using quasar physics to improve the vlbi reference frame · −300 −200 −100 0 100 200 300...

Using quasar physics to

improve the VLBI reference frame

with: Lucia Plank, Jamie McCallum, Rob Schaap (UTAS) Johannes Böhm, Hana Krásná (TU Wien) Jing Sun (Shanghai Astronomical Observatory)

Stas Shabala University of Tasmania

Outline

①  What are these radio sources we look at ?

②  Quantifying the effects of quasar structure

③  VieVS structure simulator §  Simulation strategy §  Effect on the reference frame (Station positions : mm level)

④  Mitigation strategies Stas Shabala -‐ REFAG14

Uncooperative quasars What you want them to be ² Bright point sources ² Fixed in space and Sme

Image: N. Reid

What they are ² Supermassive black holes ² Jets è structure ² Evolve on human Smescales

Image: NASA

Image: C. Mueller

Lister et al. (2009) Stas Shabala -‐ REFAG14

Inner regions of an Active Galactic Nucleus

Stas Shabala -‐ REFAG14

Image: BU / A. Marscher

core jet jet origin

One jet towards us (Doppler boosted) -‐> seen One jet away from us (D. deboosted) -‐> unseen

1 mas @ z=1

uv plane

separation

²  interferometers measure correlated amplitude and phase

u (Mh)v

(Mh)

−300 −200 −100 0 100 200 300−300

−200

−100

0

100

200

300

−10

−8

−6

−4

−2

0

2

4

6

8

10

distances measured in units of wavelength λ=c/ν

N-E plane projected in source direction

Visibility phase

uv plane

separation

²  interferometers measure correlated amplitude and phase ²  image çFourierTransformè

amplitudes/phases in uv plane

u (Mh)

v (M

h)

−300 −200 −100 0 100 200 300−300

−200

−100

0

100

200

300

−10

−8

−6

−4

−2

0

2

4

6

8

10

N-E plane projected in source direction

Lister et al. (2009)

Simulated source

Right Ascension (mas)

Dec

linat

ion

(mas

)

−2 −1 0 1 2−2

−1.5

−1

−0.5

0

0.5

1

1.5

2

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1


Visibility phase

u (Mh)

v (M

h)

−300 −200 −100 0 100 200 300−300

−200

−100

0

100

200

300

−10

−8

−6

−4

−2

0

2

4

6

8

10N-E plane projected in source direction


different location in uv plane depending on frequency and baseline

Source structure in geodetic VLBI ² Group delay = (1 / 2π) (Δ phase / Δ frequency)

ª  phase depends on projected structure as seen by a given baseline

ª  function of: •  baseline •  observing time •  amount and direction of structure

ª effect is different at each of 8 sub-bands at X-band (because frequencies are slightly different)

² Hence group delay (= slope across band) changes


Visibility phase

u (Mh)

v (M

h)

−300 −200 −100 0 100 200 300−300

−200

−100

0

100

200

300

−10

−8

−6

−4

−2

0

2

4

6

8



different location in uv plane depending on frequency and baseline

Structure phase on 9000 km baseline

8200 8300 8400 8500 8600 8700 8800 8900 9000−8

−7

−6

−5

−4

−3

−2

Frequency (MHz)

Phas

e (d

egre

es)

slope = structure delay


Visibility phase

u (Mh)

v (M

h)

−300 −200 −100 0 100 200 300−300

−200

−100

0

100

200

300

−10

−8

−6

−4

−2

0

2

4

6

8



different location in uv plane depending on frequency, baseline and time

-0.04

-0.02

0

0.02

0.04

8.2 8.4 8.6 8.8

phas

e / r

adia

ns

frequency / GHz

structure delay = 11 ps(3.3 mm light travel time)

structure delay = 6.9 ps(2.1 mm light travel time)

jet orthogonal to baseline

Hobart - HartRAOHobart - Katherine

Jet – baseline orientation

How a source is observed is important

slope = extra (structure) delay


1.  Measure incorrect posiSon •  PosiSon offset (X, Y, Z) from “correct” value

2.  Measure different posiSons for different schedules

•  Scaier in staSon posiSons for different baseline/schedule combinaSons (even with same quasars)

è increased rms for posiSons derived from different schedules

3.  MulSple observaSons inconsistent with each other

•  Larger formal uncertainSes, within a single session

3 effects of quasar structure

slope changes with projected structure

slope ≠ 0


Vienna VLBI Solware (VieVS) •  Simulate geodeSc observaSons •  Process simulated observaSons è staSon / source posiSons, EOPs

