A recent paleoseismic study used a new method to identify the timing of past earthquakes that occurred in the Monterey Bay area prior to the 1906 San Francisco earthquake.
By Jeng Hann Chong, Cal State University Northridge
Citation: Chong, J.H., 2020, Tree rings and statistics team up to date earthquakes, Temblor, http://doi.org/10.32858/temblor.100
Tree rings tell scientists a lot beyond just how old a tree was. They show the conditions the tree was subject to for every year of its life, including droughts and fires. Scientists can also use charcoal from trees to date old earthquakes but relying solely on the radiocarbon dating of charcoal of a long-lived tree can be messy. Now, researchers have come up with a new statistics-based method of dating charcoal samples in stratigraphic layers to estimate the ages of past earthquakes that gets around the challenges. Results from one California site indicate two large earthquakes occurred within 68 years of the 1906 San Francisco quake, but that the northern San Andreas Fault also went several hundred years without large earthquakes.
Studying prehistoric earthquakes
One common method to determine the ages of past earthquakes is radiocarbon dating. Paleoseismologists collect carbon-based materials like charcoal of burnt plants from different stratigraphic layers to determine the layers’ ages. The age of a past earthquake is determined based on the age of the uppermost stratigraphic layer that was cut by the fault.
However, charcoal dates can be misleading due to the inbuilt ages — the time since the charcoal was formed and buried, says Ashley Streig, a paleoseismologist at Portland State University and lead author of the new study in the Bulletin of the Seismological Society of America. Redwood trees are known for their long lifespan and this poses a problem to radiocarbon dating as the charcoal could originate from the inner part of the tree that does not reflect the actual age since it was burned down. The ages can be hundreds of years off, in fact, Streig and her team reported. So, Streig and her colleagues developed a new method that incorporates several statistical analyses to provide more accurate ages of prehistoric earthquakes, still using radiocarbon dating.
The team studied a specific section of the northern San Andreas Fault in the Santa Cruz Mountains called Hazel Dell. Streig had previously trenched the area and found that two distinct earthquakes occurred sometime between 1840 to 1906 and 1815 to 1895 along the same section of the fault as the 1906 San Francisco quake (Streig et al., 2014).
Wiggle-matching analysis
Streig and her colleagues used a model that employs a dating technique called “wiggle-matching” to a redwood tree stump excavated from the site. The wiggle-matching analysis is a method used to match the radioactive carbon in a sample to the atmosphere through time. This tree stump provides an improved understanding as to when the earthquakes happened because it narrows down the oldest possible age of the earthquakes. The team matched the ages of the growth rings from the tree stump to the atmospheric carbon with time and identified the age of the outermost bark to the late 1700s. This confirmed the 2014 findings.
Addressing inbuilt ages
Streig and her colleagues also ran a “lognormal trim” model to address the problem of having inbuilt ages. This model estimates the mean inbuilt ages using a probability density function and uses those ages to correct for an accurate age of the charcoal samples.
The two 1800s earthquake ages calculated using the lognormal trim model closely resemble this study’s dendroconstrained model that combines the wiggle-matching analysis with the ages of charcoal and macrofossils. The team pinpointed that one quake occurred in 1890 and the other in 1838 by comparing to written historical records from missions in the region.
The lognormal trim model also can be used at locations where detrital charcoal samples are available, Streig says. However, this model is not suitable for leaves, pinecones or grasses, she says, because grow and die each year.
Implications of earthquake ages
“Paleoseismology has its own challenges [that are] brought up nicely in this paper,” says Katherine Scharer, a paleoseismologist with the U.S. Geological Survey who was not involved in this study. The authors did a careful job of using the wiggle-matching as a matching tool to compare with the ages from the charcoal samples, Scharer says.
“For the Santa Cruz Mountain section of the [San Andreas] Fault, we are certain that it ruptured post-European settlement in the area,” Streig says. Her team’s models showed that three large earthquakes occurred within 68 years (in 1838, 1890 and 1906). The models also showed that this section of the San Andreas went several hundred years without large earthquakes, which calculated a lowered average recurrence interval of the last four earthquakes to about 150 years.
If these models are correct, Streig says, the earthquake record may be younger along the northern San Andreas Fault. However, she adds, a longer earthquake record along this section of the fault is needed to better understand the behavior of earthquake rupture.
“This is one of the first studies that give a perspective on quantitative inbuilt ages,” says Scott Bennett, a paleoseismologist at the U.S. Geological Survey who was not part of this study. This study shows that there is a possibility that previous paleoseismic studies in these redwood forested areas will have to be reevaluated, Bennett adds.
Check your earthquake risk at Temblor.
Further Reading
Streig, A. R., T. E. Dawson, and R. J. Weldon II (2014). Paleoseismic evidence of the 1890 and 1838 earthquakes on the Santa Cruz Mountains section of the San Andreas Fault, near Corralitos, California, Bull. Seismol. Soc. Am. 104, 285–300, doi: 10.1785/ 0120130009
Streig, A. R., R. J. Weldon, II, G. Biasi, T. E. Dawson, D. G. Gavin, and T. P. Guilderson (2020). New Insights into Paleoseismic Age Models on the Northern San Andreas Fault: Charcoal Inbuilt Ages and Updated Earthquake Correlations, Bull. Seismol. Soc. Am. XX,1–13, doi: 10.1785/0120190307