Tuning the cosmic instrument: robust cosmology through combined probes
Abstract: As wide-field surveys yield increasingly precise data, multiprobe analyses offer significant advantages. In this work, we use our previously developed framework for jointly analyzing cosmic microwave background (CMB) and large-scale structure data. We analyze combinations of three CMB (Planck PR3, Planck PR4, and ACT+WMAP) datasets, DESI Y1 Baryon Acoustic Oscillation (BAO) data, and a $9\times 2$pt low-$z$ dataset comprising KiDS-1000, BOSS DR12, and Planck CMB lensing/Integrated Sachs Wolfe (including all cross-correlations). We first assess internal consistency, finding a mild ($<2\sigma$) tension between CMB and low-$z$ datasets in the full parameter space and hints of systematics in Planck PR3 and KiDS-1000. We then derive constraints in $\Lambda\mathrm{CDM}$ and, motivated by recent DESI results, dynamical dark energy ($w_0w_a\mathrm{CDM}$) and free neutrino mass extensions. In $\Lambda \mathrm{CDM}$, we derive a novel $9\times2$pt constraint of $S8=0.777{+0.17}_{-0.17}$ and find strong consistency among CMB datasets. In $w_0w_a\mathrm{CDM}$, adding low-$z$ to CMB+BAO tightens $(w_0,w_a)$ constraints by 50\% (in figure-of-merit terms) in our baseline combination of Planck PR4 + low-$z$ + BAO. The posterior accommodates a cosmological constant ($w_0 = -1, w_a = 0$) within $1\sigma$, in contrast to the $\sim2\sigma$ preference for evolving dark energy from CMB+BAO alone. For neutrino masses, our baseline dataset yields a systematics-robust constraint of $M_\nu<0.12\mathrm{eV}$ in $\nu\Lambda\mathrm{CDM}$. Allowing dynamical dark energy and free neutrino mass ($\nu w_0w_a\mathrm{CDM}$) broadens and shifts the neutrino mass posterior higher, yielding a $1.8\sigma$ constraint ($M_\nu=0.16{+0.09}_{-0.09}\mathrm{eV}$) in our baseline. Our analysis demonstrates the power of multiprobe analyses for assessing tensions, identifying systematics and providing robust constraints.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.