# Introduction

Filaments are ubiquitous in the interstellar medium. They have been identified in molecular gas for some time (Wiseman et al., 1996; Bally et al., 1987), and with higher resolution and sensitivity surveys they are being found throughout the Galactic interstellar medium (ISM) (Arzoumanian et al., 2011; Molinari et al., 2010). More recently the filaments have been identified in the Galactic ISM in atomic form (Clark et al., 2014; McClure-Griffiths et al., 2006). The properties of the HI filaments and their relationship to the molecular filaments largely remains to be determined. Filaments appear to precede star formation in the cold ISM and are thought to be linked to a process active in the gas cloud (citation not found: Andre_2014) Wareing et al., 2016). The low density atomic filaments may represent the starting point of the filament formation process, i.e., the convergence of the gas that forms the molecular filaments. On the other hand, the structures may be created naturally in the ISM with no direct connection to the gravitationally dominated molecular filaments. The primary feature of the HI filaments that has been highlighted so far is their alignment with the Galactic magnetic field (Clark et al., 2014; McClure-Griffiths et al., 2006; Clark et al., 2015; (citation not found: Kalberla_16). This infers the filaments are linked to an aligned ionized gaseous component. (citation not found: Kalberla_16) characterized the filamentary structure he found in the all-sky HI survey with $$11-16$$resolution, though it was all of the structures found after completing an unsharp mask on the data and therefore does not isolate individual filaments. (Clark 2014) identified the orientation of the linear HI features in the GALFA-HI data used here (4and 0.18 resolution) using a rolling hough transform (RHT). The RHT rolls across the image with a circular window and identifies structures that follow the diameter of the circle at the range of possible angles. This method gives alignment information, but does not extract the individual fibers in the cube to define them further as physical structures. It also requires setting a preferred velocity binning before applying the RHT. The (McClure-Griffiths 2006) focussed on denser structures in a cloud detected in HI absorption towards the Galactic Center. Though distinct from the diffuse HI filaments being examined in emission here, they serve as another useful comparison point.

For the molecular filaments, numerous properties have been extracted and studied (REFS). One property in particular that has garnered much attention is the width of the filaments. This width could determine..... Though there was some initial excitement on the filaments having a uniform width of approximately 0.1 pc (Arzoumanian et al., 2011), this has since been found to ADD PANAU PAPER INFO

The velocity structure of the atomic filaments indicates if the filaments are actively flowing across their length, or potentially rotating along the short axis. If no velocity gradient is evident, the dispersion is representative of the level of turbulence in the gas and the temperature of the gas. The width

http://arxiv.org/abs/1608.02601 find similar results to C14 that the filaments are aligned with the Galactic magnetic field using a three dimensional shock compression model to represent the Local Bubble. They find that the likely formation of the HI fibers is the turbulent shear strain consistent with (Hennebelle 2013). Quoted from Panapoulo paper for reference In studies of sub/trans-Alfv´enic turbulence, where the magnetic field is dynamically important, filament orientations with respect to the large scale ordered field, depend on whether gravity is important. In simulations where gravity is not taken into account (e.g. Falceta-Gon¸calves et al. 2008) or structures are gravitationally unbound (Soler et al. 2013), filaments are parallel to the magnetic field. filaments in wind blown bubble sim ref http://adsabs.harvard.edu/abs/2011ApJ...731...13N

Properties papers Striations in the Taurus molecular cloud: Kelvin-Helmholtz instability or MHD waves? http://arxiv.org/abs/1606.08858
arXiv:1607.00005 Date: Thu, 30 Jun 2016 20:00:05 GMT (4386kb)

Title: The magnetic field and dust filaments in the Polaris Flare Authors: G. V. Panopoulou, I. Psaradaki, K. Tassis Categories: astro-ph.GA astro-ph.SR Comments: 14 pages, 12 figures, accepted by MNRAS
In diffuse molecular clouds, possible precursors of star-forming clouds, the effect of the magnetic field is unclear. In this work we compare the orientations of filamentary structures in the Polaris Flare, as seen through dust emission by Herschel, to the plane-of-the-sky magnetic field orientation ($$\rm B_{pos}$$) as revealed by stellar optical polarimetry with RoboPol. Dust structures in this translucent cloud show a strong preference for alignment with $$\rm B_{pos}$$. 70 filaments (within 30$$^\circ$$). We explore the spatial variation of the relative orientations and find it to be uncorrelated with the dust emission intensity and correlated to the dispersion of polarization angles. Concentrating in the area around the highest column density filament, and in the region with the most uniform field, we infer the $$\rm B_{pos}$$ strength to be 24 $$-$$ 120 $$\mu$$G. Assuming that the magnetic field can be decomposed into a turbulent and an ordered component, we find a turbulent-to-ordered ratio of 0.2 $$-$$ 0.8, implying that the magnetic field is dynamically important, at least in these two areas. We discuss implications on the 3D field properties, as well as on the distance estimate of the cloud.
( https://arxiv.org/abs/1607.00005 , 4386kb)

