The day after tomorrow

No, it's not a movie. This is a certainty. It will take billions to prevent it. Global warming is melting the ice cap. The sea is coming to get us, and when it does it won't even be worth fishing in

At 7.27pm on Saturday, January 31, 1953, a steam train bound for King's Lynn came to a sudden halt at Hunstanton railway station. It had hit a bungalow. All down the east coast, from Lincolnshire to Kent, the night was full of shocks. So many things were displaced, it was like a visitation from poltergeists. In Gorleston, the cinema manager halted a showing of High Noon to warn his audience that furniture was floating from their houses. In Heacham, beach huts landed half a mile inland near the station. People dreaming of goose-feather beds found themselves on gale-whipped rooftops, fighting for their lives while 9ft of seawater ripped out their parlours.

It was not like the Asian tsunami - not a sudden knockout that came from nowhere and that nobody could have expected. The North Sea has always been an edgy neighbour, best kept behind walls. The floods of 1953 are remembered as England's worst-ever natural disaster. From Lincolnshire to Kent, more than 300 people died, 24,000 houses were flooded and 40,000 people evacuated. But "worst ever" is a superlative example of memory loss. There were catastrophic sea floods in 1938, 1928, 1897, 1703, 1663, 1570, 1287, 1236 and 1099. In AD38, it was said, 10,000 people drowned along the east coast and in the Thames estuary.

Any day, the surge could come again. The wind piles in from the west, shovelling thick purple cloud over the faraway coast of Lincolnshire. I stand on the southern edge of the Wash, ground zero in the great inundation of 1953, looking out across this shallow, choppy inlet of the North Sea. In pubs all along the Norfolk coast, walls are hung with 50-year-old photographs showing telegraph poles jutting through swells, or boats calling at bedroom windows. On walls and bridges far inland, there are markers showing how high the water rose. On the Environment Agency's new national flood map, a blue tint shows how far it might reach next time. In the frontline villages, sirens stand ready to sound retreat.

Every year the sea here rises by six millimetres. A quarter of this is due to the long, slow tilting of the land after the last ice age; the rest is rising sea, fed by meltwater and thermal expansion as the climate warms. It might not sound much, but think what 6mm adds up to when spread across the entire expanse of ocean. Think what it will add up to in 50 years (nearly a foot). Think what such a weight of water will add to the height and power of the waves; how much further it will drive on up the valleys. In 1953 the storm surge raised sea levels in north Norfolk by 3.43 metres and at Southend by 2.74 metres. Metres! This was colossal - even now, a rise of just 60 centimetres is classified by the Met Office as a "surge event", with flood warnings issued via the Environment Agency when it coincides with a high tide.

At the time, the 1953 surge was regarded as something likely to recur on average only once every 120 years. But times and perceptions change. Sea levels have risen, storms have increased in both violence and frequency, and - though coastal defences have been strengthened - coastal dwellers near shore level are like Neapolitans under Vesuvius. It's not a question of if, but when.

The when and the what, however, wriggle through your fingers like salmon. It's a safe prediction that the nasty neighbour is going to get nastier, but the Intergovernmental Panel on Climate Change (IPCC) offers a bewilderingly wide range of possibilities. Global mean sea level in the 21st century, it says, could rise by anything between 9 and 88cm. The number of uncertainties - volume of greenhouse-gas emissions, rate of ice-melt, accuracy of climatologists' computer models, effectiveness of coastal defences - creates such a cacophony of statistical noise that you can't tell which tocsin is ringing the loudest. Around the UK, the prediction is that storms at sea will not only grow and multiply but also change track, with the result that some places (the Bristol Channel, for example) may look forward to a calmer future while others (the southern half of the North Sea) are in for a smiting.

The Met Office's expert on sea-level rise, Dr Jason Lowe of the Hadley Centre, is working on a "medium-high" emissions scenario that projects changes into the 2080s. By then, he says, a severe surge at high tide will be a much wilder beast than the equivalent storm today. Even in the relatively safe haven of the Bristol Channel it could be 20cm higher. In the English Channel it could be half a metre, and in the howling wastes of the southern North Sea an awesome 1.1 to 1.2 metres. These particular figures, admittedly, are for storms of such severity that they are likely to occur only once every 50 years. But here, too, the horizons are dissolving and re-forming.

"In the southern North Sea," says Lowe, "by the 2080s, a typical return period for what is now a 150-year event will be seven or eight years."

Add to flood risk the companion peril of erosion, and entire coasts are barnacled with anxiety. A few miles to the east and south of me, villages are poised almost literally in suspension as their cliffs tumble house by house into the sea. A couple of hundred yards to my left, two Environment Agency bulldozers are straining for the umpteen-hundredth time to reshape the shingle bank that protects the freshwater marsh, coast road and Salthouse village. Less than a fortnight later, in mid-February, their work will be undone by a storm.

Round the corner in the Thames estuary, the Office of the Deputy Prime Minister (ODPM) plans 85,000 new homes along 40 miles of flood plain. Most of this, most of the time, is kept dry by flood defences, but in February the Association of British Insurers fired a maroon. Existing defences were unlikely to be adequate in a globally warmed future, it said, and the threat in some places was so serious that houses would have to be redesigned like tropical fishermen's huts, with all their living space upstairs. Unless its members could be reassured that the tide would be tamed, the new homes would be uninsurable.

The Thames estuary is a perfect example of marine perfidy. The water may be rising, but the graph is irregular. Since its completion in October 1982, the Thames barrier has been closed against surges 90 times - a rough average of four a year. But in some years (1984, 1986, 1989, 1991, 1997) there were none at all. In the record year of 2003 there were 18. The graph shows an alpine profile of needle-sharp peaks and troughs but the trend is towards progressively higher peaks, with the three busiest years all coming since 2000. At the same time, between 1999 and 2002, the mean sea level at Sheerness rose 10cm above Met Office predictions. Londoners quaked like Venetians. How long would the barrier hold? How long before the Underground flooded? How long before the sewers, water mains, power and communications systems would be washed away? Around the water margins, straight questions float on seas of equivocation. As if to underline the uncertainty, 2004 passed with only two closures of the barrier and the water at Sheerness dropped back to its predicted level. Nobody knows why.

And yet we are encouraged to believe that the drowning of Westminster is about as likely as the Queen getting caught in a beam trawl. No part of the UK will be more heavily defended than London. The existing defences could hold out for decades, possibly even for centuries, but the risk of catastrophe becomes progressively less remote. With ODPM breathing down its neck, the Environment Agency is working on a long-term management plan for the estuary that will employ every kind of defence short of intervention from St Peter. The existing barrier will be upgraded or replaced when it reaches the end of its design life in 2030; riverside defences will be heightened; parks and nature reserves will be designed to act as spillways and storage lakes; drainage systems will be enlarged; new buildings will be made flood-resistant. There might even be a new outer barrage incorporating a road and a tidal power plant.

The agency will take its time - the first draft of its detailed plan is not expected until 2008 - but the stakes could hardly be higher. More than 1.25m people already live or work in the flood plain, and £80 billion worth of property stands at risk of damage by flood. In the flood zone are 68 London Underground and Docklands Light Railway stations, 30 mainline railway stations, three world-heritage sites, eight power stations, 16 hospitals and 400 schools. With magnificent understatement, the Environment Agency predicts that the cost of protecting all these, as well as the thousands of new homes to be built along the estuary, will be "substantial" - £4 billion is the current best estimate - and the question of who pays for it all (government? Property owners? Developers?) is not the least of the problems in search of an answer.

If the agency gets its calculations right, and the sea behaves as it is supposed to, London will flood on average no more than once every 1,000 years. Given that even the IPCC's figures are in pencil and not indelibly inked, it's not easy to know who - other than Londoners 1,000 years hence - will be in a position to say whether or not they got it right. You pour your concrete; you ride your luck.

Other places simply have no luck left to ride. Including all its kinks and curves, the English shoreline is about 9,000 kilometres long (5,580 miles), with some 1,900 kilometres (1,187 miles) protected by man-made defences. Just over half of this is to keep the water out; the rest is to stop erosion. Without improved defences, more than a million properties with a capital value of £130 billion will be at risk of coastal and tidal flooding, and another 100,000, valued at £8 billion, threatened by erosion. A further 100,000 homes are at risk around Scotland. In 70 years' time, without reinforcement, huge lumps of Norfolk, Cambridgeshire, Lincolnshire, East and North Yorkshire, Lancashire, Newport, North Somerset, East and West Sussex and Kent are likely to suffer annual damage averaging between £40,400 and £2m per acre at 2001 values.

Some prime examples can be seen from where I stand. The shingle bank, which the tide has raked out westward to Blakeney Point, is of natural origin but needs endless grooming to hold its shape. It is overtopped on average once every three years, but this could accelerate to once a year if a new strategy put forward by the Environment Agency is approved. An earlier plan to build a flood bank across the marsh failed on economic grounds. The criteria for coastal defences are pitiless and pragmatic. The Department for Environment, Food and Rural Affairs (Defra) operates a qualification system that awards points for economic benefit (property saved must be worth at least as much as the cost of protecting it), human suffering (how many lives or homes might be lost) and the value of the natural environment. Economically deprived areas will score more highly than affluent middle-class ones that are deemed better able to look after themselves. Salthouse had no hope. The idea now is to lower and flatten the shingle to make a more solid barrier against heavy surges, and to improve drainage from the marsh into the river so that flood water is returned more quickly to the sea. The underlying logic is that when the next really big one comes, nothing's going to stop it anyway.

The government has budgeted £570m for flood and coastal defences in 2005-6, though some coastal communities would be better off buying lottery tickets. The result of Defra's points system has been a competitive bidding process in which seaside villages, even those with tourist industries, have little chance against towns and cities on flood-prone inland rivers. Of the £570m, only £47m is earmarked for the coast. Factor in the cost of new defences - currently running at between £3m and £5m per kilometre, and likely to double or quadruple in the face of climate change - and you see why people are filling sandbags. It adds to the problem that no single authority has overall control. Flood plains are the responsibility of the Environment Agency. Cliffs are the local authorities'. It makes no sense. In the perpetual cycle of give and take, material from crumbling cliffs is stolen by the current to replenish the beaches that shield the flood plains further south. There are houses on the cliffs that will be lost to erosion; houses behind the beaches that are threatened by floods. One policy may not fit all, but a plethora of competing policies, each defending its own, has been disastrous.

Eastward, the shingle gives way to soft sandstone cliffs that mark the beginning of one of the most hyperactive coastlines in Europe. On the sea bottom already lie the drowned villages of Shipden, Wimpwell, Waxham Parva, Ness and Newton Cross. In neighbouring Suffolk, the medieval town of Dunwich is already half a mile out. Others reluctantly are following them over the cliff, clinging by their fingernails like sailors to a foundering wreck. Happisburgh, the most directly threatened, has lost 26 properties in 15 years, and in the next century will lose another 35, plus its caravan park, manor house and church. None of this in Defra terms justifies the cost of defending it.

It's much the same in East Yorkshire, where the Holderness coast has retreated 32 kilometres (20 miles) in a million years and 400 metres in 2000. It, too, looks out upon the watery graves of ancient villages. Famously in June 1993, Scarborough, North Yorkshire, lost the four-star Holbeck Hall hotel when the South Cliff keeled away beneath it. The Channel coast, too, is worming its way north. Chichester harbour was three miles further out when the Romans sailed into it. Studland in Dorset is losing five metres a year. Even the white cliffs of Dover are surrendering half a metre each year to winter falls.

Back in Norfolk, the solid, prosperous and seemingly impregnable village of Overstrand, with its promenade and popular holiday beach on the fringe of Cromer, will have had its heart ripped out by mid-century. Abandoned to the sea, the existing hard defences will have diminishing effect, reducing to zero in approximately 20 years. For the people who live in places like this, the agony of loss is no less than if they had been hit by bombs.

What they have actually been hit by is experts. Ironically, it is in the very act of making common cause that the responsible bodies - district councils, the Environment Agency, English Nature - have wreaked the most havoc. North Norfolk, for example, is zoned within the Anglian Coastal Authorities Group (ACAG), which reaches all the way down the east coast from the Humber to the Thames. It is one of 18 such groups that now cover 98% of the English and Welsh coastlines, and whose task is to draw up Shoreline Management Plans setting out which bits of coast they will defend, and which they will let go. The reasoning goes like this: We can't clad the entire coast in concrete, so we have to be selective. In Norfolk this means concentrating our efforts on the important coastal towns - Lowestoft, Yarmouth, Cromer, Sheringham - and, with varying degrees of delayed resistance (existing hard defences will remain effective for as long as it takes them to crumble), letting nature decide what happens to the rest. For Lowestoft to keep its beaches, it needs to be fed the sediment from Overstrand's and Happisburgh's collapsing cliffs. To fortify village frontages would be to create barren, beachless headlands as the softer coast eroded around them, and the headlands themselves would prevent the transport of sediment so that the plight of beaches would get even worse. Past lunacies are exemplified by the stretch between Eccles and Winterton, where a concrete revetment became so dangerously undermined by the churning of the tides that 150,000 cubic metres of sand now has to be piped every year from the sea bed, at a cost of £2m.

Offshore sand- and gravel-dredging infuriates coastal campaigners, who argue that it reduces tide-borne sediments and robs their beaches. It is improbable that anyone can prove this beyond doubt. Claim and counterclaim both bear their own logic, though what's certain - as the drowned villages testify - is that the problem of coastal erosion existed long before the dredging started. What is true is that people feel themselves assailed by more than just the unstoppable forces of nature. Shifts in public policy are not acts of God: they are human interventions for which nobody is responsible but the people who make them. It is only by an accident of timing that the entire burden of loss will fall upon a particular generation of homeowners, which is why North Norfolk's MP, the Liberal Democrat Norman Lamb, has raised the issue of compensation. This is a subtler move than it looks. Paying compensation would increase the public cost of abandonment and so materially alter the calculation on which the fate of each village hangs.

There may be some softening of the line as a result of a new Defra consultation exercise whose results are due at the end of the month.

In the long term, it may make it less difficult for rural communities to make their case. But one way or another the tide will have its way. The sea will make cliffs; it will make marshes, dunes and beaches. And what it makes it will take away.

To yield to its dynamic, allowing it where possible to find its own balance, may be as close to wisdom as we'll get. But wisdom should not be at the cost of fairness. If we are to armour the Thames for the future inhabitants of Prescottgrad, then let us not forget the existing populations of flood plain, estuary and cliff.