The warm neutral medium, warm ionized medium, and cool neutral medium all show strong evidence for turbulence as a process dominating their structure and motions on a wide range of scales. The spatial power spectra of density fluctuations in all three phases are consistent with a Kolmogorov slope. Turbulence in the magnetic field in the diffuse medium can also be measured through the structure function of the Faraday rotation measure. With new surveys, new analysis techniques, and new telescopes, in the next few years it will be possible to measure the structure function of the magnetic field over a similarly wide range of scales. This will give a complete picture of the turbulence as a magneto-acoustic process.

The observational record of turbulence within the molecular gas phase of the interstellar medium is summarized. We briefly review the analysis methods used to recover the velocity structure function from spectroscopic imaging and the application of these tools on sets of cloud data. These studies identify a near-invariant velocity structure function that is independent of the local environment and star formation activity. Such universality accounts for the cloud-to-cloud scaling law between the global line-width and size of molecular clouds found by Larson (1981) and constrains the degree to which supersonic turbulence can regulate star formation. In addition, the evidence for large scale driving sources necessary to sustain supersonic flows is summarized.

We summarize a continuing investigation of turbulence in high-latitude translucent molecular clouds. These low mass (~50–100 M), nearby (~100 pc), non-star forming clouds appear to be condensing out of the atomic cirrus. Unlike star-forming clouds the velocity fields in the clouds must be driven by external processes. Our detailed mapping of the clouds MBM 3,16 and 40 indicates that the dynamics in these clouds result from the combination of shear-flow and thermal instabilities, not shocks. These clouds also show coherent structures, non-Gaussian PDFs but no clear velocity-size relation. Lastly, the energetics of these clouds indicate that radiative loss may terminate the cascade before local heating takes place.

The signatures of intermittent dissipation of turbulent energy have been sought in the translucent environment of a low-mass dense core. Molecular line observations reveal a network of narrow filamentary structures, found on statistical grounds to be the locus of the largest velocity shears. Three independent properties of these structures make them the plausible sites of intermittent dissipation of turbulence: (1) gas there is warmer and more diluted than average, (2) it bears the signatures of a non-equilibrium chemistry triggered by impulsive heating due to turbulence dissipation, and (3) the power that these structures radiate in the gas cooling lines (mostly H2) is so large that it balances the total energy injection rate of the turbulent cascade, for a volume filling factor of only a few percents, consistent with other observations in the Solar Neighborhood. These filamentary structures may act as tiny seeds of gas condensation in diffuse molecular gas. They do not exhibit the properties of steady-state low-velocity magneto-hydrodynamic (MHD) shocks, as presently modelled.

Formation of GMCs is one of the most crucial issues in galaxy evolution. I will compare CO and HI in the LMC in 3 dimensional space for the first time aiming at revealing the physical connection between GMCs and associated HI gas at a ~40 pc scale. The present major findings are 1) [total CO intensity] ∝ [total HI intensity]0.8 for the 110 GMCs, and 2) the HI intensity tends to increase with the evolution of GMCs. I argue that these findings are consistent with the growth of GMCs via HI accretion over a time scale of a few × 10 Myrs. I will also discuss the role of the background stellar gravity and the dynamical compression by supershells in formation of GMCs.

Multiwavelengths studies of massive star formation regions in the LMC and SMC reveal that a second generation of stars is being formed in dense molecular clouds located in the surroundings of the massive clusters. These dense molecular clouds have survived the action of massive star UV radiation fields and winds and they appear as compact dense H2 knots in regions of weak CO emission. Alternatively, we have found that large molecular clouds, probably remnants of the parental giant molecular clouds where the first generation of stars were formed, are suffering the interaction of the winds and UV radiation field in their surfaces in the direction of the central massive cluster or massive stars. These molecular regions show 1.2 mm continuum emission form cold dust and they show embedded IR sources as determined from deep ground base JHKs imaging. The distribution of young IR sources as determined from their Mid IR colors obtained by SPITZER concentrate in the maxima of CO and dust emission. IR spectroscopy of the embedded sources with high IR excess confirm their nature as massive young stellar objects (MYSO's). Our results are suggestive of contagious star formation where triggering and induced star formation could be taking place.

Understanding interstellar dust evolution is a major challenge underlying the interpretation of Spitzer observations of interstellar clouds, star forming regions and galaxies. I illustrate on-going work along two directions. I outline the potential impact of interstellar turbulence on the abundance of small dust particles in the diffuse interstellar medium and translucent sections of molecular clouds. I present results from an analysis of ISO and Spitzer observations of the central part of 30 Doradus, looking for dust evolution related to the radiative and dynamical impact of the R136 super star cluster on its parent molecular cloud.

