The Value of Planck CMB Temperature Maps in a post-WMAP World
The previous post, in introducing the power spectrum, introduced the concept of structure on various angular scales. As such, it sets us up for looking at something quite valuable that Planck added beyond what WMAP, launched 8 years earlier, had already delivered. This extra information from Planck is on small angular scales. It is not visible when one looks at a low-resolution graphical presentation of the whole-sky map. So if you look at this:
you might be forgiven for thinking that all Planck did was repeat WMAP and confirm what WMAP found.
To see what Planck added it helps to look at some zoomed-in images such as the following ones that were given to me by Bill Jones of Princeton.
Planck and WMAP both measured the sky through different "bandpasses" that let through different frequencies of light. One of the two instruments on Planck, the straightforwardly named High Frequency Instrument (HFI), has one of its bandpasses centered on 143 GHz. Here is a map of a region of the sky, as mapped by the 143 GHz detectors:
For comparison, below is the same region of the sky as mapped by the "W-band" detectors on WMAP, with central frequency of 94 GHz:
The primary visual difference between these two maps is caused by the much higher level of noise in the WMAP map.
Even so, this comparison might give the impression that Planck improved only slightly over WMAP
because the dominant signal in both the above maps stands out well above the noise. Is anything gained by the lower noise levels in the Planck maps? To answer this question we decompose this map into contributions from fluctuations on different angular scales.
All three of these maps have the same color scale. Note that the intermediate-scale map has the highest amplitude of structure, and the small-scale map has the smallest amplitude, just as you'd expect from the power spectrum curve (see previous post). [Note: the previous power spectra we showed had "angular scale" as the x axis. Here we have the spherical harmonic multipole moment l as the x axis. A spherical harmonic with multipole moment l undergoes l oscillations every 360 degrees.]
Planck starts to improve over WMAP for angular scales that correspond to the second peak (counting from the left so the big one is the first) in the power spectrum. By the time we're at the fourth peak, there is nothing but noise left in the WMAP maps. Let's look at that right-most map from Planck HFI 143 GHz more closely:
It's a bit hard to see the signal in this map, so let's adjust the color scale to have things pop out better:
Almost all the signal here is cosmic microwave background fluctuations. Even with the huge majority of the microwave background signal filtered out (all the fluctuations on larger angular scales), this is still a very high signal-to-noise image. There are two features that jump out at the eye that are not CMB fluctuations. Do you see two concentrated areas of dark color? They are due to emission from low-redshift quasars.
What if one similarly filters a WMAP map? If you do so, the map is totally overwhelmed by noise:
We see that Planck has added a HUGE amount of information beyond what WMAP delivered.
Other maps (most notably from the South Pole Telescope and the Atacama Cosmology Telescope) also have this small-scale information, but only Planck has it from the whole sky.
I am not going to explain here how this information is valuable. I will simply assert that the information on smaller angular scales is the central reason that the Planck results are having a big impact on cosmologists, as demonstrated, for example, by the number of citations received by the Planck paper on cosmological parameters released in 2013. This paper was the most-cited cosmology paper in 2013 and again in 2014, with a current citation total exceeding 3,200.