1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2: %Comment out these lines for ApJ submission

3: \documentclass{emulateapj}

4: \lefthead{GAUDI} \righthead{{\it KEPLER} AND KUIPER}

5: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

6: \def\kms{{\rm km\,s^{-1}}}

7: \def\eq#1{equation (\ref{#1})}

8: \def\Eq#1{Eq.~\ref{#1}}

9: \def\km{\rm km}

10: \def\au{\rm AU}

11: \def\kms{\rm km~s^{-1}}

12: \def\tk{t_{\rm K}}

13: \def\rk{R_{\rm K}}

14: \def\rf{R_{\rm F}}

15: \def\rs{R_{*}}

16: \def\thk{\theta_{\rm K}}

17: \def\thf{\theta_{\rm F}}

18: \def\ths{\theta_*}

19: \def\rhof{\rho_{\rm F}}

20: \def\rhos{\rho_{\rm *}}

21: \def\kep{{\it Kepler}}

22: \def\deg{^\circ}

23: \def\rb{r_{\rm b}}

24: \def\mb{m_{\rm b}}

25: \def\muas{\mu{\rm as}}

26: \def\atc{\left<t_c\right>}

27: \def\qmin{Q_{\rm min}}

28: \def\texp{t_{\rm exp}}

29: \def\ndet{N_{\rm det}}

30: 

31: \begin{document}

32: 

33: \title{{\it Kepler} and the Kuiper Belt}

34: \author 

35: {B.\ Scott Gaudi}

36: \affil{Harvard-Smithsonian Center for Astrophysics, 60 Garden St., Cambridge, MA 02138}

37: 

38: \begin{abstract}

39: The proposed field-of-view of the {\it Kepler} mission is at an

40: ecliptic latitude of $\sim55^\circ$, where the surface density of

41: scattered Kuiper Belt Objects (KBOs) is a few percent that in the

42: ecliptic plane.  The rate of occultations of \kep\ target stars by

43: scattered KBOs with radii $r\ga10~\km$ is $\sim 10^{-6}-10^{-4}~{\rm

44: star^{-1}}~{\rm yr^{-1}}$, where the uncertainty reflects the current

45: ignorance of the thickness of the scattered KBO disk and the faint-end

46: slope of their magnitude distribution.  These occultation events will

47: last only $\sim 0.1\%$ of the planned $\texp=15$ minute integration

48: time, and thus will appear as single data points that deviate by tiny

49: amounts.  However, given the target photometric accuracy of {\it

50: Kepler}, these deviations will nevertheless be highly significant,

51: with typical signal-to-noise ratios of $\sim 10$.  I estimate that

52: $\sim 1-20$ of the $10^5$ main-sequence stars in {\it Kepler}'s

53: field-of-view will exhibit detectable occultations during its

54: four-year mission.  For unresolved events, 

55: the signal-to-noise of individual occultations

56: scales as $\texp^{-1/2}$, and the minimum detectable

57: radius could be

58: decreased by an order of magnitude to $\sim 1~\km$ by

59: searching the individual $3~{\rm sec}$ readouts for occultations.  

60: I propose a

61: number of methods by which occultation events may be differentiated

62: from systematic effects. {\it Kepler} should measure or significantly

63: constrain the frequency of highly-inclined, $\sim 10~\km$-sized KBOs.

64: \end{abstract}

65: \keywords{Kuiper Belt -- occultations -- solar system: formation -- techniques: photometric}

66: \section{Introduction\label{sec:intro}}

67: 

68: The Kuiper belt (see

69: \citealt{lj02} for a review) constitutes one of the few fossil records of the

70: formation of planets and planetesimals in the early solar system.  It

71: is therefore of great interest to determine the ensemble properties of

72: its denizens.  These properties provide clues to the physical

73: processes operating in the protoplanetary solar disk, as well as in

74: the subsequent evolution of the solar system.

75: 

76: Kuiper Belt Objects (KBOs) are currently discovered in deep optical

77: imaging surveys that are primarily limited to latitudes within $\sim

78: 10\deg$ of the ecliptic or invariable plane (see, e.g., \citealt{millis02}).  

79: The apparent $R$-magnitude of

80: a KBO at a distance of $a\simeq 40~\au$ with radius $r \simeq 100~\km$

81: is $m_R\simeq 23.4$, assuming an albedo of $4\%$.  Since the flux of a

82: KBO arises from reflected sunlight, it is proportional to $r^{2}$ and

83: $a^{-4}$.  Therefore, imaging surveys are only able to detect relatively large

84: and nearby KBOs with reasonable expenditure of resources.

85: Furthermore, in order to have the best chance of success, they do

86: not target high ecliptic latitudes, and are therefore biased against

87: KBOs with large inclinations $i$.

88: 

89: Large KBOs are known to fall into at least three dynamical classes.

90: Classical KBOs are relatively dynamically cold, with eccentricities

91: $e\la 0.3$ and $i\la 5\deg$.  Resonant KBOs are in mean

92: motion resonances with Neptune, and were likely trapped there

93: and subsequently excited to high $i$ and $e$ by the outward migration of

94: Neptune \citep{malhotra95}.  Scattered KBOs have very broad $e$ and $i$

95: distributions, and have probably been excited by some as-yet unknown process.

