High-resolution two-dimensional hydrodynamic simulations for the interstellar matter (ISM) in a galactic disk are enhanced to include explicitly star formation and the feedback e †ects from supernovae and stellar winds. A globally stable multiphase ISM is formed, in which Ðlamentary and clumpy structure is a characteristic feature. We Ðnd a new component of 106È108 K gas that is a direct consequence of the energy input from the feedback. The total supernovae rate in the system varies by an order of magnitude over a timescale of 106 yr. The evolution of the supernovae rate exhibits chaotic behavior because the star formation is triggered by supernovae explosions in the inhomogeneous interstellar medium. We also Ðnd that, in spite of its very complicated spatial structure, the multiphase ISM exhibits a one-point probability density function (pdf) that is a perfect lognormal distribution over four decades in density, 102È106pc~2. The lognormal pdf is very robust even in regions with frequent bursts of M _ supernovae. Low-density regions or cavities (\10 pc~2), on the other hand, exhibit the normal M _ Gaussian distribution. These characteristic pdfÏs are achieved over a local dynamical scale. The energy spectra are E(k) P k~3 without feedback and E(k) P k~2 including stellar energy feedback.

In order to understand the physical mechanisms underlying non-steady stellar spiral arms in disk galaxies, we analyzed the growing and damping phases of their spiral arms using three-dimensional N -body simulations. We confirmed that the spiral arms are formed due to a swing amplification mechanism that reinforces density enhancement as a seeded wake. In the damping phase, the Coriolis force exerted on a portion of the arm surpasses the gravitational force that acts to shrink the portion. Consequently, the stars in the portion escape from the arm, and subsequently they form a new arm at a different location. The time-dependent nature of the spiral arms are originated in the continual repetition of this non-linear phenomenon. Since a spiral arm does not rigidly rotate, but follows the galactic differential rotation, the stars in the arm rotate at almost the same rate as the arm. In other words, every single position in the arm can be regarded as the co-rotation point. Due to interaction with their host arms, the energy and angular momentum of the stars change, thereby causing the radial migration of the stars. During this process, the kinetic energy of random motion (random energy) of the stars does not significantly increase, and the disk remains dynamically cold. Owing to this low degree of disk heating, the short-lived spiral arms can recurrently develop over many rotational periods. The resultant structure of the spiral arms in the N -body simulations is consistent with some observational nature of spiral galaxies. We conclude that the formation and structure of spiral arms in isolated disk galaxies can be reasonably understood by non-linear interactions between a spiral arm and its constituent stars.

The structure of obscuring matter in the environment of active galactic nuclei with associated nuclear starbursts is investigated using 3-D hydrodynamical simulations. Simple analytical estimates suggest that the obscuring matter with energy feedback from supernovae has a torus-like structure with a radius of several tens of parsecs and a scale height of about 10 pc. These estimates are confirmed by the fully non-linear numerical simulations, in which the multi-phase inhomogeneous interstellar matter and its interaction with the supernovae are consistently followed. The globally stable, torus-like structure is highly inhomogeneous and turbulent. To achieve the high column densities (> 10^{24} cm^{-2}) as suggested by observations of some Seyfert 2 galaxies with nuclear starbursts, the viewing angle should be larger than about 70 degree from the pole-on for a 10^8 solar mass massive black hole. Due to the inhomogeneous internal structure of the torus, the observed column density is sensitive to the line-of-sight, and it fluctuates by a factor of order 100. The covering fraction for N > 10^{23} cm^{-2} is about 0.4. The average accretion rate toward R < 1 pc is 0.4 solar mass/yr, which is boosted to twice that in the model without the energy feedback.Comment: ApJL in press (4 pages, 3 figures) A gziped ps file with high resolution figures is available at http://th.nao.ac.jp/~wada/AGN

It has been believed that spiral arms in pure stellar disks, especially the ones spontaneously formed, decay in several galactic rotations due to the increase of stellar velocity dispersions. Therefore, some cooling mechanism, for example dissipational effects of the interstellar medium, was assumed to be necessary to keep the spiral arms. Here we show that stellar disks can maintain spiral features for several tens of rotations without the help of cooling, using a series of high-resolution three-dimensional N -body simulations of pure stellar disks. We found that if the number of particles is sufficiently large, e.g., 3 × 10 6 , multi-arm spirals developed in an isolated disk can survive for more than 10 Gyrs. We confirmed that there is a self-regulating mechanism that maintains the amplitude of the spiral arms. Spiral arms increase Toomre's Q of the disk, and the heating rate correlates with the squared amplitude of the spirals. Since the amplitude itself is limited by Q, this makes the dynamical heating less effective in the later phase of evolution. A simple analytical argument suggests that the heating is caused by gravitational scattering of stars by spiral arms and that the self-regulating mechanism in pure-stellar disks can effectively maintain spiral arms on a cosmological timescale. In the case of a smaller number of particles, e.g., 3 × 10 5 , spiral arms grow faster in the beginning of the simulation (while Q is small) and they cause a rapid increase of Q. As a result, the spiral arms become faint in several Gyrs.

We present new high resolution numerical simulations of the interstellar medium (ISM) in a central R ≤ 32 parsecs region around a supermassive black hole (1.3 × 10 7 M ⊙ ) at a galactic center. Threedimensional hydrodynamic modeling of the ISM (Wada & Norman 2002) with the nuclear starburst now includes tracking of the formation of molecular hydrogen (H 2 ) out of the neutral hydrogen phase as a function of the evolving ambient ISM conditions with a finer spatial resolution (0.125 pc). In a quasi equilibrium state, mass fraction of H 2 is about 0.4 (total H 2 mass is ≃ 1.5 × 10 6 M ⊙ ) of the total gas mass for the uniform far UV (FUV) with G 0 = 10 in Habing unit. As shown in the previous model, the gas forms an inhomogeneous disk, whose scale-height becomes larger in the outer region. H 2 forms a thin nuclear disk in the inner ≃ 5 pc, which is surrounded by molecular clouds swelled up toward h 10 pc. The velocity field of the disk is highly turbulent in the torus region, whose velocity dispersion is ≃ 20 km s −1 on average. Average supernova rate (SNR) of ≃ 5 × 10 −5 yr −1 is large enough to energize these structures. Gas column densities toward the nucleus larger than 10 22 cm −2 are observed if the viewing angle is smaller than θ v ≃ 50 • from the edge-on. However, the column densities are distributed over almost two orders of magnitude around the average for any given viewing angle due to the clumpy nature of the torus. For a stronger FUV (G 0 = 100), the total H 2 mass in an equillibrium is only slightly smaller (≃ 0.35), a testimony to the strong self-shielding nature of H 2 , and the molecular gas is somewhat more concentrated in a mid-plane. Other properties of the ISM are not very sensitive either to the FUV intensity and the supernova rate. Finally the morphology and kinematics of the circumnuclear molecular gas disks emerging from our models is similar to that revealed by recent near infrared observations using VLTI/Keck.

The probability distribution functions (PDFs) of the density of the interstellar medium (ISM ) in galactic disks and the global star formation rate (SFR) are discussed. Three-dimensional hydrodynamic simulations show that the PDFs in a globally stable, inhomogeneous ISM in galactic disks are well fitted by a single lognormal function over a wide density range. The dispersion of the lognormal PDF (LN-PDF) is larger for more gas-rich systems, whereas the characteristic density of the LN-PDF, for which the volume fraction becomes the maximum, does not significantly depend on the initial conditions. Supposing the galactic ISM is characterized by the LN-PDF, we give a global SFR as a function of average gas density, a critical local density for star formation, and the star formation efficiency (SFE). Although the present model is more appropriate for massive and geometrically thin disks ($10 pc) in inner galactic regions ( 10 3 cm À3 ) is SFE ¼ 0:001Y0:01 for normal spiral galaxies and 0.01Y0.1 for starburst galaxies. The LN-PDF and SFR proposed here could be applicable for modeling star formation on a kiloparsec scale in galaxies or numerical simulations of galaxy formation, in which the numerical resolution is not fine enough to describe the local star formation.

High-resolution, 2-D hydrodynamical simulations with a large dynamic range are performed to study the turbulent nature of the interstellar medium (ISM) in galactic disks. The simulations are global, where the self-gravity of the ISM, realistic radiative cooling, and galactic rotation are taken into account. In the analysis undertaken here, feedback processes from stellar energy source are omitted. We find that the velocity field of the disk in a non-linear phase shows a steady power-law energy spectrum over three-orders of magnitude in wave number. This implies that the random velocity field can be modeled as fully-developed, stationary turbulence. Gravitational and thermal instabilities under the influence of galactic rotation contribute to form the turbulent velocity field. The Toomre effective Q value, in the non-linear phase, ranges over a wide range, and gravitationally stable and unstable regions are distributed patchily in the disk. These results suggest that large-scale galactic rotation coupled with the self-gravity of the gas can be the ultimate energy sources that maintain the turbulence in the local ISM. We find that our models of turbulent rotating disks are consistent with the velocity dispersion of an extended HI disk in the dwarf galaxy, NGC 2915, where there is no prominent active star formation. Numerical simulations show that the stellar bar in NGC 2915 enhances the velocity dispersion, and it also drives spiral arms as observed in the HI disk.

We propose a new evolutionary model of a supermassive black hole (SMBH) and a circumnuclear disk (CND), taking into account the mass-supply from a host galaxy and the physical states of CND. In the model, two distinct accretion modes depending on gravitational stability of the CND play a key role on accreting gas to a SMBH. (i) If the CMD is gravitationally unstable, energy feedback from supernovae (SNe) supports a geometrically thick, turbulent gas disk. The accretion in this mode is dominated by turbulent viscosity, and it is significantly larger than that in the mode (ii), i.e., the CMD is supported by gas pressure. Once the gas supply from the host is stopped, the high accretion phase ($\sim 0.01- 0.1 M_{\odot} {\rm yr}^{-1}$) changes to the low one (mode (ii), $\sim 10^{-4} M_{\odot} {\rm yr}^{-1}$), but there is a delay with $\sim 10^{8}$ yr. Through this evolution, the gas-rich CND turns into the gas poor stellar disk. We found that not all the gas supplied from the host galaxy accrete onto the SMBH even in the high accretion phase (mode (i)), because the part of gas is used to form stars. As a result, the final SMBH mass ($M_{\rm BH,final}$) is not proportional to the total gas mass supplied from the host galaxy ($M_{\rm sup}$); $M_{\rm BH,final}/M_{\rm sup}$ decreases with $M_{\rm sup}$.This would indicate that it is difficult to form a SMBH with $\sim 10^{9} M_{\odot}$ observed at high-$z$ QSOs. The evolution of the SMBH and CND would be related to the evolutionary tracks of different type of AGNs.Comment: 11 pages, 11 figures, accepted for publication in Ap