231K} that 3-min 
385: umbral waves are not the source of 5-min RP waves. However, we have 
386: additionally shown that they are in fact different manifestations of the same 
387: form of wave generated by a common source at the photosphere, their 
388: differences arising from the transmitted wave power available for propagation 
389: along differently inclined field lines. As such, the observed behaviour of 
390: waves in both umbrae and penumbrae can be explained without the need for 
391: abrupt changes in either density or field orientation at the umbra/penumbra 
392: boundary as postulated by, e.g., \citet{2006A&A...456..689T, 
393: 2007A&A...463.1153T}.
394: 
395: 
396: \section{CONCLUSIONS}
397: \label{sec:con}
398: We have provided evidence that velocity signatures of RP waves observed in the 
399: He\,{\sc{i}}\,10830~\AA\ multiplet are more compatible with upward-propagating 
400: waves than with trans-sunspot waves through careful consideration of the 
401: magnetic vector. Comparing the Fourier phase differences measured between 
402: paired pixels in the photosphere and chromosphere to the dispersion relation 
403: for field-aligned acoustic waves, modified for inclined fields, points toward 
404: such waves (i.e., essentially low-$\beta$ slow modes) being responsible for 
405: the visual pattern. 
406: 
407: Initially excited by a common source at the photosphere, waves experience 
408: increasing path length to the sampling height in the chromosphere with 
409: distance into the penumbra from travelling along increasingly inclined field 
410: lines -- a scenario previously suggested by, e.g., \citet{2003A&A...403..277R} 
411: and \citet{2006RSPTA.364..313B}. For essentially constant (or weakly 
412: increasing) propagation velocities, delays of increasing magnitude will be 
413: observed in the arrival times of wavefronts at increasing radial distance 
414: through sunspot penumbrae. It is the pattern of delayed wavefronts that gives 
415: rise to the apparent outward motion of RP waves which may also explain the 
416: large range of observed wave speeds -- the horizontal ``speed'' of the delayed 
417: wavefronts at the chromosphere depends on the rate at which the magnetic field 
418: inclines out through penumbrae, permitting either sub-sonic or super-sonic 
419: horizontal ``speeds'' for different magnetic geometries. This scenario also 
420: indicates that RP waves may occur at the edges of large pores since the 
421: existence of a penumbra is not necessary to support them; only sufficiently 
422: inclined field lines are required to direct the waves laterally.
423: 
424: 
425: \acknowledgments
426: The German Vacuum Tower Telescope is operated on Tenerife by the Kiepenheuer 
427: Insitute in the Spanish Observatorio del Teide of the Instituto de 
428: Astrof\'{i}sica de Canarias. The authors wish to extend their sincere thanks 
429: to R. Centeno, M. Collados, and J. Trujillo Bueno for providing this excellent 
430: data set for our analysis.
431: 
432: 
433: \begin{thebibliography}{}
434: \bibitem[Ballesteros et~al.(1996)]{1996A&AS..115..353B}
435: Ballesteros, E., Collados, M., Bonet, J.~A., et~al. 1996, \aaps, 115, 353
436: \bibitem[Bel \& Leroy(1977)]{1977A&A....55..239B}
437: Bel, N., \& Leroy, B. 1977, \aap, 55, 239
438: \bibitem[Bogdan \& Judge(2006)]{2006RSPTA.364..313B}
439: Bogdan, T.~J., \& Judge, P.~G. 2006, Phil. Trans. Royal Soc. A, 364, 313
440: \bibitem[Centeno et~al.(2006)]{2006ApJ...640.1153C}
441: Centeno, R., Collados, M., \& Trujillo Bueno, J. 2006, \apj, 640, 1153
442: \bibitem[Christopoulou et~al.(2000)]{2000A&A...354..305C}
443: Christopoulou, E.~B., Georgakilas, A.~A., \& Koutchmy, S. 2000, \aap, 354, 305
444: \bibitem[Christopoulou et~al.(2001)]{2001A&A...375..617C}
445: ---. 2001, \aap, 375, 617
