The Edge of Memory

by Patrick Nunn

  Firstly, when a thick ice sheet develops on a continent, the weight of the ice actually causes the continent to sink; conversely, when the ice melts, the continent rises – a phenomenon known as isostatic rebound. So the coastlines of continents from which ice sheets were disappearing during deglacial times were actually rising (as was the sea’s surface) at the time, making it very challenging to isolate the precise magnitude of actual sea-level rise in such places. For this reason, oceanic islands – generally far from formerly ice-covered continents and with only narrow insular shelves – have for decades been favoured by scientists interested in measuring the exact magnitudes of postglacial sea-level changes.10

  Secondly, water from melting ice sheets on the land did not always easily reach the ocean. Sometimes natural dams formed in the narrow, deep valleys around the fringes of the former ice sheet, causing meltwater to pond behind them – much as today’s intentionally constructed dams do. In other situations, where ice sheets had created massive depressions in the centres of continents, the melted ice could not easily escape to the ocean. In both such situations, enormous meltwater lakes formed in the centres of many formerly ice-covered continents, enduring for sometimes thousands of years before typically breaching a dam or incising a bedrock barrier, which led to their rapid emptying. The meltwater floods poured huge volumes of fresh water into the oceans, often causing abrupt ocean-ecosystem changes and even circulation changes, besides sometimes raising ocean levels quite quickly over a short period of time. Such floods are also implicated in short-lived climate changes that affected the entire planet, posing huge challenges for living things while momentarily, it seems, interrupting the progress of postglacial sea-level rise and the warming that ultimately drove it.

  Dome C in Antarctica is a forbidding place, one of the coldest areas on Earth, with summer temperatures rarely warming above -25°C (-13°F) and winter temperatures often plummeting to below -80°C (-112°F). It hardly ever rains, and we are the only sapient creatures who elect to live. Scientists’ interest in Dome C comes from the fact that it is more than 3,200m (10,500ft) above sea level, one of the highest parts of this massive ice-covered continent, which means that the ice here is thicker than almost anywhere else. In addition, because ice accumulates regularly albeit very slowly in such places, Dome C sits atop one of the lengthiest and most complete archives of the climatic history of Antarctica, each layer of accumulated ice containing information about the climate of the time it was laid down. And because Dome C is so close to one end of the Earth – just 1,670km (1,038 miles) from the South Pole – and because Antarctica is surrounded by ocean, its climate registers the effects of climate changes across the entire planet. Indeed, the record from ice cores through the ice below Dome C has given us solid information about the effects of meltwater bursts on our planet’s past climate.

  In 1996, a 3,270m (10,728ft) ice core was extracted from Dome C by members of EPICA, the European Project for Ice Coring in Antarctica. Analyses of the core allowed insights into the ways in which the Antarctic climate has changed over the last 800,000 years. One of the most remarkable findings was that of temperature, proxied by the hydrogen isotope deuterium. It could be shown that particular episodes of higher temperature here coincided with those of lower temperature in the northern hemisphere. We are not talking about the major swings of temperature that distinguish ice ages from warm interglacial periods, but about shorter duration changes of perhaps a few millennia at most. The unexpected finding that changes in the climates of the northern and southern hemispheres were out of phase led ultimately to a better understanding of the effects of the catastrophic emptying of meltwater lakes.

  One of the earliest periods of rapid temperature change to be recognised during deglacial times, when the Earth was transitioning from ice-age conditions to interglacial ones, is called the Younger Dryas and lasted for perhaps 1,000 years in most parts of the world around 12,000 years ago. In much of the northern hemisphere, the start of the Younger Dryas saw an abrupt return to full-glacial conditions. Conversely, when it ended here, temperatures rose abruptly – almost 10°C (18°F) in 10 years in Greenland – after which the slower warming trend characteristic of the postglacial period resumed. The EPICA core showed something quite different, namely that within the same period of time in Antartica temperatures rose, not fell. Why was this?

  The most plausible explanation for it to date involves the thermohaline circulation of the world ocean on which we depend today (as in the past) for moderating extreme climate conditions in many places. The thermohaline circulation – or oceanic conveyor belt – is driven by variations in seawater density, controlled largely by ocean-water temperature (the thermo part of the name) and salinity (the haline bit), and has effects on every part of the ocean. Much of the thermohaline circulation occurs deep within the world’s oceans, but it comes to the surface in the North Atlantic. Here warm salty water is transported from the tropics north to polar waters, where it loses heat and sinks. Now imagine a vast meltwater lake in the centre of Canada, one of a number of outsized puddles left behind after the melting of the Laurentide ice sheet, which covered this part of North America during the LGM. The lake in question formed along the margins of the shrinking ice and may have covered an area of 440,000km2 (170,000mi2) – larger than any lake in existence today – about 13,000 years ago. As the ice-sheet margins continued melting, so a passage to the sea finally opened for this brimful lake. The resulting outburst flood saw so much fresh water enter the North Atlantic (via the Mackenzie River in the Canadian Arctic Coastal Plain) that it temporarily shut down this key part of the thermohaline circulation, soon bringing the rest of the great oceanic conveyor belt to a grinding halt. This may have been the ultimate cause of the Younger Dryas event.

  The result of the shutdown of the thermohaline circulation would have been abrupt cooling of the northern hemisphere and coeval warming of the south, exactly what polar ice-core data from places like Dome C suggest. The effects of such abrupt climate change on human societies were many and varied, but few were spared its effects. There is a school of thought that regards the Younger Dryas cooling as having forced people into experimenting with agriculture and animal domestication – the onset of cold conditions gave them little choice were they to survive. Necessity is indeed the mother of invention. In support of this suggestion, it has been noted that the earliest known ages for nascent food production in many places do appear to coincide with the Younger Dryas event.

  The effects of the Younger Dryas cooling on the sea level are generally more difficult to detect. In most parts of the world’s oceans, it seems that this rapid cooling temporarily slowed the rate at which the postglacial sea level had been rising. When the Younger Dryas ended, the rate accelerated back to what it had been before. Yet the Younger Dryas was preceded – and the majority scientific view at the moment appears to regard the juxtaposition as fortuitous rather than causal – by a rapid global rise in sea level, known as Meltwater Pulse 1A (MWP-1A). Occurring some 14,500 years ago, MWP-1A involved a sea-level rise of some 15m (50ft) in a mere 340 years.

  There are different ideas about the cause(s) of MWP-1A, with some evidence pointing to Antarctica as the source of the meltwater, and some to North America. One of the most compelling ideas is that the low ice ‘saddle’ between the two major North American ice sheets of the time – the massive Laurentide in the centre and east, and the smaller Cordilleran in the west – may have collapsed abruptly, tipping massive amounts of ice-choked water into both the North Atlantic and the Pacific Oceans, quickly reducing their surface temperatures and raising their surfaces.11

  More recently, about 8,200 years ago, the last of the great North American meltwater lakes – Lake Agassiz-Ojibway – smashed through its ice dam in Hudson Bay, generating another massive injection of fresh water to the North Atlantic and shutting down the thermohaline circulation once again – this time for some 400 years. Temperatures on land masses around the North Atlantic plunged; central Greenla
nd dropped by as much as 8°C (14.5°F) and Western Europe by perhaps 3°C (5.5°F) within this period.12 The evidence for what happened during this 8,200-Year Event is clearer than it is for the Younger Dryas, even to the extent that it has proved possible to directly measure the effects of the outburst flood on the sea level. For example, along the northern coast of the Gulf of Mexico, where the Mississippi Delta lies, the sea level rose perhaps 2.2m (7¼ft) within the 130-year duration of the 8,200-Year Event, but the immediate response was a near-instantaneous 56cm (22in) rise.13

  Ultimately no coasts in the world were immune from the effects of these rapid meltwater-fuelled bursts of rising postglacial sea level, but the record is essentially untraceable in places furthest from where they occurred. The author knows of no clear records of such rapid sea-level rise events from Australian shores, for instance, although faint hints have been found in coastal sediments dating from that time in places like Singapore, the Sunda Shelf (where island South-east Asia now lies) and Hong Kong.

  The point of this discussion is to show that postglacial sea-level rise was not a uniform process, unvarying in rate, but one that was punctuated by bursts of sudden rise, which posed the greatest challenges to living things, including our coastal-dwelling ancestors. While we are on the subject, there is a lesson for the future buried in what might appear fairly obscure musings about the distant past. For if the climate were to change to such a degree that large terrestrial ice masses – like those in Greenland and Antarctica – became unstable, slipping into the world’s ocean, this could indeed cause similarly calamitous changes along many of the world’s coasts.14 History warns us what might happen, but there is no purpose in losing sleep over it.

  Whether the sea rose in swift bursts or more slowly and uniformly, the effects of postglacial sea-level rise on the world’s coastlines varied depending on two things.

  The first of these relates to whether or not the land around which the water rose was itself moving. This makes a big difference, since should a coast already be sinking, sea-level rise will have a multiplier effect. Conversely, should an area of coastal land be rising, this will reduce – it may even reverse – the effects of sea-level rise. During postglacial times, some coasts – relieved of their thick, icy overburden – rebounded upwards, often rising faster than the sea level was rising. For people living along the coasts of Norway and Sweden, for instance, unlike those in most other parts of the world at the time, this meant that the land was emerging relative to the ocean surface – even though both were rising. As an example of how this affected coastal people in these areas, take the Viking settlement at Borre in Norway, a place acknowledged in the Norse Sagas as a royal burial site, and of such great renown among the Norse seafarers that it might come as no surprise to learn that an entire ship was buried there. Active largely in ad 600–1000, the Borre settlement now lies 5m (16ft) above the coast, raised there by isostatic rebound. It is a fossil world that is harbourless and therefore now useless for its original purposes, with the remains of wharves and jetties forming minor features in an inland landscape. Many Scandinavian coasts are still rising, still rebounding, and this action will continue to affect their inhabitants and the ways their descendants live.

  Land can sink as well as rise. Most oceanic islands are sinking, but typically so slowly that they hardly affect the observed rate at which the sea level periodically rises. A contemporary example, from a world in which there has been an upward trend of sea level for more than a century, may make more sense. Many of the world’s longest established coastal or delta cities have a long tradition of drawing fresh water from underground aquifers to quench their inhabitants’ thirst and that of their industry. Over time, extraction of groundwater causes the spaces between the particles in the delta sediments below these cities to collapse, as the water that formerly kept them buoyant is withdrawn. Even when underground aquifers have hardly been tapped for water, cities built on deltas – little more than massive piles of unconsolidated sand and gravel – also tend to sink because the sheer weight of the delta (and the city on top of it) can cause both the underlying solid crust of the Earth to deform and the loose sediments to become compacted. Collapse and compaction at a giant scale lead to gradual ground-surface sinking, a phenomenon that affects cities like Bangkok and Shanghai as much as New Orleans and Venice. For example, between 1956 and 1965, parts of Shanghai subsided at an average rate of 83mm (3¼in) each year, a rate that has somewhat lessened because of aquifer recharge but is still enough to amplify the effects of flooding in the downtown area of the city, where subsidence rates remain highest.

  The second cause of variation in the ways in which postglacial sea-level rises affected the world’s coasts is of course their character, particularly whether they rise gently, perhaps almost imperceptibly, from beneath the ocean surface, or far more boldly. Unfortunately for most people (like the author) who live on the coast, the majority favour coasts that are gently sloping – well designed, it seems, for accessing other places, for growing crops and for building cities. And why ever not? Who wants to build a settlement and roads that cling to a rocky cliffy shore, let alone try and eke out a livelihood there, when other, more salubrious options are on offer? Yet coastal plains, which are commonly low and extensive, are the most vulnerable of all coastal types to sea-level rise. Over the tens of thousands of years across which we can reconstruct the history of coastal change, it is such places that come and go most frequently, often appearing with a flourish of civilisation, then disappearing with its collapse, while coasts that are high and where there is little lowland may have cultures that endure far longer.

  The sea-level changes to which people occupying almost every part of the world’s coastline (except perhaps the Mediterranean and some parts of the North Atlantic) have become accustomed over the past few millennia are quite different to those that characterised the postglacial period, when melting land ice drove sea-level rise. For land-ice melt attributable to rising Earth-surface temperatures linked to Earth’s orbital changes – the pacemaker of the ice ages – is considered by most sea-level scientists to have stopped about six millennia ago. Yet since then the sea surface has continued to change. It has not done so by very much compared to its period of postglacial rise, yet sufficiently at times to have caused noticeable changes to coastal environments – often forcing changes to the ways in which their inhabitants live.

  It is generally considered that sea-level changes within the past 6,000 years have been temperature driven – what are termed steric changes. Within this period, temperature has been the main driver of sea-level change over decades and centuries; when temperatures fall, the sea level falls; when temperatures rise, the sea level will also rise. This close relationship between changes in temperature and sea levels is explainable by two processes. First, when you heat water it expands slightly. Thus, when the upper parts of the ocean are warmed over many years, they expand and the ocean surface rises. Conversely, when ocean-surface waters cool, they contract, occupying less space, so the sea level falls. Secondly, when temperatures rise, there is a net loss of land ice, with the meltwater produced ending up in the ocean and causing the sea level to rise. Conversely, prolonged temperature fall will eventually cause land ice to increase in volume, lowering the ocean surface. The emphasis on land-based ice in both scenarios is important because sea (floating) ice cannot alter the sea level whether it forms or melts.

  Since this period – the last 6,000 years or so – is somewhat marginal to the main focus of this chapter, a couple of examples suffice to demonstrate the effects of comparatively small sea-level perturbations on human affairs, which have become more and more sensitive to such external pressures as we approach the present day.

  Superbly evocative examples come from some of the world’s larger river deltas, places where the history of relative sea-level changes – those reflecting both ocean-water volume changes and land-level movements – can be remarkably complex. Consider the Yangtze (Changjiang) Delta in China
, which was first occupied by people sustained by rice agriculture at least 7,000 years ago. Sea-level oscillations at the end of the postglacial transgression alternately drove people off these lands when the sea level was higher, and allowed them to reclaim their former territories when it fell subsequently. In the latter case, the new settlers favoured wet-rice cultivation in this region, which may have been the place from which rice agriculture then dispersed throughout East Asia.15 The graph below (Figure 4.2) shows neatly how sea-level changes in the Yangtze Delta can be apparently linked to the rise and fall of civilisations there several millennia ago.

  Figure 4.2 Relative sea-level changes and cultural declines (marked by shaded bands) in the Yangtze Delta, China. Note that these sea-level changes are termed relative as they include both ocean-water volume changes and changes in land level, here typically linked to alternate periods of deltaic build-up and sinking. The collapse of the Majiabang culture came about when the high sea level led to groundwater flooding; that of the Songze culture also because of flooding associated with the high sea level, but in addition the intense variability of the ocean surface; that of the Liangzhu and Maqiao because of water-table rise and the expansion of the Taihu Lakes, which caused widespread inundation.

  A similar situation once obtained in northern Europe, where alternations in the occupation of the Vistula River Delta (northern Poland), which empties into the Gulf of Gdańsk, within the last 5,000 years were closely linked to sea-level changes. The rising sea level at the start of this period created coastal wetlands, a time followed by several hundred years of comparatively stable sea level that allowed people to develop an entire system of livelihoods based on offshore fishing and hunting in deltaic forests of alder, hazel and elm. As the sea level rose again subsequently, so the delta shoreline moved landward and the people’s livelihoods became dominated by fishing and sea hunting; it seems to have been just too wet underfoot for most of the year to support similar forests. Agriculture took a hold in the area much later as a gravel barrier, which shielded much of the delta from the ocean, allowed the return of forests that were later cut down to open up land for farming.16