1: \documentclass[preprint,12pt]{aastex}

2: \usepackage{natbib}

3: \bibliographystyle{apj}

4: \newcommand{\cmt}{{\rm cm}^{-3}}

5: \newcommand{\yr}{{\rm yr}}

6: \newcommand{\pc}{{\rm pc}}

7: \newcommand{\kpc}{{\rm kpc}}

8: \newcommand{\msun}{M$_\odot$}

9: \newcommand{\kms}{km s$^{-1}$}

10: \newcommand{\Mkms}{M$_\odot$ km s$^{-1}$}

11: \newcommand{\efour}{$\times 10^4$}

12: \newcommand{\efive}{$\times 10^5$}

13: \newcommand {\apgt} {\ {\raise-.5ex\hbox{$\buildrel>\over\sim$}}\ }

14: \newcommand {\aplt} {\ {\raise-.5ex\hbox{$\buildrel<\over\sim$}}\ } 

15: \shorttitle{Feedback in Disk Galaxies}

16: \shortauthors{Shetty and Ostriker}

17: 

18: \begin{document}

19: \title{Cloud and Star Formation in Disk Galaxy Models with Feedback}

20: 

21: \author{Rahul Shetty\altaffilmark{1,2} and Eve C. Ostriker\altaffilmark{1}}

22: \altaffiltext{1}{Department of Astronomy, University of Maryland, College Park,

23:   MD 20742} 

24: \altaffiltext{2}{Harvard-Smithsonian Center for Astrophysics, 60

25:   Garden Street, Cambridge, MA 02138}

26: \email{rshetty@cfa.harvard.edu, ostriker@astro.umd.edu}

27: 

28: \begin{abstract}

29: 

30: 

31:   We include feedback in global hydrodynamic simulations in order to

32:   study the star formation properties, and gas structure and dynamics,

33:   in models of galactic disks.  In previous work, we studied the

34:   growth of clouds and spiral substructure due to gravitational

35:   instability.  We extend these models by implementing feedback in

36:   gravitationally bound clouds: momentum (due to massive stars) is

37:   injected at a rate proportional to the star formation rate.  This

38:   mechanical energy disperses cloud gas back into the surrounding ISM,

39:   truncating star formation in a given cloud, and raising the overall

40:   level of ambient turbulence.  Propagating star formation can however

41:   occur as expanding shells collide, enhancing the density and

42:   triggering new cloud and star formation.  By controlling the

43:   momentum injection per massive star and the specific star formation

44:   rate in dense gas, we find that the negative effects of high

45:   turbulence outweigh the positive ones, and {\it in net} feedback

46:   reduces the fraction of dense gas and thus the overall star

47:   formation rate.  The properties of the large clouds that form are

48:   not, however, very sensitive to feedback, with cutoff masses of a

49:   few million \msun, similar to observations.  We find a relationship

50:   between the star formation rate surface density and the gas surface

51:   density with a power law index $\sim$2 for our models with the

52:   largest dynamic range, consistent with theoretical expectations for

53:   our model of disk flaring.  We point out that the value of the

54:   ``Kennicutt-Schmidt'' index found in numerical simulations (and

55:   likely in nature) depends on the thickness of the disk, and

56:   therefore a self-consistent determination must include turbulence

57:   and resolve the vertical structure.  With our simple feedback

58:   prescription (a single combined star formation event per cloud), we

59:   find that global spiral patterns are not sustained; less correlated

60:   feedback and smaller scale turbulence appear to be necessary for

61:   spiral patterns to persist.

62: 

63: \end{abstract}

64: 

65: \keywords{galaxies: ISM -- ISM: kinematics and dynamics -- ISM: star

66:   formation -- turbulence}

67: 

68: \section{Introduction\label{introsec}}

69: A crucial intermediary for the formation of stars in the ISM is the

70: gaseous cloud.  Stars form deep within Giant Molecular Clouds (GMCs),

71: and GMCs themselves may be embedded in larger molecular and atomic

72: structures, which are referred to as giant molecular associations

73: (GMAs) and superclouds \citep{Vogel88,ElmElm83}.  The dispersal of

74: cloud gas, resulting from the ionizing radiation from newly born

75: stars, stellar winds, and supernovae, limits the lifetimes of GMCs and

76: therefore determines their net star formation efficiencies.

77: Supernovae (SN) also play a significant role in maintaining and/or

78: determining the thermal phase balance of the ISM

79: \citep{CoxSmith74,McKeeOstriker77,NormanIkeuchi89,deAvillezBreitschwerdt04},

80: and simple estimates suggest that SN may be the main source of

81: turbulence, at least in the diffuse ISM

82: \citep[e.g.][]{Spitzer78,MaclowKlessen04}.  Turbulence in both the

83: diffuse and dense ISM is in turn considered one of the primary

84: mechanisms regulating star formation

85: \citep[e.g.][]{ElmScalo04a,Ballesteros-Paredes07PPV,McKeeOstriker07}.

86: Since feedback from star formation is linked to the formation,

87: evolution, and destruction of GMCs, the overall process may be

88: self-regulating.

89: 

90: The formation and growth of clouds depends on the gravitational

91: stability of the diffuse gaseous environment.  In disk galaxies,

92: galactic rotation and thermal pressure, among other factors, act to

93: oppose the growth of self-gravitating perturbations.  The Toomre $Q$

94: parameter indicates the susceptibility of axisymmetric perturbations

95: to grow in uniform thin disks: for $Q<1$, the surface density is

96: sufficiently large for gas self-gravity to overwhelm the restoring

97: effects of Coriolis forces and pressure \citep{Toomre64}.  Non-linear

98: simulations have shown that for non-axisymmetric perturbations, and

99: including the effects of disk thickness and stellar gravity, the

100: threshold is $Q\approx1.5$ \citep{KOStone02,KO01,KO07,LiMacKlessen05}.

101: The observed drop off in star formation activity traced by H$\alpha$ at

102: large radii supports the idea that stars preferentially form in gravitationally

103: unstable regions with densities above a critical value

104: \citep{Kennicutt89,MartinKenn01}.\footnote{Deep H$\alpha$

105:   \citep{Ferguson98} and UV observations \citep{Thilker07,Boissier07}

106:   indicate that a fraction of spiral galaxies have extended outer-disk

107:   star formation, but at much lower levels than in inner disks.  In

108:   high-$Q$ environments, clouds may grow to become

109:   self-gravitating by successive inelastic collisions.}  In

110: general, magnetic fields cannot prevent but only slow the collapse of

111: gas. In conjunction with other physical mechanisms, magnetic fields

112: may in fact enhance instability, as is the case when the

113: magneto-rotational instability \citep[MRI,][]{KOStone03} is present,

114: or via the magneto-Jeans instability \citep[MJI,][]{KOStone02}.

115: 

116: 

117: Star formation must commence soon after the gas accumulates to form

118: massive clouds, because almost all GMCs contain stars

119: \citep{BlitzPPV07}.  Ionizing radiation from newly formed stars

120: subsequently dissociates the molecules, and H II region expansion

121: disperses the surrounding gas; some fraction of the gas may remain

122: molecular, but in unbound clouds.  The massive O and B stars reach the

123: end of their lifetimes in $\sim$2 - 20 Myr, with those over 8 \msun\,

124: ending as SN. The cumulative effect of feedback from all the

125: contiguously forming stars contributes to the short estimated GMC

126: lifetimes of $\sim$20 Myr \citep[e.g.][]{BlitzPPV07}.  The feedback

127: from star formation could potentially prevent the formation of stars

128: in nearby regions by driving turbulence and dispersing gas, but it

129: could potentially trigger collapse events as well.  Collisions between

130: SN blast waves can result in sufficiently large densities for gas to

131: collapse and form stars.  It is not yet understood whether (or when)

132: ``positive'' or ``negative'' feedback effects dominate; exploring this

133: issue is one of the goals of the present work.

134:   

135: Despite the host of processes that impact the formation of stars,

136: observations have shown a clear correlation between the star formation

137: rate density $\Sigma_{SFR}$ and the gas surface density $\Sigma$, with

138: power-law forms (in actively star-forming regions)

139: \begin{equation}

140: \Sigma_{SFR}\propto\Sigma^{1+p},

141: \label{KSeq}

142: \end{equation}

143: now known as Kennicutt-Schmidt laws \citep{Schmidt59,Kennicutt98}.

144: Power law indices with $1+p\approx 1-2$ have been found, depending on

145: whether the total gas mass or just the molecular gas mass is included

146: in $\Sigma$ \citep[e.g. Bigiel et al. 2008 in

147: preparation,][]{WongBlitz02,Heyer04,Schuster07,Kennicutt07,Boucheetal07}.

148: These relations have been identified for a wide range of disk galaxies

149: at low and high redshifts.  Both global and local versions of the

150: $\Sigma_{SFR}$ --$\Sigma$ relations have been explored.  In the

151: former, surface densities are globally averaged within some outer

152: radius; in the latter, averages are over radial annuli or smaller

153: regions.  A second empirical law obtained by Kennicutt is

154: $\Sigma_{SFR}\approx 0.1\Sigma/t_{orb}$, where $t_{orb}$ is the

155: local orbital time of the gas.

156: 

157: Many theoretical studies have attempted to explain the observed

158: relations between the star formation rate and the gas surface density.

159: Simple analytic prescriptions can be obtained that depend on the star

160: formation efficiency per cloud free-fall time or cloud lifetime, and

161: yield consistency with the ``orbital time'' empirical relations

162: \citep{McKeeOstriker07}.  Using global 3D numerical simulations

163: including gas self-gravity, a prescription for star formation, and

164: feedback in the form of thermal energy, \citet{TaskerBryan06} found

165: power law slopes in $\Sigma_{SFR}\propto\Sigma^{1+p}$ similar to

166: observed values.  \citet{LiMacKlessenLT05,LiMacKlessen06}, using SPH

167: simulations, found both slopes ($p\sim 0.6$) and normalization similar

168: to those in \citet{Kennicutt98} ($p\approx 0.4$).  Their simulations

169: included gravity and sink particles to track the collapsing gas, but

170: did not treat feedback.  Recently, \citet{RobertsonKravtsov08}

171: performed simulations that included molecular cooling, and found that

172: the power law indices obtained by fitting

173: $\Sigma_{SFR}\propto\Sigma^{1+p}$ are generally steeper if all of

174: the gas, rather than just molecular gas, is included; this is

175: consistent with recent observational results.

176: 

177: In this work we investigate how SN driven feedback affects subsequent

178: star formation in gas disks, including star formation rates.  We model

179: feedback with a direct momentum input, rather than using a thermal

180: energy input (when underresolved, the latter approach suffers from

181: overcooling and the resulting momentum input is too low).  Our work

182: also differs from other recent simulations in our approach to treating

183: disk thickness effects; these can be very important to determining the

184: star formation rate, but direct resolution requiring zones $<5$ pc in

185: size can be prohibitively expensive to implement in global disk

186: models.

187: 

188: The evolution of large gas clouds is also relevant to studies of

189: spiral structure.  In previous work \citep[][hereafter Paper

190: I]{ShettyOstriker06} we simulated global disks with an external spiral

191: potential, and found that gravitational instability causes gas in the

192: spiral arms to collapse to form clouds with masses $\sim 10^7$

193: M$_\odot$, similar to masses of GMAs and HI superclouds.  We found

194: that gas self-gravity is also crucial for the growth of spurs (or

195: feathers), which are interarm features that are connected to the

196: spiral arm clouds \citep[see also][]{KO02}.  Observations have shown

197: that spurs are indeed ubiquitous in grand design galaxies, and are

198: likely connected with large clouds in the spiral arms

199: \citep{Elm80,LaVigneVogelOstriker06}.  If grand design spiral

200: structure is long lasting, as hypothesized by density wave theory

201: \citep[][and references therein]{LinShu64,BertinLin96}, then feedback

202: mechanisms dispersing the spiral arm clouds must nevertheless leave

203: the global spiral pattern intact.  One of the goals of this work is to

204: assess the effect of star formation feedback in massive clouds on the

205: global spiral morphology.

206: 

207: Conversely, the spiral arms also affect the initial formation of

208: clouds, therefore also impacting the star formation process.

209: Observations show that most H$\alpha$ emission in grand design

210: galaxies occurs downstream from the primary dust lanes.  An

211: explanation for these observations is that gas is compressed as it

212: flows through the spiral potential minimum, leading to cloud

213: formation; then at some later time stars form within these compressed

214: gas clouds.  Consensus on the exact nature of spiral arm offsets has

215: not yet been reached, however, owing to both observational limitations

216: and diverse theoretical views on the star formation process.  Further,

217: the relative importance of spiral arm triggering is still not

218: completely understood.  \citet{Vogel88} found that the star formation

219: efficiency (in molecular gas) in the spiral arms of the grand design

220: galaxy M51 is only larger by a factor of a few compared with interarm

221: regions.  Other observational studies comparing star formation rates

222: in grand design spiral galaxies and those without strong spiral

223: structure found similar results \citep[see][and references

224: therein]{Knapen96,Kennicutt98AR}.  As a result, density waves may

225: primarily gather gas in the spiral arms (enhancing the ability to form

226: GMCs), but may not significantly affect the star formation efficiency

227: within any given molecular parcel.  Without a large-scale density

228: wave, a similar fraction of gas might still collapse (per galactic

229: orbit) to form clouds via other mechanisms (including large-scale

230: gravitational instabilities), but not in a coherent fashion.  Here, we

231: explore the differences in cloud formation properties in gaseous disks

232: with and without an external spiral driving mechanism.

