1: \documentclass{emulateapj}
2: %\documentclass[12pt,preprint]{aastex}
3: \usepackage{natbib}
4:
5: \slugcomment{{\sc Accepted to ApJ:} June 15, 2007}
6:
7: \citestyle{aa}
8:
9: \newcommand{\msun}{M$_\odot$}
10: \newcommand{\kms}{km s$^{-1}$}
11: \newcommand{\masyr}{mas yr$^{-1}$}
12:
13: \shorttitle{Parallactic Distance to Orion with VLBA}
14: \shortauthors{Sandstrom et al.}
15:
16: \begin{document}
17:
18: %%%%%%%%%%%%%%
19: % TITLE PAGE %
20: %%%%%%%%%%%%%%
21:
22: \title{A Parallactic Distance of $389^{+24}_{-21}$ parsecs to the
23: Orion Nebula Cluster from Very Long Baseline Array Observations}
24:
25: \author{
26: Karin M. Sandstrom,
27: J. E. G. Peek,
28: Geoffrey C. Bower,
29: Alberto D. Bolatto and
30: Richard L. Plambeck}
31:
32: \affil{
33: Department of Astronomy and Radio Astronomy Laboratory, University of
34: California at Berkeley,
35: Berkeley, CA 94720}
36:
37: \email{karin@astro.berkeley.edu}
38:
39: %%%%%%%%%%%%
40: % ABSTRACT %
41: %%%%%%%%%%%%
42:
43: \begin{abstract}
44:
45: We determine the parallax and proper motion of the flaring,
46: non-thermal radio star GMR A, a member of the Orion Nebula Cluster,
47: using Very Long Baseline Array observations. Based on the parallax, we
48: measure a distance of $389^{+24}_{-21}$ parsecs to the source. Our
49: measurement places the Orion Nebula Cluster considerably closer than
50: the canonical distance of $480\pm 80$ parsecs determined by
51: \citet{genzel81}. A change of this magnitude in distance lowers the
52: luminosities of the stars in the cluster by a factor of $\sim 1.5$.
53: We briefly discuss two effects of this change---an increase in the age
54: spread of the pre-main sequence stars and better agreement between the
55: zero-age main-sequence and the temperatures and luminosities of
56: massive stars.
57:
58: \end{abstract}
59:
60: \keywords{astrometry --- stars: distances --- stars: individual (GMR A)
61: --- open clusters and associations: individual (Orion Nebula Cluster)
62: --- techniques: interferometric}
63:
64: %%%%%%%%
65: % BODY %
66: %%%%%%%%
67:
68: \section{Introduction}
69: \label{intro}
70:
71: The Orion Nebula is the nearest example of ongoing massive star
72: formation. As such, it has an important place in our understanding of
73: this fundamental process. Observations of luminosities, masses and
74: linear scales in the Orion Nebula all depend on the distance to the
75: cluster. For many years, the most often used distance measurement to
76: Orion has been $480 \pm 80$ parsecs, as determined by
77: \citet{genzel81}. The Hipparcos mission was only able to marginally
78: detect the parallax of one star in the cluster, resulting in a
79: distance of $361^{+168}_{-87}$ pc. Other techniques of estimating the
80: distance to the cluster provide values with large uncertainties and
81: systematic errors, but which are generally consistent within the
82: Genzel et al. range.
83:
84: At the canonical distance of 480 parsecs, the annual parallax of
85: an object in the Orion Nebula would be about 2 milli-arcseconds. In
86: order to measure this parallax to high precision, angular resolution
87: far in excess of 1 milli-arsecond is necessary. The Very Long Baseline
88: Array (VLBA) can achieve this degree of angular resolution, and it
89: has thus been a very useful tool for determining parallaxes for
90: radio sources out to a few kpc (e.g.
91: \citet{brisken02,chatterjee04,hachisuka06,loinard05}).
92:
93: In order to obtain a precise measurement of the distance to Orion
94: using VLBI, the target source must be compact and persistent. A
95: number of compact radio sources have been observed in Orion, first by
96: \citet{garay87} and later by \citet{felli93a} and others. At least
97: ten of the sources detected by these authors were identified as
98: nonthermal radio stars because of their compactness (unresolved by the
99: VLA A-array) and variability \citep{felli93b}. \citet{bower03}
100: observed one of these sources, GMR A, over the course of an extreme
101: millimeter flaring event, during which its 86 GHz flux density
102: increased by a factor of 5 over a few hours. Follow-up observations
103: with the VLBA detected GMR A at 15 and 22 GHz with a compact size of
104: less than 1 milli-arcsecond. We obtained five additional epochs of
105: VLBA observations spaced over the following year to monitor the
106: astrometric motion of the source. We detect GMR A in four of those
107: five observations, indicating that this source was persistent enough
108: to allow measurement of its proper motion and parallax, thus
109: determining the distance to Orion in a model-independent manner.
110:
111: Unlike other published distance measurements to Orion, the parallactic
112: distance presented here is model-independent. \citet{genzel81}
113: determined a distance of $480 \pm 80$ pc by modeling the proper
114: motions and radial velocities of H$_2$O masers in the BN/KL region
115: with an expanding, thick shell---a technique which required
116: assumptions about the geometry of the system of masers. Most other
117: available distance measurements to Orion rely on stellar models in
118: some way. Because the maser distance is independent of stellar
119: properties it has become the canonical distance to Orion. Recently,
120: \citet{jeffries07} have found a distance of $392 \pm 32$ pc based on
121: the statistical properties of rotation in pre-main sequence stars. In
122: determining this distance, \citet{jeffries07} assume a spectral type -
123: effective temperature scale for pre-main sequence stars and a random
124: distribution of spin axes. In addition, they also must accurately
125: identify stars which are still in the accretion phase of their
126: pre-main sequence lifetime. Parallax provides a fundamental measure
127: of the distance, avoiding the systematic uncertainties associated with
128: these other techniques.
129:
130: Accurate knowledge of the distance to the Orion Nebula Cluster is
131: important for the general understanding of star-formation in the
132: region. Luminosities are proportional to distance squared
133: and the ages of pre-main sequence stars are typically determined by
134: comparing their temperatures and luminosities with theoretical models
135: (for example, \citet{palla99}). Changes in the luminosity of the
136: stars translate directly into changes in their inferred ages, and the
137: age distribution of the pre-main sequence stars in the cluster is a
138: key component of any model which attempts to explain the mode of
139: star-formation in the region.
140:
141: In Section~\ref{sec:data} we describe the VLBA observations and in
142: Section~\ref{sec:reduction} we discuss their reduction. In particular
143: we discuss the use of a dual calibrator method developed by
144: \citet{fomalont05}. Section~\ref{sec:analysis} covers the analysis of
145: our positions for the source and presents our best fit values for
146: the parallax and proper motion. In Section~\ref{sec:discussion} we
147: discuss how this measurement will affect the study of the Orion Nebula
148: and its star-formation and compare our measurement with previous
149: results.
150:
151: \section{Observations and Data Reduction}
152:
153: \subsection{VLBA Observations}\label{sec:data}
154:
155: The observations were carried out at a frequency of 15 GHz with all
156: available VLBA antennas. There were six epochs of VLBA observations:
157: an initial observation of the source in January of 2003 in the wake of
158: its flaring event and five epochs evenly spaced over one year to
159: adequately sample the entire parallactic ellipse of GMR A.
160:
161: To determine the astrometric position of GMR A, we use
162: phase-referencing to the bright quasar J0541$-$0541 which is located 1.6
163: degrees to the southeast of the target. Additionally, to account for
164: phase variation due to atmospheric gradients between J0541$-$0541 and
165: GMR A, we observe a secondary calibrator J0529$-$0519 which is 1.3
166: degrees northwest of GMR A and approximately colinear with GMR A and
167: J0541$-$0541. The data reduction utilizing these two calibrators in
168: the phase-referencing is described in the following section. Each
169: observation, excluding January 2003, consisted of alternating 40
170: second integrations on the two calibrator sources and on GMR A for 6
171: hours. The first epoch, January 2003, was part of a campaign to
172: understand the millimeter and x-ray flare from GMR A at high
173: resolution. The secondary calibrator was not included in this track,
174: which consisted of alternating observations of only GMR A and
175: J0541$-$0541 for 40 seconds each. A very bright calibrator, J0530$+$1331,
176: was observed a few times during the course of each track for use in
177: removing instrumental phase offsets and delays in each IF.