Title: Neutral hydrogen and magnetic fields in M83 observed with the SKA Pathfinder KAT-7 Authors: G. Heald, W.J.G. de Blok, D. Lucero, C. Carignan, T. Jarrett, E. Elson, N. Oozeer, T.H. Randriamampandry, L. van Zee We show that the magnetic field of M83 is similar in form to other nearby star forming galaxies, and suggest that the disk-halo interface may host a large-scale regular magnetic field.
( https://arxiv.org/abs/1607.03365

# Observational Data and Analysis Methods

The observational data are from the Galactic Arecibo L-band Feed Array HI survey 2nd data release (GALFA-HI DR2; Peek et al. 2016). The survey covers HI velocities from -650 to +650 with a 0.184 channel spacing, 4$$\arcmin$$ spatial resolution, and a sensitivity of 150 mK per 1 bin (equivalent to a 5$$\sigma$$ column density sensitivity of $$\sim1.4\times10^{18}$$ cm$$^{-2}$$). Here we focus on HI fibers in regions with absolute Galactic Latitude greater than 30 degrees to minimize the background Galactic HI that needed to be removed and examined the velocity range +50 to -50 (see example channel in Figure ).

The fibers were identified in the GALFA data with FilFinder (Koch et al., 2015); however since this program was written to find filaments in two-dimensional data, we made substantial additions to recover the velocity dimension of the fibers. FilFinder identifies filaments in images with the following pre-processing steps: 1. Flattens the image with an arctan transform to remove the effect of small bright features, 2. Smooths the image with a median filter to limit the fragmentation of filaments WHAT DID WE USE FOR SMOOTHING RADIUS? 3. Applies an adaptive threshold in that the intensity of the central pixel must be greater than median of the neighborhood around it. PATCH SIZE WE USED? 4. Remove small spurious objects with a cut on the total area of objects and applying a small median filter. We chose a total area of 600 square pixels as the size threshold. The preprocessing steps prepare a mask that can be used to identify the filaments. Structures are identified within the mask with a medial axis transform that puts objects into 1-pixel wide skeletons maximized in one direction. The final step of FilFinder is to prune the skeletons to be filaments. This involves adopting the long axis as the body of the filament and removing small offshoots in other directions. See the (Koch 2015) paper for more details on the 2D filament identification process.

To identify the velocity structure of the fibers we run FilFinder on each individual channel images and then determine the spatial overlap between fibers in adjacent channels. Specifically, the steps for identifying 3D fibers in the GALFA data with FilFinder are as follows:

• An unsharped mask with a smoothing radius of 30$$\arcmin$$ was applied to each channel to remove the diffuse Galactic emission. A data cube with a smoothing radius of 15$$\arcmin$$ was also examined for comparison and the properties of several fibers were checked outside of the FilFinder with a 45$$\arcmin$$ unsharped mask.

• FilFinder is run on the each channel of the GALFA-HI cube and the objects found are identified with a bottom left and top right corner set, and a velocity slice index.

• The area encompassing objects in adjacent channels are compared and if there is a $$>$$75% overlap between the areas of objects in the two channels they are merged as part of the same structure. The object is discontinued when there is no longer a 75% overlap in an adjacent channel.

• The final catalog of three dimensional objects was trimmed to exclude those filaments with sizes $$<5000$$ pixels and aspect ratios $$>2$$. This cut was necessary to reduce the large number of objects found and optimizes the filaments to be long, thin structures.

This is different from how the HI fibers were identified in (Clark 2014) with the Rolling Hough Transform (RHT). With the RHT filaments were identified blah.....

To extract the fiber properties from the cube we create three distinct maps: 1. an integrated intensity (moment 0) map to obtain column densities, lengths and masses, 2. an average velocity (moment 1) map to examine velocity gradients along the filaments, and 3. a velocity dispersion (moment 2) map to examine the velocity widths. These maps are created using the moment task within Miriad on the USM cubes. The actual values are obtained from these moment maps by isolating the fiber in position and extracting the typical values for the fibers. Though somewhat labor intensive, the combination of using the USM cube and isolating in position was found to be the best way to isolate the fiber from the abundant additional surrounding diffuse Galactic emission. Signatures of velocity gradients were largely examined by eye (see Figure ).

