I have seen fish and men hanging from trees,
while the cow, sheep, and horse swam in the sea.
Where wagons used to roll, there the skiff flies along with sails unfurled,
and goes and returns by unaccustomed routes.
(John Stradling. Source http://website.lineone.net/~mike.kohnstamm/flood/stradling.html)
It was one of those days. Living on the coastal flood plain is always chancy. You are only as secure as your flood defences, and on this day, those defences weren’t good enough. The sea broke through, and flooded the low-lying levels around the Bristol channel, even encircling Glastonbury. As one commentator later wrote, Men that were rich getting out of bed in the morning, were poor before noon the same day. Possessions, land, buildings, they were gone. For others it was much worse. The water came without warning, submerging farms and towns. Perhaps 2000 people died, unable to flee from the floods. More than 400 years later, this still ranks as the UK’s worst flood disaster.
Even without the flood, these were memorable times. The controversial King James was on the English throne. Guy Fawkes had just tried to blow up parliament, which for unfathomable reasons is still celebrated in the UK, on November 5. The King James bible was being written, and Shakespeare performed in London as one of the King’s Men; his plays already famous. In these exciting times, the news from the west country still caused upset.
The date of the disaster can seem unclear. Some records give it as January 20, other as January 30, and the year is given as 1606 or 1607. A few people have suggested that there must have been more than one event. But the explanation for the date confusion is prosaic. This was some time after the catholic church had initiated a new calendar, with fewer leap days and shifted by ten days. The old calendar had slowly gone out of sync with the seasons and by this time the difference had become notable. The Gregorian calender replaced the old Julian one, and was a clear improvement, but it suffered from one major draw-back: it was catholic, at a time of strong protestant movements in northern Europe. Therefore, many were slow to adopt the new enemy calendar. England in particular had held off, even in those days already being wary of new-fangled continental innovations. So Bristol still used the Julian calendar – for them the date was Jan 20. But many people in London, perhaps more progressive and certainly more international in their outlook, had started to use the Gregorian calendar in use in France and Spain – so here the date was commonly (but not invariably) listed as January 30.
How about the year? That seems harder to get wrong! But in the Middle Ages, the year did not start in January, but at Easter or the onset of spring. This had ancient roots. The Jewish calendar, the oldest for which we have written records, originally started each year in spring. (It changed later to start in the autumn, probably during to the Babylonian era as Mesopotamia had always used this system. The Jewish scriptures use both systems, some predating and others postdating the change-over.) The equinoxes are convenient, by the way, because the position of the Sun on the sky changes rapidly at these times, so the precise date can be measured easily. They are also more important than the solstices for agriculture. During the Middle Ages, the start of the year was normally taken as March 1, although this was not universally accepted. The Gregorian calendar confirmed January 1 as the start of the year. Some regions which had not accepted the new calendar did accept the new beginning. So late January was 1607 according to the new system, but still in the previous year, 1606, according to the old one. Any newspaper called ‘The Times’ would have had to specify which times.
It all happened around the Bristol channel, which separates the southwest of England from Wales and forms a funnel aimed at the heart of England, the Cotswolds. This was one of the invasion routes considered by the Spanish. The funnel narrows until it becomes the Severn estuary. The funnel shape, and the fact that it has just the right length to act as a resonator, causes an extreme tidal range, which is one of the largest in the world. Beyond Bristol, the estuary becomes the river Severn, which is the longest river in the UK (Bristol itself is on a different river, the Avon). The Severn is famous for its bore, a solitary wave running upstream with the tide. On both sides of the Bristol channel there are extensive low-lying areas. On the north side are the low-lying area around Cardiff and Swansea in Wales. On the south side are the Somerset Levels, an area still prone to river flooding today.
A number of towns had sprung up around the channel, some Roman in origin (Cardiff), some Viking posts (Swansea), some medieval trading centres (Bristol). The flood plains were extensively farmed. There were sea defences, but they were not in as good a shape as they should have been. This was in part because of the dissolution of the monasteries: these had been responsible for some of the upkeep, and the new owners did not take on these responsibilities. The flood plains were a sitting duck. This was in the Little Ice Age, and storms were more frequent than they are nowadays, with perhaps the storm tracks running further south. On this day, in the early morning of January 30, the sea rose, and kept rising. By 9 am the dikes crumbled or were overtopped. The sea did not quietly accept its boundaries – water is a jealous neighbour.
Appledore is an example of a seafront town affected by the flood. It is located on the river Taw, just downstream from Barnstaple, the first place along the Bristol Channel to feel the incoming surge. As the waves came in, houses were swept away. It appears the entire sea front was destroyed. The houses along the shore were not rebuild for a century afterwards – the memory ran deep. Appledore and Barnstaple form a low-lying inlet in a hilly landscape. The flood moved up the river and flooded the inlet but the areas a bit away from the river were uphill and stayed dry. However, Barnstaple was hit hard; a few hundred people reportedly perished here.
Walter Yonge wrote about the events at Barnstaple and elsewhere of that morning:
The 20th of Jan 1606-7, by reason of a great tempest, the sea brake in at divers places on the north side of this country, as at Barnstaple, where was much hurt done. At Bridgwater two villages near thereabouts and one market town overflown, and report of 500 persons drowned, besides many sheep, and other cattle. At Bristol it flowed so high that divers packs, which were brought thither against Paul’s fair, standing together in a common hall of the city, for such purposes, stood three foot deep in water.
Around Cardiff, on the Welsh side of the Bristol channel, many low-lying areas were flooded and the water reached into the city. The nearby, flooded areas were sparsely inhabited at the time (nowadays they are densely packed suburbs). The main church in Cardiff, St Mary’s Parish Church, became undermined by the water. Over the following years, the church was slowly abandoned in favour of a sturdier one, although it remained in use until after 1700. Further east, some populated areas were wiped out: here, only the churches, complete with flood marks, remained.
The flood came in furthest in the Somerset Levels, a flood plain protected by less than solid dykes. In other places, the surge traveled up the rivers, reaching the centre of Bristol and flooding the area around the river Severn as far as Gloucester, the very area where nowadays surfers tackle the famous Severn bore.
Based on the flood marks on the churches, the water reached over 7 meter above AOD (Ordnance Datum) on the Welsh side, and over 7.5 meter on the Somerset Levels. This is actually not as high as might have been expected, as this would have been about the peak tidal level and the surge must have been well above that. The suspicion has been raised that the land has sunk by 1-2 meters since 1600, due to the improved drainage. Others argue that the tidal range has increased but there is no evidence for this and it seems too speculative. Perhaps the people putting up the markers were conservative in their estimates.
Comparing the woodcut of the flood with the modern view of the same 12th century church (Nash, Newport) gives an indication how deep the flood was. However, one should allow for some artistic license of the maker of the woodcut. The flood marker is near the door, ‘only’ about 1.5 meter above ground level, but it is small and indistinct and it is not clear that this actually is for the 1607 flood.
Cause of the 1607 disaster
The 1607 flood made the news once more when Haslett and Bryatt proposed that it had been caused by a tsunami. The disasters in Indonesia, Sri Lanka, and Japan over the past two decades have shown the devastation a tsunami can cause: the power of water is immense, and in the cities that were affected by tsunamis, almost nothing was left standing. This video is one of the scariest of the 2011 Japan tsunamis, starting with people horrified by a distant disaster unfold and ending with blind panic.
The idea that such a tsunami could happen, and had happened, in the UK was a sobering thought. The evidence for this came from a particular description of what happened, in a document entitled ‘Gods warning to his people of England’
Then they might see & perceive a far of as it were in the Element, huge and mighty Hilles of water, tumbling one over another, in such sort as if the greatest mountaines in the world, has over-whelmed the lowe Valeys or Earthy grounds. Sometimes it so dazled the eyes of many of the Spectators, that they immagined it had bin some fogge or miste, comming with great swiftnes towards them: and with such a smoke, as if Mountaynes were all on fire: and to the view of some, it seemed as if Myliyons of thousandes of Arrowes had bin shot forth at one time, which came in such swiftnes, as it was verily thought, that the fowles of the ayre could scarcely fly so fast, such was the threatning furyes thereof.
The similarity to the Japan tsunami is striking. To get a tsunami, an earthquake needs to have happened. The problem here is that a strong tsunami requires an earthquake of magnitude 7 or higher. This is because of the vast amount of water that needs to be displaced: the originating earthquake needs to lift or drop a correspondingly large volume of the subsea rocks. But there is no record of such a strong earthquake. The Bristol channel does have earthquakes, and in fact there are two reports that there was a small earthquake there some time after the flooding, either in May or February 1607. But these earthquakes reach at most M4. Proposing a much more distant earthquake does not help: this would have produced tsunamis in many different regions, not just the Bristol channel. The M8.4 Lisbon earthquake caused a tsunami in Cornwall (although not very high as it coincided with low tide) but left the Bristol channel untouched. (The UK is actually fairly well protected against tsunamis by the extensive continental shelf which disperses the energy of any tsunami far from the coast.)
A large, local tsunami can also be caused by a major landslide. In fact the highest tsunami ever recorded was caused by a landslide, triggered by an earthquake: the 1958 Lituya Bay tsunami, with a run-up height of over 500 meter! The video is about this event. But again, there is no evidence for such a landslide within the Bristol channel.
As an aside, tsunamis can also be generated by volcanic eruptions, in a variety of ways. The volcano may suffer a flank collapse, sending a huge slide into the sea . There may be a large explosion in shallow water. And finally, a descending pyroclastic flow can push up the water They are the largest cause of volcanic fatalities. Krakatoa is a good example. The top two of the VC list of dangerous volcanoes both threaten volcanic tsunamis.
For the 1607 event, the same tract that seems to describe a tsunami later reads
But so violent and swift were the outragious waves, that pursued one an other, with such vehemencie, and the Waters multiplying so much in so short a time, that in lesse then five houres space, most part of those countreys (and especially the places which lay lowe,) were all overflowen.
Five hours is far too long for a tsunami. The Lisbon tsunami came within 15 minutes and that is typical. The lack of a clear cause, and the slow onset, argue that for the 1607 event, no tsunami was to blame.
The long duration of the flood made it rather likely that a high tide would occur during the flood, making the impact much worse. But not all high tides are the same. Twice every lunar month the Sun and the Moon pull the sea water in the same direction, and a spring tide results, higher than usual by about a third. By bad luck, this was the case on the day of the flood. (The opposite effects gives rise to the neap tide, which is a wimpish high tide.)
But not all spring tides are the same. The Earth is closer to the Sun during winter, and this makes the spring tides notably higher at this time. However, at tat time the Sun is very low in the sky from the northern hemissphere and this mitigates its effect and makes the highest tides come a bit later, February-April. Furthermore, sometimes the Moon is better aligned with the Sun. The net effect was that the tide at 9 am on Jan 30, 1607, was the highest for 4.5 years. It magnified the impact of the surge, as it came on top of an already extreme tide.
(The origin of the word ‘spring tide’ is far from clear. After all, it happens twice a month! The word ‘neap tide’ is better understood: it comes from old english, when ‘nep’ meant ‘low’. The word ‘spring tide’ may also have ancient origins. The highest tides in Europe happen around the start of spring, and this a natural explanation for the name. The comments below have more discussion on this.)
On the morning of the flood, the tide at Avonsmouth peaked at 7.86 meter above mean level (AOD). This was unusually high.
But it wasn’t only the extreme tide. The morning tide was 2.5 meters higher than that in the evening, driven up by a westerly gale. Many of the reports mention the wind:
But the yeere 1606, the fourth of King James, the ryver of Severn rose upon a sodeyn Tuesday mornyng the 20 of January beyng the full pryme day and hyghest tyde after the change of the moone by reason of a myghty strong western wynde. John Paul, Vicar of Almondsbury
This storme begane at 3 of clock in the morning and continue tyll 12 of clock on the same day Barnstaple Parish Register, 1607
The westerly wind aimed directly at the Bristol channel, raising the water levels by 2 meters or more. For comparison, the storm flood of 31 January 1953 raised the water along the North Sea coast by 3-3.5 meters, but the North Sea is far more susceptible to storm surges. It was not only the force of the wind: the long duration also played an important role. Later that day, the strong wind became easterly (or more likely north easterly), driving the waters back out.
There is other evidence that the 1607 flood was a storm surge, rather than a tsunami. Flooding was reported later that same day on the opposite side of the UK, at King’s Lynn in Norfolk, coming in from the North Sea. And also on the oppsite side of the North Sea, a severe northwesterly storm occurred in the northern Netherlands on 30 January, sinking 5 large ships and badly damaging 10 more near Texel, carrying grain to Italy; 170 sailors drowned. The storm also caused flooding in Friesland, where the sea defences broke. None of the reports from the UK describe wind damage, suggesting the winds were not as strong there as in the Netherland. Interesting is that neither do the reports mention heavy rain: it was windy, but not particularly wet weather.
Strom surges are caused not only the wind. Low air pressure contributes too, as the lesser pressure allows the water to rise, by about 10 centimeter for every 10 millibars. A deep low-pressure system in the UK may have a central pressure of 960 mbar (even lower is possible), and this can contribute 50 centimeter to the surge.
The weather report
So what actually happened? The long duration of the storm suggests a relatively slow-moving low pressure system, initially north of the Bristol channel, perhaps in the Irish sea. The northwesterly storm in the Netherlands later that day shows that the low pressure moved to Denmark or Scandinavia, rather more rapidly than the leisurely progress before. The problem with this storm track is that the wind in the Bristol Channel would not have become easterly, which one source mentions happened.
A possibility is that the North Sea storm was a separate system, a secondary depression which formed south of the UK in the frontal system of the main depression, and moved rapidly around it.
