Effects Of Autumn Soil Moisture And Snow Cover Anomalies On The Variation Of Winter AO And NAO

Besides sea surface,land surface is another important surface boundary in the climate system,and has a significant impact on atmospheric circulation and climate variation.Many studies have shown that early-season land surface boundary anomalies,especially land surface hydrology(soil moisture and snow cover)anomalies are significantly correlated with wintertime weather and climate variation.So far,most studies have examined the effects of soil moisture on local temperature and precipitation,and few studies have considered the relation between early-season soil moisture anomaly and wintertime large-scale atmospheric circulation variability.Meanwhile,many studies have suggested that early-season snow cover is significantly correlated with Arctic Oscillation in the following winter,but we want to find whether the key area of interaction and the best month for predictability have changed gradually as autumn snow cover has increased significantly.Finally,recent studies haven't put effort into incorporating soil moisture and snow cover in a system for winter large-scale atmospheric circulation variation prediction.Arctic Oscillation(AO)and North Atlantic Oscillation(NAO)are the leading modes of winter atmospheric circulation in the Northern Hemisphere,which have important implications for the Northern Hemisphere climate.Therefore,the main pruposes of this study are to:(1)examine the effect of autumn soil moisture on winter AO and NAO,(2)figure out whether the increasing trend of autumn snow cover has significant influence on wintertime AO and NAO predictability,(3)incorporate soil moisture and snow cover in a system for winter AO prediction.The datasets include soil moisture data from the ERA-Interim reanalysis,observed snow cover data from the National Snow&Ice Data Center and sea level pressure data from the NCEP-NCAR reanalysis.Research methods are as follows:(1)Perform singular value decomposition(SVD)analyses between anomaly fields of these two predictors and sea level pressure in the Northern Hemisphere and the North Atlantic,and find the key areas of interaction between these two predictors and wintertime large-scale atmospheric circulation in the Northern Hemisphere(AO/NAO).(2)Create cross-validated hindcasts of winter AO and NAO using simple linear regression models with the above two predictors,and calculate the anomaly correlation coefficient(ACC).The main conclusions of this paper are as follows:First,it is found that autumn siberian soil moisture appears as an effective source of DJF AO and JFM NAO predictability.The ACC between observed DJF AO index and predicted indices of AO based on OCT and NOV Siberian soil moisture are 0.42 and 0.54.The ACC between observed JFM AO index and predicted indices of AO based on OCT,NOV and DEC siberian soil moisture are 0.35,0.46 and 0.45.Second,we group three periods(1972/73-2001/02,1977/78-2006/07,1982/83-2011/12)to study the possible role of the increasing autumn snow cover to the prediction of winter AO and NAO.The MCA results show that a snow cover anomaly with maximum amplitude along 55°N-65°N in October and a snow cover anomaly with maximum amplitude along 45°N-55°N in November are associated with a wintertime positive AO(NAO)pattern during these three periods.But the significance between OCT snow cover and DJF AO(NAO)drops from 100%(96%)to 92%(84%)with ACC changing from 0.43(0.38)to 0.19(0.20),and the significance between NOV snow cover and DJF AO(NAO)increases from 92%(70%)to 100%(100%)with ACC changing from 0.35(0.19)to 0.58(0.50).It is found that the key area of interaction and the best month for predictability have changed gradually.The key area of interaction has changed from 60°N(Siberia and the northern part of the East European plain)to 50°N(Mongolian plateau,Baikal Lake and the southern part of the East European plain)and the best month of predictability has changed from October to November.Third,thanks to the considerable independence between soil moisture and snow cover anomalies,we find that including these two leading predictor indices(NOV soil moisture and NOV snow cover)in a standard linear regression model leads to a better winter(DJF)AO forecast than using any single predictor with ACC equaling to 0.67.On the one hand,our results show significant seasonal predictability skill of Eurasian soil moisture to large-scale atmospheric circulation in the Northern Hemisphere.On the other hand,we examine the effect of Eurasian snow cover on AO(NAO)and reveal that the key area of interaction and the best month for predictability have changed gradually.This study is a good supplement for understanding the climate effects of Eurasian snow cover,and presents a new view for winter AO and NAO predictability.