The sea is our fascination. We go out of our ways to find it, and it is where we go for holidays, spending our time lying on the beach. Close to half the world’s population lives within 100 kilometers of the coast. For those who live here, there is a good living to be had, either from farming the fertile river deltas or by harvesting the bounty of the sea itself. For a species that originated on the African savannah, our affinity to the oceans is quite remarkable. There is a tide in the affairs of men, as Shakespeare wrote.
But the sea is an inconstant neighbour, hard to pin down and impossible to contain. It rises and falls, with the waves, the tides, and the storm surges. And its onslaught comes with immense power. Living near the coast involves risk. Some people live behind natural defences, such as the sand dunes. Others build their own defences, living behind man-made barriers. And sometimes, there is little choice but to move house. All’s well that ends well, Shakespeare claimed. But in the end, the sea always wins.
The pillars of Pozzuoli
The most extreme case of sea level changes around our cities comes from Naples. Three marble pillars stand on the ruin of an old Roman square, close to sea level. When the pillars were build, they were smooth. But now, their remains show strange marks. The pillars are about 14 meters tall. The bottom 3.5 meters of the pillars are smooth. Above this there is a pock-marked layer, also about 3.5 meter deep. And above that the pillars return to being smooth.
The solution to this conundrum has been known for over 200 years. The scarred surface suffered the attention of a boring mollusc: some of the holes still have remains of the bivalve attached. But this mollusc does not live at height. When it grew on the pillars, the pillars must have been under water for it to flourish. The sea here must have reached at least seven meters above its current level. Why the smooth area at the bottom of the pillars? The best explanation is that this part was covered by rubble or sediment. The sea itself has not risen (and subsequently withdrawn) by that much. Instead, it was the land that changed. The cause is bradyseism, where water circulating underground causes periodic inflation and deflation. It happens inside large calderas – and there is a massive one here. Imagine how the people would have responded to seeing the sea approach! The area was quickly abandoned. We don’t quite know when.
And the land here is still moving. The harbour has been rising, bad news for boats and it beats the purpose of a harbour. A new set of mooring points had to be built, well below the old ones. In Naples, change is constant.
Bradyseism occurs in a few other volcanic regions. A similar, more widespread problem happens in build-up areas where too much water is being pumped from the ground, causing the land to subside. Subsidence can be problematic when the land is already close to sea level (or river level) to begin with, as is the case in many of world’s major cities. Examples of cities with increasing flood risks due to (probably) self-inflicted subsidence include Djakarta and New Orleans.
But now there is a new change. And it is not the land, but the sea itself which is changing. Sea levels have started to creep up, worldwide. The amounts are still small: over the past 25 years, the sea has come up by about 7 centimeters (or 3 inches, for those less metrically inclined). The reason is well known: melting glaciers are adding their water to the oceans. In addition, those oceans are warming – and warm water expands. For the oceans, the only way is up. We improved our standards of living with the best of intentions: better lives for all people is a admirable goal, and is perhaps underused as political slogan. But the way we went about it is now producing unintended side effects. We need to better understand what is happening, how it is happening, and where things are going: otherwise we are shaping the future with our eyes closed. Good intentions are not enough. As Shakespeare warned: Some rise by sin, and some by virtue fall.
The data shows that the sea level rise began in earnest around 1900. The IPCC finds that the rise was around 1.5 mm per year early in the century, increasing to around 2 mm per year around 1990 and 3 mm per year by 2010. The total change so far has been some 20 centimeter. The models predict that this acceleration will continue, although with large uncertainties. Sea levels may rise by anywhere between 40 cm and 1 meter by the year 2100. Beyond that, the only certainty is that the rise will become larger. The last time temperatures were at current levels, in a previous interglacial, the sea topped out at 4 to 9 meters above the current levels. The situation is not identical to that time, and we shouldn’t expect identical sea levels, but the warning is there. It is important to get the models right, to know what to prepare for.
And not all of the data is understood. Over the past 25 years, the rate of sea level rise has fluctuated, and the reason for this was not clear. It left an uncertainty over the projections for the future. But now it appears that this problem has finally been solved. And the culprit is an unexpected one: the finger is pointing at a volcano. The cause was Pinatubo.
To pin the sea
Measuring sea level is not easy. The water level is constantly changing, minute to minute, hour to hour, even month to month. And an instrument to measure sea level needs a solid surface for its mounting, something not fully compatible with the ocean. It needs, in the words of Shakespeare, One foot in sea, and one on shore.
The oldest records are from Europe, starting in 1700, first in Amsterdam and shortly after measurements also began in a few other cities. The Stockholm data set has been most useful: there, sea level rise started early, in the 19th century, and accelerated during the 20th century.
In the US, regular sea level measurements started after 1800. The first instrument used was a tide staff, mounted in sometimes inconvenient locations, if possible isolated from waves, and read off manually. Later, a float was used, with a pen attached which scratched a rotating drum, covered with paper. This removed the need for manual reading: now it was only necessary to collect and change the paper.
The next step came in 1993, when satellite measurements began. This opened up the world of the sea: no longer were measurements limited to a few coastal locations. Satellites don’t cover the polar regions well, but elsewhere their coverage is much more widespread than coastal gauges could ever give. The data that we have from 1993 onward is superior to the older data.
But even with satellites, the measurements are not easy. The required accuracy is better than a centimetre, while using a radar mounted on a box that is moving at several kilometres per second, a few hundred kilometres up. The radar signal comes back in about one millisecond, and this return time needs to be measured to an accuracy of a tenth of a nanosecond. The amount of water in the atmosphere affects the travel time, so the satellite measures the water vapour separately. Even the time it takes the signal to get from the receiver to the detector counts, and over the years this electronics can age. Each satellite has internal calibration systems, and there are several different satellites which allows for some cross-calibration. Finally, the location of the satellite needs to be known to a very high accuracy, using laser ranging and GPS.
The accuracy of the satellite data approaches 1 mm, meeting the requirements for the measurement of sea level rise per year. But there was something funny about the data.
Sea level jitters
The satellite measurements showed that the sea level rise in the mid-1990’s was 3.5 mm per year. But in the following decade, it slowed down to 2.7 mm per year, before accelerating again after 2011. This was very strange. The deceleration coincided with something which became known as the ‘pause’: a period during the 2000’s when the Earth seemed to be warming up more slowly than before or after. Explanations were sought: the atmosphere can warm more slowly if the oceans act as a bigger heat sink. The efficacy of the transfer of heat to the oceans can vary, due to the vagaries of the ocean currents and oscillations. But if more heat had been taken up by the ocean, the ocean should have warmed (and thus expand) faster, not slower. This was the opposite to what was seen.
And there was more. The tide gauge measurements from before the satellite era showed a rise of about 2.5 mm per year by the early 1990’s. This was quite a bit less than the satellites found shortly after. After 1993, the tidal gauges and satellites do agree well. It is hard to directly compare the two for a given tide gauge: you would need to be lucky that a satellite crossed directly over the tide gauge, and this is actually unlikely. If you only have a few satellites, you get good global coverage, but not local coverage. But the averages agreed well. It was as if the rate of sea level rise had jumped just when the satellites came on-line.
Church & White, in 2011, were the first to point the finger of suspicion at Pinatubo. It exploded in 1991, shortly before the time when sea level rise suddenly seemed to accelerate. However, they did not have a specific model, just a feeling. The timing seemed right, and it was a large eruption, at least for the 20th century (some other centuries had had much bigger bangs). But could a volcanic eruption really affect sea levels to such an extent? The word’s mine oyster, Pinatubo could have echoed Shakespeare.
Pinatubo in action
Pinatubo was one of the largest eruptions of the 20th century. It was also the first eruption where mass casualties were avoided because of accurate prediction of the eruption, something the USGS can take a lot of credit for. Their scientists rose to the occasion.
On June 15, 1991, 20 million tons of sulfur dioxide and ash exploded into the sky. Some of this made it into the stratosphere, and it stayed around for a year or more. Some remained lower: I remember air plane windows appearing milky white and hard to see through. Lufthansa told us that it was the volcanic sulfur eating into the windows, and that they had to replace the air plane windows far more often than before. The sulfur dixode created a partially opaque layer which intercepted the sun light. Less of the sun filtered down to the surface, and therefore the Earth cooled – a bit. We now know that volcanic ‘winters’ (it actually affects summers more than winters) can last a few years. Pinatubo depressed global temperatures by 0.2-0.5C. By 1994, the Earth had recovered.
But the sea hadn’t. This was unexpected. The sea has a much higher heat capacity than the air, and so it is much harder to change the temperature of the ocean than it is to change the air. Thus, the expectation had been that Pinatubo would have little or no effect on the ocean. Satellite measurements which could have shown better did not start until two years after the eruption. If you can’t measure it, and expect it to be negligible, the tendency is to forget about it. But men are men; the best sometimes forget, Shakespeare wrote.
Fasullo et al, in 2016, remembered, and took a closer look. The figure here is from their work. The top panel shows the albedo (the fraction of sunlight intercepted by the sulfur droplets) which their model uses to compute the effect on the seas. The model contains three different components.
The first component is that of the cooling caused by Pinatubo. Just like the earth, the sea also received less sunlight. The surface water cooled a bit, and cooler water contracts: the sea level drops. In the model of the bottom panel, this is the red curve, labelled OHC (ocean heat content); it shows an expected 6 mm drop in sea level.
The second effect is indicated by the black line, labelled PW. This is ‘precitable water’ and stands for the amount of water in the atmosphere. Colder air contains less water. As the air cooled, post-Pinatubo, the excess water which no longer cold be held by the atmosphere was returned to the oceans. This caused a slow rise in sea level of about 1 mm, followed by an even slower return to zero as the air temperature recovered from the volcano.
The final effect is the amount of water stored on the continents, in lakes, rivers, ground water, and glaciers (TWS, the green line). This varies a lot from season to season and this variability is included in the model. More terrestrial water means less ocean water. Initially the cooling caused a sea level rise, for the same reason at the PW effect: there was less rain. This amounted to about 1.5 mm. But at later times it caused a fall in sea levels, as the cooler weather allowed more water to be captured by the glaciers.
The blue line in the top panel is the sum of the parts. It shows a drop of 6 mm, beginning to slowly recover after 1995. The available tide gauge data could not have picked up such a signal. It is notable that the sea level drops lagged behind global temperatures, which were largely back to pre-eruption values by 1995. The sea had the longer memory. Doesn’t it always? “The days but newly gone, Whose memory is written on the earth” is perhaps one of Shakespeare’s few mistakes: it is better to enscribe memories in the sea.
The sea level anomaly
This 2016 paper mentions that the Pinatubo effect could indeed have caused the anomalous acceleration of sea levels seen during the 1990’s. Their detailed analysis of this was published only in the past month. It is by the same team, and it is perhaps not entirely clear why the work wasn’t finished in the earlier paper.
Accurate satellite measurements started in 1993. If the models are right, the satellite data began right at the bottom of the Pinatubo drop, the worst possible starting point when sea level was lowered by 5-7 mm. The satellite era measured the recovery from this drop, but attributed all the sea changes to the long-term sea level rise.
The figure shows the result of the calculations. The blue line is the satellite data, and it shows the anomalous fast rise up to the early 2000’s, followed by a slower rise, and re-acceleration after about 2009, partly obscured by a short-lived drop in 2011. The red line shows the same data, but now with the Pinatubo effect removed. Effectively, the sum of the three components of the model of Fasullo has been added to the data. The fast initial rise has now disappeared, and the rise remains at a more or less constant rate until 2005 when a slight acceleration sets in, albeit temporarily obscured by the 2008 and 2011 dips. Finally, a model is included for the effect of the ENSO (El Niño-Southern Oscillation)– in effect, this compensates for the El Ninos and La Ninas which also affect sea level. (This last part is the new thing included in this paper – the rest is already in the Fasullo paper.) This correction gives us the green line, which is the final curve for the global long-term sea level rise that they derive. (Green is the colour of jealousy, another association first made by Shakespeare.) The main effect of the ENSO corrections is that it removes the bumps in 1998 and 2016, which coincided with the record El Ninos of those two years. Interestingly, the dip in 2011 is reduced but not fully removed: this had been attributed to a La Nina event but perhaps there was more to it. Whatever it was, it was temporary.
After the Pinatubo effect was accounted for, the sea level rise became smooth, apart from some year-on-year fluctuations. The black line shows the new fit. In the period 1993-2000, it gives a rise of a little over 2 mm per year. This value fits well with that derived from tide gauges in the period before 1993. Problem solved.
The conclusion is that the fast rise seen in the 1990’s, and the subsequent slowing down in the 2000’s, were caused by Pinatubo. Volcanoes don’t just affect the land and air. They suppress the sea as well. Who knew that volcanoes could cause sea-change? Shakespeare foresaw their value for science when he wrote: Suffer a sea-change Into something rich and strange.
Volcanoes and the sea
Was Pinatubo unique? Probably not. The simulated plot shown here suggests that several other volcanoes may have had an impact as well. But one should be careful. There are also decade-long fluctuations which are natural, caused by changing circulations in the oceans. The Agung eruption in the plot has a very long-lasting effect. That is much more likely to trace such a fluctuation. At the current time, sea level is rising so fast that these decadal changes have little impact. In the past it was different, and it is not obvious that volcanic suppression can be distinguished from the effects of changes to ocean circulation.
We can now also look at future projections. The authors find that the sea level rise is accelerating by about 0.8 mm per decade. This is what we need to know in order to calculate how high the sea might get by the end of the century. If only the linear rate of 3.5 mm per year continues to apply, then sea level will get higher by about 30 cm compared to 2015, or 50 cm since the start of the rise around 1900. This is the ‘best case’ scenario, and it would imply that many places could manage but some rich Floridians and some less rich people in Galveston would get wet feet. But with this acceleration term included, it will get wetter: sea level will increase by 50 cm from 2015 levels, or 70 cm in total. Instead of wet feet, people will begin to get wet knees. The numbers fit well with the IPCC projections which were made before this acceleration was measured: the new results exclude the lower values of the IPCC but agree with the likely range.
How high will it go? That is very uncertain. The best estimates are that for every 1C increase in global temperatures, sea level will eventually rise by 1-3 meters. But there is a complicating effect here, because if the Greenland glaciers collapse, the sea will rise by 7 meters just from this. And that is projected to happen if temperatures go up by 2-4C. Such a collapse may take many centuries, though. We should get an inkling in the next two decades: if the observed acceleration accelerates, the likely cause is Greenland and it may be time to cash in on that seaside property. Regardless of whether this happens or not, by the time the next century comes around, the sea will rise by 1cm every year. You would not want to sell flood insurance: every decade will see another previously flood-free area which is flooded following a surge.
I should point out that these are the projections of the models. There is another technique where people look at past correlations between temperatures and sea level, and extrapolate that. Those studies invariably predict larger or faster rises, typically 1 meter or more by the turn of the century. If those calculations are correct, future studies should uncover a faster acceleration. Time will tell – as long we don’t get another Pinatubo playing havoc with the data. Time’s the king of men, and gives them what he will, not what they crave, as Shakespeare said.We need to find the facts, not just hope for the best.
The scars of Pinatubo’s eruption are still there. Where once was the immense Clark air base is now a huge sand plain, left there by the lahars. The Philippines have a unique way to turn disaster into opportunity: the sand is being sold to Singapore for construction. But the sand has also left huge cliffs which are now unstable. According to the local information signs, even a mistimed shout could bring down the walls. The caldera walls are also at risk of collapses. And it will be many years before all the scars have healed.
And now we find that Pinatubo has left scars in our climate record, by suppressing the very sea. It was even responsible for an apparent pause in the sea level rise. When the data becomes sensitive enough, the effect of volcanoes can show up in surprising places. Satellites gave us a vision of Pinatubo’s volcanic sea-change.
Let Shakespeare have the last say, in the words of his Pinatubo of suppressed rage, Hamlet:
to take arms against a sea of troubles
And by exploding, end them.
Albert, March 2018