1: %version 1: submit to AJ
2: \documentclass[12pt,preprint]{aastex} % one-column, double-spaced document:
3: % \documentclass[manuscript]{aastex}
4: % \documentclass[preprint2]{aastex} % double-column, single-spaced document
5:
6: \def \chandra{{\it Chandra}}
7: \def \hrci{{HRC-I}}
8: \def \ctrt{{ct~s$^{-1}$}}
9: \def \rfrac{{\rho_{{\rm frac}}}}
10: \def \rdiff{{\rho_{{\rm diff}}}}
11: \def \rchis{{\rho_{\chi^2}}}
12: \def \rt{{R}}
13: \def \lc{{C}}
14: \def \hz43{{HZ\,43}}
15: \def \dt{{\delta}}
16: \def \tscale{{\tau}}
17:
18: \shorttitle{Capella Variability}
19: \shortauthors{Kashyap \& Posson-Brown}
20:
21: \begin{document}
22:
23: \title{Short Timescale Coronal Variability in Capella}
24:
25: \author{
26: Vinay L.\ Kashyap\altaffilmark{1},
27: Jennifer Posson-Brown\altaffilmark{1}
28: }
29: \altaffiltext{1}{Harvard-Smithsonian Center for Astrophysics, 60 Garden St., Cambridge, MA 02138}
30:
31:
32: \begin{abstract}
33: We analyze 205~ks of imaging data of the active binary, Capella,
34: obtained with the \chandra\ High Resolution Camera Imager (HRC-I)
35: to determine
36: whether Capella shows any variability at timescales $<50$~ks. We find
37: that a clear signal for variability is present for timescales
38: $\lesssim20$~ks, and that the light curves show evidence for excess
39: fluctuation over that expected from a purely Poisson process. This
40: overdispersion is consistent with variability at the 2-7\% level, and
41: suggests that the coronae on the binary components of Capella are
42: composed of low-density plasma and low-lying loops.
43: \end{abstract}
44:
45: \keywords{ stars: activity, stars: coronae, X-rays: stars }
46:
47: \section{Introduction}
48:
49: Capella ($\alpha$~Aur; G1\,III/G8\,III) is the strongest non-solar
50: coronal source accessible to high-sensitivity high-energy
51: telescopes, and has been a common calibration target for
52: X-ray and EUV instruments such as {\sl EUVE}, \chandra,
53: XMM-{\sl Newton}, etc.
54: It is a remarkably stable source, with no discernible flaring
55: activity. Even though the emission structure has shown changes,
56: especially in the high-temperature regime (see e.g., Dupree
57: et al.\ 1996, Young et al.\ 2001), and there is considerable
58: evidence for the dominant emission to change between the G1\,III
59: primary and the G8\,III secondary (e.g., Linsky et al.\ 1998,
60: Johnson et al.\ 2002, Ishibashi et al.\ 2006), the overall
61: luminosity has remained steady over many years.
62: For instance, Argiroffi et al.\ (2003) detected a change of
63: 3\% in \chandra\ HRC-S/LETGS data over the span of a year, and found
64: no variability at timescales of $0.1-10$~ks. Recently, analysis
65: of \chandra\ ACIS-S/HETGS data (Raasen et al.\ 2007, Westbrook et al.\ 2007)
66: found variations over long timescales, such as an $\approx20\%$
67: intensity enhancement in early 2006, but no evidence for any
68: variability at timescales $<50-100$~ks.
69:
70: Variability in stellar coronae is ubiquitous, and has been detected
71: in all types of coronally active stars
72: (Stassun et al.\ 2006,
73: Caramazza et al.\ 2007,
74: Westbrook et al.\ 2007,
75: G\"{u}del 2004, and references therein)
76: and over a wide range of timescales
77: (Kashyap \& Drake 1999,
78: Favata et al.\ 2005,
79: Stassun et al.\ 2006,
80: Pease et al.\ 2006,
81: Colombo et al.\ 2007). % -- http://arxiv.org/abs/0708.2399
82: This variability can occur due to many causes, ranging from
83: cyclical dynamo variations (timescales of decades to years), to
84: rotational modulation (months to hours), to
85: active region evolution (hours to days), to
86: flaring (hours to minutes).
87: Generally, active stars are characterized by recurrent flares
88: (see e.g., G\"{u}del 2004)
89: that are recognized in X-ray and EUV light curves as sudden
90: increases in the luminosity followed by a slower decay. However,
91: as activity increases, the flares start to occur closer in time
92: to each other, and it becomes increasingly difficult to resolve
93: them in the light curve (cf.\ Kashyap et al.\ 2001). Prominent
94: flares are nevertheless detected in numerous active binaries
95: (see e.g., Osten \& Brown 1999, Osten et al.\ 2004).
96:
97: Despite being one of the more coronally active stars, with a
98: strong high temeperature emission component (Brickhouse et al.\ 2000),
99: flares have never been observed on Capella (Table~\ref{t:capella}).
100: It is unknown whether this is due to a lack of flaring activity
101: to contribute to the heating, or due to a preponderance of flares
102: such that individual events cannot be distinguished (cf.\ Kashyap
103: et al.\ 2001). Here we consider recent observations of Capella
104: made with the \chandra/\hrci\ (\S\ref{s:data}). We analyze these
105: data and find that variability indeed can be detected at short
106: timescales (\S\ref{s:analysis}), suggesting that the latter
107: explanation is more plausible.
108:
109: \begin{table}[htb!]
110: \begin{center}
111: \caption{Capella stellar properties \label{t:capella}}
112: \medskip
113: \begin{tabular}{rl}
114: \hline\hline
115: \multicolumn{2}{l}{$\alpha$~Aur / GJ 194 / HD 34029 / HIP 24608 / HR 1708} \\
116: \hline
117: (R.A., Dec)$_{ICRS 2000.0}$ & (05:16:41.3591, +45:59:52.768) \\
118: distance & 13.4 pc \\
119: %orbital inclination & 41$^\deg$ \\
120: orbital period & 104 days \\
121: Components & G1\,III / G8\,III \\
122: Separation & 109 R$_\odot$ \\
123: Mass & 2.56 / 2.69 [M$_{\odot}$] \\
124: Radius & 9.2 / 12.2 [R$_{\odot}$] \\
125: %T$_{\rm eff}$ & 5700 / 4940 [K] \\
126: M$_{\rm V}$ & 0.14 / 0.25 \\
127: $B-V$ & 0.74 / 0.87 \\
128: rotational velocity & 36 / 3 [km~s$^{-1}$] \\
129: \hline
130: \end{tabular}
131: \end{center}
132: \end{table}
133:
134:
135: %Characterizing this variability is useful for many purposes,
136: %including constraining the heating mechanism (see e.g.,
137: %Audard et al.
138: %Kashyap et al.
