False Earthquake Alerts Hit Northern California Again in 2026

The federal earthquake early warning system has misfired again, and this time it did so repeatedly.

For the second time in four months, the U.S. Geological Survey’s automated earthquake detection system triggered false alerts across Northern California, generating reports of multiple phantom quakes on Monday night. Unlike December’s predawn alarm that jolted Bay Area cellphones awake, these alerts did not push loud ShakeAlert warnings to residents. But the pattern is becoming harder to ignore.

According to reporting from Bay Area news outlets, the false alerts Monday spanned a wide geographic swath. Bogus quake reports appeared for locations offshore west of the Bay Area, near Bonny Doon in Santa Cruz County, near Point Reyes, near Yreka, and near Shaver Lake in Fresno County. The fabricated magnitudes ranged from 3.2 to 3.8. The Sacramento Bee flagged one report, relayed by the CA Earthquake Bot, showing a preliminary magnitude 3.7 quake supposedly striking 53 miles west-southwest of Bonny Doon at 8:48 p.m. Monday. Check the USGS earthquake site now, and that quake simply does not exist. The only Bay Area seismic activity logged in the relevant period was a real 2.8-magnitude event just after midnight near San Ramon, part of an ongoing earthquake swarm that began there last fall.

The USGS acknowledges its automated detection system can produce erroneous readings due to signal interference or processing errors. The agency describes such errors as rare. Monday’s cluster of false alerts in a single evening tests that characterization.

The December incident was more alarming in a literal sense. At 8:06 a.m., ShakeAlert pushed a “Critical” earthquake warning to subscribed cellphones, complete with the blaring alarm sound reserved for genuine threats. The reported quake: a 5.9 magnitude centered in Lyon County, Nevada, well over 200 miles from the Bay Area. Even a real 5.9 would have been unlikely to produce meaningful shaking here. The USGS retracted the alert quickly and posted an apology on X, acknowledging there was no M5.9 earthquake and promising to investigate why the alerts went out.

Christie Rowe, director of the Nevada Seismological Laboratory at the University of Nevada, Reno, later identified a likely culprit: a faulty power system at one seismic monitoring station in Nevada. That single malfunction apparently cascaded into data transmission problems across the entire ShakeAlert network.

The December false alarm drew sharp political reaction within days. Representative Kevin Mullin joined four other Democratic House members in writing to the USGS, warning that public trust in the agency’s real-time earthquake information is essential and that the false alarm threatened it. Governor Gavin Newsom used the episode to highlight what he described as damage from the Trump administration’s cuts to federal agencies, including the USGS.

That political pressure has not produced visible results. Monday’s multi-location false alert suggests the underlying technical vulnerabilities remain unaddressed.

This matters beyond the annoyance factor. ShakeAlert exists to give Californians seconds of warning before shaking arrives. Those seconds can mean the difference between someone bracing in a doorway or stepping into a street. The system’s value is inseparable from its credibility. Every false alarm chips at the public’s willingness to act on a real warning when one comes, and in the Bay Area, a real warning will come.

San Francisco residents already have complicated feelings about earthquake preparedness. People who lived through Loma Prieta carry the memory of how fast everything changed. Younger residents have grown up with preparedness campaigns and retrofit programs but also with a general sense that the city is perpetually not quite ready. A warning system that cried wolf twice in four months does not build the confidence this region needs.

The USGS owes the public a clear accounting of what failed Monday night, what systemic fixes are underway following December’s incident, and what the agency’s timeline looks like for shoring up the network. Federal budget pressures make that last question especially pointed. A degraded monitoring station in Nevada already disrupted an entire regional alert system once. Funding gaps elsewhere in the network could produce the same result.

San Francisco Download has submitted a records request to the USGS seeking documentation of both incidents and any internal review findings.