The discovery of the accelerating universe in the late 1990s was a watershed moment in modern cosmology, as it indicated the presence of a fundamentally new, dominant contribution to the energy budget of the universe. Evidence for dark energy, the new component that causes the acceleration, has since become extremely strong, owing to an impressive variety of increasingly precise measurements of the expansion history and the growth of structure in the universe. Still, one of the central challenges of modern cosmology is to shed light on the physical mechanism behind the accelerating universe. In this review, we briefly summarize the developments that led to the discovery of dark energy. Next, we discuss the parametric descriptions of dark energy and the cosmological tests that allow us to better understand its nature. We then review the cosmological probes of dark energy. For each probe, we briefly discuss the physics behind it and its prospects for measuring dark energy properties. We end with a summary of the current status of dark energy research.

We present an analysis of 15 Type Ia supernovae (SNe Ia) at redshift z > 1 (9 at 1.5 < z < 2.3) recently discovered in the CANDELS and CLASH Multi-Cycle Treasury programs using WFC3 on the Hubble Space Telescope. We combine these SNe Ia with a new compilation of 1050 SNe Ia, jointly calibrated and corrected for simulated survey biases to produce accurate distance measurements. We present unbiased constraints on the expansion rate at six redshifts in the range 0.07 < z < 1.5 based only on this combined SN Ia sample. The added leverage of our new sample at z > 1.5 leads to a factor of ~3 improvement in the determination of the expansion rate at z = 1.5, reducing its uncertainty to ~20%, a measurement of H(z=1.5)/H0=2.67 (+0.83,-0.52). We then demonstrate that these six measurements alone provide a nearly identical characterization of dark energy as the full SN sample, making them an efficient compression of the SN Ia data. The new sample of SNe Ia at z > 1 usefully distinguishes between alternative cosmological models and unmodeled evolution of the SN Ia distance indicators, placing empirical limits on the latter. Finally, employing a realistic simulation of a potential WFIRST SN survey observing strategy, we forecast optimistic future constraints on the expansion rate from SNe Ia.

Peculiar velocities of objects in the nearby universe are correlated due to the gravitational pull of large-scale structure. By measuring these velocities, we have a unique opportunity to test the cosmological model at the lowest redshifts. We perform this test, using current data to constrain the amplitude of the "signal" covariance matrix describing the velocities and their correlations. We consider a new, well-calibrated "Supercal" set of low-redshift SNe Ia as well as a set of distances derived from the fundamental plane relation of 6dFGS galaxies. Analyzing the SN and galaxy data separately, both results are consistent with the peculiar velocity signal of our fiducial $\Lambda$CDM model, ruling out the noise-only model with zero peculiar velocities at greater than $7\sigma$ (SNe) and $8\sigma$ (galaxies). When the two data sets are combined appropriately, the precision of the test increases slightly, resulting in a constraint on the signal amplitude of $A = 1.05_{-0.21}^{+0.25}$, where $A = 1$ corresponds to our fiducial model. Equivalently, we report an 11% measurement of the product of the growth rate and amplitude of mass fluctuations evaluated at $z_\text{eff} = 0.02$, $f \sigma_8 = 0.428_{-0.045}^{+0.048}$, valid for our fiducial $\Lambda$CDM model. We explore the robustness of the results to a number of conceivable variations in the analysis and find that individual variations shift the preferred signal amplitude by less than ${\sim}0.5\sigma$. We briefly discuss our Supercal SN Ia results in comparison with our previous results using the JLA compilation.

We revisit the effect of peculiar velocities on low-redshift type Ia supernovae. Velocities introduce an additional guaranteed source of correlations between supernova magnitudes that should be considered in all analyses of nearby supernova samples but has largely been neglected in the past. Applying a likelihood analysis to the latest compilation of nearby supernovae, we find no evidence for the presence of these correlations, although, given the significant noise, the data is also consistent with the correlations predicted for the standard $\Lambda$CDM model. We then consider the dipolar component of the velocity correlations -- the frequently studied "bulk velocity" -- and explicitly demonstrate that including the velocity correlations in the data covariance matrix is crucial for drawing correct and unambiguous conclusions about the bulk flow. In particular, current supernova data is consistent with no excess bulk flow on top of what is expected in $\Lambda$CDM and effectively captured by the covariance. We further clarify the nature of the apparent bulk flow that is inferred when the velocity covariance is ignored. We show that a significant fraction of this quantity is expected to be noise bias due to uncertainties in supernova magnitudes and not any physical peculiar motion.

