This week’s entry features a new ACT CMB lensing cross BOSS CMASS galaxies cross-correlation measurement implementing a new tSZ-cleaning procedure, a new update on the H0 tension and in particular how fifth forces could lower the TRGB-calibrated measurement (besides the usual Cepheid-calibrated one), and finally the return of black hole shadows (in asymptotically de Sitter spacetime). Enjoy and stay safe!
#1 2004.01139: The Atacama Cosmology Telescope: A CMB lensing mass map over 2100 square degrees of sky and its cross-correlation with BOSS-CMASS galaxies by Omar Darwish, Mat Madhavacheril, Blake Sherwin et al.
Cross-correlations between Cosmic Microwave Background (CMB) and large-scale structure (LSS) probes are by now relatively mature and important in cosmology. The usual rationale for why one wants to consider cross-correlations is that they are less prone to systematics, since systematics in either of the two (or more) fields one is cross-correlating usually don’t correlate with the other field and its systematics. Putting it heuristically, if I am cross-correlating field A which suffers from systematic sysA, and similarly for field B with systematic sysB, then the cross-correlation gives (A+sysA)×(B+sysB)~A×B if the previous conditions are met. This is mostly but not completely true (and the “not completely” will be important for today’s paper). One type of cross-correlation which has received a lot of attention in recent years is that between CMB lensing (or, more precisely, the CMB lensing convergence field) and galaxy surveys (or, more precisely, the galaxy overdensity field): let me call this cross-correlation Clkg, for an explanation why see Eq.(5) in Darwish et al.’s paper. Clkg is powerful as it can break the degeneracy between galaxy bias and the amplitude of matter fluctuations, degeneracy which is otherwise present when considering galaxy clustering on its own. Therefore, Clkg (particularly if measured at different redshifts) can be particularly useful to study dark energy, modified gravity, massive neutrinos, and so on. Together with my friend and colleague Elena Giusarma, we led the study of one particular Clkg measurement from cross-correlating Planck lensing and BOSS CMASS galaxies, in 1802.08694, studying massive neutrinos as well. One particularly important foreground contaminant when cross-correlating CMB lensing with low-redshift galaxies is the thermal Sunyaev-Zel’dovich (tSZ) effect, due to the interaction of CMB photons with high-energy electrons. tSZ contamination can bias the cross-correlation signal low on large scales, which could explain why a number of works (including our 1802.08694) consistently found a large-scale cross-correlation amplitude too low by about 10-20%.
This is where this week’s paper by Darwish and collaborators (this is an Atacama Cosmology Telescope [ACT] collaboration paper) comes in. In particular, they try to devise a new method to reduce the effect of tSZ contamination in cross-correlations. Initially, Darwish and collaborators tried to apply a method proposed by Madhavacheril and Hill in 1802.08230 (MH18). The way the MH18 method works is quite technical, and the interested reader should read Section 4.1 of this week’s paper, alongside of course the original MH18 work. The basic idea is to two use two different CMB maps from different surveys (in this case ACT and Planck). However, when Darwish et al. went out to apply the MH18 method on real data, they found way more noise than MH18 had forecasted. This is a result of the weights used in MH18 not being minimum-variance. Darwish et al. solve this problem by devising a novel symmetrized tSZ-cleaned estimator, which only comes at a very minimal expense in terms of cross-correlation signal-to-noise. Finally, they cross-correlate these new (mostly) tSZ-free lensing maps, obtained by using ACT and Planck data, with galaxy overdensity maps from the BOSS CMASS sample, detecting the cross-correlation at high signal-to-noise ratio (about 10 sigma). Most importantly, they find a cross-correlation amplitude which is consistent with expectations, unlike other earlier works which found a low amplitude on large scales. This lends further support to the fact that earlier measurements were contaminated by tSZ, and that the novel cleaning procedure proposed here is indeed achieving its purpose. I also found footnote 7, providing a clear explanation of why tSZ contamination biases Clkg low, particularly useful. There is no doubt that in anticipation of future data (e.g. from ACTPol, Advanced ACT, DESI, Euclid, and so on) where CMB-LSS cross-correlations will be measured at high signal-to-ratio, new and powerful methods for removing foreground contamination will be necessary. The method presented in this week’s paper might very well be a first important step in the right direction.
#2 2003.12876: Screened fifth forces lower the TRGB-calibrated Hubble constant too by Harry Desmond and Jeremy Sakstein
As I wrote in many previous entries of this column (see e.g. my Week 2, Week 4, Week 5, Week 8, Week 9, and Week 12 posts), there is a general consensus that a solution to the H0 tension should include ether or both 1) new physics which lowers the sound horizon at recombination so as to infer a higher H0 from the inverse distance ladder, or 2) systematics in the local distance ladder which, once accounted for, lower the local H0. An interesting possibility, however, is that new physics might instead play the role of systematics which lead us to overestimate the local H0. An interesting possibility in this sense was explored in the earlier 1907.03778 by Desmond and Sakstein together with Bhuv Jain. The general idea is that the presence of a fifth force can shift the Cepheid period-luminosity relation (PLR) from the Newtonian relation in a different way even across the same set of Cepheids (which can happen if their masses or environmental densities have sufficient spread). If the calibration Cepheids are screened, whereas a fraction of the cosmological ones are not, this can bias the local distance ladder H0 high. However, Cepheids aren’t the only distance indicator which has been used to calibrate the local distance ladder. Another possibility is to use the tip of the red-giant branch (TRGB), an important feature in the Hertzprung-Russell diagram. In fact, a collaboration led by Wendy Freedman used this method to find H0 of about 69.8 in 1907.05922 (although Riess and collaborators replied in Yuan et al. 1908.00903 claiming that Freedman and collaborators had overestimated extinction in the Large Magellanic Cloud, and when this is corrected for, H0 goes up to about 72.5). The question is: can the same mechanism proposed by Desmond et al. in 1907.03778 lower the TGRB-calibrated H0 too?