139: %G\"{u}del et al.)
140:
141: \section{Data \label{s:data}}
142:
143: \begin{figure}
144: %\epsscale{0.8}
145: \centerline{\includegraphics[width=5.5in,angle=90]{f1.eps}} % fig_hrc_lc_ctrt.eps
146: \caption{The combined light curve of all the \chandra/\hrci\ observations
147: of Capella, spanning 205~ks. The black histogram denotes the count rate
148: for a binning of 100~s, and overlaid on it is the count rate for a binning
149: of 500~s (green histogram). The data gaps between observations are excluded,
150: and indicated by vertical red lines (solid when the gaps are $>100$~s,
151: dashed otherwise). The data comprise 40 ObsIDs (noted at the top of
152: each segment, along with the day since 2005-dec-01 that the observation
153: started). The SIM offset at which each observation is carried out is
154: indicated at the bottom of each segment.
155: \label{f:ctrt}}
156: \end{figure}
157:
158: Capella was observed as a calibration target with the \chandra/\hrci\ over
159: two cycles from December 2005 to January 2007 (see Figure~\ref{f:ctrt}).
160: All the observations were carried out at the telescope aimpoint, but at
161: different locations on the detector corresponding to different offset values
162: of the Science Instrument Module (SIM).
163: The count rates shown in Figure~\ref{f:ctrt} have been corrected
164: for the QE (quantum efficiency) values at the observation location
165: and thus represent flat-fielded light curves matched to the QE
166: at the nominal aimpoint. The QE corrections are made separately
167: for each bin of the light curves as the source dithers across the
168: detector. Note however that this detailed correction is ignorable;
169: it causes changes of $<<1\%$ when compared with count rates corrected
170: with a QE averaged over the entire dither pattern (see also \S\ref{s:hz43}).
171: We have reduced the data using the \chandra\ software for the interactive
172: analysis of observations (CIAO v3.4) and using the most recent calibration
173: products (CALDB v3.2).
174: %which include improvements to both the \hrci\ position degapping correction (Juda 2007)
175: %and the gainmap (Posson-Brown \& Kashyap 2007).
176: %\reference{} Juda, M., 2007, {\sl CXC Calibration Memo}
177: %\reference{} Posson-Brown, J., \& Kashyap, V.L., 2007, SPIE
178: The high count rates observed ($\gtrsim20$~\ctrt), coupled
179: with the sharp point spread function (PSF; it falls to 1\% of
180: the maximum at $\approx1.3''$ away from the peak), provides an unprecedented
181: opportunity to study small changes in the X-ray brightness of Capella.
182: The source counts are extracted from a circle with radius 8x the size
183: of the PSF ($\approx10''$), and the background is locally estimated from
184: a surrounding annulus of radii ($\approx10-33''$).
185: Assuming a Raymond-Smith thermal emission model with a dominant
186: temperature component at 6~MK and a H column N$_H=10^{18}$~cm$^{-2}$,
187: we find with WebPIMMS\footnote{
188: {\tt http://asc.harvard.edu/toolkit/pimms.jsp}
189: }
190: that the counts-to-energy conversion factor is
191: $\sim~5.7-6.0\times10^{-12}$~ergs~cm$^{-2}$~ct$^{-1}$
192: for different metallicities. This suggests X-ray luminosities of Capella
193: in the 0.15-4.5~keV passband of $2.4-2.9\times10^{30}$~ergs~s$^{-1}$
194: (cf.\ L$_X=3.8\times10^{30}$~ergs~s$^{-1}$ based on {\sl Einstein}/IPC
195: observations; %\footnote{
196: %Note that this should not be considered evidence for a 30\% change
197: %in Capella's X-ray luminosity since the mid-1980's since the relative
198: %calibration between the two instruments is not known to that accuracy.
199: %Furthermore, the count-rate linearity of the \hrci\ has not been
200: %established to better than $\approx30\%$ at the high count rates
201: %observed with Capella.
202: %}
203: Strassmeier et al.\ 1993).
204:
205: %%:r /data/susfu/kashyap/Capella/munge_hrc_lc.tbl
206: %%!}sort -n -k 4,4
207: %%!}sed 's/+-/\\pm/'
208: %%!}awk '{print $1 " & " $2 " & " $3 " & " $4 " & $" $5 "$ \\\\ "}'
209: %%!}sed 's/T/ /'
210: %
211: %\begin{deluxetable}{lcccl}
212: %\tabletypesize{\scriptsize}
213: %\tablewidth{0pt}
214: %\tablecaption{ \chandra\ observations of Capella \label{t:data}}
215: %\tablehead{
216: %\colhead{ObsID} & \colhead{Start Time} & \colhead{Exposure} & \colhead{SIM Offset} & \colhead{count rate} \\
217: %\colhead{\hfil} & \colhead{[yr-mon-day hr:min:sec]} & \colhead{[s]} & \colhead{[mm]} & \colhead{[\ctrt]}
218: %}
219: %\startdata
220: %
221: %6558 & 2006-01-10 14:33:16 & 5142.0 & -57 & $20.52\pm0.068$ \\
222: %6557 & 2006-01-10 13:01:26 & 5137.9 & -54 & $20.15\pm0.067$ \\
223: %6556 & 2006-01-10 11:29:36 & 5141.9 & -51 & $20.07\pm0.066$ \\
224: %6555 & 2006-01-10 09:44:07 & 5135.6 & -48 & $20.23\pm0.066$ \\
225: %6554 & 2005-12-23 02:32:04 & 5139.8 & -45 & $21.16\pm0.067$ \\
226: %8359 & 2007-01-21 16:29:27 & 5124.5 & -42 & $22.55\pm0.069$ \\
227: %8358 & 2007-01-20 22:43:07 & 5130.1 & -39 & $22.98\pm0.069$ \\
228: %8357 & 2007-01-20 21:11:17 & 5132.1 & -36 & $22.61\pm0.068$ \\
229: %8356 & 2007-01-20 19:17:53 & 4743.4 & -33 & $22.75\pm0.071$ \\
230: %8355 & 2007-01-17 23:12:05 & 5133.3 & -30 & $21.37\pm0.066$ \\