96: 

97: The size distribution of large KBOs is known to follow the form of

98: $r^{-q}$, with $q\simeq 4$, and a surface density of objects of

99: $1~{\rm deg^{-2}}$ at $r\simeq 100~\km$.  For $r\la 100~\km$,

100: collisions will preferentially destroy KBOs, so that below a certain

101: radius $r<\rb\la 100~\km$, the size distribution is expected to

102: flatten.  This break radius is set by the dynamical and physical

103: properties of the KBOs, and is roughly where the collision time is

104: equal to the age of the system.  The first estimate of this break

105: radius, based on deep Hubble Space Telescope (HST) observations, yields

106: $\rb \sim 70~\km$ (\citealt{bern04}, hereafter B04).

107: 

108: Unfortunately, it will be difficult, in the near future, to learn

109: about the properties of small KBOs beyond what is already known using

110: direct imaging, and especially so for those with high inclinations.

111: Further progress will require next-generation synoptic surveys

112: (Pan-STARRS, \citealt{kaiser02}; The Discovery Channel Telescope, \citealt{mds03}; LSST, \citealt{tyson02}), or new methods.  One such method,

113: first proposed by \citet{bailey76}, is to search for stellar

114: occultations by foreground KBOs.  This technique has been explored by

115: several authors \citep{dyson92,bw97,cf04}, and is being implemented by

116: \citet{roques03}, and

117: the Taiwanese-American Occultation Survey (TAOS; \citealt{liang03}).

118: However, due to the low expected event rate, the first occultation

119: surveys will likely focus on low-latitude fields.  Therefore, the

120: statistics of small, high-inclination KBOs will remain meager.

121: 

122: Here I explore the possibility of measuring the surface density of

123: small KBOs with high inclinations using the NASA space mission \kep.

124: \kep's main goal is to detect transiting planets around normal stars,

125: and it is far from ideally suited to detect occultations by foreground

126: KBOs, especially given the high latitude ($55\deg$) of its target

127: field, and its long ($15~{\rm min}$) exposure time.  

128: However, due to the large number of target stars and

129: extraordinary photometric precision, I demonstrate that it may

130: nonetheless detect a significant number of occultation events.  If

131: these can be separated from systematic artifacts, {\it Kepler} should

132: be able to determine or severely constrain the frequency of small,

133: high-inclination KBOs.

134: 

135: \clearpage

136: 

137: \section{Order-of-Magnitude Estimates\label{sec:oom}}

138: 

139: I begin by providing order-of-magnitude estimates of the rate and

140: properties of KBO occultation events in \kep's field-of-view (FOV).

141: Assuming all KBOs are at the same distance, the optical depth to occultations is roughly the differential

142: angular surface density $\Sigma_r(r)={\rm d}N(r)/{\rm d}r{\rm d}\Omega$

143: of KBOs of radius $r$ in the \kep\ FOV, times the solid angle

144: $\Omega(r)$ subtended by a KBO of radius $r$, integrated over all

145: radii,

146: \begin{equation}

147: \tau \sim \int {\rm d}r \Omega(r) \Sigma_r(r).

148: \label{eqn:tausimple}

149: \end{equation}

150: Although the optical depth to KBOs at a constant distance is most

151: directly related to their size distribution, optical surveys do not

152: directly constrain this quantity.  Rather, they constrain

153: the distribution of apparent magnitudes.  I therefore switch

154: the integration variable in \eq{eqn:tausimple} to apparent 

155: $R$-magnitude $m_R$,

156: \begin{equation}

157: \tau \sim \int {\rm d}m_R \Omega(m_R) \Sigma(m_R).

158: \label{eqn:tausimplem}

159: \end{equation}

160: Assuming constant albedo and

161: distance, the radius of a KBO is related to its $R$-magnitude by

162: \begin{equation}

163: r(m_R)=r_{23}10^{-0.2(m_R-23)}.

164: \label{eqn:radmag}

165: \end{equation}

166: where $r_{23}$ is the radius of a KBO at $m_R=23$. 

167: The measured differential surface density $\Sigma(r)={\rm d}N(r)/{\rm d}m_R{\rm d}\Omega$ of all bright KBOs as a function of

168: apparent magnitude near the ecliptic has the form

169: $\Sigma(m_R)=\Sigma_{23}10^{\alpha(m_R-23)}$, where

170: $\alpha$ can be related to the power-law index of the

171: size distribution by $q=5\alpha+1$.  Assuming that the

172: latitude distribution of scattered KBOs is a Gaussian with a

173: standard deviation $\sigma_\beta$, and that a fraction $f_{\rm SK}$ of KBOs

174: detected near the ecliptic plane are scattered KBOs,

175: \begin{equation}

176: \tau  \sim f_{\rm SK} \pi \theta_{23}^2 \Sigma_{23}

177: e^{\frac{-\beta^2}{2\sigma_\beta^2}}\int {\rm d}m_R

178: 10^{(\alpha-0.4)(m_R-23)}.

179: \label{eqn:tau1}

180: \end{equation}

181: Here $\theta_{23}\equiv r_{23}/a$ is the angular size of a KBO with

182: $m_R=23$.

