1: \documentclass[apj]{emulateapj}
2:
3: %\usepackage[]{natbib}
4: \bibpunct{(}{)}{;}{a}{}{,}
5:
6: \slugcomment{in press, ApJL}
7: \shorttitle{HD 12661 Unexpected Evolution}
8: \shortauthors{Veras and Ford}
9:
10: \begin{document}
11: %
12: \title{Secular Evolution of HD 12661: A System Caught at an Unlikely Time}
13: \author{Dimitri Veras\altaffilmark{1}, Eric B. Ford\altaffilmark{1}}
14: \altaffiltext{1}{Astronomy Department, University of Florida, 211 Bryant Space Sciences Center, Gainesville, FL 32111, USA}
15: %
16: \email{veras@astro.ufl.edu}
17: %
18: \begin{abstract}
19: %
20: The eccentricity evolution of multiple planet systems can
21: provide valuable constraints on planet formation models.
22: Unfortunately, the inevitable uncertainties in the current orbital
23: elements can lead to significant ambiguities in the nature of the
24: secular evolution. Integrating any single set of orbital elements inadequately
25: describes the full range of secular evolutions
26: consistent with current observations. Thus, we combine radial velocity
27: observations of HD 12661 with Markov Chain Monte Carlo sampling
28: to generate ensembles of initial conditions for direct
29: n-body integrations. We find that any mean motion resonances
30: are quite weak and do not significantly impact the
31: secular evolution, and that current observations indicate
32: circulation or large amplitude libration of the periapses.
33: The eccentricity of the outer planet
34: undergoes large oscillations for nearly all of the allowed
35: two-planet orbital solutions. This type of secular evolution would
36: arise if planet c had been impulsively perturbed, perhaps due to
37: strong scattering of an additional planet that was subsequently
38: accreted onto the star. Finally, we note that the secular evolution
39: implied by the current orbital configuration implies that planet c
40: spends $\sim~{96}\%$ of the time following an orbit more eccentric
41: than that presently observed. Either this system is being observed
42: during a relatively rare state, or additional planets are affecting
43: the observed radial velocities and/or the system's secular
44: eccentricity evolution.
45: %
46: %
47: \end{abstract}
48:
49:
50: \keywords{celestial mechanics --- stars: individual (HD 12661) --- planetary systems: formation --- methods: n-body simulations, statistical}
51:
52: \section{Introduction}
53: %
54: The secular evolution of multi-planet systems has become a topic of
55: considerable importance for constraining planetary formation theories
56: \citep{foretal2005,adalau2006,bargre2006b,sanetal2007}. We
57: investigate the secular evolution of the two giant planets orbiting HD
58: 12661, a $\simeq~1.136 M_\odot$ star. Planet HD 12661 b has a
59: semimajor axis of $\simeq 0.83$AU and a
60: moderate eccentricity, and planet HD 12661 c
61: follows a nearly circular orbit about three times
62: further away \citep{fisetal2003}. This system has inspired several
63: dynamical studies.
64: %
65: A brief survey of their results demonstrates the importance of
66: considering the uncertainty in the current orbital configuration.
67: \cite{kisetal2002} considered coplanar edge-on systems and found that
68: the then-current best-fit solution was chaotic, behavior that is
69: likely due to close proximity to the 9:2 mean-motion resonance (MMR).
70: Both \cite{gozdziewski2003} and \cite{leepea2003} found that the
71: system was close to the 11:2 MMR and that the periapses underwent
72: large amplitude libration for the edge-on case, as well as for a broad
73: range of inclinations. \cite{zhosun2003} claimed that aligned
74: configurations were less likely to be chaotic and thus more likely to
75: be stable. \cite{gozdziewski2003} found the system could be chaotic,
76: but still stable for $\sim$Gyr thanks to a secular apsidal lock about
77: an antialigned configuration. \cite{gozmac2003} performed an
78: independent analysis of the \cite{fisetal2003} radial velocities to
79: find a better orbital fit that placed the system very near (and likely
80: in) the 6:1 MMR that could be stabilized by secular apsidal lock for a
81: broad range of inclinations. Despite the different orbital solutions,
82: both \cite{leepea2003} and \cite{gozmac2003} found that the periapses
83: were more likely to librate about an antialigned configuration, but
84: that librations about an aligned configuration were also possible,
85: particularly for systems with large inclinations. However,
86: \cite{jietal2003} reported that aligned and antialigned configurations
87: were nearly equally likely. Subsequently, \cite{butetal2006}
88: published an orbital solution with a ratio of orbital periods
89: approaching 13:2.
