SPT-3G 2018 TE/EE Bandpowers and Likelihood


This page provides data products associated with the SPT-3G measurement of the CMB EE and TE polarization power spectra over 1500 square degrees described in Dutcher et al., 2021 arXiv:2101.01684. The power spectra are presented over the angular multipole range 300 < ell <= 3000 and can be used to constrain cosmological models using the likelihood package provided below. The data products here also relate to the results of Balkenhol et al., 2021, arXiv:2103.13618.

If you have any questions regarding this data set or its use, please contact Daniel Dutcher (ddutcher_at_uchicago_dot_edu) or Lennart Balkenhol (lbalkenhol_at_student_dot_unimelb_dot_edu_dot_au).


Individual figures are available in PDF format here or bundled in a ZIP file here, covering both manuscripts. We highlight selected figures below.

SPT-3G 2018 150 GHz E-mode polarization map. The data have been filtered to remove features larger than approx. 0.5 degrees, and the map has been smoothed by a 6 arcminute FWHM Gaussian. (Figure 3 in Dutcher et al., 2021)

The minimum-variance SPT-3G EE and TE bandpowers (red) overlaid on the Planck best-fit LCDM model, along with the recent measurements from Planck (Planck Collaboration 2018-V), ACT DR4 (Choi et al. 2020), POLARBEAR (POLARBEAR Collaboration 2019), and SPTpol (Henning et al. 2018). The Planck EE bandpowers are restricted to ell < 1500. The uncertainties shown for the SPT-3G bandpowers are the square root of the diagonal elements of the covariance matrix and do not include beam or calibration uncertainties. (Figure 8 in Dutcher et al., 2021)

Marginalized constraints on LCDM parameters and the Hubble constant for the SPT-3G 2018 EE/TE, SPTpol (Henning et al. 2018), and Planck (Planck Collaboration 2018-V) datasets. SPT-3G produces consistently tighter constraits than SPTpol. The results from SPT-3G are statistically consistent with the findings of Planck. (Figure 9 in Dutcher et al. 2021)

We show samples in the H_0 vs. N_eff plane from SPT-3G 2018 chains, colored according to S_8, a parameter describing the amplitude of matter perturbations today. The color range has been chosen to match the 3 sigma range of the latest KiDS-1000 results (Heymans et al. 2020). For comparison, we also show the Planck 2D marginalized posterior probability (black lines), and the 2 sigma range of the Hubble constant measurement from the distance-ladder of Riess et al. 2020. The dotted grey line is the standard model prediction of N_eff = 3.044. Contours indicate the 68% and 95% probability regions. (Left panel of Figure 1 in Balkenhol et al. 2021)

Bandpowers for Plotting

We provide the minimum-variance combination of the SPT-3G 2018 multi-frequency EE and TE Bandpowers from observations over a four-month period in 2018. The bandpowers cover the angular multipole range 300 < ell <= 3000 in bins of delta_ell = 50 between 301 < ell < 200 and delta_ell = 100 from 2001 < ell <= 3000. The bandpowers, bandpower errors, and associated bandcenters, are provided for plotting in a text file here. Bandpowers and uncertainties are given in D_ell units of uK^2. Each row is one ell-bin. The columns are (1) ell_min, (2) ell_max, (3) TE ell_center, (4) TE, (5) sigma_TE, (6) EE ell_center, (7) EE, (8) sigma_EE. Note that these bandpowers include power from foregrounds, such as Galactic dust and point sources, though we do not significantly detect this power. Foreground terms are included in our likelihood model and priors for them may be adjusted by editing the accompanying files provided in the likelihood package below. If you have any questions regarding this data set or its use, please contact Daniel Dutcher (ddutcher_at_uchicago_dot_edu).


We provide files in a gzipped tarball here that can be added to the May 2020 version of CosmoMC to interface with the SPT-3G 2018 EE/TE bandpowers discussed above. These files include the SPT-3G likelihood, SPT-3G data (multi-frequency bandpowers, covariance matrix, bandpower window functions, beam covariance, effective band centres, and calibration covariance), batch file, a sample .ini file including the baseline priors on foreground parameters, a proposal matrix for LCDM assuming the baseline priors, and a README with instructions for how to compile CosmoMC with the SPT-3G 2018 EE/TE likelihood. If you have any questions regarding the use of this likelihood, please contact Lennart Balkenhol (lbalkenhol_at_student_dot_unimelb_dot_edu_dot_au)).

In addition, we have created a GitHub repository that contains the SPT3G Y1 likelihood code and data encapsulated in clik format SPT3G_Y1_EETE.clik allowing it to be called from any sampled that is already linked with the clik library in order to run the Planck likelihoods. The package contains instruction for usage in cosmomc and cobaya.

v1.0: March 26, 2021 - Initial version released with arXiv submission.

v2.0: September 3, 2021 - Added link to github repository, with likelihood in clik format.

v3.0: October 28, 2021 - Updated calibration and beam covariance.


(October 20, 2021): We update the final temperature and polarization calibration of the SPT-3G band powers and the beam covariance matrix. The new calibration has a negligible impact impact on cosmological parameter constraints, apart from the combined amplitude parameter, which shifts by 20% of its uncertainty when using only SPT-3G data from 1.819 ± 0.038 to 1.812 ± 0.040. Joint constraints from SPT-3G and Planck do not change. The change to the beam covariance has no measureable impact on parameter constraints. These corrections appear in v3.0 of the likelihood as well as the band powers for plotting.