M5.4 earthquake shakes Portugal
A relatively small event raises memories of past great earthquakes
Citation: Hubbard, J. and Bradley, K., 2024. M5.4 earthquake shakes Portugal. Earthquake Insights, https://doi.org/10.62481/1bd6732b
Para ler este post em português (traduzido automaticamente pelo Google), clique aqui.
Earthquake Insights is an ad-free newsletter written by two earthquake scientists. Our posts are written for a general audience, with some advanced science thrown in! To get these posts delivered by email, become a free subscriber. If you would like to support our work here, please also consider a paid subscription.
Many residents of Portugal awoke to shaking in the early morning hours of August 26, 2024: a magnitude 5.4 earthquake had struck offshore (at 05:11 local time). Nearly 6,000 people reported the shaking to the EMSC, mostly from Portugal, but also from neighboring Spain and Morocco.
While the average shaking near the epicenter was reported at intensity III-IV (weak to light), individual reports ranged widely. The higher values in the range probably mostly reflect personal anxieties or local site effects.
One person located in Casal do Sapo commented to the EMSC:
I was sleeping. I felt a big, fast tremor. I don't have any hanging lamps, so I didn't immediately understand what it was.
Indeed, hanging lamps are one of the ways that people can actually see the effects of low-intensity shaking — useful! Fortunately, for those without hanging lamps, earthquake listings can also be used to confirm whether an earthquake happened or not. For people located close to a coast, checking online sources is a particularly good idea: an earthquake below the ocean may be capable of triggering a tsunami, which could require evacuation. Shaking intensity cannot be used to determine if a tsunami is on its way: a distant earthquake might cause only minor (or even no) shaking at a particular location, but still be capable of generating a wide-ranging tsunami. Fortunately, a M5.4 is too small to be tsunamigenic.
At the time of writing, a handful of aftershocks have been detected, all of them too small to feel (<M2, offshore). To check out the latest activity in the area, view the seismic activity map at Portugal’s Institute of the Sea and Atmosphere.
Tectonic context
Why did the M5.4 earthquake happen? The answer is not simple.
Portugal lies within a broad region of very complex but slow tectonic deformation reaching from the Pyrenees to the Atlas Mountains of northern Africa, and from the Azores Islands in the Atlantic Ocean to the Mediterranean Sea. As seismic records and geological studies have accumulated over time, the picture of this tectonic region has only become more and more complicated. The latest large earthquake in this region was the deadly M6.8 earthquake in the Atlas Mountains of Morocco on September 8, 2023.
Although the active crustal deformation between Nubia (basically, Africa) and Eurasia is often labelled as ‘diffuse’, meaning widely scattered and not concentrated on specific large faults, this area looks more like a tectonic mosaic, where faults of all sizes work together to somehow get things done. In these areas, focusing on one fault might not be that useful. Instead, we may have to look at the big tectonic picture and then zoom in.
Like many regions on Earth, the geology and tectonics of the oceans is very important here.
Far to the west of Portugal, the Atlantic Ocean is growing wider each year as new oceanic crust is formed along the Central Atlantic Ridge. This ridge is one of Earth’s largest features, and is generally pretty predictable. However, at the latitude of Portugal, something strange happens. North of the Azores, the ridge spreads at ~23 mm/yr. South of the Azores, it spreads at ~19 mm/yr. That leaves a difference of ~4 mm/yr. While this might not sound like much, it’s enough to break up a tectonic plate.
A complicated fault system stretches between the Azores and Gibraltar, making up this 4 mm/yr difference in plate motion. At the Azores, the system is actually a spreading ridge: the plates are pulling apart. The Terceira Ridge spreads so slowly that it is called a hyper-slow spreading system — the slowest on Earth. To the east, this spreading ridge turns into a straight transform fault called the Gloria Fault. But looking further east of the Gloria Fault, this relatively simple plate boundary system dissolves into geological chaos. It remains extremely unclear how the Gloria Fault actually connects eastward to other structures.
So, differential opening of the Atlantic — faster in the north, slower in the south — is related to the faults that exist beneath the ocean to the south of Portugal.
