1: \def\etal{{\it et al.\thinspace}}
2: \def\eg{{\it e.g.\ }}
3: \def\etc{{\it etc.\ }}
4: \def\ie{{\it i.e.\ }}
5: \def\cf{{\it c.f.\ }}
6: \def\gsim{~\rlap{$>$}{\lower 1.0ex\hbox{$\sim$}}}
7: \def\lsim{~\rlap{$<$}{\lower 1.0ex\hbox{$\sim$}}}
8:
9:
10: \documentclass[12pt,preprint]{aastex}
11:
12: %\usepackage{setspace}
13: %\doublespacing
14:
15: \begin{document}
16:
17: \title{Tides and the Evolution of Planetary Habitability}
18:
19: \author{Rory Barnes\altaffilmark{1}, Sean N. Raymond\altaffilmark{2,3}, Brian Jackson\altaffilmark{1}, and Richard Greenberg\altaffilmark{1}}
20:
21: \altaffiltext{1}{Lunar and Planetary Laboratory, University of Arizona,
22: Tucson, AZ 85721}
23: \altaffiltext{2}{NASA Postdoctoral Program Fellow}
24: \altaffiltext{3}{Center for Astrophysics and Space Astronomy, University of Colorado, Boulder, CO 80309}
25:
26: \keywords{Extrasolar Terrestrial Planets; Habitable Zones; M stars; Gl 581}
27:
28: \begin{abstract}
29: Tides raised on a planet by its host star's gravity can reduce a
30: planet's orbital semi-major axis and eccentricity. This effect is only
31: relevant for planets orbiting very close to their host stars. The
32: habitable zones of low-mass stars are also close-in and tides can
33: alter the orbits of planets in these locations. We calculate the
34: tidal evolution of hypothetical terrestrial planets around low-mass
35: stars and show that tides can evolve planets past the inner edge of
36: the habitable zone, sometimes in less than 1 billion years. This
37: migration requires large eccentricities ($>0.5$) and low-mass stars
38: ($\lsim 0.35 M_\odot$). Such migration may have important implications
39: for the evolution of the atmosphere, internal
40: heating and the Gaia hypothesis. Similarly, a planet detected interior to the habitable zone
41: could have been habitable in the past. We consider the past
42: habitability of the recently-discovered, $\sim 5$ $M_\oplus$ planet,
43: Gliese 581 c. We find that it could have been habitable for
44: reasonable choices of orbital and physical properties as recently as 2
45: Gyr ago. However, when we include constraints derived from the
46: additional companions, we see that most parameter choices that predict
47: past habitability require the two inner planets of the system to have
48: crossed their mutual 3:1 mean motion resonance. As this crossing
49: would likely have resulted in resonance capture, which is not
50: observed, we conclude that Gl 581 c was probably never habitable.
51: \end{abstract}
52:
53: \clearpage
54: \section{Introduction}
55: After the discovery of extra-solar planets, the detection of a
56: habitable, terrestrial (rocky) planet became a real possibility. Of particular interest are planets in orbit about low-mass stars ($\lsim 0.5 M_\odot$). The lower stellar mass makes detection of lower mass planets easier and leads to habitable orbits with smaller semi-major axes (which also increase the likelihood of detection by current technologies). The
57: recent detection of putative 5 and 8 $M_\oplus$ planets around an M
58: star (stellar mass $M_* = 0.31 M_\odot$) (Udry \etal 2007) near the
59: circumstellar ``habitable zone'' (HZ) (Kasting \etal 1993; Franck \etal
60: 2000; von Bloh \etal 2007; Selsis \etal 2007; hereafter S07) suggests
61: that Doppler surveys (\ie Butler \etal 2006; Naef \etal 2007) are
62: becoming sensitive enough to detect terrestrial planets in the
63: HZ of low-mass
64: stars. Not surprisingly, M stars now appear to be attractive targets
65: for detecting habitable planets (Tarter \etal 2007).
66:
67: The \textit{habitable zone} is defined, following Kasting \etal
68: (1993), as the range of distances from a star at which liquid water
69: could be stable at the surface of a planet. The definition usually
70: implicitly assumes that the star is the main source of heat. Thus, it
71: excludes locations where other sources of heat might be available, such
72: as moons, like Jupiter's Europa, where tidal heating is a major factor
73: maintaining near-surface liquid water. In that sense, the phrase
74: \textit{habitable zone}, as usually used in astrophysics, is a
75: misnomer. Nevertheless, here we use the conventional definition, while
76: explicitly emphasizing that life could exist outside this zone.
77:
78: The detection of Gl 581 c and d (Udry \etal 2007) has demonstrated
79: that planets can form near the HZ of low-mass stars. For
80: such stars the habitable zone is so close to the star that tidal
81: effects may be important (Mardling \& Lin 2004). Tides raised between
82: the star and planet introduce torques on the planet resulting in orbital
83: evolution, usually decay of semi-major axis, $a$, and eccentricity,
84: $e$. Here we quantify the effects of tidal forces in such zones. We
85: find that Earth-sized planets with large initial eccentricities could
86: migrate out of the HZ of low-mass stars in less than 1
87: billion years.
88:
89: The proper definition of an HZ is not obvious. As
90: described in S07, the boundaries of the HZ are a complex function of
91: both stellar and planetary properties. For any individual planet, the definitions we choose may not be appropriate. Nonetheless, we will show a general trend that must be
92: considered when assessing a planet's habitability, both in the past
93: and in the future.
94:
95: We consider the evolution of terrestrial planets beginning from a
96: location that is almost assuredly habitable to a region that is
97: probably not habitable. We call the time to complete this traversal
98: the ``habitable lifetime'', $\tau_{HL}$. For relatively massive stars
99: ($\gsim 0.35 M_\odot$), the HZ is far enough away from the star that tides cannot cause significant tidal evolution\footnote{Note we only consider tidal effects after the star has reached the main sequence}. However for planets with relatively large
100: eccentricities ($\gsim 0.5$) orbiting low-mass stars ($\lsim 0.35
101: M_\odot$) tides can pull a planet out of the HZ in less than 1 Gyr. The large number of low-mass stars and significant fraction of
102: known exoplanets with large $e$ suggest that tides may limit the total
103: number of habitable planets in the galaxy.
104:
105: We assume that potentially habitable planets can form with a range of
106: masses and orbits in the HZ of low-mass stars (see Chambers [2004];
107: Raymond, Quinn \& Lunine [2007] or Lissauer [2007] for a discussion of
108: the formation of such planets). Of course, the formation of a planet
109: in the HZ does not mean a planet is, will be, or has been
110: habitable. Many other factors are required such as favorable
111: composition (Lissauer 2007), a thick atmosphere, substantial mass for
112: sustained plate tectonics ($m \gsim 0.3 M_\oplus$) (Williams, Kasting
113: \& Wade 1997; Raymond, Scalo \& Meadows 2007), which on Earth drives
114: the carbonate-silicate cycle (Walker, Hays \& Kasting 1981), and substantial
115: water content (\eg Raymond, Quinn \& Lunine 2004), both for life and to
116: facilitate plate tectonics (Regenauer-Lieb, Yuen \& Barlund 2001). Nonetheless we will call planets ``habitable'' if their orbital properties are such that they permit water to be stable on the surface.
117:
118: The recently announced Gl 581 system contains a planet, c, that is
119: probably just interior to the HZ (S07). In Table 1 we list the
120: properties of the planets in this system; $m_p$ is the planet mass, $P$ is
121: the orbital period, $a$ is the semi-major axis, $e$ is the
122: eccentricity, $\varpi$ is the longitude of periastron, and $T_{peri}$
123: is the time of periastron passage in Julian Days. Tidal theory
124: predicts planet c orbited with larger values of semi-major axis and
125: eccentricity in the past. Therefore it may have been habitable
126: in the past, but tides subsequently moved the planet into an
127: uninhabitable orbit. We find that if planet c was the only planet in
128: the system, plausible physical properties predict it was
129: habitable. However, when we consider constraints from the mutual interactions of the
130: additional planets, we find that planet c has likely never been
131: habitable.
