Periodicities in sea-level changes at Stockholm

The annual mean sea-level time series for Stockholm (Sweden) for 1825–1984 (160 data points) had a large long-term negative (almost linear, only slightly quadratic) trend. After correcting for the same, the detrended series was subjected to maximum entropy spectral analysis (MESA). Choosing selected periodicities for further multiple regression analysis, series for the first 80 yr (1825–1904) showed periodicities at T = 2.40, 5.0, 6.1, 13.5, 14.8 and 32 yr, significant at a 2 level. The three largest peaks (italicized) had amplitudes of approximately (2.5–3.0) ± 0.8 cm. The latter 80 yr (1905–1984) showed significant periodicities at T= 2.05, 2.7, 3.0, 3.6, 4.4, 5.5, 6.3, 7.7, 9.8, 20.5 and 33 yr. The three largest peaks (italicized) had amplitudes of approximately (2.0–2.5) ± 0.7 cm. The whole period of 160 yr (1825–1984) showed significant periodicities at T = 2.05, 2.9, 3.2, 4.5, 4.9, 5.6, 6.4, 7.8, 11.0, 13.7, 14.8, 29 and 43 yr. The three largest peaks (italicized) had amplitudes of approximately (1.6–1.9) ± 0.6 cm. All these significant peaks explained a variance of only about 30% or less, indicating a large random component of approximately 70%. Peaks at T = 11 yr (sunspot cycle) or T = 18.6 yr (lunar nodal term) were either absent or very weak. Most of the other peaks were transient (present in the first 80 yr or the latter 80 yr) except an uncertain quasibiennial oscillation (QBO) (T = 2–3 yr) and T = 5–6 yr and T 30 yr, which seemed to be persistent throughout the whole period. Some periodicities seem to resemble those seen in the Southern Oscillation (SO) index.