A clearer outlook

It’s a thankless task being a weatherman. While an accurate forecast barely causes a murmur, get it wrong and you may as well be personally responsible for whatever climatic disaster has befallen your audience. If any reminder was needed, just rewind two decades to Michael Fish’s infamous assertion that a hurricane was the last thing on the cards.

Fortunately, things have moved on a bit since the events of 1987. Today’s meteorologists have at their disposal an arsenal of sensors, satellites, computers and numerical techniques that provide a more detailed picture than ever before on what is happening in our atmosphere and what is likely to occur in the future.

It’s just as well, because as the technology has become more complicated, so, some would claim, have our weather patterns. And with climatic aberrations likely to become even more commonplace as the impact of global warming is felt, our forecasters must fight an increasingly advanced battle against the elements in their efforts to help us plan our daily lives.

At the forefront of this battle is the UK’s Met Office, one of the world’s most respected climate research agencies. Dr Alan Dickinson, the organisation’s director of science and technology, is the person responsible for deploying and making sense of the technology designed to improve our daily, weekly, monthly and even seasonal forecasts.

An applied mathematician with a PhD in fluid dynamics, Dickinson joined the Met Office in 1975. A variety of roles in research and IT saw him become instrumental in the group’s adoption of parallel computing operations and the use of supercomputers. Indeed, earlier this year the delivery of an NEC supercomputer put the Met Office on the list of the world’s top 500 computing sites.

Dickinson heads a team of around 400 researchers and scientists, most of whom are based at the science and technology office in Exeter. Other facilities are shared with a number of UK universities. For instance, the group operates an atmospheric research aircraft that it makes available to the academic community through the National Environmental Research Council (NERC).

At the heart of the approach, however, is the use of satellites — from the geostationery systems used to take snapshots of the weather, to orbiting satellites that deploy sounding devices to acquire more detailed information about the state of the atmosphere. This understanding, said Dickinson, is the key to making an accurate forecast. ‘The status of what the atmosphere is doing now concerns how it will evolve over the next few days.’

Some of the most useful atmospheric information for the weather forecaster is contained within clouds. Traditionally, to probe clouds, weather satellites have used infrared remote sensing techniques where an instrument measures the natural radiation signature from the atmosphere. However, these techniques are limited in the amount of information they can provide, said Dickinson. Recently developments in microwavebased sensing techniques have, he said, enabled a big step forward in the ability to probe clouds.

However, even bigger advances are heralded by the muchanticipated move towards a new generation of infrared sounding satellites that will probe the atmosphere with hitherto unheard of accuracy.

‘The next generation of satellites will enable a significant step forward in our ability to monitor the climate,’ he said. ‘They will use an infrared approach but, instead of giving a coarse description of the vertical structure of the atmosphere, will give us a much more detailed one. And by capturing the vertical structure in more detail you’re in a much better position to predict how the atmosphere’s going to evolve over the next few hours and the next few days.’

One such system, the Infrared Atmospheric Sounding Interferometer (IASI), will be deployed onboard Europe’s Metop satellites, a series of three polar-orbiting systems, the first of which will be launched early next year. According to Dickinson, IASI will allow 100-fold improvement in vertical resolution and vastly improve our understanding of the atmosphere’s temperature and humidity profile.

Another future satellite-based technique that holds much promise is the application of Doppler radar, a type of weather radar that uses the Doppler effect to measure the velocity of particles suspended in the atmosphere.

A laser-based Doppler instrument is planned onboard the 2008 launch of ESA’s Aeolus, the first-ever satellite to directly observe wind profiles from space. This mission is expected to improve the quality of weather forecasts still further, and to advance understanding of atmospheric dynamics and climate processes.

Beyond the tools used to gather atmospheric information, another crucial area of development for Dickinson’s team is the continuing improvements in the mathematical methods used to make sense of all this data. Dickinson explained that his team currently uses computational fluid dynamics (CFD) software tools similar to those employed by engineers in other disciplines to simulate and model atmospheric flow. The group now represents the world’s atmosphere on a horizontal and vertical grid made up of squares, each representing a 60km area.

Dickinson said that over the next decade his team hopes to replace this with a model that has a 1km resolution. In other words, each square kilometre of the UK will be represented by a square on the grid. One of the main things that this   will enable the MET office to do is explicitly model thunderstorms, one of the biggest causes of flash floods in the UK.

‘Thunderstorms don’t just stay in one place, they have a lifecycle of their own, and they may get triggered in their early lifecycle by being knocked into a hill somewhere and then evolve into something damaging or threatening. We will be able to give much more information on the corridor that thunderstorms will go along and the timing of them — which will be important in predicting flash floods.’

But it seems that Dickinson’s team, like so many other groups that work at the sharp end of mathematics, is at the mercy of Moore’s law. Astonishingly, the MET office’s NEC supercomputer, which has a theoretical peak power of 16 billion calculations per second per processor, simply isn’t powerful enough to run a 1km-resolution grid.

Thus Dickinson and his team find themselves waiting for computing technology to catch up with their ambitions. As we reflect on some of the mildest autumn temperatures on record, and shudder at the prospect of a harsh winter of discontent, it’s clear that today’s weather forecasters need all the help they can get.

But what is Dickinson’s winter prediction? Will it be a white Christmas or will we be eating turkey on the lawn? In a tone of voice that is as far removed from the hysterical tabloid reporter as it is from the cheery mateyness of the TV weatherman, Dickinson adopts the measured, logical cadence of the mathematician.

‘The outlook is for a colder winter than we’ve experienced over the past five to 10 years — but we are forecasting it with only a 65 per cent probability.’