1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass{emulateapj}
3:
4: \begin{document}
5:
6: \shorttitle{Parameters and Predictions for HD~17156b}
7: \shortauthors{Irwin et al.}
8:
9: \title{Parameters and Predictions for the Long-Period Transiting Planet HD~17156b}
10:
11: \author{Jonathan Irwin, David Charbonneau\altaffilmark{1} and Philip Nutzman}
12: \affil{Harvard-Smithsonian Center for Astrophysics, 60 Garden St., Cambridge, MA 02138}
13: \email{jirwin at cfa dot harvard dot edu}
14:
15: \author{William F. Welsh and Abhijith Rajan}
16: \affil{Dept of Astronomy, San Diego State University, 5500 Campanile Dr., San Diego, CA 92182}
17:
18: \author{Marton Hidas, Timothy M. Brown and Timothy A. Lister}
19: \affil{Las Cumbres Observatory, 6740 Cortona Dr. Suite 102, Santa Barbara, CA 93117}
20:
21: \author{Donald Davies}
22: \affil{23819 Ladeene Ave., Torrance, CA 90505}
23:
24: \author{Gregory Laughlin and Jonathan Langton}
25: \affil{UCO/Lick Observatory, University of California at Santa Cruz, Santa Cruz, CA 95064}
26:
27: \altaffiltext{1}{Alfred P. Sloan Research Fellow}
28:
29: \begin{abstract}
30: We report high-cadence time-series photometry of the recently-discovered
31: transiting exoplanet system HD~17156, spanning the time of transit on
32: UT 2007 October 1, from three separate observatories. We present a
33: joint analysis of our photometry, previously published radial velocity
34: measurements, and times of transit center for 3 additional events.
35: Adopting the spectroscopically-determined values and uncertainties for
36: the stellar mass and radius, we estimate a planet radius of $R_{p} =
37: 1.01 \pm 0.09 \, R_{\rm Jup}$ and an inclination of $i =
38: 86.5^{+1.1}_{-0.7}$~degrees. We find a time of transit center of
39: $T_{c} = 2454374.8338 \pm 0.0020$~HJD and an orbital period of $P =
40: 21.21691 \pm 0.00071$~days, and note that the 4 transits reported to
41: date show no sign of timing variations that would indicate the
42: presence of a third body in the system. Our results do not preclude
43: the existence of a secondary eclipse, but imply there is only a 9.2\%
44: chance for this to be present, and an even lower probability (6.9\%)
45: that the secondary eclipse would be a non-grazing event. Due to its
46: eccentric orbit and long period, HD~17156b is a fascinating object for
47: the study of the dynamics of exoplanet atmospheres. To aid such
48: future studies, we present theoretical light curves for the variable
49: infrared emission from the visible hemisphere of the planet throughout
50: its orbit.
51: \end{abstract}
52:
53: \keywords{planetary systems -- stars: individual (HD~17156) -- techniques: photometric}
54:
55: \section{Introduction}
56: \label{intro_sect}
57:
58: Twenty-eight transiting planets are now reported in the
59: literature\footnote{See {\tt
60: http://www.inscience.ch/transits/}}, and the doubling time
61: scale for new detections is now roughly one year. As reviewed by
62: \citet{char2007}, it is these objects that have allowed the study of
63: the physical structure and atmospheric chemistry and dynamics of gas
64: giant exoplanets, and opened the field of comparative exoplanetology.
65: However, the majority have orbital periods of a few days, due both to
66: the lower geometric probability of transits occurring as the orbital
67: period increases, and due to the limitations of ground-based
68: photometric transit surveys that strongly favor detection of
69: short-period systems. The latter problem can be circumvented by
70: searching for transits of known radial velocity planet-bearing stars,
71: putting to use the radial velocity information to constrain the
72: possible time of transit. This is the primary aim of the
73: Transitsearch.org network \citep{sea2003}\footnote{See {\tt
74: http://www.transitsearch.org}}, which, starting in 2002, has been
75: conducting photometric follow-up observations of known radial velocity
76: detected planets. This network is a prime example of successful
77: collaboration between amateur and professional astronomers.
78:
79: Although longer period planets possess a lower geometric probability to
80: transit, there exists a loop-hole for some planets on highly eccentric
81: orbits. The transit probability for a planet on an eccentric orbit,
82: with periastron near inferior conjunction, is amplified by a factor
83: \begin{equation}
84: A = \left[ {{1+e\cos(\frac{\pi}{2} - \omega)} \over{1-e^2}}\right]
85: \end{equation}
86: where $e$ is the eccentricity, and $\omega$ is the argument of
87: pericenter \citep{sea2003}. An amplification factor near 2.9 for the
88: planet orbiting HD~17156 \citep{fischer2007}, along with an extremely
89: short time window for possible transits, brought the system to the
90: attention of the Transitsearch.org network. This was rewarded by the
91: detection of transits as recently announced by \citet{bar2007}.
92:
93: This unique system contains a $3.1$-$M_{\rm Jup}$ transiting planet in
94: a $21.2$ day highly-eccentric orbit ($e = 0.67$) about a bright ($V =
95: 8.2$) G0V star. \citet{gil2007} have recently presented refined
96: estimates of the system parameters based on new photometric data.
97: Even among planets with periods beyond 5 days, where the timescale for
98: tidal circularization quickly grows to exceed the age of most systems,
99: HD~17156b's eccentricity is unusually high, thus making it an
100: interesting test case for models of planetary migration. Preliminary
101: measurements of the Rossiter-McLaughlin effect
102: \citep{rossiter24,mclaughlin24,gw2007,winnasp2007} by \cite{nar2007}
103: show evidence for a
104: misalignment between the stellar spin axis and the planetary orbit
105: axis, which may indicate a migration mechanism involving planet-planet
106: scattering \citep[see e.g.][]{chat2007}, or Kozai migration under
107: perturbation by a yet undetected stellar companion or second planet
108: \citep{fab2007,wu2007}. The large eccentricity and long period also
109: make HD~17156 a particularly attractive target for several additional
110: follow-up studies. First, it presents a unique opportunity for the
111: study of the structure and dynamics of a gas-giant atmosphere under
112: strongly varying illumination, through infrared monitoring of the
113: planetary emission (e.g. \citealt{har2006}; \citealt{cow2007};
114: \citealt{knut2007a,knut2007b}). Second, by monitoring
115: successive times of transit, the presence of additional bodies in the
116: system can be detected or constrained from the resulting perturbations
117: on the orbit of HD~17156b (\citealt{hm2005}; \citealt{agol2005};
118: \citealt{sa2005}).
