astro-ph0501093/ms.tex
1: % ms.tex
2: % Revised version
3: % (4-Jan-05)
4: 
5: \documentclass[12pt,preprint]{aastex}
6: 
7: \shorttitle{Observations of Doppler Shift Oscillations}
8: \shortauthors{Mariska}
9: 
10: \begin{document}
11: 
12: \title{Observations of Solar Flare Doppler Shift Oscillations
13:   with the Bragg Crystal Spectrometer on {\it Yohkoh}}
14: \author{John T. Mariska}
15: \affil{E.\ O.\ Hulburt Center for Space Research, Code 7673,
16: Naval Research Laboratory, Washington, DC 20375}
17: \email{mariska@nrl.navy.mil}
18: 
19: \begin{abstract}
20: Oscillations in solar coronal loops appear to be a common
21: phenomenon.  Transverse and longitudinal oscillations have been
22: observed with both the {\it Transition Region and Coronal
23: Explorer} and Extreme Ultraviolet Imaging Telescope imaging
24: experiments.  Damped Doppler shift oscillations have been
25: observed in emission lines from ions formed at flare temperatures
26: with the Solar Ultraviolet Measurements of Emitted Radiation
27: Spectrometer.  These observations provide valuable diagnostic
28: information on coronal conditions and may help refine our
29: understanding of coronal heating mechanisms.  I have initiated a
30: study of the time dependence of Doppler shifts measured during
31: flares with the Bragg Crystal Spectrometer (BCS) on {\it Yohkoh}.
32: This Letter reports the detection of oscillatory behavior in
33: Doppler shifts measured as a function of time in the emission
34: lines of \ion{S}{15} and \ion{Ca}{19}.  For some flares, both
35: lines exhibit damped Doppler shift oscillations with amplitudes
36: of a few km~s$^{-1}$ and periods and decay times of a few
37: minutes.  The observations appear to be consistent with
38: transverse oscillations.  Because the BCS observed continuously
39: for almost an entire solar cycle, it provides numerous flare data
40: sets, which should permit an excellent characterization of the
41: average properties of the oscillations.
42: \end{abstract}
43: 
44: \keywords{Sun: flares --- Sun: corona --- Sun: oscillations ---
45: Sun: X-rays, gamma rays}
46: 
47: \section{INTRODUCTION}
48: 
49: Sequences of images made with the {\it Transition Region and
50: Coronal Explorer} ({\it TRACE}) show that a small subset of
51: large-scale loops appear to respond to nearby flares by
52: exhibiting damped transverse oscillations
53: \citep[e.g.,][]{Aschwanden99,Schrijver02}.  Damped Doppler shift
54: oscillations have also been detected over active regions with the
55: Solar Ultraviolet Measurements of Emitted Radiation (SUMER)
56: experiment on the {\it Solar and Heliospheric Observatory} ({\it
57: SOHO}) \citep[e.g.,][]{Kliem02,Wang03}.  These Doppler shifts are
58: primarily observed in emission lines from ions formed at
59: temperatures greater than 6 MK, and thus are probably associated
60: with flare-like events.
61: 
62: The study of these oscillations is fast developing into a new
63: technology for diagnosing the the physical conditions in the
64: loops and, hence, placing further constraints on possible coronal
65: heating mechanisms \citep[e.g.,][]{Roberts02,Roberts03}.
66: 
67: Detection of Doppler shifts in high-temperature emission lines
68: with SUMER suggests that they might also be detectable with
69: instruments capable of measuring Doppler shifts in soft X-ray
70: lines, such as the Bragg Crystal Spectrometer (BCS) on the {\it
71: Yohkoh} spacecraft.  \citet{Seely84} have in fact detected some
72: evidence for oscillations with a period of about 10 minutes in
73: the position of the \ion{Ca}{19} resonance line during the decay
74: of a limb flare observed on 1980 November 13 with the SOLFLEX
75: crystal spectrometer on the {\it P78-1} satellite.