233: 

234: In this work we are interested in the effect of feedback from star

235: formation in large clouds, such as GMCs and GMAs, on the star

236: formation rate, as well as on the overall dynamics and subsequent

237: cloud formation in galaxy disks.  This work extends the models

238: presented in Paper I: numerical hydrodynamic simulations of global

239: disks with gas self-gravity.  With the resolutions of our models,

240: massive GMAs do not fragment into smaller GMCs, so significant energy

241: input is required to unbind the gas in these concentrations.  If this

242: energy is provided by star formation feedback, multiple massive stars

243: would be needed to destroy the GMAs.  In this work, we model feedback

244: by considering the impact on large clouds of single energetic events.

245: In practice, this could represent multiple correlated SN; this can

246: also be considered simply as an expedient but cleanly parameterizable

247: feedback model at one extreme of the range of event

248: correlation\footnote{In future work, we intend to explore how the

249:   degree of feedback correlation affects the results.}.  We then study

250: the resulting nature of the turbulent gaseous disk, as well as the

251: formation and evolution of the clouds that form in the turbulent

252: medium.  In the next section, we describe our numerical simulation

253: approach, including model parameters and the feedback algorithm.  We

254: then present and analyze our simulation results in $\S$\ref{simres}.

255: In $\S$\ref{discsum} we discuss our results in the context of other

256: work, and summarize our conclusions.

257: 

258: \section{Modeling Method}

259: 

260: \subsection{Basic Hydrodynamic Equations}

261: 

262: To study the growth and destruction of clouds in a gaseous disk, we

263: simulate the evolution of the gaseous component by integrating the

264: equations of hydrodynamics.  As in Paper I, we include the

265: gravitational potential of the gas.  Our models are two-dimensional,

266: except that vertical structure of the disk is included in the

267: calculation of self-gravity, embodied in a function $f(z)$ (see below

268: and the Appendix).  The governing hydrodynamic equations, including

269: self-gravity, are:

270: 

271: \begin{equation}

272:   \frac{\partial \Sigma}{\partial t} + \nabla \cdot (\Sigma {\bf v} ) =

273:   0,

274: \label{cont}

275: \end{equation}

276: \begin{equation}

277:   \frac{\partial {\bf v}}{\partial t} + {\bf v}\cdot\nabla {\bf v} +

278:   \frac{1}{\Sigma}\nabla \Pi =  - \nabla(\Phi_{ext} + \Phi) - \frac{v_c^2}{R}, 

279: \label{force}\

280: \end{equation}

281: \begin{equation}

282:   \nabla^2\Phi = 4\pi G f(z) \Sigma.

283: \label{Poisson}

284: \end{equation}

285: 

286: Here, $\Sigma$, ${\bf v}$, and $\Pi$ are the gas surface density,

287: vertically averaged velocity, and vertically integrated pressure,

288: respectively, and $v_c$ is the unperturbed circular orbital velocity.  

289: For simplicity, we assume an isothermal equation of

290: state, so that $\Pi = c_s^2\Sigma$, where $c_s$ is the sound speed.

291: The term $\Phi$ is the gaseous self-gravitational potential.  To grow

292: gaseous spiral arms, we include an external spiral potential

293: $\Phi_{ext}$ to model the perturbation produced by the

294: non-axisymmetric stellar distribution, which is specified at time $t$

295: in the inertial frame, by

296: 

297: \begin{equation}

298: \Phi_{ext}(R,\phi;t)=\Phi_{ext,0}\cos[m \phi - \phi_0(R) - m\Omega_pt]

299: \label{spiralpot}

300: \end{equation}

301: where $m$, $\phi_0(R)$, and $\Omega_p$ are the number of arms,

302: reference phase angle, and spiral pattern speed, respectively.  We

303: only consider models with a constant pitch angle $i$, so that 

304: \begin{equation}

305: \phi_0(r) = -\frac{m}{\tan\,i}\ln(R) + constant.

306: \label{refang}

307: \end{equation}

308: 

309: \subsection{Model Parameters}

310: Similar to Paper I, the sound speed $c_s$ and rotational velocity

311: $v_c$ are constant in space and time, $c_s = 7$ \kms, and $v_c$ = 210

312: \kms.  We adopt the code unit of length $L_0 = 1$ kpc.  Using $c_s$ as

313: the code unit for velocity, the time unit $t_0 = L_0 / c_s = 1.4

314: \times 10^8$ years, which corresponds to one orbit $t_{orb} = 2

315: \pi/\Omega_0$ at a fiducial radius $R_0 = L_0 v_c / 2\pi c_s = 4.77$

316: kpc.  Our results will scale to other values of $R_0$ and $L_0$ with

317: the same ratio, as well as to models with the same ratio $v_c / c_s =

318: 30$.

319: 

320: In Paper I, we explored different external spiral potential strengths,

321: \begin{equation}

322: F\equiv\frac{\Phi_{ext,0}m}{v_c^2\tan i}

323: \label{eps}

324: \end{equation}

325: which is the ratio of the maximum radial perturbation force to the

326: radial force responsible for a constant rotational velocity $v_c$.  We

327: found that spurs form in disks with strong external potential

328: strengths.  Since one of our objectives is to assess the evolution of

329: the spurs in disks including feedback, here we only simulate disks

330: with $F$ = 10\%, for both 2 arm and 4 arm spiral galaxies ($m$=2 and

331: $m$=4).  The corotation radius of 25 $L_0$ corresponds to 25 kpc and a

332: pattern speed of 8.4 km s$^{-1}$ kpc$^{-1}$, for spiral models using

333: our fiducial parameters.  We also simulate disks with no external

334: spiral forcing.

335: 

336: In our computation of gas self-gravity, we include the effect of the

337: thickness of the disk via $f(z)$, which also acts as softening.  We

338: assume a Gaussian vertical gas distribution, with scale height $H

339: \propto R$, so the disk flares at larger radii (see Appendix).  For a

340: given surface density, the effective midplane density is given by

341: $\rho_0=\Sigma /( H\sqrt{2\pi})$.  As described in Paper I and

342: \citet{KO07}, including the effect of thickness provides an important

343: stabilizing effect on the disk.  For most of our simulations, we use

344: $H/R$ = 0.01.

345: 

346: As in Paper I, the Toomre parameter $Q_0 \equiv \kappa_0 c_s/(\pi G

347: \Sigma_0)$ and the surface density $\Sigma_0$ at $R_0$ are related by:

348: \begin{equation}

349: \Sigma_0 = \frac{2\sqrt{2} c_s^2 }{G L_0 Q_0 }

350: =

351: \frac{32}{Q_0}\, {\rm M}_\odot\, {\rm pc}^{-2}

352: \left(\frac{c_s}{7\,{{\rm km\, s}^{-1}}}\right)^2

353: \left( \frac{L_0}{{\rm kpc}}\right)^{-1}.

354: \label{inits}

355: \end{equation}

356: For flat rotation curves, the epicyclic frequency $\kappa = \sqrt2

357: \Omega = \sqrt2 v_c/R$.  Our models initially have $\Sigma \propto

358: R^{-1}$, so that $Q$ is constant for the whole disk.

359: 

360: 

361: \subsection{Numerical Methods}

362: Since this work is an extension of previous work, we refer the reader

363: to Paper I for a description of the cylindrical-symmetry version of

364: the ZEUS code \citep{StoneNorman92I,StoneNorman92II} that we use to

365: carry out our simulations.  We use a parallelized version of the

366: hydrodynamic code and gravitational potential solver, allowing us to

367: increase the number of zones in the grid relative to the models of

368: Paper I.  For our standard grid we set the azimuthal range to 0 -

369: $\pi/2$ radians and the radial range to 4 - 11 kpc.  We implement

370: outflow and periodic boundary conditions in the radial and azimuthal

371: directions, respectively.  These models have 1024 radial and 1024

372: azimuthal zones.  Since the radial grid spacing is logarithmic, the

373: resolution varies: the linear resolution in each dimension ($\Delta

374: R$, $R \Delta \phi$) varies from $\sim$(4 pc, 6 pc) in the innermost

375: region to $\sim$(11 pc, 17 pc) at the outer boundary.  These high

376: resolutions allow the Truelove criterion \citep{Truelove97} to be

377: satisfied throughout the simulation as gas collapses to form self

378: gravitating clumps.

379: 

380: In this work we use a different method to compute the gravitational

381: potential from that in Paper I.  Here, we use a method derived from

382: that described by \citet{Kalnajs71} in polar coordinates \citep[see

383: also][]{BT87}.  This method employs the convolution theorem for a disk

384: decomposed into logarithmic spiral arcs.  We implement softening to

385: account for the non-zero thickness of the disk.  We describe the

386: method in detail in the Appendix.

387: 

388: We note that for simulations with the standard grid and including a

389: spiral potential, the limit in azimuth requires that $m$=4 (4 arms).

390: However, we also explore some models with $m$=2 patterns, with the

391: azimuthal range 0 - $\pi$, using twice as many azimuthal zones than

392: the standard grid so that the physical resolution of both simulations

393: are equivalent.

394: 

395: \subsection{Feedback: Event Description and Algorithm \label{feedsec}}

396: Equations (\ref{cont})-(\ref{Poisson}) only describe the flow as gas

397: responds to self gravity, and to the external spiral perturbation, if

398: one is present.  However, those equations do not describe any feedback

399: that would occur after a self-gravitating cloud forms and fragments

400: into smaller-scale structures, ultimately forming stars with a range

401: of masses.  In the real ISM of galaxies, clouds are dispersed by the

402: combination of photo-evaporation by UV radiation from massive stars,

403: and the ``mechanical'' destruction by expanding HII regions and SN.

404: 

405: We include in our simulations a very simple feedback prescription by

406: implementing ``feedback events,'' each representing momentum input

407: from a number of SN (or, alternatively, multiple overlapping expanding

408: HII regions).  The specific SN rate, $R_{SN}$, averaged over all mass

409: $M_{dense}$ above a chosen threshold density in a galaxy is

410: \begin{equation}

411: R_{SN} = \frac{{\rm Number\, of\, Supernovae}}{M_{dense} \cdot {\rm time}},

412: \label{SNrate}

413: \end{equation}

414: where $N_{SN}$ is the number of supernovae.  When this rate is applied

415: to an individual dense cloud of mass $M_{cl}$ with a lifetime

416: $t_{cl}$, the average number of SN in the cloud will be

417: \begin{equation}

418: N_{SN} = R_{SN}\cdot M_{cl}\cdot t_{cl}.

419: \label{Nsn}

420: \end{equation}

421: If the total mass of stars of all masses formed per single SN is

422: $M_{SN}$, and the star formation efficiency over a cloud lifetime is

423: $\epsilon_{SF}$, then

424: \begin{equation}

425: N_{SN} = \epsilon_{SF} \frac{M_{cl}}{M_{SN}}.

426: \label{Nsnalt}

427: \end{equation}

428: Equating expressions (\ref{Nsn}) and (\ref{Nsnalt}), the mean cloud

429: lifetime is

430: \begin{equation}

431: t_{cl} = \frac{\epsilon_{SF}}{R_{SN}M_{SN}}.

432: \label{tcloud}

433: \end{equation}

434: 

435: In a given time interval $\delta t$, such as the time between

436: successive computations in the numerical evolution, the probability

437: $P$ that a cloud (of mean lifetime $t_{cl}$) is destroyed is $\delta t /

438: t_{cl}$.  Thus,

439: \begin{equation}

440: P = \delta t \cdot R_{SN} M_{SN} / \epsilon_{SF}

441: \label{SNprob}

442: \end{equation}

443: In our algorithm, clouds are defined as regions above a chosen density

444: threshold.  If a particular zone is a local density maximum, that zone

445: is selected as the center of the feedback event.  For any such

446: identified cloud, a star formation event is initiated with a

447: probability per timestep given by equation (\ref{SNprob}).  In each

448: cloud that is determined to undergo feedback, gas is evenly spread out

449: in a circular region with a prescribed bounding radius.  Gas in each

450: zone in the circular region is assigned an outward velocity (relative

451: to the center) to expand the feedback ``bubble.''  A constant

452: azimuthal velocity is also added such that total galactocentric

453: angular momentum is conserved.  The velocity profile inside the bubble

454: is proportional to the distance from the bubble center.  For most of

455: our simulations, we choose the radius of the feedback bubble to be 75

456: pc, which corresponds to 12-23 pixels, depending on the radial

457: location.  In this way, the initially collapsing cloud gas is forced

458: back into the surrounding ISM.

