Adam Ginsburg's blog

Articles by Adam Ginsburg

• Regrid an image (or cube) using CASA

  In order to regrid an image, you need some target reprojection. The simplest way to acquire such a reprojection is by extracting it from another image, e.g. a FITS image

  (I use >>> to indicate the CASA <#>: prompt here because it allows you to copy & paste the code into CASA)

  >>> header = imregrid(imagename='file.image', template='get')
  

  This header will be a python dictionary with two keys, csys, which specifies the coordinate system, and shap, which gives the shape.

  If you want to write up your own FITS header and you don't have a FITS file with those exact coordinates, you can use the header_to_template convenience tool in casa_tools.

  >>> from casa_tools import header_to_template
  >>> from astropy.io import fits
  >>> fits_header = fits.Header.fromtextfile('my_header.hdr')
  >>> header = header_to_template(fits_header)
  

  In order to regrid the file, you then need to pass in the image name and template to imregrid.

  >>> imregrid(imagename='file.image',
  ...          template=header,
  ...          output='file_regrid.image')
  

  If your input image's header did not include a 'telescope' keyword, CASA will complain:

  2013-12-18 16:46:33 SEVERE imregrid::ImageRegrid::regrid (file /opt/casa/release-4_1_0-release-1/darwin64/include/casacore/images/Images/ImageRegrid.tcc, line 118) Cannot find the observatory name UNKNOWN in the CASA

  You can add the keyword to the header yourself:

  >>> imhead('file.image',
  ...        mode='put',
  ...        hdkey='telescope',
  ...        hdvalue='Arecibo')
  

  References

  • Sebastian Muller's presentation (page 16)
  • imregrid docs

---

• Converting a FITS cube to a CASA Image

  If you want to perform analysis on a FITS file in CASA, you first need to import it into .image format.

  importfits('file.fits','file.image')

  If this works, great! You can move on. CASA will treat NaN values in an image as 'masked'.

  Dealing with Errors

  There may be issues with FITS headers. CASA respects a large number of header keys (reference).

  There are many ways to edit a FITS header. One of the most straightforward to use is edhead. Otherwise, one can use CASA's imhead

  CASA Keyword	FITS keyword	Description
  beammaj	BMAJ	Major axis of the clean beam
  beammin	BMIN	Minor axis of the clean beam
  beampa	BPA	Position angle of the clean beam
  bunit	BUNIT	Brightness unit (K, Jy/beam, etc)
  cdeltn	CDELTn	Pixel size, nth axis  (max n is 4)
  crpixn	CRPIXn	Pixel coordinate of reference point, nth axis
  crvaln	CRVALn	Pixel location of reference point, nth axis
  ctypen	CTYPEn	Axis name, nth axis. For FITS, this includes the projection
  cunitn	CUNITn	Pixel units, nth axis
  datamax	DATAMAX	Maximum pixel value in image
  datamin	DATAMIN	Minimum pixel value in image
  date-obs	DATE-OBS	Date of the observation
  equinox	EQUINOX	Reference frame for directional coordinates
  imtype		Image type: intensity,
  minpos		Position of the minimum value (world unit)
  minpixpos		Same in pixel (array)
  maxpos		Position of the maximum value (world unit)
  maxpixpos		Same in pixel (array)
  object	OBJECT	Source name
  observer	OBSERVER	Observer name
  projection	CTYPEn	Image projection ('SIN','TAN', or 'ZEA')
  reffreqtype		Reference frame for the spectral coordinates
  restfreq	RESTFREQ	Rest Frequency
  shape	NAXISn	Number of pixels along each axis
  telescope	TELESCOP	Telescope name

  What does it mean if you get this sort of error?

  CASA <1>: importfits('file.fits','file.image') 2013-12-18 13:06:18 WARN importfits::FITSCoordinateUtil::fromFITSHeader The wcs function failures are too severe to continue ... 2013-12-18 13:06:18 WARN importfits::ImageFITSConverterImpl::FITSToImage (file /var/rpmbuild/BUILD/casapy-stable/casapy-stable-42.0.26945/casacore/images/Images/ImageFITSConverter.tcc, line 71) No proper coordinate system defined in FITS file. Using dummy linear system instead.

  First, go to the casapy-yyyymmdd-hhmmss.log file and look at the errors.

  2013-12-18 13:06:18 INFO importfits::FITSCoordinateUtil::fromFITSHeader celfix incurred the error Inconsistent or unrecognized coordinate axis types 2013-12-18 13:06:18 INFO importfits::FITSCoordinateUtil::fromFITSHeader spcfix incurred the error Inconsistent or unrecognized coordinate axis types 2013-12-18 13:06:18 INFO importfits::FITSCoordinateUtil::fromFITSHeader cylfix incurred the error Inconsistent or unrecognized coordinate axis types

  In my case, the error turned out to be that CTYPE3 was set to RADI-LSR, while it should be VELO-LSR to be recognized by the CASA system. The velocity convention is, unfortunately, lost.

---

• Notes from the ISM-PP conference: Day 5

  Linda Tacconi

  High-redshift

  • Outflows with mass-loading parameters 1-10
  • velocity dispersions 25-80 km/s + seen in both CO and H-alpha

  Avishai Dekel: Violent Disk Instability

  Galaxies form at nodes in web at high z

  • disk surrounded by messy "interaction region"
  • disk outflows observed
  • Toomre unstable clumps ~10^9 msun + dynamically dominant (unlike MW clumps) + violent instability, not secular
  • long time spent in disks
  • Stellar-dominated galaxies at z~1 stop "VDI" + bulge stabilizes inner part of galaxy
  • "clumps" are rapidly spinnign
  • both VDI and "ex-situ" dark matter clumps
  • universal SF law: divide by t_ff + t_ff is either Toomre tff or gmc tff
  • "redshift 2 is the only important phase, forget about z=0"
  • Steve: How does "universal law" fit with Galactic CMZ? + Ignorance. Ask Mark.
  • Rahul: Toomre Q threshold only works for infinitely thin disk. Shouldn't threshold be lower because of thick disk?
  • Goldreich & Linden-Bell: 30% effect. Elmegreen addresses this.

---

• Notes from the ISM-PP conference Day 4

  Alberto Bolatto

  GMC lifetime: 100 Myr (Scoville 79, Koda) Size-linewidth relation:

  • Heyer 2001: small, flat dispersion clouds
  • outer galaxy clouds contained by external pressure

  Dust in IGM hints at molecular Galactic outflows

  Jin Koda: Molecular to Total gas ratio

  inner 1 kpc 100% molecular

  • 20-30% arm/interarm variations
  • molecules within interam and in arms, but clumps only in arms + Sawada 2012 + Power-Law tails in arms, but not in interarms
  • GRS more clumpy in arms
  • arm/interarm comparison in M51 + 2-3x increase in T or rho in spiral arms
  • Q Diederich: interarm clouds short-lived? Not self-gravitating?
  • A: Molecular gas is surviving, small clumps are self-gravitating
  • Why is BDI higher at tangent point?
  • What about Bania's clump?

  Superbubbles as a physical process in the ISM 15': Martin Krause

  Al26 is much faster than spiral arms

  • Too fast orbital velocity...

  My question: What is the scale height? What directions is the Al26 going?

  • expand 100 pc in all directions except back along the line of sight + Are the velocities consistent with the stellar velocity distribution?

