Elements of uncertainty

24 Jan 2022 03:25 PM

Nothing is certain in life, goes the old saying, but death and taxes. To which should probably be added, “and the weather.”

Storm Arwen was identified days before it arrived on the UK’s shores

The United Kingdom is affected by changeable weather conditions because of its close proximity to the Atlantic. In addition, the country lies in the mid-latitudes, on the fringes of the Eurasian landmass with tropics to the south and the Arctic to the north. All these factors create dynamic tensions on our weather patterns.

The science of forecasting has improved greatly in recent years. Less-developed technology and meteorological science made it difficult to foresee the ferocity of the so-called Great Storm of 1987 (much of the damage was in fact caused by the then relatively unknown phenomena, a “Sting Jet”). However, by November 2021 Storm Arwen had been predicted, named and warnings disseminated days before the storm arrived. Thirty years ago, it was hard to provide an accurate forecast more than 24 hours in advance – now we can be accurate for up to a week.

Part of this improvement is not only driven by improvements to technology but also the collection of data. Over 200 billion observations, from all points of the globe, are now received daily at the Met Office from satellites, radar, weather stations, ocean buoys, weather balloons and ship observations. The constant collection of data ensures the most up-to-date data is used by our weather models.

Having collated the data, we then do something called perturbing the data –making small adjustments to provide a variety of possible outcomes. Once we have 30 different outcomes, we compare these against one another to check for cluster patterns. For example, if 15 of the 30 data sets predict rain in Glasgow, we can state with reasonable confidence that there will be a 50% chance of rain.

Although forecasting the weather is now more accurate than ever before an element of uncertainty will always remain. However, we live in a world where uncertainty is unpopular. When you consider how the likes of aviation, travel, agriculture, leisure, the military, the insurance industry and so many other sectors are affected by the weather, eliminating uncertainty is not just desirable – it can be essential. The question is – how can we reduce the element of uncertainty?

A key part of meteorology is analysing how weather systems impact on each other. A good example of this uncertainty has been the recent unusual weather across the Northern Hemisphere. While Colorado suffered major bushfires due to drought and high winds, Washington DC and Japan experienced deep snow, and much of Europe was unusually warm with the UK recording its highest ever New Year’s Day temperatures.

Paul Davies, Met Office Principal Fellow, explains how these seemingly disparate events are connected:

“These events arise because we are locked into a ‘circumglobally teleconnection pattern’,” says Davies. “This describes how an area of active tropical thunderstorms in the western Pacific interacts with the Asian Pacific jet, bringing colder weather into the heart of the US, and in turn increasing the strength of the Atlantic Jet Steam, driving associated weather fronts into the UK.”

The more we understand the relationship between global drivers – including El Niño, La Niña, the Madden-Julian Oscillation (MJO) and phases of the sun – the more certain we can be about the future.  

“A good analogy would be a football season,” explains Paul Davies. “By looking at the results over the past few seasons, we can make fairly accurate predictions about the outcome of the title. But on a more granular level, there is still uncertainty. Even the Premier League holders can occasionally lose a match.”

What we can do however is reduce the level of uncertainty by continuing to update the systems we use to collect data and study how systems interact. As technology continues to improve, and Artificial Intelligence makes it easier to run multiple possibilities for every situation, forecasts will continue to improve. The future for weather forecasting is bright. It’s a shame we can’t always say the same about the weather!