The Benavente earthquake: when the Tagus valley shook in 1909

The Infante D. Luís observatory in Lisbon logged Friday 23 April 1909 as the only day of fog that month. Around ten in the morning, people in the villages near what would be the epicentre heard rumbling underground. At the time, a writer in the civil engineers' journal recorded later, they put it down to a revolution in Setúbal.1 source1 source

The noises went on through the day. At about five in the afternoon came a far louder one, and then a strong shock that was felt across Portugal and much of Spain. It lasted only a few seconds. The one seismograph working on the Portuguese mainland, in Coimbra, showed it went on for no more than 30. Seismologists today put the start of the rupture at about 17:39 Greenwich Mean Time, 5:03 pm local time.4 sources4 sources

In Benavente, a town of 3,557 people on the river Sorraia, houses of rammed earth and adobe folded in on themselves. The parish church came down. So did a new hospital, built to modern rules and not yet opened. At the pharmacy the front wall fell away and the upper floor dropped onto the ceiling below, yet inside, a local paper marvelled, not a single bottle had moved from its shelf.3 sources3 sources

That night many people across the region refused to sleep indoors, even where their houses still stood. In Benavente, survivors pulled an image of Nossa Senhora da Paz from the rubble, set it up in the Praça Anselmo Xavier with other saints and prayed in front of it. The mayor, whose own house had collapsed, camped with his wife in a shelter made of blankets, sheets and a bedspread.1 source1 source

Benavente, Samora Correia and Santo Estêvão sit on the left bank of the Tagus, where the river spreads into the Lezíria, a flat plain of alluvial mud and sand laid down by floods. In 1909 it was thinly peopled farming country, and almost every house was built of rammed earth, adobe or brick. About 6,300 people lived in the area that would be hit hardest.2 sources2 sources

Portugal was still a monarchy, though not a secure one. Manuel II had been proclaimed king on 2 February 1908, the day after his father, King Carlos, and his elder brother were shot dead in Lisbon. Republican and monarchist newspapers were at war with each other, and both would soon use the earthquake against the other side.2 sources2 sources

Earthquakes were on people's minds. At the end of December 1908 one had flattened Messina, in Sicily, killing thousands, and towns across the Ribatejo had held street collections for the Italian victims. People also remembered the shocks felt in Lisbon, the Ribatejo and the Alentejo in 1903, and, above all, the great Lisbon earthquake of 1755.1 source1 source

The worst damage fell in a triangle between Benavente, Samora Correia and Salvaterra de Magos, and on the village of Santo Estêvão. Paul Choffat, the Swiss-born geologist who led the official study, rated the shaking there at X on the twelve-degree Mercalli–Cancani scale then in use. In Benavente, Samora Correia and Santo Estêvão about 80% of the houses could no longer be lived in.3 sources3 sources

An intensity scale grades the shaking at each place, in Roman numerals, from what people felt and what happened to buildings, so one earthquake has many intensities. Magnitude is different: a single figure for the size of the earthquake itself. Nobody gave the 1909 earthquake a magnitude until 1956.3 sources3 sources

The flood plain behaved strangely. Long cracks opened in the fields and spat out water and sand, a sign that the waterlogged ground had briefly turned to liquid. The sand looked black when it came up wet and dried white, so on freshly ploughed land the cracks could be seen from far away. The same thing happened again in the strongest aftershocks, on 4 May and 2 August.2 sources2 sources

Farther out the damage thinned but did not stop. On the other bank, houses along the river at Alverca, Alhandra and Vila Franca de Xira were damaged. In Santarém the main corridor of the seminary was wrecked, the railway track was knocked out of line where blocks from the old walls fell on it, and some terrified people jumped from windows. In Lisbon chimneys came down and façades cracked.2 sources2 sources

Why did so many buildings fail? Contemporary engineers blamed soft alluvial ground and building methods as much as the shaking. Almost everything in these towns, a local paper noted, was made of rammed earth, adobe or brick, often weighed down with heavy stone trim. The aftershocks went on for months. Teves-Costa and colleagues, combining Choffat and Bensaude's list with the daily bulletins in the Diário do Governo, count up to 183.2 sources2 sources

The figures are close but not identical. José de Oliveira Ferreira Diniz, whose study of the earthquake was finished in December 1909, wrote that 46 people died and 35 men, women and children were badly injured. The summary of the official report by Teves-Costa and colleagues gives 46 dead and 75 injured, with Benavente accounting for 30 of the dead and 35 of the injured. Fonseca and Vilanova, and Canora and colleagues, give 47 dead.5 sources5 sources

For an earthquake that left whole towns uninhabitable, that is a small toll. Writers at the time and seismologists since give the same reasons: it struck in late afternoon, when most of the people of these farming towns were still in the fields, and the area was thinly populated.2 sources2 sources

Doctors and fire brigades reached the ruined towns quickly, mostly in private vehicles lent for the purpose; pharmacists sent help too, and the telegraph carried the news across the country. To stop looting, Benavente was guarded by a hundred police and soldiers of Caçadores 6, a light infantry unit, with thirty men of the 3rd Artillery. Some thefts happened anyway.1 source1 source

On 24 April, the day after the earthquake, Manuel II visited Benavente and Salvaterra de Magos and spoke to people who had lost their homes. He also chaired the national relief commission, the Grande Comissão Nacional de Socorros. The historian Maria Gabriela Mota Marques reads both as the king's attempt to show that the monarchy cared and was in charge.1 source1 source

The first job was shelter. Landowners and councils sent timber for huts: the farmers' guild and the council of Chamusca, upriver, sent twelve carpenters and a large load of wood. Others gave food, clothes and transport. The district governor of Santarém managed the funds, which were shared out at meetings in Lisbon, with most of the money going to the worst-hit towns.1 source1 source

Most of the money came from ordinary people. Towns held bandos precatórios, street processions with collecting tins, as well as subscriptions, bullfights and charity evenings. Local papers printed who gave what, from the hundreds or thousands of réis given by landowners to the small sums of labourers. The Spanish government sent 100,000 pesetas. The ducal house of Cadaval helped Muge, and the landowner Luiz Sommer helped Coruche.1 source1 source

Not everyone came to help. Benavente, Salvaterra and Samora Correia became a Sunday outing for sightseers from Lisbon. And the old quarrel over religion flared. Some clergy preached that the earthquake was a punishment on atheists, freemasons and people who skipped mass; republican papers answered that a God of love would not kill poor innocents.1 source1 source

The government moved quickly on paper. The Minister of Public Works named two commissions of engineers and public-works officials, by a ministerial order of 29 April printed in the civil engineers' journal; Fonseca and Vilanova put the appointment three days after the earthquake. One was to draw up rules for rebuilding the ruined towns and for new buildings across the country; the other was to study the earthquake itself.3 sources3 sources

Benavente's council proposed a national loan at 4.5% interest to rebuild the town and its villages, open to every council that joined. Owners would repay in twenty yearly instalments and pay no stamp duty or property tax for ten years if they rebuilt by December 1911. Azambuja asked the government for interest-free credit on similar terms. Mota Marques does not say whether either scheme went ahead.1 source1 source

Relief came in two stages: first huts, scattered to reduce the risk of epidemics in a district whose terrain and climate made it prone to disease, then new houses. Private donors added their own projects; a Lisbon firm set out to build a group of ten houses. Complaints followed. Some said aid went only to a small area, and that much of it paid for repairs to houses landlords rented out, not for the tenants.1 source1 source

By then the monarchy had fallen and Portugal was a republic, but four and a half years after the earthquake Benavente still had its camp of huts. Researchers on the SEISMIC-V project, which studies Portugal's traditional earthquake-resistant building, conclude that the official reconstruction plans of 1909, unlike Pombal's after 1755, do not seem to have been carried out in full. Because state aid was slow, most rebuilding was left to private owners.2 sources2 sources

What they built looked much like what had fallen. The typical post-1909 house in Benavente is a long, single-storey building with a gabled roof. Its front wall, though, has an inner timber frame filled with stone or adobe, a simpler version of the gaiola pombalina, the wooden cage designed for Lisbon after 1755. Benavente's historic centre still has its old triangular plan, around the site of the destroyed parish church.3 sources3 sources

The scientific commission was led by Paul Choffat, a Swiss geologist working for the Portuguese Geological Services, with the petrographer Alfredo Bensaude, later the founder of Lisbon's Instituto Superior Técnico. Portugal had almost no instruments, so they did what the Swiss Seismological Commission did and asked people.2 sources2 sources

Questionnaires went out by post, with free return, to local officials, priests, teachers and postal and telegraph clerks. In Lisbon 25% of 240 came back; in the provinces 58% of 897, including 91 from telegraph operators. Choffat complained that answers were late, contradictory and written by people who had never paid attention to such things, at a time when about three-quarters of Portuguese could not read or write.1 source1 source

From the answers and newspaper reports Choffat drew a map of how strongly the earthquake was felt across the country. Together with his questionnaires, it showed that the shock had come from under the land, not from the sea as people had assumed for Portuguese earthquakes. Ferreira Diniz, working alone from the Coimbra seismogram, reached the same conclusion and traced the source along a line through Salvaterra, Benavente and just east of Samora.3 sources3 sources

Abroad the earthquake was well recorded. Within months the Spanish seismologist Navarro Neumann had gathered data from 51 stations, and researchers have since recovered 31 potentially useful seismograms from 14 stations, from Spain to Sweden. Those records let modern researchers work out its moment magnitude, a measure based on the physical properties of the earthquake.2 sources2 sources

The early figures, 6.6 to 7.3, were worked out from the worst damage, from station bulletins or from a single seismogram. The modern moment magnitudes, about 6.0 to 6.2, come from digitised seismograms and, in Cabral's case, the intensity map. Surface-wave magnitudes from the same old records come out a little higher, at 6.3 and 6.35, and Fonseca and Vilanova's 2010 paper has M 6.3 in its title. Even at 6.0 it was the largest earthquake on land in the Iberian Peninsula in the twentieth century.3 sources3 sources

The earthquake also pushed Portugal towards its own seismograph network. In January 1910 a meeting called by the king chose the Infante D. Luís observatory in Lisbon as the central station, and in 1910 a Mainka seismograph was installed at the Lisbon University observatory, followed by simple seismoscopes at three weather stations on the mainland and one in Madeira. A network covering the whole mainland came only after the earthquake of 28 February 1969, which has its own article.2 sources2 sources

The rupture never reached the surface. The best modern solution, by Daniel Stich and colleagues in 2005, points to a reverse fault, where squeezing pushes one block of rock up over another, trending roughly east-north-east, about 10 kilometres down. No mapped fault at the surface matches it, so geologists think it was a blind thrust, a fault buried under the thick sediments that fill the valley.3 sources3 sources

There are candidates. One is the hidden southern end of the Azambuja fault, which shows near Azambuja as a 15-kilometre scarp and probably runs on south under the river's alluvium. Another is a buried structure linking it with the Vila Franca de Xira fault. João Cabral and colleagues trenched a half-metre step in the flood plain west of Benavente, but the deformation they found was of doubtful tectonic origin.2 sources2 sources

The Lower Tagus Valley has the highest seismic hazard in western Iberia, and its faults could produce earthquakes up to about magnitude 7. They move slowly, by a fraction of a millimetre a year. Since 1970 instruments have recorded more than 700 small ones in the valley, only 41 of them felt and only four above magnitude 4.2 sources2 sources

In 1909 the hardest-hit area had about 6,300 inhabitants. By 2017 the same towns had about 28,300, and the area holds infrastructure of national importance. Greater Lisbon, which shook in 1909 too, had almost 3 million people at the 2011 census.1 source1 source

Fonseca and Vilanova, writing a century after the event, judged that a repeat of 1909 would be a major hazard and that public awareness of it was low. Canora and colleagues note that faults inside plates often go quiet for long periods and then produce clusters of earthquakes, so a century of calm says little about what comes next.2 sources2 sources