  Star Formation in molecular gas outside galaxies: Ute Lisenfeld

  Galaxies star forming, extragalactic gas with high column & high velocity dispersion not

  • colliding galaxies: intergalactic region full of molecular gas

  Andreas Schruba: CARMA + IRAM Andromeda survey

  • All-galaxy survey with IRAM + 175 hr, 335 mosaic for "Brick 15"
  • most clouds unbound or pressure bound
  • size-linewidth relation: still lots of scatter
  • sigma^2/r vs surface brightness (same as Heyer 2009 plot) + no correlation with any parameters...
  • clouds in strongest star-forming regions have largest LOS velocity differences

  Rahul Shetty: KS fitting

  Linear relation indicates constant depletion timescale

  • super-linear: efficiency increases with increasing density
  • linear: constant efficiency with gas density + GMCs all have similar/same properties + t_dep = 2 Gyr
  • CO traces individual clouds
  • sublinear indicates non-star-forming CO
  • no universal or representative depletion time + depletion time increases with CO
  • Adam Leroy: Surface density of H2 doesn't match surface density of molecular clouds
  • Jin: Also found superlinear. Systematic bias?

  Javier Cardiel: Two regimes of SF

  Looking for triggering

  • density decreases with size increasing?
  • HII regions and molecular clouds have same size-mass relations

  Steve Longmore: ISM in Galactic Centers

  dolphin riding? Where is MW in hubble tuning fork?

  • dense gas with no star formation driven inwards by bar
  • Meidt M51 PAWS

  Molinari Ring: closed orbit... M83 is a twin Jay Gallagher's talk: SFR much, much higher in NGC 253 than MW Discrepancy between Farhad, Steve, and Sandstrom

  • TOTAL gas: CMZ agrees
  • Geometry: Farhad assumed pancake, Steve assumed ring

  Kruijssen & Longmore similarity to z=2-3...

  • CMZ clouds "indistinguishable" in a few parameters from high-Z

  Turbulent pressure provides the support in the CMZ

  • environmentally dependent density threshold
  • Q Rowan: Constant gradient in cloud properties along the ring?
  • Q Semadeni: Turbulence can't provide support.
  • Q Keto: Nothing can supply the turbulence
  • Q Falgarone: HCO+ absorption, daughter of CO+. Could trace sites of turbulence in the medium. Low density. Line width comes from intense shear.

  Tim Davis: ISM in early-type galaxies

  ISM in CMZs different

  • some are predominantely molecular (no HI)
  • cluster members have different NII/CII ratios
  • lower star forming efficiency
  • shear increases depletion time
  • Q Burkhart: Origin of gas. Replenishment from stars?
  • A: Based on gas vs star kinematics, at least 50% must be accreted.
  • Q Kruijssen: Stripping would happen if disk extends to optical radius. Does it? Stars should be stripped too
  • A: Don't extend to optical radius, but more truncated in clusters
  • Q Leroy: Where does tff claim come from...?
  • A: assume all mol clds have 85 msun/pc^2...

---

• Notes from the ISM-PP conference Day 3

  Mordecai Mac-Low: Stellar Feedback

  SNe explode away from their birthplaces

  • ionizing radiation contains 10x more energy than SNe
  • outer disk SF inadequate to explain kinetic energy observed
  • "almost constant velocity dispersion in HI in local universe"
  • ionization escape determined within 1pc (Peters) + fragmentation starvation
  • At high mass, high escape velocity, there is less unbound material
  • stellar winds relatively unimportant because of their rarity + B-stars give SNe, but not wind

  Radiation Pressure

  • more effective if more scattering + Krumholz & Matzner, Fall: Momentum Driven (0-1 scatterings) + Thomson, Murray, Hopkins: Energy Driven (many-scattering)
  • Multi-scattering fails: RT instability + scattered radiation never dominant
  • Lopez 2013: HII-region pressure always dominant; hot gas slightly lower
  • Q Fan of CR driven winds. Galactic winds are multi-phase. Thoughts?
  • A:
  • Q Thomas Henning: What about positive feedback?
  • A: Will be addressed in next talk. But, positive feedback is a 10% effect

  Molecular cloud formation in stellar feedback flows: Observing, demonstrating, quantifying: Joanne Dawson

  HI GASS + 12CO

  • effect of supershells on molecular gas fraction + shells are more molecular than surroundings
  • LMC: supergiant shells have comparable molecular gas fraction to the rest of the gas + Large scale: No effect + small scale: drive up local molecular gas fraction
  • OB cluster feedback: Not a major driver, but an important secondary role
  • Q Zinnecker: How did 30 Dor form?

  Disruption of GMCs by photoionization and stellar winds - setting the stage for supernovae. Jim Dale

  The time between first O-stars and first SNe

  • mass loss is a strong effect of escape speed
  • Leakage? Approximate winds as momentum fluxes + leakage rate constant across M/R space
  • Effect of 1st SN? + no effect/small effect for high-mass clouds

  Do you have predictions for velocity of knots as a function of distance or time? Rocket effect? Can "Rocket Effect" drive out gas?

  The evolution of molecular clouds under the influence of ionizing radiation 15': Matthias Gritschneder

  Pipe Nebula affected by feedback O vs B star

  • B-star does not generate structure + heats shell + stays very smooth
  • O-star compresses

  Feedback-driven turbulence in the multi-phase ISM 15': Andrea Gotto (MPA)

  Cannot reproduce HI velocity dispersion

  Understanding ultracompact H II regions 15': Thomas Peters

  HII region morphologies

  • no relation between age of star and size of HII region
  • probabilistic flux decrements
  • SgrB2: 1 shrinking, 2-3 growing + predictions & observations match exactly
  • ALMA outflow predictions
  • K 3-50A VLA + CARMA observations + Pamela Klaasen 14.7 GHz obs of K3-50a + ionized gas entraining molecular gas
  • models do not include dust absorption

  ALMA proposal for Sgr B2? Prediction for 3mm observability?

  An analytic model for the dynamics of the ionized outflows of massive protostars 15': Keto

  Observations that aren't explained by expanding shell

  • supersonic velocities in HII region
  • SED indicates density gradients
  • Modified Hollenbach photoevaporating disk model + velocity gradient of RRL not tied to disk or outflow + "Ballistic Infall" model
  • Q: Look similar to simulations of champagne flows.

  Deciphering the violent interaction between very massive stars and their natal clouds in the Carina Nebula Complex 15': Thomas Preibisch

  Carina:

  • HAWK-I NIR stellar population study
  • CCCP Chandra 10,000 stars
  • 60,000 msun in gas
  • 10^6 msun total... half cold
  • 300,000 msun of CO
  • 1% of total SF in galaxy
  • G0 ~100-1000
  • hot wind gas (X-ray) flows out: anticorrelated with 70um Herschel
  • Q How do you know SFE is increasing (currently 3%)?
  • A: We see it going on now. Stellar mass is increasing. Efficiency must increase further.
  • Q Nicola Schneider: What if the pillars are eroded filaments? Revealed, not triggered?
  • A: Same mechanism? (unclear)
  • Q Hennebelle: Where is most of the mass? only 1/3 in CO?
  • A: 1/2 - 2/3 still in atomic gas

  Models for the circumstellar medium of massive runaway stars 15': Dominic Meyer

  Bow shocks from runaway RSGs

  • v*/vwind >> 1 -> stable bow shock + slow stars -> unstable

  Dynamics of H II regions around exiled O stars 15': Jonathan Mackey

  25% of O-stars ejected

  • once in diffuse ISM, easier to feedback on Galactic scales
  • Zeta Ophiuci d~112pc, proper motion known, LOS velocity unknown
  • bow shock is 5 degrees! across, ~10pc

---

• Notes from the ISM-PP conference: Day 2

  Collapse of molecular clouds and the IMF 40': Matthew Bate (University of Exeter)

  Analytical models predict dependence on density, mach number, etc

  • Accretion stopping time + dynamics
  • kinetic feedback, b-fields

  Generating the IMF

  • Fragmentation to produce "stellar groups" + how fragmentation proceeds is not importan
  • Only knob that changes the IMF is the initial Jeans mass of the cloud
  • process is always: start with jeans mass, accrete up or get ejected
  • gets multiplicity right
  • Thermal feedback
  • Radiative feedback -> less fragmentation, especially less disk frag
  • radiative feedback erases dependence of IMF on initial conditions + fewer objects -> less dynamics -> less frequent ejection
  • Bate 2012 "turning point" - can reproduce IMF + can try for a predictive theory of SF
  • metallicity comparison + same IMF over 300x difference in Z
  • Krumholz: very high temperatures suppress fragmentation, -> high-mass IMF + outflows "alleviate overheating problem"
  • B-fields increase accretion, enhance thermal feedback
  • Q Hennebelle: What is competitive accretion? How do you conclude that it is the mechanism going on in the sim?
  • A: Competitive accretion is stopping time.
  • Q about terms - competitive accretion governs high mass, but term applied to low mass
  • A: All are competing, but likelihood of losing...
  • Q Paul Clark: Gas-Dust coupling happends at 10^5. Are you including metallicity dependence of coupling? I don't believe metallicity results.
  • A: No. We assume dust is always dominant. Large scale temperatures don't matter. Temperatures matter on the small, self-gravitating scales
  • Clark: Initial cores would have been more massive.
  • Bate: No, not determined by temperature, only by turbulence.
  • Q Zinnecker: Competing theory of turbulent cores. Compare them.
  • A: Fabian Heitsch did that well yesterday. Turbulent core model relies on turbulence providing isotropic pressure.
  • Q Kruijssen: Prediction is that dynamical times vs. disk survival times. Boundedness of cluster? group? has big effect. YMCs don't have top heavy. How do you make bottom-heavy IMFs?
  • A: "Can't believe I'm saying this..." possible you can get bottom-heavy imf with extreme turbulence. Mass reservoirs stripped away by high-mass flows. Prevents long-time accretion.

  Tracing the fragmentation of OMC1-north filament with the Submillimeter Array 15: Paula Teixera

  with Satoko Takahashi, Luis Zapata, Paul Ho

  • Cores detected in OMC1N NH3 filaments
  • characteristic separation scales
  • "core pods" instead of "clumps"
  • OMC1N has smaller nearest-neighbor separation than OMC2/OMC3
  • Q: Direction of B-field?
  • A: Perpendicular to filament
  • Q: Scales can be interpreted as two different fragmentation scales
  • Q Falgarone: Do you include short-spacing in the maps?
  • A: No.
  • Q: Bias.
  • A: not on short scales.

  Filament formation and star formation regulation in collapsing molecular clouds 15': Vazquez-Semadeni

  Case for globally contracting molecular clouds

  • "turbulent infall" rather than "turbulence" + no "turbulent support"
  • "turbulence is overrated" ?
  • collapse is pressureless
  • "whole thing collapsing into global potential well..."
  • velocity jumps around local potential wells
  • "two-filament appearance" due to presence of wells. + Hacar gaussian fits yield 2 filaments (incorrectly)
  • How address SFR too high for global collapse? + ionization feedback -> cloud destruction + 10% efficient... THIS IS STILL AN ORDER OF MAGNITUDE TOO HIGH

  So are all molecular clouds gravitationally collapsing? Why aren't they forming stars on a free-fall time?

  • Q Keto: Your model is different. Stars forming with accretion/flows along filaments.
  • A: No one has implied filaments are static... no contradiction.

  The structure and star-forming fate of the Galactic centre cloud G0.253+0.016 15': Katharine Johnston

  Stats...

  • hot gas, cool dust
  • no >B0 stars
  • SCUBA+SMA image + column density PDF + plateau at low column density (could this be pixelation?)
  • shocks in the south
  • hint of cloud-cloud collision
  • cloud-cloud collision comparison: Anathpindika et al 2011 + many claims that super star clusters form from cloud-cloud collisions
  • ring view is reversed + center of galaxy doesn't have to be ring, could be spirals or disk
  • Molinari ring fails
  • Q Xu: Single cloud... surprised you find only one cloud.
  • A: data only cover one cloud...
  • Q Kainulainen: Comment on low SFR. If follow analytic formulations, care about relative density contrast. Despite enormous mean density, the relative density contrast is low. Not surprised by low SFR.

  Laboratory Studies of Dust Formation and Processing 40': Cornelia Jäger

  There are many types of dust

  • "The ISM is the most dangerous place for dust."
  • 49um band of Forsterite
  • 69um band of crystalline forsterite + these can be used as a dust thermometer
  • beta-T anticorrelation + SiO2 absorption changes strongly with temperature + crystalline material has no beta-T, but amorphous DOES have a beta-T
  • generation of fullerenes in the lab at T>3500K
  • dust growth
  • Helium droplet experiment + "helium droplet beam" + superfluid helium shot through a "pickup oven" at 0.4 K + forms SiO clusters + "barrierless" reactions
  • laser evaporation of particles, condense on a substrate
  • ion-induced processing
  • irradiated silicates lead to iron inclusion
  • Q: Beta/temperature. Observations... beta is effective beta, non-constant temperature is problematic. Observed temperatures are much smaller.
  • A: Yes, there is still dependence at low temperature, but weaker.
  • Q: Are there any features in the mm that we can use to distinguish the species in the submm?
  • A: not yet.

  Combining experimental techniques for comprehensive astrophysical case studies 15': Holger Kreckel (MPIK, Heidelberg)

  Gas phase chemistry

  • terminate branches with recombination with free electrons
  • Cryogenic Storage Ring + electrostatic field + strong vacuum, low (10K) temperature
  • C- beam driven electrically, dissociate electron with a laser + "Don't normally need 2 kW continuous lasers except for death stars."
  • studying cosmic ray ionization rate + CR ionization is missing some rate coefficients
  • Q Glover: Why only go to 40K?
  • A: Divergence of beam limits temperature

  Laboratory studies on electron collisions of atomic and molecular ions 15': Andreas Wolf (MPIK, Heidelberg)

  HD formation

  • need cooled vacuum chamber to match ISM conditions

  Chemical processes in the ISM: Gas and molecules 30': Simon Glover (ZAH/ITA University of Heidelberg)

  Noble gas molecule in ISM: ArH+ in SNR Why are molecular clouds cool?

  • Molecular clouds are cool because of dust shielding, NOT CO cooling + C+, CI are just as efficient as CO + CO allows you to cool from 20 to 10K at n~10^2-10^3
  • production of high-density material drives CO formation

  Chemistry as a tracer of physical conditions in the ISM

  • CR rate measurement tool: H3+ -> CR rate + if you can measure C+ + diffuse sightlines have values >> "canonical dark cloud value" = 10^-17
  • HF: most evil substance imaginable - seep through skin, melt bones + exothermic: most fluorine lands in HF + depleted in dense gas
  • CH+ puzzler: formation slow, destruction fast + overabundant by 3-4 orders of magnitude
  • Isotopologue chemistry for CO + Fractionation: O prefers to be in 13CO. Can convert 12CO->13CO + photodissociation: less self shielding in 13CO, 13CO->12CO
  • Uncertainties: + Need good rate coefficients + Chemistry not in steady state: memory effects + PDR modeling done with GCs treated as slabs
  • Q: Are there molecules for which the uncertainties are likely small enough now?
  • A: H2 cooling doesn't play a significant role in the thermal balance of the cloud
  • Q: Matthew Bate: Can observers use H2 (HF?) to deal with dark gas?
  • A: Not clear yet.
  • Q: Deuteration important? Influence... things?
  • A: Timer for core collapse
  • Q Edith Falgarone: Disagree. We know C-shocks are heating gas to few hundred K, contributes to C+ emission. Need additional heating source. Balance not between ionization...?
  • A: Depends on assumptions of photoelectric heating efficiency. We find shock dissipation not important in dense stuff. H2 may be important in diffuse gas.

