When Science Goes Wrong (10 page)

Trees fell on roads and rail tracks in unimaginable numbers, making them impassable. Among the millions of Britons who found themselves trapped in their homes or unable to get to work were Anita Hart and her husband; at least their son’s hurricane warning had caused them to cancel their holiday. ‘Our caravan would have blown over,’ Anita said. Indeed, many caravan parks near the coast were scenes of complete devastation.

Trees also fell on power lines, causing outages. In fact, by 3.30am nearly the entire southeast of England, including most of London, had been plunged into darkness that lasted until dawn. With nowhere to send their electricity, power stations had to shut down abruptly. Many communications links were broken, either by trees falling on lines or by the collapse of broadcasting towers.

Although the immediate harm of the treefalls lay in the damage they caused, it was the sheer loss of trees that most affected Britons in the longer run. South-eastern England had not experienced such a violent windstorm in nearly three centuries; thus many trees had had time to grow to maturity and beyond without ever being severely tested by the elements. Majestic, ancient trees were a beloved and seemingly permanent feature of the English landscape, but that changed in the Great October Storm. Among the places hard hit were the Royal Botanical Gardens at Kew in West London, Britain’s premier arboretum and botanical garden. Five hundred trees – about one-third of the garden’s stock – were destroyed, and an equal number were severely damaged. Among the trees lost were many rare specimens, such as a cherry-bark elm (
Ulmus villosa)
– a species that is being harvested out of existence in its native Kashmir. Even greater damage occurred at the Royal Botanical Gardens’ second site at Wakehurst Place in Sussex, where more than half the trees were downed. Other arboretums, parks, private gardens, and woodlands suffered major losses. There was also severe damage to commercial forests: losses equalled about two years of lumber production for the entire United Kingdom.

Treefalls were the most prevalent type of damage done by the storm, but homes also suffered damage from chimney falls, stripping of roofs and sidings, and the like. One in six homeowners in southeast England filed insurance claims after the storm, and total insured losses (for Britain alone) amounted to £1.4 billion.

For meteorologists, in particular, it was a memorable night. Some were at their homes and lived through experiences like the one described by HD Lawes earlier. One meteorologist lamented that he slept soundly through the weather event of the century. Others were on duty. Those at Bracknell had to cope with a power outage that halted the Cyber 205 in its tracks around 4am. Although any forecast for mainland Britain was now superfluous, the computer was also programmed to provide individualised, automated forecasts for the oil rigs in the North Sea. Thus, as the storm roared toward the rigs, the night staff had to hand-process, type and fax the forecasts – an operation that took many hours. (One rig narrowly escaped disaster as a ship that had broken away from its moorings bore down on it: the ship was pulled away just in time.)

Most of the Bracknell day staff failed to show up, so the night staff had to continue on duty. One person who did show up was Ewen McCallum. He had prepared a summary of the week’s forecasting for a conference that day. This task had been making him nervous all week because he was new on the job and also because there had been no weather of interest to talk about. When he arrived, his boss said, ‘Never mind the bloody conference, the shit has hit the fan and you’re going to be on in an hour to explain it!’

In fact, it soon became clear that the Met Office would have a lot of explaining to do, because the storm provided a field day for the media, especially the tabloid press.

WHY WEREN’T WE WARNED? screamed a headline. The
Daily Express
spoke of the ‘complete, shameful devastation of the credibility of those smugly useless TV weather people’. The more staid
Independent
spoke of the ‘dead silence from the frog-spawn watchers,’ and the
Guardian
described the Met Office staff as being ‘at the top of everyone’s list of duffers’. Many papers ran stories alleging that the French and Dutch meteorologists got the forecast right, unlike their incompetent British counterparts.

The quotes just mentioned are taken from an analysis of press responses published by Met Office staff. The analysis also mentioned that the Met Office received hundreds of letters from members of the public, the overwhelming majority of which expressed ‘support for the Office and appreciation of its services’. This is a bit too selfserving to believe, but Britons do have a reputation for siding with underdogs. At least one highly critical letter did arrive. It was from the BBC, which accused the Met Office of damaging the corporation’s reputation by failing to provide adequate warning of the storm.

Luckily for the Met Office, the Great Storm was quickly driven off front pages by an even less-heralded disaster. The following Monday was Black Monday, when stock markets collapsed around the world. The London market lost £50 billion in total share value on that day alone, and by the end of the month more than a quarter of the London market’s value had been wiped out. Many Britons whose homes had escaped the wrath of the storm now faced even worse damage to their bank accounts.

Of course, there had to be an investigation into what went wrong with the storm forecast, and who better to conduct it than the Met Office itself? In fact, the person who was given the main job of reviewing the forecasting process was the head of Central Forecasting, Martin Morris. Not surprisingly, the report laid the blame on a factor that no one at Bracknell could be held accountable for – the shortage of data from the Atlantic. In fact, about the only person who was singled out for criticism was Michael Fish (for his ‘particularly unfortunate’ ‘no hurricane’ comment), but even he wasn’t mentioned by name.

By way of bolstering their assertions about the cause of the erroneous forecast, the meteorologists described how they repeated the run of the fine-mesh model that had been most in error (the one that ran after midnight on October 15 and which influenced Fish’s ‘no hurricane’ forecast at lunchtime). This time they added the extra data that had not been available at the time. According to the report, the model now came up with a good prediction of the storm’s track and intensity. But the researchers admitted that they achieved success ‘by selecting values of the adjustable parameters’. When researchers model events that happened in the past, almost any model can be made to predict the actual outcome if one twiddles the knobs for long enough: this, it seems, is what the Bracknell group did.

