Melting Glaciers: What the Ice Is Actually Doing, and Whether the Cities Drown

In 2006 An Inconvenient Truth put an animation in front of a very large audience: water rising over Lower Manhattan until it covered the ground where the World Trade Center had stood. It was effective, it got repeated everywhere, and it spawned a decade of maps showing Miami, Shanghai and Mumbai coloured blue.

That image is now what most people mean when they talk about sea level rise. A city going under, the way a ship goes under.

The measurements do not say that. They say something less dramatic to look at and considerably harder to live with, and the gap between the two has been useful to absolutely nobody.

The ice is going, and that part is not in dispute

Start with the photograph at the top of this piece. It shows Muir Glacier in Glacier Bay, Alaska, in 1893, as a wall of ice across the water with icebergs in front of it.

As recently as the mid-1980s Muir was still a tidewater glacier, calving bergs from an ice face about ninety metres high. It is not one now. The terminus is around seven hundred metres wide, it has pulled back out of the water, and in the repeat photographs taken from the same spots decades apart the ice simply leaves the frame and vegetation grows where it was.

Muir is unusually well documented rather than unusual. Glaciers have been retreating worldwide since the middle of the nineteenth century, and the rate has increased. None of that is seriously argued about.

What the sea has actually done

Between 1901 and 2018 global mean sea level rose by roughly twenty centimetres. Since the 1970s the rate has averaged about 2.3 millimetres a year. For the decade from 2013 to 2022 it reached 4.62 millimetres a year.

Two things follow from those numbers. The first is that the rise is real and it is speeding up, roughly doubling in rate within living memory. The second is the scale: after more than a century of it, the total is about the length of a shoe.

Where it comes from is split fairly evenly. Between 1993 and 2018, melting ice sheets and glaciers accounted for about 44 percent of the rise, and thermal expansion, seawater simply taking up more room as it warms, for about 42 percent. That second one surprises people. A large share of sea level rise has nothing to do with ice at all.

What it is projected to do

The IPCC's Sixth Assessment gives ranges rather than a number, because the answer depends on emissions. By 2100, on a low emissions path, 32 to 62 centimetres. On a high emissions path, 65 to 101 centimetres.

So the serious projection for the end of the century is somewhere between a third of a metre and about a metre, with deeper uncertainty beyond that if the Antarctic ice sheet behaves worse than expected. There is also a portion already committed, because the ocean responds to warming slowly and the rise between now and 2050 is largely locked in whatever anyone does about emissions this decade.

A metre is a lot of water spread over a planet. It is not a submerged skyline. Nothing in the assessed literature has Manhattan going under.

So why does everyone picture a drowned city

Because the projections get read as a new permanent water line, and that is the wrong mental model.

Sea level does not arrive as a flat rise onto which nothing else happens. It arrives as a new baseline underneath the tides and the storms, and those are what actually put water in the street. Raise the floor by half a metre and every high tide starts half a metre higher, and every storm surge starts half a metre higher, so the flood that used to overtop the defences once a century begins overtopping them once a decade, then once a year, then most months.

This is already happening. High tide flooding, water in the streets on a clear day with no storm at all, has become far more common in the past several decades in places that used to see it rarely.

That is the real mechanism, and it is worse than it sounds precisely because it is boring. A city does not get a single dramatic night it can rebuild after. It gets a slow increase in the number of days the tunnel floods, the basement floods, the substation trips and the insurance gets repriced, until living there costs more than it returns. People leave over decades, in ones and twos, for reasons they describe as personal.

The part that is missing from the maps

There is a second variable that the blue-shaded map does not include, and in several cities it is the larger one. The sea is rising. The land is also falling.

Jakarta is the case that makes it unarguable. Northern districts of the city have been subsiding for decades, and the principal cause is not climate at all. It is groundwater. A large city that never built out piped water at the necessary scale pumped what it needed from beneath itself, the aquifers compacted, and the ground went down. The rates measured there dwarf the few millimetres a year the sea is contributing.

Indonesia's response has been to move the capital. That is the scale of the problem, and note what it implies: the most convincing example of a major city being lost to water in the modern era is substantially a story about wells and pipes, not about ice.

Jakarta is not alone in this. Subsidence from groundwater and hydrocarbon extraction, and from the simple weight of construction on soft delta sediment, is a significant part of the relative sea level change in a number of coastal cities. It is also, unlike the thermal expansion of the entire ocean, something a city government can actually do something about within one political term.

Where the frightening version is roughly correct

Everything above is about cities with metres of elevation and the money to build defences. For some places the alarmist framing is not alarmist.

The Maldives is the lowest country on Earth. Its average natural ground level is about 1.5 metres above the sea and its highest natural point is around 2.4 metres. More than eighty percent of its land sits less than a metre above the water. Put the high-end projection for 2100 next to those numbers and there is no clever engineering argument left, only a question about how much of a country can be raised, walled or abandoned, and who pays. Tuvalu, Kiribati and the Marshall Islands face the same arithmetic.

The same is true, at enormous scale, of the great river deltas. Tens of millions of people live within a metre or two of sea level in the Ganges-Brahmaputra, the Mekong and the Nile deltas, on land that is also subsiding and where the defences are thin.

So the honest position is not that the sea is a minor problem. It is that the danger is very unevenly distributed, and that the places carrying most of it are not the skylines in the animation.

What the honest version sounds like

The people who say the ice is fine are wrong, and the photographs settle it.

The people who say New York will be underwater are also wrong, and the assessed projections settle that. What they have done, with good intentions, is set the bar at an image so extreme that when it fails to arrive the whole subject looks exaggerated, which is a gift to nobody.

The accurate version is duller and harder to argue with. The sea is rising, it is rising faster than it was, a good deal of the next few decades is already committed, the damage comes as flooding that gets more frequent rather than water that arrives and stays, and in some of the worst-affected cities the ground is dropping faster than the water is climbing for reasons that have their own separate fix.

Muir Glacier is still worth looking at. Not because its ice raised the sea by any measurable amount on its own, but because somebody stood in Glacier Bay in 1893 and photographed a wall of ice that is not there now, and there is no version of that photograph in which everything is fine.

Sources

• Wikipedia, "Sea level rise" (roughly 20 cm of rise from 1901 to 2018, about 2.3 mm a year since the 1970s, 4.62 mm a year for 2013 to 2022, the 44 percent ice and 42 percent thermal expansion split for 1993 to 2018, the IPCC AR6 ranges of 32 to 62 cm and 65 to 101 cm by 2100, the lag that commits rise to 2050, and increasing high tide flooding).
• Wikipedia, "Muir Glacier" (a tidewater glacier calving from a 90 metre ice face as recently as the mid-1980s, now about 0.7 km wide at the terminus, and the repeat photography showing the ice leaving the frame).
• Wikipedia, "Retreat of glaciers since 1850" (the worldwide pattern of glacier retreat and its acceleration).
• NOAA, "What is high tide flooding?" (flooding on clear days without storms, and why a rising baseline increases its frequency).
• Wikipedia, "Jakarta" (land subsidence along the northern coast, groundwater extraction and urban development among its causes, and the resulting coastal flood risk).
• Wikipedia, "Nusantara, Indonesia" (the relocation of Indonesia's capital away from Jakarta).
• Wikipedia, "Maldives" (the lowest country on Earth, average natural ground level about 1.5 metres and maximum about 2.4 metres, with more than 80 percent of its land under a metre).
• Wikipedia, "An Inconvenient Truth" (the 2006 film and its sea level sequence).