  Chemical processes in the ISM: Dust 30' Speaker: Thomas Henning (MPIA Heidelberg)

  Small particles heat the gas Formation/destruction balance

  • AGB star dust formation not efficient enough to balance dust destruction
  • NIR extinction law: Fritz et al 2011 towards GC + dust towards GC is different
  • Particles in mol clds grow to few microns at most

  Why is there dust in the ISM?

  • Cold ISM formation?
  • Or core-collapse SNe? + Formation agrees with predictions, destruction too much...
  • Low-temperature dust formation? + works for SiO2 in the lab

  Grain models:

  • composite grains with different size distributions
  • mixture of graphitic and silicate grains...
  • Q: SNe still not enough for other galaxies.
  • A: Our galaxy needs 0.5 per SNe. More of a problem for low metallicity.
  • Q Falgarone: ???
  • Q Zinnecker: Dusty wolf-rayet stars. Also dust factories?
  • A: Not in normal environments. O-stars may be a solution for high-Z galaxies.
  • Q: How much oxygen in dust?
  • A: More in silicate.

  Galactic dust as seen by Planck 15': Marta Alves

  All-sky dust temperature map

  • outer galaxy cold, inner galaxy warmish
  • increase of dust temperature towards poles
  • dust opacity anticorrelated with temperature + not correlated with luminosity
  • Planck extinction maps
  • Flattening of SED at mm wavelengths
  • Q Rahul: Have you accounted for degeneracies in fits?
  • A: Beta map was observed at low-resolution, then did a second run of the fit to extrapolate down to high resolution

  Formation signatures and carbon budget of molecular clouds 15': Henrik Beuther

  Signatures for cloud formation

  • CI, CO, C+ observations of 4 IRDCs
  • G11.11 IRDC + ionized lines very narrow
  • [CII]/[CI]/CO = 1/5/70 + ionized carbon measurement unknown

  Svitlana Zhukovska

  • Q Zinnecker: Gas to dust ratio varies from dense to diffuse gas (Meixner LMC).
  • A: Yes,there is dependence

  Dependence of star formation on ISM properties 40': Adam Leroy (NRAO Charlottesville)

  "Two Bottleneck" view

  • Atomic Gas -> Molecular Gas -> Dense Gas
  • cloud formation is the first regulating step
  • more molecular = more star forming
  • IR/CO vs HCN/CO... weakly correlated
  • HCN/IR approximately constant over wide range of scales
  • factor of 2 scatter in SFE

  Sarah Ragan

  • Q Semadeni: different surface density -> different evolutionary stages? Scaling law depends on how you define objects

  Amy Stutzki

  • Q Semadeni: Higher lum implies higher SFR?
  • A: lum higher, but maybe that means shorter accretion time.

---

• Notes from the ISM-PP conference

  The thickness distribution of interstellar filaments: Evangelia Ntormousi

  Filament decomposition

  • RAMSES MHD with/without ambipolar diffusion (AD) + effect of ambipolar diffusion is smoothing + AD builds very slight secondary peak
  • Q Crutcher: How strong B-fields? AD driven by turbulence (or what drives it)?
  • A: Yes, turbulence drives it. B-fields ???
  • Q MacLow: Alfven waves suppressed by AD, but magnetosonic go through & drive small flows. Oishi & MacLow: no break in power-spectrum at small scales. Consistent with small difference.
  • A: Yes, maybe smaller scale motions.
  • Q Keto: Keep B-fields coherent. But don't need to keep them coherent. How does B-field keep filaments coherent? Argues that they only last a crossing time.
  • A: magnetic tension, resists shear. Filaments would reexpand without B-fields. We don't see observed reexpansion.

  Anomalous extinction: The degeneracy between dust composition and geometry 15': Peter Scicluna (ESO)

  Forward scattering by dust

  • direct forward scattering favored
  • flatter extinction curve when optically thick
  • Q Keto: Wavelength dependence of scattering? Mostly absorption -> IR reemission?
  • A: absorption is treated. Heating.
  • Q Star outside scattering into LOS?
  • A: Yes, but probably not as significant

  Latest GREAT results from SOFIA 15': Hans Zinnecker

  SOFIA summary... 1000 hrs/year

  • Germany DLR 20% + US

  GREAT results

  • M17sw C+, CO 13-12 + CII more distributed
  • G5.89 CO 11-10 outflow
  • Inverse P-Cygni at NH3 1810 GHz in G34.26+0.15
  • OH ground state absorption towards W49 + Wiesemeyer 2012 AA 542 L7
  • OD detected towards IRAS1629A
  • CND in dust with SOFIA: Requenna-Torres
  • Southern targets: NGC 3603, LMC
  • CO & C+ maps of NGC3603 (cool!)

  Q Vazquez-Semadeni: What happens when plane hits turbulence? A: Counterweight keeps telescope pointed in the right place

  Molecular cloud properties 40': Alvaro Hacar

  Skipped

  The structure of molecular clouds from 1000 AU to Orion 15': Joao Alves

  Extinction Mapping (in Orion)

  • Stars at 2um are colorless

  • Bias: not sensitive to densest part of cloud + NICEST accounts for background population

  • PIPE: high temperatures in the low-column stuff

  • Barnard 68 is a bonnor ebert - 2001 result.. + TP role of B68!

  • Most BE cores are stable (Lada 2008) + pressure confined

  • CMF claim... + "same thing seen by Andre 2011"

  • Burkert Alves 2009: merging cores

  • Hacar 2013: multiple LOS velocity components along a single filament + friends-of-friends velocity components: subsonic + broad line consists of spaghetti + most filaments have no cores

  • summary of Lada, Lombardi led papers

  • Corona Australis cloud (Avles et al submitted) + PDF very lognormal with a tiny power-law excess + Kainulainen did the PDF work...

    • sometimes loglog gives insight...

    • mentioned my poster as rejecting lognormal (oops - that result changed since I wrote the abstract...)

  • NICEST future? + GNICER + do it to other galaxies?

    • nice extinction mapping in CNR of M83

  • Bally CMZ... + M83 analogy?

  • Q Herschel maps: larger dynamic range

  • A: VLT + Spitzer -> A_V ~ 100

  • Q: Features in PDF are clearly associated with individual objects. PDFs extremely useful...

  • A: Still throwing away second dimension. Justin Bieber has same PDF as Pipe Nebula. Hat tip to Chris Beaumont

  • Q Falgarone: Search for thinning of filaments? Change of spinning?

  • A: 1km/s/pc still...

  • Q Vazquez-Semadeni: Is B68 inside an HII region or not?

  • A: No. Herschel data shows tail of warm dust being blown away. Some external agent removed filament B68 was born in. External pressure is higher outside B68 than in, e.g., Taurus.

  • Q Zinnecker: B68. Structure more complex? Niel Boch? Not isothermal?

  • A: BE are isothermal; fit cores well a lot. BE still decent representation.

  • Q Klessen: Degeneracy distribution highly degenerate. Possible to reproduce from turbulence without contained sphere.

  • A: No change in B68. Long, subsonic things.

  • "Danger is applying one model to everything." -Klessen

  • "Shouldn't apply turbulence to everything." -Alves

  • Keto: Non-isothermal = factor of 2? Numerical models say doesn't strongly affect dynamics, still acts BE-like.

  • Vasquez-Semadeni: We are taking turbulence too far! But these are not stable either.

  Herschel view of mol cld structure & SF: Nicola Schneider

  Unscheduled talk.

  • massive stars at junctinos of filaments
  • DR21, Taurus + "striations" correlated with B-fields
  • Herschel PDF goes to higher density
  • ChamII: subtract sources, get lognormal (kinda) + "weak" slope difference in power law tails
  • compressed shells -> double-peak, broadened PDF
  • Q Joao: 2 power laws. Coincide with OB stars. Could it be unaccounted for temperature increase?
  • A:
  • Q Kainulainen Cham II: Subtracted bound cores.
  • A: A_V > 10 is core collapse
  • Q: in PDFs, are the pixels all independent, or does the PDF from a single bright source contribute to many bins?
  • A: Too many pixels

  Properties of interstellar filaments observed with Herschel and 3D magnetic field structure derived from the polarization parameters observed with Planck 15': Doris Arzoumanian

  Constant filament width

  • 0.1 pc

  B-fields in filaments

  • geometry of field can lead to depolarization
  • Q Crutcher: Polarization fraction is much higher than observed in cores. Very minor effect.
  • A: Cores would be lower.

  Magnetic Fields in Bok globules 15': Gesa Bertrang

  Supercritical: B-fields play no role

  • comparison of NIR and submm poln
  • VLT/ISAAC poln
  • Q Zinnecker: IR poln vector vs B-field. Radiative acceleration vs B-field alignment?
  • Q Keto: How do you know surrounding gas is associated with core?
  • A: We don't see anything else in the images. The globules are very isolated

  "Bok Globules" sounds very like "Buckyballs"

  Effect of turbulence on the density statistics of molecular clouds: an observational view: J Kainulainen

  Density structure dominated by turbulent motions

  • Yields a lognormal function
  • powerlaw tail from gravity
  • assume that 2D pdf can be used to yield 3D pdf
  • Can't use background stars at N kpc... + high dynamic range: 3-120x10^21 cm^-2 + 2" res
  • avoid LOS contamination by using a column cutoff
  • use 8 clouds to determine b
  • "First direct observation determination of b"
  • Dense gas mass fraction + lognormal PDF -> exponential DGMF + IRDCs have greater "fraction" of high density gas
  • COMMENT: You CAN estimate the volume density.
  • Yes, I agree, I can do it too.
  • Q: Vazquez-Semadeni. Low b suggests more solenoidal than compressive.
  • A: B-field squishes PDF.
  • A: b-parameter drives SF. B-field close second.
  • Q: How do you handle projection effects?
  • A: Not observationally, but simulations seem to show we're doing OK
  • Q Hennebelle: Equation of state. Higher adiabatic index leads to different PDF.
  • A: Should repeat experiment...
  • Q Nicola Schneider: How do you get Mach number? I get much lower Mach number. Why is there no clearly defined power-law tail?
  • A: Mach #: line width, assume temperature, -> 3D vel dispersion.
  • A: We need quantitative comparison between Herschel & extinction map. Depends on scale. We could fit with power-laws. But, looks like lognormal..
  • A: MAYBE powerlaw tails due to gravity. Maybe not! Maybe young IRDCs not dominated by gravity.

  Filamentary Structures in the ISM 15: Rowan Smith

  Arepo!

  • time dependent chemistry
  • molecular cloud factory
  • filamentary structures generated
  • filaments examined with DiSPERSE + can generate filaments with shallow profiles with or without B-fields
  • How do 2D filaments match 3D filaments?
  • DiSPERSE connects maxima: Cores are forced onto filaments!
  • Question for the audience: How do you fit a filament with a gaussian?
  • 3D reasonably consistent with 2D? + but major degeneracy between R-flat and p
  • Q Hennebelle: Why do you need B-fields for a shallow profile?
  • A: You don't need B-fields for shallow profiles.
  • (more conversation that was probably important but I missed it)
  • Q: filaments embedded in hot medium, not same as mol cloud filaments
  • Q: Xu - We found a filamentary wisp just like what you saw
  • Q: Adam Leroy: Why did you pick a particular number for the CO?
  • A: If my sensitivity to CO is a certain value, how much gas do I miss? Cumulative plots help avoid "threshold"
  • Q Adam: Would you make this back up by filling beam with faint CO?
  • A: Haven't looked at beam sizes yet.
  • Q Zinnecker: How do you get H2 at such low CO values? Such low extinction, shouldn't H2 go away?
  • A: Very well self-shielding

  Turbulence in the ISM 30': Fabian Heitsch

  Turbulent mixing -> serious resolution issues?

  • Fragmentation rather than support
  • lognormals are easy to generate

  How is turbulence driven? (how does it arise?)

  • Hydrodynamic eqns -> dispersive and curly components 1. Gradients in the velocity field (shear) 2. Angular momentum conservation term: vorticity increases when gas compressed. Make something smaller -> spin up 3. Pressure / density misaligned -> turbulence. Thermal instability.
  • "Turbulence in the ISM is a consequence"

  Turbulence decays

  • On a dynamical timescale? + lifetime can be extended somehow...
  • Drivers? + Expansion of shells + Global graviational instability
  • Vazquez-Semadeni simulation: gravitational collapse drives turbulence in a "core"

  How is turbulence driven in the models?

  • Local: Driven / fourier forcing + Choose amplitude in fourier space (Kolmogorov or Burgers) + Uniform random phase in fourier space + transform to real + apply forcing at every timestep (at constant luminosity)
  • Large scales feed smaller scales
  • Problems: + phases should be coherent (HII region shells are coherent) + driving is volume-filling! (bow shocks!) + only makes sense if accretion timescale longer than crossing time (this is OK) + periodic box: uncertain jeans mass, virial parameter.
  • Alternative: cloud formation by colliding flows.

  Support vs Fragmentation

  • assume turbulence is an extra support parameter + if true, one core should form 1 star instead of 100 (? this doesn't make sense to me) + Energy from out->in. How does this yield support? + Turbulence isotropic on small scales? Not true. Most energy on largest scales. + mildly supersonic turbulence CAN support cores... M=10 doesn't

  Turbulence doesn't work, so...?

  • turbulent support is line splitting... no hair splitting
  • turbulence REALLY leads to fragmentation
  • Q Jouni Kainulainen: Should we (observers) stop measuring CO line widths and calling it turbulent energy?
  • A: Don't stop measuring. Turbulence support can't come from gravity-driven turbulence.
  • Q: Assumed anisotropic turbulence. In B-fields, turbulence could be 2D.
  • A: Yes, but need dynamically dominant B-fields
  • Q Phillip Girichidis: Detailed description of how to model turbulence in simulations. Matters whether acceleration or force.
  • A: We know what the difference is...
  • A Vazquez-Semadeni: Force -> acceleration depends on density. Only get a lognormal when you have acceleration, not force. Force -> dense regions accelerated less -> powerlaw.

  Molecular cloud formation in converging flows 30': Patrick Hennebelle (SAp/CEA Saclay)

  Converging flows

  • convert WNM -> CNM in "converging flows" (not shocks?)
  • alternative: bistable medium..
  • polytropic EOS -> powerlaw + 2-phase is strongly non-lognormal + but they consist of two separated lognormals
  • molecular clouds are 2-phase: HI and H2 spatially coincident
  • Mass spectrum "Higher order statistics"
  • mass-size-velocity dispersion relations "consistent with larson relation" + scatter still huge
  • Accretion of HI is adequate in early stages + maybe feedback is needed in later stages?
  • Magnetization is mass-dependent?

  SF in colliding flows?

  Filaments in MHD turbulence

  • don't need converging flows to get filaments
  • filaments from intersection of shocked sheets? + HD vs MHD: filaments form in HD, but they live longer in MHD
  • filament identification: inertia matrix -> eigenvectors -> filament direction
  • MHD -> greater elongation + B-field keeps filaments more coherent + B-fields weaken shocks, therefore shocks are not the formation driver of filaments
  • Q:Burkhert Coherence length scale in galaxies
  • Vazquez-Semadeni: Colliding flows are just a representation for any style of converging flow (e.g., grav instability)
  • Jin Koda: 500pc scale cloud formation, would expect GMC to follow flow. But, we find retrograde & prograde.
  • A: Why do you expect spin? Local turbulent motions... at 500 pc, local turbulence comparable to shear.
  • Q Keto: If cloud initially forms as Jeans mass by grav frag process, so converging flows isn't really the explanation.
  • A: converging flows generated by gravitational collapse are still converging flows.
  • Q Klessen: distribution of spins arise naturally (because of curvature?)

  On the characteristic mass of stars in stellar clusters 15': Paul Clark

  Try to take low density clouds (~100) and collapse them

  • Bate 2008 showed radiation limited fragmentation
  • heating/coolring rate strong function of density + photoelectric heating dominant
  • ISM physics in AREPO
  • vary turbulent driving (b=0.3, 0.5, 1)
  • take into account shielding
  • compressive, weak/strong G0: more small-scale filaments in low-G0 field
  • solenoidal: SF takes longer, more shreddy + solenoidal vs compressive take very different amts of time to create stars
  • difference in mass function at high mass + compressive with high G0 forms more massive stars + compressive case leaves gaps: more internal heating
  • systematically offset from Chabrier IMF: deficit of low-mass stars + not forming enough low-mass stars + Why? sinks prevent binary formation

  Q: Is the solenoidal driving less mass to high density? i.e., is the density distribution still the driver of SF?

  • Q Zinnecker:
  • A: isothermal doesn't work (especially at high densities)
  • Q What about binaries?
  • A: Yes, that's it.

  Photoionization of the diffuse ionised gas in an MHD supernova-driven turbulent Interstellar Medium Jo Barnes (University of St Andrews)

  1st year phd

  • explaining [N II]/Halpha line ratios, etc.
  • scattered light + HII region -> dust -> us + diffuse gas -> dust -> us
  • scattered/total ~ 0.5 at midplane. at super high latitudes, can be 20% again
  • Q: Could low scale height be because SNe are not blown up together?
  • A: We use average SN rate, but arms should be enhanced

  The molecular richness of diffuse ISM: a tracer of turbulent dissipation 15': Edith Falgarone

  Discussing diffuse medium

  • large range in physical scale of clouds
  • 5 order of magnitude scatter in transfer rate of kinetic energy. + No trends with scale.
  • All forms of Larson relation fail at scales <0.1 pc
  • intermittency... most dissipation occurs in very small volume + dissipation occurs on filamentary structures + resolve pairs of CO-emitting regions corresponding to the dissipative high velocity shears
  • CH+ cation - known for 70 years, but poorly understood + highly endothermic formation + rapidly destroyed by H2, in ~1yr + requires extremely efficient formation rate + detected to be far too abundant
  • M82 LOS: detected in high latitude cloud! + gigantic inverse P-cygni from M82 + infalling HVC tracer?
  • also detected CH+ to another galaxy at high latitude
  • Associated with HI (same line shape)
  • highly non-equilibrium chemistry
  • "TDR" = Turbulence Dissipation Region models

  Q: CH+ seen drastically enhanced in diffuse clouds. Should we expect gigantic abundance variations in dense molecular clouds too?

  • Q CH+ + B-fields?
  • Q Frequency of events?
  • A: 1% of gas

---

• Late Night Ideas, Oct 6 2013

  Coffee at 7 pm usually isn't a problem...

  1. Dendrogram analysis of turbulent simulations. What do the structures look like? Mostly for visualization in a talk, but perhaps the start of a paper. Are the high-density substructures too small to collapse? Too hot?
  2. A deep GBT+AO observation of the center of the G49 40km/s cloud. Can we perform the same measurement as I did originally against a uniform backlight? Will that provide new information
  3. Corollary to 2: In simulations, what effect does a varying size and brightness backlight have on the inferred density?
  4. Measure the distance to the G49 cloud with NIR extinction. Keep working with Jonathan Foster on that...
  5. Test the Cepheid / RR Lyrae distance method on the G49 cloud. It's close enough that the targets should be bright. Start off with 1-hour pilot (DD?) on the most distant star with a "guaranteed" CO detection, then propose a longer project to begin using the bisector approach to determine its true distance.
  6. Eta Car ALMA. Basically, same thing, but now invoke the SN 1987A CO detection. Eta is the closer laboratory, and it must have similar shock physics...
  7. H3+ in the CMZ with Juergen. Talk to Tom Geballe about H3+ and CO absorption. Is it possible to measure CO absorption to red giants?

---

• Installing MultiNest on Mac OS X 10.7

  Installing multinest & pymultinest has proved non-trivial.

  First, you probably can't build using the default mac gcc; a gcc version >4.5 is probably needed. The error I got installing MultiNest only turned up one useful result on google... it was a question I had asked on the hyperion github page.

  To get around that, I installed the hpc compilers (4.7 since I'm on OS X 10.7) into /usr/local/hpc/ to keep them distinct from the system gcc, which is needed for many other things.

  I then installed the openMPI compilers to /usr/local/openmpi:

  ./configure --prefix=/usr/local/openmpi
  make
  sudo make install
  

  Then I tried installing pymultinest

  dor ~/repos/MultiNest cmake/master$ mkdir build
  dor ~/repos/MultiNest cmake/master$ cd build/
  dor ~/repos/MultiNest/build cmake/master$ which mpif90
  /usr/local/openmpi/bin/mpif90
  dor ~/repos/MultiNest/build cmake/master$ which gfortran
  /usr/local/hpc/bin//gfortran
  dor ~/repos/MultiNest/build cmake/master$ which gcc
  /usr/local/hpc/bin//gcc
  dor ~/repos/MultiNest/build cmake/master$ cmake -DCMAKE_{C,CXX}_FLAGS="-arch x86_64" -DCMAKE_Fortran_FLAGS="-m64" ..
  -- The Fortran compiler identification is GNU
  -- The C compiler identification is GNU 4.7.1
  -- The CXX compiler identification is GNU 4.7.1
  -- Check for working Fortran compiler: /usr/local/hpc/bin/gfortran
  -- Check for working Fortran compiler: /usr/local/hpc/bin/gfortran  -- works
  -- Detecting Fortran compiler ABI info
  -- Detecting Fortran compiler ABI info - done
  -- Checking whether /usr/local/hpc/bin/gfortran supports Fortran 90
  -- Checking whether /usr/local/hpc/bin/gfortran supports Fortran 90 -- yes
  -- Checking whether C compiler has -isysroot
  -- Checking whether C compiler has -isysroot - yes
  -- Checking whether C compiler supports OSX deployment target flag
  -- Checking whether C compiler supports OSX deployment target flag - yes
  -- Check for working C compiler: /usr/local/hpc/bin/gcc
  -- Check for working C compiler: /usr/local/hpc/bin/gcc -- works
  -- Detecting C compiler ABI info
  -- Detecting C compiler ABI info - done
  -- Checking whether CXX compiler has -isysroot
  -- Checking whether CXX compiler has -isysroot - yes
  -- Checking whether CXX compiler supports OSX deployment target flag
  -- Checking whether CXX compiler supports OSX deployment target flag - yes
  -- Check for working CXX compiler: /usr/local/hpc/bin/g++
  -- Check for working CXX compiler: /usr/local/hpc/bin/g++ -- works
  -- Detecting CXX compiler ABI info
  -- Detecting CXX compiler ABI info - done
  -- Looking for Fortran dgemm
  -- Looking for Fortran dgemm - found
  -- Looking for include file pthread.h
  -- Looking for include file pthread.h - found
  -- Looking for pthread_create
  -- Looking for pthread_create - found
  -- Found Threads: TRUE
  -- A library with BLAS API found.
  -- Looking for Fortran cheev
  -- Looking for Fortran cheev - found
  -- A library with LAPACK API found.
  -- Detected gfortran, adding -ffree-line-length-none compiler option.
  -- Found MPI_C: /usr/local/openmpi/lib/libmpi.dylib;/usr/lib/libm.dylib
  -- Found MPI_CXX: /usr/local/openmpi/lib/libmpi_cxx.dylib;/usr/local/openmpi/lib/libmpi.dylib;/usr/lib/libm.dylib
  -- Found MPI_Fortran: /usr/local/openmpi/lib/libmpi_f90.a;/usr/local/openmpi/lib/libmpi_f77.dylib;/usr/local/openmpi/lib/libmpi.dylib;/usr/lib/libm.dylib
  -- Configuring done
  -- Generating done
  -- Build files have been written to: /Users/adam/repos/MultiNest/build
  dor ~/repos/MultiNest/build cmake/master$ make
  Scanning dependencies of target multinest_mpi_shared
  [  1%] Building Fortran object src/CMakeFiles/multinest_mpi_shared.dir/utils.f90.o
  [  3%] Building Fortran object src/CMakeFiles/multinest_mpi_shared.dir/utils1.f90.o
  [  5%] Building Fortran object src/CMakeFiles/multinest_mpi_shared.dir/kmeans_clstr.f90.o
  [  6%] Building Fortran object src/CMakeFiles/multinest_mpi_shared.dir/xmeans_clstr.f90.o
  [  8%] Building Fortran object src/CMakeFiles/multinest_mpi_shared.dir/posterior.F90.o
  [ 10%] Building Fortran object src/CMakeFiles/multinest_mpi_shared.dir/priors.f90.o
  [ 12%] Building Fortran object src/CMakeFiles/multinest_mpi_shared.dir/nested.F90.o
  [ 13%] Building Fortran object src/CMakeFiles/multinest_mpi_shared.dir/cwrapper.f90.o
  Linking Fortran shared library ../../lib/libmultinest_mpi.dylib
  [ 13%] Built target multinest_mpi_shared
  Scanning dependencies of target multinest_mpi_static
  [ 15%] Building Fortran object src/CMakeFiles/multinest_mpi_static.dir/utils.f90.o
  [ 17%] Building Fortran object src/CMakeFiles/multinest_mpi_static.dir/utils1.f90.o
  [ 18%] Building Fortran object src/CMakeFiles/multinest_mpi_static.dir/kmeans_clstr.f90.o
  [ 20%] Building Fortran object src/CMakeFiles/multinest_mpi_static.dir/xmeans_clstr.f90.o
  [ 22%] Building Fortran object src/CMakeFiles/multinest_mpi_static.dir/posterior.F90.o
  [ 24%] Building Fortran object src/CMakeFiles/multinest_mpi_static.dir/priors.f90.o
  [ 25%] Building Fortran object src/CMakeFiles/multinest_mpi_static.dir/nested.F90.o
  [ 27%] Building Fortran object src/CMakeFiles/multinest_mpi_static.dir/cwrapper.f90.o
  Linking Fortran static library ../../lib/libmultinest_mpi.a
  [ 27%] Built target multinest_mpi_static
  Scanning dependencies of target multinest_shared
  [ 29%] Building Fortran object src/CMakeFiles/multinest_shared.dir/utils.f90.o
  [ 31%] Building Fortran object src/CMakeFiles/multinest_shared.dir/utils1.f90.o
  [ 32%] Building Fortran object src/CMakeFiles/multinest_shared.dir/kmeans_clstr.f90.o
  [ 34%] Building Fortran object src/CMakeFiles/multinest_shared.dir/xmeans_clstr.f90.o
  [ 36%] Building Fortran object src/CMakeFiles/multinest_shared.dir/posterior.F90.o
  [ 37%] Building Fortran object src/CMakeFiles/multinest_shared.dir/priors.f90.o
  [ 39%] Building Fortran object src/CMakeFiles/multinest_shared.dir/nested.F90.o
  [ 41%] Building Fortran object src/CMakeFiles/multinest_shared.dir/cwrapper.f90.o
  Linking Fortran shared library ../../lib/libmultinest.dylib
  [ 41%] Built target multinest_shared
  Scanning dependencies of target multinest_static
  [ 43%] Building Fortran object src/CMakeFiles/multinest_static.dir/utils.f90.o
  [ 44%] Building Fortran object src/CMakeFiles/multinest_static.dir/utils1.f90.o
  [ 46%] Building Fortran object src/CMakeFiles/multinest_static.dir/kmeans_clstr.f90.o
  [ 48%] Building Fortran object src/CMakeFiles/multinest_static.dir/xmeans_clstr.f90.o
  [ 50%] Building Fortran object src/CMakeFiles/multinest_static.dir/posterior.F90.o
  [ 51%] Building Fortran object src/CMakeFiles/multinest_static.dir/priors.f90.o
  [ 53%] Building Fortran object src/CMakeFiles/multinest_static.dir/nested.F90.o
  [ 55%] Building Fortran object src/CMakeFiles/multinest_static.dir/cwrapper.f90.o
  Linking Fortran static library ../../lib/libmultinest.a
  Error copying file "/Users/adam/repos/MultiNest/build/src/kmeans_clstr.mod" to "/Users/adam/repos/MultiNest/modules/".
  make[2]: *** [../lib/libmultinest.a] Error 1
  make[1]: *** [src/CMakeFiles/multinest_static.dir/all] Error 2
  make: *** [all] Error 2
  

  The approach that worked:

  mkdir -p build lib modules bin/chains
  cd build
  cmake -DCMAKE_{C,CXX}_FLAGS="-arch x86_64" -DCMAKE_Fortran_FLAGS="-m64" ..
  make
  sudo make install
  cd ../lib
  ln -s libmultinest.dylib libmultinest.so
  export DYLD_LIBRARY_PATH=$DYLD_LIBRARY_PATH:/usr/local/hpc/lib:/usr/local/openmpi/lib/
  export LD_LIBRARY_PATH=/Users/adam/repos/MultiNest/lib
  

  At this point, multinest imports but doesn't run the tests...

---

• Tuesday Afternoon Talks

  4:15-4:35 pm Tobias Fritz (Germany): Nuclear Cluster of the Milky Way (15 + 5 minutes)

  

  Model fitting of nuclear cluster: Sersic, etc.

  • unusually bright nuclear cluster for its size

  Globular cluster origin of nuclear cluster

  Questions:

  • Q: Round component consistent with inspiraling globular cluster. What about spectral properties?
  • A: No. Metallicity difficult.

  4:35-4:55 pm Tuan Do (Canada): Measuring the physical properties of the Milky Way Nuclear Star Cluster with 3D stellar kinematics (15 + 5 minutes)

  

  Dynamics of CMZ stars

  • Massive black hole affects dynamics of a cluster
  • testing presence of cusps + old stars have no cusp + young stars have a cusp, but not dynamically relaxed
  • what is the true stellar density profile? + can't determine by counting + kinematics can help + old stars randomly distributed + young stars concentrated
  • did the stats right + no binning + individual likelihoods + unfortunately, this is new
  • flat slope, no anisotropy
  • correlation between parameters, particularly the slope and BH mass
  • edge of core poorly constrained
  • infalling IMBH or SMBH destroys stellar cusp + observational constraints on IMBHs poor
  • QUESTION: Do the fits improve at all if you impose a GC distance? + Answered: Can merge with PDF from stellar orbits

  Question

  • Q: Inconsistent with 7.7 kpc distance to CMZ?
  • A: no, consistent within 1-sigma
  • Q: How do you contrain R_0 with this approach?
  • A: radial velocities give absolute scale. Proper motions must also have some scale. Match the scales to get R_0
  • Q: Don't use any information outside 0.5 parsecs. Shouldn't you include it? I get the right mass if I do
  • A: Will help, but velocity dispersion dominated by nearest stars
  • Q: Anisotropy in cluster. Get rid of red supergiants somehow? Tangential or radial?
  • A: Hard to know. Very little anisotropy.

  4:55-5:15 pm Mark Wardle (Australia): Star Formation within 0.5 pc of SgrA* (15 + 5 minutes)

  

  "Highest I've ever been for a talk"

  High densities needed to overcome tidal shear

  • big cloud -> Hoyle-Littleton accretion + self-intersecting orbit on the backside of the cloud
  • central hole because accreted?
  • high magnetic pressure in disk: 10x greater than gas
  • B-fields suppress fragmentation
  • Black hole takes bite out of cloud
  • optically thick disk: sigma T_eff^4 cooling + heating due to starlight, accretion
  • twisting knobs to figure out where accretion disk turns into fragmenting disk + inside, magnetically active / supported + outside (0.04 pc) fragmenting + further out: grav stable
  • maybe high accretion rate drove fermi bubble
  • accretion rate consistent with 10^6 Msun/Gyr -> build Sgr A*

  Questions

  • Q Mitch: Grav unstable region globally unstable?
  • A: Has to do with the disk height. h/r very small, not an issue
  • Q Fred Lo: We don't see any of this observationally. All history?
  • A: Yes, this all happened a few Myr ago.
  • Q: Eddington accretion: what happens to rest of star formation and the disk?
  • A: I think it's OK - star formation 10^4-10^5 years, but dump time 10^5-10^6 yrs
  • Q: Bonnel & Rice did simulations. Why need more?
  • A: B-fields? Eqn of state? Radiative cooling? Crazy enough initial conditions?
  • A Farhad: didn't include radial distribution of stars
  • Q: How do you start with 0-angular-momentum cloud?
  • A: Lots of junk around, maybe collisions do it. I don't have a good answer. Depends on how big a fluffy cloud you use.
  • Q Fred Lo: What observational signature remains?
  • A: Maybe kinematics

  5:15-5:35 pm Andrea Ghez (USA): Probing General Relativity with Short Period Stars at the Galactic Center (15 + 5 minutes)

  Why continue studies of GC stars?

  • Test GR

  • Role of BH in galaxies

  • Depth of potential 100x greater, on mass scale 10^6 larger than other tests

  • simplest tests + Relativistic redshift (easy) + Precession of periapse

  • SO-2 "star of the show" + next close approach in 2018.5 + relativistic signal ~200 km/s + need very accurate keplerian orbit first + radial velocity from BrG

  • absolute reference frame is tough + masers give the reference frame, but require larger field

  • full orbital coverage with astrometry + 13-parameter model

  • Systematic drift in reference frame + There is a major source of systematic error + Spatial variation in PSF (AO issue) is the worst systematic + drift term prevents bias + 5 km/s uncertainty projected by 2018

    • 5-sigma GR detection

  • GEMS to make reference frame

  • TMT could do 2-yr orbits

  Questions

  • Q Ostriker: Could the SMBH be sloshing? Can you allow for that? What about acceleration?
  • A: Mark Reid did calculation. Limit ~3 km/s for present BH velocity. Problem: assume linear velocities, but that is not always a safe assumption. If BH accelerating, would create "nonphysical" accelerations.
  • Q: Many things spiral into center, will lead to oscillations in BH motion
  • Q Stocke: How big an effect is blending with unseen stars?
  • A: We simulate it. But, trying to simulate something we don't understand. <0.1 mas
  • Q: Contributions from resonant relaxation, encounters with other stars.
  • A: Skipped the slide that shows this. For GR, insignificant, but for precession of periapse, have to worry.
  • Q: What about the velocity component?
  • A: small effect on velocity.

  Discussion Section

  • Q: IMFs shallower than salpeter.
  • A Lu: Hard to decouple from dynamical history. push on spectroscopy to lower masses. Decouple young and old populations
  • A Fritz: instrument sensitivity limit... VLT can detect CO bandheads...?
  • How well do we know it's 2 populations, or maybe it's just 1 population? Assumptions disagree, not measurements
  • Stocke: Background is changing in center. Can you use fluctuation analysis to determine brightness of lower mass stars?
  • Lu: Very challenging for young stars because they're a small fraction of the total luminosity. Foal(?) et al did good work on older star populations
  • Arches cluster: Hui Dong found 4 runaways of same mass as the 12 stars in the Arches cluster. Maybe as many runaways as cluster stars? Half of cluster stars kicked out. Do dynamical models predict this?
  • [silence]
  • Find massive stars in "other 3/4" (I'm lost...). Need spectroscopy
  • Ghez: Role of binary stars in ejections.
  • Missing half of cluster mass?
  • Ghez: Role of interaction with black hole? Clusters not evolving in isolation
  • Arches 2.5 Myr old. Largest stars could just start... evolving?
  • Most difficult objects to explain are LBVs. LBVs outside of cores of clusters. How do you get very massive binaries out of cores? All LBVs found are outside cores.
  • Role of binaries. New evolutionary models. LBVs that must be <2 Myr, WCs must be >4 Myr. Contradiction?
  • New evolutionary models say many stars come from rejuvenation. Massive blue stragglers.
  • In situ formation? What fraction should be expected to be formed in situ?
  • How confident are we about mass-luminosity relation for stars?
  • Lu: On main sequence, confident, but post-main-sequence is uncertain. We drop them in our analysis. Pre-main sequence is a problem in Arches. Arches fitted with only main-sequence stars
  • very difficult with rotation included
  • Distance to the Galactic center "decreasing at an alarming rate". Does determination conflict with any other indicators? What does that mean?
  • Ghez: Mark Reid likes lower R_0. All consistent within uncertainties
  • Do simulations for S stars include interactions between stars?
  • Ghez: Not in most recent version, but in an earlier version
  • How does this test of GR compare with other tests of GR around Sgr A*?
  • Ghez: Event Horizon Telescope, GRAVITY.
  • Different systematics.
  • Are HST data consistent with BH at center?
  • Tuan Do: Same reference frame. BH defined to be at center. Velocity can drift, but not position

---

Page 1 / 6 »