119:
120: The purpose of our paper is three-fold: First, we seek to refine the
121: estimates of the system parameters that were only poorly constrained
122: by the light curves of \citet{bar2007}. Second, we document the time
123: of center of transit and search for transit timing variations. Third,
124: we evaluate the likelihood that a secondary eclipse will be observable
125: for HD~17156, and present theoretical predictions of the infrared phase
126: variations as might be detected with the {\it Spitzer Space
127: Telescope}. We elected to perform an independent analysis from that
128: of \citet{gil2007}, and therefore do not use their orbital parameters
129: as constraints, since at the time of writing these are still subject
130: to change as their paper has not yet been accepted for publication.
131: We do however make use of the time of mid-transit presented in their
132: Table 1.
133:
134: We begin by presenting a description of our observations
135: of the transit event on UT 2007 October 1 in \S \ref{obs_sect}. We
136: then present our combined analysis of the extant radial velocity
137: measurements (\S \ref{rv_sect}) and photometric data (\S
138: \ref{phot_sect}), and our resulting estimates of the system
139: parameters and likelihood of a secondary eclipse. In \S
140: \ref{disc_sect}, we compare our constraints on the planet mass and
141: radius with planetary structural models, summarize the current
142: constraints on the presence of transit timing variations in the
143: system, and present model predictions for the planetary infrared light
144: curve. We conclude in \S \ref{summ_sect} with a discussion of
145: compelling avenues for future research.
146:
147: \section{Observations}
148: \label{obs_sect}
149:
150: Using the orbital period given by \citet{fischer2007} and the time of
151: mid-transit measured by \citet{bar2007}, we predicted that an event
152: would occur on UT 2007 October 1 with transit center at UT 7:53, which
153: was well-situated for observatories in the southwestern United States.
154: We present below a description of the data we gathered from three such
155: observatories, and summarize the calibration of the raw frames and
156: extraction of the time series for each.
157:
158: \subsection{Mount Laguna Observatory}
159:
160: We observed HD~17156 with the Smith 0.6~m telescope at Mount Laguna
161: Observatory using a Bessell $R$ filter and an SBIG STL-1001E CCD
162: camera (giving a scale of $0\farcs4$/pixel at the $f/20$ Cassegrain
163: focus), using BD+71~168 ($V = 9.57$, B8 spectral type) as our local
164: comparison star. The integration time for each exposure was 4~s, with
165: a 4.3s readout
166: time and a net cadence of 8.3~s. For each frame, we calculated the
167: time at mid-exposure and corrected this time to heliocentric Julian
168: Day. We dark-subtracted and flat-fielded the raw images in IRAF, and
169: then performed aperture photometry using a circular aperture with
170: radius 12.8~arcsec. The use of the large aperture insulated our
171: analysis against seeing-induced aperture losses, as the seeing varied
172: from 2.1~arcsec to 2.7~arcsec over the course of the night.
173: Clouds were present at several times during the observations,
174: especially after mid-transit. To avoid systematic effects due to
175: cloud-induced differential extinction of the different color stars, we
176: rejected observations for which the calibration star showed more than
177: $8\%$ light loss. The resulting 2148 points were binned 20:1 and
178: uncertainties were estimated from the RMS scatter in each bin divided
179: by $\sqrt{20}$. Typical uncertainties for the binned points were
180: $0.3-0.7 \%$.
181:
182: \subsection{Torrance California}
183:
184: We observed a field centered on HD~17156 using a commercial $0.25$m
185: Meade LX200 telescope and a Meade Deep Sky Imager Pro II CCD camera.
186: We employed an f/3.3 focal reducer to increase the field-of-view to
187: $\sim 20 \times 30$ arcmin. We gathered $10\,345$ exposures through
188: an Edmund Optics IR-pass filter (approximating the conventional
189: $I$-band), between UT 4:20 and 13:25. The integration time for each
190: image was 2~s. Weather conditions were clear. Approximately $10\%$
191: of the frames were lost due to an intermittent issue with the CCD
192: readout, and due to wind shake, and the guide star was lost
193: temporarily at 6:10 UT, resulting in a large pointing shift ($\sim 5\
194: {\rm arcmin}$).
195:
196: These images were processed using the software described in
197: \citet{i2007a}, which was originally developed for the Monitor project
198: \citep{aigrain2007}. Individual frames were processed by subtracting
199: a master dark and dividing by a master flat field. We then ran the
200: pipeline source detection software (see \citealt{i2007a}) on all of
201: the data frames, choosing a single frame taken in good conditions for
202: use as a reference. For each image, we derived an astrometric
203: solution employing an implementation of the triangle matching
204: algorithm of \citet{groth86} to cope with shifts in the telescope
205: position (these were $\sim 10$ pixels peak-to-peak, in a periodic
206: fashion, presumably due to errors in the telescope worm drives), and
207: field rotation due to polar misalignment of the mount.
208:
209: We generated light curves using the method described in
210: \citet{i2007a}, using the nearby star BD+71 168 as the comparison
211: source for the differential photometry. We found that the normalized
212: time series data gathered before and after the large image motion at
213: UT 6:10 have an photometric offset of roughly 3\%. Since the data
214: obtained before UT 6:10 were well before the ingress, we simply
215: trimmed them from the time-series; similarly, we trimmed the data that
216: occurred more than 0.2 days after the transit midpoint. We also
217: trimmed single large outliers by identifying all points that deviated
218: from the median of the time-series by more than 8\%. The typical
219: photometric errors in the final, trimmed time-series are 2\% per data
220: point, as determined from the RMS variation from the data gathered
221: after egress. We then binned this time series by a factor of 15 (and
222: assumed that the errors in the binned data were reduced according to
223: Poisson statistics); this binning was necessary to keep the
224: computational time for the model fitting (\S \ref{phot_sect}) to a
225: manageable duration, but the binning did not degrade the quality of
226: the fits.
227:
228: \subsection{Las Cumbres Observatory Global Telescope}
229:
230: We also observed the event using a $0.4$m Meade RCX400 telescope on a
231: custom mount, temporarily located in the car park of the Las Cumbres
232: Observatory Global Telescope (LCOGT) offices in Goleta, California.
233: We used an SBIG STL-6303E CCD camera to image a $20\times 30$~arcmin
234: field with 0.6-arcsec pixels. 380 useful images were taken between UT
235: 5:30 and 11:00, with the airmass decreasing from 1.6 to a minimum of
236: 1.25 near UT 10:00. No autoguiding was used, and the field drifted by
237: $\sim 100$ arcsec during the run. The observations were made through
238: the ``Red'' filter from the LRGBC filter set provided with the camera
239: by SBIG,\footnote{\tt
240: http://www.sbig.com/large\_format/filterchart\_large.htm} with $\sim
241: 90$\% transmission in the range 580--680~nm. Exposure times varied
242: from 20 to 30 seconds. The readout time was 22 seconds. The telescope
243: was defocused to avoid saturating the target star, which was the
244: brightest in the field. The amount of defocus was increased at about
245: UT 06:30, shortly before the transit ingress, but this did not
246: significantly affect the relative photometry at that point.
247:
248: Bias and dark subtraction and flat fielding were done using standard
249: IRAF tasks. HD~17156 and ten reference stars were measured using the
250: DAOPHOT aperture photometry package, with apertures of radius
251: 10.6~arcsec. The light curve was divided by the average of the
252: reference light curves to remove the effects of varying atmospheric
253: extinction and focus changes. The standard deviation of the (unbinned)
254: points outside the transit was 0.4\%.
255:
256: The scatter of these data is lower than for the Mount
257: Laguna $0.6$m observations, despite the smaller aperture, since the
258: larger field-of-view permits the use of multiple comparison stars. In
259: the Mount Laguna data, the signal-to-noise ratio of the final light
260: curve was limited by that of the comparison light curve, due to the
261: use of a comparison star fainter than the target. By using multiple
262: comparison stars for the Las Cumbres data, it was possible to attain a
263: higher signal level in the comparison light curve, such that the
264: overall noise was dominated by the target, hence giving a net overall
265: improvement in signal-to-noise ratio despite the use of a smaller
266: aperture telescope. Observing conditions were also somewhat worse at
267: Mount Laguna, and this along with the use of defocus to render the
268: observations insensitive to seeing variations, and to average out
269: pixel-to-pixel flat fielding errors, also contribute to reducing
270: the noise level in the Las Cumbres data.
271:
272: \section{Analysis of the Radial Velocity Data}
273: \label{rv_sect}
274:
275: Our orbital model was obtained by jointly fitting the
276: \citet{fischer2007} radial velocity data (converted to HJD) in
277: conjunction with the following four observed central transit times,
278: numbered $N = -1,0,2,3$ (where we denote our observation on UT 2007
279: October 1 as event $N = 0$): $T_{c,-1} = {\rm HJD}\, 2454353.61 \pm
280: 0.02$ \citep{bar2007}, $T_{c,0} = {\rm HJD}\,2454374.8338 \pm 0.0020$
281: (this paper; see \S \ref{phot_sect}), $T_{c,2} = {\rm
282: HJD}\,2454417.2645 \pm 0.0021$ \citep{nar2007}, and $T_{c,3} = {\rm
283: HJD}\,{2454438.4835}^{+0.0009}_{-0.0025}$ \citep{gil2007}.
284:
285: For a Keplerian orbit, the instantaneous stellar radial velocity is
286: given by
287: $V_{{\rm mod},i} = K \, [\cos(f_{i} + \omega)+e \, \cos\omega]$,
288: where $f_{i}$ is the true anomaly of the planet at time $t_i$.
289: If we assume an edge-on orbit, the time of central transit occurs when
290: $f_{t} = {\pi/2}-\omega$.
291: Given the precision of the available photometry, this approximation
292: to $f_{t}$ is excellent.
293: The true anomaly, $f$, is in turn related to the
294: the eccentric anomaly, $E$, via
295: $E = 2\arctan{ \sqrt{(1-e)/(1+e)} \tan(f/2) }$,
296: so that a central transit, $T_{c,i}$, occurs at a fixed interval
297: $\Delta T = P(E-e\sin{E})/{2\pi}$
298: following the epoch of a periastron passage, $T_{p,i}$.
299:
300: Given $N_{RV}$ radial velocity measurements (here, $N_{RV} = 33$) and
301: $N_{T_c}$ central
302: transit measurements, the goodness-of-fit function of a particular
303: orbital model is calculated as:
304: \begin{equation}
305: \chi^{2} = \sum_{i=1}^{N_{RV}}\left({ V_{{\rm mod},i}- V_{{\rm obs},i}
306: \over{\sigma_{i}}}\right)^{2}
307: + \sum_{i=0}^{N_{Tc}}\left({ T_{p,i} + \Delta T-T_{c,i}
308: \over{\sigma_{i}}}\right)^{2}
309: \end{equation}
310: We use a Levenberg-Marquardt algorithm to minimize $\chi^{2}$. The
311: best-fitting orbital parameters are listed in Table \ref{tbl:params}.
312: To obtain the quoted uncertainties, a simple bootstrap procedure was
313: used. An aggregate of 1000 alternate datasets was created by (1)
314: redrawing the radial velocities with replacement, and (2) sampling
315: $T_c$ values for the four transits by drawing from the Gaussian
316: distributions implied by the quoted errors. Fits to each of these
317: data sets were obtained, and the resulting distributions of orbital
318: parameters were used to derive the $68.3\%$ confidence intervals for
319: each parameter. In Table \ref{tbl:params}, we have also given in
320: brackets the standard deviations of these parameter distributions.
321: The latter are larger than the former due to the existence of outliers
322: in the bootstrap sample. These are generated because the relative
323: importance of each radial velocity data point in constraining the
324: derived orbital parameters depends on the orbital phase as a result of
325: the highly eccentric orbit: the points close to periastron have the
326: largest effect on the fit. Consequently, when the bootstrapping
327: procedure exchanges a point close to periastron, the parameters
328: estimated from this particular realization of the procedure show
329: sizable variations. We therefore prefer the $68.3\%$ confidence
330: intervals, since these are robust to the presence of the outliers,
331: which are in turn the result of the bootstrap procedure not being
332: strictly applicable to data such as these which are not independent
333: and identically-distributed.
334:
335: %\clearpage
336: \begin{deluxetable}{lc}
337: \tabletypesize{\normalsize}
338: \tablecaption{\label{tbl:params} Fitted Parameters for HD~17156.}
339: \tablewidth{0pt}
340: \tablenotetext{a}{For
341: the quantities $P$, $e$, ${\omega}$, $K$ and $T_p$, $68.3\%$
342: confidence intervals are quoted, with the standard deviation from
343: bootstrapping in brackets. The latter are larger due to outliers in
344: the bootstrap sample generated by the high sensitivity of the fitted
345: parameters to the velocities near periastron passage. As discussed
346: in the text, the $68.3\%$ confidence intervals are preferred.}
347:
348: \tablehead{
349: \colhead{Parameter\tablenotemark{a}} & \colhead{Value}
350: }
351:
352: \startdata
353: $P$ & $21.21691 \pm 0.00071\ (0.00094) \ {\rm d}$ \\
354: $e$ & $0.670 \pm 0.006\ (0.020)$ \\
355: ${\omega}$ & $121{\fdg}3 \pm 0{\fdg}9\ (1{\fdg}8)$ \\
356: $K$ & $273.8 \pm 3.4\ (16.3) \ {\rm m \, s^{-1}}$ \\
357: $T_p$ & $2453738.605 \pm 0.024\ (0.036) \ {\rm HJD}$ \\
358: \\
359: $T_c$ & $2454374.8338 \pm 0.0020 \ {\rm HJD}$ \\
360: $i$ & $86{\fdg}5^{+1\fdg1}_{-0{\fdg}7}$ \\
361: $R_{p}/R_{\star}$ & $0.070 \pm 0.003$ \\
362: $R_{p}$ & $1.01 \pm 0.09 \ R_{\rm Jup}$ \\
363: $R_{\star}$ & $1.47 \pm 0.08 \ R_{\Sun} $ \\
364: $\rho$ & $3.8^{+0.8}_{-1.1}$ g cm$^{-3}$ \\
365: \\
366: $f_1$ & $0.9989 \pm 0.0025$ \\
367: $k_1$ & $-0.0006 \pm 0.0015$ \\
368: $f_2$ & $1.006 \pm 0.003$ \\
369: $k_2$ & $0.0049 \pm 0.0022$ \\
370: $c_1$ & $-0.0024 \pm 0.0005$ \\
371: $c_2$ & $0.0173 \pm 0.0029$ \\
372: $c_3$ & $0.204 \pm 0.064$ \\
373: \enddata
374:
375: \end{deluxetable}
376: %\clearpage
377:
378: We note that the addition of the transit timing data provides a very
379: strong constraint on the orbital period, but has only a minor effect
380: on the other orbital parameters, which are in excellent agreement with
381: the solution given by \citet{fischer2007}.
382:
383: \section{Photometric Analysis}
384: \label{phot_sect}
385:
386: We pooled together the three photometric data sets to fit a model
387: light curve parametrized by the orbital and physical properties of the
388: HD~17156 system. Our model incorporates 4 orbital parameters
389: initially determined through the fit to the RV data: the period $P$,
390: eccentricity $e$, time of periastron $T_{p}$, and argument of
391: pericenter $\omega$. Given this orbit, we calculate a baseline light
392: curve for a quadratically limb-darkened star (characterized by
393: the stellar mass $M_{\star}$, radius $R_{\star}$, and 2 limb darkening
394: coefficients for each photometric band pass) that is being occulted by
395: a planet ($r_{p} = R_{p}/R_{\star}$) orbiting at an inclination $i$.
396: To calculate the baseline light curve, we employ the analytic formulae
397: of \citet{ma2002}, together with the quadratic limb-darkening
398: coefficients tabulated by \citet{claret2000,claret2004}, for
399: $T_{\mathrm{eff}}$ = 6000~K, [Fe/H] = $0.2$, $\log{g} = 4.0$ (based on
400: the stellar parameters given by \citealt{tak2007}). We adopted
401: $R$-band coefficients for the Mount Laguna data, SDSS $r$-band
402: coefficients for the Las Cumbres data, and $I$-band coefficients for
403: the Torrance California data. In the latter two cases, the filters
404: used for the observations only approximate $r$ and $I$, but the
405: approximation is more than suitable for our purposes (in particular,
406: the goodness-of-fit statistic described below is negligibly affected
407: by errors in the assumed bandpass). The baseline flux is then modified
408: by observational correction factors unique to each dataset (7
409: additional parameters described below).
410:
411: Nominally, our model fixes $P$, $e$, $\omega$, 6 limb darkening
412: coefficients, and $M_{\star}$, although we iteratively update the first
413: three of these parameters as the RV fit is apprised of the transit
414: timing resulting from the light curve fit. Ultimately, the change in
415: these three parameters from the iterative update process had
416: negligible effect on the quality of the light curve fit and had
417: negligible consequences for the stellar and planetary properties
418: determined by the transit analysis. Note that although $T_p$ is a
419: fit-variable in our transit analysis, the value that is really being
420: constrained by the photometry is the time of central transit, $T_c$,
421: which is related to a degenerate combination of $P$, $e$, $\omega$,
422: and $T_p$. We fix $M_{\star}$ to the value $1.2 M_{\odot}$
423: \citep{fischer2007}, which was determined by matching spectroscopic
424: observations to stellar evolution tracks \citep{tak2007}. Here, the
425: uncertainty in $M_{\star}$, for which the 95\% confidence interval is
426: $1.1-1.3\ M_{\odot}$, has only weak impact on the light curve fit. This is
427: because, as described below, we adopt an external constraint on the stellar
428: radius.
429:
430: Our light curve model employed 11 free parameters: $T_p$, the
431: planet-star radius ratio $r_p$, $R_{\star}$, $i$, and 7 additional
432: parameters related to observational corrections. The Torrance and
433: Mount Laguna fluxes were each multiplied by airmass corrections of the
434: form $f \times \exp(-k a)$, where $f$ is a normalization factor, $a$
435: is the airmass, and $k$ is the extinction coefficient. We note, however,
436: that including the Mount Laguna airmass correction had little effect on the quality of fit (indeed, the determined $f$ and $k$ were consistent with 1 and 0
437: respectively; see Table \ref{tbl:params}). We attempted
438: the same airmass correction for the Las Cumbres data, but found that a quadratic function
439: of time significantly improved the quality of fit. We adopted a
440: $\chi^2$ function as our goodness-of-fit statistic, with an additional
441: term reflecting a spectroscopic prior on $R_{\star}$, which is
442: approximately Gaussian with $\sigma = 0.085$ (G. Takeda, personal
443: communication). This additional constraint proved necessary, as our
444: data could not independently determine the stellar radius. The
445: goodness-of-fit function is
446: \begin{equation}
447: \chi^2 = \sum_{i}{ \left(
448: \frac{f_{mod}(i)-f_{obs}(i)}{\sigma_i}\right)^2 } + \frac{(R_{\star}/R_{\Sun}-1.47)^2}{0.085^2}
449: \end{equation}
450: where $f_{mod}(i)$ is the calculated flux at the time of the $i^{th}$
451: data point, $f_{obs}(i)$ is the $i^{th}$ flux measurement, and
452: $\sigma_i$ is the uncertainty of the $i^{th}$ flux measurement
453: (derived as described for each data-set in \S \ref{obs_sect}). Using
454: an IDL implementation of the \texttt{amoeba} algorithm (e.g. see
455: \citealt{press92}), we performed an initial $\chi^2$ minimization over
456: the space of free parameters in our model. Using the results, we then
457: rescaled the $\sigma_i$ so that the reduced $\chi^2$ equals 1
458: separately for each dataset. We performed iterations of transit model
459: fitting in conjunction with radial velocity fitting, using the derived
460: transit timing to inform the radial velocity fit, which in turn
461: updated the $P$, $e$, and $\omega$ used for the transit fit. As
462: mentioned above, the parameter update had negligible effect on the
463: derived stellar and planetary properties. The data, corrected for the
464: airmass and instrumental effects, are shown in Figure~\ref{lc} along
465: with the best-fitting solution.
466:
467: %\clearpage
468: \begin{figure}
469: %\epsscale{0.9} % AASTeX
470: \epsscale{1.1}
471: \plotone{f1.eps}
472: \caption{Binned light curve of HD~17156, with the airmass and instrumental
473: corrections described in \S 4 applied. \emph{Top:} Mount Laguna $R$-band
474: photometry. \emph{Middle:} Las Cumbres photometry in a filter approximating
475: $r$. \emph{Bottom:} Torrance CA photometry in a filter approximating $I$.
476: Dashed lines show the model fit simultaneously to the light curve and radial
477: velocity data, as discussed in the text.}
478: \label{lc}
479: \end{figure}
480: %\clearpage
481:
482: The uncertainty in the transit time was assessed by finely stepping
483: $T_p$ through values near its best fitting value and calculating the
484: minimum $\chi^2$ at each step, allowing the other free parameters to
485: vary. The 1-$\sigma$ upper and lower limits were assessed by noting
486: the $T_p$ values at which $\Delta \chi^2 = 1$. Given the other
487: orbital parameters, we mapped each $T_p$ into the corresponding time
488: of central transit, $T_c$. The resulting best fitting $T_c$ and
489: uncertainties are reported in Table \ref{tbl:params}. Over the range
490: of allowed $T_p$, we found very little correlation between it and the
491: best fitting stellar/planetary parameters. For this reason, we decided to
492: fix $T_p$ at its best fitting value for the Markov Chain Monte Carlo
493: (MCMC) analysis described below. Note that by fixing $T_p$, the
494: MCMC code only needs to perform a full Keplerian orbit calculation
495: once, rather than at every step in the chain, thus dramatically
496: speeding up the computation.
497:
498: To estimate the uncertainty in the 10 remaining model parameters, we
499: used an MCMC algorithm following the basic recipe described, for
500: example, in \citet{winn2007}, with whom we share the terminology used
501: below. We produced 10 chains, each of $10^6$ points, starting from
502: independent initial parameter values. The ``jump function'' was tuned
503: so that $\sim 25 \%$ of jumps were executed, and the first $10 \%$ of
504: each chain was discarded to avoid the effect of the initial condition.
505: We combined the chains, and in Table \ref{tbl:params} we report the
506: median value of each parameter, and assign uncertainties by taking the
507: standard deviation of that parameter (except $i$, for which we report
508: the $68.3\%$ confidence interval). In Figure~\ref{mcmc}, we show the results
509: of this MCMC analysis. One of the major benefits of MCMC is that
510: probability distributions for derived quantities, such as the
511: secondary eclipse impact parameter (in the bottom right panel of
512: Figure~\ref{mcmc}), are produced directly. Here, the impact parameter was
513: taken to be the minimum projected separation between planet and star
514: center in the units of the stellar radius.
515:
516: %\clearpage
517: \begin{figure}
518: %\epsscale{1.0} % AASTeX
519: \epsscale{1.1}
520: \plotone{f2.eps}
521: \caption{Posterior probability distributions for $R_{\star}$, the
522: planet-star radius ratio, $R_p$, and the secondary eclipse impact
523: parameter, based on $10^7$ MCMC samples. The secondary eclipse impact
524: parameter is the minimum projected separation between planet and star
525: center (near time of superior conjunction), in units of the
526: $R_{\star}$. The dark shaded region shows where full secondary
527: eclipses occur, while the light shaded region shows where grazing
528: eclipses occur.}
529: \label{mcmc}
530: \end{figure}
531: %\clearpage
532:
533: \section{Discussion}
534: \label{disc_sect}
535:
536: %Comparing the parameters in Table \ref{tbl:params} with those in Table
537: %1 of \citet{gil2007} indicates that although the values are
538: %comparable within the quoted uncertainties in the majority of cases,
539: %the uncertainties reported by \citet{gil2007} are substantially
540: %smaller than those from our analysis, and indeed, seem to be
541: %unrealistically small. Although reduced uncertainties in the
542: %parameters derived from the light curves are expected since they
543: %analysed photometic data of somewhat better quality, as an example,
544: %the estimates of the orbital parameters do not appear to incorporate
545: %the propagated uncertainty in the value of the stellar mass, which is
546: %degenerate with the parameters derived from the fitting of the radial
547: %velocities. Given that the same radial velocity data were used in
548: %both cases, this discrepancy must be due to the method used for
549: %analysis.
550: %
551: %Given the possibility of future revisions to the \citet{gil2007}
552: %parameters, they were not used as constraints in the fitting in this
553: %work, as discussed in \S \ref{intro_sect}, and are not considered in
554: %the discussion which follows.
555:
556: \subsection{Planetary mass, radius and density}
557:
558: The planet's mass of $3.13 \pm 0.11$ M$_{\rm Jup}$ \citep{fischer2007}
559: is the third most massive among the known transiting planets,
560: surpassed only by HD~147506b (8.6 $M_{\rm Jup}$; also known as HAT-P-2b;
561: \citealt{bakos2007}) and XO-3b (13.2 M$_{\rm Jup}$; \citealt{jk2007}).
562: We find that the mean density of HD~17156 is $\rho=3.8^{+0.8}_{-1.1}$
563: g cm$^{-3}$, substantially higher than other transiting extrasolar
564: planets, again with the exception of the remarkable planet HD~147506b
565: ($\rho$=11.9 g cm$^{-3}$; \citealt{loeil2007}). HD~17156b begins to
566: bridge the gap between HD~147506b and the other transiting planets.
567:
568: The presence of a large rocky or metallic core has been suggested to
569: account for ``hot Jupiters'' with high densities. For HD~17156b, the
570: models of \citet{boden2003} predict a radius of 1.11 $R_{\rm Jup}$
571: given the mass of HD~17156b, which is $\sim 1-\sigma$ larger than the
572: radius derived from the light curve fitting, or equivalently, the
573: measured density is higher than the models would predict. However,
574: for such a massive planet, this discrepancy is difficult to resolve by
575: invoking the presence of a solid core, since the dependence of
576: planetary radius on the presence of a core is predicted to be very
577: weak in this mass regime (e.g. \citealt{boden2003}; \citealt{bur2007};
578: \citealt{fortney2007a}). For example, \citet{boden2003} predict that
579: the addition of a 30 $M_{\earth}$ core reduces the radius by only
580: $\sim$ 0.01 $R_{\rm Jup}$ for a $\sim$ 3 $M_{\rm Jup}$ planet. This
581: scenario would therefore require an extremely massive core to account
582: for the measured radius of HD~17156b.
583:
584: Although a large rocky or metallic core can account for planets
585: with high densities, the low density (i.e. large radius) planets such
586: as TrES-4 \citep{mand2007} and WASP-1 \citep{cc2007} seem
587: to defy explanation. One possible mechanism for keeping planetary radii
588: from shrinking during planetary evolution is to pump energy into the
589: planet via tidal interactions (e.g. \citealt{mard2007}). Such
590: orbital energy transfer mechanisms depend strongly on the orbital
591: eccentricity, and HD~17156b represents a good test-case for these
592: theories, having a significantly eccentric orbit ($e=0.67$). However,
593: the radius of HD~17156b is not in any way exceptional when compared to
594: the existing extrasolar planets with close to circular orbits. This
595: leads one to speculate that tidal energy transfer via orbital dynamics
596: may not play a substantial role in planetary radius evolution, or that
597: in massive planets such as HD~17156b a mechanism is acting to allow
598: the planets to more rapidly evolve to smaller radii. We note,
599: however, that the recently-announced massive planet XO-3b
600: \citep{jk2007} also has an eccentric orbit, but in contrast to
601: HD~17156b, its radius is very large ($1.92 R_{\rm Jup}$). Thus the
602: challenges posed by newly-discovered exoplanets to theoretical models
603: of their physical structure seem to continue unabated.
604:
605: \subsection{Search for transit timing variations}
606:
607: The availability of multiple measured times-of-transit allows the
608: presence of additional bodies in the HD~17156 system to be detected or
609: constrained, by searching for the influence of their gravitational
610: perturbations on the orbit of HD~17156b (\citealt{hm2005};
611: \citealt{agol2005}; \citealt{sa2005}).
612:
613: In Figure~\ref{oc}, we plot the observed minus calculated times of center of
614: transit for the 4 events reported in the literature, including the
615: event described in this paper. We find that these results do not yet
616: reveal any evidence for timing variations that would indicate the
617: presence of a third body in the system. Due to the long orbital
618: period, HD~17156 may be more amenable to a search for such variations.
619: However, the large orbital eccentricity implies a large region for
620: which dynamical stability considerations would exclude the presence of
621: such planets.
622:
623: %\clearpage
624: \begin{figure}
625: %\epsscale{1.0} % AASTeX
626: \epsscale{1.1}
627: \plotone{f3.eps}
628: \caption{Observed minus calculated times of transit for the four
629: transit events discussed in this work. The dashed line shows the
630: prediction assuming the values of $T_c$ and $P$ from Table
631: \ref{tbl:params}. Since the value of $P$ was determined from a
632: simultaneous fit to the transit times and radial velocity data, it
633: does not represent the best fit to the transit times alone.}
634: \label{oc}
635: \end{figure}
636: %\clearpage
637:
638: \subsection{Prospects for infra-red observations}
639:
640: Our photometric analysis directly addresses the probability that
641: HD~17156b undergoes secondary eclipse, which is only possible for
642: certain combinations of $R_{\star}$ and $i$. We find a 9.2\% chance
643: for secondary eclipses to occur, but only a 6.9 \% chance for
644: non-grazing eclipses. If, indeed, HD~17156b undergoes eclipses, $i$
645: becomes locked rather tightly, thus greatly diminishing the allowed
646: volume of parameter space for the stellar and planetary properties.
647: Under this constraint, we find that the new best fitting values for
648: ($i$,$R_{\star}$,$R_p$) are
649: ($88{\fdg}7~^{+1{\fdg}0}_{-0{\fdg}1}$, $1.34 \pm
650: 0.03~R_{\odot}$, $0.89 \pm 0.04~R_{\mathrm{Jup}}$). These
651: considerations serve to motivate a search for the secondary eclipses.
652:
653: HD~17156b's large eccentricity leads to a 25-fold variation in
654: received stellar flux over the course of the $21.2169\,{\rm d}$
655: orbital period. This dramatic variation in illumination should produce
656: complex weather at the planet's photosphere. In addition, the planet
657: is subject to strong tidal forces during its periastron passage. The
658: planet is therefore expected to be in pseudo-synchronous rotation, in
659: which the planetary spin will be roughly synchronous with the orbit
660: during the interval surrounding close approach. The theory of Hut
661: (1981) predicts a pseudo-synchronous spin period, $P_ {\rm spin}$ given
662: by
663: \begin{equation}
664: P_{\rm spin}={
665: 1+{15\over{2}}e^{2}+{45\over{8}}e^{4}+{5\over{16}}e^{6}
666: \over{(1+3e^2+{3\over{8}}e^{4})(1-e^{2})^{3/2}}}P_{\rm orbit}=91.3\,
667: {\rm hr}
668: \end{equation}
669: Knowledge of $P_{\rm spin}$, $M_{\rm p}$, $R_{\rm p}$, and the
670: time-dependent pattern of received stellar flux make it possible to
671: compute global climate models for the planet (e.g. \citealt{sg2002};
672: \citealt{cho2003}; \citealt{burkert2005}; \citealt{cs2005,cs2006};
673: \citealt{ll2007}; \citealt{fortney2007b}; \citealt{ddl2007}). These
674: models, in turn, can be used to obtain predictions of the planet's
675: infrared photometric light curve at various wavelengths. Model light
676: curves can then be compared with observations using the {\it Spitzer
677: Space Telescope} (see, e.g., \citealt{char2005};
678: \citealt{deming2005}; \citealt{har2006};
679: \citealt{knut2007a,knut2007b}; \citealt{cow2007}).
680:
681: We adopt the climate model of \citet{ll2007} and apply it to
682: HD~17156b, using the known orbital and physical properties of the
683: planet. The model uses a compressible 2D hydrodynamical solver
684: \citep{ada1999} with a one-layer, two-frequency radiative transfer
685: scheme. It is important to emphasize that our hydrodynamical model
686: (like those of other workers in the field) is undergoing rapid
687: development. Indeed, a primary goal for obtaining physical
688: observations of planets under time-varying irradiation conditions is
689: to provide physical guidance to the numerical codes.
690:
691: Our model predictions are shown in Figure~\ref{ir}, which, for reference,
692: includes a model of the planet, the orbit, and the star plotted to
693: scale (top left). The hydrodynamical model assumes a solar-composition H--He
694: atmosphere for the planet, and is run for several orbits to achieve
695: equilibrium, at which point we track the predicted $8\mu$m light curve
696: for a full orbital period starting at apastron, by assuming a black
697: body spectrum for each surface element in the model, and integrating
698: the resulting $8\mu$m flux over the visible hemisphere of the planet.
699: The right column of Figure~\ref{ir} shows a series of five global
700: temperature plots, each corresponding to the thermal appearance of the
701: planet from the Earth. The figures are equally spaced in time by
702: one-quarter of an orbit (127.3 hours).
703:
704: %\clearpage
705: \begin{figure}
706: %\epsscale{0.8} % AASTeX
707: \epsscale{1.1}
708: \plotone{f4.eps}
709: \caption{{\it Upper left hand panel:} The orbital geometry of
710: HD~17156b. The small dots show the position of the planet at 2.4 hour
711: intervals throughout the 21.217~day orbit, and show the planet and
712: star drawn to scale ($R_{\star}=1.2 \, R_{\odot}$). The locations
713: of the planet at successive time intervals equal to one quarter of an
714: orbit (1-5) are marked. The small orange spheres indicate the 91.3~hour
715: predicted spin frequency of the planet, which is expected to be
716: in pseudo synchronous rotation.
717: {\it Right hand panels:} Temperature maps of the planet at locations 1-5,
718: as viewed from the Earth. The temperature scale is mapped to a black body
719: color map that approximates the planet's visual appearance in its own
720: intrinsic radiation.
721: {\it Lower left panel:} Predicted $8~{\mu}$m light
722: curve for the planet during the course of one planetary orbit.}
723: \label{ir}
724: \end{figure}
725: %\clearpage
726:
727: The predicted 8$\mu$m flux from the planet is shown in the lower-left
728: corner of the figure, and features a rapid rise from a baseline of
729: $F_{\rm p}/ F_{\star} \sim 1.7\times 10^{-4}$ to $F_{\rm p}/F_{\star}
730: \sim 5\times 10^{-4}$ during the $\sim 30$ hour interval following
731: periastron. The magnitude of this rise is comparable with (but
732: slightly less than) that expected for HAT-P-2b and HD~80606b (Langton
733: \& Laughlin 2007).
734:
735: \section{Concluding remarks}
736: \label{summ_sect}
737:
738: HD~17156 thus represents a highly interesting system for follow-up
739: studies in the mid infra-red (e.g. using the {\it Spitzer Space
740: Telescope}), even if it does not undergo secondary eclipses.
741: Ground-based observations should also be pursued to continue the
742: search for transit timing variations, and hence to constrain the
743: presence of additional bodies in the system, which, given the high
744: orbital eccentricity of the planet HD~17156b, may show interesting
745: dynamical properties and place constraints on planetary formation and
746: evolution scenarios. This discovery represents an important success
747: for the Transitsearch.org project, and highlights the potential
748: rewards of collaboration between distributed networks of amateur
749: astronomers and the professional community.
750:
751: \acknowledgments We thank Mauro Barbieri and the transit discovery
752: team for alerting us to the October transit opportunity, and Genya
753: Takeda for discussions regarding the spectroscopic determination of
754: stellar properties. DC gratefully acknowledges funding from the David
755: and Lucile Packard Foundation Fellowship for Science and Engineering.
756: WFW acknowledges support from Research Corporation. Observations
757: obtained at MLO made use of the High Performance Wireless Research and
758: Education Network sponsored by the NSF ANIR division under grant
759: ANI-0087344 and the University of California San Diego. This research
760: has made use of NASA's Astrophysics Data System (ADS), the SIMBAD
761: database, operated at CDS, Strasbourg, France, NASA's SkyView and
762: Dr.~John Thorstensen's SkyCalc software.
763:
764: Finally, we thank the anonymous referee for a thorough and detailed
765: report which has helped to improve the paper.
766:
767: \begin{thebibliography}{}
768: \bibitem[\protect\citeauthoryear{Adams \& Swartztrauber}{1999}]{ada1999}
769: Adams, J.C. \& Swartztrauber, P. N. 1999, Monthly Weather Review, 127, 1872
770: \bibitem[\protect\citeauthoryear{Agol et al.}{2005}]{agol2005}
771: Agol, E., Steffen, J., Sari, R., \& Clarkson, W.\ 2005, \mnras, 359, 567
772: \bibitem[\protect\citeauthoryear{Aigrain et al.}{2007}]{aigrain2007}
773: Aigrain, S., Hodgkin, S., Irwin, J., Hebb, L., Irwin, M., Favata,
774: F., Moraux, E. \& Pont, F. 2007, \mnras, 375, 29
775: \bibitem[\protect\citeauthoryear{Bakos et al.}{2007}]{bakos2007}
776: Bakos, G. \'{A}., et al.\ 2007, \apj, 670, 826
777: \bibitem[\protect\citeauthoryear{Barbieri et al.}{2007}]{bar2007}
778: Barbieri, M., et al.\ 2007, \aap, 476, L13
779: \bibitem[\protect\citeauthoryear{Bodenheimer et al.}{2003}]{boden2003}
780: Bodenheimer, P., Laughlin, G. \& Lin, D. N. C. 2003, ApJ, 592, 555
781: \bibitem[\protect\citeauthoryear{Burkert et al.}{2005}]{burkert2005}
782: Burkert, A., Lin, D. N. C., Bodenheimer, P. H., Jones, C. A. \& Yorke, H. W. 2005, \apj, 618, 512
783: \bibitem[\protect\citeauthoryear{Burrows et al.}{2007}]{bur2007}
784: Burrows, A., Hubeny, I., Budaj, J. \& Hubbard, W. B. 2007, \apj, 661, 502
785: \bibitem[\protect\citeauthoryear{Charbonneau et al.}{2005}]{char2005}
786: Charbonneau, D., et al.\ 2005, \apj, 626, 523
787: \bibitem[\protect\citeauthoryear{Charbonneau et al.}{2007}]{char2007}
788: Charbonneau, D., Brown, T.~M., Burrows, A., \& Laughlin, G.\ 2007,
789: in Protostars and Planets V, ed. B. Reipurth, D. Jewitt \& K. Keil
790: (Tucson: University of Arizona Press), 701
791: \bibitem[\protect\citeauthoryear{Chatterjee et al.}{2007}]{chat2007}
792: Chatterjee, S., Ford, E.~B., \& Rasio, F.~A.\ 2007, ApJ, submitted (astro-ph/0703166)
793: \bibitem[\protect\citeauthoryear{Cho et al.}{2003}]{cho2003}
794: Cho, J. Y.-K., Menou, K., Hansen, B. M. S. \& Seager, S. 2003, \apjl, 587, L117
795: \bibitem[\protect\citeauthoryear{Claret}{2000}]{claret2000}
796: Claret, A. 2000, A\&A, 363, 1081
797: \bibitem[\protect\citeauthoryear{Claret}{2004}]{claret2004}
798: Claret, A. 2000, A\&A, 363, 1081
799: \bibitem[\protect\citeauthoryear{Collier Cameron et al.}{2007}]{cc2007}
800: Collier Cameron, A., et al.\ 2007, \mnras, 375, 951
801: \bibitem[\protect\citeauthoryear{Cooper \& Showman}{2005}]{cs2005}
802: Cooper, C. S. \& Showman, A. P. 2005, \apj, 629, 45
803: \bibitem[\protect\citeauthoryear{Cooper \& Showman}{2006}]{cs2006}
804: Cooper, C. S. \& Showman, A. P. 2006, \apj, 649, 1048
805: \bibitem[\protect\citeauthoryear{Cowan et al.}{2007}]{cow2007}
806: Cowan, N.~B., Agol, E., \& Charbonneau, D.\ 2007, \mnras, 379, 641
807: \bibitem[\protect\citeauthoryear{Deming et al.}{2005}]{deming2005}
808: Deming, D., Seager, S., Richardson, L.~J., \& Harrington, J.\ 2005, \nat, 434, 740
809: \bibitem[\protect\citeauthoryear{Dobbs-Dixon \& Lin}{2007}]{ddl2007}
810: Dobbs-Dixon, I. \& Lin, D. N. C. 2007, preprint (arXiv:0704.3269)
811: \bibitem[\protect\citeauthoryear{Fabrycky \& Tremaine}{2007}]{fab2007}
812: Fabrycky, D., \& Tremaine, S.\ 2007, \apj, 669, 1298
813: \bibitem[\protect\citeauthoryear{Fischer et al.}{2007}]{fischer2007}
814: Fischer, D.~A., et al.\ 2007, \apj, 669, 1336
815: \bibitem[\protect\citeauthoryear{Fortney et al.}{2007a}]{fortney2007a}
816: Fortney, J. J., Marley, M. S. \& Barnes, J. W. 2007a, \apj, 659, 1661
817: \bibitem[\protect\citeauthoryear{Fortney et al.}{2007b}]{fortney2007b}
818: Fortney, J. J., Cooper, C. S., Showman, A. P., Marley, M. S. \& Freedman, R. S. 2007b, \apj, 652, 746
819: \bibitem[\protect\citeauthoryear{Gaudi \& Winn}{2007}]{gw2007}
820: Gaudi, B. S., Winn, J. N. 2007, ApJ, 655, 550
821: \bibitem[\protect\citeauthoryear{Gillon et al.}{2007}]{gil2007}
822: Gillon, M., Triaud, A.~H.~M.~J., Mayor, M., Queloz, D., Udry, S., \& North, P.\ 2007, A\&A, submitted (arXiv:0712.2073)
823: \bibitem[\protect\citeauthoryear{Groth}{1986}]{groth86}
824: Groth, E. J. 1986, AJ, 91, 1244
825: \bibitem[\protect\citeauthoryear{Harrington et al.}{2006}]{har2006}
826: Harrington, J., Hansen, B.~M., Luszcz, S.~H., Seager, S., Deming, D., Menou, K., Cho, J.~Y.-K., \& Richardson, L.~J.\ 2006, Science, 314, 623
827: \bibitem[\protect\citeauthoryear{Holman \& Murray}{2005}]{hm2005}
828: Holman, M.~J., \& Murray, N.~W.\ 2005, Science, 307, 1288
829: \bibitem[\protect\citeauthoryear{Johns-Krull et al.}{2007}]{jk2007}
830: Johns-Krull, C. M., et al. 2007, ApJ, in press (arXiv:0712.4283)
831: \bibitem[\protect\citeauthoryear{Knutson et al.}{2007a}]{knut2007a}
832: Knutson, H. A., Charbonneau, D., Allen, L. E., Burrows, A., \&
833: Megeath, S. T. 2007a, ApJ, in press (arXiv:0709.3984)
834: \bibitem[\protect\citeauthoryear{Knutson et al.}{2007b}]{knut2007b}
835: Knutson, H.~A., et al. 2007b, \nat, 447, 183
836: \bibitem[\protect\citeauthoryear{Loeillet et al.}{2007}]{loeil2007}
837: Loeillet B., et al. 2007, A\&A, submitted (arXiv:0707.0679)
838: \bibitem[\protect\citeauthoryear{Irwin et al.}{2007}]{i2007a}
839: Irwin, J., Irwin, M., Aigrain, S., Hodgkin, S., Hebb, L., \& Moraux,
840: E. 2007, MNRAS, 375, 1449
841: \bibitem[\protect\citeauthoryear{Langton \& Laughlin}{2007}]{ll2007}
842: Langton, J., \& Laughlin, G.\ 2007, ApJ, in press (arXiv:0711.2106)
843: \bibitem[\protect\citeauthoryear{Mandel \& Agol}{2002}]{ma2002}
844: Mandel, K., \& Agol, E. 2002, ApJ, 580, 171
845: \bibitem[\protect\citeauthoryear{Mandushev et al.}{2007}]{mand2007}
846: Mandushev, G., et al.\ 2007, \apjl, 667, L195
847: \bibitem[\protect\citeauthoryear{Mardling}{2007}]{mard2007}
848: Mardling, R.~A.\ 2007, \mnras, 382, 1768
849: \bibitem[\protect\citeauthoryear{McLaughlin}{1924}]{mclaughlin24}
850: McLaughlin, D. B. 1924, ApJ, 60, 22
851: \bibitem[\protect\citeauthoryear{Narita et al.}{2007}]{nar2007}
852: Narita, N., Sato, B., Ohshima, O., \& Winn, J.~N.\ 2007, PASJ, submitted (arXiv:0712.2569)
853: \bibitem[\protect\citeauthoryear{Press et al.}{1992}]{press92}
854: Press, W. H., Teukolsky, S. A., Vetterling, W. T. \& Flannery, B. P.
855: 1992, Numerical Recipes in Fortran 77: The Art of Scientific
856: Computing (Cambridge, UK: Cambridge University Press)
857: \bibitem[\protect\citeauthoryear{Rossiter}{1924}]{rossiter24}
858: Rossiter, R. A. 1924, ApJ, 60, 15
859: \bibitem[\protect\citeauthoryear{Seagroves et al.}{2003}]{sea2003}
860: Seagroves, S., Harker, J., Laughlin, G., Lacy, J., \& Castellano, T.\ 2003, \pasp, 115, 1355
861: \bibitem[\protect\citeauthoryear{Showman \& Guillot}{2002}]{sg2002}
862: Showman, A. P. \& Guillot, T. 2002, A\&A, 385, 166
863: \bibitem[\protect\citeauthoryear{Steffen \& Agol}{2005}]{sa2005}
864: Steffen, J.~H., \& Agol, E.\ 2005, \mnras, 364, L96
865: \bibitem[\protect\citeauthoryear{Takeda et al.}{2007}]{tak2007}
866: Takeda, G., Ford, E. B., Sills, A., Rasio, F., Fischer, D. A., \&
867: Valenti, J. A. 2007, ApJS, 168, 297
868: \bibitem[\protect\citeauthoryear{Winn}{2007}]{winnasp2007}
869: Winn, J. N. 2007, in ASP Conf. Ser. 366, Transiting Extrapolar
870: Planets Workshop, ed. C. Afonso, D. Weldrake, \& Th. Henning (San
871: Francisco: ASP), 170
872: \bibitem[\protect\citeauthoryear{Winn et al.}{2007}]{winn2007}
873: Winn, J. N., Holman M. J. \& Roussanova A. 2007, ApJ, 657, 1098
874: \bibitem[\protect\citeauthoryear{Wu et al.}{Wu, Murray \& Ramsahai}{2007}]{wu2007}
875: Wu, Y., Murray, N.~W., \& Ramsahai, J.~M.\ 2007, \apj, 670, 820
876: \end{thebibliography}
877:
878: \end{document}
879: