Zurück zu den Preprints des Jahres 2000


Long-term behaviour

by Lehmann, A., Rode, M..

Series: 2000-08, Preprints

62M10 Time series, auto-correlation, regression, etc.
62G07 Density estimation
86A05 Hydrology, hydrography, oceanography

This study analyses weekly data samples from the Elbe
river at Magdeburg between 1984 and 1996 to investigate the changes in
metabolism and water quality in the Elbe river since the German
reunification in 1990. Modeling water quality variables by univariate
time series models such as autoregressive component models and ARIMA
models reveals the improvement of water quality due to the reduction
of waste water emissions since 1990. The models are used to determine
the long-term and seasonal behaviour of important water quality
parameters. A new procedure for testing the significance of a sample
correlation coefficient is discussed. The cross-correlation analysis
is applied to hydrophysical, biological, and chemical water quality
variables of the Elbe river since 1990. Special emphasis is laid on
the detection of spurious sample correlation coefficients.

Water quality,