Image by David H. Lipp/US Air National Guard

Predicting the future trajectory and death toll of the coronavirus pandemic has been difficult, especially with different teams of researchers interpreting different data sets in myriad ways.

That’s why scientists at University of Massachusetts Amherst built an ensemble model that combines eight of the most reputable coronavirus projections, WNYC reports. Unfortunately, as the model grows more refined, it’s narrowing in on a devastating prediction: COVID-19 will kill 110,000 people in the U.S. by June 6.

As of this story’s publication on May 13, there have been 82,391 coronavirus-related deaths in the U.S. — a number that makes up roughly 28 percent of the global death toll. That means that in just about three weeks, the numbers suggest that the coronavirus will claim almost 30,000 more lives in the U.S.

These models are predictions based on existing data — there’s no guarantee they’ll hold up, and policy decisions could help or hinder the public death and suffering. But researchers told WNYC that the ensemble model will be far more robust than any individual one, lending credibility to the grim outlook.

“Individual models are being changed every week. They’re sensitive to the last observed data in different ways,” Nicholas Reich, the Amherst biostatistician who created the model, told WNYC. He added that with an ensemble “there’s a certain consistency and robustness. You’re not quite, sort of flapping in the wind.”