Oral-presentation at
European Geophysical Society 2000, Nice 2000
Regional analysis and representation of sea level variability from altimetry and climate data in the tropical Pacific
1Institute of Physical Geodesy;
Darmstadt University of Technology
The El Nino / Southern Oscillation (ENSO) phenomena has become a subject of research in the last years.
Topex/Poseidon altimetry data are used in the interval between September 1992 and August 1999.
One of the objectives is applying CCA to the prediction of short-term climate fluctuations, to explore both the features of the optimzed linear relationships and the level of the accompanying skill.
Introduction
The Southern Oscillation is not a stationary phenomena. This means that the background state is not constant.
Two major SO components have been documented in various studies: quasi-biennial and low-frequencies components (Trenberth 1976, Barnett, 1991). The importance of the annual cycle is also crucial, as the SO persistence depends on the annual cycle (Trenberth and Shea 1987) and therefore the interaction between annual cycle and SO is important for understanding and predicting ENSO. On the other hand it has been noticed that the strengh of the annual cycle is regulated by ENSO.
The temporal structure of ENSO has been less well documented than its spatial structure due to the limited time-length of observations. To study the El Nino phenomena long records of observables are necessary. Choosing the sea level height from altimetry as observable an interval of time of maximum 15 years is available for investigation.
The El Nino 1997 has been studied regionally (Wang, 1999 and Groten et al., 2000) analysing only two years of Topex/Poseidon (T/P) altimetry data and correlating the sea level height with sea surface temperature, atmospheric pressure and wind speed. Statistical methods have been used to isolate the principal modes of variability. During this short interval of time dominated by the El Nino event the dominant components of variability corresponds clearly to this events for all the variables.
Here Topex/Poseidon altimetry data from 1992 to 1999 are used to investigate the sea level height variability in the Equatorial Pacific and its temporal and spatial structures.
Method of analysis
Principal component Analysis (PCA) and Canonical Correlation Analysis (CCA) methods are used. A first attempt with wavelet analysis has been done.
Predictor and predictand data are detrended and then standardized so that all variables have temporal stationarity and are weighted equally to produce a correlation-based CCA (Barnston and Ropelewski, 1992).
Bibliography
More Abstracts