178:
179: \subsection{Data Reduction}\label{sec:reduction}
180:
181: The data reduction was performed in AIPS. The initial
182: calibrations proceeded as described in the AIPS cookbook for
183: VLBA data reduction. These include amplitude calibration and
184: fringe-fitting to determine the instrumental single-band delay
185: (slope of phase vs. frequency introduced by the instrument) and
186: the residual phase delays and rates introduced by the atmosphere and
187: inaccuracies in the correlator model.
188:
189: Recently discovered errors in the Earth Orientation Parameters (EOPS)
190: used by the VLBA correlator for all epochs excluding January 2003 were
191: corrected using the task CLCOR as described in the VLBA Memo
192: 69\footnote{VLBA Test Memo 69, October 6, 2005
193: \url{http://www.vlba.nrao.edu/memos/test/test69memo/index.html}}. The
194: data were corrected for the single- and multi-band delays with the
195: task FRING. The target source is detected at 5-$\sigma$ or greater,
196: in four out of the six epochs using only phase-referencing to the main
197: calibrator J0541$-$0541. In the October 2004 observation, GMR A is
198: only detected after ATMCA correction as described below. In April
199: 2004, there are a number of significant peaks in the image because the
200: flux from GMR A has been scattered by poor phase coherence, probably
201: due to fluctuations in tropospheric water vapor. We have omitted the
202: April 2004 data from the following analysis because of its poor
203: quality.
204:
205: In order to improve the image quality and remove some systematic
206: errors in the position of the target source we used dual-source
207: phase-referencing as implemented in the AIPS task ATMCA
208: \citep{fomalont05}. This corrects for phase gradients across the sky
209: due to tropospheric effects. The correction can be done in a number
210: of ways, but in our case we used interpolation between two calibrator
211: sources placed on either side of the target. Prior to ATMCA we also
212: correct for the effects of structure in our calibrators through
213: self-calibration cycles on both sources. It is clear from the images
214: that both calibrators have resolved structure, which makes the
215: self-calibration cycle important for accurate imaging of the target
216: source. We note that the structure of the main calibrator appears to
217: be approximately constant over the time period of our observations. We
218: see at most a few percent variation in the brightness distribution of
219: J0541$-$0541 comparing between epochs with clean components all
220: restored to the lowest resolution of the five observations. Because
221: of the lack of variation in the source structure and the high
222: signal-to-noise of the calibrator observations, the self-calibration
223: cycle should be sufficient to correct systematic errors in the phase
224: solutions due to the structure of the calibrator. After ATMCA
225: calibration we see an increase in the compactness of GMR A of up to 17
226: percent. Additionally, in the October dataset, the significance of
227: the GMR A detection changes from $<4$ to 5 sigma. The effects of the
228: dual-source calibration are summarized in Table~\ref{tab:atmca}.
229:
230: \input{tab1.tex}
231: \input{tab2.tex}
232: \input{tab3.tex}
233:
234: After ATMCA, cleaned images of the sources were made using the task
235: IMAGR. In each epoch, we measured the position of GMR A with
236: reference to the main calibration source J0541$-$0541 for which we
237: assume the J2000.00 position R.A. $5^{\rm h}41^{\rm m}38\fs084106$,
238: Dec. $-5^\circ41\arcmin49\farcs42841$. The position assumed for the
239: secondary calibrator, J0529$-$0519 in J2000.00 is R.A. $5^{\rm
240: h}29^{\rm m}53\fs532715$, Dec. $-5^\circ19\arcmin41\farcs61564$.
241: This position is used in the course of ATMCA calibration. The five
242: images of GMR A are shown in Figure~\ref{fig:gmra}, centered on the
243: maximum point in each image. The images of the calibrators
244: J0541$-$0541 and J0529$-$0519 are shown in Figures~\ref{fig:j0541}
245: and~\ref{fig:j0529}, respectively. In these figures the positions are
246: relative to the maximum of the image in the December 2003 observation.
247:
248: It is clear from the integrated flux densities listed in
249: Table~\ref{tab:gmra} that we observe significant variability in GMR A
250: at 15 GHz. This observation is in line with previous studies that
251: show large changes in the source flux density
252: \citep{felli93b,zapata04,bower03}. \citet{bower03} postulate that the
253: structure of GMR A consists of a compact, highly variable source that
254: the VLBA detects and an extended, $\sim 5$ mJy envelope that is
255: resolved out. Our observations certainly indicate that the source
256: detected by the VLBA is highly variable.
257:
258: \begin{figure*}[h]
259: \centering
260: \epsscale{1.0}
261: \plotone{f1.eps}
262: \caption{Images of GMR A for all epochs. All images are after ATMCA
263: calibration, except for the January 2003 image, as described in the
264: text. Each image is centered on the brightest pixel of that image.
265: The contour levels are -2.5, 2.5, 4, 7, 10, 15 and 20 times the noise
266: level of the individual image. The dotted contours are the negative
267: values. The noise levels are listed in Table~\ref{tab:gmra}. The
268: sythesized beam is shown in the lower left corner.}
269: \label{fig:gmra}
270: \end{figure*}
271:
272: \begin{figure*}
273: \centering
274: \epsscale{1.0}
275: \plotone{f2.eps}
276: \caption{Images of the primary calibrator at each epoch
277: with ATMCA calibration, except for the January 2003 observation which
278: does not have ATMCA calibration but has been through an amplitude and
279: phase self-calibration cycle. The images are centered on the location
280: of the brightest pixel in the December 2003 image. The contour levels
281: are -5, 5, 20, 50, 100, 200, 500 times the noise level of the
282: individual image. The synthesized beam is shown in the lower left of
283: each image. Negative contours are shown with dotted lines. In the
284: January 2003 image there are no pixels less than -5 times the noise
285: level.}
286: \label{fig:j0541}
287: \end{figure*}
288:
289: \begin{figure*}
290: \centering
291: \epsscale{1.0}
292: \plotone{f3.eps}
293: \caption{Images of the secondary calibrator, with ATMCA calibration.
294: The images are centered on the location of the brightest pixel in the
295: December 2003 image. The contours are -5, 5, 10, 20, 40 and 60 times
296: the noise level of the individual image, The synthesized beam is shown
297: in the lower left of each image. Negative contours are shown
298: with dotted lines.}
299: \label{fig:j0529}
300: \end{figure*}
301:
302: \subsubsection{Structure in the VLBI Images of GMR A}
303:
304: The images of GMR A shown in Figure~\ref{fig:gmra} show resolved
305: structure. At the distance we measure to the cluster, stellar
306: photospheres should be unresolved. We can estimate the photospheric
307: radius of GMR A from the bolometric luminosity of 6 L$_\odot$ and the
308: spectral type K5 V as determined by \citet{bower03}, given the
309: effective temperature - spectral type calibration for pre-main
310: sequence stars of \citet{cohen79}. This yields a radius of $\sim 4.3$
311: R$_\odot$, which at the distance of Orion is $\sim 50$ $\mu$as, well
312: below the resolution of our observations. Since there is resolved
313: structure present in the images, we must consider its source in
314: determining the position of GMR A. Scattering by turbulent
315: interstellar plasma can cause angular broadening. We estimate the
316: effects of interstellar scattering on our images using the
317: \citet{cordes02} model, which predicts an angular broadening scale at
318: 15 GHz of $\sim 3 \ \mu$as for the line of sight to the calibrators
319: and $\sim 0.05 \ \mu$as for GMR A. The measured size of our
320: calibrators are consistent with very little or no interstellar
321: scattering, in agreement with the model predictions. Some of the
322: structure in our images is most likely due to atmospheric calibration
323: errors. In particular, the correction from ATMCA applies mostly in
324: the Right Ascension dimension due to the positioning of our two
325: calibrator sources East-West relative to GMR A. It is the case that
326: most of the images are elongated more in the Declination dimension
327: compared to the beam shape, suggesting atmospheric effects.