90: %
91: \cite{bargre2006a} found that this orbital solution places the
92: system very near the the boundary between librating and
93: circulating modes of secular evolution.
94: %
95: The revised orbital solutions for HD 12661 b \& c and the lack of
96: consensus regarding the system's secular evolution both motivate an
97: updated dynamical study. Further, the historical range of orbital
98: configurations illustrates the importance of properly accounting for
99: uncertainties in orbital determinations.
100:
101: \cite{gozdziewski2003} found that the classical secular theory gave
102: only a crude approximation to the secular evolution due to large
103: eccentricities and the 11:2 MMR. Both \cite{verarm2007} and
104: \cite{libhen2007} caution against using low-order secular theories to
105: model the system's behavior. \cite{leepea2003},
106: \cite{rodgal2005}, and \cite{libhen2007} found that either the
107: octupole or high-order Laplace-Lagrange approximation described the
108: secular evolution of HD 12661 well and that the secular dynamics is
109: not significantly affected by the 5:1, 11:2, and 6:1 MMRs. Although
110: we do not expect near-resonant terms to significantly affect the
111: secular evolution, we use direct n-body integrations to
112: % be sure we
113: account for the full dynamics, including all possible MMRs.
114:
115: We determine the mode(s) of secular evolution
116: consistent with observations and the requirement of long-term
117: stability. Our study improves upon previous studies by combining a
118: Bayesian analysis of an updated set of radial velocity observations
119: with direct n-body integrations to examine HD 12661's stability and
120: secular evolution. We characterize the full range of orbital
121: histories that are consistent with observations and reject solutions
122: that do not exhibit long-term orbital stability. Thus, we can make
123: quantitative statements about the relative probability of different
124: modes of secular evolution.
125:
126:
127:
128: \section{Methods}
129:
130:
131: We reanalyze an updated set of radial velocity observations
132: (Wright et al.\ 2008) assuming that the stellar velocity is
133: the superposition of two
134: Keplerian orbit plus noise. First, we perform a brute force search
135: over parameter space to verify that there are no qualitatively different and
136: comparably likely solutions. Working in a Bayesian framework, we
137: generate an ensemble of $\simeq~5\times10^5$ orbital solutions using
138: Markov Chain Monte Carlo (MCMC;
139: \citealt{ford2005,ford2006,gregory2007a,gregory2007b}). Each state of
140: the Markov chain includes the orbital period ($P$), velocity amplitude
141: ($K$), eccentricity ($e$), argument of pericenter measured from the
142: plane of the sky ($\omega$), and mean anomaly at a given epoch ($u$)
143: for each planet. We use the priors, candidate transition functions,
144: automated step size control, and convergence tests described in
145: \cite{ford2006}. In particular, we assume a prior for the velocity ``jitter'' ($\sigma_j$) of $p(\sigma_j) \propto [1+\sigma_j/(1 {\mathrm m s^{-1}})]^{-1}$.
146:
147: We randomly select subsamples of orbital solutions to
148: be investigated with long-term n-body integrations and consider a
149: range of possible inclinations ($i$) and ascending nodes ($\Omega$).
150: We generate each orbit from an isotropic distribution (i.e., uniform
151: in $\cos{i}$ and $\Omega$) and divide the resulting systems into bins
152: based on the relative inclination between the two orbital planes
153: ($i_{\rm rel}$). The bins are
154: %
155: 1) $i_{\rm rel}=0^{\circ}$ (coplanar),
156: 2) $0^{\circ} \le i_{\rm rel} \le 30^{\circ}$,
157: 3) $30^{\circ} \le i_{\rm rel} \le 60^{\circ}$,
158: 4) $60^{\circ} \le i_{\rm rel} \le 90^{\circ}$,
159: 5) $90^{\circ} \le i_{\rm rel} \le 120^{\circ}$,
160: 6) $120^{\circ} \le i_{\rm rel} \le 150^{\circ}$, and
161: 7) $150^{\circ} \le i_{\rm rel} \le 180^{\circ}$.
162: %
163: We integrate 2,000 systems in each of the first four bins and 500 in
164: each of the last three bins. We can reweight our simulations so as to
165: determine the probability of the different modes of secular evolution
166: for various assumed distributions of relative inclinations.
167: %
168: From each set ($P,K,e,\omega,i,\Omega,u$), we generate the planet mass
169: ($m$) and semi-major axis ($a$) using a Jacobi coordinate system
170: \citep{leepea2003}.
