We present a statistical study of the distribution and physical properties of cold dense material in and around the inner Galactic Plane near infrared bubbles as catalogued by the Milky Way Project citizen scientists. Using data from the ATLASGAL 870 um survey, we show that 48 +/- 2% of all cold clumps in the studied survey region (|l| <= 65 degrees, |b| <= 1 degree) are found in close proximity to a bubble, and 25 +/- 2% appear directly projected towards a bubble rim. A two-point correlation analysis confirms the strong correlation of massive cold clumps with expanding bubbles. It shows an overdensity of clumps along bubble rims that grows with increasing bubble size, which shows how interstellar medium material is reordered on large scales by bubble expansion around regions of massive star formation. The highest column density clumps appear resistent to the expansion, remaining overdense towards the bubbles' interior rather than being swept up by the expanding edge. Spectroscopic observations in ammonia show that cold dust clumps near bubbles appear to be denser, hotter and more turbulent than those in the field, offering circumstantial evidence that bubble-associated clumps are more likely to be forming stars. These observed differences in physical conditions persist for beyond the region of the bubble rims.

The all-Galaxy CO survey of Dame, Hartmann, & Thaddeus (2001) is by far the most uniform, large-scale Galactic CO survey. Using a dendrogram-based decomposition of this survey, we present a catalog of 1064 massive molecular clouds throughout the Galactic plane. This catalog contains $2.5 \times 10^8$ solar masses, or $25^{+10.7}_{-5.8} \%$ of the Milky Way's estimated H$_2$ mass. We track clouds in some spiral arms through multiple quadrants. The power index of Larson's first law, the size-linewidth relation, is consistent with 0.5 in all regions - possibly due to an observational bias - but clouds in the inner Galaxy systematically have significantly (~ 30%) higher linewidths at a given size, indicating that their linewidths are set in part by Galactic environment. The mass functions of clouds in the inner Galaxy versus the outer Galaxy are both qualitatively and quantitatively distinct. The inner Galaxy mass spectrum is best described by a truncated power-law with a power index of $\gamma=-1.6\pm0.1$ and an upper truncation mass $M_0 = (1.0 \pm 0.2) \times 10^7 M_\odot$, while the outer Galaxy mass spectrum is better described by a non-truncating power law with $\gamma=-2.2\pm0.1$ and an upper mass $M_0 = (1.5 \pm 0.5) \times 10^6 M_\odot$, indicating that the inner Galaxy is able to form and host substantially more massive GMCs than the outer Galaxy. Additionally, we have simulated how the Milky Way would appear in CO from extragalactic perspectives, for comparison with CO maps of other galaxies.

The very long and thin infrared dark cloud "Nessie" is even longer than had been previously claimed, and an analysis of its Galactic location suggests that it lies directly in the Milky Way's mid-plane, tracing out a highly elongated bone-like feature within the prominent Scutum-Centaurus spiral arm. Re-analysis of mid-infrared imagery from the Spitzer Space Telescope shows that this IRDC is at least 2, and possibly as many as 5 times longer than had originally been claimed by Nessie's discoverers (Jackson et al. 2010); its aspect ratio is therefore at least 300:1, and possibly as large as 800:1. A careful accounting for both the Sun's offset from the Galactic plane ($\sim 25$ pc) and the Galactic center's offset from the $(l^{II},b^{II})=(0,0)$ position shows that the latitude of the true Galactic mid-plane at the 3.1 kpc distance to the Scutum-Centaurus Arm is not $b=0$, but instead closer to $b=-0.4$, which is the latitude of Nessie to within a few pc. An analysis of the radial velocities of low-density (CO) and high-density (${\rm NH}_3$) gas associated with the Nessie dust feature suggests that Nessie runs along the Scutum-Centaurus Arm in position-position-velocity space, which means it likely forms a dense `spine' of the arm in real space as well. The Scutum-Centaurus arm is the closest major spiral arm to the Sun toward the inner Galaxy, and, at the longitude of Nessie, it is almost perpendicular to our line of sight, making Nessie the easiest feature to see as a shadow elongated along the Galactic Plane from our location. Future high-resolution dust mapping and molecular line observations of the harder-to-find Galactic "bones" should allow us to exploit the Sun's position above the plane to gain a (very foreshortened) view "from above" of the Milky Way's structure.