328:
329: However, there is also more persistent structure, mainly obvious when
330: GMR A is bright. \citet{bower03} noted this structure and found that
331: the January 2003 observation could be equally well represented by a
332: Gaussian extended relative to the beam shape or by two unresolved
333: sources with a separation of 0.8 mas. In the December 2003 and 2004
334: observations as well, we see structures that are better represented by
335: two unresolved sources or an extended Gaussian. We considered the
336: possibility of binarity to explain the resolved structure and we note
337: that if GMR A were a binary, orbital motion would introduce scatter
338: into our solution for the parallax and proper motion. Using the
339: positions determined by fitting two unresolved sources to the images,
340: we attempted to find a mass ratio which gave a center of mass position
341: between the two purported sources that decreased the scatter in the
342: parallax and proper motion solution. Because there is no consistent
343: way to distinguish between the two sources, we also searched through
344: the 16 possible permutations of primary and secondary components. For
345: each of these permutations, we also attempted to fit a relative visual
346: binary orbit using a technique based on that of \citet{hartkopf89}.
347: We were unable to find a consistent solution that improved the scatter
348: of our parallax and proper motion using these two complementary
349: techniques. Therefore, it seems unlikely that the structure in our
350: images is due to a binary companion.
351:
352: A more likely explanation for the structure in our images is large
353: scale magnetic features associated with the star. These have been
354: observed with VLBI in other magnetically active weak-line T Tauri
355: stars \citep{phillips91,andre92} with sizes of
356: up to 10-20 R$_*$. A separation of 0.8 mas at the distance we measure
357: to Orion corresponds to $\sim 15$ R$_*$, which is comparable to
358: structures observed on other weak-line T Tauri stars. We do know that
359: GMR A has a $\sim 2$ kG photospheric magnetic field \citep{bower03}
360: and experiences a high level of magnetic flaring activity as seen in
361: x-ray, millimeter and radio observations, so it seems reasonable to
362: expect large scale magnetic structures around the star. The effect
363: of the positional jitter introduced in our measurements by imperfectly
364: calibrated atmosphere and by source structure is to increase the error
365: bars in the distance determination.
366:
367: \subsubsection{Astrometric Positions of GMR A}
368:
369: If GMR A were unresolved and we could account for the effects of the
370: atmosphere by observing a known point source, the best technique for
371: determining the positions of the target would be to fit a fixed-size
372: Gaussian to the images. The size of the Gaussian would take into
373: account the blurring effects of the atmosphere. However, since GMR A
374: appears to have resolved structure which may vary with time and
375: neither of our calibrators are unresolved, the best technique
376: for determining the source positions is to fit variable size
377: Gaussians to the images. We did this using the AIPS task JMFIT,
378: allowing the dimensions and position angle of the fit Gaussian to
379: vary. Table~\ref{tab:fits} lists the dimensions of the beam for each
380: observation and the properties of the best fit Gaussian. The
381: positions derived from these fits are listed in Table~\ref{tab:gmra}.
382: As a check, we have also fit the images with a Gaussian constrained to
383: have the shape and size of the beam. The positions we find from these
384: two techniques are consistent within their error bars, however the
385: positions we determine from fitting the constrained Gaussian have
386: smaller error bars, increasing the $\chi^2$ of our parallax and proper
387: motion solution, though it yields the same distance.
388:
389: \input{tab4.tex}
390:
391: In the case of the January 2003 observation, ATMCA can not be applied,
392: as only one calibrator was observed. To account for any systematic
393: differences in the positions determined with and without ATMCA, we
394: increased the uncertainty of the January 2003 observation by adding
395: the r.m.s. effect of ATMCA on the other three observations in
396: quadrature with its observed position errors (85 $\mu$as in R.A. and
397: 18 $\mu$as in Dec.).
398:
399: \section{Astrometric Analysis}\label{sec:analysis}
400:
401: The position of GMR A as a function of time is determined by its
402: position, parallax and proper motion in the following way:
403: \begin{equation}
404: \alpha(t) = \alpha_0 + \mu_{\alpha}t + \pi f_{\alpha}(\alpha,\delta,t)
405: \end{equation}
406: \begin{equation}
407: \delta(t) = \delta_0 + \mu_{\delta}t + \pi f_{\delta}(\alpha,\delta,t)
408: \end{equation}
409: Here $\mu_{\alpha}$ and $\mu_{\delta}$ are the proper motions in Right
410: Ascension and Declination, respectively, and $\pi$ is the parallax.
411: $f_{\alpha}$ and $f_{\delta}$ are the parallactic displacements for a
412: source at a distance of 1 pc at the position of GMR A. The parallactic
413: displacements are calculated based on the formulae presented in the
414: U.S. Naval Observatory Almanac.
415:
416: To fit to these five positions (ten data points) we employed a
417: $\chi^2$ fit to five parameters: position in R.A. and Dec, proper
418: motion in R.A. and Dec., and parallax. The best fit solution had a
419: reduced $\chi^2$ value of $\chi_R^2 = 10.2$. This indicates that our
420: error bars on the position are approximately $\chi_R = 3.2$ times
421: larger than the formal errors from Gaussian fitting, mostly due to
422: systematic effects from tropospheric variations and variability of the
423: source structure. To determine the error bars on the best fit
424: parallax and proper motion we proceeded in the following way: first,
425: we added systematic errors in quadrature to the error bars listed in
426: Table~\ref{tab:gmra} to achieve $\chi_R^2 = 1$; we then did a Monte
427: Carlo simulation in which we added offsets, drawn from a Gaussian
428: distribution centered at zero with a width representing the total
429: positional errors from the previous step, to the measured positions
430: and fit for the astrometric parameters. The resulting distributions
431: of parallax and proper motion were well-represented by Gaussians,
432: allowing a straightforward determination of the 1-$\sigma$ errors.
433:
434: The best approach to adding in systematic errors to achieve $\chi_R^2
435: = 1$ is not well-defined, so we approach the problem with a few
436: different techniques. To start, we added in quadrature the same
437: systematic error in the R.A. and Dec dimensions till we reached
438: $\chi_R^2 = 1$. This gave a parallax of $2.53 \pm 0.18$ mas
439: (equivalent to a distance of $395^{+30}_{-26}$ pc). However, it is
440: most likely not the case that the systematic errors are the same in
441: R.A. and Dec. We know that our correction from ATMCA mostly applies
442: in the Right Ascension dimension, due to the positioning of our
443: calibrators East-West relative to the target. To address the
444: non-uniformity of the systematic errors we added in quadrature error
445: ellipses with varying axial ratios and determined the axial ratio for
446: which the total area of the systematic error ellipse necessary to
447: achieve $\chi_R^2 = 1$ was smallest. As expected based on the
448: positions of our calibrator sources, the necessary systematic error
449: ellipse is larger in the Declination dimension by a factor of 2.5. To
450: reach $\chi_R^2 = 1$ the geometric mean of the axes of the error
451: ellipse with these dimensions was 0.17 mas. This more realistic
452: appraisal of our systematic errors gives a parallax of $2.57 \pm 0.15$
453: mas ($389^{+24}_{-21}$ pc). It is also possible that the systematic
454: errors are epoch dependent. In this situation it is not clear how
455: best to add in systematic errors unless we base them on the beam shape
456: or the errors from Gaussian fitting (which are propotional to each
457: other in theory, $\sigma \sim 0.5
458: \frac{\theta_{\mathrm{FWHM}}}{\mathrm{SNR}}$). To explore this
459: possibility we have scaled up the error ellipses from Gaussian fitting
460: to achieve $\chi_R^2 = 1$. This technique gives a parallax of $2.61
461: \pm 0.14$ mas ($383^{+22}_{-20}$ pc), consistent with the other
462: techniques.
463:
464: We note that the astrometric parameters obtained through these various
465: techniques are robust to our treatment of the errors. We consider
466: adding the fixed axial ratio error ellipse to be the most realistic
467: appraisal of the systematic errors given the positioning of the
468: calibrators. This technique yields a parallax of $\pi = 2.57 \pm
469: 0.15$ mas, which corresponds to a distance of $389^{+24}_{-21}$ pc.
