Characteristics,causes And Climate Impacts Of The Interannual Variation Of Autumn Tibetan Plateau Snow Cover

Snow cover is an important component of the Earth climate system.Snow cover variability has a siginificant influence on the surface energy budget and the hydrological cycle.The Tibet Plateau(TP)has a high altitude and is covered by snow over the whole year.The TP can impose a climate impact on the surrounding regions and even the globe through thermodynamic and dynamic effects.With the establishment of observation stations on the TP and the development of satellite observations,the climate effects of the snow cover variation over TP have been widely discussed.However,previous studies mainly focus on the winter-spring snow cover over TP,and thus it is still unclear about the characteristic of the autumn snow cover interannual variation over TP,the causes of autumn snow cover interannual variation over TP,and the climate influences of the interannual variation of autumn snow cover over TP.In this study,the characteristics,causes,and climate impacts of the interannual autumn Tibetan Plateau(TP)snow cover changes are investigated through observation analysis and numerical experiments.The results are summarized as the following.(1)In autumn,high mean snow cover extent(SCE)values appear over the western and eastern TP respectively.The first empirical orthogonal function mode(EOF1)of the SCE is dominated by variations over the central-eastern TP.A significant change of autumn SCE over TP is detected in the mid-1990 s.Thus,analysis is performed for two subperiods,namely,1979-1994(P1)and 1995-2018(P2).(2)In both P1 and P2,snowfall from late summer and autumn contribute to the variation in autumn SCE over the central-eastern TP.The contribution of snowfall is larger in P1 than P2.The atmospheric general circulation plays a different role in P1 and P2.In P1,the autumn SCE-related atmospheric circulation anomalies over East Asia are relatively weak and not significant.In contrast,during P2,associated with greater-than-normal autumn snow over the central-eastern TP,a pronounced wave train-like atmospheric pattern is observed to originate from the North Atlantic Ocean,cross the mid-latitude Eurasian continent and reach the TP.The negative atmospheric low system associated with this wave train dominates the upper level of the TP,leading to significant anomalous upward air motion,which favors snowfall anomalies over the TP and contributes to the formation of the central-eastern TP autumn snow anomalies.Further analysis shows that during P2,the SSTA over the North Atlantic Ocean can impose a perturbation on the overlying atmosphere,which mainly develops by obtaining energy from the mean flow through baroclinic energy conversion.(3)In both P1 and P2,the local effect of the anomalous positive SCE is a cooling on the overlying atmosphere.A comparison of local energy budgets associated with the SCE reveals that the local cooling effect of the anomalous positive SCE is more significant in P1 than P2.Results show that during P1,associated with the anomalous positive SCE,a negative geopotential height anomaly emerges at the upper level of the troposphere over the northeastern TP.This negative anomaly moves northeastward with time and imposes an anomalous vorticity perturbation near the East Asia Westerly Jet(EAWJ)core in the following seasons.This vorticity perturbation propagates downstream,forming a Rossby wave train-like pattern along the wave guide of the westerly jet.The vorticity perturbation crosses the North Pacific Ocean and reaches the North American(NA)region,leading to significant positive and negative SAT anomalies over western and eastern NA,respectively.In contrast,during P2,the anomalous positive SCE has a relatively weak local cooling effect on the overlying atmosphere and the atmospheric response tends to locate south of the TP.The vorticity forcing related to the SCE around the EAWJ is also much weaker in P2 than in P1.The following propagation of the perturbation is also weak and is mostly to the south along the west coast area of NA in the following winter,which has a weak impact on the surface air temperature(SAT)over NA during this period.(4)In this study,three machine learning(ML)models and autumn SCE over TP are used to forecast the winter surface air temperature over NA.The performance of three ML models is also compared with two Canadian operational dynamic models.Overall,three ML models show comparable skill to two dynamic models,and a five-model ensemble can significantly improve the skill of two dynamics models over the central and western NA.Besides,the additional use of the TP SCE predictor can indeed contribute to improving the seasonal forecast skill of the winter SAT over NA.The spatial distribution of the improved areas,however,is mode dependent.The results of this study reveal that the autumn snow cover over TP can impose a broad climate impact on earth.