Angelica de OliveiraCosta

1
Modeling Diffuse Galactic Radio Emission from 10
MHz to 100 GHZ
Angelica de Oliveira-Costa
Collaborators M.Tegmark, B.Gaensler, J.Jonas,
T.Landecker P.Reich
2
Modeling Diffuse Galactic Radio Emission from 10
MHz to 100 GHZ
3000
Angelica de Oliveira-Costa
Collaborators M.Tegmark, B.Gaensler, J.Jonas,
T.Landecker P.Reich
3
Modeling Diffuse Emission from 10 MHz to 100 GHz
Galactic Contamination
To do a good job on removing foregrounds, we need
to understand their frequency and scale
dependence, frequency coherence, and better
characterize their non-Gaussian behavior.
Objectives
Accurate modeling and subtraction of Galactic
foreground contamination in order to correct the
measured 21 cm power spectrum.
Unique opportunity to understand the Galactic
emission process between 10 MHz to 100 GHz.
Jelic Zaroubi (2007)
GEM Collaboration
Emission Mechanisms
4
Modeling Diffuse Emission from 10 MHz to 100 GHz
Galactic Contamination
To do a good job on removing foregrounds, we need
to understand their frequency and scale
dependence, frequency coherence, and better
characterize their non-Gaussian behavior.
Emission Mechanisms
Objectives
Modeling of Galactic foreground contamination
helps to optimize experimental design, scan
strategy data analysis pipelines. Also, it
helps to quantify mitigate contamination from
distant sidelobes (Judds talk).
GEM Collaboration
5
Modeling Diffuse Emission from 10 MHz to 100 GHz
Galactic Contamination
To do a good job on removing foregrounds, we need
to understand their frequency and scale
dependence, frequency coherence, and better
characterize their non-Gaussian behavior.
Emission Mechanisms
Objectives
Modeling of Galactic foreground contamination
helps to optimize experimental design, scan
strategy data analysis pipelines. Also, it
helps to quantify mitigate contamination from
distant sidelobes (Judds talk).
GEM Collaboration
6
Modeling Diffuse Emission from 10 MHz to 100 GHz
7
Modeling Diffuse Emission from 10 MHz to 100 GHz
0.013
0.017
0.010
0.026
0.035
0.038
0.022
0.081
0.045
0.150
0.085
0.400
0.408
0.176
0.404
23
1.42
0.820
2.33
41
33
94
61
8
Modeling Diffuse Emission from 10 MHz to 100 GHz
0.010
0.026
0.035
0.022
0.081
0.045
0.150
0.085
0.408
23
1.42
0.820
2.33
41
33
94
61
9
Modeling Diffuse Emission from 10 MHz to 100 GHz
0.010
0.026
0.035
0.022
0.081
0.045
0.408
23
1.42
2.33
41
33
94
61
10
Modeling Diffuse Emission from 10 MHz to 100 GHz
0.010
0.026
0.022
0.081
0.045
0.408
23
1.42
2.33
41
33
94
61
11
Modeling Diffuse Emission from 10 MHz to 100 GHz
0.010
0.026
0.022
0.081
0.045
0.408
23
1.42
2.33
41
33
94
61
12
Modeling Diffuse Emission from 10 MHz to 100 GHz
0.010
0.045
Comparison between different methods tested
0.045
89
22
Scaled Haslam
PCA Spline
16
Polynomial Fit
Spline only
5
10 freq calc
13
Modeling Diffuse Emission from 10 MHz to 100 GHz
0.010
Principal Component Analysis
1st
80
1st Comp
3rd Comp
2nd Comp
2nd
19
free-free
3rd
synchrotron
0.6
spinning dust
dust
14
Modeling Diffuse Emission from 10 MHz to 100 GHz
0.010
Principal Component Analysis
1st Comp
3rd Comp
2nd Comp
free-free
synchrotron
spinning dust
dust
Scale lg(T/1K)
15
Modeling Diffuse Emission from 10 MHz to 100 GHz
0.010
http//www.space.mit.edu/home/angelica/gsm
16
Modeling Diffuse Emission from 10 MHz to 100 GHz
0.010
Comparison between data model, how good is our
fit?
The predicted map is constructed without using
the data from that original frequency.
Relative rms error in the sky region 123456.
17
Modeling Diffuse Emission from 10 MHz to 100 GHz
0.010
Movies
logT
spectral index, b
Running of b, g
Movies, MWA range
logT
spectral index, b
Running of b, g
18
Modeling Diffuse Emission from 100 GHz to 3000
GHz
41
33
23
0.408
94
61
94, FDS
240mm
60mm
100mm, SFD
100mm
140mm
Ha, Fink
Predictions for a given map
19
Modeling Diffuse Emission from 100 GHz to 3000
GHz
41
33
23
94
61
94, FDS
240mm
60mm
100mm, SFD
100mm
140mm
Ha, Fink
Predictions for a given map
20
Modeling Diffuse Emission from 100 GHz to 3000
GHz
0.010
21
Conclusions
We have presented a global sky model for 10 MHz
to 3000 GHz Galactic emission derived from all
publicly available total power large-area radio
surveys, digitized with OCR when necessary and
compiled to a uniform format. Both data
compilation and our software for returning a
predicted all-sky map are available at
http//www.space.mit.edu/home/angelica/gsm
Our ability to measure cosmological parameters
using 21 cm tomography will only be as good as
our understanding of the foregrounds.
To do a good job on removing foregrounds, we need
to improve/better model their emission from the
MHz to the GHz range, to understand their
frequency and scale dependence, frequency
coherence, and better characterize their
non-Gaussian behavior.
We are currently extending this work to include
extragalactic point sources and galactic
polarized emission.
22
Modeling Diffuse Emission from 100 GHz to 3000
GHz
41
33
23
0.408
94
61
94, FDS
240mm
60mm
100mm, SFD
100mm
140mm
25mm
12mm
4.9mm
Ha, Fink