132:
133: Tides strong enough to induce significant orbital evolution can also heat the interior of a planet through tidal working. The same type of heating causes Io's intense vulcanism (Peale, Cassen \& Reynolds 1979). Therefore we might expect an increased heating rate on
134: the planets we consider here as well. The orbital evolution we
135: consider could also have implications for the evolution of a habitable
136: planet's atmosphere. In some cases, the orbit-averaged flux can increase
137: by a factor of 2 as the orbit shrinks (assuming constant stellar luminosity).
138:
139: In $\S$ 2 we review tidal theory, present our definition of the HZ,
140: and describe how we determine the habitable lifetime. In $\S$ 3 we
141: present our results for hypothetical terrestrial planets. In $\S$ 4 we
142: describe how our model applies to the Gl 581 c planet. Finally, in
143: $\S$ 5 we draw our conclusions and discuss directions for future
144: research.
145:
146: \section{Methodology}
147:
148: \subsection{Tidal Theory}
149: We evaluate the effects of tides using the classical, second order
150: equations assembled by Goldreich \& Soter (1966) (see also Kaula
151: 1964; Greenberg 1977):
152: \begin{equation}
153: \label{eq:adot}
154: \frac{da}{dt} = -\Big(21\frac{\sqrt{GM_*^3}R_p^5k}{m_pQ_p}e^2 + \frac{9}{2}\frac{\sqrt{G/M_*}R_*^5m_p}{Q'_*}\Big)a^{-11/2}
155: \end{equation}
156: \begin{equation}
157: \label{eq:edot}
158: \frac{de}{dt} = -\Big(\frac{21}{2}\frac{\sqrt{GM_*^3}R_p^5k}{m_pQ_p} + \frac{171}{16}\frac{\sqrt{G/M_*}R_*^5m_p}{Q'_*}\Big)a^{-13/2}e
159: \end{equation}
160: where $a$ is semi-major axis, $G$ is Newton's gravitational constant,
161: $m_p$ is the mass of the planet, $k$ is the planet's Love number,
162: $Q_p$ is the planet's tidal dissipation function, $Q'_*$ is the star's
163: tidal dissipation function divided by two-thirds its love number,
164: $R_p$ is the planet's radius, and $R_*$ is the star's radius. The
165: first terms in these equations are due to the tide raised on the
166: planet by the star, and the second terms are due to tides raised on
167: the star by the planet. Eqs.\ (\ref{eq:adot} -- \ref{eq:edot}) are
168: only valid if the planet's orbital period is less than the star's
169: rotation period, otherwise the signs of the second two terms will
170: change. For the cases we consider, this condition is irrelevant
171: because the second terms of Eqs.\ (\ref{eq:adot} -- \ref{eq:edot}) are
172: 6 -- 7 orders of magnitude smaller than the first. This difference is
173: largely due to the mass difference between a terrestrial planet and a
174: star. Note that tidal theory predicts decay for both $a$ and $e$. As
175: tides slowly change a planet's orbit, the planet may move into or out
176: of the HZ.
177:
178: The equations above were derived under the implicit assumption that $e$
179: values are small. Here we apply them to cases that include conditions
180: with large $e$ values. Thus higher-order corrections may potentially be
181: important. Several models of tidal evolution that are not restricted to
182: small values of e have been derived (Hut 1981; Eggleton, Kiseleva \& Hut
183: 1998; Mardling \& Lin 2002). However, such models must include specific
184: assumptions about how a body responds to the ever-changing tidal
185: potential. The tidal forcing involves components with a various
186: frequencies, but the nature of the response of a planet or star is
187: uncertain. It is not clear how the tidal phase lag depends on the
188: frequency of each harmonic. This issue becomes increasingly critical when
189: $e$ is large and a wide range of different comparable-amplitude frequencies
190: are involved. For a model valid to a high order in $e$, the validity of
191: parameterizing tidal effects through a single constant $Q$ is questionable.
192: Moreover, tidal heating could lead to core/mantle melting and a time-dependent $Q$ value. Not enough is known about the actual response of real bodies to evaluate
193: these higher-order effects. Therefore, to the best of current knowledge,
194: Eqs.\ (1 -- 2) above are reasonable representations of tidal evolution. As
195: the theory matures, the effect of higher-order terms could be included
196: and this study should be revisited. Note that the higher order theories
197: that have been derived all predict faster evolution than Eqs. (1 -- 2)
198: (all the higher order terms have the same sign as those in Eqs. (1 -- 2), so their predictions for the habitable lifetime, $\tau_{HL}$, would
199: be even shorter than those we derive below. In that sense, our approach
200: provides a conservative estimate of the tidal effects on terrestrial
201: planets.
202:
203: Eqs.\ (\ref{eq:adot} -- \ref{eq:edot}) are only valid when the star's
204: orbit-averaged torque on the planet's tidal bulge is zero. When this situation occurs, the planet is said to be ``tidally locked'' to its star. For a
205: planet on an eccentric orbit, the torque is greatest at pericenter at
206: which point the star is moving with an angular velocity greater than
207: the mean motion. Therefore the rotational frequency of the planet
208: $\Omega$ is larger than its orbital frequency $n$ (tidal locking does not mean that one side of the planet faces the star for all times, this situation only occurs for a circular orbit). To second order in $e$ these two
209: frequencies are related by the following equation,
210: \begin{equation}
211: \label{eq:spinlock}
212: \Omega = n(1 + \frac{19}{2}e^2)
213: \end{equation}
214: (Murray \& Dermott 1999; Goldreich 1966). Note that this equation neglects effects of a permanent quadrupole moment (we ignore this effect as their is currently no hope of measuring it in an exoplanet). Even for modest eccentricities, the spin and orbital frequencies can be significantly different due to the coefficient in front of the $e^2$ term. A planet
215: with an eccentricity of 0.32 has a spin period half that of its
216: orbital period.
217:
218: The time for a planet to become tidally locked is given by
219: \begin{equation}
220: \label{eq:timelock}
221: t_{lock} = \frac{8m_pQ_pa^6}{45GM_*^2kr_p^3}\Delta\Omega
222: \end{equation}
223: (Goldreich \& Soter 1966; Rasio \etal 1996), where $\Delta\Omega$ is the difference between the initial spin, and the spin given by Eq.\ (\ref{eq:spinlock}). For an Earth-mass, terrestrial planet ($k=0.3, Q_p = 100$) planet
224: orbiting a 0.3 $M_\odot$ star with $e=0$ $a=0.1$ AU, this timescale is about $10^6$ years. We
225: consider planets initially with much larger eccentricities, but
226: eccentricity probably does not affect this timescale significantly
227: (Peale 1977). This time is very short compared to other timescales associated with planet formation, and we may assume
228: the planets are tidally locked to the star as soon as they are formed.
229:
230: In order to determine $\tau_{HL}$ for possible planets, we consider a
231: range of values of $M_*$, $m_p$, $a$ and $e$. We will set $Q'_* =
232: 10^{5.5}$, $Q_p = 21.5$, and $k = 0.3$, reasonable choices based on
233: limited observational constraints for stars (Ogilvie \& Lin
234: 2007; Jackson, Greenberg \& Barnes 2007) and the Earth (Dickey \etal 1994; Mardling \& Lin 2004). However, the
235: time-averaged value of $Q$ of the Earth is
236: probably closer to 100 (Lambeck 1977), which is similar to the values
237: of other terrestrial planets in the Solar System (Yoder
238: 1995). Nonetheless we choose $Q_p = 21.5$ to be consistent with past
239: work (Mardling \& Lin 2004).
240:
241: From
242: our choices of $M_*$ and $m_p$ we must determine radii. For the
243: planet, we will use the scaling relationship
244: \begin{equation}
245: \label{eq:rp}
246: \frac{R_p}{R_\oplus} = \Big(\frac{m_p}{M_\oplus}\Big)^{0.27}
247: \end{equation}
248: (Valencia, O'Connell \& Sasselov 2006). For the stellar radius, we use an empirical
249: relation derived from observations of eclipsing binaries (Gorda \&
250: Svechnikov 1999):
251: \begin{equation}
252: \label{eq:rs}
253: \log_{10} \frac{R_*}{R_\odot} = 1.03\log_{10} \frac{M_*}{M_\odot} + 0.1,
254: \end{equation}
255: where $R_\odot$ and $M_\odot$ are the solar radius and mass,
256: respectively. Note that Eq.\ (\ref{eq:rs}) is only valid when $M_*
257: \lsim M_\odot$.
