Notes from MCP Meeting - 23-24 April 2012

In attendance: Jim Braatz, Mark Reid, Jim Condon, Fred Lo, Feng Gao, Wei Zhao


  • Mark reviewed Bayesian fitting formalism

Bayes fitting

  • How does VLBI sensitivity affect H0 measurement?
    • One good VLBI track (GBT+VLBA+EB) gives reliable maser spots to ~ 10 mJy ( This is a cutoff of ~ 7-sigma detections)
    • Most of our previously observed galaxies have essentially ~4 good tracks: get maser spots to ~ 5 mJy
    • With JVLA we expect ~ 50% improvement in VLBI sensitivity so with 4 good tracks we get maser spots to ~ 3 mJy
    • Critical plot for Aug 1 proposal: Show a spectrum with a horizontal line at ~ 10 mJy and ~ 3 mJy to demonstrate how many more lines we can get, esp, at systemic velocities, with the improved sensitivity.
    • Need multiple tracks closely-spaced in time to avoid ambiguities in averaging widely-spaced (in time) data.
    • Note 3 mJy noise improves position accuracy of strong lines and also allows us to detect more lines, and in an ideal case this extends the velocity range mapped (more important)
    • However, it may be ok to average data widely spaced in time if the disk geometry is appropriate, i.e. masers at a given velocity always have approx the same position on the sky and same accel, even when observed over many years. Some evidence that this is the case for U3789 (all 3 VLBI maps look the same)
    • How does 3 mJy sensitivity affect Bayes H0 result?

  • Next, we need to show how this improved sensitivity and more velocity coverage of VLBI-mapped systemic lines improves the H0 measurement. Best ways to approach this question are:
    • Starting with UGC 3789 results, degrade the data to simulate the same galaxy ~ 2X farther away. Refit the degraded data. How much worse is it?
    • Generate fake data for a simulated disk, complete with accelerating systemic features, and fit with Bayes program. Critical to make fake spectra modeled on a real galaxy, and it must have systemic maser features that extend the mappable velocity range, with 10 mJy < flux < 3 mJy

  • Can we use HI recession velocity as a prior on Bayes fitting?
    • Jim B has GBT data on a few of the maser disk galaxies -- he will circulate sample spectra
    • Need JVLA data to get recession velocities accurate enough. Is this feasible?

  • Can we detect the continuum, and if so, can we use the position of the continuum to constrain the position of the dynamical center?

  • How much difference in H0 would it make to provide a more accurate v0 prior?
    • Try the disk fitting program with artificially tight prior on v0.

Accel fitting

  • How do we assess the need for better acceleration data or different cadence?
    • Average 3 epochs into 1 using a rolling boxcar average
      • NGC 6323 and J0437 are good candidates for this experiment; Feng and Jim will work on this.
    • Test importance of cadence by dropping every other epoch from one of our good cases, and refit the accels.
    • Need to write software that gives us a better look at results of the accel fitting program, along the lines of Jim B's tools for plotting fitted accels over "old style" plots

  • Modifications requested to the new acceleration fitting program:
    • print residuals for each solution
    • try to fit a single acceleration for a given velocity range
    • try to fit a parametrized acceleration a(v) = a0+k(v-v0)

Disk Physics

  • Do we understand the physics of the disk and mechanism for making masers?
    • Fred will investigate some of these questions during his sabbatical
    • Why are systemic and h.v. lines comparable in intensity?
      • Evidence for saturation?
    • Alignment of maser clouds leads to visible maser spots?
    • What happens to maser line width during a flare?
      • Several good cases to investigate: NGC 1194; Mrk 1419. Good project for Eugenia.

Aug 1 proposal planning

  • Discussion of priorities for VLBI
    • Three periods to consider
      1. Fall 2012 (GBT+EB+VLBA)
      2. Winter 2013 (JVLA+GBT+EB+VLBA)
      3. Fall 2013 (GBT+EB+VLBA)

  • We mainly discussed period 2 (with JVLA) and we decided on priorities:
    • (1a) NGC 5765 : 8 tracks x 6 hours
    • (1b) J0437+2456 : 8 tracks x 6 hours
    • (2) ESO558-G009 : 8 tracks x 6 hours

  • We discussed other galaxies but decided they were lower priority.
    • Mrk 1419
    • NGC 6323
    • NGC 6264
      • Need status update from Cheng-Yu
    • UGC 3789
    • IC 2560
    • NGC 1194
    • NGC 2273
    • UGC 6093

  • What about Fall 2012 and Fall 2013?
    • Get lots of tracks on N5765.
    • Get more BH candidates?

Improving our VLBI methods

  • Dependence of VLBI imaging on absolute maser position for self-cal tracks
    • This is calculable from first principles - we think the position offset of a maser spot relates either to absolute freq difference or "video" offset
    • Easy way to test this is to insert a position offset using CLCOR on a maser with good position, then reimage
    • VLA positions may not be sufficient for galaxies that we intend to observe for distance measurement

  • Discussion of the need for geodetic blocks in our VLBI observations
    • Mark says we don't need these for self-cal tracks; We do still need them for phase-ref tracks
    • Saves ~ 2h per track for self-cal tracks. Can get clock error and residual delay from hourly observations of the delay calibrator (needs to be <~ 10 degrees from target and >~ 500 mJy continuum source). We already observe these in our tracks.
    • Note, required structure of self-cal VLBI tracks is then:
 - Fringe finder/bandpass calibrator (very strong, northern target, need not be close to target)
 - Loop (1 hr) : Delay cal (~ 5 mins) then Target (~ 55 mins broken into two ~ 30 min segments)
 - Fringe Finder/bandpass cal
 - Delay cal

-- JimBraatz - 2012-04-25
Topic revision: r3 - 2012-04-26, JimBraatz
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding NRAO Public Wiki? Send feedback