towards the zero-spacing: mosaicing and single dish
TRANSCRIPT
![Page 1: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/1.jpg)
Towards the zero-spacing: Mosaicing and Single Dish CombinationNaomi McClure-Griffiths The Australian National University
![Page 2: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/2.jpg)
Outline• Big images = mosaicing • What is the zero-spacing problem?
– Impacts on large-scale emission – Imaging artefacts – Total-power
• Solutions – Concept – Cross-calibration – Recipes
2
![Page 3: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/3.jpg)
Why Mosaic?
• Wide-field imaging: – Interested in source that is larger than
primary beam, θ > λ / D
• Large scale structure: – Interested in structure on scales larger than
that sampled by the shortest baseline: θ > λ/dmin
3
![Page 4: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/4.jpg)
• Insert ASKAP image of SMC in continuum
4
• Insert ASKAP image of SMC in HI
Two different reasons to mosaic:
![Page 5: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/5.jpg)
An Interferometer samples angular scales:
5
d
D
![Page 6: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/6.jpg)
An Interferometer samples angular scales:
5
d
D
![Page 7: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/7.jpg)
Mosaicing Fundamentals
• Background theory: – Ekers & Rots (1979) pointed out that one can
think of a single dish as a collection of sub-interferometers.
6
diameter DD
“Fourier coverage”
*Ron showed us this in his first talk
![Page 8: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/8.jpg)
7
![Page 9: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/9.jpg)
7
![Page 10: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/10.jpg)
7
![Page 11: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/11.jpg)
7
![Page 12: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/12.jpg)
Mosaicing Fundamentals
d
D D
dd - D
“Fourier coverage”
d + D
• Extending this formalism to interferometers, we find that an interferometer doesn’t just measure angular scales θ =λ / d it actually measures λ / (d – D) < θ < λ / (d + D)
8
d + D
d - D
![Page 13: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/13.jpg)
Mosaicing Fundamentals• But you can’t get all that extra info from a single
pointing – As with a single dish, you have to scan to get the extra
“spacings” • Ekers & Rots showed that you can recover this
extra information by scanning the interferometer • The sampling theorem states that we can gather
as much information by sampling the sky with a regular, Nyquist spaced grid (Cornwell 1988)
9
![Page 14: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/14.jpg)
• Insert ASKAP image of SMC in continuum
10
• Insert ASKAP image of SMC in HI
Joint deconvolutionLinear mosaic
![Page 15: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/15.jpg)
Joint Deconvolution Approach• Form a linear combination of the individual pointings, p:
• Here σp is the noise variance of an individual pointing and A(l) is the primary response function of an antenna
• W(l) is a weighting function that suppresses noise amplification at the edge of mosaic (amongst other things)
11
Sault, Brouw, & Staveley-Smith (1996)
![Page 16: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/16.jpg)
Mosaicing: Joint Approach
• Joint dirty beam depends on antenna primary beam:
• Use all u-v data from all points simultaneously
– Extra info gives a better deconvolution
12
![Page 17: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/17.jpg)
But…– Maximum angular scale even with
misaiming is:
13
D
dmin
![Page 18: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/18.jpg)
14
That’s the zero-spacing problemASKAP, 8h
km-3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3
km-3
-2.5
-2-1
.5-1
-0.5
00.
51
1.5
2
Natural UV coverageNatural UV coverage
baseline ubaseline ubase
line
vba
selin
e v
![Page 19: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/19.jpg)
14
That’s the zero-spacing problemASKAP, 8h
km-3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3
km-3
-2.5
-2-1
.5-1
-0.5
00.
51
1.5
2
Natural UV coverageNatural UV coverage
baseline ubaseline ubase
line
vba
selin
e v
km-1.8 -1.6 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
km-1
.4-1
.2-1
-0.8
-0.6
-0.4
-0.2
00.
20.
40.
60.
81
1.2
1.4
Natural UV coverageNatural UV coverage
baseline ubaseline ubase
line
vba
selin
e v
![Page 20: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/20.jpg)
14
That’s the zero-spacing problemASKAP, 8h
km-3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3
km-3
-2.5
-2-1
.5-1
-0.5
00.
51
1.5
2
Natural UV coverageNatural UV coverage
baseline ubaseline ubase
line
vba
selin
e v
km-1.8 -1.6 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
km-1
.4-1
.2-1
-0.8
-0.6
-0.4
-0.2
00.
20.
40.
60.
81
1.2
1.4
Natural UV coverageNatural UV coverage
baseline ubaseline ubase
line
vba
selin
e v
m-450 -400 -350 -300 -250 -200 -150 -100 -50 0 50 100 150 200 250 300 350 400
m-3
50-3
00-2
50-2
00-1
50-1
00-5
00
5010
015
020
025
030
0
Natural UV coverageNatural UV coverage
baseline ubaseline ubase
line
vba
selin
e v
![Page 21: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/21.jpg)
The zero-spacing problem
• So if the source is large compared to
there’s a problem that:
1. Limits ability to recover large-scale structure 2. Causes image artefacts around extended
objects 3. Prevents total flux measurements
15
✓max
⇠ �/dmin
![Page 22: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/22.jpg)
16
![Page 23: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/23.jpg)
16
Echidna
![Page 24: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/24.jpg)
16
Echidna FT(Echidna)
![Page 25: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/25.jpg)
17
Echidna
missing zero-spacing
FT
![Page 26: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/26.jpg)
17
Echidna
missing zero-spacing
FT
FT-1
![Page 27: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/27.jpg)
17
Echidna
missing zero-spacing
FT
FT-1
![Page 28: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/28.jpg)
18
Echidna
“single dish”
FT
![Page 29: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/29.jpg)
18
Echidna
“single dish”
FT
FT-1
![Page 30: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/30.jpg)
18
Echidna
“single dish”
FT
FT-1
![Page 31: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/31.jpg)
Image Artefacts• Cause of
“negative bowls”
• Cleaning algorithms try to recover this without information
19
Braun &
Walterbos (1985)
Fourier domain Spatial domain
![Page 32: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/32.jpg)
20
Negative bowls
from the CGPS
![Page 33: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/33.jpg)
21
Total flux problems
Helfer et al (2003)
![Page 34: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/34.jpg)
21
Total flux problems
Helfer et al (2003)
![Page 35: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/35.jpg)
Recovering large-scale diffuse emission
22
McClure-Griffiths et al (2003)
![Page 36: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/36.jpg)
Recovering large-scale diffuse emission
22
McClure-Griffiths et al (2003)
![Page 37: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/37.jpg)
Solution Methods:
• Basic methods: – Combine dirty data in u-v plane then image
and deconvolve – Image, deconvolve, then combine
• Variants: – Combine during maximum entropy (“joint”
deconvolution) – Combine during deconvolution with single-dish
as model (“default” method)
23
![Page 38: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/38.jpg)
24
Using a single-dish as an interferometer
Ekers & Rots (1979)
IDsd(l,m) = I(l,m) ⇤Bsd(l0,m0)
V 0sd(u, v) = V (u, v)⇥ bsd(u, v)
Single dish gives “visibilities” from zero-spacing to D
![Page 39: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/39.jpg)
Cross-calibration• Scale factor to link
flux scales:
• Measure intensities in overlapping u-v space
25
Interferometer data
Single dish data
Overlapping region
u
v
fcal =Sint
Ssd
fcal =IintIsd
![Page 40: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/40.jpg)
Combine Dirty Images, then Deconvolve
• Combine the dirty images, and
• where the ratio of beam solid angles gives
• and similarly combine the beams
26
IDint IDsd
IDcomb
=IDint
+ ↵ fcal
IDsd
1 + ↵
↵ =⌦int
⌦sd
– Implemented in miriad’s mosmem, casa clean(?)