446: \bibitem[Collados(2003)]{2003SPIE.4843...55C}
447: Collados, M.~V. 2003, Proc. SPIE, Volume 4843, 55
448: \bibitem[De Pontieu et~al.(2004)]{2004Natur.430..536D}
449: De Pontieu, B., Erd{\'{e}}lyi, R., \& James, S.~P. 2004, \nat, 430, 536
450: \bibitem[Georgakilas et~al.(2000)]{2000A&A...363..306G}
451: Georgakilas, A.~A., Christopoulou, E.~B., \& Koutchmy, S. 2000, \aap, 363, 306
452: \bibitem[Giovanelli(1972)]{1972SoPh...27...71G}
453: Giovanelli, R.~G. 1972, \solphys, 27, 71
454: \bibitem[Jensen \& Orrall(1963)]{1963ApJ...138..252J}
455: Jensen, E., \& Orrall, F.~Q. 1963, \apj, 138, 252
456: \bibitem[Kobanov et~al.(2006)]{2006SoPh..238..231K}
457: Kobanov, N.~I., Kolobov, D.~Y., \& Makarchik, D.~V. 2006, \solphys, 238, 231
458: \bibitem[Kobanov \& Makarchik(2004)]{2004A&A...424..671K}
459: Kobanov, N.~I., \& Makarchik, D.~V. 2004, \aap, 424, 671
460: \bibitem[Krijger et~al.(2001)]{2001A&A...379.1052K}
461: Krijger, J.~M., Rutten, R.~J., Lites, B.~W., et~al. 2001, \aap, 379, 1052
462: \bibitem[Lagg et~al.(2004)]{2004A&A...414.1109L}
463: Lagg, A., Woch, J., Krupp, N., \& Solanki, S.~K. 2004, \aap, 414, 1109
464: \bibitem[Lagg et~al.(2007)]{2007A&A...462.1147L}
465: Lagg, A., Woch, J., Solanki, S.~K., \& Krupp, N. 2007, \aap, 462, 1147
466: \bibitem[Lites(1984)]{1984ApJ...277..874L}
467: Lites, B.~W. 1984, \apj, 277, 874
468: \bibitem[Mart{\'{\i}}nez Pillet et~al.(1999)]{1999ASPC..183..264M}
469: Mart{\'{\i}}nez Pillet, V., Collados, M., S{\'{a}}nchez Almeida, J., 
470:   et~al. 1999, in ASP Conf. Ser. 183: High Resolution Solar Physics: Theory,
471:   Observations, and Techniques, ed. T.~R. Rimmele, K.~S. Balasubramaniam,
472:   \& R.~R. Radick, 264
473: \bibitem[Rouppe van der Voort et~al.(2003)]{2003A&A...403..277R}
474: Rouppe van der Voort, L.~H.~M., Rutten, R.~J., S{\"{u}}tterlin, P.,
475:   Sloover, P.~J., \& Krijger, J.~M. 2003, \aap, 403, 277
476: \bibitem[R\"{u}edi et~al.(1995)]{1995A&A...293..252R}
477: R\"{u}edi, I., Solanki, S.~K., \& Livingston, W.~C. 1995, \aap, 293, 252
478: \bibitem[Sivaraman(1973)]{1973SoPh...33..333S}
479: Sivaraman K.~R. 1973, \solphys, 33, 333
480: \bibitem[Tziotziou et~al.(2006)]{2006A&A...456..689T}
481: Tziotziou, K., Tsiropoula, G., Mein, N., \& Mein, P. 2006, \aap, 456, 689
482: \bibitem[Tziotziou et~al.(2007)]{2007A&A...463.1153T}
483: ---. 2007, \aap, 463, 1153
484: \bibitem[Wikst{\o}l et~al.(2000)]{2000ApJ...531.1150W}
485: Wikst{\o}l, {\O}., Hansteen, V.~H., Carlsson, M., \& Judge, P.~G. 2000, 
486:   \apj, 531, 1150
487: \bibitem[Zirin \& Stein(1972)]{1972ApJ...178L..85Z}
488: Zirin, H., \& Stein, A. 1972, \apjl, 178, L85
489: \end{thebibliography}
490: 
491: 
492: \begin{figure*}
493: \centering
494: \includegraphics[width=15.6cm]{small-f1_color.eps}
495: \caption{Space-time plots of: continuum intensity at 10825.7$\pm$0.3~\AA\ 
496: ({\emph{a}}), absolute magnetic field strengths in Si\,{\sc{i}} ({\emph{b}}) 
497: and He\,{\sc{i}} ({\emph{e}}), LOS velocities in Si\,{\sc{i}} ({\emph{c}}) and 
498: He\,{\sc{i}} ({\emph{f}}), and magnetic inclinations in solar coordinates from 
499: Si\,{\sc{i}} ({\emph{d}}) and He\,{\sc{i}} ({\emph{g}}). Si\,{\sc{i}} and 
500: He\,{\sc{i}} velocities both have linear background trends removed; 
501: Si\,{\sc{i}} velocities are scaled up by a factor of 8 to the dynamic range of 
502: the He\,{\sc{i}} velocities. Upper solid white lines enclose a lightbridge in 
503: the umbra, while the lower white (black) line marks the umbral/penumbral 
504: (penumbral/quiet Sun) boundary. Temporal averages of the parameters (RMS 
505: values for LOS velocities) are shown in the right-most panels for both 
506: Si\,{\sc{i}} and He\,{\sc{i}} (solid and dotted curves, respectively). Regions 
507: of spatial pixels paired between Si\,{\sc{i}} and He\,{\sc{i}} (see 
508: \S~\ref{subsubsec:atm_cou}) are marked by horizontal dotted lines in panels 