76: 
77: In this Letter I show that damped Doppler shift oscillations are
78: present in some of the flares observed with the BCS.
79: 
80: \section{BCS OBSERVATIONS}
81: 
82: The {\it Yohkoh} BCS has been described in detail by
83: \citet{Culhane91}.  Four bent crystals cover narrow wavelength
84: ranges centered on emission lines of \ion{Fe}{26}, \ion{Fe}{25},
85: \ion{Ca}{19}, and \ion{S}{15}.  Radiation diffracted by each
86: crystal strikes a one-dimensional position-sensitive proportional
87: counter, with the \ion{Fe}{26} and \ion{Fe}{25} crystals sharing
88: one detector and the \ion{Ca}{19} and \ion{S}{15} crystals
89: sharing a second detector.  All four detectors view the entire
90: Sun.  For this study I use data from the \ion{Ca}{19} and
91: \ion{S}{15} channels.  Examples of the appearance of the spectra
92: in these two channels are available in, for example,
93: \citet{Doschek92}.
94: 
95: To search for Doppler shift oscillations in the BCS data, I have
96: initiated an examination of the temporal behavior of the physical
97: characteristics of a large number of the flares observed by the
98: BCS over its 10-year lifetime.  Fitting the BCS data using
99: synthetic spectra yields estimates of the temperature, emission
100: measure, nonthermal broadening, and Doppler shift of the emitting
101: plasma as a function of time.  The best fit parameters are
102: determined by computing synthetic spectra for an isothermal
103: plasma and then varying the fitting parameters to minimize the
104: $\chi^2$ statistic.
105: 
106: In most cases, the BCS obtains spectra in all four wavelength
107: channels every 3~s.  The \ion{Fe}{26} channel rarely contains
108: useful data, and obtaining Doppler shifts from the \ion{Fe}{25}
109: data is made difficult by the complexity of the spectra in that
110: channel.  Thus, this study is restricted to data from the
111: \ion{S}{15} and \ion{Ca}{19}.  Early or late in a flare, the
112: instrument count rates are small, resulting in spectra that are
113: difficult to fit.  For this survey, I have therefore summed
114: individual spectra until a minimum of 10,000 counts are
115: accumulated in the \ion{S}{15} and \ion{Ca}{19} channels.  This
116: generally results in excellent fits to the data and assures that
117: each data point in the resulting time series of physical
118: parameters will have roughly the same statistical significance.
119: It does, however, mean that the data points in the two derived
120: time series for each flare will not generally be at identical
121: times.
122: 
123: I have analysed data for more than 100 flares observed by the
124: BCS.  Most of the data examined to date cover the first year of
125: the mission.  In addition I have examined flares listed in
126: \citet{Schrijver02} for which there are useful BCS data.  A
127: significant number of the flares observed (roughly 25\%) show
128: evidence for oscillatory behavior.  In the following section, I
129: show the results for one event that exhibited evidence of damped
130: Doppler shift oscillations.
131: 
132: \section{RESULTS}
133: 
134: Figure~\ref{fig:bcs_data} shows the temporal behavior of the
135: count rates and the physical parameters resulting from fits to
136: the individual spectra for the \ion{S}{15} and \ion{Ca}{19}
137: wavelength channels in one flare that exhibits damped oscillatory
138: behavior.  This C2.5 flare took place at S16W88 on 1992 October
139: 2, and is one of the nonocculted limb flares in the survey of
140: occulted and nonocculted limb flares performed by
141: \citet{mariska99}.  Nonocculted flares in that study refers to
142: events for which images obtained with the {\it Yohkoh} Soft X-Ray
143: Telescope (SXT) showed that the footpoints of the flaring loops
144: were not occulted by the solar limb.  At the peak of the flare,
145: the region of strong emission in the SXT images occupies only a
146: few pixels above the solar limb, suggesting that the hot plasma
147: detected with the BCS is confined to a relatively small volume.