But that’s not all! There is another, older ocean that is also playing a role here. Long ago, the pieces of continental crust that now make up the African and Eurasian continents were separated by a wide ocean — much wider than the existing Mediterranean — which formed after the breakup of a previous supercontinent during the Paleozoic era (think 270-340 million years ago)! That ancient ocean has now mostly disappeared: its crust has subducted into the mantle as Africa and Eurasia moved closer together, culminating in the complicated collision zone that today reaches from the Straits of Gibraltar to Turkey. However, remnants of that oceanic crust still lie beneath the Mediterranean Sea, growing ever colder and more dense over time. This old oceanic crust really wants to subduct, and will basically take any chance it can get to fall into the mantle below. The result is an extremely odd situation, where small subduction zones can form rapidly, with trenches zipping around willy-nilly as the old, cold lithosphere sinks into the Earth.
A 2024 paper by Duarte et al. shows an intriguing tectonic perspective — mapping out some of this subducted, ancient ocean crust deep within the mantle, still connected to a small trench at Gibraltar.
In this figure, the purple blob is a subducted slab that has fallen down into the mantle. From above (left panel), it looks like the slab is still kind of attached to the continents. However, a view from inside Earth (right panel) shows that the slab has mostly detached, and is still holding on to the surface only in one place. That place is the Gibraltar Trench, a tiny subduction zone that has been proposed to lie beneath the Straits of Gibraltar — right in the area where the Gloria Fault is trying to dump about 4 mm/yr of strike-slip motion into other faults.
So, earthquakes affecting Portugal, Morocco, and Spain are probably not only due to Africa-Eurasia motion, but also to deeper geodynamic features that still connect to the rigid plates at the surface.
We could dive deeper into the geology, which has been studied pretty well. However, the important point is that the major tectonic systems in this area are clearly related to the earthquakes, but we still don’t know that much about them!
Fortunately, we can also use the records of past earthquakes as a guide to the present and future.
Historical earthquakes
Although people in Portugal do not often feel shaking, the country has a long record of damaging earthquakes. We propose that these fall into two basic categories: local earthquakes (caused by rupture of faults located on or close to land), and regional earthquakes (caused by much larger ruptures located far offshore).
Local earthquakes affecting Portugal are generally caused by rupture of individual faults, and can be highly destructive if they occur close to population centers. The faults that rupture tend to be shallow, and although they may be widely felt and have locally intense shaking, areas of maximum shaking tend to be comparatively small.
Local earthquakes of magnitude between 6 and 7 struck near Lisbon in 1344, 1531 and 1909 (all dates in this post are ACE).
The record starts with a 1344 earthquake that is simply recorded in several earthquake lists, each made centuries after the fact. Comparison with later earthquakes suggests a source near Lisbon. Not much else is known about this very old event!
The better recorded 26 January 1531 Lisbon earthquake caused intensity IX shaking along the Targus River, and produced a damaging tsunami within the Targus estuary. Most of what we know about this earthquake comes from careful consideration of historical accounts, and urban archaeological studies. This earthquake is thought to have caused ~30,000 deaths, although the actual sourcing of that number isn’t clear to us.
The smaller 23 April 1909 Benavente earthquake also occurred quite close to Lisbon, causing extreme local damage as well as more widespread liquefaction. The event was far less damaging than the 1531 earthquake, and was responsible for 60 deaths. Somewhat surprisingly, this earthquake was actually recorded by early seismometers. Stich et al. (2005) found that the earthquake was recorded by 51 seismic stations distributed across Europe; 29 of these contained potentially useful seismograms.
Stich et al. were also able to digitize these seismograms and apply modern seismological methods. They found that the earthquake was caused by thrust rupture on a fault striking northeast-southwest – critical information that we can use to tie the earthquake to possible source faults. The focal mechanism is consistent with several faults mapped in the area, mostly structures that were born during the time of the dinosaurs (the Mesozoic) as extensional faults, and were reactivated in more recent times under a combination of compression and strike-slip, as Eurasia and Africa have shouldered more closely against each other.
The activity and seismic potential of on-land faults in Portugal is still a topic of active research.
The southern coast of Portugal was struck by a devastating local earthquake on the afternoon of 27 December, 1722, which produced a tsunami and caused widespread damage to nearby coastal cities. This earthquake was apparently sourced from rupture of an offshore fault, within an extremely complex and seismically active geological setting in the Gulf of Cadiz. As we mentioned previously, the geology of that area is pretty complex, and there are several possible explanations for the cause of the earthquake.
Regional earthquakes affecting Portugal have been sourced from huge ruptures of much more mysterious faults located far offshore. These kinds of earthquakes can cause very widespread damage, and can also source huge oceanic tsunamis.
The regional earthquake record might begin with a 1356 event, which caused damage in both Lisbon and Seville and is thought to have originated at sea to the southwest of Portugal. However, as for the 1344 earthquake, actual information is pretty sparse.
The most famous regional earthquake, the 1 November 1755 Lisbon earthquake, remains the most damaging natural disaster (aside from plagues, famines, and heat waves) to have struck Western Europe. That earthquake was caused by rupture of faults offshore southern Portugal, and is estimated to have reached magnitude 7.7-9.0. A large number of compounding factors led to a huge death toll: masonry buildings were not only challenged by intense shaking that lasted between three and five minutes — an unusually long duration —, but were also located on substrates that themselves became unstable, exacerbating their collapse. A huge tsunami arrived forty minutes after the shaking, preceded by a draw-down of the ocean. Survivors of the shaking had been driven toward the shoreline by the collapsed buildings and widespread fires that were started by candles and lamps, which had been lit for All Saints’ Day. While the impact on Portugal was catastrophic, this earthquake was large enough to also impact nearby countries like Morocco and Spain. The total death toll is estimated at ~40,000-50,000.
Which fault sourced the 1755 earthquake? We don’t know. However, an interesting paper published in 2023 tackled this question anew by directly modeling the expected shaking intensities for different source faults, and comparing the models to historical records.
Silva et al. (2023) first created a map of estimated shaking intensity, using historical and environmental records of various kinds.
One of the data points is particularly fascinating to us: at Tolmo de Minateda in Spain, a cliff containing Visigoth tombs carved directly into the stone apparently collapsed during the shaking (as dated by lichen, of all things, by Pérez-López et al., 2019).
Silva et al. then used the USGS Shakemap tool to calculate the expected ground shaking intensities for different source fault scenarios, and then compared them to the estimates from the collected observations. Most of our readers will be somewhat familiar with Shakemap because we often use these kinds of estimated intensities in our posts about recent earthquakes. However, Shakemap is also useful as a forensic tool for studying past earthquakes.
In the end, Silva et al. found that rupture of any one fault in the region cannot reproduce all of the shaking observations. In their opinion, the shaking intensity observations require the collective rupture of several large faults:
This kind of multiple-fault rupture is basically the nightmare scenario for any fault system. The equivalent scenario in California would be rupture of the northern San Andreas Fault and a large part of the Cascadia subduction zone, at the same time!
While this might sound unlikely, we do know that composite ruptures do happen. Over the last decades, several highly complex ruptures in other regions have produced big earthquakes. The best example is the 2012 Mw 8.6 Wharton Basin earthquake, which was the largest strike-slip sequence ever recorded, was caused by cascading rupture across several huge faults. Many seismic hazard assessments now routinely include these kinds of multi-fault rupture scenarios.
However, one of the challenges of searching for composite ruptures is that they vastly expand the realm of possibilities. For a historical study with many possible candidate structures and no direct evidence of slip on any given fault, it is unlikely that we can confidently identify the actual sequence of events — instead, we can only exclude some scenarios. Earthquakes follow some basic guidelines, but are always surprising us by producing more slip than expected, or deeper ruptures, or faster ruptures — all things that can impact shaking intensities and tsunami generation. In the end, the preferred model of Silva et al. (2023) does seem pretty reasonable to us.
(As a side note, the 1755 earthquake is sometimes incorrectly linked with a different, devastating earthquake that occurred several weeks later in Morocco: the 27 November 1755 Meknes earthquake. The confusion arose from historical reports that confounded the two deadly events. Recent geological studies have demonstrated that the Meknes earthquake was a local earthquake; possibly encouraged by stressing from the earlier 1755 earthquake, but otherwise unrelated.)
The 1755 earthquake was soon followed by another great regional earthquake, the 31 March 1761 Lisbon earthquake (it seems like there are too many “Lisbon” earthquakes, and we might need to diversify the naming in the future — especially since these earthquakes did not occur near Lisbon!). Like the 1755 earthquake, the 1761 event was associated with widespread shaking for three to seven minutes, and a tsunami that arrived in Lisbon an hour and a half after the shaking. Some estimates place the magnitude at ~M8.5 — making this event potentially comparable to the earthquake in 1755 — but the reported death toll is in the dozens rather than the tens of thousands. Based on the unusually long shaking intensity, it is possible that a multiple-fault rupture happened a second time.
A long period of relative quiet followed, which was disrupted by the 28 February 1969 M7.8 Portugal earthquake. Although widely felt, this earthquake caused little damage. Because this event occurred within the instrumental period, we have more information about its geological origin — likely thrust slip on a fault striking NE-SW, pretty deep (~50 km) within the old oceanic lithosphere.
A closer look at the geological context of the M5.4 earthquake
OK, we have looked at the regional picture and historical earthquakes. What about the actual area of today’s event?
Although recorded seismicity is more concentrated to the south of Portugal, earthquakes also occur within Portugal, and off its western coast. The recent M5.4 earthquake is one of these. This appears to be the largest such event since 1934, when a M5.6 earthquake occurred on-land to the east. Before that, a M5.3 to the northeast in 1926 was followed about a year later by a M5.4. And of course we can’t forget the 1909 M6.0 Benavente earthquake near Lisbon.
So far, three focal mechanisms have been determined for the August 26, 2024 earthquake by three different international seismological organizations (GFZ, OCA, and INGV). They all contain some similar features, although there is some variability — it can be difficult to accurately determine focal mechanisms for events of this size, especially when they occur offshore.
All three focal mechanisms show a strike-slip component, and nodal planes (the curves separating the white and black quadrants, representing the potential source faults) that are oriented NNE-SSW and E-W. If the NNE-SSW plane represents the fault, then the earthquake exhibited left-lateral strike-slip motion. Two of the solutions also show a thrust component of slip for this nodal plane.
What do we know about the geology in the epicentral area? A 2003 study used seismic reflection imaging to map out the active faults offshore southwestern Iberian margin (Gràcia et al., 2003). They found a number of potentially active structures, including two large thrust faults oriented NNE-SSW (marked MPF and SPF in the figure below), which might be implicated in the great 18th century earthquakes. More NNE-SSW faults have been mapped through Lisbon, including potential candidates for the 1909 Benavente earthquake. In between, the mapped geology is pretty different. The main faults near the surface seem to be either left over from the breakup of Eurasia from North America, or have been active during Miocene time — millions of years ago. So it isn’t clear to us how, or if, these NNE-SSW structures extend into the epicentral region of the M5.4 earthquake.
Thus, there is an observational gap in the area of the M5.4 earthquake. If any reader has more information about structures in that gap area, please point us in the right direction! Until then, we can only conjecture about the faults in the rupture area. It seems pretty likely that the complex fault system seen in the south extends through the M5.4 epicentral area and onto shore near Lisbon. It would likely require significant further research to nail down details of these faults.
What to expect:
All earthquakes have the capacity to trigger aftershocks, and this event is no different.
In some cases, earthquakes are able to trigger other events of similar magnitude or larger. Indeed, this appears to have happened in 1926/1927 in central Portugal. Because the recent M5.4 earthquake was moderate magnitude and was located offshore, it caused little damage, so a nearby triggered earthquake of similar magnitude is probably not a significant concern.
What about the chance of a larger triggered event — like the one in 1755? Since we don’t confidently know which fault or faults ruptured in 1755, it is difficult to speculate. However, this is always a very unlikely scenario. In general, any given earthquake has about a 5% chance of being a foreshock of something larger, but “larger” here has a wide range of possibilities, the vast majority of them much, much smaller than the one in 1755.
The Portuguese government has described this earthquake as a “real test of response capabilities in the event of a catastrophe.” Indeed, we should always use small earthquakes as a reminder to revise and refresh our preparations — at both governmental and individual levels.
References:
Bradley, K., Hubbard, J., 2023. Deadly M6.8 earthquake hits Morocco. Earthquake Insights, https://doi.org/10.62481/23bce686
Chester, D.K. and Chester, O.K., 2010. The impact of eighteenth century earthquakes on the Algarve region, southern Portugal. Geographical Journal, 176(4), pp.350-370. https://doi.org/10.1111/j.1475-4959.2010.00367.x
Duarte, J.C., Riel, N., Rosas, F.M., Popov, A., Schuler, C. and Kaus, B.J., 2024. Gibraltar subduction zone is invading the Atlantic. Geology, 52(5), pp.331-335. https://doi.org/10.1130/G51654.1
Gràcia, E., Dañobeitia, J., Vergés, J. and Team, P.A.R.S.I.F.A.L., 2003. Mapping active faults offshore Portugal (36 N–38 N): implications for seismic hazard assessment along the southwest Iberian margin. Geology, 31(1), pp.83-86. https://doi.org/10.1130/0091-7613(2003)031%3C0083:MAFOPN%3E2.0.CO;2
Grandin, R., Borges, J.F., Bezzeghoud, M., Caldeira, B. and Carrilho, F., 2007. Simulations of strong ground motion in SW Iberia for the 1969 February 28 (M s= 8.0) and the 1755 November 1 (M∼ 8.5) earthquakes-II. Strong ground motion simulations. Geophysical Journal International, 171(2), pp.807-822. https://doi.org/10.1111/j.1365-246X.2007.03571.x
Hill, E.M., Yue, H., Barbot, S., Lay, T., Tapponnier, P., Hermawan, I., Hubbard, J., Banerjee, P., Feng, L., Natawidjaja, D. and Sieh, K., 2015. The 2012 Mw 8.6 Wharton Basin sequence: A cascade of great earthquakes generated by near‐orthogonal, young, oceanic mantle faults. Journal of Geophysical Research: Solid Earth, 120(5), pp.3723-3747. https://doi.org/10.1002/2014JB011703
Pérez-López, R., Giner-Robles, J.L., Rodríguez-Pascua, M.A., Silva, P.G., Roquero, E., Bardají, T., Elez, J. and Huerta, P., 2019. Lichenometric dating of coseismic rockfall related to the Great Lisbon Earthquake in 1755 affecting the archaeological site of “Tolmo de Minateda”(Spain). Zeitschrift für Geomorphologie, 62(2), pp.271-293. https://doi.org/10.1127/zfg_suppl/2019/0504
Silva, P.G., Elez, J., Pérez-López, R., Giner-Robles, J.L., Gómez-Diego, P.V., Roquero, E., Rodríguez-Pascua, M.Á. and Bardají, T., 2023. The AD 1755 Lisbon Earthquake-Tsunami: Seismic source modelling from the analysis of ESI-07 environmental data. Quaternary International, 651, pp.6-24. https://doi.org/10.1016/j.quaint.2021.11.006
Somoza, L., Medialdea, T., Terrinha, P., Ramos, A. and Vázquez, J.T., 2021. Submarine active faults and Morpho-Tectonics around the Iberian margins: seismic and tsunamis hazards. Frontiers in Earth Science, 9, p.653639. https://doi.org/10.3389/feart.2021.653639
Stich, D., Batlló, J., Macià, R., Teves-Costa, P. and Morales, J., 2005. Moment tensor inversion with single-component historical seismograms: The 1909 Benavente (Portugal) and Lambesc (France) earthquakes. Geophysical Journal International, 162(3), pp.850-858. https://doi.org/10.1111/j.1365-246X.2005.02680.x
Teves-Costa, P., Batlló, J. and Cabral, J., 2017. The Lower Tagus Valley (Portugal) earthquakes: Lisbon 26 January 1531 and Benavente 23 April 1909. Física de la Tierra, 29(2017), pp.61-84. https://doi.org/10.5209/FITE.57599
The idea of “cascading” quakes along multi-fault systems (San Andreas and Cascadia) is sufficient to trigger insomnia.
Articolo ben strutturato, ricco di informazioni, che consentono di avere un idea chiara sia del contesto geologico strutturale, che della storia sismotettonica dell'area del sisma.