Simulation of The Analysis of Interferometric Microwave Background Polarization Data

a
r
X
i
v
:
1
4
0
7
.
3
3
0
5
v
1

[
a
s
t
r
o
-
p
h
.
C
O
]

1
1

J
u
l

2
0
1
4
Statistical Challenges in 21st Century Cosmology
Proceedings IAU Symposium No. 306, 2014
A.F. Heavens, J.-L. Starck, A. Krone-Martins, eds.
c 2014 International Astronomical Union
DOI: 00.0000/X000000000000000X
Simulation of the analysis of interferometric
microwave background polarization data
Emory F. Bunn
1
, Ata Karakci
2
, Paul M. Sutter
3,4
, Le Zhang
5
,
Gregory S. Tucker
2
, Peter T. Timbie
5
and Benjamin D. Wandelt
4
1
University of Richmond, USA
email:[email protected]
2
Brown University, USA
3
Ohio State University, USA
4
Institut dAstrophysique de Paris, France
5
University of Wisconsin Madison, USA
Abstract. We present results from an end-to-end simulation pipeline interferometric observa-
tions of cosmic microwave background polarization. We use both maximum-likelihood and Gibbs
sampling techniques to estimate the power spectrum. In addition, we use Gibbs sampling for
image reconstruction from interferometric visibilities. The results indicate the level to which
various systematic errors (e.g., pointing errors, gain errors, beam shape errors, cross- polariza-
tion) must be controlled in order to successfully detect and characterize primordial B modes as
well as other scientic goals. In addition, we show that Gibbs sampling is an eective method
of image reconstruction for interferometric data in other astrophysical contexts.
Keywords. cosmology: cosmic microwave background, techniques: interferometric, methods:
statistical
Measurement of cosmic microwave background (CMB) anisotropy has become one of
the most powerful tools in cosmology. In recent years, researchers have built on the suc-
cess of these anisotopy measurements by studying linear polarization in the CMB. In
particular, considerable attention has been focused on the search for B-mode polariza-
tion, which has the potential to measure the energy scale of ination, along with probing
cosmology in a variety of other ways (Hu & Dodelson 2002). BICEP2 has measured a B-
mode polarization signal in the microwave sky (Ade et al. 2014), which if conrmed will
represent a major advance in cosmology. The eld eagerly awaits other measurements at
dierent frequencies and with dierent instruments.
Because the B-mode signal is expected to be very faint, control of systematic errors is
of paramount importance. An argument can be made (Timbie et al. 2006, Bunn 2007)
that interferometers provide better control of systematics than imaging telescopes. This
is one of the reasons that, for instance, the QUBIC collaboration (Ghribi et al. 2013) is
constructing an instrument based on the novel approach of bolometric interferometry.
Whether or not interferometers have better systematic error properties than imagers,
it is clear that they have dierent sensitivity to systematics. Interferometric systematic
issues have not received as much attention as systematics in imaging systems. Given
the importance of a robust characterization of CMB B modes, it seems worthwhile to
study these eects in detail. For these reasons, we have performed detailed simulations of
interferometric observations of CMB polarization, in order to characterize the eects of
various systematic errors on the reconstruction of the polarization power spectra (Karakci
et al. 2013a, Karakci et al. 2013b, Zhang et al. 2013).
We have developed tools that generate visibilities for interferometers with arbitrary an-
1
2 E.F. Bunn et al.
0
5
10
15
20
25
50 100 150 200 250 300 350
C
l E
E
l(
l+
1
)
/ 2
[
K
2
]
l
EE Angular Power Spectrum
GS
ML
Input
Realization
0
0.01
0.02
0.03
0.04
0.05
50 100 150 200 250 300 350
C
l B
B
l(
l+
1
)
/ 2
[
K
2
]
l
BB Angular Power Spectrum
GS
ML
Input
Realization
g=0.1 Antenna Gain Errors (EE & BB)
50 100 150 200 250 300 350
Multipole l
0.01
0.10
C
l
/

l
=5x10
4
Beam Cross Polarization (EE & BB)
50 100 150 200 250 300 350
Multipole l
0.001
0.010
0.100
C
l
/

l
Figure 1. Top: Mean posterior band powers obtained by Gibbs sampling (GS) are shown in
black. The maximum-likelihood (ML) band powers are shown in blue. Dark and light grey
indicate 1 and 2 uncertainties on the Gibbs sampling results. Binned power spectra of the
signal realization and input power spectra are shown in pink and red. Bottom: Examples of
the eects of systematic errors, obtained by both ML (triangles) and GS (dots) are shown.
Solid and dashed lines correspond to experiments that interfere linear and circular polarization
states respectively. Results are shown for the EE (red) and BB (blue) power spectra. For further
details, see Karakci et al. (2013b).
tenna placement, beam shape, etc., from HEALPix sky maps. We can include the eects
of a wide variety of systematic errors, including beam shape and pointing errors, cross-
polarization, gain errors, etc. We then estimate power spectra from these visibilities in two
independent ways, via maximum-likelihood estimation (Hobson & Maisinger 2002)and
Gibbs sampling (Larson et al. 2007, Sutter et al. 2012).
Figure 1 show the results of simulations that are described in detail in Karakci et al.
(2013a), Zhang et al. (2013), and Karakci et al. (2013b). These results are for a simulated
interferometer consisting of a 2020 close-packed array of feedhorns with a Gaussian
beam width of 5
and separation D = 7.89. We include sky rotation for observations

from the South Pole, with average noise per visibility of 0.015 K. Figure 1 shows that
we correctly reconstruct the power spectra and illustrates the eects of introducing some
systematic errors. Order-of-magnitude agreement is found between the results of these
simulations and a simple semi-analytic approach (Bunn 2007). Far more detailed results
may be found in the papers cited above.
Gibbs sampling provides simultaneous samples of the power spectrum and the signal
map. We have shown that these signal map samples provide excellent image reconstruc-
tion from visibility data in other (non-CMB) contexts (Sutter et al. 2014). Figure 2 shows
some of the results of this work.
On the left of Figure 2 we show sample images taken from the CASA user guide
(Jaeger 2008). These images were then observed with a simulated interferometer con-
sisting of 12 randomly-placed antennas, each with a beam size of 0.075 times the image
width. We assumed 6 hours of observation with a signal-to-noise ratio of 10 per visibility.
The center panel shows the mean reconstructed image calculated via Gibbs sampling,
multiplied by the primary beam. The right panel shows the result of
1
reconstruction in
Simulation of CMB interferometric data 3
Figure 2. Examples of image reconstruction. The left panel shows test images from the CASA
user guide, multiplied by the primary beam of our simulated observations. The center panel
shows the result of image reconstruction using Gibbs sampling, and the right shows 1 recon-
struction. Adapted from Sutter et al. (2014).
the pixel basis, which has been shown (Wiaux et al. 2009) to have similar performance
to the widely-used CLEAN algorithm (Hogbom 1974). By a variety of quantitative mea-
sures, Gibbs reconstruction performs better than this proxy for CLEAN reconstruction
(Sutter et al. 2014).
This research was funded by the following awards from the National Science Founda-
tion: AST-0908902, AST-0908844, AST-0908900, AST-0908855, AST-0927748.
References
Ade, P.A.R. et al. 2014, Phys. Rev. Lett., 112, 241101
Bunn, E.F. 2007, Phys. Rev. D, 75, 083517
Ghribi, A., et al. 2013, JLTP, 173
Hobson, M.P. & Maisinger, K. 2002, MNRAS, 334, 569
Hogbom J.A. 1974, A&AS, 15, 417
Hu W. & Dodelson, S. 2002, ARAA, 40, 171
Jaeger et al. 2008, in Argyle, R. W., Bunclark, P. S., & Lewis J. R., eds, Astronomical Data
Analysis Software and Systems XVII, Vol. 394 of Astronomical Society of the Pacic Con-
ference Series, The Common Astronomy Software Application (CASA), p. 623
Karakci, A., Sutter, P.M., Zhang, L., Bunn, E.F., Korotkov, A., Timbie, P., Tucker,G.S., &
Wandelt, B.D. 2013, ApJS, 204, 10
Karakci, A., Zhang, L., Sutter, P.M., Bunn, E.F., Korotkov, A., Timbie, P., Tucker, G.S., &
Wandelt B. D. 2013, ApJS, 207, 14
Larson, D.L., Eriksen, H.K., Wandelt, B.D., Gorski, K.M., Huey, G., Jewell, J.B., & ODwyer
I.J. 2007, ApJ, 656, 653
Sutter, P.M., Wandelt, B.D., & Malu, S.S. 2012, ApJS, 202, 9
Sutter P. M., et al. 2014, MNRAS, 438, 768
Timbie, P.T., et al. 2006, New Astron. Revs, 50, 999
Wiaux, Y., Jacques, L., Puy, G., Scaife, A.M.M., & Vandergheynst P. 2009, MNRAS, 395, 1733
Zhang, L., Karakci, A., Sutter, P.M., Bunn, E.F., Korotkov, A., Timbie P., Tucker, G.S., &
Wandelt, B.D. 2013, ApJS, 206, 24

Simulation of The Analysis of Interferometric Microwave Background Polarization Data

Uploaded by

Copyright:

Available Formats

Simulation of The Analysis of Interferometric Microwave Background Polarization Data

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Simulation of The Analysis of Interferometric Microwave Background Polarization Data

Uploaded by

Copyright:

Available Formats

a

and separation D = 7.89. We include sky rotation for observations

You might also like