183: 

184: Assuming that objects at the break magnitude

185: produce occultations above the detection limit, 

186: there are two regimes to consider.  If the slope of the magnitude

187: distribution above the break magnitude $m_R>\mb$ ($r<\rb$) is

188: $\alpha_2 \la 0.4$, then the integral in \eq{eqn:tau1} is dominated by

189: objects near $m_b$, and is 

190: $\sim \Delta m_R 10^{(\alpha-0.4)(m_b-23)}$, where $\Delta m_R\simeq 

191: 0.6|\alpha-0.4|^{-1}\simeq 3$ is roughly the full-width half-max of the

192: distribution.  In the other regime, if $\alpha_2\ge 0.4$, then the

193: integral in \eq{eqn:tau1} formally diverges.  In practice, the

194: lower limit in radius, or 

195: upper limit in magnitude, $m_{\rm max}$, will then be set by

196: where the occultation is too brief or too shallow to be detectable.

197: As I demonstrate in \S\ref{sec:rate}, the former generally turns out to be satisfied

198: first. 

199: In this regime, the integral in \eq{eqn:tau1} is dominated by

200: objects with $m_R\sim m_{\rm max}$, and thus is $\sim \Delta m_R

201: 10^{(\alpha-\alpha_2)(\mb-23)+(\alpha_2-0.4)(m_{\rm max}-23)}$. 

202: 

203: A simple boxcar occultation of duration $2t_c$ has a 

204: signal-to-noise $Q$ of,

205: \begin{equation}

206: Q={2\atc \gamma^{1/2}}{t_{\rm exp}^{-1/2}},

207: \label{eqn:sstn}

208: \end{equation}

209: where $\gamma=2.46 \times 10^{5}~{\rm s^{-1}}$ is the rate at which

210: photons will be collected by \kep\ from a star with $m_*=12$.  This

211: expression holds when the occultation duration is less than the

212: exposure time.  For circular KBOs and equatorial occultations, 

213: the duration is simply $t_c=\thk/\mu$, 

214: where $\thk$ is the angular

215: radius of the KBO, $\mu=v/a$ is the proper motion of the KBO,

216: and $v\sim v_\oplus =30~\kms$ is the relative velocity of the KBO,

217: which is primarily due to the reflex motion of the

218: Earth.  This yields $t_c \simeq 4~{\rm s}~10^{-0.2(m_R-23)}$.

219: For objects near the break of $m_b=26$,

220: $t_c\simeq 1~{\rm s}$, and for the planned \kep\ exposure time of $\texp=15~{\rm min}$, a typical target star with apparent magnitude $m_*=14$ yields

221: $Q\sim 13$.  The minimum acceptable signal-to-noise of $Q\sim 7$

222: (justified in \S\ref{sec:rate}) is reached at $m_{\rm max}\simeq 27.4$, 

223: where $r\simeq 16~{\rm km}$, and $t_c\simeq 0.5~{\rm s}$.

224: 

225: Adopting fiducial values for the various input parameters, I now

226: estimate the optical depth in the two regimes.  Surveys of bright KBOs

227: find $\alpha\simeq 0.6$ and $\Sigma_{23}\sim 1~{\rm

228: deg^{-1}~mag^{-1}}$ (e.g., \citealt{lj02}).  The properties of 

229: the scattered KBO

230: population are poorly known due to severe selection biases, and

231: estimates of these properties should therefore be taken with caution.

232: I adopt $\sigma_i =20\deg$ (\citealt{brown01}, hereafter B01; 

233: \citealt{tjl01}), $f_{\rm SK}\sim 0.1$, and 

234: a common distance of $a=42~\au$.  Assuming a

235: 4\% albedo, this yields $r_{23}\simeq 123~\km$, and 

236: $\theta_{23}\simeq 4~{\rm mas}$.   For classical KBOs, the

237: break radius is expected to occur near $m_b\sim 24$ \citep{ps04}, as is

238: observed (B04).  However, scattered KBOs may have longer collision

239: times, and thus the break radius may be smaller.  I adopt

240: $m_b\simeq 26$, as B04 find for their `excited' subsample.

241: Finally, the FOV of \kep\ is at $\beta\simeq 55\deg$. Altogether, this yields

242: $\tau_{\rm br} \sim 10^{-13}$

243: for the break-limited regime, and

244: $\tau_{\rm sn} \sim 3 \times 10^{-13}10^{1.4\alpha_2}$

245: for the signal-to-noise limited regime.

246: 

247: The event rate $\Gamma$ is related to the optical depth by,

248: \begin{equation}

249: \Gamma = \frac{2\tau}{\pi\atc},

250: \label{eqn:erate}

251: \end{equation}

252: where $\atc$ is the mean time scale of the occultation events. 

253: Assuming that $\atc$ is equal to the time scale at the peak of

254: the size distribution, this yields $\Gamma_{\rm br} \simeq 2\times 10^{-6}~{\rm yr^{-1}}$

255: for the break-limited regime, and

256: $\Gamma_{\rm sn} \simeq 10^{-6}~{\rm yr^{-1}} 10^{1.4\alpha_2}$

257: for the signal-to-noise limited regime.