231: %8354 & 2007-01-17 21:40:15 & 5117.3 & -27 & $21.38\pm0.065$ \\
232: %8353 & 2007-01-17 20:08:25 & 5128.8 & -24 & $21.48\pm0.065$ \\
233: %6553 & 2005-12-20 01:39:43 & 5187.0 & -21 & $21.26\pm0.064$ \\
234: %6552 & 2005-12-25 10:08:33 & 5146.2 & -18 & $21.70\pm0.065$ \\
235: %6551 & 2005-12-25 08:36:43 & 5142.7 & -15 & $21.81\pm0.065$ \\
236: %6550 & 2005-12-25 07:04:53 & 5146.9 & -12 & $21.91\pm0.065$ \\
237: %6549 & 2005-12-25 05:14:59 & 5140.4 & -9 & $21.72\pm0.065$ \\
238: %6548 & 2005-12-23 01:00:14 & 5137.2 & -6 & $20.92\pm0.064$ \\
239: %6547 & 2005-12-22 23:20:27 & 5141.3 & -3 & $21.02\pm0.063$ \\
240: %6559 & 2006-01-10 16:05:06 & 5140.7 & 0 & $19.69\pm0.061$ \\
241: %8360 & 2007-01-21 18:15:43 & 5102.4 & 0 & $21.82\pm0.065$ \\
242: %6540 & 2005-12-06 16:01:25 & 5138.5 & 3 & $22.15\pm0.065$ \\
243: %6541 & 2005-12-07 20:58:02 & 4756.7 & 6 & $21.83\pm0.068$ \\
244: %6542 & 2005-12-07 22:39:40 & 5109.5 & 9 & $21.46\pm0.065$ \\
245: %6543 & 2005-12-06 17:52:33 & 5146.6 & 12 & $22.58\pm0.067$ \\
246: %6544 & 2005-12-06 19:24:23 & 5158.8 & 15 & $22.71\pm0.067$ \\
247: %6545 & 2005-12-19 22:12:06 & 5139.1 & 18 & $21.43\pm0.065$ \\
248: %6546 & 2005-12-20 00:05:46 & 5140.4 & 21 & $21.66\pm0.066$ \\
249: %8352 & 2007-01-17 18:36:36 & 5129.6 & 24 & $21.36\pm0.065$ \\
250: %8351 & 2007-01-17 17:04:46 & 5129.6 & 27 & $21.15\pm0.065$ \\
251: %8350 & 2007-01-17 15:32:55 & 5129.8 & 30 & $21.07\pm0.065$ \\
252: %8349 & 2007-01-17 14:01:05 & 5131.5 & 33 & $21.21\pm0.065$ \\
253: %8348 & 2007-01-17 12:29:16 & 5129.9 & 36 & $21.11\pm0.065$ \\
254: %8347 & 2007-01-17 10:57:25 & 5132.1 & 39 & $20.80\pm0.065$ \\
255: %8346 & 2007-01-17 09:25:36 & 5131.8 & 42 & $21.16\pm0.066$ \\
256: %8345 & 2007-01-17 07:42:33 & 5130.1 & 45 & $21.03\pm0.066$ \\
257: %8344 & 2007-01-15 06:10:28 & 5236.3 & 48 & $21.17\pm0.065$ \\
258: %8343 & 2007-01-15 04:36:53 & 5236.6 & 51 & $21.31\pm0.066$ \\
259: %8342 & 2007-01-15 03:03:18 & 5236.5 & 54 & $21.49\pm0.066$ \\
260: %8341 & 2007-01-15 01:15:47 & 5192.1 & 57 & $21.76\pm0.068$ \\
261: %
262: %%06558 & 2006-01-10T14:33:16 & 5142.1 & -57 & $17.14\pm0.058$ \\
263: %%06557 & 2006-01-10T13:01:26 & 5137.9 & -54 & $17.33\pm0.058$ \\
264: %%06556 & 2006-01-10T11:29:36 & 5141.9 & -51 & $17.65\pm0.059$ \\
265: %%06555 & 2006-01-10T09:44:07 & 5135.6 & -48 & $18.16\pm0.059$ \\
266: %%06554 & 2005-12-23T02:32:04 & 5139.9 & -45 & $19.17\pm0.061$ \\
267: %%08359 & 2007-01-21T16:29:27 & 5124.5 & -42 & $20.68\pm0.064$ \\
268: %%08358 & 2007-01-20T22:43:07 & 5130.1 & -39 & $21.36\pm0.065$ \\
269: %%08357 & 2007-01-20T21:11:17 & 5132.1 & -36 & $21.26\pm0.064$ \\
270: %%08356 & 2007-01-20T19:17:53 & 4743.4 & -33 & $21.50\pm0.067$ \\
271: %%08355 & 2007-01-17T23:12:05 & 5133.3 & -30 & $20.46\pm0.063$ \\
272: %%08354 & 2007-01-17T21:40:15 & 5117.3 & -27 & $20.72\pm0.064$ \\
273: %%08353 & 2007-01-17T20:08:25 & 5128.8 & -24 & $20.92\pm0.064$ \\
274: %%06553 & 2005-12-20T01:39:43 & 5187.0 & -21 & $20.80\pm0.063$ \\
275: %%06552 & 2005-12-25T10:08:33 & 5146.2 & -18 & $21.33\pm0.064$ \\
276: %%06551 & 2005-12-25T08:36:43 & 5142.7 & -15 & $21.54\pm0.065$ \\
277: %%06550 & 2005-12-25T07:04:53 & 5146.9 & -12 & $21.82\pm0.065$ \\
278: %%06549 & 2005-12-25T05:14:59 & 5140.4 & -9 & $21.75\pm0.065$ \\
279: %%06548 & 2005-12-23T01:00:14 & 5137.2 & -6 & $20.95\pm0.064$ \\
280: %%06547 & 2005-12-22T23:20:27 & 5141.3 & -3 & $21.16\pm0.064$ \\
281: %%06559 & 2006-01-10T16:05:06 & 5140.7 & 0 & $19.81\pm0.062$ \\
282: %%08360 & 2007-01-21T18:15:43 & 5102.4 & 0 & $21.97\pm0.066$ \\
283: %%06540 & 2005-12-06T16:01:25 & 5138.5 & 3 & $22.14\pm0.066$ \\
284: %%06541 & 2005-12-07T20:58:02 & 4756.7 & 6 & $21.81\pm0.068$ \\
285: %%06542 & 2005-12-07T22:39:40 & 5109.5 & 9 & $21.38\pm0.065$ \\
286: %%06543 & 2005-12-06T17:52:33 & 5146.6 & 12 & $22.20\pm0.066$ \\
287: %%06544 & 2005-12-06T19:24:23 & 5158.8 & 15 & $22.26\pm0.066$ \\
288: %%06545 & 2005-12-19T22:12:06 & 5139.1 & 18 & $21.01\pm0.064$ \\
289: %%06546 & 2005-12-20T00:05:46 & 5140.4 & 21 & $21.04\pm0.064$ \\
290: %%08352 & 2007-01-17T18:36:36 & 5129.6 & 24 & $20.63\pm0.063$ \\
291: %%08351 & 2007-01-17T17:04:46 & 5129.6 & 27 & $20.48\pm0.063$ \\
292: %%08350 & 2007-01-17T15:32:55 & 5129.8 & 30 & $20.55\pm0.063$ \\
293: %%08349 & 2007-01-17T14:01:05 & 5131.5 & 33 & $20.50\pm0.063$ \\
294: %%08348 & 2007-01-17T12:29:16 & 5129.9 & 36 & $20.30\pm0.063$ \\
295: %%08347 & 2007-01-17T10:57:25 & 5132.1 & 39 & $20.00\pm0.062$ \\
296: %%08346 & 2007-01-17T09:25:36 & 5131.8 & 42 & $20.14\pm0.063$ \\
297: %%08345 & 2007-01-17T07:42:33 & 5130.1 & 45 & $19.89\pm0.062$ \\
298: %%08344 & 2007-01-15T06:10:28 & 5236.3 & 48 & $19.91\pm0.062$ \\
299: %%08343 & 2007-01-15T04:36:53 & 5236.6 & 51 & $19.90\pm0.062$ \\
300: %%08342 & 2007-01-15T03:03:18 & 5236.5 & 54 & $19.88\pm0.062$ \\
301: %%08341 & 2007-01-15T01:15:47 & 5192.1 & 57 & $19.87\pm0.062$ \\
302: %
303: %%\hline
304: %%\end{tabular}
305: %%\end{center}
306: %%\end{table}
307: %\enddata
308: %\end{deluxetable}
309:
310: \section{Analysis and Discussion \label{s:analysis}}
311:
312: \subsection{Autocorrelation \label{s:autocorr}}
313:
314: \begin{figure}
315: %\centerline{\includegraphics[]{fig_hrc_acorr.eps}}
316: %\epsscale{0.8}
317: \plotone{f2.eps} % fig_hrc_acorr.eps
318: \caption{Autocorrelation in the light curve of Capella. The autocorrelation
319: for a light curve binned by 100~s is shown as the dark blue curve, computed
320: for different lag times. The expected autocorrelation for a light curve
321: with no variability is 0 for all non-zero lag times, and the 95\% and 99\%
322: uncertainty bounds on it are shown as horizontal lines (green and red
323: respectively). The bounds are computed via Monte Carlo simulations of
324: a flat light curve with the same mean and number of bins. There is
325: clear evidence for variability at timescales $\lesssim20$~ks. In contrast,
326: the autocorrelation for a light curve constructed by offsetting it by the
327: mean count rate in each ObsID (pale blue curve) shows no evidence of
328: variability.
329: \label{f:autocorr}}
330: \end{figure}
331:
332: It is clear from the \hrci\ light curve (Figure~\ref{f:ctrt}) that
333: Capella undergoes slow changes in its luminosity over timescales of
334: weeks and months, with count rates ranging from $\approx20-23$~\ctrt.
335: For instance, note the drop in intensity between
336: ObsIDs 6552 and 6555, which are separated by 16~days,\footnote{
337: The SIM offset also changes from $-18$ to $-48$~mm between these
338: two observations, and it could be argued that uncalibrated differences
339: in the QE uniformity may account for the drop in intensity. However,
340: the similarity of count rates between observations 6558 and 6559,
341: both of which were carried out on the same day, but at SIM offsets
342: of $-57$ and $0$~mm respectively, indicates that the drop in intensity
343: is real.
344: }
345: and the difference between ObsIDs 6559 and 8360, which are done
346: $\approx1$~yr apart, but are both carried out at the same detector
347: location (thus precluding calibration differences as a factor; see
348: \S\ref{s:hz43} below).
349:
350: This impression is confirmed by an autocorrelation analysis;
351: we construct count rate light curves $\rt(t_i;\dt)$ at various
352: binning sizes $\dt$ and compute the autocorrelation
353: $$ %\begin{equation}
354: P(t_k;\dt) = \frac{\sum_{i=1}^{N_\dt-k} (\rt(t_i;\dt) - \nu(\dt))(\rt(t_{i+k};\dt)-\nu(\dt)) }{ \sum_{i=1}^{N_\dt} (\rt(t_i;\dt)-\nu(\dt))^2} \,,
355: $$ %\end{equation}
356: where
357: $
358: \nu(\dt) %= (1/N)\sum_{i=1}^{N_\dt}\rt(t_i;\dt)
359: $
360: is the average
361: count rate in the $N_\dt$ bins in the light curve. The autocorrelation for
362: $\dt=100$~s is shown in Figure~\ref{f:autocorr} as the dark blue curve
363: (curves for other bin sizes are similar), along with estimates of the
364: 95\% and 99\% uncertainties for each lag time $t_k$. These
365: uncertainties are computed via Monte carlo simulations, by constructing
366: 1000 light curves with the same number of bins as in $\rt(t_i;\dt)$, as
367: Poisson deviates for an unvarying source intensity of $\nu(\dt)\cdot\dt$.
368: The autocorrelation drops linearly until it becomes indistinguishable from
369: statistical noise at a lag time of $\approx20$~ks; this is the typical
370: signature of variability which occurs at timescales $\lesssim20$~ks,
371: such that count rates that are separated by longer timescales are
372: essentially uncorrelated.
373: At larger lag times, $\sim30-60$~ks, the autocorrelation appears to rise
374: again, but this is not distinguishable from statistical noise.\footnote{
375: At even larger lag times, $t_k>65$~ks, the autocorrelation drops further
376: and becomes significantly $<0$, suggesting that the count rates are
377: anticorrelated at large timescales, i.e., the intensity tends to
378: fluctuate over long temporal separations.
379: However, these values are not physically
380: meaningful, since the data gaps between observations are large and the
381: lag time ceases to be a useful construct. Note that typical observation
382: times are 5~ks, and the observation times for contiguous ObsIDs ranges
383: from 12 to 62~ks, with a median of 19~ks (see Figure~\ref{f:ctrt}).
384: % tsp=(t.TSTOP)[icap] & tst=(t.TSTART)[icap] & os=sort(tst) & tst=tst[os] & tsp=tsp[os]
385: % print,tst[1:*]-tsp & tst2=tst & tsp2=tsp & ti_clean,tst2,tsp2 & ti_write,tst2,tsp2
386: % print,median([18564.,11642,18937,18140,23582,29481,24404,61967,18718,13145])
387: }
388: %Note that the ACIS-S/HETGS data probed these timescales, and found
389: %no signal for variability; from the anticorrelation analysis, it
390: %appears that any variability that exists at these timescales are
391: %generally hard to detect.
392:
393: This is consistent with the results found by Raasen et al.\ (2007) and
394: Westbrook et al.\ (2007), who found similar variations over similar
395: timescales. However, at a counts intensity level of $\approx2$~\ctrt\
396: (obtained from ACIS-S/HETG dispersed counts) they were unable to detect
397: any variability at timescales $\lesssim50$~ks corresponding to the
398: durations of the observations,
399: even using sophisticated algorithms such as the one described by
400: Gregory \& Loredo (1992).
401: Here, we observe the source with a counts intensity an order of magnitude
402: higher, and are thus able to investigate the variability at shorter
403: timescales. Note that Argiroffi et al.\ (2003) also find no variability
404: at timescales $<10$~ks, but again, the HRC-S/LETGS data they rely on
405: has count rates of $\lesssim3$~\ctrt, which is too small to detect the
406: existence of variability on Capella.
407:
408: Because the \hrci\ observations are done in short segments, it is useful to
409: consider the effect of removing large timescale effects on the
410: autocorrelation. We thus reconstruct the light curve by offsetting
411: that computed in each ObsID by the average intensity of the source
412: during that ObsID (i.e., for each segment the average count rate is
413: set to $0$), and recompute the autocorrelation. This has the
414: effect of completely removing variations at timescales greater
415: than $\approx\frac{1}{2}$ of the typical exposure time, and will reveal
416: any variability that may exist at very small timescales. The result
417: of this is shown as the pale blue curve in Figure~\ref{f:autocorr}.
418: This is everywhere consistent with no variability. Because the typical
419: observation time is 5~ks, this suggests that there is no variability
420: variability on Capella at timescales $\lesssim3$~ks. Note that offsetting
421: the light curves by different amounts at different times, as we have
422: done in this exercise, introduces additional statistical uncertainty
423: into the results because of a non-stationary bias, and therefore
424: the power of the test to detect a variability signal is decreased.
425: In order to determine whether there does exist variability over
426: timescales of $\sim5$~ks, we test the dispersion of the fluctuations
427: (see \S\ref{s:overdisp}).
428:
429: \subsection{Overdispersion \label{s:overdisp}}
430:
431: %K-S tests on each ObsID are inconclusive. A large fraction of
432: %the p-vals show significance (7/40==17.5\% at $<0.05$, 15/40==37.5\% at $<0.1$)
433:
434: The large count rates of Capella seen with the \hrci\ afford
435: us the capability to analyze the light curve at short timescales.
436: In order to test the constancy of the intensity within each ObsID,
437: we carry out a Kolmogorov-Smirnoff (K-S) test on the photon arrival time
438: data in each ObsID. A large fraction of the datasets show evidence
439: that the null hypothesis of no variation within an observation can be
440: rejected: 17\% have $p<0.05$, and 37\% have $p<0.1$. This is not
441: strong evidence for short timescale variability, but the repeated
442: rejection of the null at a frequency larger than expected indicates
443: the possible existence of intermittent variability. Also note that
444: the K-S test generally has low power in detecting small, slow
445: variations, and a better test is required. For that, we consider
446: whether the observed fluctuations in the light curve are consistent
447: with statistical deviations. Such a test is also of value in
448: establishing the magnitude of the residual calibration errors (see
449: \S\ref{s:hz43}).
450:
451: \begin{figure}
452: \centerline{\includegraphics[width=5.5in,angle=90]{f3.eps}} % capella_rms.eps
453: \caption{Overdispersion in the light curve in each Capella ObsID.
454: The overdispersion measures $\rfrac$ (top), $\rdiff$ (middle), and
455: $\rchis$ (bottom) are calculated for each ObsID for different
456: values of the light curve bin sizes $\dt=25,50,100$~s, and
457: are denoted by the thin vertical lines grouped around the SIM
458: offset for that observation. The lines are offset from each
459: other for clarity and have $\dt$ increasing from left to right,
460: and the measured values for each ObsID are connected by dark lines.
461: The vertical lines represent the $\pm3\sigma$ error bars for the
462: {\sl null}, determined from Monte Carlo simulations of a model
463: without any intrinsic variability but matching the count rate and
464: exposure time of the observation.
465: The values expected for the null model are shown for each
466: $\rho_{(\cdot)}$ as the horizontal dashed line.
467: The overdispersion measures computed for the combined \hz43\
468: data (for a binning that matches $\dt=25$~s for Capella) is
469: shown as the pale blue band whose width corresponds to the
470: $\pm3\sigma$ error bounds determined the same way as for
471: Capella.
472: \label{f:overdisp}}
473: \end{figure}
474:
475: The \hrci\ light curve (Figure~\ref{f:ctrt}) shows numerous sharp
476: fluctuations similar to that expected from Poisson fluctuations.
477: We have tested whether these fluctuations are consistent with such
478: a picture, and conclude that they are not; the observed fluctions are
479: invariably {\sl overdispersed} compared to the expected Poisson
480: deviations (see Figure~\ref{f:overdisp}).
481: In particular, given a counts light curve $\lc(t_i;J,\dt)$ for ObsID $J$,
482: where $t_i$ are the $N_\dt$ time bins resulting from choosing a bin of
483: size $\dt$, we summarize the light curve with its mean
484: $
485: \mu_J(\dt) %= (1/N_\dt) \sum_{i=1}^{N_\dt} \lc(t_i;J,\dt)
486: $
487: and variance
488: $
489: \sigma_J^2(\dt) %= \frac{1}{N_\dt-1} \sum_{i=1}^{N_\dt} [\lc(t_i;J,\dt)-\mu_J(\dt)]^2 \,,
490: $
491: and compute three measures of overdispersion:
492: \begin{mathletters}
493: \begin{eqnarray}
494: \rfrac &\equiv& \frac{\sigma_J(\dt)}{\sqrt{\mu_J(\dt)}} \,, \\
495: \rdiff &\equiv& \frac{\sigma_J^2(\dt)-\mu_J(\dt)}{\mu_J(\dt)} \,, \\
496: \rchis &\equiv& \sum_{i=1}^{N_\dt} \frac{[\lc(t_i;J,\dt)^2-\mu_J(\dt)]^2}{N_\dt\mu_J(\dt)} \,.
497: \end{eqnarray}
498: \end{mathletters}
499: If the light curve fluctuations are fully explained as Poisson
500: fluctuations, we must have
501: $\rfrac\approx1$, $\rdiff\approx0$, and $\rchis\approx1$,
502: since the variance of a Poisson process is equal to its mean.\footnote{
503: Note that unlike the autocorrelation analysis above, here we use
504: counts light curves in order to maintain the correspondence
505: with a Poisson statistical process; the effect of QE non-uniformity
506: is negligible, as described in \S\ref{s:data}, and as demonstrated
507: explicitly in \S\ref{s:hz43}.
508: }
509: In contrast, if an additional process is operating to cause
510: variations in the source intensity, we must have
511: $\rfrac>1$, $\rdiff>0$, and $\rchis>1$.
512:
513: \begin{table}[htb!]
514: \begin{center}
515: \caption{Average overdispersion \label{t:overdisp}}
516: \medskip
517: \begin{tabular}{lccc}
518: \hline\hline
519: $\dt$~[s] & $\rfrac$ & $\rdiff$ & $\rchis$ \\
520: \hline
521: 25 & 1.35$\pm$0.15 & 0.84$\pm$0.40 & 1.83$\pm$0.40 \\
522: 50 & 1.41$\pm$0.31 & 1.08$\pm$1.26 & 2.06$\pm$1.25 \\
523: 100 & 1.50$\pm$0.43 & 1.44$\pm$1.87 & 2.39$\pm$1.84 \\
524: \hline
525: \end{tabular}
526: \end{center}
527: \end{table}
528:
529: For the \hrci\ dataset, computing the overdispersion measures
530: for all ObsIDs, for a variety of time binning sizes
531: ($\dt=25,50,100$~s), we find that the latter condition holds
532: (Figure~\ref{f:overdisp}; see also Table~\ref{t:overdisp})
533: for the majority of the cases.
534: In each case, we compute the error bars on $\rho_{(\cdot)}$ via Monte Carlo
535: simulations of the null model, and thereby determine the significance
536: of each measurement of the overdispersion measure by calibrating the
537: statistic independently for the specific values of $\{\mu,N_\dt\}$.
538: Thus, we find that in most of the observations, the overdispersion is
539: significant, and hence that the observed variations in the
540: Capella light curves cannot be explained as due only to Poisson
541: fluctuations.
542: In particular, we find
543: $\rfrac\approx1.37-1.45$,
544: $\rdiff\approx0.87-1.09$, and
545: $\rchis\approx1.86-2.05$
546: as the range of the medians over the entire dataset for the
547: three different bin sizes considered. The overdispersions
548: tend to increase with $\delta$, as is expected because the
549: relative statistical error decreases with increasing counts.
550: The averages and the standard deviations are reported
551: in Table~\ref{t:overdisp} for each $\dt$, and show that the
552: result is robust over different methods and binning sizes.
553: The measured points are in every case except one greater
554: than the nominal value for the null, and the large standard
555: deviations arise from large deviations upward.
556: Thus we conclude that Capella exhibits variability over
557: durations that characterize the observations, i.e., at
558: timescales $\sim5$~ks. More sophisticated analyses (which
559: are in progress) are necessary to fully characterize this
560: variability.
561:
562: These measurements suggest that the intrinsic variability has
563: an effect that is of similar magnitude to the Poisson process,
564: and that we are able to detect it primarily because of the
565: large count rates and because the large number of independent
566: datasets allows us to check that all give similar results.
567: Our analysis does not allow a direct calculation of the intensity
568: variations that result in these overdispersion values, but
569: assuming a model with small deviations, we estimate that
570: intensity variations of the order
571: $\frac{\Delta\lc}{\lc}\approx0.02-0.07$ are
572: consistent with the observed $\rho_{(\cdot)}$.
573: % ct=22.*[25,50,100] & sig=sqrt(ct)
574: % for f=0.002,0.08,0.002 do print,f,sqrt(sig^2+(f*ct)^2)/sig
575: Since such intensity variations are of the same order as the
576: expected Poisson fluctuations in the light curves, it is not
577: surprising that methods such as the Kolmogorov-Smirnoff test
578: and Fourier analysis fail to detect the existence of the subtle
579: variability in the data.
580: Similarly, Argiroffi et al.\ (2003) place a $3\sigma$ limit
581: on the variability at $<5-10\%$, based on \chandra\ HRC-S/LETGS
582: data where Capella has a count rate of $\approx2.7$~\ctrt.
583: With the \hrci, count rates $\approx8$ times higher are observed,
584: which allows us to detect the existence of such variability
585: at $>3\sigma$ (Figure~\ref{f:overdisp}).
586:
587: \subsection{Calibration \label{s:hz43}}
588:
589: As the source position changes on the detector with each observation,
590: and even during a single observation as it dithers across the
591: detector, the source passes over regions with different quantum
592: efficiencies (QE). Variations in the QE will lead to corresponding
593: differences in the observed count rate, and may produce a false
594: signal of variability. In comparing the source intensities from
595: different ObsIDs, we have taken the QE variations into account
596: using the current best estimate of the \hrci\ QE map, which is
597: estimated to be accurate to $\approx5\%$ across the detector
598: (Posson-Brown \& Donnelly 2004; Figure~\ref{f:ctrt} suggests that
599: the relative error in the QE map is considerably lower).
600: While computing the overdispersion measures however, we do not
601: make such corrections in order to not bias the statistical
602: estimates. Instead, we estimate the contribution of the local
603: variations in QE to the overdispersion by comparing the variances
604: in the light curves generated with and without correcting for
605: this effect: the
606: observed $\sigma(\dt)$ increases by $\lesssim3\%$, which is
607: negligible compared to the magnitude of the observed overdispersions
608: (Figure~\ref{f:overdisp}).
609: As a further test, we have carried out Fourier transforms of
610: the counts light curves; these show no evidence of a periodic
611: signal, as would be present at frequencies corresponding to
612: the dither periods if QE variations were to make a significant
613: contribution to the variability.
614: In addition, we place a direct limit on the residual calibration
615: uncertainty by comparing the analysis of Capella with that of
616: a known non-varying source, \hz43.
617:
618: % NOTE: acc. to Napiwotzki et al.\ (1993), HZ 43B is a M3.5 star
619: % at 42000 RSun away from HZ 43A, which is at 58 pc from Earth.
620: % This means an angular separation of
621: % ((f.RSun*42e3)/(58.*f.pc))*(180./!pi)*(3600.)=3.36 arcsec,
622: % easily separable with Chandra.
623: % \reference{} Napiwotzki, R., Barstow, M.A., Fleming, T., Holweger, H., Jordan, S., \& Werner, K., 1993, A\&A, 278, 478,
624:
625: \hz43\ is a H-rich DA white dwarf with a temperature
626: $T_{\rm eff}\approx50$~kK (Dupuis et al.\ 1998). It has no known
627: or expected variability, and thus serves as a comparison target to
628: verify the effect of the residual calibration uncertainty. The
629: source has been observed numerous times as a calibration target
630: by \chandra\ in the \hrci+LETGS configuration, and has thus far
631: been observed for a total of 31.3~ks at the nominal aimpoint.
632: The observed count rate is 3.6~\ctrt, and coincidentally the
633: accumulated counts
634: approximately matches the number of counts produced in any given
635: Capella ObsID, making it an excellent proxy to test for the
636: existence of any instrument-based variability.\footnote{
637: \hz43\ has also been observed at other times and other locations
638: on the \hrci, but these observations do not have sufficient counts
639: for a useful comparison with Capella data. Furthermore, the
640: off-axis pointings also have large PSFs, which again precludes
641: direct comparisons.
642: %There does exist one set of observations at $+10'$ off-axis
643: %($+30$~mm SIM offset)
644: %that contains a total of $\approx100,000$ counts within a
645: %$10''$ radius, but the QE at this location is suspected to
646: %be decreasing (Posson-Brown, Donnelly, \& Pease 2002), and
647: %the \hz43\ observations span too long a time interval to
648: %serve as useful datasets for calibration.
649: }
650: We measure
651: the overdispersion in \hz43\ at various bin sizes, set such
652: that the same number of counts are expected in each bin as
653: for the Capella data. For instance the Capella light curve
654: binned at $\dt=25$~s is approximately matched by a \hz43\ light
655: curve binned at $\dt=140$~s; i.e., absent intrinsic variations,
656: and assuming that the magnitude of the local QE variations are
657: spatial scale independent, both curves should have identical
658: statistical properties.
659: We find that the \hz43\ overdispersion measures are fully
660: consistent with there being no intrinsic variability whatsoever
661: (see Figure~\ref{f:overdisp}), and thus serve to confirm
662: the detection of overdispersion in the Capella data.
663:
664: %HZ43 plot here
665: %
666: %\begin{figure*}
667: %\centerline{\includegraphics[width=6in,angle=90]{hz43_rms.eps}}
668: %\caption{Overdispersion in the light curve in each HZ~43 ObsID.
669: %\label{f:hz43_overdisp}}
670: %\end{figure*}
671:
672: \subsection{Coronal Structure}
673:
674: Stellar coronae are generally considered to be analogous to the
675: solar corona, in the sense that the X-ray emission arises from
676: optically thin, collissionally excited plasma, which is organized
677: in active regions by magnetic fields and is probably heated by
678: magnetic reconnection events. In the case of Capella, the
679: energetics indeed support this view: the surface area of both
680: components is $\approx100$ times that of the Sun (Table~\ref{t:capella}),
681: and the X-ray luminosity is correspondingly higher, at $\approx100$
682: times the solar luminosity at the peak of its activity cycle. This
683: is consistent with the picture of a solar like atmosphere, with the
684: coronae dominated by loops in active regions that cover a large
685: fraction of the surface. However, the temperatures on Capella's
686: coronae are significantly hotter, and the strong emission component
687: at $T\sim6$~MK has no corresponding structure on the Sun.
688:
689: We have established above (\S\S\ref{s:autocorr},\ref{s:overdisp})
690: that Capella exhibits variability over timescales ranging from
691: $\tscale\sim5-20$~ks.
692: While the longer timescale variability could arise simply due
693: to slow evolution of active regions, the variability at shorter
694: timescales ($\gtrsim5$~ks) is likely due to the dynamical balance
695: of heating events and cooling. Because Capella maintains its luminosity
696: to within a few percent, it must be that heating and cooling events
697: are in balance for the most part, and therefore the variability
698: timescale must be matched to the total coronal heating rate changes,
699: and hence to the radiative cooling timescales.
700: (For typical coronal densities, conductive cooling is not a factor
701: at these timescales.)
702: We can estimate the physical characteristics of the Capella coronae
703: based on this correspondence. The cooling timescale,
704: \begin{equation}
705: \tau = \frac{k_{\rm B} T}{n_e \Lambda(T)} \,,
706: \end{equation}
707: where $n_e$ is the electron number density, and $\Lambda(T)$ is
708: the power emitted by a unit volume of the plasma at temperature $T$.
709: Adopting a coronal temperature of $\log{T}=6.8$, and a
710: metallicity $Z=0.6$ (see Brickhouse et al.\ 2000), the power
711: emitted is $\Lambda\approx2.3\times10^{-23}$~ergs~cm$^{3}$~s$^{-1}$,
712: and we estimate the coronal number density
713: \begin{equation}
714: n_e = 4 \times 10^{9}~\tau_{10}^{-1} ~~ {\rm cm^{-3}} \,,
715: \end{equation}
716: where $\tau_{10}=\frac{\tau}{(10~{\rm ks})}$.
717: This is consistent with the limit $n_e\lesssim10^{10.2}$~cm$^{-3}$
718: found by Ness et al.\ (2003) based on an analysis of the
719: Ne\,IX triplet in \chandra\ and XMM-{\sl Newton} grating data.
720: For the observed
721: luminosity (see \S\ref{s:data}) and the adopted temperature,
722: the volume emission measure $EM\approx10^{53}$~cm$^{-3}$.
723: Assuming that the emission is spread uniformly across the
724: surface of both stars, and adopting a visible surface area 100
725: times that of the Sun, we estimate the height of the corona
726: \begin{equation}
727: h_{\rm est}\approx1.3\times10^{9}~\tau_{10}^2~f_{\rm fill}^{-1} ~~{\rm cm} \,,
728: \end{equation}
729: where $f_{\rm fill}$ is the surface filling fraction of
730: active regions.
731: This is comparable to the typical loop sizes seen on the Sun,
732: but is small compared to the density scale height on Capella,
733: $h_{\rm scale}\approx8\times10^{11}$~cm.
734: %(The density scale
735: %height also implies a lower limit, $n_e>4\times10^{8}$~cm$^{-3}$,
736: %which incidentally excludes conductive cooling as a dominant
737: %factor in the energy balance, since such densities imply a
738: %conductive cooling timescale of $\sim1$~Ms.)
739:
740: We assume here that the entire corona is part of the balance
741: between the heating and cooling occuring at the detected
742: variability timescales. It is possible however that
743: there are multiple components of emission in the corona: the
744: small timescale variability may be due to a small portion of
745: the corona, e.g., as in X-ray bright points on the Sun, and
746: the bulk of the emission may arise in a lower density component
747: that cools slowly and is heated correspondingly slower and
748: reaches to a height comparable to the density scale height.
749: However, this scenario appears to be ruled out by the analysis
750: of high-resolution EUV spectra by Sanz-Forcada, Brickhouse,
751: \& Dupree (2003), who suggest that the bulk of the plasma on
752: Capella is at high densities. Note that while we have confined
753: our attention to $\tscale=5-20$~ks, we cannot formally rule out
754: variability at shorter timescales (more sophisticated analyses
755: are in progress). Such variability will act to increase the
756: estimate for the coronal density $n_e$ and decrease the height
757: of the emission layer $h_{\rm est}$. We speculate that the
758: most probable scenario for emission on Capella is one where it
759: is dominated by numerous low-lying activity sites.
760:
761: %$n_e=4\times10^9~\tscale_{10}^{-1}$~cm$^{-3}$
762: %
763: %$g_*\approx650$~cm~s$^{-2}$
764: %
765: %$h_{\rm X}\sim10^9~\tscale_{10}^2$~cm
766: %
767: %$h_{\rm dens} \sim 8\times10^{11}$~cm
768:
769: \section{Summary \label{s:summary}}
770:
771: We have analyzed a deep \chandra/\hrci\ observation of the
772: active binary Capella to detect variability at timescales
773: shorter than 50~ks. Capella is a highly stable coronal
774: source on which hitherto no short term variability had been
775: seen, despite its relatively high activity level.