We present Brut, an algorithm to identify bubbles in infrared images of the Galactic midplane. Brut is based on the Random Forest algorithm, and uses bubbles identified by >35,000 citizen scientists from the Milky Way Project to discover the identifying characteristics of bubbles in images from the Spitzer Space Telescope. We demonstrate that Brut's ability to identify bubbles is comparable to expert astronomers. We use Brut to re-assess the bubbles in the Milky Way Project catalog, and find that 10-30% of the objects in this catalog are non-bubble interlopers. Relative to these interlopers, high-reliability bubbles are more confined to the mid plane, and display a stronger excess of Young Stellar Objects along and within bubble rims. Furthermore, Brut is able to discover bubbles missed by previous searches -- particularly bubbles near bright sources which have low contrast relative to their surroundings. Brut demonstrates the synergies that exist between citizen scientists, professional scientists, and machine learning techniques. In cases where "untrained" citizens can identify patterns that machines cannot detect without training, machine learning algorithms like Brut can use the output of citizen science projects as input training sets, offering tremendous opportunities to speed the pace of scientific discovery. A hybrid model of machine learning combined with crowdsourced training data from citizen scientists can not only classify large quantities of data, but also address the weakness of each approach if deployed alone.

The physical properties of molecular clouds are often measured using spectral-line observations, which provide the only probes of the clouds' velocity structure. It is hard, though, to assess whether and to what extent intensity features in position-position-velocity (PPV) space correspond to "real" density structures in position-position-position (PPP) space. In this paper, we create synthetic molecular cloud spectral-line maps of simulated molecular clouds, and present a new technique for measuring the reality of individual PPV structures. Our procedure projects density structures identified in PPP space into corresponding intensity structures in PPV space and then measures the geometric overlap of the projected structures with structures identified from the synthetic observation. The fractional overlap between a PPP and PPV structure quantifies how well the synthetic observation recovers information about the 3D structure. Applying this machinery to a set of synthetic observations of CO isotopes, we measure how well spectral-line measurements recover mass, size, velocity dispersion, and virial parameter for a simulated star-forming region. By disabling various steps of our analysis, we investigate how much opacity, chemistry, and gravity affect measurements of physical properties extracted from PPV cubes. For the simulations used here, our results suggest that superposition induces a ~40% uncertainty in masses, sizes, and velocity dispersions derived from 13CO. The virial parameter is most affected by superposition, such that estimates of the virial parameter derived from PPV and PPP information typically disagree by a factor of ~2. This uncertainty makes it particularly difficult to judge whether gravitational or kinetic energy dominate a given region, since the majority of virial parameter measurements fall within a factor of 2 of the equipartition level alpha ~ 2.

The linewidth (sigma) - size (R) relationship has been extensively measured and analysed, in both the local ISM and in nearby normal galaxies. Generally, a power-law describes the relationship well with an index ranging from 0.2-0.6, now referred to as one of "Larson's Relationships." The nature of turbulence and star formation is considered to be intimately related to these relationships, so evaluating the sigma-R correlations in various environments is important for developing a comprehensive understanding of the ISM. We measure the sigma-R relationship in the Central Molecular Zone (CMZ) of the Galactic Centre using spectral line observations of the high density tracers N2H+, HCN, H13CN, and HCO+. We use dendrograms, which map the hierarchical nature of the position-position-velocity (PPV) data, to compute sigma and R of contiguous structures. The dispersions range from ~2-30 km/s in structures spanning sizes 2-40 pc, respectively. By performing Bayesian inference, we show that a power-law with exponent 0.3-1.1 can reasonably describe the sigma-R trend. We demonstrate that the derived sigma-R relationship is independent of the locations in the PPV dataset where sigma and R are measured. The uniformity in the sigma-R relationship suggests turbulence in the CMZ is driven on the large scales beyond >30 pc. We compare the CMZ sigma-R relationship to that measured in the Galactic molecular cloud Perseus. The exponents between the two systems are similar, suggestive of a connection between the turbulent properties within a cloud to its ambient medium. Yet, the velocity dispersion in the CMZ is systematically higher, resulting in a coefficient that is nearly five times larger. The systematic enhancement of turbulent velocities may be due to the combined effects of increased star formation activity, larger densities, and higher pressures relative to the local ISM.

Despite over 30 years of study, the mass-area relationship within and among clouds is still poorly understood both observationally and theoretically. Modern extinction datasets should have sufficient resolution and dynamic range to characterize this relationship for nearby molecular clouds, although recent papers using extinction data seem to yield different interpretations regarding the nature and universality of this aspect of cloud structure. In this paper we try to unify these various results and interpretations by accounting for the different ways cloud properties are measured and analyzed. We interpret the mass-area relationship in terms of the column density distribution function and its possible variation within and among clouds. We quantitatively characterize regional variations in the column density PDF. We show that structures both within and among clouds possess the same degree of "universality", in that their PDF means do not systematically scale with structure size. Because of this, mass scales linearly with area.

We present a study on the shells (and bubbles) in the Perseus molecular cloud using the COMPLETE survey large-scale 12CO(1-0) and 13CO(1-0) maps. The twelve shells reported here are spread throughout most of the Perseus cloud and have circular or arc-like morphologies with a range in radius of about 0.1 to 3 pc. Most of them have not been detected before most likely as maps of the region lacked the coverage and resolution needed to distinguish them. The majority of the shells are coincident with infrared nebulosity of similar shape and have a candidate powering source near the center. We suggest they are formed by the interaction of spherical or very wide-angle winds powered by young stars inside or near the Perseus molecular cloud -a cloud that is commonly considered to be mostly forming low-mass stars. Two of the twelve shells are powered by high-mass stars close to the cloud, while the others appear to be powered by low or intermediate mass stars in the cloud. We argue that winds with a mass loss rate of about 10^-8 to 10^-6 M_sun/yr are required to produce the observed shells. Our estimates indicate that the energy input rate from these stellar winds is similar to the turbulence dissipation rate. We conclude that in Perseus the total energy input from both collimated protostellar outflows and powerful spherical winds from young stars is sufficient to maintain the turbulence in the molecular cloud. Large scale molecular line and IR continuum maps of a sample of clouds will help determine the frequency of this phenomenon in other star forming regions.

We apply Support Vector Machines -- a machine learning algorithm -- to the task of classifying structures in the Interstellar Medium. As a case study, we present a position-position velocity data cube of 12 CO J=3--2 emission towards G16.05-0.57, a supernova remnant that lies behind the M17 molecular cloud. Despite the fact that these two objects partially overlap in position-position-velocity space, the two structures can easily be distinguished by eye based on their distinct morphologies. The Support Vector Machine algorithm is able to infer these morphological distinctions, and associate individual pixels with each object at >90% accuracy. This case study suggests that similar techniques may be applicable to classifying other structures in the ISM -- a task that has thus far proven difficult to automate.

We estimate the fidelity of solar neighborhood (D < 100 pc) catalogs soon to be derived from Pan-STARRS astrometric data. We explore two quantities used to measure catalog quality: completeness, the fraction of desired sources included in a catalog; and reliability, the fraction of entries corresponding to desired sources. We show that the main challenge in identifying nearby objects with Pan-STARRS will be reliably distinguishing these objects from distant stars, which are vastly more numerous. We explore how joint cuts on proper motion and parallax will impact catalog reliability and completeness. Using synthesized astrometry catalogs, we derive optimum parallax and proper motion cuts to build a census of the solar neighborhood with the Pan-STARRS 3 Pi Survey. Depending on the Galactic latitude, a parallax cut pi / sigma pi > 5 combined with a proper motion cut ranging from mu / sigma mu > 1-8 achieves 99% reliability and 60% completeness.

The winds and radiation from massive stars clear out large cavities in the interstellar medium. These bubbles, as they have been called, impact their surrounding molecular clouds and may influence the formation of stars therein. Here we present JCMT observations of the J=3-2 line of CO in 43 bubbles identified with Spitzer Space Telescope observations. These spectroscopic data reveal the three-dimensional structure of the bubbles. In particular, we show that the cold gas lies in a ring, not a sphere, around the bubbles indicating that the parent molecular clouds are flattened with a typical thickness of a few parsecs. We also mapped 7 bubbles in the J=4-3 line of HCO+ and find that the column densities inferred from the CO and HCO+ line intensities are below that necessary for "collect and collapse" models of induced star formation. We hypothesize that the flattened molecular clouds are not greatly compressed by expanding shock fronts, which may hinder the formation of new stars.

We investigate the ability to discern between lognormal and powerlaw forms for the observed mass function of dense cores in star forming regions. After testing our fitting, goodness-of-fit, and model selection procedures on simulated data, we apply our analysis to 14 datasets from the literature. Whether the core mass function has a powerlaw tail or whether it follows a pure lognormal form cannot be distinguished from current data. From our simulations it is estimated that datasets from uniform surveys containing more than approximately 500 cores with a completeness limit below the peak of the mass distribution are needed to definitively discern between these two functional forms. We also conclude that the width of the core mass function may be more reliably estimated than the powerlaw index of the high mass tail and that the width may also be a more useful parameter in comparing with the stellar initial mass function to deduce the statistical evolution of dense cores into stars.

We present arcsecond resolution mid-infrared and millimeter observations of the center of the young stellar cluster AFGL961 in the Rosette molecular cloud. Within 0.2 pc of each other, we find an early B star embedded in a dense core, a neighboring star of similar luminosity with no millimeter counterpart, a protostar that has cleared out a cavity in the circumcluster envelope, and two massive, dense cores with no infrared counterparts. An outflow emanates from one of these cores, indicating a deeply embedded protostar, but the other is starless, bound, and appears to be collapsing. The diversity of states implies either that protostellar evolution is faster in clusters than in isolation or that clusters form via quasi-static rather than dynamic collapse. The existence of a pre-stellar core at the cluster center shows that that some star formation continues after and in close proximity to massive, ionizing stars.