At one time, astronomy professor Larry Molnar made a strong statement in 2017 – he and his team identified a binary star in the constellation Cygnus as an unambiguous candidate for a merger and explosion in the near future. Object KIC 9832227 is a pair of stars about 1800 light-years from Earth and has a narrow 11-hour orbit. This first-of-its-kind prediction caught the attention of an international audience and caused a stir in academia.
The interest prompted Molnar's peers to delve deeper into this issue and test the prediction. And now, 18 months later, a group of researchers led by Quentin Sossia, a graduate student at San Diego State University, published an article in The Astrophysical Journal Letters, which revises the predicted Molnar merger and says that it will not happen. And Molnar has already agreed with this assessment.
“Exact science makes reliable predictions,” Molnar said. “There were a few other studies that tried to change our design, but we could resist criticism. But this person really was able to defeat me, and I see some advantages in this. This shows that science is able to correct any judgment. '
Molnar's forecast was data bound. The binary orbit is oriented in such a way that the stars in turn eclipse each other from the point of view of the Earth. The predictions used the measured times of minimum light (mean eclipse) from all available sources. From 2013 to 2016, the Calvin Observatory was used for an extensive series of measurements. Archival measurements from other observatories have been found every year from 2007 to 2013. This was rounded off with one very early 1999 measurement from the Northern Hemisphere Variation Survey (NSVS). In addition, since the prediction was made public, new data from the Calvin Observatory that follows the predicted trajectory has been received over the course of the year.
However, Sosya and his team investigated observational gaps between 1999 and 2007 by analyzing previously unpublished archived data made in 2003 for NASA's Ames Vulcan project. They were surprised to find that eclipses occur half an hour later than expected from the merger hypothesis. This made them overestimate Molchan's estimates of the eclipse time. They have confirmed many values since 2007, but found that the 1999 NSVS was one hour late.
The inappropriate NSVS value was traced back to a typographical error in the document published to describe the 1999 data. The paper distorted the eclipse time by exactly 12 hours. This, in turn, brought Molnar's calculation to one orbit (11 hours) plus one hour. The change in the status of what was happening in the period from 1999 to 2003 changed the forecast itself. Thus, the agreement between last year's measurements and Molnar's published prediction should now be seen as a coincidence, not confirmation.
