Magnitude 7.4 shock ruptures a fault in the Longitudinal Valley of eastern Taiwan

POSTED ON APRIL 4, 2024 BY TEMBLOR

A magnitude 7.4 earthquake rocked much of Taiwan, causing death and damage. The eastern side of the island, near the epicenter, was especially affected.

By Ross S. Stein, Ph.D., Temblor, Inc., Shinji Toda, Ph.D., IRIDeS, Tohoku University, Japan, Chung-Han Chan, Ph.D., E-DREaM, National Central University, Taiwan, and Volkan Sevilgen, M.Sc., Temblor, Inc.

Citation: Stein, R.S., Toda, S., Chan, C.-H., and Sevilgen, V., 2024, Magnitude 7.4 shock ruptures a fault in the Longitudinal Valley of eastern Taiwan, Temblor, http://doi.org/10.32858/temblor.338

Author’s note: The U.S. Geological Survey (USGS) assigned the mainshock as magnitude 7.4. The Central Weather Bureau of Taiwan assigned the event a magnitude 7.2. Locations and depths also differ by about 15 kilometers. In the figures, we use both versions depending on the data source, but in the text, we refer to the event as magnitude 7.4.

On April 3, 2024, at 7:58 a.m. local time (April 2, 2024, 23:58 UTC), a magnitude 7.4 earthquake struck near the city of Hualien, in eastern Taiwan. The earthquake was followed by a magnitude 6.5 aftershock a mere 13 minutes later. The U.S. Geological Survey (USGS) reports a relatively shallow depth of about 35 kilometers (or about 22 miles).

As of this writing, ten people have died according to Taiwan’s National Fire Agency. More than 1,000 injuries have been reported, and first responders have been engaging in search and rescue missions to find people trapped by landslides or stuck in damaged structures. Thus far, authorities report that 28 buildings have collapsed.

In this preliminary report, we probe what happened, what we can learn, and what might happen in the upcoming days and weeks.

Forces of nature on display in Taiwan

Few places on Earth are as seismically active as eastern Taiwan, which sits atop the edge of the Eurasian Plate. Here, the Philippine Sea Plate is crashing into the Eurasian Plate at 90 millimeters per year, about twice the Cascadia motion in the Pacific Northwest of the United States.

The Longitudinal Valley Fault, a thrust fault that runs along much of Taiwan’s east coast, absorbs about half of this motion (Figure 1). Thus, the fault moves at about 50 millimeters (2 inches) per year), shoving a coastal sliver of the island upward (and westward) with each successive earthquake. This uplift builds the Coastal Range.

Figure 1. Top: The rate of magnitude 7 and larger earthquakes in eastern Taiwan could be unmatched compared with anywhere else in the world, with 13 such quakes striking since 1920. Seven of these occurred within 40 kilometers of the April 2, 2024, epicenter. Most of the earthquakes appear to have struck on the Longitudinal Valley Fault, which is inclined, or dipping, to the east, as seen in the cross section in Figure 2. While the 1920 magnitude 8.2 shock occurred farther offshore, its location uncertainty is also much greater than for more recent quakes. Bottom: Eastern Taiwan is relentlessly battered by typhoons, accelerating the erosion of the mountains uplifted by earthquakes. Credit: top panel, Stein et al., 2024, CC BY-NC-ND 4.0, bottom panel: NOAA and Joint Typhoon Warning Center

The Longitudinal Valley Fault looks rather simple from Earth’s surface, but at depth, it is paired with the Central Range Fault, inclined in the opposite direction (dipping to the west; Figure 2). The Central Range Fault also accommodates crustal contraction, and has hosted events like the 2022 magnitude 6.5 and magnitude 6.9 Chihshang earthquakes (Tang et al., 2023). The Central Range Fault gives rise to the Central Range. At this point, it’s unclear whether the Longitudinal Valley Fault or the Central Range Fault ruptured in the earthquake. The USGS source assumes the former, but scientists in Taiwan lean toward the latter. This will be clarified in the weeks ahead.