Focusing on the revised TRGB-calibrated H0 of 1908.00903, since it sets a harder task for their model, Desmond and Sakstein show that the answer is indeed positive, provided that the Large Magellanic Cloud (LMC) is unscreened, and the fifth force is about 30%-50% as strong as Newtonian gravity (on the other hand, the Milky Way has to be screened). This is a very interesting finding which confirms that the earlier screened fifth force proposal made by Desmond et al. is definitely worth exploring further. If this mechanism truly underlies the H0 tension, there are a few consistency checks one can carry out. The mechanism requires the LMC (or precisely the red giants in the LMC) to be unscreened but the calibration Cepheids to be screened. If the whole LMC were uniformly screened, so that LMC Cepheids are as screened as LMC red giants, one would expect the LMC Cepheids-calibrated local H0 to be significantly lower. However, as Riess and collaborators showed in 1903.07603, this is not the case: this somehow challenges the fifth force explanation, unless one could argue that the degree of screening of Cepheids and red giants in the LMC is expected to be sufficiently different, in a natural way. In general, valuable information on the viability of Desmond and Sakstein’s proposal would come by comparing the value of H0 derived from different calibrators. Looking to the future, we expect future calibrations of the TRGB absolute I-band magnitude using Milky Way stars with parallaxes measured by Gaia. Since the Milky Way has to be screened, one could actually expect the so-calibrated local H0 to be lower than the inverse distance ladder value, turning the H0 tension completely around! This is a rather striking prediction, so I very much look forward to progress on the TRGB-calibration side which will enable tests of this very interesting scenario. In closing, it is worth noting that the scenario envisaged by Desmond and Sakstein does not trivially take care of the H0LiCOW high H0 measurements, and it is not clear whether it can do so at all (the interested reader should read the last paragraph of Section IV).
#3 2003.14107: A study on black hole shadows in asymptotically de Sitter spacetimes by Rittick Roy and Sayan Chakrabarti
Following the detection of the shadow of the supermassive black hole (BH) M87* by the Event Horizon Telescope, there has been a steady flow of papers revisiting the shadows of BHs beyond General Relativity (GR), or other exotic compact objects, in light of this detection (I myself contributed to this thread in 1904.12983, 1905.12421, and 1912.08231). Most of these papers focused on asymptotically flat spacetimes, mostly because this case is usually the easiest to treat. However, a perhaps even more motivated case is that where the spacetime is asymptotically de Sitter (i.e. has a positive cosmological constant), given we believe the Universe is accelerating! While this is not necessarily relevant for BHs which are close to us, it will be important for the future as it might become possible to image BHs at higher redshift, whose shadow we can expect to be affected by cosmic expansion. Computing the shadows of BHs in expanding Universes is not trivial, the two main reasons being that of finding an adequate description of a BH embedded, in an expanding universe, and next that of computing light ray trajectories in the strong gravity regime. The usual approach for computing the BH shadow with respect to a comoving observer consists of performing a McVittie transformation, and more importantly requires expressing the BH metric in isotropic form. However, a closed form for the isotropic transformation function is not always easy, or even possible, to obtain for a general BH solution in an arbitrary theory of gravity. The question then is: is it possible to do away with knowing the closed form solution for this isotropic transformation function, and still compute the shadow of a BH in an asymptotically de Sitter spacetime as seen by a comoving observer?
The answer, as Roy and Chakrabarti show in this week’s paper, is yes! The core of the approach consists in first computing the shadow (or more precisely its angular size) for a static observer, then using an aberration formula to relate it to the same quantity as viewed by a comoving observer. The aberration formula involves computing the velocity of one observer with respect to the other, which can be done once the metric functions are known. This simple procedure bypasses the need to know the exact McVittie type transformation of the BH spacetime. Roy and Chakrabarti then apply their procedure to study the shadows of some well-motivated BH solutions embedded in de Sitter space, namely the Reissner-Nordström-de Sitter BH and the 5-dimensional Gauss-Bonnet-de Sitter BH. The simple procedure presented here can be applied quite generally to other BH solutions in more than 4 dimensions. The only caveat is that the BH solution has to be spherically symmetric to begin with. This means that it is, in principle, not applicable directly to rotating (axially symmetric) BHs, for instance Kerr-de Sitter or Kerr-Newman-de Sitter BHs. I don’t know whether it is trivial to extend this procedure to cover these cases, but it is certainly something worth investigating.