148: 
149: While the total count rate, temperature, emission measure, and
150: nonthermal velocity show generally smooth behavior as a function
151: of time, the Doppler shift shows evidence of oscillatory behavior
152: in both channels.  In particular, the Doppler shift in the
153: \ion{S}{15} channel appears to show evidence for a damped
154: oscillation.
155: 
156: To determine the characteristics of the oscillations observed in
157: Figure~\ref{fig:bcs_data}, I follow the example of earlier work
158: on {\it TRACE} oscillations \citep[e.g.,][]{Aschwanden02} and fit
159: the observations with a combination of a damped sine wave and a
160: polynomial background.  For each channel I define a function of
161: the form
162: %
163: \begin{equation}
164: v(t) = A_0 \sin(\omega t + \phi)\exp(-\lambda t) + B(t)\, ,
165: \label{eq:sine_wave}
166: \end{equation}
167: %
168: where
169: %
170: \begin{equation}
171: B(t) = b_0 + b_1 t + b_2 t^2 + b_3 t^3 + \cdots
172: \end{equation}
173: %
174: is the trend in the background data.  Starting from an initial
175: guess for the fit parameters, the data in each channel were fit
176: to Equation~(\ref{eq:sine_wave}) using Levenberg-Marquardt
177: least-squares minimization \citep[e.g.,][]{Press92}.  For all the
178: flares studied the number of terms needed for the background
179: expression was generally one or two, although occasionally three
180: terms produced a better fit.
181: 
182: The Doppler shift data shown in Figure~\ref{fig:bcs_data} are
183: sufficiently noisy that it is difficult to fit them with
184: Equation~(\ref{eq:sine_wave}).  Since it is clear from the figure
185: that the period of the sine wave is on the order of a few
186: minutes, I have accumulated the BCS data for longer intervals by
187: requiring that the minimum number of counts in a channel exceed
188: 10,000 and that the accumulation time be at least 20~s.  The top
189: panels in Figure~\ref{fig:bcs_fits} show the Doppler shift data
190: for the two channels accumulated in this manner.  Also plotted on
191: the top panels is the best fit background trend.  For the
192: \ion{S}{15} spectra, a two-term expression provided the best fit,
193: while for \ion{Ca}{19}, one term was sufficient.
194: 
195: The bottom two panels of Figure~\ref{fig:bcs_fits} show the
196: background-subtracted data along with the best-fit damped sine
197: wave.  For both channels, only the regions between the vertical
198: dashed lines were used in the fitting process.  Even with the
199: longer accumulation time the data are still noisy, especially
200: those from the \ion{Ca}{19} channel.  The fits to the data,
201: however, do appear to be reasonable.  Reduced $\chi^2$ values are
202: 5.8 and 1.2 for the \ion{S}{15} and \ion{Ca}{19} channels,
203: respectively.
204: 
205: Table~\ref{tab:results} summarizes the fitting results.  While
206: the origin of time for the two channels was tied to the first
207: fitted spectrum, the values in the table for \ion{Ca}{19} have
208: been adjusted to correspond to the same initial time used for the
209: \ion{S}{15} fits.  The errors listed in the table correspond to
210: the diagonal elements in the covariance matrix for the fits.
211: Examination of the results in the table shows that both BCS
212: channels show evidence for the same oscillatory behavior in the
213: Doppler shift.  This was generally the case with the other flares
214: that showed damped oscillatory behavior.
215: 
216: \section{DISCUSSION AND CONCLUSIONS}
217: 
218: Figure~\ref{fig:bcs_data} shows that for the first several
219: minutes of the flare the BCS \ion{S}{15} and \ion{Ca}{19}
220: channels show different plasma temperatures.  The comparable
221: oscillation parameters in Table~\ref{tab:results} thus indicate
222: that the flaring structures at the two temperatures are closely
223: related.  This suggests that the oscillations are not restricted
224: to just one isolated loop.