Late-time power law expansion has been proposed as an alternative to the standard cosmological model and shown to be consistent with some low-redshift data. We test power law expansion against the standard flat $\Lambda$CDM cosmology using goodness-of-fit and model comparison criteria. We consider Type Ia supernova (SN Ia) data from two current compilations (JLA and Union2.1) along with a current set of baryon acoustic oscillation (BAO) measurements that includes the high-redshift Lyman-$\alpha$ forest measurements from BOSS quasars. We find that neither power law expansion nor $\Lambda$CDM is strongly preferred over the other when the SN Ia and BAO data are analyzed separately but that power law expansion is strongly disfavored by the combination. We treat the $R_\text{h} = ct$ cosmology (a constant rate of expansion) separately and find that it is conclusively disfavored by all combinations of data that include SN Ia observations and a poor overall fit when systematic errors in the SN Ia measurements are ignored, despite a recent claim to the contrary. We discuss this claim and some concerns regarding hidden model dependence in the SN Ia data.

We develop a general method to "self-calibrate" observations of galaxy clustering with respect to systematics associated with photometric calibration errors. We first point out the danger posed by the multiplicative effect of calibration errors, where large-angle error propagates to small scales and may be significant even if the large-scale information is cleaned or not used in the cosmological analysis. We then propose a method to measure the arbitrary large-scale calibration errors and use these measurements to correct the small-scale (high-multipole) power which is most useful for constraining the majority of cosmological parameters. We demonstrate the effectiveness of our approach on synthetic examples and briefly discuss how it may be applied to real data.

Some recent observations provide $> 2\sigma$ evidence for phantom dark energy -- a value of the dark energy equation of state less than the cosmological-constant value of $-1$. We focus on constraining the equation of state by combining current data from the most mature geometrical probes of dark energy: Type Ia supernovae (SNe Ia) from the Supernova Legacy Survey (SNLS3), the Supernova Cosmology Project (Union2.1), and the Pan-STARRS1 survey (PS1); cosmic microwave background measurements from Planck and WMAP9; and a combination of measurements of baryon acoustic oscillations. The combined data are consistent with $w = -1$ for the Union2.1 sample, though they present moderate ($\sim 1.9\sigma$) evidence for a phantom value when either the SNLS3 or PS1 sample is used instead. We study the dependence of the constraints on the redshift, stretch, color, and host galaxy stellar mass of SNe, but we find no unusual trends. In contrast, the constraints strongly depend on any external $H_0$ prior: a higher adopted value for the direct measurement of the Hubble constant ($H_0 \gtrsim 71~\text{km/s/Mpc}$) leads to $\gtrsim 2\sigma$ evidence for phantom dark energy. Given Planck data, we can therefore make the following statement at $2\sigma$ confidence: either the SNLS3 and PS1 data have systematics that remain unaccounted for or the Hubble constant is below 71 km/s/Mpc; else the dark energy equation of state is indeed phantom.

We study the effects of current systematic errors in Type Ia supernova (SN Ia) measurements on dark energy (DE) constraints using current data from the Supernova Legacy Survey (SNLS). We consider how SN systematic errors affect constraints from combined SN Ia, baryon acoustic oscillations (BAO), and cosmic microwave background (CMB) data, given that SNe Ia still provide the strongest constraints on DE but are arguably subject to more significant systematics than the latter two probes. We focus our attention on the temporal evolution of DE described in terms of principal components (PCs) of the equation of state, though we examine a few of the more common, simpler parametrizations as well. We find that the SN Ia systematics degrade the total generalized figure of merit (FoM), which characterizes constraints in multi-dimensional DE parameter space, by a factor of two to three. Nevertheless, overall constraints obtained on more than five PCs are very good even with current data and systematics. We further show that current constraints are robust to allowing for the finite detection significance of the BAO feature in galaxy surveys.