Quasar structure simulaSons •  Quasars ≠ point-‐like •  Extra structure delay per observaSon •  (mostly) Mock source catalogues

VieVS source structure simulator

VIEVS


Simulated catalogues

² Structure indices è mock quasar images SI = 1 + 2 log (τ / ps) u  Choose SI (none, 1, 2, 3, 4, ICRF2 distribution) u  2-component sources u  Also one “real” CONT11 catalogue

² Simulate realistic schedules with VieVS u  CONT11

o  15 – 29 September 2011 o  13 stations o  30 realizations of each day

u  Additional delay term due to source structure


SI 3 = 10 ps

CONT11


-0.3

-0.2

-0.1

0

0.1

0.2

0.3

Bd Hh Hb Zc Wf Wz On Ts Ny Ys Kk Tc Ft

med

ian

X po

sitio

n of

fset

/ cm

real structureno structure

Real sources – X coord

Stas Shabala -‐ REFAG14 Shabala+14, J. Geod. submitted

-0.3

-0.2

-0.1

0

0.1

0.2

0.3


med

ian

Y po

sitio

n of

fset

/ cm


Real sources – Y coord


-0.3

-0.2

-0.1

0

0.1

0.2

0.3


med

ian

Z po

sitio

n of

fset

/ cm


Real sources – Z coord


0

0.1

0.2

0.3

0.4

0.5


med

ian

posi

tion

offs

et /

cmreal structureno structure

Real sources – 3D coord offset


0

0.1

0.2

0.3

0.4

0.5


med

ian

posi

tion

offs

et /

cmreal structureno structure

Offset > 2 mm

Real sources – 3D coord offset


0

0.05

0.1

0.15

0.2

0.25

0.3

none 1 2 3 4ICRF2 real

posi

tion

/ cm

structure index

median offsetrms

uncertainty

How important is quasar structure ?

SI = 1 + 2 log (τ / ps) Stas Shabala -‐ REFAG14 Shabala+14, J. Geod. submitted

0

0.05

0.1

0.15

0.2

0.25

0.3


posi

tion

/ cm

structure index

median offsetrms

uncertainty



Higher SIs give worse positions

0

0.05

0.1

0.15

0.2

0.25

0.3


posi

tion

/ cm

structure index

median offsetrms

uncertaintytheory



0

0.05

0.1

0.15

0.2

0.25

0.3


posi

tion

/ cm

structure index

median offsetrms

uncertainty


SI = 1 + 2 log (τ / ps)

0.5 – 1 mm error


0

0.05

0.1

0.15

0.2

0.25

0.3


posi

tion

/ cm

structure index

median offsetrms

uncertainty


SI = 1 + 2 log (τ / ps)

Rms and σ underes1mate how wrong the soluSon really is


0

0.01

0.02

0.03

none 1 2 3 4ICRF2 real 0

0.001

0.002

pol r

ms

/ mas

(UT1

-UTC

) rm

s / m

s

structure index

xpolypol(UT1-UTC)

EOP rms


What happens if we include troposphere ?


0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

none 1 2 3 4 ICRF2 real

posi

tion

/ cm

structure index

median offsetrms

uncertainty

Station positions (with trop.)

SI = 1 + 2 log (τ / ps)

Troposphere dominates


0

0.02

0.04

0.06

0.08

0.1

0.12


0.001

0.002

0.003

0.004

0.005

0.006

0.007

0.008

pol r

ms

/ mas

(UT1

-UTC

) rm

s / m

s

structure index

xpolypol(UT1-UTC)

EOP rms (with troposphere)


Source structure : effects on CRF

SIMULATED

Simulation (Cont11), with troposphere & measurement noise

NO STRUCTURE STRUCTURE, SI=3



SIMULATED

Simulation (Cont11), with troposphere & measurement noise

NO STRUCTURE STRUCTURE, SI=3

MEDIAN ESTIMATED POSITION

JET DIRECTION

Systematic displacement ~ 100 µas

2.5

3.0

3.5

Structure Index

log

eµ

as

(0,1.5

)

(1.5,1

.75)

(1.75

,2)

(2,2.5

)

(2.5,3

)

(3,3.5

)

(3.5,4

)

(4,4.5

)

(4.5,5

)(5,

6)

less stable

posiS

ons

more structure


OBSERVED

More structure = worse source posiSons

Schaap et al. 2013, MNRAS, 434, 585

Problem summary ²  Quasar structure simulations with VieVS

²  mock + real quasar catalogues

²  Station positions u  offset from true values u  larger formal uncertainties u  increased rms u  important at mm level

²  EOPs and source positions also get worse with increasing structure


Good news ①  We can image quasar structure

(and make corrections)

How do we improve the Reference Frames ? (with quasar physics)

Stas Shabala -‐ REFAG14 Lister et al. (2009)


(and make corrections)

Bad news Quasars evolve (need to to this often)


Stas Shabala -‐ REFAG14 Lister et al. (2009)


②  Jet direction remains constant (avoid unfavourable baseline – jet orientation)

Bad news Quasars evolve




②  Jet direction remains constant

③  Quasars evolve ! (observe quasars when they are “well behaved”)


✗ ✓




②  Jet direction remains constant

③  Quasars evolve !

Strategies ①  Use real sources to correct observations

②  Minimize projected structure through clever scheduling

③  Include / exclude sources at appropriate times



baseline length [103 km]

1. Quasar structure corrections

Impr

ovem

ent

Raw observaSons rms

Corrected using available images

OBSERVED

CONT11 OBSERVATIONS



difference in mm !


Impr

ovem

ent

Raw observaSons rms

Corrected using available images

Improvement = raw -‐ corrected

OBSERVED



difference in mm !


Impr

ovem

ent

OBSERVED

Problems •  Dominated by

troposphere •  Many quasars don‘t

have images •  Quasars evolve Future •  Imaging from

geodeSc VLBI data

Small improvement •  45 bl beier: +0.22 mm 33 bl worse: -‐0.12 mm < 1 mm ... but promising...



difference in mm !


Impr

ovem

ent

OBSERVED

Problems •  Dominated by

troposphere •  Many quasars don‘t

have images •  Quasars evolve Future •  Imaging from

geodeSc VLBI data

Small improvement •  45 bl beier: +0.22 mm 33 bl worse: -‐0.12 mm < 1 mm ... but promising... Long baselines most improved


•  Jet direcSon is constant •  Worst case: parallel, long baselines

Stru

ctur

e de

lay

in p

s

Angle + length of baseline

Stru

ctur

e de

lay

in p

s

Angle between jet & baseline [°]

R1/R4 of Oct 2013, SI=3

%... median value & % of observations

2. Re-analysis using source structure


•  Jet direcSon is constant •  Worst case: perpendicular, long baselines

Stru

ctur

e de

lay

in p

s

Angle + length of baseline

Stru

ctur

e de

lay

in p

s

Angle between jet & baseline [°]

R1/R4 of Oct 2013, SI=3

%... median value & % of observations


Half the data is ok

¼ of data downweighted


Stru

ctur

e de

lay

in p

s

Angle + length of baseline Half the data is ok

¼ of data downweighted

Simula1on test •  Downweight 23% of

observaSons •  Structure only simulaSons Improvement •  StaSon posiSons by 22% •  Formal uncertainSes by 30%

Problem Need these observaSons to resolve troposphere !


SIMULATED Stas Shabala -‐ REFAG14

Scheduling •  Complex task •  OpSmizing for max. number of scans,

sky coverage, etc. Strategy •  Schedule w.r.t. relaSve orientaSon

and baseline length •  Simulate

First results appear promising

2. Re-scheduling w.r.t source structure

SIMULATED

R1/R4 (re-scheduled) of Oct 2013

Impr

ovem

ent


2

2.5

3

3.5

1995 2000 2005 0

0.5

1

1.5

stru

ctur

e in

dex

flux

dens

ity /

Jy

Year

SIX-band

3. Quasar evolution ! NEW ! •  Structure anS-‐correlates with flux

density è Observe sources when flux density is high (and so structure is small)

Source 1357+769


3. Quasar evolution

2

2.5

3

3.5

1995 2000 2005 0

0.5

1

1.5

stru

ctur

e in

dex

flux

dens

ity /

Jy

Year

SIX-band

! NEW ! •  Structure anS-‐correlates with flux

density è Observe sources when flux density is high (and so structure is small) •  PosiSon scaier decreases for sources

with high flux density

−2 −1 0 1 2

−2−1

01

2

RA Deviation (mas)

Dec

Dev

iatio

n (m

as)

−2 −1 0 1 2

−2−1

01

2

RA Deviation (mas)

Dec

Dev

iatio

n (m

as)

−2 −1 0 1 2

−2−1

01

2RA Deviation (mas)

Dec

Dev

iatio

n (m

as)Low flux density Medium flux density High flux density

Source 1357+769

Summary q  Quasars are not point sources

o  extra group delay ²  Baseline-, time-, frequency- dependent

o  systematic error è simply more observations won’t help

q  Quasar structure simulations with VieVS 2.2 ²  new source structure simulator module ²  mock + real quasar catalogues ²  station positions affected at mm level ²  troposphere dominates (for now)

u  Mitigation strategies v  corrections using source images in analysis v  not scheduling unfavourable jet / baseline combinations

§  avoid long baselines parallel to jet v  source selection

§  radio sources vary in structure on year timescales §  more compact when flaring (bright)

Inner regions of an Active Galactic Nucleus


Image: BU / A. Marscher

Image: Kovalev et al. (2008)

core jet jet origin

Synchrotron self-‐absorpSon è core locaSon depends on frequency è “Core ShiN”

One jet towards us (Doppler boosted) -‐> seen One jet away from us (D. deboosted) -‐> unseen

1 mas @ z=1

2 3 4 5 6 7 8 9 10 11 12

phas

e (a

rbita

ry u

nits

)

frequency / GHz

total

ionosphere

non-dispersive

Quasar structure in VGOS

How do we connect phases across 10 GHz?

-0.04

-0.02

0

0.02

0.04

2 4 6 8 10 12

phas

e / r

adia

ns

frequency / GHz

max structure delay = +/- 26 ps(7.6 mm light travel time)

max structure delay = +/- 8.2 ps(2.4 mm light travel time)

Expected delay accuracy 2-‐4 ps

Quasar structure > 20 ps

Quasars What you want them to be ² Bright point sources ² Fixed in space and Sme

Image: N. Reid

What they are ² Supermassive black holes ² 106 Smes more distant than stars

Image: NASA

Image: C. Mueller

•  Quasars are really far away (z >=1 ) (z = 1 is half the age of the Universe) •  Use quasars because… D(z = 1 quasar) = 6.7 x 109 pc D(stars at GalacSc centre) = 8 x 103 pc è quasi-‐iner1al reference frame

Structure delay (ps)


0

0.01

0.02

0.03


0.001

0.002

pol f

orm

al e

rror /

mas

(UT1

-UTC

) for

mal

erro

r / m

s

structure index

xpolypol(UT1-UTC)

EOP formal error (structure only)


0

0.02

0.04

0.06

0.08

0.1

0.12


0.001

0.002

0.003

0.004

0.005

0.006

0.007

0.008

pol f

orm

al e

rror /

mas

(UT1

-UTC

) for

mal

erro

r / m

s

structure index

xpolypol(UT1-UTC)

EOP formal error (with trop.)


0.001

0.01

0.1

1

10

6000 7000 8000 9000 10000

med

ian

posi

tion

offs

et /

cm

average distance to other stations / km

OnWzYsNyZc BdWf FtTs Hb Kk Tc Hh

no structure

Median position offset median of 15 x 30 realizations


0.001

0.01

0.1

1

10

6000 7000 8000 9000 10000

med

ian

posi

tion

offs

et /

cm



SI=1no structure

Median position offset


0.001

0.01

0.1

1

10

6000 7000 8000 9000 10000

med

ian

posi

tion

offs

et /

cm



SI=1SI=2

no structure



0.001

0.01

0.1

1

10

6000 7000 8000 9000 10000

med

ian

posi

tion

offs

et /

cm



SI=1SI=2SI=3

no structure



0.001

0.01

0.1

1

10

6000 7000 8000 9000 10000

med

ian

posi

tion

offs

et /

cm



SI=1SI=2SI=3SI=4

no structure



0.001

0.01

0.1

1

10

6000 7000 8000 9000 10000

med

ian

posi

tion

offs

et /

cm



SI=1SI=2SI=3SI=4

ICRF2no structure



0.001

0.01

0.1

1

10

6000 7000 8000 9000 10000

med

ian

posi

tion

offs

et /

cm



SI=1SI=2SI=3SI=4

ICRF2real structureno structure



0.1

0.2

0.3

0.5

1.0

2.0

6000 7000 8000 9000 10000

med

ian

posi

tion

offs

et /

cm


OnWz YsNyZc BdWf FtTs Hb Kk Tc Hh

SI=1SI=2SI=3SI=4

ICRF2real structureno structure



using quasar physics to improve the vlbi reference frame · −300 −200 −100 0 100 200 300...

Documents