INVESTIGATING THE RELATIONSHIP BETWEEN EURASIAN SNOW AND THE ARCTIC OSCILLATION WITH SATELLITE DATA AND MODELS

摘要

Recent research suggests Eurasian snow-covered arearn(SCA) influences the Arctic Oscillation (AO) via thernpolar vortex. This could be important for NorthernrnHemisphere winter season forecasting. A fairly strongrnnegative correlation between October SCA and the AO,rnbased on both monthly and daily observational data, hasrnbeen noted in the literature. While reproducing thesernprevious links when using the same data, we find nornfurther evidence of the link when using an independentrnsatellite data source, or when using a climate model.rnThe Arctic Oscillation (AO) is a hemispheric-scalernmode of climate variability that affects the NorthernrnHemisphere. The AO is linked to the stratospheric polarrnvortex, and it is thought that changes in the strength ofrnthis vortex can be felt at the surface. The ArcticrnOscillation affects the pressure pattern throughout thernentire Northern Hemisphere of the Earth, and thusrnaffects weather patterns across Northern Hemisphericrnlandmasses, such as Canada, the United States andrnEurope. For example, the positive AO phase brings mildrnand wet conditions to Northerly locations such asrnScotland and Scandinavia. Conversely, the negative AOrnphase brings mild and wet conditions to locationsin thernsubtropics, such Middle East and North Africa.rnGong et al [1] present a synopsis of published studies onrnthe relationship between snow cover and the AO. Thesernpapers focus on Eurasian October snow as a predictor ofrnthe following winter AO signature. The phenomenonrnwas investigated using observational data and anrnatmospheric general circulation model (GCM, in thisrncase ECHAM3). Focusing on Siberian snow cover,rnthese studies suggested that an anomalously high snowcoveredrnarea resulted in the weakening of the polarrnstratospheric vortex. Cohen et al [2] further proposed arnmechanism for the coupling.rnThe mechanism [2] is as follows:rn1. The area of Siberia that is covered in snowrndramatically increases throughout October.rn2. This sudden increase in snow cover results in thernmass cooling of air aloft. This results in anrnintensification of the Siberian High. This in turn resultsrnin a further decrease in temperatures.rn3. Mass cooling in the mountainous regions of Asiarnresults in an increase in the upward motion of Rossbyrnwaves (as described by the Wave Activity Flux).rn4. The breaking of these waves once they reach thernstratosphere results in the weakening of the polarrnstratospheric vortex and associated warming of thernstratospheric polar cap. 5. The weakening of the polarrnstratospheric vortex is also associated with zonal meanrnwind and geopotential height anomalies in thernstratosphere. These anomalies descend down into therntroposphere, eventually reaching the surface.rn6. This downward propagation results in thernmanifestation of a strong negative AO phase at thernsurface.rnCohen et al [3] subsequently proposed that rather thanrnthe mean monthly amount, the rate of change of SCArnduring October (termed the Snow Advance Index) is arnbetter predictor of DJF AO index than the mean OctoberrnSCA. In this work, we investigate the predictability ofrnthe AO based on both monthly and daily data.rnPrevious studies referenced here have used thernRobinson et al / Rutgers snow lab snow cover productrnas the observational data source (further details inrnsection 2.1 below). In this work, we use an independentrnsatellite data source to see if we can detect the samernsignal, alongside a coupled atmosphere-ocean GCM.rnHardiman et al [4] investigated the ability of GCMs tornrepresent this mechanism, and found that all thosernstudied failed to capture the observed correlations. Wernuse a more up-to-date model, contributing to thernCoupled Model Intercomparison Project 5.rnThis project was supported by Dan Guertin and CarolynrnAtkins at EDF Trading. Any improvement in seasonalrnforecasting for winter conditions in Europe wouldrnbenefit energy companies such as EDF Energy andrnwould give energy trading companies such as EDFrnTrading a competitive advantage.