Figure 2. This cross section is simplified from Tang et al. (2023) and depicts what’s happening underground at a location south of the April 2, 2024, magnitude 7.4 earthquake. Here, two thrust faults accommodate much of the 90 millimeters per year of tectonic contraction. This accommodation occurs by a combination of large earthquakes and aseismic creep (Thomas et al., 2014). Oddly, the faults join near the ground surface in the Longitudinal Valley, so as the Coast Range (to the east) and Central Range (to the west) are pushed upward, the triangular block between and beneath them is pushed downward. Just below the cross section, the Eurasian Plate is being subducted from the west, forming yet another surface on which large earthquakes can strike. Credit: Stein et al., 2024, CC BY-NC-ND 4.0

Eastern Taiwan is a place where, through earthquakes and cyclones, the forces of nature coalesce and do battle. Eroded sediments from both of the uplifted ranges blanket the intervening Longitudinal Valley, which is controlled by the fault zones. Along the faults, rocks are ground to a fine powder by earthquakes, and those sediments are among the most erodible. This competition between tectonic uplift, which in part results from the cumulative effect of earthquakes, and erosion, which attempts to grind down mountains via processes like landslides and storm-generated runoff, prevents eastern Taiwan from attaining a great elevation (Stolar et al., 2007).

Earthquake early warning

Taiwan Central Weather Administration operates a national earthquake early warning system. A Palert system developed by academics operates independently. Each system has its strengths.

The national earthquake early warning system worked well and so may have saved lives. The initial calculated magnitude was low at magnitude 6.2, but updated to 6.8 as more station information came in. That’s still low than the final magnitude, but closer.

For towns more than 50 kilometers from the epicenter, people received a warning before the strong shaking began. But the shaking was stronger in Taipei, 125 kilometers to the north, than the warning system anticipated, probably because the fault rupture propagated in that direction, causing the seismic wave amplitudes to be greater but briefer. Nevertheless, the shaking in Taipei was not strong enough to damage well-built structures.

Unfortunately, Haulien, the largest affected city, was in the so-called late-alert zone of the warning system, meaning that an alert would have come after shaking began. Haulien lies just 30 kilometers north of the mainshock’s epicenter, which is located on or just offshore (Figure 3). Seismic waves travel at about 2 to 3 kilometers per second. If the earthquake had been detected immediately at the epicenter (which would only happen if a seismic station was directly on top of the epicenter), then early warning systems would have about 10 to 15 seconds to detect and verify the event, calculate the predicted shaking, and deliver alerts. However, this was not possible because the nearest seismic station was 15 kilometers from the hypocenter at depth, which means that the first wave would have reached that station about 7 seconds after the rupture began. At this point, the process of detection and calculation of shaking would have taken approximately 5 to 6 seconds. Taking into account the time needed for delivering alerts, this may explain why the residents of Hualien do not appear to have received an alert in time (pers. comm., Prof. Yih-Min Wu, National Taiwan University).

Figure 3. Taiwan operates several earthquake early warning systems. All the earthquakes on the day of the mainshock that triggered a warning as part of the Palert system are shown as purple circles on this map. The smaller colored circles are the local shaking intensities across Taiwan that are associated with the April 2 mainshock. The density of these observations is perhaps only surpassed in Japan. The Palert system is accessible in real time from the Temblor-E-DREaM app (app.temblor.net). Credit: Stein et al., 2024, CC BY-NC-ND 4.0

Hazard assessment

Over the past decade, a consortium of scientists from Taiwan (Taiwan Earthquake Model, now under the direction of the Earthquake Disaster & Risk Evaluation and Management Center, or E-DREaM), constructed a new probabilistic seismic hazard assessment for the country. The magnitude 7.4 mainshock lies in an area of unusually high hazard (as seen in the inset in Figure 4, as this region is otherwise obscured by the aftershocks in the map).

Figure 4. The hazard in the vicinity of the mainshock is among the highest in Taiwan. So, in that sense, this quake was not unforeseen. The Taiwan Earthquake Model developed the hazard map shown here under the direction of E-DREaM, interactively accessible in the Temblor app (app.temblor.net). Credit: Stein et al., 2024, CC BY-NC-ND 4.0

Strength of shaking

The strength of shaking for this earthquake is largely consistent with modeled expectations for a quake of this size and depth, with one exception: A station about 12 kilometers from the presumed rupture periphery experienced a 150% g peak ground acceleration (Figure 5). Only critical buildings are designed to withstand horizontal accelerations exceeding 100% g. While 150% g is extreme for a magnitude 7.4, it was only seen at one station, so it may not be a reliable recording. Conversely, the peak ground velocities, often considered more representative of building performance, are rather low, never exceeding 70 centimeters per second.

Could this sequence grow?

Many past earthquakes that have struck along the Longitudinal Valley Fault occurred as progressive sequences. Most famously, in 1951, magnitude 7.5 and 7.1 shocks struck on one day, followed a month later by magnitude 7.3 and magnitude 7.8 shocks on the same day. Together, these earthquakes ruptured a 110-kilometer section of the fault. So, here we make a preliminary calculation of the stress transferred to surrounding faults to assess where the April 2, 2024, magnitude 7.4 shock brought faults closer to failure, and where it brought them further from failure.

One can see that central and northwestern Taiwan lie in a stress shadow (blue zone in Figure 6) indicating that they have become relatively safer as a result of the mainshock. On the other hand, the Central Range and parts of the Longitudinal Valley Fault, specifically to the north and south of the rupture, now appear to be more hazardous than before. However, this is an oversimplified view of the “source” fault rupture on which the magnitude 7.4 quake struck. So, we regard these results as preliminary and advisory.

Figure 6. Large stress increases (red colors) are calculated in the lightly populated Central Range, but also to the north and south of the rupture, which are more densely populated regions. Central Taiwan is calculated to fall under a stress shadow (blue colors), potentially reducing seismic hazard. The calculation uses Coulomb 3.4 (Toda et al., 2011) and the USGS “finite fault” source model. The receiver faults are nodal planes of background and aftershock focal mechanisms. Fault friction is assumed to be 0.4, and the most positive stress change on the two nodal planes of each mechanism is used, so the map is “red biased.” This means that the map is not an equal distribution of red and blue, but contains more red. Thus, if a region is blue, both nodal planes are calculated to have been brought further from failure (Toda and Stein, 2024a). Credit: Stein et al., 2024, CC BY-NC-ND 4.0

Taiwan was prepared

The past century of large, damaging earthquakes in eastern Taiwan prepared its residents for strong shaking. Those prior seismic events also prepared its scientists and government agencies to establish and enforce building codes, to maintain dense monitoring networks, to implement robust earthquake early warning systems, and to develop a rapid emergency response capability. As a result, though buildings fell and people died, the damage and loss of life was far lower than it would be in most other seismically active parts of the world, including many developed countries — a signal achievement.

References

Stolar, D. B., Willett, S. D., & Montgomery, D. R. (2007). Characterization of topographic steady state in Taiwan. Earth and Planetary Science Letters, 261(3-4), 421-431. https://doi.org/10.1016/j.epsl.2007.07.045

Tang, CH., Yunung Nina Lin, Hsin Tung, Yu Wang, Shiann-Jong Lee, Ya-Ju Hsu, J. Bruce H. Shyu, Yu-Ting Kuo, and Horng-Yue Chen (2023), Nearby fault interaction within the double-vergence suture in eastern Taiwan during the 2022 Chihshang earthquake sequence. Commun Earth Environ 4, 333 (2023). https://doi.org/10.1038/s43247-023-00994-0

Thomas, M. Y., J.-P. Avouac, J.Champenois, J.-C. Lee, and L.-C. Kuo (2014), Spatiotemporal evolution of seismic and aseismic slip on the Longitudinal Valley Fault, Taiwan, J. Geophys. Res., 119, 5114–5139, doi:10.1002/2013JB010603.

Toda, Shinji, Stein, R.S., Sevilgen, Volkan, and Lin, Jian (2011), Coulomb 3.3 Graphic-rich deformation and stress-change software for earthquake, tectonic, and volcano research and teaching—user guide: U.S. Geological Survey Open-File Report 2011–1060, 63 p., available at https://temblor.net/coulomb/

Toda, S., and R. S. Stein (2024a). The Role of Stress Transfer in Rupture Nucleation and Inhibition in the 2023 Kahramanmaraş, Türkiye, Sequence, and a One-Year Earthquake Forecast, Seismol. Res. Lett. 95, 596–606, 1–11, doi: 10.1785/0220230252.

Copyright

We publish our work — articles and maps made by Temblor — under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) license.

For more information, please see our Republishing Guidelines or reach out to news@temblor.net with any questions.

GTVRG e.V.

Samstag, 13. April 2024

MAGNITUDE 7.4 QUAKE ROCKS TAIWAN