776: The \hrci\ is uniquely positioned to achieve this objective
777: because it allows the measurement of large count rates
778: ($\gtrsim20$~\ctrt) with no pileup effects. The unprecedented
779: data quality allows us to test for the existence of variability
780: at timescales as short as 5~ks.
781:
782: We confirm the conclusion of Argiroffi et al.\ (2003),
783: Raasen et al.\ (2007), and Westbrook et al.\ (2007) that
784: Capella exhibits intensity variations at the $3-10\%$ level
785: over timescales of months and years, but unlike those studies
786: which were limited in statistical power due to low count
787: rates, we detect variability at timescales $<50$~ks.
788:
789: We apply numerous statistical tests such as autocorrelation,
790: overdispersion, K-S, etc., to the data, both cumulatively and
791: in individual observation segments, and find that variability
792: does exist at timescales $\tau=5-20$~ks. This suggests that
793: the coronal plasma is at a relatively low density
794: ($n_e\lesssim10^{9}$~cm$^{-3}$) and that the emission arises
795: in low-lying loops with heights $\sim10^{9}$~cm.
796:
797: \acknowledgements
798: This work was supported by the Chandra X-ray Center NASA contract
799: NAS8-39073. We thank Brad Wargelin, Frank Primini, Jeremy Drake,
800: Mike Juda, Nancy Evans, and Scott Wolk for useful discussions.
801:
802: % more sophisticated analyses are in progress
803:
804: %We confirm the primary conclusions of \citet{raa07}, who
805: %found no evidence of changes in intensity over timescales of $<100$~ks,
806: %but a large timescale variation that resulted in an increased flux
807: %early 2006. We find no evidence for variations at timescales $<50$~ks,
808: %and conclusively find that the source is variabile over larger
809: %timescales.
810: %
811: %Within each observation, there is no evidence for any variability,
812: %as is confirmed by an application of the Gregory-Loredo test
813: %\citep{gl92}: the odds that the star varies during any of the
814: %individual observation is $<NUMBER$. In contrast, combining 15 of the
815: %16 available observations (excepting ObsID 1199, a 2 ks observation)
816: %to construct a concatenated event list spanning over 350~ks, and
817: %applying the same test to this concatenated events list, we find that
818: %the odds of Capella being variable are 10:1, indicating a probability
819: %of over 0.99 that Capella is variable on timescales of years.
820:
821: \clearpage
822: \begin{references}
823:
824: %\reference{} Audard, M., G\"{u}del, M., Drake, J., \& Kashyap, V., 2000, ApJ, 541, 396
825: %\reference{} Posson-Brown, J., Donnelly, R.H., \& Pease, D., 2003, at the \chandra\ Calibration Workshop, 27-28 Oct 2003
826: \reference{} Argiroffi, C., Maggio, A., \& Peres, G., 2003, A\&A, 404, 1033
827: \reference{} Brickhouse, N.S., Dupree, A.K., Edgar, R.J., Liedahl, D.A., Drake, S.A., White, N.E., Singh, K.P., 2000, ApJ, 530, 387
828: \reference{} Caramazza, M., Flaccomio, E., Micela, G., Reale, F., Wolk, S.J., \& Feigelson, E.D., 2007, A\&A, 471, 645
829: \reference{} Colombo, J.F.A., Caramazza, M., Flaccomio, E., Micela, G., \& Sciortino, S., 2007, astro-ph/0708.2399
830: \reference{} Dupree, A.K., \& Brickhouse, N.S., 1996, in Poster Proc., IAU Symp.\ 176: Stellar Surface Structure (Wien: Institut f\"{u}r Astronomie), 184
831: \reference{} Dupuis, J., Vennes, S., Chayer, P., Hurwitz, M., Bowyer, S., 1998, ApJL, 500, L45
832: \reference{} Favata, F., Flaccomio, E., Reale, F., Micela, G., Sciortino, S., Shang, H., Stassun, K.G., \& Feigelson, E.D., 2005, ApJS, 160, 469F
833: \reference{} G\"{u}del, M., 2004, A\&ARev, 12, 71
834: \reference{} Ishibashi, K., Dewey, D., Huenemoerder, D. P., \& Testa, P., 2006, ApJ, 644, 117
835: \reference{} Johnson, O., Drake, J. J., Kashyap, V., Brickhouse, N. S., Dupree, A. K., Freeman, P., Young, P. R., \& Kriss, G.A., 2002, ApJ, 565, 97
836: \reference{} Kashyap, V., \& Drake, J.J., 1999, ApJ, 524, 988
837: \reference{} Kashyap, V.L., Drake, J.J., G\"{u}del, M., \& Audard, M., 2002, ApJ, 580, 1118
838: \reference{} Linsky, J. L., Wood, B. E., Brown, A., \& Osten, R. A., 1998, ApJ, 492, 767
839: \reference{} Ness, J.-U., Brickhouse, N.S., Drake, J.J., \& Huenemoerder, D.P., 2003, ApJ, 598, 1277
840: \reference{} Osten, R.A., Brown, A., Ayres, T.A., Drake, S.A., Franciosini, E., Pallavicini, R., Tagliaferri, G., Stewart, R.T., Skinner, S.L., \& Linsky, J.L., 2004, ApJS, 153, 317
841: \reference{} Osten, R.A., \& Brown, A., 1999, ApJ, 515, 746
842: \reference{} Pease, D.O., Kashyap, V.L., Ratzlaff, P.W., Drake, J.J., Evans, N.R., Mossman, A.E., Kim, D., Green, P.J., \& the ChaMP Collaboration, 2006, AAS-HEAD, 9, 01.01
843: \reference{} Posson-Brown, J., \& Donnelly, R.H., 2004, at the \chandra\ Calibration Workshop, 25 Oct 2004
844: \reference{} Raassen, A. J. J., Kaastra, J. S., A\&Ap, 461, 679
845: \reference{} Sanz-Forcada, J., Brickhouse, N.S., \& Dupree, A.K., 2003, ApJS, 145, 147
846: \reference{} Stassun, K.G., van den Berg, M., Feigelson, E., \& Flaccomio, E., 2006, ApJ, 649, 914
847: \reference{} Strassmeier, K.G., Hall, D.S., Fekel, F.C., \& Scheck, M., 1993, A\&AS, 100, 173
848: \reference{} Westbrook, O.W., Evans, N.R., Wolk, S.J., Kashyap, V.L., Nichols, J.S., Mendygral, P.J., Slavin, J.D., \& Waldron, W.L., 2007, ApJS, submitted
849: \reference{} Young, P. R., Dupree, A. K., Wood, B. E., Redfield, S., Linsky, J. L., Ake, T. B., \& Moos, H. W., 2001, ApJ, 555, L121
850:
851: \end{references}
852:
853: \end{document}
854:
855: