Variability and predictability of North Atlantic cold-season atmospheric river occurrence frequency in a set of high-resolution atmospheric simulations

Atmospheric rivers (ARs) are conduits of enhanced water vapor transport in the lower troposphere, accounting for around 90% of poleward water vapor transport outside of the tropics and responsible for extreme precipitation and flooding events in many regions. In this study, we explore the variability and predictability of the North Atlantic (NA) cold-season (DJFM) AR occurrence frequency during 1951-2011 and their connections to climate variations using an ensemble of 30 simulations with a 60-km-resolution atmospheric general circulation model forced by observed sea surface temperatures (SSTs). EOF analysis is applied to the ensemble mean to determine the interannual-decadal variability of the AR occurrence. The first leading mode features a north-south dipole pattern and is controlled by combinations of ENSO and NAO conditions. Specifically, concurrent El Niño and negative NAO act in a consistent way to shift ARs equatorward, whereas concurrent La Niña and positive NAO tend to displace ARs poleward. The second leading mode is dominated by spatially homogeneous variations in the AR occurrence over northern Europe, with the time series showing a significant upward trend. The upward trend can be linked to the SST warming in the Indo-western Pacific since the 1970s. The third leading mode is characterized by the concentration and dispersion of AR occurrence in the southeast-northwest direction, and appears to be associated with the Scandinavian pattern and the SST anomalies in the easternmost equatorial Pacific. Further analysis is performed to examine the internal variability in the AR occurrence. We find that AR occurrence exhibits relatively weak internal variability over the central and eastern U.S. but extremely strong internal variability over western and northern Europe. The former can be attributed to the strong SST forcing associated with ENSO, whereas the latter originates from the chaotic variations of the NAO.