Modulators of Arctic summer climate modes : from storm to global teleconnection

Abstract

In this study, we investigate several factors to modulate the summer Arctic circulation which has a significant linkage with the sea-ice variation. Using self-organizing map (SOM) clustering, the daily Arctic (north of 60 °N) mean sea level pressure (MSLP) during 1979-2017 is classified to three SOM modes through the procedure finding the optimal number. The resultant patterns are characterized by the negative Arctic mode, the dipole mode between Pacific (positive) and Atlantic (negative) sides, and the positive Arctic mode with weak negative Russia for SOM1, SOM2, and SOM3, respectively. The occurrences of the frequency of each SOM pattern computed by counting the number of days that are matched to each SOM in a year show not only the large interannual variation but also the decadal shift. Since Arctic summer is well known as the most synoptic period with the frequent storm occurrence in Arctic central Ocean, we first check the influence of storms on these variabilities in SOM occurrences. All spatial and temporal changes in Arctic MSLP are largely explained by the storm distributions in the interannual time scale. Considering the short scale of e-folidng time of all three SOM mode, we also analyze the daily evolution of atmospheric condition in the synoptic scale, and obtain the consistent results with the unfiltered daily SOM pattern and the associated storm distribution. From these, we can conclude that the Arctic summer MSLP variability is attributed to the accumulated effect of the spatio-temporal distribution of synoptic storms. Further, the possible factors to modulate summer Arctic MSLP via affecting the storm distribution in the longer time scale are investigated. Basically, it is well known that the activity of storm is strongly related to the baroclinic instability. Indeed, the storm activity related to each SOM pattern are all significantly linked with the baroclinic instability which is determined from the surface temperature distribution. Since the Arctic surface condition can be modulated by the snow cover or sea-ice condition from the preceding season, we also investigated the snow cover, sea-ice, and the SST teleconnection from the preceding winter to the following winter.