Expect聽a lot more heat in coming聽years. The world will warm at an even faster rate from 2018 to 2022 than the underlying global warming trend, according to a forecasting method so simple it can run on a laptop.
We are fairly good at predicting the weather for the next few days. Long-term forecasts of the climate are also already proving reliable. Yet what would be really useful would be detailed forecasts for the聽months and years ahead.
The chaotic nature of weather means it will never be possible to predict events like heatwaves years in advance, but some aspects of the climate, such as average yearly temperatures, are聽predictable.
Advertisement
There are now 10 groups worldwide, including the UK Met Office, collaborating on based on sophisticated supercomputer models of the planet that incorporate the laws of physics.
Forecast maps
Their suggest that the northern hemisphere will keep warming especially fast, and that Australia and South Africa could get much less rain than normal.
A few groups are also making forecasts based on much simpler 鈥渟emi-empirical鈥� methods. These look at what happened in the past, what is happening now and then use brute-force statistics to predict what might happen next. No physics is involved.
This week, a semi-empirical method was unveiled by Florian S茅vellec at the University of Western Brittany, France, and Sybren Drijfhout at the Royal Netherlands Meteorological Institute. Its predictions of rapid warming from 2018 to 2022 are broadly in line with those made by the supercomputer models.
The new method may be a useful addition to our toolkit, but it doesn鈥檛 add much. The paper suggests the possibility of making predictions on mobile devices. There is no particular need for this, given we have the internet.
This study also predicts only the average global surface temperature, which is by far the easiest thing to do and the least useful. The paper claims that 鈥渁ccurate and reliable interannual predictions of global temperatures are key for determining the regional climate change impacts鈥�. That is not the case, says Adam Scaife, head of long-range prediction at the UK鈥檚 Met Office.
The long-term warming trend certainly matters, but the small changes in average global surface temperature from year to year have no discernible impact on regional weather. For instance, 2018 is not as hot as 2016 was in terms of the global average, yet this year has seen some ferocious heatwaves.
Long-term regional forecasts would be more useful. But semi鈥慹mpirical methods cannot provide them, says Scaife. And although the supercomputer models are already making regional forecasts, the Met Office web page hosting its forecasts : 鈥淒ecadal prediction is still experimental and the forecasts should not be relied on for making decisions, particularly on regional scales.鈥�
In other words, it is certainly going to be hot 鈥� but we will have to wait and see exactly when, where and how hot.
This article appeared in print under the headline 鈥淭he 2022 climate report: it鈥檒l be hot鈥�
Nature Communications