509: {\emph{b}}-{\emph{g}}. Note oscillations in He\,{\sc{i}} field strength and 
510: inclination are not real, but result from misfitting of the Stokes profiles 
511: associated with wave shocking. Observed shock profiles require two components; 
512: the single component inversion used here retrieves weakened, more inclined 
513: fields. However, the He\,{\sc{i}} velocities retrieved still represent the 
514: general plasma motion.}
515: \label{fig:con_B_vel}
516: \end{figure*}
517: 
518: 
519: \begin{figure}
520: \centering
521: \includegraphics[width=\columnwidth]{small-f2_color.eps}
522: \caption{Phase difference spectra between He\,{\sc{i}} LOS velocity in the 
523: umbra and those at increasing distance into the penumbra. Symbol size and 
524: shading denote cross-spectral power and squared coherence, respectively. 
525: Panels show groups of phase difference spectra calculated between umbral pixel 
526: 55 and the pixel numbers listed in the lower-left corners, moving from cases 
527: concerning pixels closest to the umbra/penumbra boundary ({\emph{a}}) toward 
528: those in the middle penumbra ({\emph{h}}). The Fourier squared coherence 
529: ``noise'' level has a value of 0.2 for randomly distributed phase differences 
530: in these data.}
531: \label{fig:fft_tra_sun}
532: \end{figure}
533: 
534: 
535: \begin{figure}
536: \centering
537: \includegraphics[width=\columnwidth]{small-f3.eps}
538: \caption{Cartoon schematic space-time diagram illustrating the form of pixel 
539: coupling between the photosphere and chromosphere for the case of 
540: field-aligned, upward-propagating waves presented in 
541: \S~\ref{subsec:upw_pro_wav}. Dark (light) grids denote umbral (penumbral) 
542: pixels, while dark (light) arrows indicate magnetic lines of force (i.e., wave 
543: paths) linking back to the photospheric umbra (penumbra). Note increasing 
544: delays in wavefront arrival time at the chromospheric sampling height because 
545: of increased propagation lengths along more inclined field lines. The 
546: horizontal and vertical axes are not to scale, resulting in magnetic field 
547: inclinations that appear different to the values actually retrieved from the 
548: Stokes inversions.}
549: \label{fig:cartoon}
550: \end{figure}
551: 
552: 
553: \begin{figure}
554: \centering
555: \includegraphics[width=\columnwidth]{small-f4_color.eps}
556: \caption{Same as Fig.~\ref{fig:fft_tra_sun}, but for spatially-offset 
557: dual-height pairs of photospheric and chromospheric pixels. Curves show 
558: modified acoustic dispersion curves using the measured Si\,{\sc{i}} field 
559: inclinations.}
560: \label{fig:fft_upw_pro}
561: \end{figure}
562: 
563: 
564: \begin{figure}
565: \centering
566: \includegraphics[width=\columnwidth]{small-f5_color.eps}
567: \caption{Space-frequency plots of Si\,{\sc{i}} ({\emph{a}}) and He\,{\sc{i}} 
568: ({\emph{b}}) velocity power. Spectra from each spatial pixel have been 
569: normalized to the variance of the corresponding time series; white noise has 
570: power of unity and 18.4 is the 99.99\% significance level. Overlaid solid 
571: (dotted) curves are the acoustic cutoff modified by the Si\,{\sc{i}} 
572: (He\,{\sc{i}}) inclinations, while horizontal lines mark the same boundaries 
573: as in Fig.~\ref{fig:con_B_vel}.}
574: \label{fig:he_fft_pow}
575: \end{figure}
576: 
577: 
578: \begin{figure}
579: \centering
580: \includegraphics[width=\columnwidth]{small-f6_color.eps}
581: \caption{Space-time plots covering the sunspot umbra and penumbra. {\emph{a}}) 
582: He\,{\sc{i}} velocities retrieved by the inversion. {\emph{b}}) He\,{\sc{i}} 
583: velocities after bandpass filtering in the range $2.5-3.5$~min. {\emph{c}}) 
584: He\,{\sc{i}} velocities after bandpass filtering in the range $4.5-5.5$~min. 
585: Values in panels {\emph{a}} and {\emph{b}} are clipped to enhance wavefront 
586: visibility within the penumbra. White lines mark the umbra/penumbra boundary, 
587: while right-hand panels show unclipped RMS LOS velocity. The dotted curve 
588: included in the upper RMS velocity panel shows the RMS velocity parallel to 
589: the field, after cosine correction for the inclination of the field from the 
590: observer's LOS.}
591: \label{fig:filt_5min}
592: \end{figure}
593: 
594: 
595: \newpage
596: \clearpage
597: 
598: \begin{table}
599: \begin{center}
600: \caption{Spatial pairings between Si\,{\sc{i}} and He\,{\sc{i}} pixels 
601: and their corresponding magnetic field inclinations in solar 
602: coordinates\label{tab:pixel_pairs}}
603: \begin{tabular}{ccccc}
604: \tableline
605: \tableline
606: \multicolumn{2}{c}{Si\,{\sc{i}}}	&  \multicolumn{2}{c}{He\,{\sc{i}}}	&  Notes$^{\dagger}$\\
607: Spatial	&  Inclination	&  Spatial	&  Inclination	&  \\
608: Pixel	&  (\degr)	&  Pixel	&  (\degr)	&  \\
609: \hline
610: 56	&  172		&  55		&  165		&  a\\
611: 55	&  170		&  54		&  165		&  a\\
612: 54	&  169		&  53		&  164		&  b\\
613: 53	&  167		&  52		&  164		&  b\\
614: 52	&  165		&  51		&  163		&  c\\
615: 51	&  163		&  50		&  163		&  c\\
616: 50	&  161		&  49		&  163		&  c\\
617: 49	&  160		&  48		&  163		&  d\\
618: 48	&  158		&  47		&  163		&  d\\
619: 47	&  157		&  46		&  162		&  d\\
620: 47	&  157		&  45		&  161		&  e\\
621: 46	&  155		&  44		&  160		&  e\\
622: 45	&  153		&  43		&  158		&  e\\
623: 44	&  151		&  42		&  156		&  f\\
624: 43	&  150		&  41		&  153		&  f\\
625: 42	&  148		&  40		&  150		&  f\\
626: 41	&  146		&  39		&  147		&  g\\
627: 40	&  145		&  38		&  145		&  g\\
628: 39	&  143		&  37		&  143		&  g\\
629: 39	&  143		&  36		&  142		&  h\\
630: 38	&  141		&  35		&  139		&  h\\
631: 37	&  139		&  34		&  138		&  h\\
632: 36	&  136		&  33		&  136		&  h\\
633: 35	&  134		&  32		&  135		&  h\\
634: \tableline
635: \end{tabular}
636: \tablerefs{$^{\dagger}$ Panel of Fig.~\ref{fig:fft_upw_pro} in which the 
637: resulting Fourier phase difference spectra are overplotted}
638: \end{center}
639: \end{table}
640: 
641: 
642: 
643: \end{document}
644: