This would be great if true. Maybe they could come up with some way to predict earthquakes from this.
So...do you report that (for example) "San Francisco will be hit by an 8.5 earthquake in three hours"?
"Interesting yes. Quite a bit of prediction lore out here. I had not heard of this Lithosphere-atmosphere-ionosphere Coupling mechanism. It seems plausible. Although realize that predicting earthquakes is like curing cancer or the common cold, there are so many variations that it's essentially impossible to have a single solution. But you have to start somewhere. In this case, that is implying the release of radioactive gas, there needs to be a surface based detection to confirm that specific mechanism.
Also the proposed anomaly metrics in the paper needs to be accompanied by independent review for null hypothesis. The authors claim a 2-sigma anomaly but that is meaningless without that. But in principle since the authors seem to be using standard anomaly detection methods, if there is a nut in there, they might just find it. Doesn't look crackpot. If they can confirm any mechanism like this that would be a fantastic discovery. But yes it would still be some time before we could be sure of it's utility, eg not many false positives."
"On false positives, guy at Caltech here has story of predicting earthquake and telling disaster chief in San Fran, where it was predicted. Does chief now tell mayor to evacuate the city, or ignore it, or just happen to take the day off and drive his family for an out of town excursion?"