WARMING’S LINKS WITH DISASTERS? CHANCE IS ALSO A POWERFUL FORCE
▶ It’s easy to blame recent disasters on climate change but the stats just don’t add up. Robert Matthews reports
Will it never end? Hurricanes in the Caribbean, a huge earthquake near Mexico, devastating floods in South Asia. And those are just what our home planet has thrown at us.
In recent days the Sun has joined in, blasting the Earth with radiation in the solar equivalent of several Category 5 storms. It disrupted radio communication and produced auroral displays over North America’s southern states.
Such a litany of mayhem in so short a time may be unprecedented. It has already had a devastating impact on the lives of millions. Even those of us not directly affected have been left with an eerie sense of approaching apocalypse.
Scientists are divided about the significance of it all. Some insist that the spate of severe hurricanes in the Caribbean cannot be put down to chance. They argue that it must be the result of climate change, with global warming spawning more violent hurricanes.
Others are equally vehement that, while consistent with a warmer world, there are dangers in reading too much into recent events.
The most reliable evidence of a link comes from analysis of records dating back a century or more – and this shows no sign of a trend towards increasingly violent storms.
Demonstrating cause and effect is one of the most difficult problems in science, not least because of the ability of pure chance to produce spurious patterns.
Some years ago, a Harvard graduate student set up a computer program that scoured the web for random collections of data. Using the same techniques routinely used by researchers searching for correlations, the program found countless statistically significant but meaningless relationships.
In one example, it revealed that if the number of swimming pool accidents in the US annually were plotted out over a decade, they were strongly correlated to the number of films starring Nicolas Cage.
Of course, extend the data set far enough and the correlation vanishes. Swimming pool accidents predate even Cage’s long career.
But the problem with many data sets, such as those linking global warming to its supposed consequences, is that it remains unclear just how long it takes for the real signal to emerge from the random noise.
This is why most climate scientists remain reluctant to link the recent spate of hurricanes to global warming, despite having few doubts that global warming is real and makes such events more probable.
But what of the confluence of the other disasters, such as the worst floods in Bangladesh and neighbouring states for a decade, and the strongest earthquake to strike Mexico since 1787?
Again, the trickery of randomness is the most likely explanation. Chance events are surprisingly prone to appearing in “runs”, even when completely unconnected.
Toss a coin once a day for a month and probability theory shows that you’re likely to witness runs of about half a dozen heads or tails – enough to raise suspicions that there’s something odd going on.
Add in the sheer plethora of natural disasters that strike in any given year and the potential for runs among them becomes considerable.
In short, seeing significance in the global spate of disasters makes no sense. It’s hard enough to make a case for the recent rash of strong hurricanes being a “smoking gun” of climate change.
It’s also beside the point, because whatever their cause these disasters still happened. But a rational analysis of their effect reveals grounds for optimism about the future.
Devastating hurricanes were a feature of our planet long before the emergence of human beings. Even if the world’s governments agreed to stop burning fossil fuel tomorrow, hurricanes would remain a major threat in the tropics.
Yet in the past few decades, impressive progress has been made in mitigating impacts.
Hurricanes can no longer strike without warning, as one did in Galveston, Texas in 1900, killing about 10,000 – the deadliest in US history. Their paths can be predicted with reasonable accuracy days in advance, allowing communities at risk to be alerted.
After Hurricane Andrew in 1992, Florida’s construction codes were upgraded, making buildings far more resistant to damage and again saving lives.
It is a similar story with earthquakes. Increasing numbers of countries now insist on quake-resistant construction, thus countering the major cause of deaths in earthquake zones – homes that collapse.
Mexico City also has a quake-detection system that gives about 60 seconds of warning before an impending strike – enough time for those at risk to take cover. This helped to keep the toll from this month’s colossal quake to about 100. When the city was struck by a weaker event in 1985, it killed tens of thousands.
But there is one glaring exception to this trend – the impact of the recent floods in South Asia. The death toll in India, Bangladesh and Nepal has exceeded 1,000, and millions have been left homeless.
The root cause is neither the vagaries of nature nor the environmental impact of human activity. It is much simpler than that, and much harder to solve – poverty.
When your existence demands that you eke out a living in harm’s way, there is little that satellite imagery or computer forecasting can do to protect you.
Demonstrating cause and effect is one of the most difficult problems in science, not least because of the ability of pure chance to produce spurious patterns
Robert Matthews is visiting professor of science at Aston University, Birmingham, UK