The typical width of the fibers is an important property to derive from the observations given how it factors in to other physical properties and the predictions from simulations. The width was not easy to obtain from the integrated intensity maps as there are variations in this parameter along the lengths of the fibers. We utilized the FiLTer program of on the RHT fibers to obtain the typical widths of the HI fibers. This therefore extracts the width properties of all of the RHT identified fibers in Figure . FiLTer...ADD BRIEF FILTER DESCR SHOULD PROBABLY COMPARE THIS TO WHAT YOU WOULD EXTRACT BY EYE FOR MIN/MAX. IF PROGRAM ISN’T BEHAVING GO BACK TO THIS IDEA WITH UG

# Results

The properties of the filaments are shown in Table \ref{props.tab}. The physical properties of width, length, volume density, pressure and HI mass are derived assuming the filaments are at a distance of 100 pc. An estimate for the temperature is obtained from the linewidth of the fiber, which is derived from the average dispersion for the cloud, $T (K) = 21.8 (2.355\sigma)^2$ This is an upper limit on the temperature as any turbulent component to the dispersion is not removed.

The thermal pressure can be estimated using, $P_{\rm th}/k (K cm^{-3})= n_H T$ but this is an upper limit on the thermal pressure given the temperature is an upper limit. Given our data, a derivation of the total pressure from the dispersion is more relevant. $P = \sigma^2 1.4 m_H n_H$ THESE TWO ARE DIRECTLY RELATED SINCE VARIABLES IN BOTH ARE THE SAME, SO PROBABLY ONLY INCLUDE ONE.

# Discussion

Following (McClure-Griffiths 2006), we can constrain the strength of the magnetic field by assuming the alignment of the fibers with the field indicates it is the mechanism dominating its observed structure. The magnetic energy density should therefore be larger than the kinetic and gravitational energy densities.

relevant section from Papadoulo paper on relevant other papers In a series of papers the Planck collaboration compared the magnetic field to ISM structure across a range of hydrogen column densities (NH). Planck Collaboration Int. XXXII (2014) considered the orientation of structures in the diffuse ISM in the range of NH ∼ 1020 − 1022 and found significant alignment with the plane-of-the-sky magnetic field (Bpos). Planck Collaboration Int. XXXV (2015) found that in their sample of 10 nearby clouds, substructure at high column density tends to be perpendicular to the magnetic field, whereas at low column density there is a tendency for alignment. Studies of optical and NIR polarization, tracing Bpos in cloud envelopes, show that dense filaments within star-forming molecular clouds tend to be perpendicular to the magnetic field (Pereyra & Magalh˜aes 2004; Alves et al. 2008; Chapman et al. 2011; Sugitani et al. 2011). On the other hand, there are diffuse linear structures termed striations, that share a common smoothly varying orientation and are situated either in the outskirts of clouds (Goldsmith et al. 2008; Alves de Oliveira et al. 2014) or near dense filaments (Palmeirim et al. 2013). These structures show alignment with Bpos (van den Bergh 1956; Chapman et al. 2011; Palmeirim et al. 2013; Alves de Oliveira et al. 2014; Malinen et al. 2015). The extremely well-sampled data of Franco & Alves (2015) in Lupus I show that Bpos is perpendicular to the cloud’s main axis but parallel to neighbouring diffuse gas. There are, however, exceptions to this trend (e.g. L1506 in Taurus, Goodman et al. 1990).

Examples of gaussian filters with covariances defined by a rotation angle (labeled).

### References

1. Jennifer J. Wiseman, Paul T. P. Ho. Heated gaseous streamers and star formation in the Orion molecular cloud. Nature 382, 139–141 Nature Publishing Group, 1996. Link

2. John Bally, Antony A. Stark, Robert W. Wilson, William D. Langer. Filamentary structure in the Orion molecular cloud. ApJ 312, L45 IOP Publishing, 1987. Link

3. D. Arzoumanian, Ph. André, P. Didelon, V. Könyves, N. Schneider, A. Men’shchikov, T. Sousbie, A. Zavagno, S. Bontemps, J. Di Francesco, M. Griffin, M. Hennemann, T. Hill, J. Kirk, P. Martin, V. Minier, S. Molinari, F. Motte, N. Peretto, S. Pezzuto, L. Spinoglio, D. Ward-Thompson, G. White, C. D. Wilson. Characterizing interstellar filaments with Herschel in IC 5146. Astronomy & Astrophysics 529, L6 EDP Sciences, 2011. Link

4. S. Molinari, B. Swinyard, J. Bally, M. Barlow, J.-P. Bernard, P. Martin, T. Moore, A. Noriega-Crespo, R. Plume, L. Testi, A. Zavagno, A. Abergel, B. Ali, L. Anderson, P. André, J.-P. Baluteau, C. Battersby, M. T. Beltrán, M. Benedettini, N. Billot, J. Blommaert, S. Bontemps, F. Boulanger, J. Brand, C. Brunt, M. Burton, L. Calzoletti, S. Carey, P. Caselli, R. Cesaroni, J. Cernicharo, S. Chakrabarti, A. Chrysostomou, M. Cohen, M. Compiegne, P. de Bernardis, G. de Gasperis, A. M. di Giorgio, D. Elia, F. Faustini, N. Flagey, Y. Fukui, G. A. Fuller, K. Ganga, P. Garcia-Lario, J. Glenn, P. F. Goldsmith, M. Griffin, M. Hoare, M. Huang, D. Ikhenaode, C. Joblin, G. Joncas, M. Juvela, J. M. Kirk, G. Lagache, J. Z. Li, T. L. Lim, S. D. Lord, M. Marengo, D. J. Marshall, S. Masi, F. Massi, M. Matsuura, V. Minier, M.-A. Miville-Deschênes, L. A. Montier, L. Morgan, F. Motte, J. C. Mottram, T. G. Müller, P. Natoli, J. Neves, L. Olmi, R. Paladini, D. Paradis, H. Parsons, N. Peretto, M. Pestalozzi, S. Pezzuto, F. Piacentini, L. Piazzo, D. Polychroni, M. Pomarès, C. C. Popescu, W. T. Reach, I. Ristorcelli, J.-F. Robitaille, T. Robitaille, J. A. Rodón, A. Roy, P. Royer, D. Russeil, P. Saraceno, M. Sauvage, P. Schilke, E. Schisano, N. Schneider, F. Schuller, B. Schulz, B. Sibthorpe, H. A. Smith, M. D. Smith, L. Spinoglio, D. Stamatellos, F. Strafella, G. S. Stringfellow, E. Sturm, R. Taylor, M. A. Thompson, A. Traficante, R. J. Tuffs, G. Umana, L. Valenziano, R. Vavrek, M. Veneziani, S. Viti, C. Waelkens, D. Ward-Thompson, G. White, L. A. Wilcock, F. Wyrowski, H. W. Yorke, Q. Zhang. Clouds filaments, and protostars: The Herschel Hi-GAL Milky Way. Astronomy and Astrophysics 518, L100 EDP Sciences, 2010. Link

5. S. E. Clark, J. E. G. Peek, M. E. Putman. MAGNETICALLY ALIGNED H I FIBERS AND THE ROLLING HOUGH TRANSFORM. ApJ 789, 82 IOP Publishing, 2014. Link

6. N. M. McClure-Griffiths, J. M. Dickey, B. M. Gaensler, A. J. Green, Marijke Haverkorn. Magnetically Dominated Strands of Cold Hydrogen in the Riegel-Crutcher Cloud. ApJ 652, 1339–1347 IOP Publishing, 2006. Link

7. C. J. Wareing, J. M. Pittard, S. A. E. G. Falle, S. Van Loo. Magnetohydrodynamical simulation of the formation of clumps and filaments in quiescent diffuse medium by thermal instability. Monthly Notices of the Royal Astronomical Society 459, 1803–1818 Oxford University Press (OUP), 2016. Link

8. S. E. Clark, J. Colin Hill, J. E. G. Peek, M. E. Putman, B. L. Babler. Neutral Hydrogen Structures Trace Dust Polarization Angle: Implications for Cosmic Microwave Background Foregrounds. Phys. Rev. Lett. 115 American Physical Society (APS), 2015. Link

9. Patrick Hennebelle. On the origin of non-self-gravitating filaments in the ISM. Astronomy & Astrophysics 556, A153 EDP Sciences, 2013. Link

10. Eric W. Koch, Erik W. Rosolowsky. Filament identification through mathematical morphology. Monthly Notices of the Royal Astronomical Society 452, 3435–3450 Oxford University Press (OUP), 2015. Link