Modeled stress on the San Andreas and San Jacinto faults reached a 1,000-year high near Cajon Pass, a June study found

The highest stress value anywhere in a 1,000-year computer history of Southern California’s faults sits on the San Jacinto-Bernardino section, at 3.6 megapascals. The Mojave South stretch of the San Andreas comes in at 2.8 megapascals. Both numbers come from a simulation led by Liliane Burkhard of the University of Bern’s Physics Institute, and both describe the ground around Cajon Pass, the junction where the two fault systems nearly meet.

The result is a model output, not a gauge reading, and it dates to June 2026. The paper appeared on June 3 and Bern issued its release on June 8; a ScienceDaily repost on October 2 sent it around again, which is why the finding has been circulating as if it were new this week.

A four-dimensional earthquake cycle model

Burkhard’s team built what the University of Bern release describes as a physics-based earthquake cycle model of the San Andreas and San Jacinto faults, run across roughly 1,000 years of earthquake activity. The paper, published in the Journal of Geophysical Research: Solid Earth with Bridget Smith-Konter, Katherine Scharer and David Sandwell as co-authors, rests on 1,000 years of paleoseismic data and tracks how each earthquake shifts stress onto neighboring segments.

The conclusion is carefully scoped. In the words relayed by ScienceDaily’s summary of the release, tectonic stresses in parts of the system have reached, and in some places surpassed, the highest values found anywhere in the model’s 1,000-year history. That is a claim about the simulated record, and about parts of the system rather than every segment.

The numbers on each segment

The paper’s abstract gives Coulomb stress estimates by 2025 for three segments: 2.8 megapascals on the Mojave South segment of the San Andreas, 1.8 megapascals on the North San Bernardino segment, and 3.6 megapascals on the San Jacinto Bernardino segment. Only the last of these is named as the highest value in the whole simulation. The abstract also reports the fastest accumulation, about 1.8 megapascals per 100 years, north of Cajon Pass, with lower rates to the south.

The spread matters because the 1,000-year ceiling belongs to one section of the San Jacinto fault, while the San Andreas segment sits lower, at 2.8. Describing both faults as simultaneously at a record would overstate what the paper reports; the Bern release says instead that the two segments currently show similarly elevated stress.

Cajon Pass as an earthquake gate

Cajon Pass is where a rupture can stop or carry on. The release points to two historical cases: the 1857 Fort Tejon earthquake, magnitude 7.9, ended at the pass, while the 1812 Wrightwood earthquake ruptured through both systems. Burkhard says the decisive factor is not only how much stress has built up on a single fault but how aligned the stresses on the two fault systems are.

That alignment is what the model flags today. The release says the present configuration, with both segments similarly loaded, resembles the pattern that has historically come before joint ruptures with far greater consequences. A combined rupture would reach Los Angeles, San Bernardino, Riverside and the Coachella Valley. Burkhard describes the gate in terms of response rather than obstruction: Cajon Pass, she says, does not simply block or channel earthquakes but responds to the stress conditions on either side of it. The 1812 Wrightwood earthquake, which CNN puts at magnitude 7.5 and 40 deaths, is the case the team treats as the likely crossing, since it appears to have ruptured both systems at once.

CNN’s June 19 coverage carried the scenario size: locked segments capable of magnitude 6.7 or higher, and a joint rupture of magnitude 7.4 to 7.8. Matthew Weingarten of San Diego State University, who was not involved in the work, told the outlet that the insight is not that stress builds over time, which has long been known, but that the balance between the faults may decide whether earthquakes stay contained or expand.

The line the authors draw at prediction

Burkhard is explicit that the study “is not a prediction of when an earthquake will occur.” It offers a physics-based reading of stress conditions that planners can use. EarthSky’s June 9 account repeated the caveat and noted that the model reconstructs earthquake history from radiocarbon dating, tree-ring analysis and historical documents.

The paper is also a statement about method. Four-dimensional here means three spatial dimensions plus time: the model advances the faults through repeated earthquake cycles, so that stress on a segment today is the sum of everything that has slipped around it for a millennium. That is why the central number is a position inside a long record rather than a threshold: 3.6 megapascals is the largest value the simulation has produced, and the simulation covers only the span for which paleoseismic evidence could anchor it.

The unresolved part is timing. A stress ceiling in a 1,000-year simulation says nothing about which year a rupture comes, and Burkhard’s own framing is that emergency planning for a joint rupture should now be treated as realistic rather than worst-case.

This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.

More from Morning Overview