End-of-the-week synopses looking at the possibility of directly detecting dark energy fluctuations using anisotropies in gravitational wave luminosity distances, constraints on baryonic effects from the DES Y1 data, and revisiting constraints on the CMB temperature and whether the latter might have some bearings on the H0 tension. A heads-up that this will be the last entry for the next 5 weeks, as I’m going on holiday. Enjoy, and see you again here in mid-September!
#1 2007.13722: Detecting Dark Energy Fluctuations with Gravitational Waves by Alice Garoffolo et al.
Can we directly detect dark energy (DE) fluctuations? This feat is in some way the holy grail of cosmological searches for DE, which so far have only focused on indirect effects of DE on the clustering or growth of the large-scale structure. One common feature of DE models for cosmic acceleration beyond the cosmological constant (including modified gravity models) is the fact that the luminosity distance one infers from standard sirens such as gravitational waves (GWs) differs from that inferred from standard candles such as Type Ia Supernovae (SNeIa), both at the background level and at the perturbation level. This has been studied in a number of recent papers (such as 1712.08108 and 1912.08093), and opens up the possibility of using the anisotropies in electromagnetic (EM) luminosity distances as a cosmological probe, as studied in a number of recent papers including 1908.08951. This immediately leads to two follow-up questions: 1) can one do the same with GW luminosity distances? 2) can one combine GW and EM luminosity distance anisotropies to constrain DE models, exploiting the fact that in many such models the two are different as I wrote above?
In this week’s paper, Garoffolo and collaborators essentially address these questions, focusing on two specific DE models: a so-called designer f(R) model reproducing the LCDM background, and a generalized Brans-Dicke model (essentially a time-varying Newton’s constant). They first compute the angular power spectrum of the GW luminosity distance anisotropies, finding that the contribution of the scalar field clustering component - in other words the direct contribution of DE fluctuations - is too small to be detected in GW luminosity distance anisotropy auto-correlations alone. For this reason they turn to the construction of a joint estimator which makes us of GW and EM luminosity distance anisotropies, with the EM signal provided by SNeIa. The joint estimator given in Eq.(4) essentially tries to exploit the fact that GW and EM luminosity distances generally differ, to extract the direct contribution of DE fluctuations in a statistical sense. Garoffolo and collaborators find this to be possible in principle, although the cost to pay is a huge number of required GW events. For a sense of how huge these numbers are, have a look at Table I. These numbers are at the very least “ambitious” :) as the authors themselves write, but there are certainly avenues for improvements, as hinted to in the final paragraph of the paper. I guess this leaves ample opportunities for further exploring this very interesting avenue for direct detection of DE.
Overall, this was a very interesting paper which proposes a cool way of at least trying to directly detect DE. As the second author Marco Raveri (who already featured in Week 9 entry) wrote on Facebook, it is rather funny that a paper describing one of the very few ways of directly detecting Dark Energy goes as GR (Garoffolo-Raveri) et al.: I couldn’t agree more! It’s worth pointing out that another interesting paper with a related spirit by Natalie Hogg et al. also appeared this week: 2007.14355 (I decided not to cover both papers in order to increase the diversity of topics, and therefore resolved to cover the more ambitious proposal). As a side self-publicity note, another ambitious avenue for cosmological direct detection of DE is if DE couples to visible matter through interactions which are unscreened at the cosmological level. This is something we studied in 1911.12374 (in fact, our v1 title was much cooler, but the editor didn’t like it!), and much as this week paper’s findings, we found that directly detecting DE fluctuations in this way requires, to say the least, very optimistic assumptions. Unfortunately I don’t know what to make of the initials of the first authors VV et al.! Somewhat more optimistic results, however, were obtained in 2004.14661 (which I covered in my Week 18 entry). While it looks like directly detecting DE is awfully hard, I seriously hope that one day we will be able to do it.
#2 2007.15026: Dark Energy Survey Year 1 Results: Constraining Baryonic Physics in the Universe by the DES collaboration
With big data come big responsibilities. As we slowly enter into the era of sub-% precision cosmology, issues which were earlier just optional nuisances (of the “we’ll leave it for future work“ type) are now concerns we can no longer ignore, if we want to fully exploit the wealth of cosmological information encoded in state-of-the-art surveys. For example, we need to be able to predict the matter power spectrum to the ~1% level out to wavenumbers k~10 h*Mpc^-1 for upcoming Stage IV experiments. On these scales, the linear power spectrum is a horrible approximation to reality, and it becomes imperative to model what is usually referred to via the umbrella term “baryonic physics/feedback/(g)astrophysics“. These are baryonic effects operating in galaxies and redistributing or in many cases blowing out matter, and causing uncertainties in the matter power spectrum at the level of 10%. There are many ways to deal with baryonic physics. The most conservative strategy is simply to discard all modes where baryonic effects are important, or in other words to not use small-scale information (I don’t think this would make funding agencies very happy). This is what the Dark Energy Survey (DES) collaboration did in their year 1 (Y1) analysis in 1708.01538. While this is a very conservative and safe approach, it comes at the cost of excluding many (most?) data points, which still contain a huge deal of untapped cosmological information. Another more challenging possibility, which the community has slowly been moving towards in recent years (see for instance the excellent review 1905.06082) is that of modelling baryonic effects, informed by complex hydrodynamical simulations, semi-analytical prescriptions, and whatnot. There is some disagreement across simulations as to the form and magnitude of these effects, but also a general consensus that by now we can’t do without modelling baryonic effects.
In this week’s paper, led by Hung-Jin Huang, the DES collaboration takes important steps in this sense. In particular, they incorporate state-of-the-art baryonic effects (informed by various simulations) into their so-called 3×2pt analysis, which makes use of cosmic shear, galaxy clustering, and cross-correlations thereof. Since different simulations do not always agree on the form and magnitude of baryonic effects (in some cases simply because the processes simulated are not the same!), a flexible way of spanning the space of baryonic effects is required, and the DES collaboration does it via a principal component (PC) analysis of summary statistics extracted from these simulations - you can see the first three PCs in Fig. 2). The DES collaboration then tries to jointly constrain cosmology and baryonic effects. More precisely, what is constrained is Q1, the amplitude of the first PC, as the use of only one PC is found to be sufficient. Combining DES data with Planck and BAO data (without temperature information from Planck, due to a possible tension with DES), the collaboration finds Q1~1.42. To someone with little or no knowledge of baryonic effects such as myself, this number doesn’t mean much, which makes Fig. 15 very illuminating in this sense! This value of Q1 indicates a preference for weak to moderate baryonic effects, and excludes one of the most extreme AGN feedback hydrodynamical scenario, the so-called cosmoOWLS-AGN scenario with minimum AGN heating temperature of 10^8.7 K, at ~2.8σ. I think this is a very important paper as it shows that already with current data (with all due caveats) we can start to probe which prescriptions for baryonic scenarios best describe reality, which in turn can aid and inform the design of survey analyses pipelines for upcoming experiments.
#3 2007.13789: Is there evidence for a hotter Universe? by Carlos Bengaly, Javier Gonzalez, and Jailson Alcaniz
If you have been following my blog regularly you probably remember the paper 2005.10656 by Ivanov et al. which tried to recast the H0 tension into a T0 tension, with T0 the monopole CMB temperature. I covered this paper in my Week 21 entry. Now, if you recall the conclusions, it is basically impossible to solve the H0 tension invoking a colder CMB, because of both the exquisite sub-% level measurements of T0 from COBE/FIRAS in astro-ph/9605054, and the important role of BAO measurements in breaking the otherwise present H0-T0 degeneracy (another way of rephrasing the geometrical degeneracy in this context). Intriguingly, the later 2006.16149 by Bose and Lombriser instead tried to solve the H0 tension (and other tensions including the anomalous lensing amplitude inferred by Planck temperature data alone, the sigma8 tension, and the low CMB quadrupole moment) by invoking a spatially open and hotter Universe. I’ll admit I haven’t read this paper in detail to have a good sense of whether it is feasible or not, but I’ll give it the benefit of the doubt. In any case, the point is that regardless of whether a colder or a hotter CMB can have some bearings on the H0 tension, it is interesting and timely to consider whether measurements of the CMB temperature other than COBE/FIRAS are in agreement with the latter.
In this week’s paper, Bengaly and collaborators use currently available measurements of T(z), the temperature of CMB photons as a function of redshift, to determine T0 and assess its consistency with COBE/FIRAS. They use various (more than 100) measurements of T(z) up to redshift ~1, all of them coming from SZ measurements. In a first instance, the determination of T0 is performed parametrically, using the well-known relation seeing T(z) scaling as (1+z) in the absence of cosmic opacity or photon non-conservation (again, there is some relation to the paper by Hogg et al. 2007.14355 I mentioned above). The resulting T0 is in excellent agreement with COBE/FIRAS. In a second instance, a non-parametric Gaussian Processes regression reconstruction of T(z) is performed, but the result is virtually unchanged, with T0 still in excellent agreement with COBE/FIRAS. Interestingly, as shown in Fig. 1, the resulting reconstructed T0 is in moderate tension with both the colder T0 suggested by Ivanov et al. (IAL20 in the plot) and the hotter T0 suggested by Bose and Lombriser (BL20 in the plot). While invoking a different T0 to address the H0 tension is an appealing idea, as it would not require new physics, I suspect that T0 has little or nothing to say concerning the H0 tension. This is why I really liked this week’s paper which, while performing a relatively simple analysis, confirms this suspicion.