Cosmologists can now proclaim with confidence (but with some surprise too) that, in round numbers, our universe consists of 5% baryons, 25% dark matter, and 70% dark energy. It is indeed embarrassing that 95% of the universe is unaccounted for: even the dark matter is of quite uncertain nature, and the dark energy is a complete mystery.
The network of key arguments is summarised in Figure 1. Historically, the supernova evidence came first. But had the order of events been different, one could have predicted an acceleration on the basis of CDM evidence alone; the supernovae would then have offered gratifying corroboration (despite the unease about possible poorly-understood evolutionary effects).
Figure 1. The network of arguments that point towards a flat Universe dominated by 'dark energy'.
Our universe is flat, but with a strange mix of ingredients. Why should these all give comparable contributions (within a modest factor) when they could have differed by a hundred powers of ten?
In the coming decade, we can expect advances on several fronts. Physicists may well develop clearer ideas on what determined the favouritism for matter over antimatter in the early universe, and on the particles that make up the dark matter. Understanding the dark energy, and indeed the big bang itself, is perhaps a remoter goal, but ten years from now theorists may well have replaced the boisterous variety of ideas on the ultra-early universe by a firmer best buy. They will do this by discovering internal inconsistencies in some contending theories, and thereby narrowing down the field. Better still, maybe one theory will earn credibility by explaining things we can observe, so that we can apply it confidently even to things we cannot directly observe. In consequence, we may have a better insight into the origin of the fluctuations, the dark energy, and perhaps the big bang itself.
Inflation models have two generic expectations; that the universe should be flat and that the fluctuations should be gaussian and adiabatic (the latter because baryogenesis would occur at a later stage than inflation). But other features of the fluctuations are in principle measurable and would be a diagnostic of the specific physics. One, the ratio of the tensor and scalar amplitudes of the fluctuations, will have to await the next generation of CMB experiments, able to probe the polarization on small angular scales. Another discriminant among different theories is the extent to which the fluctuations deviate from a Harrison-Zeldovich scale-independent format (n = 1 in the usual notation); they could follow a different power law (i.e. be tilted) , or have a `rollover' so that the spectral slope is itself a function of scale. Such effects are already being constrained by WMAP data, in combination with evidence on smaller scales from present-day clustering, from the statistics of the Lyman alpha absorption-line `forest' in quasar spectra, and from indirect evidence on when the first minihalos collapsed, signalling the formation of the first Population III stars that ended the cosmic dark age.
In parallel, there will be progress in `environmental cosmology'. The new generation of 10-metre class ground based telescopes will give more data on the universe at earlier cosmic epochs, as well as better information on gravitational lensing by dark matter. And there will be progress by theorists too. The behaviour of the dark matter, if influenced solely by gravity, can already be simulated with sufficient accuracy. Gas dynamics, including shocks and radiative cooling, can be included too (though of course the resolution isn't adequate to model turbulence, nor the viscosity in shear layers). Spectacular recent simulations have been able to follow the formation of the first stars. But the later stages of galactic evolution, where feedback is important, cannot be modelled without parametrising such processes in a fashion guided by physical intuition and observations. Fortunately, we can expect rapid improvements, from observations in all wavebands, in our knowledge of galaxies, and the high-redshift universe.
Via a combination of improved observations, and ever more refined simulations, we can hope to elucidate how our elaborately structured cosmos emerged from a near-homogeneous early universe.