Perhaps aware that an internal investigation might not satisfy the Met Office’s critics, the Secretary of State for Defence commissioned a review of the report’s findings by a well-known mathematician, Peter Swinnerton-Dyer, and a university-based meteorologist, Robert Pearce. These two were somewhat freer in their comments than were the authors of the internal report. They criticised the duty forecasters at Bracknell for following the computer models too closely and for failing to recognise a situation in which the models were likely to underestimate the strength of the winds. Swinnerton-Dyer and Pearce also compared the British Meteorological Office unfavourably with its French and Dutch counterparts. The French meteorologists, they said, were better educated and trained and had faster computers, and as a result did a better (though still imperfect) job of forecasting the October storm. (Jacques Siméon of Météo France confirmed this to me.) Some of Swinnerton-Dyer’s and Pearce’s recommendations were implemented: most notably, the Met Office was promised a new supercomputer, which was finally installed in 1991.

Thomas Jung, the ECMWF scientist who recently reanalysed the 1987 storm, believes that weaknesses in the Met Office’s computer models, and not problems with the data, were responsible for the erroneous forecasts. When Jung applied the ECMWF’s current model to the then-available data, he got an excellent prediction of the track and intensity of the storm for several days beforehand (though for some reason the prediction broke down during the final few hours before the storm). Jung tested the sensitivity of the model’s prediction to slight variations in the data, such as might be caused by erroneous or absent measurements at some locations. He did this by running the model 50 times, each time using slightly different starting conditions. The model robustly forecast the storm in spite of these perturbations. This kind of ‘ensemble forecasting’ is now standard practice, because it gives the forecasters a measure of how confident they can be in the model’s predictions. Thus it helps avoid situations like the one that occurred in 1987, when the models were in conflict but there was no objective way to assess which one was closest to the truth.

How did the Great October Storm measure up to other historic windstorms? The previous great storm of this type to strike southern England occurred on November 26 and 27, 1703. The effects were described by novelist Daniel Defoe (of
Robinson Crusoe
fame) in his 1704 book
The Storm, or a Collection of the Most Remarkable Casualties and Disasters which happen’d in the late Dreadful Tempest both by Sea and Land.
The storm followed very much the same track across southern England as did its 1987 counterpart, and was probably similar in magnitude, but it did much greater damage on account of lower building standards in those times. Some details recorded by Defoe, such as the stripping of the lead roof of Westminster Abbey, suggest that winds may have been even stronger than in 1987. About 8,000 people died: these included the Bishop of Bath and Wells, who was ‘found with his brains dash’d out’, and 1,500 sailors who drowned when a dozen ships of Queen Anne’s navy were sunk.

Actually, windstorms nearly as powerful as the 1987 event cross the British Isles quite frequently, but they nearly always affect areas much farther to the north, such as the Hebrides Islands or the Cairngorm Mountains of Scotland. These attract very little attention because they cause next to no damage. ‘They just blow the sheep backward,’ as Bill Giles put it. Remarkably, another very powerful windstorm swept across England just 27 months after the Great October Storm. The storm of January 25, 1990 was accurately forecast, but because it struck during the day and affected a wider area, it caused more casualties: 39 people were killed in Britain and about 50 in Europe. Another three million trees were destroyed in the UK, adding to the toll of the 1987 storm.

The most disastrous European windstorm on record occurred on January 16, 1362. Known as the Grote Mandrenke or ‘Great Drowning’, it raised a wall of seawater that surged for miles across the coastal lands of the Netherlands and neighbouring countries. At least 25,000 people drowned and the coastline was permanently altered. The loss of life would doubtless have been much greater, had not the Black Death already killed off about half the population.

The Great October Storm is now receding into history. A generation of Britons has grown up with no memory of the storm itself, nor of how southern England looked before it struck. Perhaps an occasional young person may wonder why Ashdown Forest barely merits being called a forest, or why a town with but one ancient oak would call itself Sevenoaks, but that is it. And gradually, new trees are replacing the old.

Michael Fish retired in 2004. In spite of his performance on the day of the Great Storm, he received numerous honours for his long service to the BBC. These included an honorary degree from City University, which may have made up for his lack of an earned degree at that institution. The Queen named him a Member of the Most Excellent Order of the British Empire. And after leaving government service he was free to capitalise financially on his fame – or notoriety. A few months later he was on the radio advertising central heating systems. ‘A woman rang to say there was a hurricane on the way,’ he told listeners. ‘Well, I couldn’t care less, because I have this new wonderfully efficient Worcester central heating boiler, so whatever happens I’ll be as warm as toast and have loads of hot water.’ In 2006, perhaps yearning to be back in front of the cameras, Fish tried out for ITV’s
Celebrity X Factor
. He failed to win a spot after Simon Cowell poured faint praise on his rendition of ‘Singing in the Rain.’

VOLCANOLOGY: The Crater of Doom

VOLCANOES DEMAND RESPECT, but they don’t always get it.

In 1993, when geologist Stanley Williams led a party of scientists to their deaths in the crater of an active volcano, he triggered an eruption of controversy and blame.

The volcano in question was Galeras (“the galleons”), a 14,000ft-high peak in the
cordilleras
of south-western Colombia. On its eastern flank, just four miles from the summit but 5,000 feet below it, lies the city of San Juan de Pasto. Although it has a population of more than 300,000 and is the capital of the Colombian department of Nariño, Pasto is a fairly sleepy provincial town that is largely cut off from the bustling metropolitan centres to the north. If anything keeps the citizens of Pasto on their toes, it is the rumblings and occasional eruptions of their local volcano.