171: %
172:
173: We used the hybrid symplectic integrator of {\tt Mercury}
174: \citep{chambers1999} to integrate each set of initial conditions for
175: at least 1 Myr. Based on a smaller series of 10 Myr integrations, we
176: found that the vast majority of instabilities are manifest within 1
177: Myr. We classified systems as ``unstable'' if, for either planet,
178: $a_{\rm max}-a_{\rm min} > \tau a_i$, where $a_{\rm max}$ and $a_{\rm
179: min}$ represent the maximum and minimum values of the semimajor axis,
180: $a_i$ is the initial value of the semimajor axis, and $\tau=0.3$. We
181: discarded each set of initial conditions that was found to be
182: unstable, and analyzed the properties of the remaining ``stable''
183: systems. Although a small fraction of our ``stable'' systems might
184: exhibit instability if integrated for much longer timescales, our
185: criteria avoids miscategorizing a system exhibiting bounded chaos
186: (e.g., \citealt{gozdziewski2003}) as unstable. We manually verified
187: that the above criteria gives reasonable results and that various
188: choices of $\tau\in[0.10,0.30]$ make just a few percent difference in
189: the number of simulations labeled as stable. The percent of stable
190: systems in our 7 bins are: 1) 100\%, 2) 99.7\%, 3) 93.3\%, 4)
191: 16.4\%, 5) 0.2\%, 6) 16.7\%, and 7) 98.6\%. We independently affirmed
192: the trend exhibited by these stability percentages by performing a
193: smaller, additional set of simulations by using an ensemble of initial
194: conditions generated from a different MCMC code that accounted for planet-planet
195: interactions with a Hermite integrator.
196:
197:
198:
199: \section{Results}
200:
201:
202:
203: Several previous studies of HD 12661
204: (e.g. \citealt{jietal2003,leepea2003,zhosun2003}) and other planetary
205: systems \citep{chietal2001,malhotra2002,foretal2005,bargre2006b} have
206: highlighted the importance of the apsidal angle
207: ($\Delta\varpi=\Omega_b+\omega_b-\Omega_c-\omega_c$), which is useful
208: for describing the secular dynamics of systems with small relative
209: inclinations.
210: %
211: Since we consider systems with a wide range of relative inclinations,
212: we instead focus on $\Delta\varpi'$, the angle between the periastron
213: directions projected onto the invariable plane. Typically, the
214: apsidal angle is classified as circulating, librating about $0^\circ$,
215: or librating about $180^\circ$. By inspecting plots of apsidal angle
216: evolution for individual systems, we find that some systems spend most
217: of the time librating about one center, but occasionally the apsidal
218: angle circulates for a short period of time.
219: Therefore, we consider two summary statistics describing the secular
220: evolution: the root mean square (RMS) of $\Delta\varpi'$ and the mean
221: absolute deviation (MAD) of $\Delta\varpi'$ about the libration
222: center, where the ``libration center'' is defined as the angle about which
223: the RMS $\Delta\varpi'$ is minimized.
224: %
225: % Note that this definition does not strictly require the apsidal angle to pass through the value of the center.
226: %
227: For
228: a system undergoing small amplitude libration about either center, the
229: MAD and RMS deviation are of order the libration amplitude. For a
230: system undergoing uniform circulation, the MAD approaches 90$^{\circ}$ and the
231: RMS deviation approaches 103.92$^{\circ}$. These
232: statistics can be used to identify the mode of secular
233: evolution in clear cut cases and provide a quantitative measure that is well-defined even for
234: systems with complex evolution.
235:
236: We calculate both measures using orbital elements measured every $1000$ yr.
237: Although the libration amplitude (maximum absolute
238: deviation of $\Delta\varpi'$) can be sensitive to the frequency of
239: sampling and leads to ambiguities, both RMS $\Delta\varpi'$ and MAD
240: $\Delta\varpi'$ are more robust statistics that allows us to describe
241: the secular evolution with a simple quantitative measure, regardless
242: of whether the system is librating, circulating, or switching between
243: regimes due to short-term perturbations.
244: %
245: %This method may also be expanded to any other secular or resonant angle.
246: %
247: This robustness is particularly important for studying systems where
248: there is little distinction between the librating and circulating
249: regimes, due to one planet's orbit periodically becoming nearly
250: circular.
251:
252:
253:
254: For each n-body integration,
255: we determine the libration center and calculate both the RMS and MAD
256: of $\Delta\varpi'$ about that center.