258:
259: \subsection{The Eccentric Habitable Zone}
260: For circular orbits, the HZ is defined as the range of orbits in which
261: a terrestrial planet, with an atmosphere favorable for habitability, can support liquid
262: water on the surface (Kasting \etal 1993; S07). However, extra-solar
263: planet orbits are often highly eccentric. For these orbits,
264: Williams \& Pollard (2002) showed that, despite large variations in
265: surface temperature, the key for long-term climate stability is the
266: orbit-averaged flux,
267: \begin{equation}
268: \label{eq:flux}
269: F = \frac{L}{4\pi a^2\sqrt{1-e^2}},
270: \end{equation}
271: (Williams \& Pollard 2002; Adams \& Laughlin 2006), where $F$ is the orbit-averaged flux and $L$ is the stellar luminosity. Note, however, that potential water loss at
272: periastron and potential cloud freeze-out at apoastron have not been
273: accounted for, despite their implications for habitability. We define
274: the eccentric habitable zone (EHZ) as the range of orbits for which a
275: planet receives as much flux over 1 orbit as a planet on a circular
276: orbit in the ``classical'' HZ (Kasting \etal 1993; S07). We must also scale the
277: location of the EHZ to the luminosity of the stars we are
278: considering. Lower mass stars have lower luminosities, and hence the
279: EHZ's are closer to the star. We therefore define the inner and outer
280: edges of the EHZ, $l_{in}$ and $l_{out}$ (in AU), respectively, as
281: \begin{equation}
282: \label{eq:lin}
283: l_{in} = (l_{in\odot} - a_{in}T_* - b_{in}T_*^2)\Big(\frac{L}{L_\odot}\Big)^{1/2
284: }(1-e^2)^{-1/4},
285: \end{equation}
286: and
287: \begin{equation}
288: \label{eq:lout}
289: l_{out} = (l_{out\odot} - a_{out}T_* - b_{out}T_*^2)\Big(\frac{L}{L_\odot}\Big)^{1/2
290: }(1-e^2)^{-1/4},
291: \end{equation}
292: where $l_{in}$ and $l_{out}$ are the inner and outer edges of the HZ, respectively, in AU, $l_{in\odot}$ and $l_{out\odot}$ are the
293: inner and outer edges of the HZ in the solar system, respectively, in AU, $a_{in} = 2.7619
294: \times 10^{-5}$ AU/K, $b_{in} = 3.8095 \times 10^{-9}$ AU/K$^2$, $a_{out} = 1.3786 \times 10^{-4}$ AU/K, and $b_{out} = 1.4286 \times 10^{-9}$ AU/K$^2$ are empirically
295: determined constants, $T_* = T_{eff} - 5700$ K, and $L_\odot$ is the solar luminosity (S07). Terms involving $T_*$ relate stellar intensity at different wavelengths and atmospheric windows of predicted habitable planets. Eq.\ (\ref{eq:lin} -- \ref{eq:lout}) are only valid for $T_{eff} < 3700$ K. $T_{eff}$ is the
296: effective temperature of the star, given by
297: \begin{equation}
298: \label{eq:teff}
299: T_{eff} = \Big(\frac{L}{4\pi\sigma R_*^2}\Big)^{1/4}
300: \end{equation}
301: where $\sigma$ is the Stefan-Boltzman constant.Therefore a planet with
302: a given $e$ is on the inner edge of the EHZ if $a = l_{in}$. Note that as $e$ changes so do the locations of $l_{in}$
303: and $l_{out}$.
304:
305: The values of $l_{in\odot}$ and $l_{out\odot}$ are therefore the key
306: parameters in the identification of the edges of the EHZ. We
307: consider three criteria identified in S07: 1) 0\% cloud cover, 2) 50\%
308: cloud cover, and 3) 100\% cloud cover\footnote{Note that S07 also give
309: values of $l_{in\odot}$ and $l_{out\odot}$ based on assumptions
310: regarding the past and present habitability of Venus and Mars coupled
311: with stellar evolution, but we will ignore those models here.}. S07
312: give values of $l_{in\odot}$ of $\sim 0.89$, $\sim 0.72$, and
313: $\sim 0.49$ AU for the three possibilities, respectively. For the outer
314: edge, S07 give $l_{out\odot} = 1.67, 1.95,$ and 2.4 AU, for the three cloud cover models, respectively. We
315: arbitrarily choose the 50\% cloud cover case to be the limits of the
316: EHZ. This choice is the middle of the possibilities, and roughly
317: corresponds to the cloud cover on the Earth.
318:
319: The stellar mass is related to the luminosity in the following way,
320: \begin{equation}
321: \label{eq:luminosity}
322: \lambda = 4.101\mu^3 + 8.162\mu^2 + 7.108\mu + 0.065,
323: \end{equation}
324: where $\lambda$ = log$_{10} (L/L_\odot)$ and $\mu$ = log$_{10}
325: (M_*/M_\odot)$ (Scalo \etal 2007). With these definitions, we plot the shape of the EHZ
326: as a function of stellar mass and initial eccentricity, $e_0$, in
327: Fig.\ \ref{fig:hz}. As both $e_0$ and $M_*$ increase, the location of
328: the EHZ moves out and grows wider. If, for a certain star, $T_{eff} < 3700$ K, we set it to 3700 K in Eqs.\ (\ref{eq:lin} -- \ref{eq:lout}) as done in S07.
329:
330: \subsection{Determination of the Habitable Lifetime}
331: The habitable lifetime, $\tau_{HL}$, depends on the initial orbit of a
332: planet. Rather than explore all possible orbits, we will consider one
333: type of starting location: the inner 0\% cloud cover boundary (\ie
334: $l_{in\odot} = 0.89$ AU). Although this choice
335: of initial condition is arbitrary, it demonstrates how tides will
336: change the orbits, and, hence, habitability. As stated in the previous
337: section, we consider a planet uninhabitable when it crosses the 50\%
338: cloud cover boundary ($l_{in\odot} = 0.72$ AU). We vary the planet's initial eccentricity $e_0$ and $M_*$ in
339: increments of 0.01 and 0.01 $M_\odot$, respectively, and consider four different planet masses: 5, 1,
340: 0.5, and 0.1 M$_\oplus$.
341:
342: In order to calculate $\tau_{HL}$ we numerically integrate Eqs.\
343: (\ref{eq:adot} - \ref{eq:edot}) and (\ref{eq:lin}) with a timestep of
344: 10,000 years. Convergence tests showed that this timestep is two
345: orders of magnitude smaller than necessary to produce reliable
346: results. We integrate the tidal equation until either $a < l_{in}$(50\% cloud cover) or
347: $\tau_{HL} > 10$ Gyr.
348:
349: \section{Habitable Lifetimes}
350: In this section we present results for $\tau_{HL}$ as defined in $\S$
351: 2.3. Fig.\ \ref{fig:thz4} shows contour lines of $\tau_{HL}$ for planets
352: with masses of 5, 1, 0.5 and 0.1 M$_\oplus$. There are two prominent
353: features in each panel. First is the abrupt change in $\tau_{HL}$ from
354: 10 Gyr to less than 1 Gyr between $e_0 = 0.53$ and 0.54. This feature
355: occurs because the time to circularize the orbit is less than 10 Gyr. When $e$ reaches zero, the
356: evolution of $a$ effectively stops, and the planet is stuck on one
357: side of the boundary or the other. Orbits with $e_0 \lsim 0.53$ become
358: circular with $a$ just larger than $l_{in}$. However, when $e_0 \gsim
359: 0.54$, the orbits circularize with $a$ slightly less than $l_{in}$.
360:
361: These different types of evolution are shown in Fig.\ \ref{fig:evol}.
362: Note that the evolution of both cases is very similar, but the final
363: relationship between $a$ and $l_{in}$ is such that one planet remains
364: in the EHZ, but the other does not. Note that the EHZ changes because
365: $e$ is changing. The difference is therefore not due to some fundamental
366: feature of tidal theory (other than changes in semi-major axis become
367: negligible for small eccentricities), rather
368: our definition of the EHZ leads to the different values of
369: $\tau_{HL}$.
