How BAO measurements can fail to detect quintessence

We model the nonlinear growth of cosmic structure in different dark energy models, using large volume N-body simulations. We consider a range of quintessence models which feature both rapidly and slowly varying dark energy equations of state, and compare the growth of structure to that in a universe with a cosmological constant. The adoption of a quintessence model changes the expansion history of the universe, the form of the linear theory power spectrum and can alter key observables, such as the horizon scale and the distance to last scattering. The difference in structure formation can be explained to first order by the difference in growth factor at a given epoch; this scaling also accounts for the nonlinear growth at the 15% level. We find that quintessence models which feature late $(z<2)$, rapid transitions towards $w=-1$ in the equation of state, can have identical baryonic acoustic oscillation (BAO) peak positions to those in $\Lambda$CDM, despite being very different from $\Lambda$CDM both today and at high redshifts $(z \sim 1000)$. We find that a second class of models which feature non-negligible amounts of dark energy at early times cannot be distinguished from $\Lambda$CDM using measurements of the mass function or the BAO. These results highlight the need to accurately model quintessence dark energy in N-body simulations when testing cosmological probes of dynamical dark energy.