225: 
226: The measured values for the periods (3.93 and 3.57~min) and decay
227: times (3.52 and 4.26~min) fall near the low side of the range of
228: values listed for the 26 oscillating loops observed with {\it
229: TRACE} by \citet{Aschwanden02}.  Following \citet{Wang03}, we
230: define the maximum displacement amplitude for each fit as
231: %
232: \begin{equation}
233: A = A_0/(\omega^2 + \lambda^2)^{1/2} \, .
234: \end{equation}
235: %
236: This yields displacements of 416 and 295~km for \ion{S}{15} and
237: \ion{Ca}{19}, respectively.  These values are less than 1~arcsec.
238: Thus, if the oscillations are due to a position change in the
239: magnetic structure confining the flaring plasma, the change would
240: be invisible at the roughly 3~arcsec spatial resolution of the
241: SXT.  This inferred excursion also falls near the low side of the
242: range listed by \citet{Aschwanden02}.  All the above values fall
243: below the ranges listed by \citet{Wang03} for Doppler shift
244: oscillations observed with SUMER.
245: 
246: The measured values for the oscillations shown in the the table
247: and figures are typical of those determined for the other flares
248: with well-defined oscillations.  Those flares tend to be at or
249: near the solar limb, which, since BCS measures only line-of-sight
250: Doppler shifts is where one would expect to most clearly see
251: oscillatory behavior.  Many other flares in the sample show
252: evidence for significant Doppler shift fluctuations, but the data
253: are not easily fitted with a model such as that used in
254: Equation~(\ref{eq:sine_wave}).  This would suggest that there are
255: many cases where different structures in the flaring plasma are
256: moving in an uncoordinated fashion, leading to a complex signal
257: in the BCS, which observes all the flaring plasma.
258: 
259: Since the BCS has no imaging capability and integrates over all
260: the plasma contributing photons to the emission observed in each
261: wavelength channel, it is not possible to determine from the
262: Doppler shift data alone whether we are observing transverse or
263: longitudinal oscillations.  The periods and decay times measured
264: for the event shown in this Letter and those measured for other
265: events in the study are more typical of those seen in the {\it
266: TRACE} observations \citep{Aschwanden02}---suggesting that the
267: BCS observations are detecting transverse oscillations.  On the
268: other hand, the ratio of the decay time to the period is roughly
269: 0.9, more typical of the results obtained by \citet{Wang03}.
270: 
271: Unfortunately, none of the flare-associated events listed in
272: \citep{Wang03} had useful BCS data.  A number of the events
273: analyzed by \citep{Aschwanden02} have useful BCS observations,
274: and those events are included in a more comprehensive analysis of
275: the BCS Doppler shift observations, which is currently in
276: progress \citep{Mariska05}.  The results of that analysis should
277: provide a firmer estimate of the range of oscillation properties
278: seen in the BCS data and help determine whether they are
279: transverse or longitudinal oscillations.
280: 
281: \acknowledgments
282: I thank G.~A. Doschek, H. Hudson, and H.~P. Warren for their
283: comments on the manuscript.  This research was supported by
284: ONR/NRL 6.1 basic research funds.
285: 
286: \begin{thebibliography}{}
287: 
288: \bibitem[Aschwanden et al.(2002)]{Aschwanden02} Aschwanden,
289:   M. J., DePontieu, B., Schrijver, C. J., \& Title, A. M. 2002,
290:   \solphys, 206, 99
291: 
292: \bibitem[Aschwanden et al.(1999)]{Aschwanden99} Aschwanden,
293:   M. J., Fletcher, L., Schrijver, C. J., \& Alexander, D. 1999,
294:   \apj, 520, 880
295: 
296: \bibitem[Culhane et al.(1991)]{Culhane91} Culhane, J.~L.~et al.\
297:   1991, \solphys, 136, 89
298: 
299: \bibitem[Doschek et al.(1992)]{Doschek92} Doschek, G.~A.~et al.\
300:   1992, PASJ, 44, L95.
