Presentation

Toward Enhanced Seasonal Forecasts Using Machine learning: Understanding the Polar Vortex Variability
DescriptionIn both hemispheres, the atmosphere above polar regions is characterized by a vortex of winds, whose intensity fluctuates seasonally and peaks during winter. In average once every two years, a sudden warming of the stratosphere disturbs the wind field causing displacements and deformations of the vortex. The impact of these events, known as Sudden Stratospheric Warmings (SSW), is not limited to the stratosphere but also strongly influence the weather at the earth surface up to three months after their occurrence. Such influence increases the predictability of weather and can be exploited to produce better long range forecasts. Understanding the structure of these perturbations is key to their prediction, and so to improve the quality of seasonal weather forecasts. However, studying the stratosphere requires to analyse large quantity of data with complex interactions. For this reason, the conditions yielding to SSW, the different states of the polar vortex, and their interactions are still not fully understood. In this talk, we conduct of thorough analysis of different stratospheric parameters, use supervised learning to find potential SSW precursors and quantify predictability of such perturbations.
SlidesPDF
TimeThursday, 8 July 202112:00 - 12:30 CEST
LocationJean Calvin
Event Type
Minisymposium
Domains
CS and Math
Emerging Applications
Climate and Weather
Solid Earth Dynamics