Mystery Gap: Connecting Earthquake Faults near San Francisco, California, Requires Many Approaches
San Pablo Bay, a northern extension of California’s San Francisco Bay, hides its secrets well. For decades, scientists have speculated about whether two of the region’s earthquake faults—the Hayward and Rodgers Creek faults—connect beneath the bay.
Now, USGS scientists have arrived at an answer by combining several lines of evidence—detailed images of sediment layers just below the bay floor, measurements of the density and magnetization of sediment and rock deep beneath the bay, and modeling of fault deformation. They report their findings in “Missing link between the Hayward and Rodgers Creek faults,” published October 19, 2016, in Science Advances.
Why does it matter?
The longer the stretch of fault that breaks during an earthquake, the stronger the quake. When two faults are close to one another, the earthquake can jump from one to the other, making the rupture longer and the shaking stronger. When two faults are directly connected, it’s even easier for earthquake rupture to continue from one fault to the next.
A break along the combined length of the Hayward and Rodgers Creek faults could produce a major earthquake of magnitude 7.4. That earthquake would release more than five times the energy released by the 1989 magnitude 6.9 Loma Prieta earthquake, which caused about $6 billion in damage and killed 63 people.
To estimate the earthquake hazard posed by the Hayward and Rodgers Creek faults, scientists need to understand whether and how the faults connect. Previous work showed that the two faults approach each other closely beneath San Pablo Bay, but their exact relationship remained a mystery.
A difficult place to map
Mapping with sound energy from large research vessels has illuminated details about many faults hidden underwater. (For example, see “Striking New Seafloor Imagery of the Queen Charlotte-Fairweather Fault in the Gulf of Alaska,” this issue.) But San Pablo Bay poses challenges for such mapping: The bay is mostly less than 2 meters (7 feet) deep—too shallow for most research vessels equipped with acoustic mapping systems. What’s more, widespread natural gas in the bay sediment scatters sound energy, making it difficult to image more than a few meters below the bay floor.
USGS researchers approached the first problem by towing a “chirp subbottom profiler” from a shallow-draft boat. The chirp profiler releases pulses of high-frequency sound (“chirps”) that penetrate the seafloor and bounce off sediment layers below. The profiler then records the sound that bounces back. Computer processing turns the echoes into a vertical view of the layers, called a seismic-reflection profile. This method is similar to doctors using ultrasound to see what’s inside the human body.
The high-frequency chirps allowed the system to detect small vertical bends and breaks in the otherwise flat sediment layers beneath San Pablo Bay. (See the chirp seismic-reflection profile, below.) These subtle deformations told the researchers that they were seeing an active fault.
The fault strand they had discovered appeared to be an extension of the Hayward fault, but it was not following the path they had expected.
“Where the faults enter San Pablo Bay, the Hayward from the south and the Rodgers Creek from the north, their orientations suggest that they’ll run parallel to one another, separated by a space, or ‘stepover,’ of about 5 kilometers [3 miles],” said Janet Watt, USGS research geophysicist and lead author of the study. “So when we went out to map, we thought we were going to map details of the stepover—like minor fractures that might enable an earthquake rupture to cross from one fault to the other.”
As the data accumulated, however, the researchers saw that the Hayward fault strand they were mapping bends slightly to the right as it traverses San Pablo Bay, heading toward the Rodgers Creek fault.
They realized that they might be looking at a fault bend. The distinction is important because fault bends affect earthquake hazards differently than stepovers. These differences include how likely a break on one fault will continue on to the next, how much slip (movement of rock on either side of the fault) will occur, and where ground shaking will be strongest.
Natural gas pockets in the bay mud prevented imaging below about 5 meters (16 feet) beneath the bay floor. The researchers wanted to look deeper to determine if the active fault strand they had discovered really is part of a major fault.
Gravity and magnetism show the fault goes deep
The force of gravity at Earth’s surface varies depending on the density of rocks below. These gravity variations are too tiny for humans to feel, but sensitive instruments can measure them. Watt and her team compiled measurements collected earlier by the USGS to make a gravity map of San Pablo Bay (below, left). Dense metamorphic rocks west of the fault create higher gravity values; lower gravity values to the east come from sedimentary rocks. A sharp change in gravity values lies close to locations where seismic-reflection profiles revealed the Hayward fault.
The rocks beneath our feet also have their own tiny magnetic fields. In general, differences in the strength of rock magnetization reflect differences in the abundance of magnetite, a common mineral. (Magnetite grains often form dark streaks in California’s beach sand.)
To measure changes in the magnetic field beneath San Pablo Bay, USGS coauthor David Ponce mounted a magnetometer on a pole extending from the bow of the boat. From those measurements, he constructed a map of magnetic field variations (below, right). In general, warm colors indicate a stronger magnetic field (more magnetite in underlying rock and sediment), and cool colors a weaker magnetic field (less magnetite). The Hayward fault seen in the seismic-reflection profiles follows the northeast edge of the highest magnetic-field values. To the trained eye, the map also reveals subtle details about how rocks have been bent and broken by the fault.
Both the gravity and magnetic maps show that the active fault identified in the chirp seismic-reflection profiles extends kilometers below the bay floor, to depths where earthquake ruptures begin.
Deformation modeling to close the gap
The maps also show a gap of about 7 kilometers (4 miles) between the north end of the Hayward fault seen in chirp profiles and the south end of the Rodgers Creek fault mapped on land. The researchers suspected that the Hayward fault continues across the gap, but marshy land and extremely shallow water prevented direct mapping. A sharp step in gravity values and an abrupt change in topography—from hilly on the west to relatively flat on the east—marks the path where the researchers thought that the two faults might connect.
To test their hypothesis, USGS coauthor Tom Parsons created two computer simulations of the faults and the pulling and pushing pressures (stresses) on them. In one model, he assumed the 7-kilometer gap between the faults. In the second model, he assumed that the faults connect along the bending path (dashed white line) shown in the gravity map.
Using a fault-movement rate of 9 millimeters (one-third inch) per year from measurements on land, Parsons applied stresses to each fault simulation. The first model showed that stresses on disconnected faults would likely produce a break connecting them in the area where the researchers had mapped the Hayward fault beneath San Pablo Bay. The second model showed that stresses on connected faults would produce deformations like those seen in the chirp seismic-reflection profiles, including secondary faults and a zone of subsidence (a sunken area) in the northern part of the bay.
Now the scientists had seismic-reflection profiles, gravity, magnetism, and computer simulations supporting the same conclusion. Rather than the stepover they had expected, the researchers had discovered a direct connection between the Hayward and Rodgers Creek faults.
“What we found is a fault bend,” said Watt.
This new information will likely lead to better estimates of how the faults might rupture together in a large earthquake. Better estimates can improve earthquake forecasts, like the ones in the USGS Fact Sheet “Earthquake Outlook for the San Francisco Bay Region 2014–2043.”
The next step is to determine if the Hayward and Rodgers Creek faults have actually ruptured together in the past—and when, and how often. The answers could lie in thick, gooey mud.
Learning more from mud
From October 17–21, 2016—coincidentally, the same week the “Missing Link” paper was published—Watt and her team collected 20 core samples of muddy underwater sediment along the newly mapped extension of the Hayward fault in San Pablo Bay. They plan to use radiocarbon dating of benthic foraminifera (one-celled microorganisms) in the core samples to determine the ages of sediment bends and breaks near the fault. These dates, plus more seismic profiling planned for fall 2017, should help unravel the history of earthquakes in the bay.Stay tuned!
in this issue: