Statistical Estuary Data Analysis in Models and Measurements - Some Methods and their Limitations

摘要

This paper considers the impact of people on a complex natural system like an estuary and how this influence can be assessed. The methods of mathematical modelling and analysis of measured data are briefly introduced and their advantages and shortcomings highlighted. The paper looks into statistical methods for time series analysis, using the example of the water level - a parameter that represents a mixed times series signal consisting of harmonic and non-harmonic components. Some spectral methods and the regression method are then chosen from the wide selection of methods available and their limitations and prerequisites are discussed. A procedure combining both methods is presented and applied to the example of the Elbe gauge in Hamburg-St. Pauli. The focus is on the method and its limitations; the results obtained are of secondary importance and are not discussed here. Reference is made to a relatively recent method of analysing non-stationary and nonlinear phenomena (Hilbert Huang Transform). The application of this method would, however, imply abandoning the classical view of the tidal wave as consisting of partial tides. Since it is hardly possible any more for one single person or institution to be familiar with the vast selection of analytical methods or to apply them, the analysis of and research into complex systems will require cooperative approaches in the future.