The data made available through this page has been updated. The most recent version of this data may be accessed through /product/map/current/
The best-estimate WMAP temperature power spectrum of the CMB is produced from a weighted combination of the 28 cross power spectra. For multipole moment l < 100, the combined spectrum is formed from a simple weighted average of the 15 V and W band cross-power spectra. For l > 100, all 28 cross power spectra are used to form the combined spectrum, but different weighting schemes are used for three separate l regimes.
The complete (unbinned) TT power spectrum data are provided as an ASCII table containing the following columns:
- the multipole moment l
- the power spectrum in units of uK^2
- the corresponding diagonal elements of the covariance matrix
- the portion of the diagonal element of the covariance matrix attributed to measurement errors
- the portion of the diagonal element of the covariance matrix attributed to cosmic variance, assuming the best-fit running index model described in Spergel et al. (2003).
The diagonal elements provide an estimate of the error, but should not be used as exact error bars. The complete Fisher matrix should be used for that purpose.
A binned version of the power spectrum is also available, in a separate file whose format and content are similar to those described above for the unbinned data. The choice of l bins is the same as that used in e.g., Figure 8 of Hinshaw et al. (see below).
More details can be found in:
Release Notes:
2/11/2003:
Initial release of Combined TT Power Spectrum, V1.0.
2/21/2003:
Combined TT Power Spectrum, V1.1, revised column 3 to estimate cosmic
variance error from the best-fit model C_l instead of from the data
C_l. This gives uncertainties as in e.g., Fig. 10 of Hinshaw et
al. (2003). Also added columns 4 and 5.
Binned TT Power Spectrum, V1.0, new product.
10/31/2003:
Binned TT Power Spectrum, V1.1, added two new columns documenting
the lower and upper values of multipole moment 1 used to form each
bin. These new columns were inserted after column 1 of version 1.0
and the rest of the V1.0 data columns were re-numbered but no changes
were made to their content.