470: The proper motions of GMR A in R.A. and Dec. from this solution are
471: $\mu_{\alpha} \cos{\delta} = 1.89 \pm 0.12$ \masyr \ and $\mu_\delta =
472: -1.67 \pm 0.19 $ \masyr. At the distance we measure, the transverse
473: velocity is $v_t = 4.65 \pm 0.39$ \kms. Figure~\ref{fig:solution}
474: shows a plot of our solution and Table~\ref{tab:params} lists the
475: measured astrometric parameters and their uncertainties.
476:
477: \input{tab5.tex}
478:
479: \begin{figure*}
480: \centering
481: \epsscale{0.9}
482: \plotone{f4.eps}
483: \caption{The measured positions of GMR A with the best fit parallax
484: and proper motion. The diamonds represent the predicted
485: position of GMR A for each observation. The error bars on the
486: measured positions are scaled as described in the text. The dashed
487: line is the proper motion of the source, with the parallactic motion
488: subtracted. The parallax corresponds to a distance of
489: $389^{+24}_{-21}$ pc.}
490: \label{fig:solution}
491: \end{figure*}
492:
493: \section{Discussion}\label{sec:discussion}
494:
495: The Orion Nebula is part of a very large and very complex
496: star-formation region (see \citet{genzel89} for a review of the large
497: scale structure). The implications of a measurement of the distance
498: to one star thus depend sensitively on where that star is located
499: relative to the stellar associations and molecular gas in this region.
500: The available evidence very strongly constrains GMR A to be a part of
501: the Orion Nebula Cluster, located within a few parsecs of the
502: Trapezium stars but embedded in the molecular cloud which is currently
503: being disrupted by the ionizing radiation from $\Theta^1$ Ori C. In
504: the following paragraphs we will present the evidence for placing GMR
505: A in the Orion Nebula Cluster and then proceed to compare our distance
506: to previous measurements and, finally, briefly discuss some of the
507: implications our distance measurement has for the general study of
508: this important region.
509:
510: \subsection{The Membership of GMR A in the Orion Nebula Cluster}
511:
512: \citet{bower03} observed GMR A, following its intense outburst
513: detected with the BIMA interferometer, in the near-IR with NIRC and
514: NIRCSPEC on the Keck telescopes. While there is no optical source
515: coincident with the position of GMR A, there is an infrared source
516: which they identify as a weak-line T Tauri star embedded in molecular
517: gas. GMR A was also detected in the COUP survey as a highly variable
518: x-ray source (COUP J053551.8$-$052149) \citep{feigelson02,getman03}
519: behind a gas column density of N$_{\mathrm{H}} = 10^{22.3}$ cm$^{-2}$.
520: The combination of the spectral classification as a K5 V star and a
521: bolometric luminosity of 6 L$_\odot$ \citep{bower03} suggests a very
522: young age for this star (around 1 Myr, based on the pre-main sequence
523: tracks of \citet{palla99}) as does its location embedded in molecular
524: gas.
525:
526: The molecular gas in Orion A, located behind the H II region created
527: by the Trapezium stars, has a very high column depth $A_V \approx
528: 50-100$ mag, essentially providing a wall behind which no background
529: sources are detected. Near-IR surveys have shown that the projected
530: spatial distribution of optically visible and extincted, near-IR
531: sources are very similar \citep{hillenbrand98} and that around 50\% of
532: sources are only visible in the infrared. The three-dimensional
533: distribution of the Orion Nebula Cluster has thus been interpreted as
534: fairly spherical with about half of that sphere still embedded in the
535: molecular cloud out of which it formed \citep{hillenbrand98}. The
536: projected distribution of stars has a radius of $\sim 3$ pc around
537: $\Theta^1$ Ori C. GMR A is located 1.95 arcminutes away from the
538: center of the cluster, which corresponds to a projected distance of
539: 0.22 parsecs at our measured distance to the cluster. The x-ray
540: absorption column density measured to GMR A indicates that it must be
541: embedded in the molecular gas, but it cannot be more than a few
542: parsecs deep. The lack of foreground and background star-forming
543: regions, the proximity of GMR A to the Trapezium on the sky, the fact
544: that it is embedded in molecular gas, and its very young age
545: convincingly place GMR A as a member of the Orion Nebula Cluster.
546:
547: In addressing the membership of a stellar cluster, a typical technique
548: is to compare the velocity of a star to the average velocity and
549: dispersion of the cluster. There have been a number of measurements
550: of these quantities for the ONC with proper motion and radial velocity
551: studies. Recent spectrocopic observations of stars in the cluster by
552: \citet{sicilia-aguilar05} show a mean heliocentric velocity of 25
553: \kms, with a $\sim 2$ \kms \ uncertainty in the zero point and a
554: dispersion of $\sigma = 2.3$ \kms. \citet{bower03} measured a
555: heliocentric radial velocity of $14\pm 5$ \kms \ for GMR A. Given the
556: combined uncertainties in these measurements and the cluster radial
557: velocity dispersion, GMR A is consistent with the radial velocity of
558: the cluster.
559:
560: Most proper motion studies of clusters, including that of
561: \citet{jones88} which surveyed nearly 1000 stars in Orion, measure
562: relative proper motions. In order to compare our measurement, which
563: is an absolute proper motion, to the results of these studies, we must
564: first find the absolute proper motion of the ONC to subtract from our
565: values. This value has been measured by a number of authors using
566: different astrometric catalogs. \citet{baumgardt00} use the Hipparcos
567: catalog and measure a mean proper motion of $\mu_\alpha \cos(\delta) =
568: 1.73 \pm 0.40$ \masyr \ and $\mu_\delta = -0.47 \pm 0.27$ \masyr \ for
569: the ONC. \citet{kharchenko03} used the ASCC catalog (which contains
570: data from Hipparocs, Tycho and the USNO catalogs) and found very
571: similar values $\mu_\alpha \cos(\delta) = 2.02 \pm 0.74$ \masyr \ and
572: $\mu_\delta = -0.19 \pm 0.66$ \masyr. Later, \citet{kharchenko05}
573: improved their determination of the absolute proper motion with an
574: expanded astrometric catalog and found $\mu_\alpha \cos(\delta) = 1.96
575: \pm 0.31$ and $\mu_\delta = -0.77 \pm 0.46$ \masyr. All of these
576: values are consistent within their error bars, so we proceed in our
577: analysis using the recent value of \citet{kharchenko05}. We measure a
578: proper motion for GMR A of $\mu_\alpha \cos(\delta) = 1.89 \pm 0.12$
579: \masyr \ and $\mu_\delta = -1.67 \pm 0.19$ \masyr. These values are
580: quite similar to the cluster mean. We can compare the residual
581: velocity after subtracting the mean to the velocity dispersion of the
582: cluster. The velocity dispersion of the ONC has been measured by a
583: number of authors. The largest survey of relative proper motions was
584: carried out by \citet{jones88}. They found a one-dimensional proper
585: motion dispersion of $\sim 1.1$ \masyr \ for all of the cluster stars
586: in their survey (I magnitude of 16 or higher). Additionally, they
587: found a trend of decreasing velocity dispersion with increasing mass
588: (and therefore, magnitude) for stars in the ONC.
589: \citet{hillenbrand98} show that for stars with masses between 0.1 and
590: 0.3 \msun \ the dispersion is $<\sigma> \approx 1.26$ \masyr \ and for
591: stars between 1 and 3 \msun \ it decreases to $1.00$ \masyr. The
592: proper motion of GMR A is less than 1-$\sigma$ from the cluster
593: average indicating a very high membership probability.
594:
595: At the distance we measure to the ONC, a velocity of 1 \masyr \ is
596: equivalent to 1.85 \kms. Relative to the cluster, GMR A is moving
597: $0.13 \pm 0.61$ \kms \ West, $1.67 \pm 0.93$ \kms \ South, and $-11.0 \pm
598: 5.4$ \kms \ along the line of sight. We note that there are large
599: uncertainties in the radial velocity, both for GMR A and for the
600: cluster as a whole. However, even at $-11$ \kms \ relative to the
601: cluster, GMR A would only have moved $\sim 10$ pc over its lifetime.
