The rate at which the universe expands has been changing since the Big Bang, but this is not expected in the currently accepted model of cosmology. Still, it seems this discrepancy definitely exists and new Hubble observations confirm that the mismatch between the model and observations is very much real.

The team performed detailed observations of Cepheid variable stars, a special class of stars that have been used to estimate intergalactic distances. The team's work has strengthened those distance estimates and improved the value of the Hubble constant, the measurement of the expansion rate of the universe. The study has been accepted for publication in The Astrophysical Journal and is currently available on arXiv.

The value remains 9 percent higher than what's expected using data from the cosmic microwave background, radiation emitted 380,000 years after the Big Bang. The new measurement also shows that the likelihood that the discrepancy is just an error is now one in 100,000, a significant improvement on the one in 3,000 suggested by previous measurements.

“This mismatch has been growing and has now reached a point that is really impossible to dismiss as a fluke. This is not what we expected,” project leader Adam Riess, Nobel Laureate and Professor of Physics and Astronomy at Johns Hopkins University, said in a statement.

“This is not just two experiments disagreeing,” he explained. “We are measuring something fundamentally different. One is a measurement of how fast the universe is expanding today, as we see it. The other is a prediction based on the physics of the early universe and on measurements of how fast it ought to be expanding. If these values don’t agree, there becomes a very strong likelihood that we’re missing something in the cosmological model that connects the two eras.”

The exact mechanism that is making this happen is currently unclear. It might require some new physics. Researchers are interested in reducing the uncertainty surrounding the Hubble constant. This was 10 percent in 2001, 5 percent in 2009, and is 1.9 percent in the current study. Their goal is to bring this uncertainty down to 1 percent.

Cepheid variables are stars that pulsate with a well-defined stable period, which is linked to the star's true luminosity. This relationship was discovered by American astronomer Henrietta Swan Leavitt. With that value, it is possible to estimate the distance of these objects, and they have served as useful calibration tools for further distances obtained with supernovae. By expanding our understanding of Cepheid variables, we can improve all the distance estimates calibrated using these stars.