258: 

259: \kep\ will have

260: $N_*\sim 10^5$ main-sequence stars in its field-of-view, and will have

261: a lifetime of $T=4$ years, so the number of expected events is $N_{\rm

262: det}=N_*\Gamma_{\rm br} T \simeq 0.8$ for the break-limited regime.  

263: For the signal-to-noise limited regime, and 

264: a faint-end slope of $\alpha_2=0.6$, $\Gamma_{\rm sn}\simeq 7.7\times 10^{-6}~{\rm yr^{-1}}$, and the number of expected events is 

265: $N_{\rm det}=N_*\Gamma_{\rm sn} T \simeq 3$.  

266: 

267: 

268: \section{Evaluation of Event Rate}\label{sec:rate}

269: 

270: I now provide a somewhat more accurate estimation of the occultation

271: rate in \kep's FOV, taking into account the distribution of KBO radii,

272: inclinations, and velocities; the number, magnitude, and angular size

273: distribution of the target stars; and the suppression of

274: signal-to-noise in occultation events due to diffraction and finite

275: source size effects.  Readers who are not interested in these details

276: can skip to \S\ref{sec:discussion} without significant loss of continuity.

277: I will continue to make a number of simplifying

278: assumptions, including that all KBOs are located at the same distance

279: of $a=42~\au$.  

280: 

281: The observable optical depth to a star of a given magnitude $m_*$ and

282: angular size $\ths$ is

283: \begin{equation}

284: \tau(m_*,\ths)=\int {\rm d}m_{R} \Sigma(\beta,m_{R})\Omega(m_{R})

285: \Theta(Q-\qmin),

286: \label{eqn:tauacc}

287: \end{equation}

288: where $\beta$ is the ecliptic latitude of the FOV, $\qmin$ is the

289: minimum detectable signal-to-noise, and $\Theta(x)$ is the Heaviside

290: step function.

291: 

292: The surface density $\Sigma(\beta,m_{R})$ of KBOs of a given $m_R$ at

293: an ecliptic latitude $\beta$ can be written as,

294: $\Sigma(\beta,m_{R})=\Sigma(0,m_{R})f_\Sigma(\beta)$, where

295: $f_\Sigma(\beta)$ is the surface density distribution of KBOs

296: normalized to the ecliptic plane.  For a population with an observed

297: inclination distribution at the ecliptic plane of $f_e(i)$, this is

298: given by (B01),

299: \begin{equation}

300: f_\Sigma(\beta)= \int_\beta^{\pi/2} {\rm d}i \frac{f_e(i)\sin

301: i}{(\sin^2i-\sin^2\beta)^{1/2}}\times\left[ \int_0^{\pi/2}{\rm d}i

302: f_e(i)\right]^{-1}.

303: \label{eqn:ldist}

304: \end{equation}

305: An inclination distribution of the form

306: $f_e(i)=\exp(-i^2/2\sigma_i^2)$, with a standard deviation $\sigma_i=20\deg

307: \pm 4\deg$, provides a reasonable fit to the observed distribution of

308: scattered KBOs (B01; \citealt{tjl01,millis02}). For this form, the denominator of \eq{eqn:ldist} is

309: equal to $(\pi/2)^{1/2}\sigma_i {\rm Erf}(\pi/2\sqrt{2}\sigma_i)$.  The

310: \kep\ FOV extends over $10\deg$ in ecliptic latitude.  Since

311: the surface density varies by a factor of $\sim 2$ over this range

312: for reasonable values of $\sigma_i$, I average $f_\Sigma(\beta)$  over $50\deg\le \beta \le

313: 60\deg$, assuming a uniform distribution in $\beta$.

314: 

315: 

316: For unresolved events ($2t_c \le \texp$), I write the 

317: signal-to-noise $Q$ of an occultation as,

318: \begin{equation}

319: Q={2 t_c\gamma^{1/2} }{t_{\rm

320: exp}^{-1/2}}f_{u_0}f_{Q}10^{-0.2(m_*-12)},

321: \label{eqn:snexp}

322: \end{equation}

323: whereas, for resolved events ($2t_c \ge \texp$),

324: \begin{equation}

325: Q=(2t_c\gamma)^{1/2} f_{u_0}f_{Q} 10^{-0.2(m_*-12)},

326: \label{eqn:snexpres}

327: \end{equation}

328: Where, as before, $t_{\rm exp}$ is the exposure time and

329: $\gamma$ is the photon collection rate at $m_*=12$.  The term 

330: $f_{u_0}$ is the reduction in signal-to-noise due to

331: non-equatorial occultations, which is

332: $f_{u_0}=\int_0^1 {\rm d}u_0(1-u_0^2)^{1/2}=0.79$ for unresolved events,

333: and $f_{u_0}=\int_0^1 {\rm d}u_0(1-u_0^2)^{1/4}=0.87$ for resolved

334: events.

335: 

336: Significant

337: suppression of the signal-to-noise due to diffraction and finite-source

338: effects is expected 

339: whenever the angular radius $\thk$ of the KBO is

340: comparable to, or smaller than, either the angular size of the star

341: $\ths$, or the Fresnel angle $\thf=\sqrt{\lambda/2\pi a}$, where

342: $\lambda$ is the wavelength of observations.  

