What’s happened to the weather?
Image credit: Science Photo Library
The weather has taken a coronavirus hit. The planes that provide the data are stuck on the ground. What is the forecast for forecasting?
While we were growing used to skies blissfully free of noisy aeroplanes during lockdown, it was a mixed blessing for meteorologists. Every day they were accustomed to receiving millions of weather observations from commercial flights – information on wind speed or air temperature some 10km above the Earth at a given point. As UK aviation braced for troubled times, fewer flights meant less information to feed into the supercomputers that calculate weather with ever greater precision.
As we all know, British weather is “noisy and unpredictable”, says former Met Office and BBC meteorologist John Hammond, co-founder of weathertrending.com. “We have huge variation one year to the next.”
Accurate weather reports can keep the country running, particularly in times of uncertainty where changing conditions are more apparent – and sometimes they can even mean the difference between life and death.
April’s weather was settled, sunny and predictable – which meant the lack of data from planes wasn’t as critical as in disturbed patterns. But quiet skies have left a little hole in meteorologists’ data and shed light on the delicate science of weather forecasting. To meet this shortfall, they’ve been launching more weather balloons than usual – but this is an expensive stopgap.
Getting it right is critical, particularly amid a seemingly increasing amount of extreme weather, says Andrew Charlton-Perez, professor of meteorology at the University of Reading. Predicting heatwaves, which kill old and vulnerable people, or droughts, which lead to bush fires, or severe rainfall, can save lives. “Last summer vulnerable people in France (which recorded a record 45.9°C in June) were provided with air conditioning or taken where it was cooler, and this is arguably as important as tackling issues such as Covid-19. We’re trying to find the sweet spot to take action where necessary. Information amid uncertainty helps you take those decisions.”
Hour-by-hour forecasting matters too. Will a disturbed jet stream prompt a bank holiday event to be a washout? Does the roof need to go on at Wimbledon (sadly irrelevant this year)? Should councils grit the roads? Will a supermarket sell out of barbeque meat on a sunny weekend? Will impending snow prompt a run on long-life goods in local shops? Will there be leaves on the tracks?
Unreliable forecasts have long been a standing joke – BBC weatherman Michael Fish was never allowed to forget his famous dismissal of reports of an impending hurricane before the 1987 storm ripped through the south of England, killed 18 people and caused an estimated £2bn of damage.
And in August 1979, the Fastnet Race – from Cowes to Fastnet Rock south of Ireland and back to Plymouth – went ahead before an unusually vicious storm wrought havoc on the famous yachting event with the death of 19 people. Meteorologists had underestimated the force of the winds and mountainous seas.
Short-term forecasts are now more accurate than ever, say meteorologists. “As a ballpark guide,” says Jim Trice, head of observations projects at the Met Office, “we’ve been able to predict an extra day ahead per decade with the same accuracy – but there’s always scope to improve.”
While it’s not a ‘eureka’ science, weather forecasting is one of the most complex, says Charlton-Perez. Computer models describe the intricate physics that dictates how air and water vapour flow around the planet. It’s an area crying out for fresh blood, from students to technologists and engineers looking to change course. “Users of weather forecasts don’t realise the technical sophistication of the systems that underpin them,” he says. If forecasts are to continue to improve, the sector needs a qualified, innovative workforce.
Today’s predictions are largely ensemble forecasts, derived from different models and estimates of the state of the atmosphere today. “So a typical ensemble might have three models that each make around 50 forecasts,” says Charlton-Perez. “Say if you have 150 forecasts, and 100 of them predict above-normal rainfall, there’s a 2/3 probability it will be above normal This is an attempt to model the ‘chaos’ effect in the atmosphere. Small errors in our estimates today can result in large differences in the forecast a week from today.”
And the Met Office’s supercomputer system – which occupies a substantial space at headquarters in Exeter – is one of the most powerful in the world dedicated to weather and climate. Capable of 14,000 trillion calculations per second (14 petaflops), it crunches 215 billion weather observations from satellites, radars and weather stations every day. Weather scientists are constantly tweaking equations and tinkering with the input of millions of observations. In February 2020, the UK government pledged up to£1.2bn for latest supercomputing > < technology for more accurate predictions of climate change.
As well as staff dotted around the world, from Scotland to Cyprus to Antarctica and more, there are some 300 unmanned Met Office weather stations placed around the UK, monitoring windspeed, temperature, humidity and pressure, sometimes visibility. Since the 1987 storm, about ten weather buoys have been installed to the west in the Atlantic and North Sea so the UK can have better sight of what is coming – weather ships are a thing of the past, says Trice. And then there are 15 huge ‘golf ball’ radar stations found on local vantage points around the country. These emit electromagnetic pulses and the echoes are stitched together by meteorologists to build a picture of rainfall across the UK.
Hay fever sufferers are served by some more analogue technology – the Met Office collaborates with universities, using data from a network of UK pollen traps, which feed into daily pollen-count forecasts.
Moreover, Met Office weather balloons are launched regularly around the UK, carrying instruments known as radiosondes high into the air with sensors to measure temperature and humidity.
When commercial airlines are operating as usual, the Met Office might receive some seven million separate observations of windspeed and direction supplied from sensors and directly from messages intercepted from planes to air traffic control radar. “Though obviously volumes have dropped drastically since lockdown,” says Trice.
Around Europe, there are also devices to measure lightning strikes and triangulate their position, while the engaging blitzortnung.org shows real-time lightning storms on a world map. Satellites have overhauled weather forecasting and lifted it to a new level of precision.
Microwave sounders on satellites operated by the intergovernmental European Organisation for the Exploitation of Meteorological Satellites (Eumetsat) assess how much water vapour is present in the atmosphere and Met Office dishes beside the Exeter base track observations.
Back on the ground, the Met Office works with a network of voluntary climate observers, from local government to research institutions to members of the public, who record rainfall. Rain gauges have grown more sophisticated, and the Met Office regularly monitors sensors and outstations for accuracy.
Technologists at the Met Office are also keeping an opportunistic eye on autonomous cars, which could in the future provide a rich source of local weather information. “It’s early days but we could use sensors on vehicles to tell us a lot,” says Trice. On a simple level, he explains, use of windscreen wipers could feed back information on rainfall, while external car cameras and lidar, which uses laser pulses to create a 3D model of the environment around the car, could indicate intensity of rainfall and visibility. “It’s something we are working on with industry.”
As jet streams rush along some 11km above the Earth’s surface, and the oceans store and discharge heat, and we pump millions of tonnes of carbon into the atmosphere every year, meteorologists take a long and short-term view.
Forecasts for the century are sobering – these incorporate climate science and surface in the predictions of the United Nations Intergovernmental Panel on Climate Change. More than three billion people will be living in “near-unliveable” temperatures by 2070, says a May 2020 report conducted by the University of Exeter and international scientists. Longer droughts, higher temperatures and heavier rainfall are all a result of climate change, and while there is some natural variability, manmade climate change leads to changes in frequency, intensity, duration and timing of extreme and unprecedented weather.
There are seasonal forecasts which are produced every month. These and longer-term forecasts use models similar to those used for daily forecasts, says Charlton-Perez, but will incorporate the impact of a broader range of factors, such as changes in ocean and ice.
Sub-seasonal forecasts – an active area of research – look between three to five weeks into the future. Medium-range looks one to two weeks ahead – our sunny spring at the start of lockdown was well forecast – and short-range forecasts try to predict hour by hour with ever greater reliability.
While pollution during lockdown has undoubtedly dropped, this is not thought to have affected the weather. A May study by the Met Office’s Hadley Centre, which researches climate change, says the global drop in emissions due to the pandemic will have very little impact upon the amount of CO2 in the atmosphere, which is still rising.
“But [the fall in emissions] is a drop in the ocean,” says meteorologist Hammond. “We’ve been pumping carbon into the atmosphere for centuries. The oceans store most of the heat, and they take forever to warm and cool. They act as huge storage heaters, so flicking the switch for five minutes in the global system will make very little difference.”
Unpredictable weather day-to-day, year-to-year, makes communicating longer-term climate change difficult. 30 June 2020 will probably be very different from 30 June 2019. “People live from month to month and it’s hard to have a longer-term perspective,” says Hammond – and he’s staking his reputation on a forthcoming run of cold winters, noting: “There’s a loose relationship between solar activity and the severity of our winters.”
Every Friday, Charlton-Perez and colleagues meet to discuss the weather, and it seems to him that extreme events are cropping up with ever greater frequency. Winter 2019-20 has seen three storms barrelling through the UK, and the wettest February on record, as well as the highest December temperature on record, 18.7°C, and a record-breaking heatwave last July.
And the planet is 1°C warmer than it was 200 years ago, before the industrial revolution. “That’s indisputable,” says Hammond. “Even if we have a winter dominated by a cold northerly wind, it will be not quite as cold as it used to be, and summer will feel just a little bit hotter, all other things being equal. The atmosphere is just that bit warmer.”
While the Ancient Greeks and other civilisations kept a close eye on the weather, modern forecasting didn’t begin until the invention of instruments to measure the atmospheric conditions. The first thermometer was created by Galileo in the 1590s, using water and a glass. Half a century later, Evangelista Torricelli observed that the height of a column of mercury in a glass tube rose and fell in line with fine and poor weather – and went on to invent the barometer. French physicist Blaise Pascal hiked up a mountain with a column of mercury and found air pressure was lower at the top, theorising that a vacuum sat above.
But until modern forecasting, farmers and others who worked outside had to rely on clues such as cloud formations and animal behaviour. Between 1855 and 1860, some 7,400 ships were wrecked off the UK coasts, and Admiral Robert FitzRoy believed lives could be saved. He went on to found what is now the Met Office and began issuing storm warnings in 1860. The following year he published the first world weather forecast on 1 August 1861, making use of the electric telegraph. London was forecast a mild 62°F, clear with a south westerly force 3 breeze.
The first mathematical prediction was produced in the early 1900s by the father of numerical forecasting, Lewis Fry Richardson. It took six weeks to complete by solving equations on paper. Even then, his first forecast was inaccurate as data was corrupted, but his pioneering method was proven to be correct and he’s revered by today’s meteorologists.
The first meteorological satellite
Sixty years ago, on 1 April 1960, Nasa launched a weather satellite equipped with television cameras and video recorders. The Television Infrared Observation Satellite (TIROS-1) provided forecasters with the first ever aerial view of developing clouds. Orbiting at 450 miles above Earth, the satellite, which operated for 78 days, sent back more than 19,000 fuzzy but usable images – including one revealing a typhoon 1,000 miles east of Australia – and proving the value of space-based observation.
The weather and D-Day
The lives of a million soldiers, and arguably the future of Europe, depended upon the accuracy of weather forecasts ahead of D-Day in 1944. Clear skies and calm seas were critical, along with low tide near sunrise. Meteorologists advised that the invasion of Normandy should fall between 5 and 7 June, or two weeks later. Originally set for 5 June, General Eisenhower delayed for a day after worsening conditions were forecast. But during 4-5 June, forecasters predicted a temporary break in the weather and the invasion went ahead on 6 June, amid strong winds and rough seas. Had the allies waited until 19 June, they may have foundered, as a severe storm hit the Channel on that day.
Sign up to the E&T News e-mail to get great stories like this delivered to your inbox every day.