As an example, here is a synoptic chart for 3 Dec 1994, before a minor surge occurred along the Severn estuary. On this day, the main low pressure system was located south of Iceland, but a minor disturbance had formed to its southwest, and this disturbance rapidly moved around the low pressure system. It did not develop into a major storm. However, this kind of situation could describe the events of 1607, if the secondary disturbance came in north of the Bristol channel, and moved southeast followed by a northeasterly track through the North Sea to Denmark. It would have moved south of the Wash (thus causing northeasterly winds here, leading the flooding at King’s Lynn), and north of the Netherlands causing the northwesterly storm there. The increasing wind may indicate the system deepened along this track.
Other possibilities exist, of course. The lack of more weather descriptions from elsewhere in the UK makes it difficult to know exactly what happened, and in any case weather systems are never carbon copies. Flooding in the Severn estuary increases if there is a long ‘fetch’ for the water, meaning wind patterns that sweep up the water from south of Ireland towards the Bristol channel. This can place a low pressure system over Ireland or just to the north. If such a system tracked across the UK, it could have caused the North Sea storm as well. However, this would not predict a change to easterly winds in the Somerset Levels. It is of course possible that this claim, based on a single source, is wrong.
The Great British tsunamis
Although the UK is fairly well shielded from tsunamis, there have been exceptions. The worst tsunami to hit the UK during its habitation hit Scotland, around 8000 years ago. A big landslide of Norway caused a wall of water 20 meters high which overran much of the Scottish coast, leaving a layer of sand up to half a meter thick. Some of the Shetland islands may have been inundated completely. Further south, the coast was not yet in its current location. In the middle of the North Sea was a marshy, inhabited region known as Dogger bank. Nowadays it is 20 meter below the sea. The bank must have been wiped clean by the tsunami.
The only other confirmed UK tsunami was in 1755, followed the M8.4 Lisbon earthquake. The tidal wave was 2-3 meters high in Cornwall, but little damage was done as the wave came in at low tide; the wave did not travel further up along the coast.
Several other tsunami-like events have been reported, most recently in 2011 and 1929. All appear to be caused by weather. Thunderstorms at sea can push up large waves, which pile up along the coast just like a tsunami. The 1929 wave reportedly reached several meters high, but it hit at low tide. It coincided with a sudden gale. These meteo-tsunamis tend to occur in summer, mainly June or July.
What happened before can happen again. Risk analyses must take this into account, and preferably plan for events worse than anything that has already happened. The Bristol channel area is much busier than it was before: more than a million people live near it. The flood defences have improved, and a similar event as in 1607 may be manageable. But could worse happen?
Adding an extreme tide an an extreme storm surge gives a possible sea height of 9 meter above AOD at the tip of the Bristol channel. Modeling has been done for such a situation, which in places would overwhelm the existing defences. Extensive flooding around Cardiff and Swansea would occur, and the material loss could exceed 10 billion pound. The question is how realistic such a situation is. There has been flooding around the Bristol channel a number of times in the past century; in the wet winter of 2014 it was due to the incessant rain, but at other times it was caused by storm surges, the highest of which reached 1.5 meter. The surge level is therefore not out of the question. The danger is when this coincides with a spring tide. Such a coincidence of course is rare. Is the combined risk acceptable? As sea levels rise, the risks increase, and what is rare now may become common within a century.
But now there is more at stake. Somerset is the location of Hinkley point, where a new nuclear power station (Hinkley Point C) is to be build. Power stations need access to water for cooling, and so they are build close to water, but obviously above the flood line. The Hinkley site has a low (but non-zero) flood risk. However, with 0.5 to 1 meter of sea level rise, the flooding risk will become significant. Fingers crossed.
After 400 years, the 1607 remains famous. The idea that it was due to a tsunami does not stand up the scrutiny: it was a storm surge, badly timed to coincide with the highest tide for four years. Even so, the flood marks in the churches are not as high as might have been expected. Is this due to the sea being lower or the land higher 400 years ago? Or are the markers not always correct? There are aspects we do not understand well. It is clear that this could happen again. How good are the defences? In the UK, sea defences are judged on economic benefit. In the Netherlands, they are treated as defensive weapons where any defeat is disastrous. Different attitudes give different outcomes.
In the end, it is all about the weather. Whether the sea can conquer the land is up in the air.
Albert, June 2017
The Bristol Channel floods of 1607 – reconstruction and analysis. Kevin Horsburgh and Matt Horritt, 2006, Weather, 61, 272
1607 Bristol Channel Floods: 400-Year Retrospective RMS SPECIAL REPORT, 2007.
The Severn Tsunami? The Story of Britain’s Greatest Natural Disaster. 2007, Mike Hall
Historic tsunami in Britain since AD 1000: a review. S. Haslett, and E. Bryant, 2008, Natural Hazards and Earth System Sciences, 8, 587–601
Updated 20 June 2017, to correct some points raised in the comments