![Page 41: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/41.jpg)
27
Beam combination
+ =
ATCA Parkes ATCA+Parkes
Bcomb
=B
int
+ ↵Bsd
1 + ↵
![Page 42: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/42.jpg)
Combination after deconvolution
• “Feathering” technique combines deconvolved image in Fourier space
where
28
Vcomb
(k) = !0(k)Vint
(k) + fcal
!00(k)V 0sd
(k)
!0(k) + !00
(k) =1p2⇡
exp
✓�✓2intk
2
4 ln 2
◆
![Page 43: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/43.jpg)
29
Weighting functions
![Page 44: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/44.jpg)
29
Weighting functions
![Page 45: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/45.jpg)
29
Weighting functions
![Page 46: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/46.jpg)
30
FT
FT
FT-1=fcal+
x
Image, deconvolve, then combine
Implemented in miriad’s immerge, casa feather
![Page 47: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/47.jpg)
Combination during Deconvolution: I• Use the SD image as a “default” in
deconvolution – Implemented in miriad’s mosmem, casa’s clean and
ASKAPsoft (at least it was!)
31
Normal Clean SD default
![Page 48: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/48.jpg)
Combination during Deconvolution: I• Use the SD image as a “default” in
deconvolution – Implemented in miriad’s mosmem, casa’s clean and
ASKAPsoft (at least it was!)
31
Normal Clean SD default
![Page 49: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/49.jpg)
Combination during Deconvolution: I• Use the SD image as a “default” in
deconvolution – Implemented in miriad’s mosmem, casa’s clean and
ASKAPsoft (at least it was!)
31
Normal Clean SD default
![Page 50: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/50.jpg)
Combination during Deconvolution: II• Jointly deconvolve both images using a
maximum entropy technique • Where we maximise
• subject to:
32
{ = �X
i
Ii ln
✓IiMie
◆
X
i
�IDint �Bint ⇤ I
2
i< N �2
int
X
i
⇢IDsd �
Bsd ⇤ Ifsd
�2
i
< M �2sd
– Implemented in miriad’s mosmem
![Page 51: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/51.jpg)
33
Rathborne et al (2015)
ALMA + Mopra
![Page 52: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/52.jpg)
Some Complications• Noise matching • Need big single dish for overlapping u-v coverage:
– rule-of-thumb Dsd~2*dmin
• Cross-calibration very-much subject to ratio of beam sizes
• Single dish image not larger than interferometer or aliasing
• Single dish is not well-defined: – elevation effects, sidelobes (<70% efficiency) – certainly not a perfect Gaussian beam so any method that
deconvolves SD suffers, e.g. joint deconvolution – Brightness temperatures vs Jy/Bm
34
![Page 53: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/53.jpg)
35
The ALMA + ACA solution
From J Ott
![Page 54: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/54.jpg)
Summary• Widefield imaging can include the desire to recover
extended emission – Mosaic-ing can help with this
• Lack of “zero”-spacing in interferometers leads to: – lack of sensitivity to large scale emission – imaging artefacts (negative bowls) – inability to measure total flux
• Can be solved by combining interferometer with single dish data via – “joint” or “default”deconvolution – “feathering” – combination then deconvolution
36
![Page 55: Towards the zero-spacing: Mosaicing and Single Dish](https://reader033.vdocuments.net/reader033/viewer/2022041617/6253456d9b3a192eec1ffc32/html5/thumbnails/55.jpg)
Some useful literature• Stanimirovic (2002) ASP Conf. Series 278 • Sault & Killeen (2003) Miriad Users Manual • Holdaway (1999) ASP Conf. Series 180 • Ekers & Rots (1979) Image Formation, IAU Coll 49, 61. • Cornwell (1988) A&A, 202, 316. • Cornwell (1989) ASPC 6. • Cornwell, Holdaway & Uson (1993) A&A, 271, 697. • Sault, Staveley-Smith & Brouw (1996) A&A Suppl., 120, 375. • Holdaway (1998) ASPC 180, ch.20. • Subrahmanyan (2004) MNRAS, 348, 1208. • Bhatnagar, Golap & Cornwell (2005) ASPC 347, 96
37