background (CMB) using ultrasensitive radio telescopes on the ground, balloons, and spacecraft. With a combination of these and other observations, astrophysicists have shown that the geometry of space is approximately flat, that the age of the universe is 13.7 billion years, and that there is nearly five times as much matter in a dark, invisible form as in normal matter that can turn into visible stars. The past decade also saw strong affirmation of the remarkable discovery that the expansion of the universe is accelerating.

We can now say that there is a ubiquitous and ethereal substance called dark energy that is expanding the fabric of space between the galaxies at ever faster speeds and that accounts for 75 percent of the mass-energy of the universe today. The effects are so tiny on the scale of an experiment on Earth that the only way forward is to use the universe at large as a giant laboratory.

Two complementary approaches to understanding dark energy have been considered by this survey: one on the ground and the other in space. On the ground, the proposed LSST would provide optical imaging of brighter galaxies over half the sky every few days. It would build up measurements of galaxy images that are distorted by (weak) gravitational lensing and detect many relatively nearby super-novae. From space, the proposed WFIRST would produce near-infrared images of fainter galaxies over smaller areas and observe distant supernovae. It would also provide near-infrared spectroscopy for sensitive baryon acoustic oscillation measurements. What has become clear over the past few years is that instead of just considering dark energy in different regimes, LSST and WFIRST will actually be quite synergistic, and observations from one are essential to interpreting the results of the other. In particular, by working together, they would provide the powerful color information needed for redshift¹¹ estimation. The properties of dark energy would be inferred from the measurement of both its effects on the expansion rate and its effects on the growth of structure (the pattern of galaxies and galaxy clusters in the universe). In doing so it should be possible to measure deviations from a cosmological constant¹² larger than about a percent. Massively multiplexed spectrographs in intermediate-class and large-aperture ground-based telescopes would also play an important role.

Second, and most remarkably, it is now possible to contemplate observing the earliest moments of the universe. Another source of gravitational radiation may be the most intriguing of all. The patterns in the CMB are theoretically consistent with what could have been laid down during the first instants after the big bang during an