The paradox arises from the fact that two French researchers, publishing this article on July 20 in the American journal Science, retrospectively analyzed data from GPS satellites at 5-minute intervals during the 48 hours preceding 90 major earthquakes that occurred over the past 20 years. They found that where the data was most accurate, tiny shifts in the tectonic plates could actually be detected: an average of two hours before the earthquake could be detected by seismographs. Currently, alerts are measured in minutes at most.
The downside is that the level of precision required is far beyond what is traditionally expected of GPS: in theory, it should be able to detect movement on the order of tenths of a millimeter.
History also teaches caution, seismologist Roland Bürgmann reminds us in a comment on the article: Since the 1970s, the past has been characterized by numerous attempts to discover the “signs” of earthquakes. The problem has always been that these warning signs could seldom be clearly distinguished from normal seismic “background noise”.
However, what Quentin Bletery and Jean-Mathieu Nocquet of the University of the Côte d’Azur in Nice propose – an analysis of the “slipping” of tectonic plates – could turn out to be a “signal” different from the rest. But if at the same time it turns out that the required precision is so high, then the technology for such detections is not yet within our reach …