The Western Baltic sea ice season in terms of a mass-related severity index 1879–1992 (II ). Spectral characteristics and associations with the NAO, QBO, and solar cycle

Abstract

In this sequel paper, the investigation of the ice winter severity in theWestern Baltic is continued in the frequency domain. Spectral analysis of the time series of the mass-related ice index (VAΣ) reveals prominent quasicycles with periods of 2.3, 5.8 and 7.8 years. The same cycles stand out in the variance spectrum of the winter North Atlantic Oscillation (NAO) index. Variations in the quasibiennial Q32 and intermediate I105 periodicity ranges (indices give the range limits in years) account for 68.2% of the linear correlation (−0.46) between both time series. A maximum entropy subinterval analysis shows that the spectral variance composition of the VAΣ series undergoes significant temporal changes. The relative contribution of the Q32 range is down from 52.7% during 1899±20 to almost half of this value during 1972±20. The significance of the I105 and of short periodic oscillations in the S53 range has in turn increased from about 15% each to 36.8 and 30.7%, respectively. Predictions of ice winter severity based on the spectral characteristics of the full record are thus bound to fail. Some predictive skill for severe ice winters, which occurred with an overall frequency of 24.6%, may be realized from an apparent association with the 11-year solar activity cycle. This association has been most pronounced since the 1950s, about the time the 8-year rhythm started dominating variability. 9 out of 10 severe ice winters eventuated in phase with either high (3) or low solar activity (6). Moreover, almost all of the most severe ice winters occurred when the phase of the quasibiennial stratospheric wind oscillation (QBO) at the Equator was east (west) and, at the same time, solar activity was low (high). The peaks in the VAΣ spectrum after WWII reflect the dominance of singular 8-year ice events of modulated strength. This quasicyclic recurrence of strong ice winters in conjunction with the apparent statistical Solar Cycle–QBO–Ice Winter Severity relationship may be cautiously used in ice winter forecasting.