Project Supported by

NSF, National Science Foundation
KICP, Kavli Institute for Cosmological Physics
USAP, United States Antarctic Program
Raytheon Polar Support Corporation


Related Websites

CBI, Cosmic Background Imager
DASI, Degree Angular Scale Interferometer
WMAP, Wilkinson Microwave Anisotropy Probe




 Overview 




This page provides data products associated with the SPT measurement of the CMB power spectrum described in astroph/1210.7231, Story et al 2012. The power spectrum is presented over the multipole range 650 < ell < 3000 and can be used to constrain cosmological models.
If you have any questions regarding this data set or its use, please contact Kyle Story (kstory_at_uchicago_edu).


 Figures 




Individual figures are available in PDF format here or bundled in a ZIP file here. Below we show a couple of example figures.

The SPT power spectrum is shown here.
The peak at ell~800 is the third acoustic peak.
The bandpower errors shown do not include beam or calibration uncertainties.
(This is the left panel of Figure 3 in Story et al. 2012.)


The SPT bandpowers, WMAP bandpowers, and bestfit LCDM theory spectrum shown with dashed (CMB) and solid (CMB+foregrounds) lines. The bandpower errors do not include beam or calibration uncertainties.
(This is Figure 4 in Story et al. 2012.)



 Bandpowers 




Update: Jun 11, 2014:
Files compatible with the latest (May 2014) version of cosmomc can be downloaded here in a gzipped tarball. These files can be used without recompiling cosmomc. The included README has instructions.
The bandpowers, covariances and window functions are here in a gzipped tarball.
An alternative form of the newdat file with the calibration uncertainty moved from the covariance matrix to line 10 can be found here.
These are the SPT 150 GHz Bandpowers from 2008 to 2011 observations. The bandpowers cover the multipole range 650 < ell < 3000 in bins of deltaell = 50.
Note that these bandpowers include power from foreground components such as SZ and point sources. The contributions from these components must be accounted for in order to properly analyze the SPT bandpowers, as described in Section 6.1 of Story et al. For convenience we give the ellspace templates for these components (in "units" of D_ell=C_ell*ell*(ell+1)/2PI and normalized to equal 1 at ell=3000):
The prior on the amplitude of the SZ (tSZ+kSZ) component (template) is D3000 = 5.5 +/ 3.0 uK^2.
The prior on the amplitude of the Poisson point source component (template) is D3000 = 19.3 +/ 3.5 uK^2.
The prior on the amplitude of the clustered point source component (template) is D3000 = 5.0 +/ 2.5 uK^2.
An alternative form with the foreground priors placed in the newdat file is here in a gzipped tarball. Note that chains in the paper did not use this form, and that the implemented prior no longer gaurantees each of the three foreground terms is positive definite. We expect the latter effect to be small. This newdat file may be simpler to use since it is compatible with an unmodified version of cosmomc  there is no requirement to add the three SPT foreground terms to cosmomc.
The foreground amplitudes are also required to be positive.
A text file with the bandpowers for plotting can be found here.
Each row is one ellbin. The columns are (1) ell_center (2) bandpower for 150x150, (3) bandpower uncertainty for 150x150. Each ell bin has the same width: 50.
If you have any questions regarding this data set or its use, please contact Kyle Story (kstory_at_uchicago_edu).


 Parameter Tables and Markov Chains 




We have uploaded a raw dump of compressed (tar.gz) chains. Each tar file corresponds to a single dataset/cosmological model combindation. These files can be browsed here.
We will add documentation and a more userfriendly interface soon.


 Likelihood Code 




Please see the Likelihood section here. The new bandpowers and foreground templates can be used with the Keisler et al. 2011 likelihood code for cosmomc.
These modifications to cosmomc are not necessary if the newdat file including the foreground priors linked above is used.
The SPTCL cluster likelihood used in some chains can be found here.