602:
603: There is one measurement of the proper motion of GMR A in the
604: literature by \citet{gomez05}. They determined the absolute
605: proper motions of 35 radio sources in Orion with archival VLA data
606: spanning 15 years. Although we have much higher angular resolution in
607: our observations, their long time baseline makes the comparison
608: useful. They measure the absolute proper motion of the cluster to be
609: $\mu_\alpha \cos(\delta) = 0.8 \pm 0.2$ \masyr \ and $\mu_\delta
610: \cos(\delta) = -2.3 \pm 0.2$ \masyr. These values are quite different
611: from others in the literature. The absolute proper motion for GMR A
612: they find is $\mu_\alpha \cos(\delta) = -2.02 \pm 1.87$ \masyr \ and
613: $\mu_\delta = 1.15 \pm 1.93$ \masyr. Our measurement is only
614: consistent with these values at the 2-sigma level. They
615: find a substantially higher velocity dispersion for the cluster than
616: found in optical surveys and theorize that the population of radio
617: sources might have larger random velocities. However, using our
618: proper motion, GMR A has a very typical velocity compared to the
619: optical sources.
620:
621: To summarize, GMR A is a very likely member of the Orion Nebula Cluster
622: based on its youth, proximity to the cluster on the sky, proper
623: motion, and location embedded in molecular gas. Having established
624: its place in the Orion region, we now compare our distance measurement
625: to those in the literature.
626:
627: \subsection{Comparison with Previous Measurements}
628:
629: Despite its importance in our understanding of star-formation, the
630: distance to the Orion Nebula Cluster is quite uncertain. Till
631: recently there has only been one model-independent distance
632: measurement---the marginially significant parallax from Hipparcos of
633: the star HD 37061 which corresponds to a distance of
634: $361^{+168}_{-87}$ parsecs \citep{bertout99}.
635:
636: Many authors have estimated the distance to the cluster by fits to the
637: upper main sequence, a procedure which is complicated by systematic
638: uncertainties in the models, the variable background and extinction of
639: the Orion Nebula, and the region over which stars are included in the
640: analysis, among other issues. Some results with this technique are
641: those of \citet{penston73} who found a distance of $363^{+26}_{-24}$
642: parsecs, using infrared photometry to characterize each star's
643: extinction individually; a later, more detailed analysis by
644: \citet{penston75} who found $\sim 400 \pm 20$; \citet{warren78} found
645: a distance of $483^{+57}_{-51}$ parsecs; and \citet{anthony-twarog82}
646: who reanalyzed the data from \citet{warren78} found a distance of
647: $434^{+21}_{-19}$ parsecs. These results alone---some of which make
648: use of the same data, others which use the same techniques---show a
649: spread of more than 100 parsecs, illustrating the difficulty of such
650: measurements. Many authors have also attempted statistical techniques
651: which combine the proper motion and radial velocity distributions of
652: the stars with assumptions regarding cluster expansion or contraction.
653: Two early examples of this technique are \citet{strand58} who find a
654: distance of 525 parsecs and \citet{johnson65} who find 380 parsecs.
655:
656: Although the Hipparcos mission could only marginally detect the
657: parallax of a single star in the ONC, combination of the astrometric
658: measurements of many stars has yielded some results for Orion.
659: \citet{wilson05} analyzed the aggregate astrometric motions of stars
660: foreground and background to the Orion A in three distinct regions of
661: the cloud. They estimate a distance of $465^{+75}_{-57}$ parsecs to
662: the ONC region. This distance is larger than what we measure, but is
663: consistent with our measurement at the one-sigma level.
664: \citet{brown94} also used arguments regarding foreground and
665: background stars, but instead comparing their reddening and 100 $\mu$m
666: fluxes to estimate distances to the near and far edges of the cloud of
667: 320 and 500 parsecs.
668:
669: The orbital parameters of binary systems can be used in some cases to
670: determine very accurate distances. \citet{stassun04} have analysed an
671: eclipsing binary 0.3 degrees south of $\Theta^1$ C Ori and found it to
672: be at $419 \pm 21$ parsecs. However, its status as a member of the
673: ONC is somewhat uncertain because of its relatively old age and
674: location relative to the cluster. The distance to this binary
675: agrees with our distance to the ONC. Recent work by
676: \citet{kraus07} on the visual binary system containing $\Theta^1$Ori C
677: yields two equally good orbital solutions which their data cannot yet
678: distinguish. Assuming a luminosity-mass relationship for the stars
679: they find distances of $384 \pm 11$ and $434 \pm 12$ parsecs from the
680: two orbits. Further astrometric and spectroscopic observations of
681: this system may provide a very accurate distance in the future.
682:
683: \citet{genzel81} combined observations of the proper motions and
684: radial velocites of H$_2$O masers in the BN/KL region with the
685: assumption of a spherical, uniformly expanding, thick shell and found
686: a distance of $480 \pm 80$ parsecs. Because of its relative precision
687: and independence from stellar evolution calculations, this distance
688: has become the canonical distance to the Orion Nebula Cluster.
689: Indeed, many studies of the ONC population and star-formation make use
690: of this distance without consideration of its substantial
691: uncertainties. Our distance measurement is 20\% closer than that
692: of \citet{genzel81}, though within their combined uncertainties.
693: However, our distance is more precise and does not depend on the
694: assumption of geometries for expanding maser sources which has been
695: shown to be considerably more complex in many cases than what these
696: authors assume \citep{greenhill05}.
697:
698: Very recently, \citet{hirota07} measured the annual parallax of one H$_2$O
699: maser spot in the Orion BN/KL outflow using VERA (VLBI Exploration of
700: Radio Astronomy). The parallax they measure corresponds to a distance
701: of $437 \pm 19$ parsecs using only the motion in Right Ascension, and
702: $445 \pm 42$ when they solve for the parallax using both the Right
703: Ascension and Declination. Note that these error bars are only
704: statistical, and the latter value agrees with ours at the 1-$\sigma$
705: level. Parallax determinations from maser spot motion can be
706: problematic, because of intrinsic structure changes and/or spot
707: acceleration. For the spot in question \citeauthor{hirota07} detect
708: changes in its emission line profile over their two years of
709: observation, which likely indicate changes in the structure of the
710: source. Stellar parallax measurements are less vulnerable to this
711: type of systematic uncertainty. If the difference between our VLBA
712: measurement and the \citeauthor{hirota07} VERA determination is real,
713: and the VERA error bars are accurate, it may argue for a larger
714: separation between the BN/KL region and the Orion Nebula Cluster along
715: the line of sight.
716:
717: Finally, we compare our distance to the recent work of
718: \citet{jeffries07} who have used the statistical properites of
719: pre-main sequence star rotation to determine a distance of $392 \pm
720: 32$ parsecs to the ONC. This techniques involves the assumption of
721: random spin orientations for the stars in the cluster and makes use of a
722: spectral type - effective temperature scale for pre-main sequence
723: stars. From the whole sample analyzed by \citet{jeffries07}, a
724: distance of $440 \pm 34$ parsecs is derived. However, excluding stars
725: which show evidence of accretion---a factor which makes the necessary
726: radius determination unreliable---lowers the distance to $392 \pm 32$
727: parsecs. Our distance agrees very well with this latter value.
728: \citet{jeffries07} note that the distance they determine is very
729: sensitive to the assumed spectral type - effective temperature scale,
730: such that distance scales with the square of the effective
731: temperature. In addition, it is clear that the exclusion of Classical
732: T Tauri stars, which show evidence for accretion, significantly
733: changes the derived distance, indicating that an accurate account of the
734: various sources of systematic error is extremely important to this
735: technique.
736:
737: To summarize, we present a distance to the ONC of $389^{+24}_{-21}$
738: parsecs using the fundamental technique of parallax. In contrast to
739: previous distance estimates, this measurement does not rely on
740: modeling of stellar evolution, cluster dynamics, or calibrations of
741: effective temperature for pre-main sequence stars. In comparing our
742: measurement to previous values, we find it to be consistent with many
743: prior measurements within their substantial error bars. We find good
744: agreement with recent values from \citet{jeffries07}, \citet{kraus07}
745: and \citet{stassun04}, but independent of the various assumptions
746: those authors are forced to make.
747:
748: \subsection{Implications of a Closer Distance to the ONC}
749:
750: Many studies of the Orion Nebula assume the distance of
751: \citet{genzel81} or similar. Our distance is 20\%
752: closer. In luminosity terms, this change means luminosities are a
753: factor of 1.5 lower than previously claimed, a result that is
754: especially important for the determination of pre-main sequence star
755: ages, which scale with luminosity as $t\propto L^{-3/2}$. Therefore,
756: at a distance of 390 parsecs, the stars are nearly twice as old as
757: they would be at 480 parsecs. This scaling of age with luminosity is
758: true only for fully convective pre-main sequence stars that have
759: contracted by a substantial amount from their initial radii (see, for
760: instance, \citet{palla99} for further discussion). Near the birthline
761: this assumption breaks down and the ages will not be affected to the
762: same degree. A decrease in luminosity will therefore age the entire
763: population of pre-main sequence stars, but not uniformly, increasing
764: the age spread of the population.
765:
766: The distribution of stellar ages in Orion is the basis for many
767: theories describing star-formation in the region and massive,
768: clustered star-formation in general. These theories can be broken
769: down into two general categories---those that purport that star
770: formation in the ONC happened suddenly and quickly
771: (e.g. \citet{hartmann01,hartmann03}) and those that argue that star
772: formation has been occuring over a longer timescale, but has been
773: accelerating in recent times (e.g. \citet{palla00,tan06,huff06}).
774: The age spread of cluster members is a fundamental measurement of the
775: timescale of star-formation in the cluster. A larger age spread, as
776: would result from the decrease in luminosity, tends to favor models
777: where star-formation has occurred over a more extended period of time.
778:
779: Another interesting feature of the stellar population in Orion is that
780: the high mass stars seem to fall above the expected zero-age main
781: sequence for the cluster \citep{palla99,hillenbrand98}.
782: \citet{palla99} argue that this must be due to a systematic problem in
783: determining the luminosities, effective temperatures or some
784: combination thereof. These analyses assumed a distance of 470
785: parsecs, so there is indeed a systematic offset---according to our
786: measurement these luminosities are too high by a factor of 1.5. A
787: shift of this magnitude brings the luminosities of the high mass stars
788: into much better agreement with the zero-age main sequence
789: predictions.
790:
791: \section{Conclusions}\label{sec:conclusions}
792:
793: We have monitored the astrometric motion of the flaring, non-thermal
794: radio star GMR A over the course of two years with the Very Long
795: Baseline Array. We determine from these data the proper motion and
796: parallax of the star. Based on its young age, its proximity to the
797: center of the Trapezium on the sky, the consistency of its proper
798: motion with that of the cluster and its location embedded in molecular
799: gas, the probability that GMR A is a member of the Orion Nebula
800: Cluster is very high, and thus the distance we determine based on its
801: parallax is representative of the cluster as a whole. We find the ONC
802: is at a distance of $389^{+24}_{-21}$ parsecs, nearly 100 parsecs
803: closer than the canonical distance of 480 parsecs determined by
804: \citet{genzel81}. The distance presented here is in good agreement
805: with recent work by \citet{jeffries07}, with the advantage of being
806: independent of assumptions about stellar properties. A closer
807: distance has important implications for the study of star-formation in
808: the Orion Nebula, one of the most well-studied sites of massive star
809: formation, most notably the increase in the ages and the age spread of
810: the pre-main sequence stars in the cluster. The decrease in
811: luminosity also brings the more massive stars into better agreement
812: with the zero-age main-sequence.
813:
814: Further VLBA observations of other radio stars in the ONC may overcome
815: the limitations of a single star distance and begin to probe the depth
816: of the cluster. In addition, suitable targets for this type of
817: observation should be found in most clusters with substantial pre-main
818: sequence populations. Future VLBI observations of magnetically
819: active, pre-main sequence stars in these clusters could provide
820: precise, fundamental distance measurements to many nearby star-forming
821: regions.
822:
823: \acknowledgements
824:
825: The authors would like to thank the referee for thorough and helpful
826: comments. We would also like to thank Steve Stahler and Reinhard
827: Genzel for sharing their expertise on the Orion region. KMS
828: acknowledges support from an NSF Graduate Research Fellowship and
829: would like to thank Franck Marchis and Jason Wright for helpful
830: discussions. The research of JEGP was is supported in part by NSF
831: grant AST04-06987. The National Radio Astronomy Observatory is a
832: facility of the National Science Foundation operated under cooperative
833: agreement by Associated Universities, Inc.
834:
835: {\it Facilities:} \facility{VLBA ()}
836:
837: %%%%%%%%%%%%%%
838: % REFERENCES %
839: %%%%%%%%%%%%%%
840:
841: %\bibliographystyle{apj}
842: %\bibliography{apj-jour,gmra_refs}
843: \begin{thebibliography}{45}
844: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
845:
846: \bibitem[{{Andre} {et~al.}(1992){Andre}, {Deeney}, {Phillips}, \&
847: {Lestrade}}]{andre92}
848: {Andre}, P., {Deeney}, B.~D., {Phillips}, R.~B., \& {Lestrade}, J.-F. 1992,
849: \apj, 401, 667
850:
851: \bibitem[{{Anthony-Twarog}(1982)}]{anthony-twarog82}
852: {Anthony-Twarog}, B.~J. 1982, \aj, 87, 1213
853:
854: \bibitem[{{Baumgardt} {et~al.}(2000){Baumgardt}, {Dettbarn}, \&
855: {Wielen}}]{baumgardt00}
856: {Baumgardt}, H., {Dettbarn}, C., \& {Wielen}, R. 2000, \aaps, 146, 251
857:
858: \bibitem[{{Bertout} {et~al.}(1999){Bertout}, {Robichon}, \&
859: {Arenou}}]{bertout99}
860: {Bertout}, C., {Robichon}, N., \& {Arenou}, F. 1999, \aap, 352, 574
861:
862: \bibitem[{{Bower} {et~al.}(2003){Bower}, {Plambeck}, {Bolatto}, {McCrady},
863: {Graham}, {de Pater}, {Liu}, \& {Baganoff}}]{bower03}
864: {Bower}, G.~C., {Plambeck}, R.~L., {Bolatto}, A., {McCrady}, N., {Graham},
865: J.~R., {de Pater}, I., {Liu}, M.~C., \& {Baganoff}, F.~K. 2003, \apj, 598,
866: 1140
867:
868: \bibitem[{{Brisken} {et~al.}(2002){Brisken}, {Benson}, {Goss}, \&
869: {Thorsett}}]{brisken02}
870: {Brisken}, W.~F., {Benson}, J.~M., {Goss}, W.~M., \& {Thorsett}, S.~E. 2002,
871: \apj, 571, 906
872:
873: \bibitem[{{Brown} {et~al.}(1994){Brown}, {de Geus}, \& {de Zeeuw}}]{brown94}
874: {Brown}, A.~G.~A., {de Geus}, E.~J., \& {de Zeeuw}, P.~T. 1994, \aap, 289, 101
875:
876: \bibitem[{{Chatterjee} {et~al.}(2004){Chatterjee}, {Cordes}, {Vlemmings},
877: {Arzoumanian}, {Goss}, \& {Lazio}}]{chatterjee04}
878: {Chatterjee}, S., {Cordes}, J.~M., {Vlemmings}, W.~H.~T., {Arzoumanian}, Z.,
879: {Goss}, W.~M., \& {Lazio}, T.~J.~W. 2004, \apj, 604, 339
880:
881: \bibitem[{{Cohen} \& {Kuhi}(1979)}]{cohen79}
882: {Cohen}, M. \& {Kuhi}, L.~V. 1979, \apjs, 41, 743
883:
884: \bibitem[{{Cordes} \& {Lazio}(2002)}]{cordes02}
885: {Cordes}, J.~M. \& {Lazio}, T.~J.~W. 2002, ArXiv Astrophysics e-prints
886:
887: \bibitem[{{Feigelson} {et~al.}(2002){Feigelson}, {Broos}, {Gaffney}, {Garmire},
888: {Hillenbrand}, {Pravdo}, {Townsley}, \& {Tsuboi}}]{feigelson02}
889: {Feigelson}, E.~D., {Broos}, P., {Gaffney}, III, J.~A., {Garmire}, G.,
890: {Hillenbrand}, L.~A., {Pravdo}, S.~H., {Townsley}, L., \& {Tsuboi}, Y. 2002,
891: \apj, 574, 258
892:
893: \bibitem[{{Felli} {et~al.}(1993{\natexlab{a}}){Felli}, {Churchwell}, {Wilson},
894: \& {Taylor}}]{felli93a}
895: {Felli}, M., {Churchwell}, E., {Wilson}, T.~L., \& {Taylor}, G.~B.
896: 1993{\natexlab{a}}, \aaps, 98, 137
897:
898: \bibitem[{{Felli} {et~al.}(1993{\natexlab{b}}){Felli}, {Taylor}, {Catarzi},
899: {Churchwell}, \& {Kurtz}}]{felli93b}
900: {Felli}, M., {Taylor}, G.~B., {Catarzi}, M., {Churchwell}, E., \& {Kurtz}, S.
901: 1993{\natexlab{b}}, \aaps, 101, 127
902:
903: \bibitem[{{Fomalont}(2005)}]{fomalont05}
904: {Fomalont}, E.~B. 2005, in ASP Conf. Ser. 340: Future Directions in High
905: Resolution Astronomy, ed. J.~{Romney} \& M.~{Reid}, 460--+
906:
907: \bibitem[{{Garay} {et~al.}(1987){Garay}, {Moran}, \& {Reid}}]{garay87}
908: {Garay}, G., {Moran}, J.~M., \& {Reid}, M.~J. 1987, \apj, 314, 535
909:
910: \bibitem[{{Genzel} {et~al.}(1981){Genzel}, {Reid}, {Moran}, \&
911: {Downes}}]{genzel81}
912: {Genzel}, R., {Reid}, M.~J., {Moran}, J.~M., \& {Downes}, D. 1981, \apj, 244,
913: 884
914:
915: \bibitem[{{Genzel} \& {Stutzki}(1989)}]{genzel89}
916: {Genzel}, R. \& {Stutzki}, J. 1989, \araa, 27, 41
917:
918: \bibitem[{{Getman} {et~al.}(2003){Getman}, {Feigelson}, {Garmire}, {Murray}, \&
919: {Harnden}}]{getman03}
920: {Getman}, K.~V., {Feigelson}, E.~D., {Garmire}, G., {Murray}, S.~S., \&
921: {Harnden}, Jr., F.~R. 2003, \iaucirc, 8068, 2
922:
923: \bibitem[{{Gomez} {et~al.}(2005){Gomez}, {Rodriguez}, {Loinard}, {Lizano},
924: {Poveda}, \& {Allen}}]{gomez05}
925: {Gomez}, L., {Rodriguez}, L.~F., {Loinard}, L., {Lizano}, S., {Poveda}, A., \&
926: {Allen}, C. 2005, \apj, 635, 1166
927:
928: \bibitem[{{Greenhill} {et~al.}(2005){Greenhill}, {Chandler}, {Reid}, \&
929: {Humphreys}}]{greenhill05}
930: {Greenhill}, L.~J., {Chandler}, C.~J., {Reid}, M.~J., \& {Humphreys}, E.~M.~L.
931: 2005, in Protostars and Planets V, 8614--+
932:
933: \bibitem[{{Hachisuka} {et~al.}(2006){Hachisuka}, {Brunthaler}, {Menten},
934: {Reid}, {Imai}, {Hagiwara}, {Miyoshi}, {Horiuchi}, \& {Sasao}}]{hachisuka06}
935: {Hachisuka}, K., {Brunthaler}, A., {Menten}, K.~M., {Reid}, M.~J., {Imai}, H.,
936: {Hagiwara}, Y., {Miyoshi}, M., {Horiuchi}, S., \& {Sasao}, T. 2006, \apj,
937: 645, 337
938:
939: \bibitem[{{Hartkopf} {et~al.}(1989){Hartkopf}, {McAlister}, \&
940: {Franz}}]{hartkopf89}
941: {Hartkopf}, W.~I., {McAlister}, H.~A., \& {Franz}, O.~G. 1989, \aj, 98, 1014
942:
943: \bibitem[{{Hartmann}(2003)}]{hartmann03}
944: {Hartmann}, L. 2003, \apj, 585, 398
945:
946: \bibitem[{{Hartmann} {et~al.}(2001){Hartmann}, {Ballesteros-Paredes}, \&
947: {Bergin}}]{hartmann01}
948: {Hartmann}, L., {Ballesteros-Paredes}, J., \& {Bergin}, E.~A. 2001, \apj, 562,
949: 852
950:
951: \bibitem[{{Hillenbrand} \& {Hartmann}(1998)}]{hillenbrand98}
952: {Hillenbrand}, L.~A. \& {Hartmann}, L.~W. 1998, \apj, 492, 540
953:
954: \bibitem[{{Hirota} {et~al.}(2007){Hirota}, {Bushimata}, {Choi}, {Honma},
955: {Imai}, {Iwadate}, {Jike}, {Kameno}, {Kameya}, {Kamohara}, {Kan-ya},
956: {Kawaguchi}, {Kijima}, {Kim}, {Kobayashi}, {Kuji}, {Kurayama},
957: {Manabe}, {Maruyama}, {Matsui}, {Matsumoto}, {Miyaji}, {Nagayama},
958: {Nakagawa}, {Nakamura}, {Oh}, {Omodaka}, {Oyama}, {Sakai}, {Sasao},
959: {Sato}, {Sato}, {Shibata}, {Shintani}, {Tamura}, {Tsushima}, \&
960: {Yamashita}}]{hirota07}
961: {Hirota}, T., {Bushimata}, T., {Choi}, Y.~K., {Honma}, M., {Imai}, H.,
962: {Iwadate}, K., {Jike}, T., {Kameno}, S., {Kameya}, O., {Kamohara}, R.,
963: {Kan-ya}, Y., {Kawaguchi}, N., {Kijima}, M., {Kim}, M.~K., {Kobayashi},
964: H., {Kuji}, S., {Kurayama}, T., {Manabe}, S., {Maruyama}, K., {Matsui},
965: M., {Matsumoto}, N., {Miyaji}, T., {Nagayama}, T., {Nakagawa}, A.,
966: {Nakamura}, K., {Oh}, C.~S., {Omodaka}, T., {Oyama}, T., {Sakai}, S.,
967: {Sasao}, T., {Sato}, K., {Sato}, M., {Shibata}, K.~M., {Shintani}, M.,
968: {Tamura}, Y., {Tsushima}, M., \& {Yamashita}, K. 2007, ArXiv e-prints,
969: 705
970:
971: \bibitem[{{Huff} \& {Stahler}(2006)}]{huff06}
972: {Huff}, E.~M. \& {Stahler}, S.~W. 2006, \apj, 644, 355
973:
974: \bibitem[{{Jeffries}(2007)}]{jeffries07}
975: {Jeffries}, R.~D. 2007, ArXiv Astrophysics e-prints
976:
977: \bibitem[{{Johnson}(1965)}]{johnson65}
978: {Johnson}, H.~M. 1965, \apj, 142, 964
979:
980: \bibitem[{{Jones} \& {Walker}(1988)}]{jones88}
981: {Jones}, B.~F. \& {Walker}, M.~F. 1988, \aj, 95, 1755
982:
983: \bibitem[{{Kharchenko} {et~al.}(2003){Kharchenko}, {Pakulyak}, \&
984: {Piskunov}}]{kharchenko03}
985: {Kharchenko}, N.~V., {Pakulyak}, L.~K., \& {Piskunov}, A.~E. 2003, VizieR
986: Online Data Catalog, 808, 291
987:
988: \bibitem[{{Kharchenko} {et~al.}(2005){Kharchenko}, {Piskunov}, {R{\"o}ser},
989: {Schilbach}, \& {Scholz}}]{kharchenko05}
990: {Kharchenko}, N.~V., {Piskunov}, A.~E., {R{\"o}ser}, S., {Schilbach}, E., \&
991: {Scholz}, R.-D. 2005, \aap, 438, 1163
992:
993: \bibitem[{{Kraus} {et~al.}(2007){Kraus}, {Balega}, {Berger}, {Hofmann},
994: {Millan-Gabet}, {Monnier}, {Ohnaka}, {Pedretti}, {Preibisch}, {Schertl},
995: {Schloerb}, {Traub}, \& {Weigelt}}]{kraus07}
996: {Kraus}, S., {Balega}, Y.~Y., {Berger}, J.~., {Hofmann}, K.~., {Millan-Gabet},
997: R., {Monnier}, J.~D., {Ohnaka}, K., {Pedretti}, E., {Preibisch}, T.,
998: {Schertl}, D., {Schloerb}, F.~P., {Traub}, W.~A., \& {Weigelt}, G. 2007,
999: ArXiv Astrophysics e-prints
1000:
1001: \bibitem[{{Loinard} {et~al.}(2005){Loinard}, {Mioduszewski},
1002: {Rodr{\'{\i}}guez}, {Gonz{\'a}lez}, {Rodr{\'{\i}}guez}, \&
1003: {Torres}}]{loinard05}