370:
371: The apparently sudden transition from long to short habitable
372: lifetimes can be shown by considering how $a$ and $e$ are related:
373: \begin{equation}
374: \label{eq:ae}
375: \ln\Big(\frac{a}{a_0}\Big) = e^2 - e_0^2,
376: \end{equation}
377: where $a_0$ and $e_0$ are the initial values of $a$ and $e$, if the effects of the tide raised on the star by the planet is
378: negligible (which is the case here) (Jackson, Greenberg \& Barnes 2007). If we assume $e=0$ (\ie the tidal evolution has
379: effectively ended, and $t = \infty$), $a = l_{in}$(50\% cloud cover)
380: and $a_0 = l_{in}$(0\% cloud cover), then the value of $e_0$ that
381: solves Eq.\ (\ref{eq:ae}) should correspond to the critical value ($e_0 \sim 0.535$)
382: seen in Fig.\
383: \ref{fig:thz4} (note that $a_0$ is also a function of $e_0$ and Eq.\
384: [\ref{eq:ae}] must be solved numerically). In the cases we consider,
385: this equation is independent of mass. The values of $a$ and $a_0$ have
386: the same dependence on luminosity (and hence mass, \cf Eq.\
387: [\ref{eq:luminosity}]), and, for the stars we consider, we have set
388: $T_{eff} = 3700$ in order to follow the example of S07. Therefore we
389: expect the critical value of $e_0$ to be independent of both stellar
390: and planetary mass. Solving Eq.\ (\ref{eq:ae}) for the values of
391: $l_{in}$ we have chosen, we find the critical value of $e_0$ is
392: 0.53490, in agreement with Fig.\ \ref{fig:thz4}.
393:
394: The second feature of the contours is the downturn that occurs for
395: $e_0 \gsim 0.65$, in other words, $\tau_{HL}$ gradually increases with
396: $e_0$. This dependence follows from the change in the EHZ at larger
397: values of $e_0$, as shown in Fig.\ \ref{fig:hz}. At these values not
398: only is the EHZ wider, but $a_0$ is also larger. Therefore planets' orbits
399: will evolve more slowly (note the $a$ dependence in Eqs.\ [\ref{eq:adot} --
400: \ref{eq:edot}]), and have further to go before they reach $l_{in}$(50\% cloud cover).
401:
402:
403: We also see in Fig.\ \ref{fig:thz4} that lower mass planets tend to
404: have longer habitable lifetimes. This relationship follows from the
405: dependence of $da/dt$ on $m_p$ in Eqs. (\ref{eq:adot} --
406: \ref{eq:edot}). Low-mass stars, like the ones considered here, are
407: likely to host low-mass planets (Laughlin, Bodenheimer \& Adams 2004; Ida \& Lin 2005; Raymond, Scalo \& Meadows 2007), and therefore disks around M
408: stars may preferentially form planets with relatively long habitable
409: lifetimes. However, they may not contain enough volatiles for life
410: because the impacts that deliver them may have large enough energies that the volatiles will be lost to the planet (Lissauer 2007).
411:
412: Models of the evolution of habitable atmospheres have focused on planets with
413: static orbits (\eg Kasting \& Catling 2003). Figure \ref{fig:thz4}
414: shows that such a model may be inappropriate for determining
415: habitability and biosignatures of planets around M stars (Segura \etal
416: 2005; S07). In Fig.\ \ref{fig:flux} we plot the orbit-averaged stellar
417: flux for a planet with
418: $\tau_{HL} = 4.5$ Gyr orbiting a 0.2 $M_\odot$ star (assuming the
419: stellar luminosity is constant). This figure also shows how the spin
420: frequency of the planet $\Omega$ evolves relative to its mean motion $n$,
421: see Eq.\ (\ref{eq:spinlock}). The tidally evolving planet does not
422: orbit such that only one side faces the star until the orbit has
423: become circular at $t \approx 6$ Gyr.
424:
425: Note as well that there exists a family of orbits with $\tau_{HL}
426: \approx 4.5$ Gyr, about the age of the Earth. Therefore these
427: calculations suggest that if a planet with the correct ingredients for
428: life formed with that orbit, and life subsequently developed in a
429: manner similar to the Earth, that life would be eliminated by the
430: ultimate global warming event: The passage of the planet through the
431: inner edge of the HZ due to tidal evolution.
432:
433: \section{Application to the Gl 581 System}
434: In this section we apply the ideas of $\S$ 2 to the recently
435: announced Gl 581 planetary system (Udry \etal 2007). This system
436: contains three planets with orbits and masses listed in Table
437: 1. Planet c is probably interior to the 50\% cloud cover HZ, and
438: planet d is probably exterior (S07). The age of the star is
439: uncertain. Its kinematics and metallicity suggest that it is at least a few Gyr old (Bonfils \etal 2005; Delfosse \etal 1998), but its very low X-ray luminosity (Delfosse \etal 1998) indicates it may be as old as 8 --
440: 10 Gyr old (Ribas, private communication; Selsis, private
441: communication). We therefore tidally evolve planet c backward for 10 Gyr by
442: changing the signs of Eqs.\ (\ref{eq:adot} -- \ref{eq:edot}) (Jackson, Greenberg \& Barnes 2007). First we consider how Gl 581 c would evolve if
443: it were an isolated planet, then we consider how the additional
444: companions constrain its evolution and the planets' physical
445: properties.
446:
447: \subsection{Past Evolution of Gl 581 c}
448: If Gl 581 c were the only planetary companion to its host star, then
449: its past tidal evolution is adequately modeled by Eqs.\ (\ref{eq:adot} --
450: \ref{eq:edot}). For this particular system, we have specific
451: constraints on physical and orbital properties. The planet's radius
452: $R_c$ lies in the range $1.6 \lsim R_c \lsim 2 R_\oplus$, assuming the
453: observed minimum mass is the actual mass (S07; Valencia, O'Connell \& Sasselov 2006; Fortney, Marley \& Barnes 2007; Valencia, Sasselov \& O'Connell 2007a, 2007b; Sotin, Grasset \& Moquet 2007), and its best-fit eccentricity
454: $e_c$ is 0.16, although any value between 0 and 0.3 is about equally likely
455: (Udry \etal 2007). For tidal parameters, we appeal to values in our
456: Solar System. The current values of $k$ and $Q$ for the Earth are
457: about 0.3 and 21.5, respectively, from Lunar Laser Ranging (Dickey
458: \etal 1994; Mardling \& Lin 2004), but over the lifetime of the
459: Earth-moon system, the $Q$ value is probably close to 100 (Lambeck
460: 1977), similar to the value of Mars (Yoder 1995). The
461: star's mass, radius, luminosity and effective temperature are $0.31
462: M_\odot, 0.38 R_\odot, 0.0135 L_\odot$ and 3200 K, respectively
463: (Bonfils \etal 2005). Planet c is currently inside the 100\% cloud
464: cover HZ, but such an atmospheric condition is unlikely (S07). We
465: therefore seek to identify physical and orbital parameters that allow Gl 581 c to have been habitable in the past, \ie inside the 50\% cloud cover HZ.
466:
467: First we consider the effects of varying eccentricity. In Fig.\
468: \ref{fig:e} we show how long ago Gl 581 c would have been inside the
469: 50\% cloud cover HZ, as a function of its current eccentricity, for 2
470: different, plausible values of $Q_c$, assuming $k = 0.3$. At low
471: $e_c$ values, the time is longer because tidal evolution is slower
472: with smaller $e$, while at larger values the time is longer due to the strong coupling between $a$ and $e$, and because the
473: edge of the HZ is farther out, \cf Eqs.\ (\ref{eq:adot} -- \ref{eq:edot})
474: and (\ref{eq:lin}). Therefore
475: there exists a critical value of $e_c$ that minimizes the time since the
476: planet was habitable. In this case, that value is 0.32 in which case
477: the planet would have been inside the 50\% cloud cover HZ as recently
478: as 3 Gyr ago.
479:
480: Next we consider how the planet's radius constrains past
481: evolution. For most cases we considered, the Gl 581 c planet could
482: have been inside the 50\% cloud cover HZ for many Gyr, assuming a
483: system age of 10 Gyr. In Fig.\ \ref{fig:rp} we plot the evolution of
484: the EHZ and $a_c$ for three different values of $R_p$, and assuming
485: its current eccentricity is 0.16. All the cases plotted assume $Q_c =
486: 21.5, k = 0.3$. The shading represents the 100\%, 50\%, and 0\% cloud
487: cover models, and white is outside the HZ, as in Fig.\
488: \ref{fig:hz}. Even for the extreme case $r_c = 2 R_\oplus$, the planet
489: could never have been in the 0\% cloud cover HZ. In this case, it is
490: hard to imagine how $Q_c$ could be so low, since the planet is most
491: likely a ``water world'' (Raymond, Quinn \& Lunine 2004; L\'eger \etal
492: 2004) and for such a planet dissipation of tidal energy would be less
493: efficient ($Q$ would be larger). If the system
494: is 10 Gyr old, this case predicts planet c was habitable for 8 Gyr and
495: tides would have sterilized planet c 2 Gyr ago.
496:
497: \subsection{Possible Interactions Between Planets b and c}
498: So far we have ignored interactions between planets b and c, but as we run
499: backwards in time, we need to remember that their mutual gravitational interactions will cause the eccentricities to oscillate on a timescale of $\sim 10^3$ years, and we must consider the possibility that as their
500: periods change, they could hit a mean motion resonance. The first
501: significant resonance would be the 3:1 ratio of orbital periods. If
502: they ever crossed this resonance, the planets would have almost
503: assuredly been captured into the resonance (Lee \& Peale 2002). Such a
504: scenario would invalidate the conclusions of the previous section because the system is not observed to be in resonance today.
505:
506: Planet b is also subject to tidal forces so we must account for its
507: tidal evolution. Currently its outer 3:1 resonance lies at $\sim
508: 0.085$ AU, and therefore if it did not experience significant
509: evolution, then planet c has orbited inside that distance since the
510: system formed. If, however, planet b experienced significant tidal
511: evolution, then both planets could have migrated such that the 3:1 was
512: always instantaneously beyond the orbit of planet c. The value of
513: $Q'_b$ is unknown for the Neptune-mass planet. If the planet is
514: gaseous ($k = 3/2$), then $Q_b \sim 10^6$ (Jackson, Greenberg \& Barnes
515: 2007), but if it is terrestrial, then the value is probably closer to
516: 100. In Fig.\ \ref{fig:qb} we plot the evolution of $a_b$ and $e_b$
517: for values of $Q_b$ between $10^4$ and $10^6$. We assume the planet
518: has the same bulk density as Neptune, and therefore the radius of
519: $3.74 R_\oplus$. This choice is somewhat arbitrary, but is consistent
520: with the radius of the Neptune-mass planet GJ 436 (Gillon \etal 2007;
521: Deming \etal 2007). The evolution depends strongly on $Q_b$, and in
522: order for planet c to have been habitable and avoided the 3:1
523: resonance, then $Q_b \lsim 5\times 10^4$.
524:
525: In fact, $Q_b$ must be less than $4\times 10^4$ in order to avoid the
526: 3:1 resonance, as shown in Fig.\ \ref{fig:3_1}. This figure plots the
527: evolutions of b and c (without mutual interactions, and $Q_c = 21.5,
528: r_c = 1.8 R_\oplus$), and the instantaneous 3:1 resonance of planet
529: b. We see that in this case planet c just misses the 3:1 resonance.
530:
531: To this point, we have not considered the secular interactions between
532: the planets. Secular interactions cause the eccentricities and
533: orientations of the orbits to oscillate with time (see \eg Barnes \&
534: Greenberg [2006] for a review of secular theory). In Fig.\
535: \ref{fig:secular} we show the secular interactions for the currently
536: observed system, excluding planet d. This simulation was performed
537: with HNBODY\footnote{Publicly available at
538: http://janus.astro.umd.edu/HNBody}, and includes general
539: relativistic effects. Our integration reveals that the current
540: values of the eccentricities are near their extrema. We presume that
541: using the observed values provides a reasonable estimate of the tidal
542: evolution, however we may be overestimating the tidal effects on
543: planet c, and underestimating the tidal effects on planet b. A more sophisticated treatment that combines the secular and tidal evolutions was beyond the scope of this investigation (see $\S$ 5).
544:
545: Tidal theory predicts the eccentricities would be larger in the
546: past. The amplitude of the eccentricity oscillations (due to gravitational interactions between planets) scales with $e$
547: (among other quantities), and eccentricity oscillations would
548: therefore have been larger in the past. Larger eccentricity values
549: can often lead to dynamical instability (\eg Barnes \& Quinn 2004;
550: Barnes \& Greenberg 2007). Therefore we also require that the
551: mutual, tidal evolutions of planets b and c predict dynamically stable orbits. Such a consideration shows that if $Q_c$ is small
552: enough to exclude a 3:1 resonance capture (\ie $< 4 \times 10^4$),
553: then the orbits of planets b and c would have been unstable 10 Gyr
554: ago. This instability arises from the large $e$ values the planets had
555: then ($e_b = 0.74, e_c = 0.47$).
556:
557: Even if we begin the backward integration with the two planet's
558: average eccentricity (over the secular cycle), planet c was probably
559: not habitable. Planet c's average eccentricity is lower than its
560: observed value, and it has probably experienced less tidal evolution than
561: predicted above. Therefore it is unlikely it could have ever been
562: inside the 50\% cloud cover EHZ. On the other hand, b's average
563: eccentricity is larger than the current value, and its tidal evolution
564: would have probably been more significant. In that case, tidal evolution is
565: more likely to predict dynamical instability. We conclude that the
566: requirements that planets b and c avoid the 3:1 resonance and be
567: stable at all times cannot be satisfied if $Q_c = 21.5$ and $k=0.3$.
568:
569: However, if the two planets have properties similar to analogous bodies in the
570: Solar System, then the system would have been stable, and never in a
571: 3:1 resonance. In Fig.\
572: \ref{fig:bestrewind} we show the evolution of the 2 planets assuming
573: the inner planet is Neptune-like ($Q_b = 10^6, R_b = 3.74 R_\oplus$),
574: and planet c is a typical terrestrial planet ($Q_c = 100, k = 0.3, R_c =
575: 1.8 R_\oplus$) (Yoder 1995). In this case, the 3:1 resonance crossing
576: occurred nearly 10 Gyr ago, and the system 10 Gyr ago is stable
577: (assuming low amplitude apsidal libration). This scenario invokes the
578: simplest assumptions, and is therefore the
579: preferable explanation. We conclude that Gl 581 c is probably a terrestrial planet with $Q \sim 100$, but was never inside the
580: 50\% cloud cover EHZ.
581:
582: \section{Conclusions}
583: The detection of a terrestrial planet around a low-mass star is insufficient to
584: determine that planet's past and future habitability. The tidal forces
585: between planet and star can significantly change the orbits, and hence
586: limit the habitable lifetime. Planets detected in the HZ with large
587: eccentricities may be bound for hotter temperatures and ultimately a
588: global extinction. On the other hand, planets detected interior to the HZ
589: may have been habitable in the past. Gl 581 c was most likely not
590: habitable in the past, but $\S$ 4.1 shows that if its companion planets
591: were on different orbits, past habitability would have been possible.
592:
593: For plausible physical and orbital properties of hypothetical
594: terrestrial planets, tides may evolve the planets' orbits past the
595: inner edge of the HZ over a timescale comparable to the age of the
596: Earth. Therefore in order for planets to be habitable long enough for complex
597: life to develop, they must form with eccentricities low enough that
598: tides don't eventually make them inhospitable to life. Alternatively
599: planets that form beyond $l_{out}$ may evolve into the habitable zone,
600: but the ``flux'' of planets out, through the inner edge of the EHZ, is larger
601: than that of planets in, through the outer edge, because of the very steep
602: dependence of $da/dt$ on $a$. Therefore, on average and assuming
603: uniform distribution in $a$, tides tend to reduce the total number of habitable
604: planets around M stars in the galaxy.
605:
606: Our work has shown that tides may shorten habitable lifetimes of
607: planets that form with eccentricities larger than about 0.5. Although
608: the formation mechanism of such planets is unknown, the existence of
609: giant planets with such eccentricities suggests even terrestrial-sized
610: planets may form with similar values. However, the majority of known
611: exoplanets probably formed with eccentricities below 0.5 (see, \eg Jackson,
612: Greenberg \& Barnes 2007), thus most habitable lifetimes
613: will not be significantly shortened due to tides.
614:
615: This work predicts that the vast majority of terrestrial planets in
616: HZs around low-mass stars ($\lsim 0.2 M_\odot$) will be detected on nearly
617: circular orbits because the time to circularize the orbit is less than 1 Gyr, \cf Fig.\ \ref{fig:evol}. Should space-based transit
618: missions like COROT and Kepler find terrestrial planets on non-circular orbits around old
619: (more than a few Gyr), low-mass stars ($< 0.35 M_*$) in small orbits
620: ($a\lsim 0.1$ AU), then that system is likely to contain additional companions (Mardling \& Lin 2004) which pump up eccentricities due to mutual interactions.
621:
622: The increasing stellar flux on a planet due to orbital decay could
623: impact its atmospheric conditions and composition. Previous models of
624: habitable atmospheres have considered planets at a single semi-major
625: axis value (\ie Kasting \etal 1993; Kasting \& Catling 2003; Segura
626: \etal 2005; Tinetti, Rashby \& Yung 2006; Kiang \etal 2007). The tidal change
627: in the orbit of a habitable planet may affect the relative abundances
628: of biosignatures, such as the simultaneous presence of oxygen and
629: methane (Sagan \etal 1993). Future work on the evolution of a
630: habitable atmosphere and the identification of biosignatures should
631: consider the change in stellar flux due to tidal evolution.
632:
633: When a planet is detected such that it was most likely in the HZ in
634: the past, it would be interesting to determine if it did, in fact, support life in the past. In the foreseeable future,
635: such a determination will have to be made through analysis of disk-averaged
636: spectra of planets (Tinetti \etal 2006; Tinetti, Rashby \& Yung 2006; Kaltenegger, Traub \& Jucks 2007). Future work on the evolution of habitable atmospheres should
637: explore the possibility of detecting signatures of extinct life on
638: planets around M stars.
639:
640: Future work should also address the determination of habitable lifetimes in multiple
641: planet systems as well as incorporate higher
642: order corrections to the tidal equations. Such improvements will
643: require substantial advancement in our understanding of the
644: deformations of solid and gaseous bodies due to tides. The Gl 581
645: systems demonstrates the need to develop a theory that describes two
646: tidally damped orbits that also experience significant mutual
647: interactions (note that Mardling [2007] has derived expressions for systems in which 1 planet experiences tidal evolution). However, the results of $\S$ 4.2 suggest that such improvements are unlikely to change our
648: assessment for the Gl 581 system.
649:
650: Planets with masses $\lsim 0.3
651: M_\oplus$ are likely too small to support plate tectonics, which are
652: thought to be necessary for life (Williams, Kasting \& Wade 1997; Raymond, Scalo \& Meadows 2007). This work suggests some of these smaller planets may be heated
653: internally by tides, such as in the Galilean satellites (Peale, Cassen \& Reynolds 1979). In some cases, in which $\tau_{HL}$ is long
654: enough for life to develop, the tidal working of the body may supply
655: enough heat to drive plate tectonics after radiogenic heating has become negligible. Therefore these low-mass planets, previously thought to be
656: uninhabitable, may, in fact, be good locations for life. Future work should
657: explore this possibility.
658:
659: Perhaps the most distressing aspect of this work (from a SETI
660: perspective) is the prediction that planets can be habitable long
661: enough for complex life to develop, but then that life is extinguished
662: by tides. Yet this work suggests that such a ``tidal extinction'' may
663: occur on some planets around low-mass stars.
664:
665: The results presented here may have significant implications in the
666: context of the Gaia hypothesis (Lovelock \& Margulis 1974), which
667: proposes that biology can change the physical conditions on a planet,
668: such as its atmosphere, so as to promote the continuing existence of
669: life. In other words, evolution favors organisms that help maintain a
670: habitable environment. According to that hypothesis, an inhabited
671: planet may avert a tidal extinction due to its organisms developing
672: adaptations that not only change themselves, but also change the atmosphere such that
673: the sterilizing effects of increased stellar radiation are
674: mitigated. Therefore the Gaia hypothesis predicts that we may observe
675: biosignatures on planets that are on orbits interior to the nominal
676: HZ, if they were formerly in the HZ. In that case we might need to
677: refine the definition of the HZ into two branches: the ``physical'' HZ
678: (\ie the definition proposed by Kasting, Whitmire \& Reynolds [1993]
679: that applies to planets before life gains a foothold), and the
680: ``biologically-extended'' HZ (\ie a larger region in which evolved
681: life can stave off sterilization). For example, planet Gl 581 c was
682: probably never in the physical HZ, but if it once had been, then it
683: might have remained habitable even after reaching its current orbit,
684: according to the Gaia hypothesis. The detection of other similar new
685: planets may provide observational tests of the Gaia hypothesis. The
686: detection of such a planet would be a landmark event, not only in the
687: detection of life beyond the Solar System, but because it will reveal
688: fundamental features of life, evolution, and planetary habitability.
689:
690: We conclude that tides can move habitable planets into non-habitable
691: orbits and described how this phenomenon is relevant for astrobiological models. Future work into the search for habitable worlds, through
692: planet-hunting, atmospheric modeling, geophysical modeling, etc.,
693: should bear in mind the effects of tidal evolution of planets around M
694: stars.
695:
696: \medskip
697: This work was funded by PG\&G grant NNG05GH65G. S.N.R. was supported by an appointment to the NASA Postdoctoral Program t the University of Colorado Astrobiology Center, administered by Oak Ridge Associated Universities through a contract with NASA.
698:
699: \references
700: Adams, F.C. \& Laughlin, G. (2006) Effects of Secular Interactions in Extrasolar Planetary Systems. \textit{Astrophys. J.} 649, 992-1003.\\
701: Barnes, R. \& Greenberg, R. (2006) Extrasolar Planetary Systems Near a Secular Separatrix. \textit{Astrophys. J.} 638, 478-487.\\
702: ---------------. (2007) Stability Limits in Resonant Planetary Systems. \textit{Astrophys. J. Lett.}, 665, L67-L70.\\
703: Barnes, R. \& Quinn, T.R. (2004) The (In)stability of Planetary Systems. \textit{Astrophys. J.} 611, 494-516.\\
704: Bonfils, X. \etal (2005) The HARPS search for southern extra-solar planets. VI. A Neptune-mass planet around the nearby M dwarf Gl 581. \textit{Astron. \& Astrophys. Lett.} 443, L15-L18.\\
705: Butler, R.P. \etal (2006) Catalog of Nearby Exoplanets. \textit{Astrophys. J.} 646, 505-522.\\
706: Chambers, J.E. (2004) Planetary accretion in the inner Solar System. \textit{Earth \& Plan. Sc. Lett.} 223, 241-252.\\
707: Delfosse, X. \etal (1998) Rotation and chromospheric activity in field M dwarfs. \textit{Astron. \& Astrophys.} 331, 581-595.\\
708: Deming, D.(2007) Spitzer Transit and Secondary Eclipse Photometry of GJ 436b. \textit{Astrophys. J.} 667, L199-L202.\\
709: Dickey, J.O. \etal (1994) Lunar Laser Ranging - a Continuing Legacy of the Apollo Program. \textit{Science} 265, 482-490.\\
710: Eggleton, P.P \etal (1998) The Equilibrium Tide Model for Tidal Friction. \textit{Astrophysical J.} 499, 853-870.\\
711: Fortney, J.J., Marley, M.S. \& Barnes, J.W. (2007) Planetary Radii across Five Orders of Magnitude in Mass and Stellar Insolation: Application to Transits. \textit{Astrophys. J.} 659, 1661-1672.\\
712: Franck, S. \etal (2000) Determination of habitable zones in extrasolar planetary systems: Where are Gaia's sisters? \textit{J. of Geophys. Research} 105, 1651-1658.\\
713: Greenberg, R. (1977) Orbit-orbit resonances among natural satellites. In \textit{Planetary Satellites}, edited by J. Burns. Arizona UP, Tucson, AZ.\\
714: Gillon, M. \etal (2007) Accurate Spitzer infrared radius measurement for the hot Neptune GJ 436b. \textit{Astron. \& Astrophys.} 471, L51-L54.\\
715: Goldreich, P. (1966) Final spin states of planets and satellites. \textit{Icarus} 71, 1-7.\\
716: Goldreich, P. \& Soter, S. (1966) Q in the Solar System. \textit{Icarus} 5, 375-389.\\
717: Hut, P. (1981) Tidal evolution in close binary systems. \textit{Astron. \& Astrophys.} 99, 126-140.\\
718: Jackson, B., Greenberg, R. \& Barnes, R. (2007) Tidal Evolution of Close-in Extrasolar Planets. \textit{Astrophys. J.}, submitted.\\
719: Kaltenegger, L., Traub, W.A., \& Jucks, K.W. (2007) Spectral Evolution of an Earth-like Planet. \textit{Astrophys. J.} 658, 598-616.\\
720: Kasting, J.F. \& Catling, D. (2003) Evolution of a Habitable Planet. \textit{Ann. Rev. of Astron. \& Astrophys.} 41, 429-463.\\
721: Kasting, J.F., Whitmire, D.P., \& Reynolds, R.T. (1993) Habitable Zones around Main Sequence Stars. \textit{Icarus} 101, 108-128.\\
722: Kaula, W.M. (1964) Tidal Dissipation by Solid Friction and the Resulting Orbital Evolution. \textit{Rev. Geophysics} 2, 661-685.\\
723: Kiang, N.Y. \etal (2007) Spectral Signatures of Photosynthesis. II. Coevolution with Other Stars And The Atmosphere on Extrasolar Worlds. \textit{Astrobiology} 7, 252-274.\\
724: Lambeck, K. (1977) Tidal Dissipation in the Oceans: Astronomical, Geophysical and Oceanographic Consequences. \textit{Phil. Trans. of the Royal Soc. of London A} 287, 545-594.\\
725: Laughlin, G., Bodenheimer, P. \& Adams, F.C. (2004) The Core Accretion Model Predicts Few Jovian-Mass Planets Orbiting Red Dwarfs. \textit{Astrophys. J. Lett.} 612, L73-L76.\\
726: Lee, M.-H. \& Peale, S.J. (2002) Dynamics and Origin of the 2:1 Orbital Resonances of the GJ 876 Planets. \textit{Astrophys. J.} 567, 596-609.\\
727: L\'eger, A. \etal (2004) A new family of planets? ``Ocean-Planets''. \textit{Icarus} 169, 499-504.\\
728: Lissauer, J.J. (2007) Planets Formed in Habitable Zones of M Dwarf Stars Probably Are Deficient in Volatiles. \textit{Astrophys. J. Lett.} 660, L159-L162.\\
729: Lovelock, J.E. \& Margulis, L. (1974) Homeostatic tendencies of the Earth's atmosphere. \textit{Origins of Life and Evolutions of Biospheres} 5, 93-103.\\
730: Mardling, R.A (2007) Long-term tidal evolution of short-period planets with companions. \textit{MNRAS}, accepted.\\
731: Mardling, R.A. \& Lin, D.N.C. (2002) Calculating the Tidal, Spin, and Dynamical Evolution of Extrasolar Planetary Systems. \textit{Astrophys. J.} 573, 829-844.\\
732: -----------. (2004) On the Survival of Short-Period Terrestrial Planets. \textit{Astrophysical J.} 614, 955-959.\\
733: Naef, D. \etal (2007) The HARPS search for southern extra-solar planets. IX. Exoplanets orbiting HD 100777, HD 190647, and HD 221287. \textit{Astron. \& Astrophys.} 470, 721-726.
734: Ogilvie, G.I. \& Lin, D.N.C. (2007) Tidal Dissipation in Rotating Solar-Type Stars. \textit{Astrophys. J.} 661, 1180-1191.\\
735: Peale, S.J. (1977) Rotation Histories of the Natural satellites. In \textit{Planetary Satellites}, edited by J. Burns. Arizona UP, Tucson, AZ.\\
736: Peale, S.J., Cassen, P, \& Reynolds ,R.T. (1979) Melting of Io by tidal dissipation. \textit{Science} 203, 892-894.\\
737: Rasio, F.A. \etal (1996) Tidal Decay of Close Planetary Orbits. \textit{Astrophys. J.} 470, 1187-1191.\\
738: Raymond, S.N., Quinn, T. \& Lunine, J.I. (2004) Making other earths: dynamical simulations of terrestrial planet formation and water delivery. \textit{Icarus} 168, 1-17.\\
739: ------------. (2007) High-Resolution Simulations of The Final Assembly of Earth-Like Planets. 2. Water Delivery And Planetary Habitability. \textit{Astrobiology} 7, 66-84.\\
740: Raymond, S.N., Scalo, J. \& Meadows, V. (2007) A decreased probability of habitable planet formation around low-mass stars. \textit{Astrophys. J.} 669, 606-614.\\
741: Regenauer-Lieb, K., Yuen, D.A. \& Barlund, J. (2001) The Initiation of Subduction: Criticality by Addition of Water? \textit{Science} 294, 578-580.\\
742: Ribas, I., private communication.\\
743: Sagan, C. \etal (1993) A Search for Life on Earth from the Galileo Spacecraft. \textit{Nature} 365, 715-721.\\
744: Scalo, J. \etal (2007) M Stars as Targets for Terrestrial Exoplanet Searches And Biosignature Detection. \textit{Astrobiology} 7, 85-166.\\
745: Selsis, F., private communication.\\
746: Selsis, F. \etal (2007) Habitable planets around the star Gl581? \textit{Astron. \& Astrophys.}, accepted.\\
747: Segura, A. \etal (2005) Biosignatures from Earth-Like Planets Around M Dwarfs. \textit{Astrobiology} 5, 706-725.\\
748: Sotin, C., Grasset, O., \& Moquet, A. (2007) Mass/radius for extrasolar Earth-like planets and ocean planets. \textit{Icarus}, in press.\\
749: Tarter, J.C \etal (2007) A Reappraisal of The Habitability of Planets around M Dwarf Stars. \textit{Astrobiology} 7, 30-65.\\
750: Tinetti, G., Rashby, S. \& Yung, Y.L. (2006) Detectability of Red-Edge-shifted Vegetation on Terrestrial Planets Orbiting M Stars. \textit{Astrophys. J. Lett.} 644, L129-L132.\\
751: Tinetti, G. \etal (2006a) Detectability of Planetary Characteristics in Disk-Averaged Spectra. I: The Earth Model. \textit{Astrobiology} 6, 34-47.\\
752: -----------. (2006b) Detectability of Planetary Characteristics in Disk-Averaged Spectra II: Synthetic Spectra and Light-Curves of Earth. \textit{Astrobiology} 6, 881-900.\\
753: von Bloh, W. \etal (2007) The habitability of super-Earths in Gliese 581. astro-ph/07053758.\\
754: Udry, S. \etal (2007) The HARPS search for southern extra-solar planets XI. Super-Earths (5 \& 8 $M_\oplus$) in a 3-planet system. \textit{Astron. \& Astrophys. Lett.} 469, L43-L47.\\
755: Valencia, D., Sasselov, D.D., \& O'Connell, R.J. (2007) Radius and Structure Models of the First Super-Earth Planet. \textit{Astrophys. J.} 656, 545-551.\\
756: Walker, J.C.G., Hays, P.B. \& Kasting, J.F. (1981) A negative feedback mechanism for the long-term stabilization of Earthy surface temperature. \textit{J. Geophys. Res.} 86, 9776-9782.\\
757: Williams, D.M., Kasting, J.E. \& Wade, R.A. (1997) Habitable moons around extrasolar giant planets. \textit{Nature} 385, 234-236.\\
758: Williams, D.M. \& Pollard, D. (2002) Earth-like worlds on eccentric orbits: excursions beyond the habitable zone. \textit{Int. J. of Astrobiology} 1, 61-69.\\
759: Yoder, C.F. (1995) Astrometric and geodetic properties of Earth and the solar system. In \textit{Global Earth Physics. A Handbook of Physical Constants}, edited by T. Ahrens, American Geophysical Union, Washington, D.C.\\
760:
761: \clearpage
762: \begin{center}Table 1: Best-Fit Orbital Elements of the Gl 581 Planetary System
763: \end{center}
764: \begin{tabular}{ccccccc}
765: \hline
766: Planet & $m$ (M$_{\oplus}$) & $P$ (d) & a (AU) & $e$ & $\varpi$ ($^o$) & $T_{peri}$ (JD)\\
767: \hline\hline
768: b & 15.7 & 5.363 & 0.041 & 0.02 & 273 & 2452998.76\\
769: c & 5.03 & 12.932 & 0.073 & 0.16 & 267 & 2452993.38\\
770: d & 7.7 & 83.6 & 0.25 & 0.2 & 295 & 2452936.9\\
771: \end{tabular}
772:
773: \clearpage
774: Figure 1 -- Semi-major axes that are habitable as a function of
775: stellar mass and eccentricity from Eqs.\ (\ref{eq:lin} --
776: \ref{eq:luminosity}). The shading represents
777: different choices of cloud cover for the limits of the HZ. Darkest
778: gray assumes no cloud cover, medium gray assumes 50\% cloud cover,
779: light gray assumes total cloud cover, and white is uninhabitable.
780:
781: Figure 2 -- Contours of $\tau_{HL}$ of terrestrial planets, that begin
782: at the inner edge of the 0\% cloud cover EHZ, as a function of stellar
783: mass and initial eccentricity $e_0$. The terrestrial planets have a mass of
784: 5 $M_\oplus$ (top left), 1 $M_\oplus$ (top right), 0.5 $M_\oplus$
785: (bottom left), and 0.1 $M_\oplus$ (bottom right), and a radius
786: determined by Eq.\ (\ref{eq:rp}). Planets with initial eccentricities
787: below 0.53, or orbiting stars with a mass greater then 0.35 $M_\odot$
788: are habitable for at least 10 Gyr. Lower mass planets tend to have
789: longer habitable lifetimes.\\
790:
791: Figure 3 -- Evolution of the semi-major axes (solid line) for two hypothetical planets relative to EHZ boundaries (shown by the shading, \cf Fig.\ \ref{fig:hz}). These planets
792: are in orbit about a 0.15 $M_\odot$ star. \textit{Top:} Evolution of a
793: planet with $e_0 = 0.52$. This planet's semi-major axis decreases
794: until its eccentricity reaches zero after 750 million years, at which
795: point the tidal evolution effectively stops. The tidal evolution
796: stopped just before the planet crossed the inner edge of the EHZ, and
797: the planet will be habitable indefinitely. \textit{Bottom:}
798: Evolution of a planet with $e_0 = 0.54$. Although very similar to the
799: other planet's evolution, this planet's larger initial eccentricity
800: results in more torque so that when $e$ reaches zero, the planet is
801: just interior to the 50\% cloud cover HZ. The habitable lifetime for this planet is $7.5 \times 10^8$ years.\\
802:
803: Figure 4 -- \textit{Top:} Orbit-averaged stellar flux
804: received by a tidally evolving planet relative to that of the Earth
805: $F_\oplus$ (solid line), see Eq.\ (\ref{eq:flux}), as well as at periastron and
806: apoastron (dotted lines). The planet orbits a 0.2 M$_\oplus$ star with
807: initial orbital elements of $a_0 = 0.0877$, and $e_0 = 0.84$. The tidal
808: lifetime of the planet, 4.5 Gyr in this case, corresponds to the age of
809: the Earth, and is shown by the dashed line. Note, however, the extreme
810: difference between flux at periastron and apoastron (a factor of
811: 200). \textit{Bottom:} Ratio of the rotation frequency of the tidally
812: evolving planet to its instantaneous mean motion, \cf Eq.\
813: (\ref{eq:spinlock}). Initially the planet rotates nearly 10 times faster
814: than it revolves, but after $\sim 6$ Gyr, the orbit has circularized and
815: $\Omega = n$.\\
816:
817: Figure 5 -- Time since Gl 581 c would have been habitable
818: as a function of its current eccentricity. We set $R_c = 1.8
819: R_\oplus$, and consider $Q$ values of 100 and 21.5. The $Q = 21.5$
820: case permits habitability as recently as 3 Gyr ago.\\
821:
822: Figure 6 -- Evolution of Gl 581 c for three different possible radius
823: values. The shading represents different definitions of the HZ, \cf Fig.\ \ref{fig:hz}.\\
824:
825: Figure 7 -- Evolution of $a_b$ and $e_b$ for various choices of
826: $Q_b$, assuming $k = 3/2$. The evolution is negligible unless $Q_b \lsim 10^5$.\\
827:
828: Figure 8 -- Apsidal evolution of the Gl 581 b and c planets, see Table
829: 1. \textit{Top:} The apsidal oscillation is
830: circulation. \textit{Bottom:} Evolution of $e_b$ (solid line) and
831: $e_c$ (dashed line). The orbits are currently near the extrema of
832: their eccentricity values.\\
833:
834: Figure 9 -- Evolution of planets b (dashed line), c (solid line) and
835: the instantaneous location of b's outer 3:1 mean motion resonance
836: (dotted line) for $Q_b = 4\times 10^4, k = 3/2$. In this case planet c would have
837: just avoided the 3:1 resonance. The shading corresponds to different definitions of the HZ, \cf Fig.\ \ref{fig:hz}. 10 Gyr ago the configuration was
838: unstable.\\
839:
840: Figure 10 -- Evolution of planets b (short-dashed line), c (solid line) and the instantaneous location of b's outer 3:1 mean motion resonance (dotted line) for $Q_b = 10^6, R_b = 3.74 R_\oplus, k = 3/2$ and $Q_c = 100, R_c = 1.8 R_\oplus, k = 0.3$. In our model, this evolution is the most believable, and prevents planet c from crossing the 3:1 mean motion resonance and reaching the 50\% cloud cover habitable zone (medium grey). This evolution permits stable interactions 10 Gyr ago.
841:
842: \clearpage
843: \begin{figure}
844: \plotone{hz.ps}
845: \figcaption[]{\label{fig:hz} \small{}}
846: \end{figure}
847:
848: \clearpage
849: \begin{figure}
850: \plotone{thz4.ps}
851: \figcaption[]{\label{fig:thz4} \small{}}
852: \end{figure}
853:
854: \clearpage
855: \begin{figure}
856: \plotone{evol.ps}
857: \figcaption[]{\label{fig:evol} \small{}}
858: \end{figure}
859:
860:
861: \clearpage
862: \begin{figure}
863: \plotone{flux.ps}
864: \figcaption[]{\label{fig:flux} \small{}}
865: \end{figure}
866:
867: \clearpage
868: \begin{figure}
869: \plotone{tago.ps}
870: \figcaption[]{\label{fig:e} \small{}}
871: \end{figure}
872:
873: \clearpage
874: \begin{figure}
875: \plotone{rp.eps}
876: \figcaption[]{\label{fig:rp} \small{}}
877: \end{figure}
878:
879: \clearpage
880: \begin{figure}
881: \plotone{gl581b.q.ps}
882: \figcaption[]{\label{fig:qb} \small{}}
883: \end{figure}
884:
885: \clearpage
886: \begin{figure}
887: \plotone{gl581bc.secular.ps}
888: \figcaption[]{\label{fig:secular} \small{}}
889: \end{figure}
890:
891: \clearpage
892: \begin{figure}
893: \plotone{gl581.3_1.ps}
894: \figcaption[]{\label{fig:3_1} \small{}}
895: \end{figure}
896:
897: \clearpage
898: \begin{figure}
899: \plotone{gl581.bestrewind.ps}
900: \figcaption[]{\label{fig:bestrewind} \small{}}
901: \end{figure}
902:
903: \end{document}
904: