The outer limits of weather prediction

Weather varies dramatically across the globe depending on many factors, including distance from the equator and poles, proximity to oceans and the time of day and year. Forecasting the development of weather systems that bring rain and cloud, sun and heat, is therefore a major challenge for meteorologists working in numerical weather prediction.

The prediction of weather has improved rapidly since the advent of computing. Over the past 50 years, meteorological forecasting has moved from being rooted in empirical knowledge and basic understanding of atmospheric physics to the advanced forecasting – using complex supercomputed simulations of the atmosphere – that we have today.

Like future climate prediction, weather forecasting is constantly developing and increasing in predictive accuracy. Such developments have many applications: protecting the lives of people who work at sea, aiding the agricultural sector and informing the insurance industry, to name but a few.

The long view

Dr Colm Clancy, a meteorologist at Met Éireann’s Research, Environment and Applications Division, specialises in numerical weather prediction.

"Forecasting skill has improved at a rate of approximately a day per decade, meaning a five-day forecast now is as accurate as a four-day forecast was 10 years ago," he says.

Weather forecasts have improved through advances in two key areas. Firstly, continuing research contributes to the mathematical and physical understanding of how weather systems work. Secondly, the realisation of large supercomputing power has allowed for complex simulations of the atmosphere.

“A number of factors have contributed to these improvements,” he says. “These include increased knowledge and understanding of atmospheric phenomena, the availability of better and more weather observations, improved mathematical methods in the forecast models and a consistent increase in computational power.”

Through model development and increasing weather observation, current research is set to improve forecasts even further.

“In terms of model development, we want to run the forecasts at higher and higher spatial resolutions” says Clancy. “This requires further research on the physical processes occurring at small scales, and also brings challenges related to computational resources.”

Such computational resources include parallel supercomputing, whereby multiple processors simultaneously work on different parts of a problem.

One key issue in predicting the weather relates to our ability to forecast beyond the near future – beyond 14 days – because “weather system chaos” kicks in. Chaos is the process through which forecasts become unreliable with time, limiting our ability to forecast long-term weather.

“Beyond the two-week point we are reaching the limits of predictability, which is largely caused by our imprecise knowledge of the current weather conditions,” says Clancy.

What this means is that it is impossible to know the exact current weather conditions (necessary to run model forecasts) at every point across the world. Our global weather observation network – which measures things such as temperature and pressure – suffers from a lack of complete spatial coverage. We cannot accurately measure the temperature at every location on Earth, at each height in the atmosphere, at every given moment.

Things are improving, however. “Over recent decades there has been a huge increase in the amount of observational data available, particularly with the introduction of satellite measurements,” he says.

With these improvements, long-term forecasting, which includes seasonal forecasts (predicting summers, for example), has become an active area of research.

Reading the oceans

Clancy says “future seasonal forecasts might use information on atmospheric or oceanic phenomena, which are known to affect weather patterns”. An example of this would be the El Niño phenomenon, which was active earlier this year.

“But for precise and detailed weather forecasts, the predictability limit of two weeks remains, and anything beyond cannot yet be trusted.”

So with further research into the physics of atmospheric processes, and a better understanding of how our atmosphere is dependent on ocean behaviour, over the coming decades we can expect the accuracy of weather forecasts to increase up to the 10-day mark.

Such future developments in numerical weather prediction will help to improve and save lives. Nevertheless, as we move into a future affected by a quickly changing climate, the weather on this particular Atlantic island will ultimately remain variable, and wet.

Conor Purcell PhD is a science and nature writer who regularly contributes to The Irish Times. He can be found on Twitter @ConorPPurcell and some of his other articles at cppurcell.tumblr.com

WET WET WET: THE IRISH FORECASTING CHALLENGE

Predicting Ireland’s changeable weather is a formidable challenge. Rainfall can be intermittent or intense and lasting for entire days, and is dependent on the changing size and height of clouds moving across the island.

Our clouds and rain systems develop out over the Atlantic and are brought to us by westerly winds (moving eastwards) in the form of low-pressure systems.

These pressure systems are continental-scale air masses that move over the ocean and land, determining the temperature we feel, the hours of sunshine we enjoy and the rain we endure.

“We live in a very active weather zone, which we call the mid-latitudes,” says Met Éireann’s Colm Clancy. “Here the polar air coming from the north meets the tropical air from the south to help form the weather fronts which are such a regular feature of our weather charts. Additionally we live at the edge of a large landmass, with an extensive ocean stretching away to the west.

“These two features combine to produce weather in Ireland which is highly changeable.”

Compared with Spain, for example, which is dominated by more stable weather characterised by high-pressure systems, Irish weather is far less predictable.

Unlike low-pressure systems, which push air upwards to heights where water vapour condenses into water droplets, forming clouds, high-pressure systems push air downwards over the land with little opportunity for water drops and clouds to form. Because of this, areas experiencing high pressure are typically dry, whereas this island is wet.