459: 

460: In our simulation, we only consider the isothermal expansion of the

461: clouds, since we assume an isothermal equation of state.  Thus, we can

462: only consider the net energy input at a stage when expansion of the

463: shell has become strongly radiative.  Numerical simulations show that

464: for a single SN of energy $E_{SN} \approx 10^{51}$ ergs, the radial momentum

465: during the radiative stage is $P_{rad} \approx 3-5\times10^{5}$

466: M$_\odot$ \kms \citep{Chevalier74,Cioffietal88}.  During the

467: subsequent evolution of the bubble, the shell momentum $P_{sh}$ is

468: conserved, and is equal to $P_{rad}$. Wind-driven and pressure-driven

469: H II region bubbles similarly are accelerated to reach a final

470: momentum $P_{rad}$.

471: 

472: For a total number of massive stars formed given by equation

473: (\ref{Nsnalt}), and assuming correlation in time, the total momentum

474: applied to the shell is

475: \begin{equation}

476: P_{sh} = N_{SN} P_{rad} = \epsilon_{SF} \frac{M_{cl}}{M_{SN}} P_{rad}

477: \label{SNmom}

478: \end{equation}

479: The shell velocity $V_{sh}$ is

480: \begin{equation}

481: V_{sh} = P_{sh}/M_{sh} = \epsilon_{SF} 

482: \frac{M_{cl}}{M_{sh}} \frac{P_{rad}}{M_{SN}}.

483: \label{shellv}

484: \end{equation}

485: Here, $M_{sh}$ is the sum of $M_{cl}$ and any ambient gas in the

486: (circular) feedback region.  Assuming $M_{SN}$ = 100 M$_\odot$ and

487: $\epsilon_{SF}=0.05$, for $P_{rad} = 3\times10^{5}$ \Mkms, $V_{sh}$ =

488: 150 \kms $\times (M_{cl}/M_{sh})$.

489: 

490: Given our feedback prescription, the two key parameters are the

491: probability per unit time for cloud destruction (eq. [\ref{SNprob}]),

492: and the momentum input in the feedback event (eq. [\ref{SNmom}]).  For

493: the simulations presented here, we explore a range in the rate

494: $R_{SN}$ and in the momentum input per massive star, $P_{rad}$.  The

495: specific SN rate is set either to $R_{SN} = (10^9\,$ M$_\odot \times

496: 50\,\,{\mathrm {yr}})^{-1}= 2 \times 10^{-11}$ \msun$^{-1} \yr^{-1}$

497: (comparable to that in the Milky Way), or ten times that rate (these

498: models are denoted by $R_{SN}=1$ or 10 in Table \ref{standardmods}).

499: Since $M_{SN}/\epsilon_{SF}$ in equation (\ref{SNprob}) appears as its

500: inverse in equation (\ref{SNmom}), we fix $M_{SN} = 100 \,$ M$_\odot$

501: for all simulations, motivated by the initial mass function of

502: \citet{Kroupa01}, and explore variations in $\epsilon_{SF}$.  

503:  Scaling to fiducial values, we then have

504: \begin{equation}

505: t_{cl} = 2.5 \times 10^7 \yr 

506: \left(\frac{\epsilon_{SF}}{0.05}\right)

507: \left(\frac{R_{SN}}{2\times 10^{-11}\, {\rm M}_\odot^{-1}\, \yr^{-1} }\right)^{-1}

508: \end{equation}

509: for the typical lifetime of dense clouds.

510: The momentum $P_{rad}$ is set either to 3.4\efour\ or 3.4\efive\ \Mkms, in

511: order to allow for a range in feedback energy and out-of-plane losses

512: (venting from the galaxy) that reduce $P_{rad}$ for a given energy

513: input. 

514: 

515: We note that for low values of $P_{rad}$, the energy input will not be

516: sufficient to destroy a dense, bound cloud.  In particular, a cloud of

517: surface density $\Sigma_{cl}$ will become unbound only if $V_{sh}

518: \apgt (2G)^{1/2} (\pi \Sigma_{cl} M_{cl})^{1/4}$.  For

519: $\Sigma_{cl}=200$ \msun $\pc^{-2}$ and $M_{cl}=10^6$ \msun, the

520: minimum shell velocity is $\sim 15$\kms. For our larger value of

521: $P_{rad}$, this inequality is comfortably satisfied, but for the

522: smaller value it is not.  We indeed find that for the low $P_{rad}$

523: models, clouds are not destroyed by feedback.  For these models, then,

524: the ratio $t_{cl}\equiv \epsilon_{SF}/(R_{SN}M_{SN})$ becomes the mean

525: interval between (non-destructive) feedback events in a given cloud..

526: 

527: Before any feedback, the spiral models are executed for some time to

528: allow gas to concentrate (due to self gravity) and form clouds in

529: spiral arms.  In simulations without spiral forcing, condensations

530: begin to grow due to an initial 0.1\% density perturbation.  As a

531: result of shear, the first structures that form are large scale

532: flocculent spiral-like features, which we termed ``sheared background

533: features'' in Paper I.  Gas in these features then collapses to form

534: distinct clouds.  Thus, we wait until some threshold density is

535: reached before feedback occurs.  For most models, the threshold

536: density is $\Sigma/\Sigma_0 = 10$ (this sets the threshold $\Sigma$ at

537: 320 $M_\odot$ pc$^{-2}$ for $Q_0$=1, and 160 $M_\odot$ pc$^{-2}$ for

538: $Q_0$=2).  We hereafter refer to any contiguous structure in our

539: simulation with a density above this chosen threshold as a ``cloud,''

540: regardless of whether the given structure hosts a feedback event or

541: not.

542: 

543: We note that with our feedback prescription, the star formation rate

544: is given by the mass in dense gas (i.e. exceeding the threshold

545: surface density) times $R_{SN} M_{SN}$.  This linear relation is

546: supported by the shallow slopes of $\Sigma_{SFR}$ versus

547: $\Sigma_{mol}$ (as observed in CO emission).  In some other recent

548: work

549: \citep[e.g.][]{LiMacKlessenLT05,LiMacKlessen06,TaskerBryan06,TaskerBryan08,RobertsonKravtsov08},

550: the star formation rate is taken as equal to the mass (with density

551: above some threshold) divided by the free-fall time at that density,

552: times some efficiency factor.  Our prescription is therefore

553: equivalent to choosing a ratio of efficiency over free-fall time at

554: the surface density threshold of $\epsilon_{ff}/t_{ff}= R_{SN} M_{SN}

555: = (5\times 10^{7-8}\yr)^{-1}$.  Since the mean internal density within

556: real GMCs (which have surface densities similar to our critical

557: threshold) is $\sim 100\ \cmt$, with corresponding free-fall time of 4

558: Myr, our models would cover a range of star formation efficiencies per

559: free-fall time of $\epsilon_{ff}\sim 1-10$\%.

560: 

561: \section{Simulation Results \label{simres}}

562: 

563: We first present simulations with standard grid parameters, without

564: spiral structure.  We then show results of simulations including

565: spiral structure, as well as simulations pertaining to different

566: radial regions.

567: 

568: Table \ref{standardmods} shows the initial conditions of the standard

569: set of models we

570: present, as well as the relevant parameters controlling the feedback

571: events.  Column (1) lists each model.  Column (2) shows the initial

572: Toomre $Q$ parameter which is initially constant for the whole disk.

573: Column (3) indicates the number of arms, all with $F$=10\%.  Column

574: (4) gives the SN rate, which is required for setting the probability

575: that a feedback event occurs in a cloud (see eq.  [\ref{SNprob}]).

576: Column (5) shows the assumed star formation efficiency, and column (6)

577: gives the adopted momentum input per massive star.  

578: For these models $H/R=0.01$.

579: 

580: 

581: \subsection{Disks without Spiral Structure}

582: 

583: Figure \ref{befSN} shows a snapshot of model with $Q_0$=1, at time

584: $t/t_{orb}$ = 0.84, without an external spiral potential and before

585: any feedback.  As discussed in $\S$\ref{feedsec}, trailing features

586: grow due to the self-gravity and shear in the disk (see Paper I for

587: details).  The most dense structures grow as sheared, trailing

588: features.  It is in these regions where the first SN will occur to

589: disperse the dense gas.

590: 

591: Figure \ref{Q1ASN} shows a snapshot of model Q1A, at time $t/t_{orb}$

592: = 1.125.  For model Q1A, the SN parameters are all at the low end of

593: the range.  At the time of this snapshot, 105 feedback events have

594: occurred, in clouds which have mean $M_{cl} = 1.2\times10^6$

595: M$_\odot$.  The main difference between Figures \ref{Q1ASN} and

596: \ref{befSN} is the shape of the trailing features.  The feedback

597: events have caused the features to become fragmented at some

598: locations.  However, dispersal of gas due to feedback was not

599: sufficient to prevent or reverse the inflow of gas into the high

600: density agglomerations.  Either the SN do not occur rapidly enough, or

601: do not have enough momentum to alter the basic morphology.  Even

602: increasing both the SN rate by a factor of 10, and doubling the star

603: formation efficiency makes little difference; the strong

604: self-gravitational force from the trailing features keeps much of the

605: gas in those structures.  Increasing the SN momentum (or equivalently

606: the velocity) by up to a factor of 8 still does not significantly

607: affect the outcome: much of the gas is contained in the sheared

608: structures at any given time.

609: 

610: It is only when $P_{rad}$ is increased to 3.4\efive\ \Mkms, along with

611: increasing $R_{SN}$ by a factor of 10 and $\epsilon_{SF}$ to 0.05,

612: that we find a significant difference compared to the case Q1A, as in

613: Model Q1D shown in Figure \ref{Q1BSN}.  The velocity is sufficiently

614: large to drive gas away from the density maxima of the trailing

615: structures.  Further, the rate is high enough that a large number of

616: events occur to significantly alter the morphology, in comparison with

617: Figure \ref{Q1ASN}.  Feedback events in this model are so frequent and

618: energetic that collisions between bubbles occur.  In some instances,

619: such collisions create density enhancements that later result in more

620: collapse and subsequent feedback along the bubble interface.  At time

621: $t/t_{orb}$ = 1.125 (Fig. \ref{Q1BSN}(a)), we can still make out the

622: underlying loci of the initial structures formed by gravitational

623: instability and shear, though 537 feedback events have occurred up to

624: this time.  Yet, after an additional 26 Myr and 75 feedback events

625: (Fig. \ref{Q1BSN}(b)), the dominant large scale features do not have a

626: single pitch angle.  Further, the locations of many of the bubbles are

627: clustered.  Though gas is driven away from the initial structures

628: formed before feedback, at later times clouds form in clusters near

629: the initial density maxima, and where feedback bubbles overlap.

630: Qualitatively, the features in the disk, consisting of filaments and

631: bubbles, are similar to the global models including feedback of

632: \citet{WadaNorman01}.  We discuss the masses of the clouds in both

633: ``non-spiral'' and spiral models in the next section

634: ($\S$\ref{spiralsec}).

635: 

636: In disks with $Q_0=2$, sheared features will also grow, but need more

637: time to develop than in the $Q_0=1$ disks.  Due to its relative stability,

638: after $t/t_{orb}=2$ only a few clouds have formed.  As a result,

639: implementing feedback does not affect the majority of the disk.  To

640: study the effect of feedback in $Q_0=2$ disks, another mechanism is

641: necessary to grow clouds everywhere in the disk.  We thus simulate

642: $Q_0$=2 disks with an external spiral potential, and then implement

643: feedback to destroy the spiral arm clouds that grow.

644: 

645: \subsection{Disks with Spiral Structure \label{spiralsec}}

646: 

647: In disks with spiral structure, the stellar spiral potential acts as a

648: source of perturbation; the compression of gas as it flows through the

649: potential eventually leads to the growth of self gravitating clouds.

650: We explore the effect of feedback on the morphology of the gaseous

651: spiral arms and interarm spurs, as well as any subsequent cloud

652: formation.

653: 

654: Figures \ref{spmodsQ1}-\ref{spmodsQ2} show snapshots of models with

655: $m=4$, for $Q_0=1$ and $Q_0=2$, without any feedback.  In the spiral

656: models, the growth of spiral arm clouds occur sooner than clouds

657: formed by natural instabilities in a rotating self-gravitating

658: disk.\footnote{In our models, the amplitude of the spiral perturbation

659:   is ``turned-on'' gradually, reaching the maximum amplitude $F$ at

660:   $t/t_{orb}$ = 1.  Due to this imposed ``turn-on'' time, the growth

661:   rate of GMCs in our models is not representative of actual GMC

662:   formation timescales.}  Figures \ref{spmodsQ1}-\ref{spmodsQ2} show

663: snapshots of models without any feedback, though in the $Q=1$ snapshot

664: (Fig. \ref{spmodsQ1}) the densities have surpassed the threshold

665: density $\Sigma/\Sigma_0=10$ chosen for models with feedback.  Note

666: that while the $Q=1$ model (with strong self-gravity) shows dense gas

667: knots within the arm, the $Q=2$ model (with weaker self-gravity) shows

668: spur-like features; gas does not collapse as promptly.

669: 

670: For both models Q1SA and Q2SA, the SN momenta are insufficient to

671: offset the growth of clouds and spurs resulting from the spiral

672: potential.  After a feedback event, the dispersed cloud gas flows back

673: toward the spiral arm.  As a result, clouds continue to grow over

674: time.  Further, the spurs also continue to grow in density.  Without

675: feedback, self-gravitating spiral arm clouds cause the surrounding gas

676: to flow in with large velocities.  Eventually, the simulations have to

677: be stopped because the Courant time is too small.  The time when the

678: simulation ceases, depending partly on our choice of the minimum

679: acceptable Courant time, also depends on which clouds are (randomly)

680: selected for feedback; clouds that have produced large inflow

681: velocities would have to be dispersed for the simulation to continue

682: to evolve.

683: 

684: We again find that large SN momenta are required to sufficiently

685: disperse clouds so that immediate re-collapse does not occur.  For

686: such models, the SN rate has an effect on the number of subsequent

687: clumps formed.  Figure \ref{Q1HV} shows a snapshot of models Q1SC and

688: Q1SD, $\sim$ 21 Myr after the first feedback events.  At this time, 53

689: feedback events have occurred in model Q1SC, and 540 in model Q1SD.

690: In model Q1SC, it is clear that most, if not all, feedback events

691: originated in the spiral arms.  However, in model Q1SD, many feedback

692: events have occurred in interarm regions.  The spiral arms are not as

693: identifiable, though at this time the remnants of spurs are still

694: identifiable.  Further, model Q1SD contains many more clumps than

695: model Q1SC. The enhanced SN rate has caused the collision of more

696: shell remnants, which lead to formation of self gravitating clumps at

697: the interfaces.  In both cases, feedback events have caused gas to be

698: dispersed from the arms, eventually removing any trace of the

699: underlying spiral potential, as can be seen in Figure \ref{Q1HVlater}.

700: 

701: Figure \ref{clmasses} shows the histogram of the masses of the clouds

702: $M_{cl}$ that hosted feedback events\footnote{From equation

703:   (\ref{Nsnalt}), feedback events in model Q1A and Q1D on average

704:   consist of 300 and 350 SN, respectively.} in models Q1A, Q1D, Q1SA

705: and Q1SD.  In all cases, the maximum mass of the clouds is below

706: $10^7$ M$_\odot$, and the means and medians for the distributions lie

707: in the range $0.5-2.2\times 10^6$ \msun.  In model Q1A, most feedback

708: events have occurred in the large scale sheared features that grow due

709: to gravitational instability.  However, in model Q1D, some fraction of

710: the feedback events have occurred in regions of colliding flows.  The

711: histogram suggests that clouds formed by colliding flows have

712: characteristically lower masses than those formed in the large scale

713: sheared features.  Similarly, in model Q1SA, most feedback events have

714: occurred in the spiral arms, since most clouds form in the arms.  On

715: average, the clouds in model Q1SD have lower masses, with many clouds

716: formed due to colliding flows initiated in earlier feedback events.

717: For a power law in the mass distribution, $dN/d\log M \propto

718: M^{-\alpha}$, the distribution in the high end masses for model Q1SD

719: (below the cutoff at $\log(M)=6.4$) gives $\alpha \sim 0.6$.  This

720: slope and the upper limit in cloud masses is similar to the range and

721: the upper limit in the observed masses of GMCs \citep[see][and

722: references therein]{McKeeOstriker07}.  A histogram of the masses of

723: all clouds at any given time, normalized by the correct probability

724: $\delta t/t_{cl}$, reproduces the overall shape of the histogram of

725: the of masses of clouds with feedback.  A detailed analysis of the

726: cloud mass distribution is not appropriate here because many of the

727: lower mass clouds are not well resolved, and because higher-mass

728: clouds would be subject to turbulent fragmentation that we cannot

729: follow.  Higher resolution simulations are therefore required to

730: obtain more complete cloud mass distributions.  Nevertheless, it is

731: clear that the upper mass limits for clouds in all models are similar

732: to those in real spiral galaxies.

733: 

734: \subsection{Star Formation Properties}

735: 

736: \subsubsection{Star Formation Rates and Turbulence\label{SFR_turb}}

737: For comparison to observations, two quantities of interest are the

738: star formation rate $SFR$ and the turbulent velocity $v_{turb}$. In

739: each simulation, we record each feedback event to determine the $SFR$.

740: For some chosen time bin $\Delta t$, we compute

741: \begin{equation}

742: SFR = \epsilon_{SF} \frac{\sum M_{cl}}{\Delta t},

743: \label{SFR}

744: \end{equation}

745: where $\sum M_{cl}$ is the total mass of all gas in clumps (i.e. above

746: the chosen threshold surface density) that have undergone feedback

747: events in the chosen time interval. (Recall that the mean lifetime of

748: clouds, or the mean interval between star formation events if they are

749: non-destructive, is given by eq. [\ref{tcloud}].)

750: 

751: We define the turbulent velocity as the RMS sum of any non-circular

752: velocities, weighted by the corresponding mass:

753: \begin{equation}

754: v_{turb} = \left ( \frac{\sum (\delta{\bf

755:     v}_{i,j})^2\Sigma_{i,j} A_{i,j}}{\sum\Sigma_{i,j} A_{i,j}} \right)^\frac{1}{2},

756: \label{vturb}

757: \end{equation}

758: where $A_{i,j}$ is the area of each zone and only non-circular

759: velocity components are considered: $\delta{\bf v} = {\bf v} - v_c

760: {\bf \hat\phi}$.  Figure \ref{Q1SFR} shows the star formation rate and

761: turbulent velocity as a function of time, for the $Q_0$=1 models

762: without spiral structure.  The time bin $\Delta t$ for our $SFR$

763: calculation is 3 Myr.  In these models, the first feedback events

764: occur at time $\sim$125 Myr.  However, for the first $\sim$25 Myr

765: after feedback begins, the $SFR$ for all models is only a few

766: M$_\odot$ yr$^{-1}$.  Only $\sim$25 Myr after the first feedback

767: events does the $SFR$ substantially increase, owing to ``propagating''

768: star formation.  Further, the Q1D model with large feedback momenta

769: ($P_{rad}$ = 3.4\efive\ \Mkms) and large SN rate ($R_{SN}=10$) has the

770: $SFR$ increase to $\sim$10 M$_\odot$ yr$^{-1}$.  This occurs because

771: with large velocities and a high global rate, adjacent shells collide

772: and more clouds are formed in the interfaces, which may subsequently

773: undergo star formation.

774: 

775: The bottom panel of Figure \ref{Q1SFR} shows $v_{turb}$, for all

776: feedback models without spiral structure, together with results from a

777: simulation without any feedback.  For the later case, we just allow

778: self-gravity to grow clouds indefinitely.  When we compute $v_{turb}$

779: in the model without feedback considering only the low density gas, we

780: obtain similar values.  This suggests that, before any feedback,

781: large-scale motions from disk self-gravity and shear are the primary

782: sources of turbulence \citep[see][]{KO07}.  The models with low

783: feedback momentum continue the trend of $v_{turb}$ established by the

784: no-feedback case.  In a few instances of enhanced feedback, there is a

785: corresponding jump in $v_{turb}$.  The enhanced $SFR$ at later times

786: for the model with large SN momenta also increases levels of

787: $v_{turb}$.

788: 

789: Figures \ref{Q1SSFR} and \ref{Q2SSFR} show the $SFR$ and $v_{turb}$

790: for the spiral models with $Q_0=1$ and $Q_0=2$.  Comparing Figures

791: \ref{Q1SSFR} (with spiral structure) and \ref{Q1SFR} (without spiral

792: structure), the star formation rate is consistent to within a factor

793: of 2, although slightly larger in some of the spiral models.  The

794: general trends from the models without spiral structure are reproduced

795: in Figures \ref{Q1SSFR} and \ref{Q2SSFR}.  Earlier times are shown in

796: Figure \ref{Q1SSFR}, since the spiral arms cause gas to collapse into

797: clouds sooner.  It is clear that only in models with large SN shell

798: velocities -- and regardless of the input rate $R_{SN}$ --

799: do the turbulent velocities increase appreciably;

800: otherwise, the turbulent velocity (as we have defined it) is dominated

801: by effects from gas self-gravity.

802: 

803: It is interesting to compare results from pairs of models in which one

804: parameter is varied and the others are controlled.  Comparing models

805: Q1SE and Q1SB, both have the same $\epsilon_{SF}=0.05$ and

806: $P_{rad}=3.4 \times 10^4$ \Mkms, but the former has $R_{SN}$

807: larger by a factor 10.  The measured SFR in Q1SE is a factor $\sim 10$

808: larger than that in Q1SB, consistent with the naive expectation that

809: $SFR \propto R_{SN}$.  However, when we compare Q1SD with Q1SC, which

810: again differ in $R_{SN}$ by a factor 10, we find SFR ratios differing

811: only by a factor $\sim 4$.  This same trend is also true for models

812: Q2SD and Q2SC.  The reason for this difference in dependence on

813: $R_{SN}$ is that the E and B models have low $P_{rad}$ and low

814: turbulence levels, whereas the C and D models have higher $P_{rad}$

815: and turbulence.  Thus, stronger feedback causes the scaling of SFR to

816: depart from $SFR \propto R_{SN}$.  We note that since $SFR = R_{SN}

817: M_{SN} M_{dense}$ by definition, the ratios of specific SFR between any two

818: models differ by their ratios of $R_{SN} M_{dense}/M_{tot}$.  Thus,

819: if SFR increases at a rate less than $\propto R_{SN}$, it implies that

820: increasing $R_{SN}$ {\it decreases} the dense gas fraction $M_{dense}/M_{tot}$.

821: 

822: We can directly investigate the effect of turbulence by comparing the

823: pair Q1SB and Q1SC, which have the same $\epsilon_{SF}=0.05$ and

824: $R_{SN}$, but momentum input parameters differing by a factor 10.  As

825: noted above, this increases the turbulence level in Q1SC compared to

826: Q1SB by about 10 \kms.  It also reduces the SFR in Q1SC compared to

827: Q1SB, by a factor $\sim 2-4$.  Similarly, Q1SE has lower $P_{rad}$

828: than Q1SD, and a substantially lower turbulence level.  For this pair,

829: too, the SFR in the lower-turbulence model is higher by a factor $\sim

830: 3-5$.  As discussed in \S \ref{introsec}, in principle turbulence

831: could both enhance star formation (by creating more dense gas in

832: compressions), and suppress star formation (by destroying overdense

833: structures with rarefactions and shear flows).  Examining the

834: evolution of Q1SE indeed shows that feedback events only slightly

835: expand clouds, and collapse subsequently resumes.  On the other hand,

836: clouds in model Q1SD are completely destroyed after a single feedback

837: event.  Evidently, in the models with strong feedback-driven

838: turbulence, the rate of new cloud formation from shell collisions does

839: not compensate for the truncation of star formation when a given cloud

840: is destroyed.

841: 

842: The comparisons of (Q1SB,Q1SC) and (Q1SE,Q1SD) indicate that {\it in

843:   net}, the increase of turbulence reduces star formation.\footnote{We

844:   note that models Q2SB and Q2SC also show the same generic behavior,

845:   but with a smaller difference in the turbulence level, the

846:   suppression of star formation is also lesser.}  Since the specific SFR is

847: proportional to the dense gas fraction if $R_{SN}$ is held fixed,

848: these results imply that the dense gas mass fraction is lower when the

849: turbulence level is higher.

850: 

851: We show the relationship between the mass weighted turbulent velocity

852: and the surface density in Figure \ref{dispsig}.  Most feedback events

853: occur in high density regions.  In the higher density regions, the

854: difference in turbulent velocities (or the velocity dispersions)

855: between models with $P_{rad}$ = 3.4\efour\ \Mkms\ and $P_{rad}$ =

856: 3.4\efive\ \Mkms\ is \apgt 7 \kms.  At lower density regions, where

857: there have been fewer feedback events, the dispersions of all models

858: are comparable.

859: 

860: Figure \ref{turbspec} shows the turbulent power spectrum (power

861: $\propto v^2$) of model Q1D.  The power is shown at constant

862: wavenumbers $k_R$ and $k_\phi$.  The slopes of the power spectra range

863: from -2.5 to -3.  For models that evolve for significant amounts of

864: time, such as model Q1D, the power spectra are relatively independent

865: of time.  These results are consistent with turbulence dominated by

866: numerous shocks, or Burgers turbulence.  From Figure \ref{turbspec},

867: the amplitudes of turbulence evidently decrease at smaller scales.

868: 

869: 

870: The total turbulent amplitudes shown in Figs.

871: \ref{Q1SFR}-\ref{Q2SSFR} represent the velocity dispersion averaged

872: over the whole disk.  For the purposes of assessing turbulent

873: contributions to local disk stability, however, only the level of

874: turbulence within a Jeans length $\sim c_s^2/(G \Sigma)$ is relevant.

875: Furthermore, local observations of turbulence within the Milky Way

876: generally measure velocity dispersions on scales less than the disk

877: thickness.  Thus, it is useful to estimate the turbulent amplitudes at

878: smaller scales than the whole disk.  We do this by running a window

879: (or ``beam'') of 1 kpc or 100 pc over the map, and finding the

880: dispersion of the velocity within this window at locations separated

881: by the window size. When all zones within the window are weighted

882: equally (as is true for the velocity power spectrum), we find that the

883: mean velocity dispersions for model Q1D on scales of 1 kpc and 100 pc

884: are 18 \kms\ and 6.5 \kms, respectively.  When we weight by mass, the

885: respective velocity dispersions are 31 \kms\ and 10 \kms.  The larger

886: values obtained when weighting by mass are indicative of the

887: importance of dense expanding shells in driving the turbulence.

888: 

889: Since turbulence adds to the total momentum flux (the ram pressure

890: acts similarly to the thermal pressure), a common

891: assumption is that the sound speed $c_s$ can be replaced by

892: \begin{equation}

893: c_{eff}^2 = c_s^2 + \sigma_R^2

894: \label{c_effeqn}

895: \end{equation}

896: in the dispersion relations that characterize stability to

897: axisymmetric modes, where $\sigma_R$ is the radial component of the

898: velocity dispersion.  For models Q1SC and Q1SD, which have high

899: $P_{rad}$, we find that the mean values of $\sigma_R$ on kpc scales

900: are 17 and 18 \kms, respectively.  For the corresponding models Q1SB

901: and Q1SE that have low $P_{rad}$, on the other hand, the values of

902: $\sigma_R$ on kpc scales are 10 and 8 \kms, respectively.  Thus, the

903: values of $c_{eff}$ exceed $c_s$ by a factor 1.6 for the

904: low-turbulence models, whereas this increases to a factor 2.7 for the

905: high-turbulence models.  Our results discussed above indicate a

906: decrease in the star formation rate with increasing $c_{eff}$; we

907: discuss theoretical ideas related to this finding in \S\ref{SFpredict}

908: below.

909: 

910: 

911: 

912: \subsubsection{Kennicutt-Schmidt Law}

913: Figure \ref{QKS} shows the local star formation rate per area as a

914: function of mean surface density.  To obtain these points, simulation

915: data were binned in radius and time, of widths 1 kpc and 18 Myr,

916: respectively.  Only models with a sufficient number of points, which

917: is dependent to some degree on the number of feedback events, are

918: shown.  Best fit lines to the data points are also shown.  The rates

919: show considerable scatter, both between models with different

920: parameters, as well as among points from a given model.  However,

921: where a large dynamic range is available, as is the case for the Q1D

922: model and its extension to smaller radii (see below), a power law

923: relation $\Sigma_{SFR} \propto \Sigma^{1+p}$ is quite clear.

924: 

925: The $Q_0=1$ models, both with and without a spiral perturbation, and

926: with different feedback parameters, generally give slopes

927: $1+p\sim1-3$.  Most of the $Q_0=1$ models evolve for sufficiently long

928: times that gas in the first clouds that are formed are allowed to be

929: recycled into subsequently formed clouds several times.  The $Q_0=2$

930: models, on the other hand, give a variety of slopes, and the

931: relationship between the star formation rate and surface density is

932: not as well correlated as in the $Q_0=1$ models.  For the $Q_0=2$

933: models, the number of feedback events is insufficient to affect much

934: of the disk.  As a result, some clouds continue to collapse, and the

935: Courant condition would demand an extremely small time step; at this

936: point, we halt the simulation.  Since the stochastic feedback events

937: do not result in developed turbulence and a steady state is not

938: approached in the $Q_0=2$ models, the SFR as computed is sensitive to

939: model parameters governing the feedback events.

940: 

941: For some of our models, we have also run simulations of the inner

942: regions of disks, with radial extent $R\in0.8-2.2$ kpc.  The other

943: parameters are the same as for the standard models.  The only

944: difference here, besides the radial range, is the initial surface

945: density.  Since $Q_0$ is constant, and $\Sigma_0\propto (Q_0 R)^{-1}$,

946: the initial surface density at all radii is increased by a factor 5

947: compared to the standard models with $R\in4-11$ kpc.

948: 

949: Figure \ref{innercomp} shows the star formation rate as a function of

950: surface density for model Q1D together with the corresponding inner

951: region model.  The larger surface density does indeed lead to higher

952: star formation rates, with a slope $1+p=2.2$ that is similar to the

953: value $1+p=2.4$ of the standard model.  We find similar trends for

954: other inner disk models in comparison with the corresponding standard

955: models.  For comparison, Figure \ref{innercomp} also shows data from

956: \citet{Kennicutt98}.  Each point indicates the globally averaged star

957: formation rate for individual galaxies or their central regions (for

958: starbursts).  Though there is less scatter in the simulation points,

959: the slope of the $\Sigma_{SFR}$ - $\Sigma$ relation from the

960: simulations ($\sim$2.3) is larger than the slope from observational

961: data ($1+p\sim$1.4).  At the low $\Sigma$ end, the model results

962: overlap with the observed points.

963: 

964: \subsubsection{Predicting  Star Formation Times\label{SFpredict}}

965: The star formation (or gas depletion) time $t_{SF}$ for the whole

966: gaseous component of a galaxy is the time required for all the gas to

967: be converted to stars if the star formation proceeds as it has been

968: during a given interval $\Delta t$:

969: \begin{equation}

970: t_{SF} = \Delta t \frac{M_{tot}}{\epsilon_{SF} \sum M_{cl}},

971: \label{deptime}

972: \end{equation}

973: where $M_{tot}$ is the total mass in a given annulus.  This quantity

974: can be measured in our simulations; the summation in equation

975: (\ref{deptime}) is taken over all clouds in which a feedback event has

976: occurred, as in equation (\ref{SFR}).

977: 

978: Observationally, if the SN rate per dense gas mass $R_{SN}$ is known,

979: the star formation (or gas depletion) time can also be estimated based on

980: the total amount of gas and the portion in dense clouds as:

981: \begin{equation}

982: t^\prime_{SF} = \frac{M_{tot}}{M_{dense} M_{SN} R_{SN}},

983: \label{deptimeapp}

984: \end{equation}

985: where $M_{dense}$ is the total mass of gas above some chosen threshold

986: density.  Since $R_{SN}M_{SN} = \epsilon_{SF} / t_{cl}$ from equation

987: (\ref{tcloud}), the results of equations (\ref{deptime}) and

988: (\ref{deptimeapp}) should agree on average.  With our two parameter

989: choices $R_{SN}$ = 1 or 10 (in units 2 $\times 10^{-11}$ \msun$^{-1}$

990: $\yr^{-1}$), this implies $t^\prime_{SF} = (1\ {\rm or}\ 0.1) \times

991: (M_{tot}/M_{dense}) \times 5\times10^8\ \yr$.  We note that if the

992: star formation or gas depletion time were computed only for

993: \textit{dense} gas (with local surface density $\apgt 200$ \msun\

994: $\pc^{-2}$), then for our prescription it would simply be equal to a

995: constant, $t^\prime_{SF}(dense)=(M_{SN} R_{SN})^{-1}=5\times10^{7}$ or

996: $5\times10^{8}\ \yr$ for $R_{SN}$ = 10 or 1, respectively.

997: 

998: 

999: Figure \ref{pltdeptime} shows the star formation time in different radial

1000: annuli for Model Q1D, as a function of $\Sigma$.  The actual 

1001: times, shown by the filled symbols, are computed using equation

1002: (\ref{deptime}), after binning the simulation data in radii of 1 kpc

1003: widths and in time with $t/t_{orb}$ = 0.125 widths.  The open symbols

1004: show the predicted times by applying equation

1005: (\ref{deptimeapp}) on the same binned data.  The predicted times

1006: agree well with the actual times.  We find similar

1007: agreement with all other models.\footnote{Most other models do not run

1008:   for as long as the D models that have high feedback rates, because

1009:   some dense clumps continue to collapse without feedback, eventually

1010:   causing the Courant condition to be violated.}  We also tested the

1011: correlation between $t_{SF}$ and $t_{orb}$, and found no strong

1012: correlation.  This lack of correlation occurs because at later times

1013: the surface density profile no longer resembles the initial $R^{-1}$

1014: profile.  

1015: 

1016: What is expected, on theoretical grounds, for the value of the star

1017: formation time?  Consider the case in which gas cycles

1018: between diffuse (gravitationally unbound) and dense (gravitationally

1019: bound) components.  The diffuse component forms dense clouds at a rate

1020: $M_{diff}/t_{diff}$, and the dense clouds are returned to the diffuse

1021: component plus stars over a cloud lifetime at a rate

1022: $M_{dense}/t_{cl}$.  Here $M_{diff}$ and $M_{dense}$ are the total

1023: diffuse and dense gas masses in an annulus, with corresponding surface

1024: densities when averaged over the area of $\Sigma_{diff}$ and

1025: $\Sigma_{dense}$ (the latter is not to be confused with the surface

1026: density of an individual dense cloud, which is much higher).

1027: Similarly, $\Sigma$ is the surface density corresponding to the total

1028: mass of all the gas $M_{tot}$ in an annulus.  In equilibrium, the

1029: rates into and out of the dense component are equal, so that the star

1030: formation rate per unit area averaged over the annulus is

1031: \begin{equation}

1032: \Sigma_{SFR}=\epsilon_{SF} \frac{\Sigma}{t_{diff}+t_{cl}}=

1033: \epsilon_{SF} \frac{\Sigma_{diff}}{t_{diff}}=

1034: \epsilon_{SF} \frac{\Sigma_{dense}}{t_{cl}}=

1035: R_{SN} M_{SN} \Sigma_{dense}.

1036: \label{rateeq}

1037: \end{equation}

1038: Including all the gas, the star formation timescale using the

1039: definition of equation (\ref{deptimeapp}) (and dropping the prime) is then 

1040: \begin{equation}

1041: t_{SF}=\frac{t_{diff}+t_{cl}}{\epsilon_{SF}}=

1042: \frac{t_{diff}}{\epsilon_{SF}}\times\frac{M_{tot}}{M_{diff}}=

1043: \frac{t_{cl}}{\epsilon_{SF}}\times\frac{M_{tot}}{M_{dense}}.

1044: \label{SFtime}

1045: \end{equation}

1046: Since $t_{cl}/\epsilon_{SF}=(R_{SN} M_{SN})^{-1}= 5\times10^{7}$ or

1047: $5\times10^{8}\ \yr$ is held constant within any given model, the star

1048: formation time for all gas in an annulus is inversely proportional to

1049: the fraction of the gas above the density threshold in that annulus.

1050: If most of the gas is diffuse (as is true in our simulations), then

1051: $t_{diff}\gg t_{cl}$ and $t_{SF}\sim t_{diff}/\epsilon_{SF}$; the star

1052: formation time is set by the typical time required for diffuse gas to

1053: collect into bound clouds.

1054: 

1055: What characteristic values might be predicted for the cloud formation

1056: timescale, $t_{diff}$?  The shortest possible timescale would be that

1057: associated with the fastest-growing Jeans modes in a disk.  For a disk

1058: with semi-thickness $H$ and sound speed $c_s$, the approximate

1059: dispersion relation for in-plane modes is $\omega^2=k^2 c_s^2 - 2 \pi

1060: G \Sigma |k|/(1+|k|H)$ \citep[][Paper I]{KOStone02}. For the

1061: fastest-growing modes (which satisfy $d|\omega^2|/dk=0$) and for $H<

1062: c_s^2/(\pi G \Sigma)$ (i.e. less than the thickness of an isothermal

1063: disk bound only by its own gravity), the inverse of the growth rate is

1064: $0.3-0.5 t_J$, where $t_J= c_s/(G \Sigma)$ is the thin-disk Jeans

1065: length divided by $c_s$.  In reality, rotation, shear, and turbulence

1066: must all affect the cloud growth timescale (see below), but the Jeans

1067: time nevertheless provides a useful reference value.

1068: 

1069: Another reference value for a structure formation timescale that is

1070: frequently used is the free-fall time,

1071: $t_{ff}=(3\pi/32G\rho)^\frac{1}{2}$.  If the surface density and

1072: volume density are related via $\Sigma = \rho H \sqrt{2 \pi}$ (as for

1073: a Gaussian density distribution), then $t_{ff}=(3 \sqrt{2}

1074: \pi^{3/2}H/32G\Sigma)^\frac{1}{2}$.  For our ``thick-disk'' Poisson

1075: solver, $H\propto R$ is adopted, so that $t_{ff} \propto

1076: (R/\Sigma)^{1/2}$.  Our initial profiles follow $R\propto

1077: \Sigma^{-1}$, so that in the initial conditions $t_{ff}\propto

1078: \Sigma^{-1} \propto t_J$.  In particular, for the $Q=1$ case,

1079: $t_{ff}=0.3 t_J$ everywhere initially. Over time, however, the surface

1080: density is spatially rearranged, so that the values of $t_J$ and

1081: $t_{ff}$ are no longer strictly proportional.

1082: 

1083: Figure \ref{t_J-t_g} shows the relationships between the star

1084: formation time and the reference values $t_J$ and $t_{ff}$.  While a

1085: clear correlation is evident for both relations, we find that there is

1086: less scatter in the $t_{SF}-t_J$ relation than in $t_{SF}-t_{ff}$

1087: relation.  Further, many of the data points are consistent with a

1088: linear relationship $t_{SF}=7t_J$, as indicated in the figure.  If we

1089: compare to the prediction $t_{SF}=t_{diff}/\epsilon_{SF}$ and

1090: substitute the value $\epsilon_{SF}=0.05$ used in model Q1D, this

1091: yields $t_{diff}=0.35 t_J$, which agrees with the simple estimate

1092: described above based on self-gravitating instabilities in thick

1093: disks.  This result suggests that, provided the efficiencies of star

1094: formation in GMCs are constant and the disk is dominated by diffuse

1095: gas, the Jeans time in the diffuse gas controls the rate of star

1096: formation.  While this result is quite intriguing, a high dynamic

1097: range in a wider range of disk models is necessary to further

1098: investigate this relationship.

1099: 

1100: We note that in the dispersion relation used to predict $t_{diff}\sim

1101: t_J=c_s/(G\Sigma)$, no account was made for turbulence.  As discussed

1102: in \S \ref{SFR_turb}, the simplest phenomenological modification of

1103: this relation would simply be to substitute $c_s\rightarrow c_{eff}$

1104: (see eq. [\ref{c_effeqn}]).  The results presented in \S\ref{SFR_turb}

1105: which compare SFRs for model pairs with low and high $P_{rad}$, and

1106: hence different $c_{eff}$, are at least semi-quantitatively in support

1107: of this prescription for modifying $t_J$.  There, we found that an

1108: increase of $c_{eff}$ by a factor of $\sim$ 2 is associated with a

1109: decrease in the SFR by a factor $\sim$3.  However, the current models

1110: are not sufficient for a definitive statement.  An important objective

1111: for future work is to test the relation between $t_{SF}$ and the

1112: turbulence level using a more extensive set of models; the velocity

1113: dispersion can be varied by tuning the parameter $P_{rad}$.  A

1114: fundamental understanding of star formation in molecular-dominated

1115: regions of galaxies (where the thermal velocity dispersion is dwarfed

1116: by the turbulent value) will depend on such investigations.

1117: 

1118: Modeling truly three-dimensional disks, with the vertical dimension

1119: fully resolved, would allow for a more complete 

1120: study of the correlations between $t_{SF}$ and the two gravitational

1121: times, $t_J$ and $t_{ff}$.  Depending on the regime, vertical

1122: hydrostatic equilibrium (for an isothermal medium) may be in the limit

1123: dominated by (a) the disk's gaseous self-gravity, so that the 

1124: effective thickness of the ISM is $\Sigma/(2\rho_0)=c_s^2/(\pi G

1125: \Sigma)$, or (b), the disk's stellar gravity, so that the effective thickness 

1126: is $\Sigma/(2\rho_0)=c_s \sigma_* /(2\sqrt{\pi} G \Sigma_* )\propto 

1127: (Q_*/Q)c_s^2/(\pi G \Sigma)$.  Here,

1128: $\sigma_*$ and $\Sigma_*$ are the stellar vertical velocity

1129: dispersion and surface density, respectively, and $Q_*$ is the Toomre

1130: parameter for the stellar disk.  Using these two forms, if gas

1131: dominates the vertical gravity, then $t_{ff}\propto t_J$, whereas if

1132: the stars dominate the vertical gravity, then $t_{ff}\propto t_J(Q_*/Q)^{1/2}$.

1133: If galaxies evolve such that $Q_*/Q$ is constant, then $t_J \propto

1134: t_{ff}$ in either case; it would then be empirically difficult to

1135: establish whether  $t_J$ or $t_{ff}$ is more fundamental for

1136: determining the star formation time. With explicit three dimensional

1137: models, on the other hand, it will be possible to 

1138: study the dependence of $t_{SF}$ on $t_J$

1139: and $t_{ff}$ separately, with $Q_*/Q$ a tunable parameter.  This

1140: represents a very interesting avenue for future research.

1141: 

1142: 

1143: \section{Discussion and Summary \label{discsum}}

1144: 

1145: \subsection{Kennicutt-Schmidt Law in Simulations \label{ksdisc}}

1146: 

1147: The prescription we adopt for star formation in this paper implies a

1148: constant relation between the mass (or mean surface density) of {\it dense}

1149: gas and the rate (or mean surface density) of star formation,

1150: $\Sigma_{SFR}= R_{SN} M_{SN} \Sigma_{dense}$.  Using this

1151: prescription, we then test how the star formation rate scales with the

1152: surface density of \textit{all} the gas.  We find that our simulations

1153: are consistent with scalings $\Sigma_{SFR} \propto \Sigma^{1+p}$ for a

1154: range of power law indices, but with significant scatter.  In part,

1155: both the range of indices and the scatter in many of our models may

1156: arise from transient effects, rather than describing the behavior in a

1157: fully-developed star-forming disk.  Our simulations suggest that

1158: measured star formation properties are subject to transient effects,

1159: thus for meaningful theoretical predictions it is necessary for

1160: systems to evolve well beyond the initial state.

1161: 

1162: For our strong-feedback model that most closely reaches an equilibrium

1163: between cloud formation and destruction and has a large dynamic range

1164: of surface density, we find a fairly tight relationship between

1165: $\Sigma_{SFR}$ and $\Sigma$, with $1+p\sim 2$ (see Fig.

1166: \ref{innercomp}).  This implies the fraction of dense gas follows

1167: $M_{dense}/M_{tot}=\Sigma_{dense}/\Sigma\propto \Sigma$.  If we

1168: interpret this in terms of cloud formation/destruction equilibrium

1169: (cf. eq. \ref{rateeq}), with a constant mean cloud lifetime given by

1170: equation (\ref{tcloud}), this implies a dense gas formation time

1171: $\propto \Sigma^{-1}$.  As discussed in \S\ref{SFpredict}, our

1172: quantitative results are generally consistent with a formation time

1173: for dense gas $\propto t_J$ or $t_{ff}$, which vary (exactly or

1174: approximately) $\propto \Sigma^{-1}$ in our models.

1175: 

1176: In other recent numerical work, star formation prescriptions

1177: $\Sigma_{SFR}\propto \Sigma/t_{ff}$ have been adopted, where either

1178: all of the gas or just high-density gas is included in the right-hand

1179: side.  This would imply $\Sigma_{SFR} \propto \Sigma^{3/2}(G/H)^{1/2}$

1180: for the dependence on surface density and disk thickness.  For a disk

1181: in vertical hydrostatic equilibrium with vertical velocity dispersion

1182: $\sigma_z$, the natural thickness varies as $H\propto

1183: \sigma_z^2/(G\Sigma)$, which would imply $\Sigma_{SFR} \propto

1184: \Sigma^2 G/\sigma_z$.  Thus, a vertically-resolved disk with a constant

1185: vertical velocity dispersion would be expected to yield an index

1186: $1+p=2$.  If the disk thickness is determined not by hydrostatic

1187: equilibrium but in some other way, however, the resulting star

1188: formation rate and the index in the K-S law would depend on the

1189: numerical prescription (or physical process) that sets $H$.  In our

1190: models, we have a flared disk $H\propto R$ and set $\Sigma\propto

1191: R^{-1}$ in our initial conditions, which accounts for the index

1192: $1+p\sim 2$ that we obtain.  If, on the other hand, the value of $H$

1193: were constant in a given simulation (either by design for a

1194: two-dimensional simulation, or as a consequence of limited spatial

1195: resolution in a three-dimensional simulation), then the result would

1196: be $1+p\sim 1.5$.  Thus, limited vertical resolution can potentially

1197: artificially reduce the scaling index in the K-S relation, as measured

1198: from numerical simulations.  A fully-resolved vertical dimension is

1199: therefore required if the star formation prescription is to be based

1200: on a volume density.  In practice, the resolution requirement can be

1201: quite demanding if the disk is dominated by cold atomic or molecular

1202: gas, since $c_s^2/(\pi G \Sigma)=4\pc (T/100{\rm K})(\Sigma/10$ \msun

1203: $\, \pc^{-2})^{-1}$.  This also points to the necessity of incorporating

1204: turbulent processes in three-dimensional models, since observed cold

1205: gas is in fact dominated by turbulent rather than thermal pressure.

1206: If these turbulent effects were not included, the disk thickness would

1207: be unphysically small.

1208: 

1209: \subsection{Model Limitations and Future Prospects}

1210: 

1211: \subsubsection{Spiral Structure \label{spiral_turb}}

1212: 

1213: In spiral models, the external spiral potential is initially the

1214: primary driver for enhancing the density, leading eventually to the

1215: growth of clouds.  In models that evolve for a significant amount of

1216: time, soon after feedback and the dispersal of cloud gas the global

1217: spiral pattern is disrupted, and eventually vanishes.  With the simple

1218: feedback prescription that we have adopted, we were unable to simulate

1219: a spiral galaxy in which the global spiral pattern is maintained

1220: simultaneously as cloud gas is returned to the ISM through feedback.

1221: 

1222: If the arms truly are long lasting, then either the spiral potential

1223: is much stronger than in our models ($F>>10\%$), and/or the real

1224: feedback events are not as disruptive of structure on kpc scales.

1225: Very large $F$, however, does not appear consistent with observations

1226: of the old stellar disk \citep{RixRieke93}.  One possibility is that

1227: realistic feedback is both gentler and less correlated than the simple

1228: prescription of our current models, and as a consequence the spiral

1229: arm coherence would not be destroyed by large-scale shells.  Indeed,

1230: semi-analytic models suggest that photo-ionization may evaporate much

1231: of the mass in a typical GMC before the pressure-driven expansion of

1232: HII regions unbinds the whole cloud

1233: \citep[e.g.][]{KrumholzMatznerMcKee06}.  Those models do not include

1234: supernovae, however, which are unavoidable if a GMC survives for more

1235: than one generation of OB stars.  Still, supernovae that are less

1236: correlated in space and time than the extreme case we have considered

1237: would disperse cloud gas in smaller parcels.  Less correlated energy

1238: inputs would produce shells with diameters less than the spiral arm

1239: thickness, and could more easily leave global spiral structure intact.

1240: By studying how the resulting spiral morphology varies with the

1241: correlation of feedback energy, it will be possible to place limits on

1242: how correlated star formation is in real galaxies.

1243: 

1244: \subsubsection{Multiphase ISM}

1245: 

1246: The models discussed in this paper use the simplest possible

1247: prescription for gas thermodynamics, which is an isothermal equation

1248: of state.  Our adopted sound speed of $c_s = 7$ \kms\ corresponds to a

1249: temperature of $T \sim 10^4$ K, characteristic of the warm phase of

1250: the ISM.  We adopted this approach in order to investigate, in a

1251: controlled fashion, various separate effects that can contribute to

1252: the regulation of star formation.

1253: 

1254: In parallel with our simplified ISM thermodynamics, our approach to

1255: modeling feedback from star formation is also reduced to the most

1256: basic elements. In our models, we follow the expansion of clouds

1257: subsequent to correlated SN events.  Of course, in a real SN event,

1258: thermal energy is injected into the ISM, and it is the expansion of a

1259: very hot and very diffuse bubble of gas that drives the formation of a

1260: dense shell around it.  The cooling time in the high density shell is

1261: short, so at late stages the isothermal approximation is adequate.

1262: The cooling time of the hot interior of each individual bubble, and of

1263: the hot phase of the ISM that results from merging SN remnants, is

1264: much longer.  However, the hot phase contains only a very small

1265: fraction of the total ISM mass.  From the point of view of most of the

1266: mass in the ISM, the primary effect of SN is to inject momentum.  By

1267: adopting an isothermal equation of state, and treating feedback as

1268: providing momentum inputs, this effect is captured in an approximate

1269: way.

1270: 

1271: A significant limitation of our models is that we do not treat the

1272: cold (T$\sim$100 K) atomic component of the ISM explicitly.  Because

1273: the level of turbulence in the atomic component is comparable to the

1274: thermal velocity dispersion of warm gas \citep{HeilesTroland03}, the

1275: effective pressure in the cold medium may be comparable to the thermal

1276: pressure in the warm medium.  The dynamics associated with ``turbulent

1277: pressure'' may, however, be quite different from those resulting from

1278: micro-physical thermal pressure.  A very important direction for future

1279: work is to study directly how large-scale gravitational instabilities

1280: and spiral structure develop in multiphase, turbulent, cloudy gas.

1281: 

1282: Another limitation of our models is that they are two-dimensional

1283: (although the disk flares with radius).  This constrains feedback

1284: energy to be confined within the galaxy's midplane, and does not allow

1285: for dynamically evolving disk thickness.  In the real ISM, correlated

1286: SN may be important in driving the SN heated gas away from the

1287: midplane of the galaxy into the halo, through so-called chimneys and

1288: superbubbles \citep{NormanIkeuchi89}.  To explore the effect of this

1289: energy loss in an approximate way, in our models we consider both a

1290: ``standard'' momentum input per SN, and a momentum input reduced by a

1291: factor of ten.  However, the cycling of gas through the galactic halo

1292: has other consequences as well.  After this gas is cooled in the halo,

1293: it falls back onto the disk in the form of cloudlets

1294: \citep[e.g.][]{JoungMacLow06}.  Even though recent simulations have

1295: shown that the fraction of mass that is vertically driven is small

1296: \citep[e.g.][]{deAvillezBreitschwerdt04}, the in-falling clouds may

1297: still affect the dynamics of the disk and may also act as another

1298: source of turbulence.

1299: 

1300: In order to accurately model disks that account for the effects of SN

1301: heating, chimneys, superbubbles, and the return of halo gas onto the

1302: disk, a three dimensional grid, as well as explicit treatment of

1303: heating and cooling, are necessary.  Three dimensional simulations

1304: will also allow us to test the sensitivity of the Kennicutt-Schmidt

1305: slope to the disk thickness (which evolves in response to star

1306: formation), as discussed in $\S$ \ref{ksdisc}.  These directions are

1307: important avenues for future research.

1308: 

1309: \subsection{Summary \label{summary}}

1310: 

1311: In this paper, we consider the formation of self-gravitating

1312: structures in global models of spiral galaxies, focusing on the

1313: effects of star formation feedback.  Our numerical simulations adopt a

1314: simple, isothermal treatment of the gas, and follow the flow in the

1315: disk by integrating the hydrodynamical equations on a polar grid.  We

1316: incorporate vertical disk thickness effects within the solution of the

1317: Poisson equation, which assumes that the disk flares as $H\propto R$.

1318: The feedback model treats the specific star formation rate in gas

1319: above a given surface density threshold as a constant, $R_{SN}

1320: M_{SN}$.  Feedback is implemented by spatially-resolved radial

1321: momentum injection subsequent to star formation events; the momentum

1322: injection is proportional to the number of stars formed.  In order to

1323: explore the sensitivity of the resulting model properties to the

1324: feedback parameters, we consider a range of specific star formation

1325: rates, star formation efficiencies $\epsilon_{SF}$, and momentum

1326: injection per massive star $P_{rad}$.  We analyze the ISM spatial

1327: distribution, star formation rates, and turbulent properties of our

1328: model disks in cases with and without an externally-imposed spiral

1329: gravitational perturbation.

1330: 

1331: Our main findings are as follows:

1332: 

1333: (1) In models where $P_{rad}$ is comparable to the level expected from

1334: a supernova, clouds are destroyed by star formation events and the

1335: mean turbulence level is high.  In models where $P_{rad}$ is a factor

1336: of ten lower, to represent inefficient feedback (e.g. if SN energy is

1337: vented vertically rather than kept in the disk), the self-gravitating

1338: structures that form are not destroyed by feedback, and the turbulence

1339: levels are substantially lower.  Turbulence levels are insensitive to

1340: the star formation rate parameter $R_{SN}$ and the overall star

1341: formation rate, however.

1342: 

1343: (2) In models with strong feedback, expanding flows lead to collisions

1344: of shells, which then lead to gravitational collapse of overdense

1345: regions and further star formation events. In this sense, our models

1346: are a concrete realization of the concept of self-propagating star

1347: formation.  We find, however, that the {\it net} effect of feedback is

1348: to {\it lower} the rate of star formation.  That is, when we compare

1349: models with strong feedback (large $P_{rad}$) and weak feedback (small

1350: $P_{rad}$), the former have lower resultant star formation rates.

1351: Similarly, when we compare models (at large and fixed $P_{rad}$) that

1352: have high or low feedback event rates $R_{SN}$, the fraction of dense

1353: gas is lower when the event rate is higher.  In principle, turbulence

1354: can either enhance collapse and star formation (by inducing shell

1355: collisions) or suppress collapse and star formation (by breaking up

1356: overdense regions).  Our results show that although both effects

1357: occur, the latter dominates: star formation is in net suppressed by 

1358: feedback.

1359: 

1360: (3) For $\Sigma\sim 10-100$ \msun\ $\pc^{-2}$, the range in $\Sigma_{SFR}$

1361: for our simulations is similar to the range observed in normal disks.

1362: The slope of the Kennicutt-Schmidt scaling relation $\Sigma_{SFR}

1363: \propto \Sigma^{1+p}$ is steeper ($1+p\sim 2$) in our simulations than

1364: the slopes found from current observations at high (average) surface density.

1365: The discrepancy may be due to our assumption that the disk thickness

1366: varies with radius as $H\propto R$.  Indeed, our numerical results are

1367: consistent with the theoretical prediction that $t_{SF}\propto t_J$ or

1368: $t_{ff}$ when the gravitational times $t_J$ and $t_{ff}$ are

1369: calculated based on our model prescription.  We point out that

1370: shallower scalings of $\Sigma_{SFR}$ with $\Sigma$ would be expected

1371: if the vertical velocity dispersion increases with $\Sigma$.  This

1372: would increase the disk thickness at small radii (where $\Sigma$ is

1373: large) relative to what we have assumed, and consequently increase the

1374: gravitational times and reduce $\Sigma_{SFR}$.  

1375: 

1376: (4) Motivated by our own results, we remark that in general, the

1377: thickness of the gaseous disk in a galaxy (either observed or

1378: simulated) is important for setting the index in the Kennicutt-Schmidt

1379: relationship.  Numerical simulations must resolve the natural disk

1380: scale height (set by pressure and turbulence) if the adopted

1381: prescription for star formation depends on the volume density $\rho$

1382: of gas.  A simulation that is vertically unresolved ($H\rightarrow

1383: const.$) while adopting

1384: $\rho_{SF} \propto \rho/t_{ff}(\rho)$, and hence $\Sigma_{SF}

1385: \propto \Sigma/t_{ff}(\rho)$, will automatically yield an index

1386: $1+p=1.5$ in the K-S law since $t_{ff}^{-1}\propto

1387: (\Sigma/H)^{0.5}$.  Fundamental understanding of K-S laws requires a

1388: self-consistent determination of the dependence of $H$ on $\Sigma$.

1389: 

1390: (5) For turbulence driven by expanding shells in overdense regions, we

1391: find that the power spectra decrease with decreasing size consistent

1392: with the scalings for shock-dominated flows (``Burgers turbulence'').

1393: While typical mass-weighted velocity dispersions on kpc scales in our

1394: high-$P_{rad}$ models are 31 \kms, these decrease to 10 \kms\ on 100

1395: pc scales.  Radial and azimuthal components of the velocity dispersion

1396: in a given scale are comparable.

1397: 

1398: (6) For all of our models, the maximum masses of dense clouds that form

1399: are several million \msun, consistent with observations of the upper

1400: cutoff in GMC/GMA mass distributions in local group galaxies.  In

1401: models with strong turbulence, such that self-gravitating

1402: condensations can form in colliding flows, a wider range of cloud

1403: masses results, with a lower peak in the distribution (but similar

1404: upper cutoff).  Higher resolution simulations will allow for a more

1405: detailed analysis of the mass distributions.

1406: 

1407: (7) Within the context of the feedback prescription and parameters for

1408: our current set of models, we find that global spiral patterns are not

1409: maintained.  For low $P_{rad}$, insufficient momentum is injected to

1410: overdense structures so that arm clouds continue to collapse,

1411: eventually depleting the surrounding spiral arm gas.  For high

1412: $P_{rad}$, large-scale expanding shells form and the global spiral

1413: structure is destroyed as cloud gas is dispersed.  We conclude that

1414: highly-correlated star formation, which is the limit that we adopt in

1415: the present models, is incompatible with long-lived spiral structure.

1416: It will be interesting to determine, by comparing spiral morphology

1417: with results from models adopting differing feedback prescriptions,

1418: what constraints are placed on the spatial and temporal correlation of

1419: star formation feedback in real galaxies.

1420: 

1421: 

1422: \acknowledgements 

1423: 

1424: This work was supported by grants 1278889 (NASA/{\it Spitzer}), and

1425: AST 0507315 (NSF).  Computations were performed on clusters supported

1426: by the Center for Theory and Computation in the Astronomy Department

1427: at the University of Maryland.  For much of the data analysis and

1428: visualization, we have made use of {\tt NEMO} software

1429: \citep{Teuben95}.  We are also grateful to the anonymous referee for

1430: helpful comments.

1431: 

1432: \appendix

1433: \section{Appendix}

1434: In the Appendix of Paper I, we described two methods to solve

1435: Poisson's equation numerically on a polar grid; both methods employ

1436: Fast Fourier Transforms (FFTs).  One method sums the potential from

1437: concentric rings, as described by \citet{Miller76}.  The other method

1438: employs a coordinate transformation from polar coordinates to a

1439: Cartesian-like coordinate system.  The former method is exact, but

1440: computationally expensive, and the latter is an approximation, but

1441: computationally efficient.

1442: 

1443: Here, we describe another FFT based method that is exact, and more

1444: efficient than the \citet{Miller76} method.\footnote{As in Paper I, we

1445:   again make use of the freely-available FFTW software

1446:   \citep{FFTW05}.}  

1447: The basic scheme is described in \citet{Kalnajs71}

1448: and \citet{BT87}; we describe a modification of Kalnajs's method that

1449: includes the effect of nonzero disk thickness $H$, which also acts as

1450: softening.

1451: 

1452: The potential $\Phi$ at each position

1453: $(R,\phi)$ on the disk, at $z$=0, is

1454: 

1455: \begin{equation}

1456: \Phi(R,\phi,z=0) = -G \int dR' \int d\phi' \int dz' \frac{R' f(z',R',\phi')\Sigma(R',\phi')}{[R^{'2}+R^2-2RR'\cos(\phi-\phi')+z^{'2}]^\frac{1}{2}}.

1457: \label{phi}

1458: \end{equation}

1459: Here, $G$ is the usual gravitational constant, $\Sigma$ is the total

1460: surface density,  and the 

1461: function $f=\rho(z',R',\phi')/\Sigma(R',\phi')$

1462: describing the vertical profile of the volume 

1463: density must be normalized,

1464: $\int_{-\infty}^{\infty}dz' f(z',R',\phi') = 1$.  

1465: Substituting $u'\equiv \ln R'$, and $\zeta' =

1466: z'/\sqrt{2}R'$ in equation (\ref{phi}), the potential reduces to

1467: \begin{equation}

1468: \Phi(R,\phi,z=0) = -Ge^u \int du' \int d\phi' \int d\zeta'

1469: \frac{ e^{u'-u} e^{u'} f(\zeta',u',\phi')\Sigma(u',\phi')}

1470: {[e^{u-u'}(\cosh(u-u')-\cos(\phi-\phi'))+\zeta^{'2}]^\frac{1}{2}}.

1471: \label{phi_sub}

1472: \end{equation}

1473: If $R' \rho(z',R',\phi')/\Sigma(R',\phi')= 

1474: e^{u'}f(\zeta',u',\phi')\equiv g(\zeta')$ 

1475: is a function of

1476: $\zeta'$ only (see below), we

1477: can define 

1478: \begin{equation}

1479: I(u'-u,\phi'-\phi) \equiv e^{u'-u}\int d\zeta' 

1480: \frac{ g(\zeta')}{[e^{u-u'}(\cosh(u-u')-\cos(\phi-\phi'))+\zeta^{'2}]^\frac{1}{2}}.

1481: \label{Idef}

1482: \end{equation}

1483: Using the definition of $I$ in equation (\ref{phi_sub}),

1484: we obtain $\Phi$ as a two dimensional convolution:

1485: \begin{equation}

1486: \Phi(R,\phi,z=0) = -Ge^u \int du' \int

1487: d\phi'\Sigma(u',\phi')I(u'-u,\phi'-\phi).

1488: \label{phi_noz}

1489: \end{equation}

1490: Applying the Fourier convolution theorem to equation 

1491: (\ref{phi_noz}), the gravitational potential can

1492: be computed by taking the Fourier transform of $\Sigma$ to obtain

1493: $\hat\Sigma$, and then taking the inverse Fourier transform of the

1494: product of $\hat\Sigma$ and $\hat I$, where $\hat I$ is the Fourier

1495: transform of $I$.  In hydrodynamic simulations, $\hat I$ can be

1496: computed once at the beginning of the simulation run, so that only two

1497: FFTs need to be performed at each timestep, FFT($\Sigma$) and

1498: FFT$^{-1}$($\hat\Sigma \hat I$).

1499: 

1500: 

1501: The function $I$, and therefore its convolution $\hat I$, depends on

1502: the normalized vertical distribution function $g(\zeta)$.

1503: For the specific case of a Gaussian vertical density distribution

1504: (which holds if the vertical gravity is dominated by that of the

1505: stellar disk), 

1506: \begin{equation}

1507: f(z^{},R^{}) = \frac{e^{-z^{2}/2H^2(R^{})} }{\sqrt{2\pi H^2(R^{})}}.

1508: \label{fgaus}

1509: \end{equation}

1510: For a disk that flares as $H(R^{}) \propto R^{}$, we define

1511: ${\cal H} = H(R^{})/R^{}$, so that

1512: \begin{equation}

1513: e^{u^{}}f(\zeta^{},R^{}) = \frac{e^{-(\zeta^{}/{\cal

1514:       H})^2}}{\sqrt{2\pi}{\cal H}}\equiv g(\zeta^{}).

1515: \label{fthick}

1516: \end{equation}

1517: Similarly, if the vertical density follows a ${\rm sech}^2$ distribution

1518: (true if the gaseous self-gravity dominates), then

1519: $g(\zeta^{})=(2{\cal H})^{-1}{\rm sech}^2(\zeta^{}\sqrt{2}/{\cal

1520:   H})$.

1521: 

1522: 

1523: 

1524: For our simulations, we adopt the Gaussian profile; this yields the following

1525: explicit expression for $I$:

1526: \begin{equation}

1527: I(u'-u,\phi'-\phi) \equiv \frac{e^{u'-u}}{\sqrt{2\pi}{\cal

1528:     H}} \int_{-\infty}^{\infty} d\zeta' 

1529: \frac{ e^{-(\zeta'/{\cal H})^2}}{[e^{u-u'}(\cosh(u-u')-\cos(\phi-\phi'))+\zeta^{'2}]^\frac{1}{2}}.

1530: \label{Iexp}

1531: \end{equation}

1532: 

1533: Finally, we comment on the assumption $H(R)\propto R$ which enables

1534: the three-dimensional gravitational integral to be written as a

1535: two-dimensional convolution.  If the stellar disk dominates gravity,

1536: then for an isothermal disk the vertical density distribution is

1537: Gaussian with $H/R = c_s Q_* (c_{*,z}/c_{*,R})/2v_c$, so values of

1538: $c_s/v_c$, $Q_*$, and $c_{*,z}/c_{*,R}$ that are independent of radius

1539: imply constant $H/R$.  Similarly, if gas is the dominant component for

1540: vertical gravity, $H = c_s^2/\pi G \Sigma$, so that $H/R = c_s

1541: Q/\sqrt{2}v_c$.  If both the Toomre $Q$ parameter and $c_s/v_c$ are

1542: independent of $R$, then $H/R$ = constant.  For self-gravitating

1543: gaseous disks, if $Q=1$ and $v_c/c_s$=30, then $H/R = 0.02$.

1544: Including stellar gravity typically reduces $H$ by a factor of $\sim2$

1545: \citep[e.g][]{KOStone02}.

1546: 

1547: For the simulations described in this paper, we use $H/R\equiv {\cal

1548:   H}=0.01$ in equation (\ref{Iexp}).  We have tested other values of

1549: $\cal H$, and find that our results are not sensitive to the exact

1550: value.  However, large changes significantly affect the rate of

1551: growth of self-gravitating perturbations.

1552: 

1553: 

1554: \bibliography{ref}

1555: 

1556: \begin{deluxetable}{cccccc} 

1557:   \tablewidth{0pt} \tablecaption{Parameters for

1558:     Standard\tablenotemark{a} Models} \tablehead{ \colhead{Model} &

1559:     \colhead{$Q_0$} & \colhead{$m$} &

1560:     \colhead{$R_{SN}$\tablenotemark{b} } & \colhead{$\epsilon_{SF}$}

1561:     & \colhead{$P_{rad}$ ($10^5$ \Mkms)} \\

1562:     \colhead{(1)} & \colhead{(2)} & \colhead{(3)} & \colhead{(4)} &

1563:     \colhead{(5)} & \colhead{(6)} } \startdata

1564:   Q1A & 1 & 0 & 1 & 0.025 & 0.34  \\

1565:   Q1B & 1 & 0 & 1 & 0.05 & 0.34  \\

1566:   Q1D & 1 & 0 & 10 & 0.05 & 3.4  \\

1567:   Q1SA & 1 & 4 & 1 & 0.025 & 0.34  \\

1568:   Q1SB & 1 & 4 & 1 & 0.05 & 0.34  \\

1569:   Q1SC & 1 & 4 & 1 & 0.05 & 3.4  \\

1570:   Q1SD & 1 & 4 & 10 & 0.05 & 3.4 \\

1571:   Q1SE & 1 & 4 & 10 & 0.05 & 0.34  \\

1572:   Q2SA & 2 & 4 & 1 & 0.025 & 0.34  \\

1573:   Q2SB & 2 & 4 & 1 & 0.05 & 0.34  \\

1574:   Q2SC & 2 & 4 & 1 & 0.05 & 3.4  \\

1575:   Q2SD & 2 & 4 & 10 & 0.05 & 3.4  \\

1576:   \enddata {\singlespace \tablenotetext{a}{\footnotesize

1577:       1024$\times$1024 zones; R $\in$ 4-11 kpc; $\phi \in$

1578:       0-$\frac{\pi}{2}$ radians} \tablenotetext{b}{\footnotesize

1579:       Units of $2\times 10^{-11} {\rm M}_\odot^{-1}\, {\rm yr}^{-1}$, i.e.

1580:       number of SN per 50 years per $10^9$M$_\odot$} }

1581: \label{standardmods}

1582: \end{deluxetable}

1583: 

1584: \begin{figure}

1585: \includegraphics[angle=-90,scale=0.75]{f1.ps}

1586: \caption{Density snapshots of $Q_0=1$ non-spiral model before any

1587:   feedback, at time $t/t_{orb}$ = 0.84.  Gray scale is in units of

1588:   $\log(\Sigma/\Sigma_0)$.}

1589: \label{befSN}

1590: \end{figure}

1591: 

1592: \begin{figure}[c]

1593: \includegraphics[angle=-90,scale=0.75]{f2.ps}

1594: \caption{Density snapshots of model Q1A, at time $t/t_{orb}$ = 1.125.

1595:   Gray scale is in units of $\log(\Sigma/\Sigma_0)$.}

1596: \label{Q1ASN}

1597: \end{figure}

1598: 

1599: \begin{figure}

1600: \includegraphics[angle=-90,scale=0.75]{f3.ps}

1601: \caption{Density snapshots of model Q1D, at time $t/t_{orb}$ = 1.125

1602:   (a) and at time $t/t_{orb}$ = 1.375 (b).  Gray scale is in units of

1603:   $\log(\Sigma/\Sigma_0)$.}

1604: \label{Q1BSN}

1605: \end{figure}

1606: 

1607: \begin{figure}

1608: \includegraphics[angle=-90,scale=0.75]{f4.ps}

1609: \caption{$Q_0=1$ spiral model, without feedback, at $t/t_{orb}=0.675$.

1610:   Gray scale is in units of $\log(\Sigma/\Sigma_0)$. }

1611: \label{spmodsQ1}

1612: \end{figure}

1613: 

1614: \begin{figure}

1615: \includegraphics[angle=-90,scale=0.75]{f5.ps}

1616: \caption{$Q_0=2$ spiral model, without feedback, at $t/t_{orb}=1.04$.

1617:   Gray scale is in units of $\log(\Sigma/\Sigma_0)$. }

1618: \label{spmodsQ2}

1619: \end{figure}

1620: 

1621: \begin{figure}

1622: \includegraphics[angle=-90,scale=0.75]{f6.ps}

1623: \caption{Models Q1SC (left) and Q1SD (right) at $t/t_{orb}=0.73$.  

1624: Gray scales are in units of $\log(\Sigma/\Sigma_0)$. }

1625: \label{Q1HV}

1626: \end{figure}

1627: 

1628: 

1629: \begin{figure}

1630: \includegraphics[angle=-90,scale=0.75]{f7.ps}

1631: \caption{Models Q1SC (left) and Q1SD (right) at $t/t_{orb}=1.15$.  

1632: Gray scales are in units of $\log(\Sigma/\Sigma_0)$. }

1633: \label{Q1HVlater}

1634: \end{figure}

1635: 

1636: \begin{figure}

1637: \plottwo{f8a.ps}{f8b.ps}

1638: \caption{Cloud masses in models with strong (thick lines) and weak

1639:   (thin lines) feedback.

1640: Left: Histogram of $M_{cl}$ in models Q1D (thick) and Q1A

1641:   (thin), up until time $t/t_{orb}$ = 1.125 (see Figs. \ref{Q1ASN} -

1642:   \ref{Q1BSN}).  The mean (median) $M_{cl}$ for models Q1A and Q1D are

1643:   1.2$\times10^6$ (0.8$\times10^6$) and 0.7$\times10^6$

1644:   (0.5$\times10^6$) M$_\odot$, respectively. Right: Histogram of

1645:   $M_{cl}$ in models Q1SD (thick) and Q1SA (thin), up until time

1646:   $t/t_{orb}$ = 0.73 (model Q1SD is shown in Fig. \ref{Q1HV}).  The

1647:   mean (median) $M_{cl}$ for models Q1SA and Q1SD are 2.2$\times10^6$

1648:   (1.9$\times10^6$) and 0.8$\times10^6$ (0.6$\times10^6$) M$_\odot$,

1649:   respectively.}

1650: \label{clmasses}

1651: \end{figure}

1652: 

1653: 

1654: \begin{figure}

1655: \includegraphics[angle=-90,scale=0.5]{f9.ps}

1656: \caption{$SFR$ (top) and $v_{turb}$ (bottom) for models with $Q_0$=1,

1657:   without spiral structure.  The values in parentheses in the legend

1658:   are the SN rate parameter $R_{SN}$, the star formation efficiency

1659:   $\epsilon_{SF}$, and SN momentum $P_{rad}$ (in \Mkms) of each model.

1660:   The large open squares in the bottom panel are the turbulent velocities

1661:   for a simulation without any feedback.}

1662: \label{Q1SFR}

1663: \end{figure}

1664: 

1665: \begin{figure}

1666: \includegraphics[angle=-90,scale=0.5]{f10.ps}

1667: \caption{$SFR$ (top) and $v_{turb}$ (bottom) for models with $Q_0$=1,

1668:   as in Figure \ref{Q1SFR}, but with spiral structure.}

1669: \label{Q1SSFR}

1670: \end{figure}

1671: 

1672: \begin{figure}

1673: \includegraphics[angle=-90,scale=0.5]{f11.ps}

1674: \caption{$SFR$ (top) and $v_{turb}$ (bottom), as in Figure

1675:   \ref{Q1SSFR}, but for models with $Q_0$=2, with spiral structure.}

1676: \label{Q2SSFR}

1677: \end{figure}

1678: 

1679: \begin{figure}

1680: \includegraphics[angle=-90,scale=0.5]{f12.ps}

1681: \caption{Mass weighted turbulent velocities vs. mean surface density

1682:   of Q1S models, averaged in annuli of widths 1 kpc, and in the time

1683:   interval $t/t_{orb} \in 100 - 116$ Myr.  Turbulent velocities are

1684:   only shown from annuli and time intervals within which feedback

1685:   events have occurred.}

1686: \label{dispsig}

1687: \end{figure}

1688: 

1689: \begin{figure}

1690: \includegraphics[angle=-90,scale=0.5]{f13.ps}

1691: \caption{Turbulent power spectra of Model Q1D. Power is shown at

1692:   constant $k_R$ (left) and constant $k_\phi$ (right) (each slice is 

1693:   along the minimum nonzero value of the respective $k$).  To obtain the

1694:   dimensions of $k_\phi$ ($k_\phi = m n_\phi/R$), we use the mean

1695:   radius of the grid.  Best fits for values between $\log(k) \in 0 -

1696:   1$ gives slopes of -2.9 (left) and -2.4 (right).}

1697: \label{turbspec}

1698: \end{figure}

1699: 

1700: \begin{figure}

1701: \includegraphics[angle=0,scale=0.75]{f14.ps}

1702: \caption{Schmidt law for models in Table \ref{standardmods}.  Each

1703:   point is obtained by binning the simulation data in radii, with 1

1704:   kpc widths, and in time, with 18 Myr widths.  Only annuli and times

1705:   with at least 1 feedback event are included.  The lines are the best

1706:   fit to the points; with their slopes ranging from 1.8 - 3.8 (top),

1707:   1.1 - 3.0 (middle), and -2.0 - 1.9 (bottom).}

1708: \label{QKS}

1709: \end{figure}

1710: 

1711: \begin{figure}

1712: \plottwo{f15a.ps}{f15b.ps}

1713: \caption{Left: Star formation rate vs. surface density for model Q1D

1714:   (triangles; $R\in$ 4 - 11 kpc), as well as the corresponding model of

1715:   the inner region (crosses; $R\in$ 0.8 - 2.2 kpc).  Best fit lines

1716:   for each model are also shown, with slopes of 2.4 for the 4 - 11 kpc

1717:   model and 2.2 for the 0.8 - 2.2 kpc model.  Right: Triangles and

1718:   crosses from figure on the left are shown, along with globally

1719:   averaged observational data from \citet{Kennicutt98}: circles show

1720:   normal spirals, with best fit slope of 1.3, and diamonds show IR starburst

1721:   sources, with best fit slope of 1.4.}

1722: \label{innercomp}

1723: \end{figure}

1724: 

1725: \begin{figure}

1726: \includegraphics[angle=-90,scale=0.5]{f16.ps}

1727: \caption{Star formation times from model Q1D, as a function of mean

1728:   surface density.  Different symbols correspond to different time

1729:   bins, of width 18 Myr.  Simulation data are also binned in radii

1730:   with widths 1 kpc.  Filled symbols show actual depletion times (eq.

1731:   [\ref{deptime}]), and open symbols show predicted depletion times

1732:   (eq. [\ref{deptimeapp}]), for each annular and temporal bin.  }

1733: \label{pltdeptime}

1734: \end{figure}

1735: 

1736: \begin{figure}

1737: \includegraphics[angle=-90,scale=0.5]{f17.ps}

1738: \caption{Star formation times from model Q1D, as a function of Jeans

1739:   time (top) and free-fall time (bottom).  Simulation data are binned

1740:   as described in the caption to Figure \ref{pltdeptime}, with the

1741:   innermost and outermost annuli excluded.  The dashed lines, shown

1742:   for comparison, have slopes of 1.}

1743: \label{t_J-t_g}

1744: \end{figure}

1745: 

1746: \end{document}