A new statistical analysis of type 1a supernovae observations has failed to find substantial statistical evidence that the rate of expansion of the universe has been increasing over time. Instead, the calculations are consistent with a universe that is expanding at a mostly constant rate – something that could be at odds with the popular lambda-cold dark matter (ΛCDM) model of cosmology.
Type 1a supernovae are exploding stars that play an important role in astronomy as “standard candles” that emit the same type and quantity of light. This means that the distance to a supernova can be worked out simply from its brightness in the sky.
Prior to the late 1990s, cosmologists had assumed that the expansion of the universe should either be constant over time, or slowing down. But then a team led by Saul Perlmutter and another team led by Adam Riess and Brian Schmidt noticed that the rate of expansion of the universe has been increasing. The teams found that more than 50 distant type 1a supernovae are fainter than expected for their measured redshift.
The expansion of the universe causes the light from a supernova to be shifted to longer wavelengths when observed on Earth. This redshift tells astronomers how quickly the supernova was moving away from us when the explosion occurred – which gives us the rate of the expansion of the universe at that time.
Surprise discovery
The surprise discovery was evidence that the expansion of the universe has been accelerating. It earned Perlmutter, Riess and Schmidt the 2011 Nobel Prize for Physics and led physicists to speculate that this acceleration was driven by an unseen entity called dark energy.
The evidence for accelerated expansion is marginal
Subir Sarkar, University of Oxford
Since then, further independent evidence for the accelerating expansion has come to light in measurements of the cosmic microwave background (CMB) and observations of galaxies. Indeed, the accelerating expansion of the universe has become a pillar of the most popular theory of cosmology, ΛCDM, where Λ is the cosmological constant that describes the acceleration.
Hundreds of other type 1a supernovae have been observed since the 1990s, but now some physicists are beginning to doubt whether these observations support an accelerating expansion. Subir Sarkar of the University of Oxford in the UK, Jeppe Nielsen of the Niels Bohr International Academy in Denmark and Alberto Guffanti of Italy’s University of Turin have done a statistical analysis of data from 740 type 1a supernovae and concluded “that the data are still quite consistent with a constant rate of expansion”.
Overly simple
The difference between the trio’s study and previous analyses is how variations in supernovae light are dealt with. While all type 1a supernovae are nearly identical, astrophysicists know that there are important differences that must be accounted for. Sarkar and colleagues argue that the statistical techniques adopted for previous studies are too simple and not appropriate for the growing set of observational data.
Using a technique that Sarkar describes as “industry standard statistics,” the trio took a different approach to dealing with variations in the supernovae. They concluded that the deviation from a constantly expanding universe is less than about 3σ, which is a relatively poor statistical significance. “The evidence for accelerated expansion is marginal,” says Sarkar, who believes that the ΛCDM model needs rethinking.
Roberto Trotta of Imperial College London does not go that far, pointing out that there is other independent and strong evidence for the accelerating expansion. However, he acknowledges that the evidence for acceleration in type 1a observations does not appear to be as robust as previously thought. Trotta – who has developed a new statistical method for analysing type 1a data that is different than Sarkar’s – says that astronomers are poised to observe thousands of new type 1a supernovae and must be prepared to adopt more rigorous statistical techniques to analyse them.
The analysis is described in Scientific Reports.