257: %
258: We discover that the RMS $\Delta\varpi'$ ranges from
259: $68^{\circ}-85^{\circ}$ and MAD $\Delta\varpi'$ ranges from
260: $63^{\circ}-82^{\circ}$, depending on the relative inclination
261: (see Table \ref{TabLib}). The
262: libration centers are preferentially anti-aligned for
263: prograde, nearly-coplanar systems, and transition to
264: almost entirely aligned for near-coplanar retrograde systems.
265: %
266: The stability is highly dependent on initial relative
267: inclination; very few highly inclined retrograde systems were
268: stable, indicating that non-secular perturbations influence
269: the long-term dynamics for some relative inclinations.
270: Although large values of RMS $\Delta\varpi'$ and MAD $\Delta\varpi'$
271: may be consistent with non-uniform circulation,
272: visual spot checks of individual systems all suggest libration.
273: %
274: %We present aggregated results
275: %
276: %in Table \ref{TabLib} and Fig.\ \ref{FigApsial}.
277:
278:
279: %with libration
280: %centers about $0^{\circ}$ (``aligned'') or $180^{\circ}$
281: %(``anti-aligned'')
282: %subdivided according to relative inclination. As $i_{\rm rel}$
283: %increases, there is a systematic shift towards alignment of the
284: %apsidal angle. For high relative inclinations, the very large
285: %libration amplitudes indicate a circulating apsidal angle.
286: %
287:
288: %
289: %Figure \ref{FigApsial} plots the RMS and MAD of the apsidal angle and
290: %their standard deviations for the stable systems as a function of the
291: %planets' mean $\sin i$. Each panel reports results for systems with a
292: %given range of relative inclination. Data are shown only for bins of
293: %relative and mean inclination which contain at least 3 stable systems.
294: %The plots demonstrate that the allowed systems have large libration
295: %amplitudes. While the libration amplitude is best constrained for low
296: %values of $i_{rel}$, the libration amplitude is largely independent of
297: %both $\sin i$ and $i_{rel}$.
298:
299: If a system of two planets on nearly circular orbits is excited
300: impulsively, then the secular evolution will follow the boundary
301: between the circulating and librating regimes and cause one orbit to
302: repeatedly return to a nearly circular orbit \citep{malhotra2002}. In
303: some systems, this secular evolution can be used to constrain the
304: system's formation \citep{foretal2005}. \cite{bargre2006b}
305: %
306: found that the \cite{butetal2006} orbital solution implies the
307: eccentricity of HD 12661 c comes particularly close to zero, as
308: parameterized by their $\epsilon$ parameter (Eq. 1 of
309: \citealt{bargre2006b}). Because they considered only the published
310: orbital solution in an edge-on, coplanar orientation, they were unable
311: to asses the finding's robustness to uncertainties in the orbital
312: parameters or non-edge-on systems.
313: %
314: In order to determine what fraction of stable systems consistent with
315: observations result in one planet returning to a nearly circular
316: orbit, we calculate four statistics from each of our n-body
317: integrations. The first two are $c_1 \equiv e_{min,c}/e_{max,c}$ and
318: $c_2 \equiv e_{min,b}/e_{max,b}$, which represent the ratio of the
319: minimum to maximum eccentricity for each planet. The
320: third ($c_3$) is equal to
321: $\left[ 2 {\rm min} (e_b e_c) \right]/
322: (x_{max} - x_{min} + y_{max} - y_{min})$, as defined by
323: \cite{bargre2006b}, where $x \equiv e_b e_c
324: \sin\left(\Delta\varpi\right)$ and $y \equiv e_b e_c
325: \cos\left(\Delta\varpi\right)$.
326: %
327: The fourth is
328: %
329: $c_4 \equiv {\rm min}(e_b e_c) /
330: {\rm max}(e_b e_c)$,
331: which is a similar measure that includes the
332: eccentricities of both planets in a
333: rotationally symmetric manner that
334: is independent of the systems' orientation (unlike $c_3$).
335:
336:
337: %########################################################################
338: \begin{figure}
339: \epsscale{0.7}
340: \plotone{f1.eps}
341: %\centerline{\psfig{figure=figures/distributions.eps,width=3.25truein,height=3.25truein}}
342: \figcaption{Values of $c_1$ (dots), $c_2$ (triangles), $c_3$
343: (squares), and $c_4$ (crosses) and their standard deviations,
344: for coplanar systems, binned according to the sine of the inclination of the invariable plane.
345: %mass-weighted mean of the $\sin i$ factors.
346: \label{FigCvsIlos}
347: \label{schem2}}
348: \end{figure}
349: %########################################################################
350:
351:
352: %########################################################################
353: \begin{figure}
354: \epsscale{0.7}
355: \plotone{f2.eps}
356: %\centerline{\psfig{figure=figures/exceed.eps,width=3.25truein,height=3.25truein}}
357: \figcaption{Fraction of time for which each planet's eccentricity exceeded
358: a given threshold (x-axis) averaged over allowed orbital solutions.
359: The curves with dots show results for the coplanar case, but assuming an
360: isotropic distribution of inclinations relative to the plane of the sky. The dashed vertical
361: lines indicate the median eccentricity values (0.031 and 0.361) for the planets'
362: initial eccentricities, and their 5\%-95\% uncertainty ranges are
363: indicated by the horizontal solid segments. HD 12661 c's orbital eccentricity
364: exceeds its current value for approximately 96\% of its evolution.
365: \label{FigEHisto}
366: \label{schem1}}
367: \end{figure}
368: %########################################################################
369:
370:
371: Figure \ref{FigCvsIlos} plots the mean and standard deviation of each
372: measure ($c_1$ - dots, $c_2$ - triangles, $c_3$ - squares, $c_4$ -
373: crosses) for coplanar systems which have been binned according to the
374: sine of the inclination of the invariable plane.
375: The values
376: $c_1$, $c_3$ and $c_4$ all maintain values $< 0.1$ for all
377: bins of relative inclination for prograde systems, and
378: $< 0.15$ for retrograde systems. The large values of
379: $c_2$ (triangles) imply that HD 12661 b maintains a moderately
380: eccentric orbit throughout the planet's evolution.
381: Regardless, the measure $c_4$
382: demonstrates that the system lies near the boundary of libration and
383: circulation. For the vast majority of systems, $c_3$ is over one
384: order of magnitude greater than 0.003 (as found by \citealt{bargre2006b}).
385: %
386:
387:
388: The finding that planet c undergoes large eccentricity oscillations is
389: remarkable since it currently has an eccentricity of only
390: $\simeq~0.02$. We investigate this observation further by plotting the fraction
391: of simulation time that each planet's eccentricity exceeded a given
392: value for the coplanar systems (Fig.\
393: \ref{FigEHisto}). The dotted vertical lines indicate representative
394: values for the planets' initial eccentricities.
395: HD 12661 c's orbital eccentricity exceeds its current
396: value for approximately 96\% of its evolution. Thus, the
397: near-circular orbit we now observe is a rare occurrence.
398:
399: \section{Conclusion}
400: %
401: Several previous studies had suggested HD 12661 b \& c are in a
402: secular apsidal lock and are undergoing large amplitude librations.
403: However, previously, even the orbital period of HD 12661 c was poorly
404: determined. Now that radial velocity observations span over two
405: orbital periods of HD 12661 c, the current orbital period of planet c is
406: well-constrained. An exhaustive search for
407: libration of all possible 11:2, 6:1, and 13:2 resonant angles
408: found that in only a few percent of all simulations performed, large
409: amplitude libration lasts over several $10^5$
410: yr. The current orbital elements of the two
411: giant planets imply that the system lies near the boundary of
412: circulating and librating regimes, regardless of the unknown orbital
413: inclinations. This behavior causes the eccentricity of planet c to
414: spend most of its time with a significant eccentricity ($e_c\ge~0.2$). By computing the projected apsidal angle on the invariable plane, we find that the RMS $\Delta\varpi'$ ranges from $68^{\circ}-85^{\circ}$ and MAD $\Delta\varpi'$ ranges from $63^{\circ}- 82^{\circ}$, for any stable initial relative inclination,
415: and planet c's eccentricity would be
416: greater than its current value $96\%$ of the time.
417:
418: The most straightforward interpretation is that the eccentricity of
419: HD 12661 b was excited by an impulsive perturbation and planet c's
420: eccentricity is periodically excited via secular perturbations,
421: similar to the history of the $\upsilon$ And c \& d system
422: \citep{malhotra2002,foretal2005}. Such an impulse could be the result
423: of planet-planet scattering involving at least one additional planet
424: \citep[e.g.][]{rasfor1996,weimar1996}. The
425: current eccentricity of planet b suggests that the putative scattered
426: planet would have had a mass roughly half that of planet b
427: \citep{forras2008}. If the additional planet had been scattered outwards,
428: then there is a significant chance that it would have altered the
429: secular evolution of planet c \citep{bargre2006b}. Given the
430: semi-major axis of planet b, it is possible that the scattered planet
431: was accreted onto the star. If the instability occurred after the star
432: had reached the main sequence, then this instability could contribute to the
433: star's large surface high metalicity. One potential concern for this
434: mechanism is whether the timescale for the apoastron of the scattered
435: planet to be lowered would be significantly shorter than the
436: $\simeq~10^4-10^5 $yr timescale for secular
437: evolution of planets b \& c.
438:
439: There is an alternative interpretation of our results. One might
440: think that we are unlikely to observe a system at such a rare time,
441: casting doubt upon the orbital fit and thus the inferred secular evolution.
442: One would expect that one of the 30 currently observed
443: multiple planet systems is transiently passing through such
444: an unusual state. However, there are several ways in which
445: a system could be ``unusual'' so that the standard caveats of
446: {\em a posteriori} statistics apply.
447: Given the current radial velocity observations, there
448: are few alternatives. We consider it unlikely that our global search
449: missed a qualitatively different pair of orbits due to our assumption
450: of non-interacting Keplerian orbits. More plausibly, one or more
451: undetected planets could be affecting the orbital solution and/or the
452: secular evolution of the system. After accounting for planets b \& c,
453: we find no statistically significant periodicities in the current
454: radial velocity observations. Current observations are consistent
455: with a small long-term acceleration, but zero acceleration falls
456: within the 95\% credible interval. Further, models with a long-term
457: acceleration do not result in significantly different orbital
458: parameters. Our Bayesian analysis assuming two planets and
459: uncorrelated Gaussian noise constrains the jitter to be
460: $\simeq~3.4\pm~0.7$m/s. Further radial velocity observations can test
461: whether a portion of this jitter is due to additional planets.
462:
463:
464:
465:
466: \acknowledgments{We thank the referee for insightful comments, and
467: Debra Fischer and Geoff Marcy for providing
468: updated radial velocity observations prior to publication. This
469: research was supported by NASA/JPL RSA1326409. Some of the data
470: analyzed were obtained at the W. M. Keck Observatory, which is
471: operated as a scientific partnership among the California Institute of
472: Technology, the University of California, and the National Aeronautics
473: and Space Administration. The Observatory was made possible by the
474: generous financial support of the W. M. Keck Foundation.
475: %
476: We acknowledge the University of Florida High-Performance Computing
477: Center for providing computational resources and support.
478: %
479: }
480: %
481: \begin{thebibliography}{}
482: %
483:
484: \bibitem[Adams \& Laughlin(2006)]{adalau2006}
485: Adams, F.~C., \& Laughlin, G.\ 2006, \apj, 649, 1004
486:
487: \bibitem[Barnes \& Greenberg(2006a)]{bargre2006a}
488: Barnes, R., \& Greenberg, R.\ 2006a, \apjl, 647, L163
489:
490: \bibitem[Barnes \& Greenberg(2006b)]{bargre2006b}
491: Barnes, R., \& Greenberg, R.\ 2006b, \apjl, 652, L53
492:
493: \bibitem[Butler et al.(2006)]{butetal2006}
494: Butler, R.~P., et al.\ 2006, \apj, 646, 505
495:
496: \bibitem[Chambers(1999)]{chambers1999}
497: Chambers, J.~E.\ 1999, \mnras, 304, 793
498:
499: \bibitem[Chiang et al.(2001)]{chietal2001}
500: Chiang, E.~I., Tabachnik, S., \& Tremaine, S.\ 2001, \aj, 122, 1607
501:
502: \bibitem[Fischer et al.(2003)]{fisetal2003}
503: Fischer, D.~A., et al.\ 2003, \apj, 586, 1394
504:
505: \bibitem[Ford(2005)]{ford2005}
506: Ford, E.~B.\ 2005, \aj, 129, 1706
507:
508: \bibitem[Ford(2006)]{ford2006}
509: Ford, E.~B.\ 2006, \apj, 642, 505
510:
511: \bibitem[Ford et al.(2005)]{foretal2005}
512: Ford, E.~B., Lystad, V., \& Rasio, F.~A.\
513: 2005, \nat, 434, 873
514:
515: \bibitem[Ford \& Rasio(2008)]{forras2008}
516: Ford, E.~B., \& Rasio, F.~A.\ 2008,
517: \apj, 686, 621
518:
519: \bibitem[Go{\'z}dziewski(2003)]{gozdziewski2003}
520: Go{\'z}dziewski, K.\ 2003, \aap, 398, 1151
521:
522: \bibitem[Go{\'z}dziewski \& Maciejewski(2003)]{gozmac2003}
523: Go{\'z}dziewski, K., \& Maciejewski, A.~J.\ 2003, \apjl, 586, L153
524:
525: \bibitem[Gregory(2007a)]{gregory2007a} Gregory, P.~C.\ 2007a, \mnras, 381, 1607
526:
527: \bibitem[Gregory(2007b)]{gregory2007b} Gregory, P.~C.\ 2007b, \mnras, 374, 1321
528:
529: \bibitem[Ji et al.(2003)]{jietal2003}
530: Ji, J., Liu, L., Kinoshita, H., Zhou, J.,
531: Nakai, H., \& Li, G.\ 2003, \apjl, 591, L57
532:
533: \bibitem[Kiseleva-Eggleton et al.(2002)]{kisetal2002}
534: Kiseleva-Eggleton, L., Bois, E., Rambaux, N.,
535: \& Dvorak, R.\ 2002, \apjl, 578, L145
536:
537: \bibitem[Lee \& Peale(2003)]{leepea2003}
538: Lee, M.~H., \& Peale, S.~J.\ 2003, \apj, 592, 1201
539:
540: \bibitem[Libert \& Henrard(2007)]{libhen2007}
541: Libert, A.-S., \& Henrard, J.\ 2007, \aap, 461, 759
542:
543: \bibitem[Malhotra(2002)]{malhotra2002}
544: Malhotra, R.\ 2002, \apjl, 575, L33
545:
546: \bibitem[Rasio \& Ford(1996)]{rasfor1996}
547: Rasio, F.~A., \& Ford, E.~B.\ 1996, Science, 274, 954
548:
549: \bibitem[Rodr{\'{\i}}guez \& Gallardo(2005)]{rodgal2005}
550: Rodr{\'{\i}}guez, A., \& Gallardo, T.\ 2005, \apj, 628, 1006
551:
552: \bibitem[S{\'a}ndor et al.(2007)]{sanetal2007}
553: S{\'a}ndor, Z., Kley, W., \& Klagyivik, P.\ 2007, \aap, 472, 981
554:
555: \bibitem[Veras \& Armitage(2007)]{verarm2007}
556: Veras, D., \& Armitage, P.~J.\ 2007, \apj, 661, 1311
557:
558: \bibitem[Weidenschilling \& Marzari(1996)]{weimar1996}
559: Weidenschilling, S.~J., \& Marzari, F.\ 1996, \nat, 384, 619
560:
561: \bibitem[Wright et al.(2008)]{wrightetal2008}
562: Wright, J.R., Upadhyay, S., Marcy, G.W., Fischer, D.A., Ford, E.B.\ 2008, \apj, submitted.
563:
564: \bibitem[Zhou \& Sun(2003)]{zhosun2003}
565: Zhou, J.-L., \& Sun, Y.-S.\ 2003, \apj, 598, 1290
566:
567:
568: \end{thebibliography}
569:
570:
571: %########################################################################
572: %\begin{figure}
573: %\epsscale{0.7}
574: %\plotone{f1.eps}
575: %\centerline{\psfig{figure=figures/stdev.eps,width=3.25truein,height=3.25truein}}
576: %\figcaption{The root mean square (dots) and mean absolute deviation
577: %(squares and triangles) about a libration center, and their standard
578: %deviations (vertical bars) as a function of the mass-weighted mean
579: %inclinations of both planets, binned into $9^{\circ}$ intervals. The
580: %top panel is for coplanar systems, and the next three panels are for
581: %systems with $0^{\circ} \le i_{rel} \le 30^{\circ}$, $30^{\circ} \le
582: %i_{rel} \le 60^{\circ}$, and $60^{\circ} \le i_{rel} \le 90^{\circ}$,
583: %respectively. Triangles indicate systems with planets on retrograde
584: %orbits (with same range of $\sin i_{rel}$), and data are plotted only
585: %for bins with at least three stable systems.
586: %\label{FigApsial}}
587: %\end{figure}
588: %########################################################################
589:
590: %\clearpage
591:
592: %########################################################################
593: %\begin{figure}
594: %\epsscale{0.7}
595: %\plotone{f1.eps}
596: %\centerline{\psfig{figure=figures/distributions.eps,width=3.25truein,height=3.25truein}}
597: %\figcaption{Values of $c_1$ (dots), $c_2$ (triangles), $c_3$
598: %(squares), and $c_4$ (crosses) and their standard deviations,
599: %for coplanar systems, binned according to the sine of the inclination of the invariable plane.
600: %mass-weighted mean of the $\sin i$ factors.
601: %\label{FigCvsIlos}
602: %\label{schem2}}
603: %\end{figure}
604: %########################################################################
605:
606: %\clearpage
607:
608: %########################################################################
609: %\begin{figure}
610: %\epsscale{0.7}
611: %\plotone{f2.eps}
612: %\centerline{\psfig{figure=figures/exceed.eps,width=3.25truein,height=3.25truein}}
613: %\figcaption{Fraction of time for which each planet's eccentricity exceeded
614: %a given threshold (x-axis) averaged over allowed orbital solutions.
615: %The curves with dots show results for the coplanar case, but assuming an
616: %isotropic distribution of inclinations relative to the plane of the sky. The dashed vertical
617: %lines indicate the median eccentricity values (0.031 and 0.361) for the planets'
618: %initial eccentricities, and their 5\%-95\% uncertainty ranges are
619: %indicated by the horizontal solid segments. HD 12661 c's orbital eccentricity
620: %exceeds its current value for approximately 96\% of its evolution.
621: %\label{FigEHisto}
622: %\label{schem1}}
623: %\end{figure}
624: %########################################################################
625:
626:
627: \begin{deluxetable}{ c c c c c c c c c c c}
628: \label{TabLib}
629: \tabletypesize{\scriptsize}
630: \tablecaption{Results of N-body Integrations}
631: \tablewidth{0pt}
632: \tablehead{
633: \colhead{$i_{rel}$} &
634: \colhead{(RMS) aligned fraction} &
635: \colhead{(MAD) aligned fraction} &
636: \colhead{RMS $\Delta\varpi'$} &
637: \colhead{MAD $\Delta\varpi'$} &
638: \colhead{Stable} &
639: \colhead{Isotropic}
640: }
641: \renewcommand{\arraystretch}{1.5}
642: \startdata
643: %\tableline
644: $0^{\circ}$ & $35.9\%$ & $31.7\%$ & $84.9^{\circ} \pm 8.5^{\circ}$ & $81.1^{\circ} \pm 7.8^{\circ}$ & $100\%$ & \nodata
645: \\
646: $0^{\circ}-30^{\circ}$ & $39.3\%$ & $35.2\%$ & $83.8^{\circ} \pm 9.4^{\circ}$ & $79.5^{\circ} \pm 8.6^{\circ}$ & $99.7\%$ & 6.7\%
647: \\
648: $30^{\circ}-60^{\circ}$ & $56.8\%$ & $51.8\%$ & $84.2^{\circ} \pm 10.4^{\circ}$ & $79.7^{\circ} \pm 9.8^{\circ}$ & $93.3\%$ & 18.3\%
649: \\
650: $60^{\circ}-90^{\circ}$ & $56.0\%$ & $52.0\%$ & $83.8^{\circ} \pm 10.5^{\circ}$ & $79.6^{\circ} \pm 9.9^{\circ}$ & $16.4\%$ & 25.0\%
651: \\
652: $90^{\circ}-120^{\circ}$ & --- & --- & --- & --- & $0.2\%$ & 25.0\%
653: \\
654: $120^{\circ}-150^{\circ}$ & $95.2\%$ & $95.2\%$ & $74.7^{\circ} \pm 12.1^{\circ}$ & $69.3^{\circ} \pm 11.1^{\circ}$ & $16.7\%$ & 18.3\%
655: \\
656: $150^{\circ}-180^{\circ}$ & $99.6\%$ & $99.2\%$ & $68.4^{\circ} \pm 12.8^{\circ}$ & $63.0^{\circ} \pm 12.1^{\circ}$ & $98.6\%$ & 6.7\%
657: \\ \hline
658: Prograde, isotropic & & & & & 69.8\% & 80.0\% \\
659: Retrograde, isotropic & & & & & 38.5\% & 50.0\% \\
660: %\tableline
661: \enddata
662: % Numbers: 2000, 1986, 1858, 334, 1, 114, 474
663: %\tablecomments{\label{TabLib} \small{}}
664: %
665: \tablecomments{\label{TabLib} \small{
666: Columns 2 and 3 each list the RMS and MAD percent of stable systems with a libration center of $0^\circ$ instead of $180^\circ$ for the range of relative inclinations in column 1. Columns 4 and 5 report the RMS and MAD variations for systems which librate around the most common center for that range of initial relative inclinations. Column 6 lists the percent of stable systems, and Column 7 lists the percent of systems with the given range of inclination for an isotropic distribution. The bottom two rows present aggregated stability statistics for the isotropic prograde and retrograde distributions of orbits. 2000 simulations were performed for each prograde relative inclination bin, and 500 simulations for each retrograde bin.}}
667: %
668: \end{deluxetable}
669:
670:
671:
672: \end{document}
673:
674:
675:
676:
677: