Climate Change Research Spotlight

Using Multiple Models to Improve Seasonal Forecasting

The first study to examine the ability of a suite of general circulation models to predict sudden warmings in Earth’s stratosphere highlights the potential for improving Northern Hemisphere forecasts.

Source: Geophysical Research Letters


By

Disturbances in the stratosphere, the atmospheric layer that stretches from 10 to 50 kilometers above Earth’s surface, have the potential to dramatically affect weather on the ground. Chief among these are sudden stratospheric warmings, episodes of wind reversals during which temperatures in the polar stratosphere can increase more than 50°C within a few days. Because these sudden warmings are often linked to severe cold outbreaks across densely populated regions of Eurasia, more accurately predicting such events could help mitigate their impacts on people and infrastructure.

To better understand the ability of general circulation models to accurately forecast sudden stratospheric warmings, Karpechko et al. compared how well the nine subseasonal-to-seasonal models comprising the S2S ensemble predicted one of the most recent Northern Hemisphere events. This sudden warming, which began around 12 February 2018, was followed by an intense, multiweek cold snap across northern Eurasia.

Of the nine models the team evaluated, two correctly predicted the occurrence of the sudden stratospheric warming 13 days prior to its onset, and all but one accurately forecast the event 4 days in advance. The results indicate the primary factor that limited each model’s ability to predict the warming at least 10 days ahead was the proper location of a high-pressure system over Russia’s Ural Mountains. According to the authors, this “Ural high” appears to play an important role in the atmospheric propagation of planetary-scale waves, phenomena associated with the displacement of the stratospheric polar vortex and ensuing periods of unusually cold surface temperatures, although the exact role of the stratospheric event on the surface conditions remains unclear. The authors also emphasized the successful prediction of the multiweek cold period, beyond the timescale of weather forecasts.

As the first study to examine the ability of a suite of subseasonal-to-seasonal forecast models to predict sudden stratospheric warmings, this paper highlights the ensemble’s potential to collectively improve the forecasting skills of general circulation models. The encouraging results suggest that with future refinements, the S2S ensemble will be able to advance the science of subseasonal, monthly-scale forecasting, particularly in the Northern Hemisphere’s heavily populated midlatitudes. (Geophysical Research Letters, https://doi.org/10.1029/2018GL081091, 2018)

—Terri Cook, Freelance Writer

Citation: Cook, T. (2019), Using multiple models to improve seasonal forecasting, Eos, 100, https://doi.org/10.1029/2019EO119927. Published on 15 April 2019.
Text © 2019. The authors. CC BY-NC-ND 3.0
Except where otherwise noted, images are subject to copyright. Any reuse without express permission from the copyright owner is prohibited.