1004: {Loinard}, L., {Mioduszewski}, A.~J., {Rodr{\'{\i}}guez}, L.~F.,
1005: {Gonz{\'a}lez}, R.~A., {Rodr{\'{\i}}guez}, M.~I., \& {Torres}, R.~M. 2005,
1006: \apjl, 619, L179
1007:
1008: \bibitem[{{Palla} \& {Stahler}(1999)}]{palla99}
1009: {Palla}, F. \& {Stahler}, S.~W. 1999, \apj, 525, 772
1010:
1011: \bibitem[{{Palla} \& {Stahler}(2000)}]{palla00}
1012: ---. 2000, \apj, 540, 255
1013:
1014: \bibitem[{{Penston}(1973)}]{penston73}
1015: {Penston}, M.~V. 1973, \apj, 183, 505
1016:
1017: \bibitem[{{Penston} {et~al.}(1975){Penston}, {Hunter}, \& {Oneill}}]{penston75}
1018: {Penston}, M.~V., {Hunter}, J.~K., \& {Oneill}, A. 1975, \mnras, 171, 219
1019:
1020: \bibitem[{{Phillips} {et~al.}(1991){Phillips}, {Lonsdale}, \&
1021: {Feigelson}}]{phillips91}
1022: {Phillips}, R.~B., {Lonsdale}, C.~J., \& {Feigelson}, E.~D. 1991, \apj, 382,
1023: 261
1024:
1025: \bibitem[{{Sicilia-Aguilar} {et~al.}(2005){Sicilia-Aguilar}, {Hartmann},
1026: {Szentgyorgyi}, {Fabricant}, {F{\H u}r{\'e}sz}, {Roll}, {Conroy}, {Calvet},
1027: {Tokarz}, \& {Hern{\'a}ndez}}]{sicilia-aguilar05}
1028: {Sicilia-Aguilar}, A., {Hartmann}, L.~W., {Szentgyorgyi}, A.~H., {Fabricant},
1029: D.~G., {F{\H u}r{\'e}sz}, G., {Roll}, J., {Conroy}, M.~A., {Calvet}, N.,
1030: {Tokarz}, S., \& {Hern{\'a}ndez}, J. 2005, \aj, 129, 363
1031:
1032: \bibitem[{{Stassun} {et~al.}(2004){Stassun}, {Mathieu}, {Vaz}, {Stroud}, \&
1033: {Vrba}}]{stassun04}
1034: {Stassun}, K.~G., {Mathieu}, R.~D., {Vaz}, L.~P.~R., {Stroud}, N., \& {Vrba},
1035: F.~J. 2004, \apjs, 151, 357
1036:
1037: \bibitem[{{Strand}(1958)}]{strand58}
1038: {Strand}, K.~A. 1958, \apj, 128, 14
1039:
1040: \bibitem[{{Tan} {et~al.}(2006){Tan}, {Krumholz}, \& {McKee}}]{tan06}
1041: {Tan}, J.~C., {Krumholz}, M.~R., \& {McKee}, C.~F. 2006, \apjl, 641, L121
1042:
1043: \bibitem[{{Warren} \& {Hesser}(1978)}]{warren78}
1044: {Warren}, Jr., W.~H. \& {Hesser}, J.~E. 1978, \apjs, 36, 497
1045:
1046: \bibitem[{{Wilson} {et~al.}(2005){Wilson}, {Dame}, {Masheder}, \&
1047: {Thaddeus}}]{wilson05}
1048: {Wilson}, B.~A., {Dame}, T.~M., {Masheder}, M.~R.~W., \& {Thaddeus}, P. 2005,
1049: \aap, 430, 523
1050:
1051: \bibitem[{{Zapata} {et~al.}(2004){Zapata}, {Rodr{\'{\i}}guez}, {Kurtz}, \&
1052: {O'Dell}}]{zapata04}
1053: {Zapata}, L.~A., {Rodr{\'{\i}}guez}, L.~F., {Kurtz}, S.~E., \& {O'Dell}, C.~R.
1054: 2004, \aj, 127, 2252
1055:
1056: \end{thebibliography}
1057:
1058: \clearpage
1059:
1060: %%%%%%%%%%
1061: % TABLES %
1062: %%%%%%%%%%
1063:
1064: %\input{tab1.tex}
1065:
1066: %\input{tab2.tex}
1067:
1068: %\input{tab3.tex}
1069:
1070: %\input{tab4.tex}
1071:
1072: %\input{tab5.tex}
1073:
1074: %%%%%%%%%%%
1075: % FIGURES %
1076: %%%%%%%%%%%
1077:
1078: %\begin{figure}
1079: %\centering
1080: %\epsscale{1.0}
1081: %\plotone{f1.eps}
1082: %\caption{Images of GMR A for all epochs. All images are after ATMCA
1083: %calibration, except for the January 2003 image, as described in the
1084: %text. Each image is centered on the brightest pixel of that image.
1085: %The contour levels are -2.5, 2.5, 4, 7, 10, 15 and 20 times the noise
1086: %level of the individual image. The dotted contours are the negative
1087: %values. The noise levels are listed in Table~\ref{tab:gmra}. The
1088: %sythesized beam is shown in the lower left corner.}
1089: %\label{fig:gmra}
1090: %\end{figure}
1091:
1092: %\begin{figure}
1093: %\centering
1094: %\epsscale{1.0}
1095: %\plotone{f2.eps}
1096: %\caption{Images of the primary calibrator at each epoch
1097: %with ATMCA calibration, except for the January 2003 observation which
1098: %does not have ATMCA calibration but has been through an amplitude and
1099: %phase self-calibration cycle. The images are centered on the location
1100: %of the brightest pixel in the December 2003 image. The contour levels
1101: %are -5, 5, 20, 50, 100, 200, 500 times the noise level of the
1102: %individual image. The synthesized beam is shown in the lower left of
1103: %each image. Negative contours are shown with dotted lines. In the
1104: %January 2003 image there are no pixels less than -5 times the noise level.}
1105: %\label{fig:j0541}
1106: %\end{figure}
1107:
1108: %\begin{figure}
1109: %\centering
1110: %\epsscale{1.0}
1111: %\plotone{f3.eps}
1112: %\caption{Images of the secondary calibrator, with ATMCA calibration.
1113: %The images are centered on the location of the brightest pixel in the
1114: %December 2003 image. The contours are -5, 5, 10, 20, 40 and 60 times the
1115: %noise level of the individual image, The synthesized beam is
1116: %shown in the lower left of each image. Negative contours are shown
1117: %with dotted lines.}
1118: %\label{fig:j0529}
1119: %\end{figure}
1120:
1121: %\begin{figure}
1122: %\centering
1123: %\epsscale{0.9}
1124: %\plotone{f4.eps}
1125: %\caption{The measured positions of GMR A with the best fit parallax
1126: %and proper motion. The diamonds represent the predicted
1127: %position of GMR A for each observation. The error bars on the
1128: %measured positions are scaled as described in the text. The dashed
1129: %line is the proper motion of the source, with the parallactic motion
1130: %subtracted. The parallax corresponds to a distance of
1131: %$389^{+24}_{-21}$ pc.}
1132: %\label{fig:solution}
1133: %\end{figure}
1134:
1135: \end{document}
1136: