## Rees-Sciama effect

Anais Rassat
Posts: 24
Joined: October 20 2004
Affiliation: LASTRO (EPFL)
Contact:

### Rees-Sciama effect

From what I understand the Rees-Sciama effect on the CMB can be thought of as a non-linear, secondary ISW. Is it large enough to be observed, and has it ever been?

Anais.

Aurelien Fraisse
Posts: 11
Joined: September 27 2004
Affiliation: Princeton University

### Rees-Sciama effect

Hi Anais,

The Rees-Sciama effect is indeed the non-zero second order of the ISW effect on small scales.
See, e.g., http://www.nature.com/nature/journal/v2 ... 7511a0.pdf.

It has not been observed yet, but this should become possible in the next few years, e.g., by cross-correlating the kSZ effect and cosmic shear.
This was recently discussed in a paper by Dore, Hennawi and Spergel: astro-ph/0309337.

Cheers,
-Aurelien.

Syksy Rasanen
Posts: 119
Joined: March 02 2005
Affiliation: University of Helsinki

### Rees-Sciama effect

The Rees-Sciama effect (caused by the time-dependence of the gravitational potential inside non-virialised non-linear structures) has not been observed (though see below). The RS effect from a typical distribution of non-linear objects was estimated to be at most 10^{-6} in astro-ph/9506048. Studies of local structures (which happen to be somewhat non-typical) in the 1990s also concluded that the effect due to the Great Attractor at 40-60 h^{-1} Mpc (which was thought to be responsible for most of the kinetic dipole) is at most 10^{-6}.

However, recent observations suggest that the dipole velocity instead comes from the Shapley Supercluster, around 130-180 h^{-1} Mpc away, so the Rees-Sciama effect could be 10^{-5}, at the same level as the intrinsic anisotropies. Dominik Schwarz and me had the idea that this might explain the preferred direction in the CMB. As we (and Aleksandar Rakic) discuss in astro-ph/0601445 (you'll find references for the above statements there), this however has the wrong observational signature. The RS effect would lead to the quadrupole and the octopole to line up with the dipole, i.e. to have extra power in the m=0 component, whereas the quadrupole and octopole are in fact observed to be planar (i.e. favor the m=\pm l components).

In astro-ph/0509039 the claim was made that the lensing of the CMB dipole is equivalent to the RS effect from local structures, and that this could explain the observed preferred direction. It's true that if you add contamination which is orthogonal to the dipole, the statistical significance of the anomalies goes down. (Though I don't understand is how the imprint is orthogonal to instead of aligned with the dipole when described in terms if the RS effect.) This has been discussed in another thread, http://cosmocoffee.info/viewtopic.php?t ... ht=pancake .

Anais Rassat
Posts: 24
Joined: October 20 2004
Affiliation: LASTRO (EPFL)
Contact:

### Rees-Sciama effect

Yes I read your paper, which brought up the question I had. You say that if the dipole originates from Shapley, then the RS effect would be 10^-5, ie the same order as the other anisotropies, so wouldn't it have been observed?

Is non-observation of RS proof that the dipole doesn't originate from Shapley?

Anais.

Syksy Rasanen
Posts: 119
Joined: March 02 2005
Affiliation: University of Helsinki

### Rees-Sciama effect

We say that if there is a major contribution to the dipole from Shapley (as seems to be the case), then the RS effect can be of the order 10^{-5}. This number is based on scaling arguments and comparison to calculations of the Great Attractor, and one should make a more careful estimate before saying something definite about the relation of the dipole and local structures.

Aurelien Fraisse
Posts: 11
Joined: September 27 2004
Affiliation: Princeton University

### Rees-Sciama effect

[It could be observed] by cross-correlating the kSZ effect and cosmic shear. This was recently discussed in a paper by Dore, Hennawi and Spergel: astro-ph/0309337.
Sorry, what I meant is "cross-correlating the convergence of the CMB ($\kappa$) with its temperature". It seems to be the only way of getting rid of all the known effects other than the RS. This was not discussed in the reference above, but in part in a paper by Verde and Spergel: astro-ph/0108179. I haven't found a more detailled analysis (e.g., including the halo model) yet.

Cheers,
-Aurelien.