The ground electronic state of neutral atomic carbon has three fine-structure levels. In the interstellar medium, the relative populations of the upper two levels are established by collisional excitations (and de-excitations) balanced against spontaneous radiative decay. Consequently, the fractions of C I in the upper two levels indicate acceptable combinations of local temperature and density, which in turn indicate the approximate thermal pressures of the medium. We can measure the values of these fractions and how they vary from one location to the next by observing the multiplets of C I seen in absorption in the ultraviolet spectra of hot stars.

We have identified 102 stars for which the HST MAST archive has E140H STIS spectra that are suitable for measuring the absorption features of C I at velocity resolutions of 3 kms−1(or better). A special analysis method developed by Jenkins & Tripp (2001) permits determinations of the amounts of C I in each of the three levels as a function of radial velocity over a wide dynamic range in column density, since several multiplets of vastly different strengths can be considered simultaneously.

The C I data reveal that the much of the diffuse, cold, neutral medium has pressures that are distributed in an approximately log-normal fashion, spread over a range 1000 < p/k < 104 cm−3 K (FWHM), but with low level tails outside this range. The dispersion of pressures increases slightly for gases that have radial velocities outside the expected range for quiescent material along each line of sight. This link to the kinematics of the gas is consistent with the picture that pressure fluctuations are driven by the dynamics of a turbulent medium. If the gas is a single medium that is being driven by turbulent forces, its barytropic index (slope of log p vs. log n) is more than 0.9, which is inconsistent with the value 0.72 for material that is expected to be in thermal equilibrium. Slightly less than one part in a thousand of the gas is at pressures of order or greater than ~105cm−3 K and seems to nearly always accompany the gas at normal pressures.

It is shown that a number of key observations of the Galactic ISM can be understood, if it is treated as a highly compressible and turbulent medium, energized predominantly by supernova explosions (and stellar winds). We have performed extensive numerical high resolution 3D hydrodynamical and magnetohydrodynamical simulations with adaptive mesh refinement over sufficiently long time scales to erase memory effects of the initial setup. Our results show, in good agreement with observations, that (i) volume filling factors of the hot medium are modest (typically below 20%), (ii) global pressure is far from uniform due to supersonic (and to some extent super-Alfvénic) turbulence, (iii) a significant fraction of the mass (~60%) in the warm neutral medium is in the thermally unstable regime (500 < T < 5000 K), (iv) the average number density of Ovi in absorption is 1.81 × 10−8 cm−3, in excellent agreement with Copernicus and FUSE data, and its distribution is rather clumpy, consistent with its measured dispersion with distance.

We have performed numerical simulations which were designed to further our understanding of the turbulent interstellar medium (ISM). Our simulations include a multi-phase thermodynamic model of the ISM, magnetic fields, and sheared rotation, allowing us to study the effects of the magnetorotational instability (MRI) in an environment containing high density cold clouds embedded in a warm, low density, ambient medium. These models have shown that the MRI is indeed a significant source of turbulence, particularly at low mean densities typical of the outer regions of the Milky Way, where star formation rates are low, but high levels of turbulence persist. Here, we summarize past findings, as well as our most recent models which include vertical stratification, allowing us to self-consistently model the vertical distribution of material in the disk.

Turbulence is believed to be important to star formation both within GMCs (affecting the IMF and the SFE), and on larger scales in the ISM (affecting GMC formation rates and properties). The traditional view of the ISM attributes most of the turbulent driving to stellar sources – in particular, supernovae and HII regions. However, evidence suggests that sources other than star formation must contribute significantly to large-scale turbulent driving in the diffuse ISM, thus also affecting the turbulence that GMCs inherit. I review recent investigations of alternative sources proposed for driving ISM turbulence, including thermal instabilities, the magnetorotational instability, large-scale self-gravitating modes, and instabilities in spiral shocks. I summarize results based on numerical simulations regarding the levels of turbulence that can be driven, and how these amplitudes depend on galactic conditions. This recent work shows that, under certain circumstances, very large-amplitude (trans-sonic with respect to the warm gas) motions and magnetic fields can be driven even without stellar energy inputs. Since turbulence can either trigger or limit star formation, understanding these potentially large non-stellar driving sources is crucial for a developing a theory of star formation regulation in the Milky Way and other disk galaxies.

HI observations at sufficiently high spatial and velocity resolution have revealed a wealth of structures such as shells and bubbles in the ISM of late–type galaxies. These structures are filled with metal–enriched, coronal gas from SNe which, through overpressure, powers their expansion. Material swept up by these expanding shells can go “critical” and form subsequent (secondary or propagating) star formation. Shells that grow larger than the thickness of the gas layer will blow out of the disk, spilling enriched material into the halo (or in the case of violent starbursts, the Intergalactic Medium). We review what has been achieved to date and present some first results of a major project based on THINGS (The HI Nearby Galaxy Survey), which aims to extend studies of the ISM in galaxies to 34 nearby systems (<10 Mpc), all observed to the same exacting standards (resolution 6″ × 5 km s−1, or better; typical detection threshold of ~5 × 1019 atom cm−2).

The Magellanic System harbors >800 expanding shells of neutral hydrogen, providing a unique opportunity for statistical investigations. Most of these shells are surprisingly young, 2–10 Myr old, and correlate poorly with young stellar populations. I summarize what we have learned about shell properties and particularly focus on the puzzling correlation between the shell radius and expansion velocity. In the framework of the standard, adiabatic model for shell evolution this tight correlation suggests a coherent burst of star formation across the whole Magellanic System. However, more than one mechanism for shell formation may be taking place.

We briefly discuss different methods used to identify HI shells in TB datacubes. Then we give results for our automatic method applied to LDS and LAB HI surveys of the Milky Way (2nd and 3rd quadrants). We fit the radial distribution of HI shells (the exponential profile with the scale length of 3 kpc) and the size distribution (the power law with the index of 2.1). We compare the distribution of identified HI shells with HII regions and study the differences between identifications in the 2nd and 3rd quadrants.

The spatial power spectrum of the HI 21 cm intensity in the Small Magellanic Cloud (Stanimirovic et al. 1999) is a power law over scales as large as those of the SMC itself. It was interpreted as due to turbulence by Goldman (2000) and by Stanimirovic & Lazarian (2001). The question is whether the power spectrum is indeed the result of a dynamical turbulence or is merely the result of a structured static density. In the turbulence interpretation of Goldman (2000) the turbulence was generated by the tidal effects of the last close passage of the LMC about 0.2 Gyr ago. The turbulence time-scale was estimated by Goldman to be 0.4 Gyr, so the turbulence has not decayed yet. Staveley-Smith et al. (1997) observed in the SMC about five hundreds of HI super shells. Their age is more than an order of magnitude smaller than the turbulence age. Therefore, if the turbulence explanation holds, their observed radial velocities should reflect the turbulence in the gas in which they formed. In the present work we analyze the observed radial velocities of the super shells. We find that the velocities indeed manifest the statistical spatial correlations expected from turbulence. The turbulence spectrum is consistent with that obtained by Goldman (2000).

In order to elucidate star formation in the Large Magellanic Cloud, a complete survey of the molecular clouds was carried out by NANTEN. In this work, we compare 230 giant molecular clouds (GMCs), whose physical quantities are well determined, with young clusters and Hii regions. We find that about 76% of the GMCs are actively forming stars or clusters, while 24% show no signs of massive star or cluster formation. Effects of supergiant shells (SGSs) on the formation of GMCs and stars are also studied. The number and surface mass densities of the GMCs are higher by a factor of 1.5–2 at the edge of the SGSs than elsewhere. It is also found that young stellar clusters are more actively formed in the GMCs facing to the center of the SGSs. These results are consistent with the previous studies by Yamaguchi et al. and suggest the formation of GMCs and the cluster is triggered by dynamical effects of the SGSs.

Interstellar superbubbles generated by multiple supernova explosions are common in star-forming galaxies. They are the most obvious manifestation of mechanical feedback, and are largely responsible for transferring both thermal and kinetic energy to the interstellar medium from the massive star population. However, the details of this energy transfer remain surprisingly murky when individual objects are studied. I will summarize what we currently know about candidate dominant processes on these scales.

The gravitational instability of shells and supershells and its relevance to the triggering of the star formation in a turbulent ISM is discussed. The IMF of the self-gravitating clumps formed out of shells is computed and triggering time is compared with the shell dissolution time due to chaotic motions.

The aim of the present study, based on literature data, is to find signatures of the giant radio continuum structure Loop III on the nearby interstellar medium, and search for molecular cloud and star formation, possibly triggered by its expansion. The preliminary results are as follows: (1) The 3D map of the Local Bubble, published by Lallement et al. (2003) suggests that Loop III is probably more distant than the early models had indicated. (2) The molecular clouds at high galactic latitudes in the 2nd Galactic quadrant are probably associated with the neutral/molecular wall of Loop III. (3) Star formation in Lynds 1333 and Lynds 1082 (GF 9) might have been triggered by the expansion of Loop III. (4) The supernova(e), whose explosion produced Loop III, might have been located in the SU Cas association.

Our goal is to evaluate the role of triggering effects on the star formation and early stellar evolution by presenting a statistically large sample of cloud and low-mass YSO data. We conducted large area surveys (ranging from 400 square-degree to 10800 square-degree) in optical, NIR and FIR. The distribution of the ISM and low-mass YSOs were surveyed. A relative excess was found statistically in the number of dense and cold core bearing clouds and low mass YSOs in the direction of the FIR loop shells indicating a possible excess in their formation.