343: I determine $f_Q$ by numerically calculating occultation light curves

344: for various values of $\rhof\equiv \thf/\thk$ and $\rhos\equiv\ths/\thk$, integrating over a bandpass

345: centered at $\lambda=500~{\rm nm}$ with a full width of $25\%$.  I

346: then determine the suppression in signal-to-noise of these curves relative to a

347: boxcar complete occultation with duration $2t_c$.  

348: For unresolved events, for which diffraction and finite-source effects

349: may be important, this can be approximated as

350: \begin{equation}

351: f_Q=\frac{1}{4}

352: {\rm Erfc}\left(\frac{\log{\rhos}-0.52}{0.42}\right)

353: {\rm Erfc}\left(\frac{\log{\rhof}}{0.65}\right).

354: \label{eqn:fq}

355: \end{equation}

356: For a mid-F dwarf at $\sim 1~{\rm kpc}$ (typical of Kepler target

357: stars), $\ths \simeq 6~\muas$, while for $\lambda =500~{\rm nm}$ and

358: $a=42~\au$, $\thf\simeq 23~\muas$.  Therefore the diffraction effects

359: are typically expected to be more important than the finite size of

360: the stars.  The radius where $\thk=\thf$ is $r \sim 1~\km$, or $m_R

361: \sim 34$; the radius where $\thk=\ths$ for the example above is $r\sim

362: 0.26~{\rm km}$, or $m_R \sim 37$.  As both of these magnitudes are

363: considerably fainter than the expected

364: cutoff in $Q$ due to photon noise for the nominal \kep\ exposure time,

365: the corrections due to diffraction and finite source effects are

366: small in this case.

367: 

368: The final ingredient in calculating $Q$ is the radius crossing

369: time $t_c$.  I determine the relative velocity $v$ of a KBO at a given

370: ecliptic longitude and latitude by $v=|{\bf v}_{\perp,\oplus}-{\bf

371: v}_K|$, where ${\bf v}_{\perp,\oplus} ={\bf v}_\oplus- ({\bf v_\oplus}

372: \cdot {\hat {\bf n}}) {\hat {\bf n}}$, ${\hat {\bf n}}$ is the unit

373: line-of-sight vector to the KBO, and $v_K\equiv v_\oplus(a/\au)^{-1/2}$.  I assume circular orbits for the

374: Earth and KBO, and that the KBO is observed at a latitude equal to its

375: inclination.  I then average the proper motion $\mu=v/a$ over

376: ecliptic longitude.  This yields $\left<\mu\right>=3.19''~{\rm

377: hr^{-1}}$, and $t_c=3.70~{\rm s}$ for $r=100~\km$.

378: 

379: The average radius crossing time $\atc$ is given by,

380: \begin{equation}

381: \left<t_c(m_*,\ths)\right>=\int {\rm d}m_{R} \Sigma(\beta,m_{R})

382: \frac{\thk(m_{R})}{\left<\mu\right>} \Theta(Q-\qmin)

383: \label{eqn:atc}

384: \end{equation}

385: $$

386: \times \left[\int {\rm d}m_{R} \Sigma(\beta,m_{R}) \Theta(Q-\qmin)\right]^{-1}.

387: $$

388: This can be combined with the optical depth $\tau(m_*,\ths)$ to yield

389: the event rate,

390: \begin{equation}

391: \Gamma(m_*,\ths)=\frac{2\tau(m_*,\ths)}{\pi\left<t_c(m_*,\ths)\right>}.

392: \label{eqn:erateacc}

393: \end{equation}

394: Finally, the total number of detected occultations can be found by

395: integrating over the magnitude distribution ${\rm d}N_*/{\rm d}m_*$

396: and angular size distribution ${\rm d}N_*/{\rm d}\ths$ of the stars in

397: \kep's FOV,

398: \begin{equation}

399: \ndet= T \int {\rm d}m_* \frac{{\rm d}N_*}{{\rm d}m_*} \int {\rm

400: d}\ths \frac{{\rm d}N_*}{{\rm d}\ths} \Gamma(m_*,\ths).

401: \label{eqn:rate}

402: \end{equation}

403: I adopt the joint apparent magnitude -- spectral type distribution

404: derived by \citet{jd03} and presented in their Table 1; this

405: distribution is in turn based on the Besan\c{c}on Galactic model

406: \citep{robin03}, assuming an extinction of $1~{\rm mag~kpc^{-1}}$.

407: Adopting the relations between absolute magnitude, physical radii, and

408: spectral type for main-sequence sequence stars from \citet{allen76},

409: the corresponding distribution of angular radii can also be derived.

410: 

411: \begin{figure}

412: \epsscale{1.7} 

413: \plotone{f1.eps}

414: \vskip-0.3in

415: \caption{\label{fig:one} 

416: The contours show the number of expected occultation events

417: with signal-to-noise $>7$ in \kep's field-of-view, as a function of

418: the standard deviation $\sigma_i$ of the inclination distribution in degrees,

419: and the faint-end slope $\alpha_2$ of the surface density versus

420: magnitude relation.  The shaded region shows the approximate $95\%$

421: confidence limits on these parameters.  Each panel shows the number of events

422: for a different exposure time.  (a) Nominal \kep\ exposure time of

423: $\texp=15\min$. (b) Exposure time equal to the readout time

424: of 3 seconds.}

425: \end{figure}

426: 

427: For the KBO magnitude distribution $\Sigma(0,m_{R})$ near the ecliptic

428: plane, I use the results from B04.  They compile data from several

429: surveys, including their own HST results, and

430: fit the surface density distribution to a double power-law form,

431: \begin{equation}

432: \Sigma(m_R)=(1+c)\Sigma_{23}\left[10^{-\alpha_1(m_R-23)}+c10^{-\alpha_2(m_R-23)}\right]^{-1},

433: \label{eqn:sigmadouble}

434: \end{equation}

435: where $c\equiv 10^{(\alpha_1-\alpha_2)(m_b-23)}$.

436: For their `excited' subsample, which has

437: $i>5\deg$, and $d<28\au$ or $d>55\au$, they find $\alpha_1=0.66$, $m_b=26.0$, and

438: $\Sigma_{23}=0.39~{\rm mag^{-1}~deg^{-2}}$.  I estimate

439: that $\sim 1/3$ of these are scattered KBOs based on the

440: relative number of scattered to resonant KBOs in the B01 analysis.

441: The faint-end slope

442: $\alpha_2$ is poorly constrained for this sample, but is bounded at

443: $68\%$c.l.\ by $\alpha_2\la 0.1$ and $95\%$c.l.\ by $\alpha_2<0.4$.

444: 

445: To calculate the rate of detected events, I adopt a signal-to-noise

446: threshold of $\qmin=7$.  The total number of flux measurements made

447: by \kep\ over its lifetime will be $\sim N_* T

448: t_{\rm exp}^{-1} \sim 4\times 10^9$, therefore requiring a false alarm

449: probability of $\la 10^{-10}$.  One expects $\la 1$ false alarm

450: for $Q>7$.  At the break magnitude $m_{\rm eq}$, the

451: signal-to-noise is $Q\ga 10$ for $\texp=15~{\rm min}$.  

452: Therefore, the results are not

453: sensitive to the choice of $\qmin$ for $\qmin \la 10$, as long as

454: objects near the break magnitude dominate the integral of the

455: magnitude distribution, i.e.\ as long as $\alpha_2\la 0.4$.

456: 

457: Figure \ref{fig:one}(a) shows number of detected occultation events in

458: \kep's FOV, as a function of the dispersion in the inclination

459: distribution $\sigma_i$, and the faint-end slope $\alpha_2$.  The

460: change in the shape of the contours near $\alpha_2\sim 0.4$ is due to

461: the transition from the break limited to signal-to-noise limited

462: regimes described in the previous section.  For $\alpha_2\la 0.4$, the

463: event rate depends mainly on $\sigma_i$, whereas for $\alpha_2\ga 0.4$, it

464: depends primarily on $\alpha_2$.  Very few events are expected for

465: $\sigma_i\la 10\deg$ regardless of $\alpha_2$.  This is due to the

466: exponential suppression of the number of KBOs in the \kep\ FOV for the

467: assumed Gaussian distribution of inclinations.  For the fiducial value

468: of $\sigma_i=20\deg$, and $-1.0 \la \alpha_2\la 0.2$, the number of events is $\sim 2$.

469: This is a factor of a few larger than the rough estimate in the

470: previous section for the break-dominated regime.  This difference

471: can be attributed to differences in the assumed number of target stars, 

472: surface density at the ecliptic, and suppression at the latitude

473: of the \kep\ FOV.  The event rate

474: increases steeply for increasing $\alpha_2\ga 0.4$.  For $\alpha_2=0.6$,

475: $\ndet\simeq 8~{\rm yr^{-1}}$, again a factor of a few larger than

476: the rough estimate. 

477: 

478: 

479: \section{Discussion}\label{sec:discussion}

480: 

481: The shaded area in Figure \ref{fig:one} shows roughly the 95\%

482: confidence limit on the standard deviation $\sigma_i$ of the inclination

483: distribution from B01, and the faint-end slope $\alpha_2$

484: from B04.  In this region, the predicted number of events

485: varies from $\ll 1$ to nearly $20$.  Marginalizing $\ndet$ over

486: $\sigma_i$ and $\alpha_2$, assuming Gaussian uncertainty distributions

487: in both parameters with mean and dispersion of $20\deg$ and $4\deg$ for

488: $\sigma_i$, and $-0.5$ and $0.6$ for $\alpha_2$, yields a best

489: estimate of the number of expected events of $\left< \ndet \right>\sim 3$.  If no events

490: are detected during the mission lifetime, a 95\% confidence upper

491: limit can be placed that should correspond to the $\ndet=3$ contour,

492: which would exclude roughly half the currently allowed region of

493: $\sigma_i$ and $\alpha_2$.

494: 

495: Although the expected number of events in the 95\% c.l.\ region of

496: $\sigma_i$ and $\alpha_2$ is relatively small, it is important to keep

497: in mind that these parameters might be considerably in error, as all

498: previous surveys have had relatively strong selection biases against

499: scattered KBOs.  In particular, the constraint on $\alpha_2$ from B04

500: is applicable to their `excited' sample, the majority of which are

501: probably resonant KBOs.  