301: 
302: \bibitem[Kliem, et al.(2002)]{Kliem02} Kliem, B., Dammasch,
303:   I.~E., Curdt, W., \& Wilhelm, K.\ 2002, \apjl, 568, L61
304: 
305: \bibitem[Mariska(2005)]{Mariska05} Mariska, J.~T. 2005, \apj, in
306:   preparation
307: 
308: \bibitem[Mariska \& McTiernan(1999)]{mariska99} Mariska, J. T.,
309:   \& McTiernan, J. M. 1999, \apj, 514, 484
310: 
311: \bibitem[Press et al.(1992)]{Press92} Press, W. H., Teukolsky,
312:   S. A., Vetterling, W. T., \& Flannery, B. P. Numerical Recipes
313:   in C (2d ed.; Cambridge: Cambridge Univ. Press)
314: 
315: \bibitem[Roberts(2002)]{Roberts02} Roberts, B. 2002, \solphys,
316:   193, 139
317: 
318: \bibitem[Roberts(2003)]{Roberts03} Roberts, B. 2003, in SOHO13,
319:   Waves and Small Scale Transient Events in the Solar Atmosphere:
320:   A Joint View from SOHO and TRACE, ESA SP-547
321: 
322: \bibitem[Schrijver et al.(2002)]{Schrijver02} Schrijver, C. J.,
323:   Aschwanden, M. J, \& Title, A. M. 2002, \solphys, 206, 69
324: 
325: \bibitem[Seely \& Feldman(1984)]{Seely84} Seely, J. F., \&
326:   Feldman, U. 1984, \apj, 280, L59
327: 
328: \bibitem[Wang et al.(2003)]{Wang03} Wang, T.~J., Solanki, S.~K.,
329:   Curdt, W., Innes, D.~E., Dammasch, I.~E., \& Kliem, B.\ 2003,
330:   \aap, 406, 1105
331: 
332: \end{thebibliography}
333: 
334: \clearpage
335: 
336: \begin{deluxetable}{lcc}
337: \tablecaption{Fitting Results\label{tab:results}}
338: \tablewidth{0pt}
339: \tablehead{
340: \colhead{Parameter} & \colhead{\ion{S}{15}} &
341: \colhead{\ion{Ca}{19}}}
342: \startdata
343: $A_0$ (km s$^{-1}$) & $11.3 \pm 1.9$ & $8.76 \pm 2.78$ \\
344: $\omega$ (rad min$^{-1}$) & $1.60 \pm 0.06$ & $1.76 \pm 0.17$ \\
345: $\phi$ (rad) & $-2.43 \pm 0.05$ & $-2.86 \pm 0.41$ \\
346: $\lambda$ (min$^{-1}$) & $0.284 \pm 0.090$ & $ 0.235 \pm 0.130$ \\
347: \enddata
348: \end{deluxetable}
349: 
350: \clearpage
351: 
352: \begin{figure*}
353: \plotone{f1.eps}
354: \caption{Temporal behavior of the temperature, emission measure,
355:   nonthermal broadening velocity, and Doppler shift derived from
356:   the BCS \ion{S}{15} and \ion{Ca}{19} observations for the 1992
357:   October 12 flare.  The zero value for the Doppler shift
358:   velocities has been set to the wavelength shift of the first
359:   fitted spectrum.}
360: \label{fig:bcs_data}
361: \end{figure*}
362: 
363: \clearpage
364: 
365: \begin{figure*}
366: \plotone{f2.eps}
367: \caption{Averaged BCS Doppler shift data and decaying sin wave
368:   fits for the \ion{S}{15} and \ion{Ca}{19} observations for the
369:   1992 October 12 flare.  The zero value for the shifts plotted
370:   in the top panels is based on an arbitrary reference system for
371:   the BCS in which a flare near the solar Equator with no Doppler
372:   shift would have a value of 0~m\AA.  Time is measured from
373:   21:50:44 UT.}
374: \label{fig:bcs_fits}
375: \end{figure*}
376: 
377: \end{document}
378: