The next few years could be "anomalously warm", according to a new study.

Researchers have developed a mathematical model to predict how average global surface air temperatures will vary over the next few years.

The results suggest that the period from 2018 to 2022 could see an increased likelihood of extreme temperatures.

The findings are published in the journal Nature Communications.

The warming caused by emissions of greenhouse gases like CO2 is not increasing at a perfectly steady rate.

In the early years of the 21st Century, scientists pointed to a hiatus in warming. But several analyses show that the five warmest years on record all have taken place since 2010.

These variations from year-to-year do not affect the long-term trend in warming temperatures.

Now, a new method for trying to predict global temperatures suggests the next few years will be hotter than expected.

Rather than using traditional climate simulation techniques, Florian Sévellec, from the CNRS in Brest, France, and Sybren S Drijfhout, from the University of Southampton, developed a statistical method to search through simulations of climatic conditions in the 20th and 21st Century and look for situations that are comparable to the present day.

Future possibilities
The team then used these climatic "analogues" to deduce future possibilities.

In particular, the anomalous warmth predicted over the next few years is due to a low probability of intense cold climatic events.

Once the algorithm is "learned" (a process which takes a few minutes), predictions are obtained in a few hundredths of a second on a laptop. In comparison, supercomputers require a week using traditional simulation methods.

Gabi Hegerl, professor of climate system science at the University of Edinburgh, who was not involved with the study, said: "The authors have tried to predict whether global climate variability will make the next years warmer or cooler overall than the mean warming trend. They have skilfully used worldwide climate model data for previous years to calculate probabilities for the next few years.

"The findings suggest it's more likely we'll get warmer years than expected in the next few years.

"But their method is purely statistical, so it's important to see what climate models predict based on everything we know about the atmosphere and the oceans. Those are more expensive to run but also use more climate physics and observational information.

She added: "These new predictions are not geared up at the moment to predict regional trends such as the hot summer this year; so they may predict how likely it is to have a global record warm year, but not a regional record summer like we've had in the UK."

For the moment, the method only yields an overall average, but scientists now would like to adapt it to make regional predictions and, in addition to temperatures, estimate rain and drought trends.