502: Therefore, the faint-end slope of scattered KBOs may differ

503: considerably from that measured by B04; this might even be expected if

504: scattered KBOs were excited early in the formation of the solar

505: system.  For $\alpha_2\ga 0.4$, the number of expected

506: events can be quite substantial.  In addition

507: to uncertainties in the properties of the known scattered KBO population, there may be additional populations of KBOs at larger distances that are

508: currently undetectable, but may be detected or constrained by \kep.

509: 

510: For unresolved events, the signal-to-noise of occultations scales as $\texp^{-1/2}$.  Therefore, the signal-to-noise of individual events

511: can be increased by simply decreasing the \kep\ exposure time.  

512: In the break-limited regime ($\alpha_2 \la 0.4$), where small

513: objects with $m_R>m_b$ do not contribute much to the optical depth, this does not

514: significantly increase 

515: the number of detectable events, although it does improve the reliability

516: of already detectable occultations with $m_R\la m_b$.

517: In the signal-to-noise limited regime $\alpha_2 \ga 0.4$, decreasing

518: the exposure time can increase the number of detectable occultations

519: dramatically.   As currently planned, \kep\ will only transmit 

520: integrations of $\texp=15~{\rm min}$, due primarily to bandwidth limitations. 

521: However, the detectors will in fact be read out every 3 seconds in 

522: order to avoid saturation by bright stars.  These $3~{\rm sec}$ readouts

523: will then be filtered and combined into $15~{\rm min}$ integrations.  

524: It may be possible, using in-flight software, to search for occultations in 

525: the individual $3~{\rm sec}$ exposures, and flag and transmit 

526: significant events.   With this possibility in mind, in 

527: Figure \ref{fig:one}(b) I show $\ndet$ as a function

528: of $\sigma_i$ and $\alpha_2$ for $\texp=3~{\rm sec}$.  As expected,

529: deep in the break-limited regime ($\alpha_2\la 0$), 

530: the number of detectable occultations is similar 

531: for $\texp=3~{\rm sec}$ and $\texp=15~{\rm min}$.

532: However, for $\alpha_2\ga 0$, the number of expected event increases

533: significantly for $\texp=3~{\rm sec}$.  In the

534: allowed region of parameter space, the number of events can be as large as $\ndet\sim

535: 600$.  Marginalizing over $\sigma_i$ and $\alpha_2$ yields $\left<

536: \ndet \right> \sim 12$.  For $\alpha_2=0.6$, and $\sigma_i =20\deg$,

537: $\ndet \simeq 10^3$.  Whether or not it is efficient or

538: feasible to search the individual $3~{\rm sec}$ integrations for

539: occultations remains to be seen. However, given the large potential 

540: increase in the

541: number of expected events, and the fact that the subsequent constraints

542: will be hard to acquire with other methods, the possibility should be

543: considered.

544: 

545: Given the fact that the occultation events are unresolved, and have a

546: small amplitude, systematic effects that could mimic occultation

547: events are a serious concern.  For a large sample of candidate events,

548: there are some signatures that may differentiate between real

549: occultations and systematics.  Occultation events should show a factor

550: of $\sim 2$ gradient in the optical depth over the $\sim 10\deg$ in

551: ecliptic latitude spanned by the \kep\ FOV.  In addition, the

552: relative velocity also varies over the course of the year; for the

553: latitude of the \kep\ FOV, this variation is $10~\kms$ peak-to-peak,

554: which corresponds to a variation in both the signal-to-noise and event

555: rate of amplitude $\sim 30\%$.  Unfortunately, given the intrinsic

556: velocity dispersion of KBOs, and the variation in signal-to-noise

557: noise due to non-equatorial occultations, it will be difficult to

558: measure these deviations unless there are a large number of events.

559: For individual events, it may be possible to distinguish between

560: occultations and systematic effects based on the detailed shape and

561: centroid of the stellar point-spread function for high signal-to-noise events

562:  \citep{jd03}.

563: Ground-based follow-up observations can be used to eliminate

564: contamination from background stars that happen to be blended with the

565: primary star.  

566: 

567: It may also be possible to directly detect the KBOs causing 

568: occultation events by imaging them in reflected light. 

569: Provided that the data are analyzed

570: in real time, candidate occultation events could be alerted within a day

571: of the event.  The signal-to-noise of the occultations should be

572: correlated with their size (e.g., Eq.\ \ref{eqn:snexp}), and therefore

573: brighter events can be preferentially selected for follow-up.  The

574: brighter candidates will have $m_R \sim m_B \simeq 26~{\rm mag}$,

575: within the range of large, ground-based telescopes.  Typical proper

576: motions are a few arcseconds per hour.  Therefore, deep imaging taken

577: at several epochs over a few days after the occultation should reveal

578: a faint object moving away from the bright target star.  

