Variable Seasonal and Subseasonal Oscillations in Sea Level Anomaly Data and Their Impact on Prediction Accuracy
Weekly sea level anomaly (SLA) maps are now available courtesy of the Archiving, Validation and Interpretation of Satellite Oceanographic (AVISO) data. Using the Fourier Transform Band Pass Filter (FTBPF) variable broadband seasonal and subseasonal oscillations were computed as a function of geographic location. Irregular amplitude and phase variations in these oscillations cause the increase of prediction errors of the SLA data for a few weeks in the future. The amplitude and phase variations of the broadband annual oscillation were computed by a combination of the FTBPF and the Hilbert transform. In order to detect the impact of irregular amplitude or/and phase variations of the annual oscillation on the SLA prediction errors, standard deviations maps of amplitude time differences as well as of the products of phase time differences and amplitudes were examined. The SLA data prediction errors in certain geographic regions of the ocean seem to be caused mainly by nonlinear behaviour of the broadband annual oscillation. The nonlinearities are probably driven by mesoscale eddies, and the significant impact on SLA prediction errors was observed in the vicinity of the western boundary currents.
KeywordsFourier transform band pass filter Prediction Satellite altimetry Sea level change
- Kosek W (1995) Time variable band pass filter spectra of real and complex-valued polar motion series. Artif Satellites 30:27–43Google Scholar
- Kosek W, Rzeszótko A, Popiński W (2006) Phase variations of oscillations in the Earth orientation parameters detected by the wavelet technique. Proc Journées 2005:121–124Google Scholar
- Kosek W (2001) Long-term and short period global sea level changes from TOPEX/Poseidon altimetry. Artif Satellites 36(3):71–84Google Scholar
- Popiński W (2008) Insight into the fourier transform band pass filtering technique. Artif Satellites 43:129–141Google Scholar
- Röske F (1997) Sea level forecasts using neural networks. Ocean Dyn 49:71–99Google Scholar