579: 

580: Provided that systematic effects can be controlled, and a reliable

581: sample of occultation events can be acquired, then the observed

582: distribution of signal amplitudes can be used to statistically infer

583: the distribution of event time scales.  This can then be converted to

584: an estimate of the optical depth.  If no events are detected, then it

585: should be possible to place limits on the surface density of KBOs in

586: the \kep\ FOV.  Therefore, \kep\ should measure or significantly constrain the

587: number of highly-inclined, $\sim 10~\km$-sized KBOs.

588: 

589: 

590: \acknowledgments 

591: I would like to thank Josh Bloom, Dave Latham, and Josh Winn for useful discussions, Matt Holman for discussions and reading the manuscript, the anonymous referee for helpful comments and suggestions, and Cheongho Han

592: for the use of his code to calculate occultation light curves. 

593: This work was supported by a Menzel Fellowship from the 

594: Harvard College Observatory.

595: 

596: \begin{thebibliography}{}

597: 

598: \bibitem[Allen(1976)]{allen76} 

599: Allen, C.~W.\ 1976, Astrophysical Quantities (London: Athlone)

600: 

601: \bibitem[Bailey(1976)]{bailey76}

602: Bailey, M.\ 1976, Nature, 259, 290

603: 

604: \bibitem[Bernstein et~al.(2004)]{bern04}

605: Bernstein, G.~M., Trilling, D.~E., Allen, R.~L., Brown, M.~E., Holman, M., Malhotra, R.\ 2004, ApJ, submitted (astro-ph/0308467) (B04)

606: 

607: \bibitem[Brown \& Webster(1997)]{bw97}

608: Brown, M.~J.~I., Webster, R.~L.\ 1997, MNRAS, 289, 783

609: 

610: \bibitem[Brown(2001)]{brown01}

611: Brown, M.~E.\ 2001, AJ, 121, 2804 (B01)

612: 

613: \bibitem[Cooray \& Farmer(2003)]{cf04} 

614: Cooray, A.~\& Farmer, A.~J.\ 2003, \apjl, 587, L125 

615: 

616: \bibitem[Dyson(1992)]{dyson92}

617: Dyson, F.~J.\ 1992, QJRAS, 33, 45

618: 

619: \bibitem[Jenkins \& Doyle(2003)]{jd03}

620: Jenkins, J.~M., Doyle, L.~R.\ 2003, ApJ, 595, 429

621: 

622: \bibitem[Kaiser et al.(2002)]{kaiser02} Kaiser, N.~et al.\ 2002, 

623: \procspie, 4836, 154 

624: 

625: \bibitem[Liang et~al.(2003)]{liang03}

626: Liang, C.-L., Rice, J.~A., de Pater, I., Alcock, C., Axelrod, T.,

627: Wang, A., \& Marshall, S.\ 2003 (astro-ph/0209509)

628: 

629: \bibitem[Luu \& Jewitt(2002)]{lj02}

630: Luu, J.~X., \& Jewitt, D.~C.\ 2002, ARA\&A, 40, 63

631: 

632: \bibitem[Malhotra(1995)]{malhotra95}

633: Malhotra, R.\ 1995, \aj, 110, 420

634: 

635: \bibitem[Millis et al.(2002)]{millis02} Millis, R.~L., Buie, 

636: M.~W., Wasserman, L.~H., Elliot, J.~L., Kern, S.~D., \& Wagner, R.~M.\ 

637: 2002, \aj, 123, 2083

638: 

639: \bibitem[Millis, Dunham, \& Sebring(2003)]{mds03} 

640: Millis, R.~L., Dunham, E.~W., \& Sebring, T.~A.\ 2003, BAAS, 203,  38.16

641: 

642: 

643: \bibitem[Pan \& Sari(2004)]{ps04}

644: Pan, M., \& Sari, R.\ Icarus, submitted (astro-ph/0402138)

645: 

646: \bibitem[Robin, Reyl{\' e}, Derri{\` e}re, \& 

647: Picaud(2003)]{robin03} Robin, A.~C., Reyl{\' e}, C., Derri{\` 

648: e}re, S., \& Picaud, S.\ 2003, \aap, 409, 523 

649: 

650: \bibitem[Roques,  Moncuquet \& Sicardy(1987)]{rms87}

651: Roques, F., Moncuquet, M., Sicardy, B.\ 1987, AJ, 93, 1549

652: 

653: \bibitem[Roques \& Moncuquet(2000)]{rm00}

654: Roques, F., Moncuquet, M.\ 2000, Icarus, 147, 530

655: 

656: \bibitem[Roques et al.(2003)]{roques03} Roques, F., Moncuquet, 

657: M., Lavilloni{\` e}re, N., Auvergne, M., Chevreton, M., Colas, F., \& 

658: Lecacheux, J.\ 2003, \apjl, 594, L63 

659: 

660: 

661: \bibitem[Trujillo, Jewitt \& Luu(2001)]{tjl01}

662: Trujillo, C.~A., Jewitt, D.~C., \& Luu, J.~X.\ 2001, AJ, 122, 457

663: 

664: \bibitem[Tyson(2002)]{tyson02} Tyson, J.~A.\ 2002, \procspie, 

665: 4836, 10 

666: 

667: \end{thebibliography}

668: 

669: 

670: \end{document}

671: 

672: 

673: 

674: 

675: 

676: 

677: =

678: