Process identification by principal component analysis of river water-quality data

Time series of nutrient concentrations and related water quality parameters taken at several locations along the River Elbe were subjected to multivar iate statistical analysis. The main question underlying this study is conce rned with whether known interactions between water quality variables can be recovered as statistically significant covariance patterns. For this purpo se, the standard technique of principal component analysis (PCA) was applie d. Raw data and deviations from an estimated seasonal cycle were analysed. In both cases, two leading patterns of covariance was obtained, one dischar ge-dependent and the other related to biological activities. Linear regress ion modelling based on discharge and temperature was used to approximately eliminate the impact of meteorological forcing; this led to a large reducti on of the seasonal component. The remaining partial variance of water-quali ty variables could be shown to be dominated by biological activities for wh ich temperature is of secondary importance. Amplitudes of the pattern relat ed to biological processes are much less correlated between different stati ons than those of the pattern induced by spatially homogenous discharge. Th e analysed covariance patterns agree well with general knowledge about basi c dynamical processes in the river. Therefore, multivariate statistical ana lysis offers an objective method to estimate the observed strengths of the given processes that involve simultaneous changes of several water-quality parameters. Such an assessment is a prerequisite when observations are to b e compared with corresponding results from process-oriented numerical model s in order to increase the knowledge about the nutrient system. A related a pplication would be to use it to identify the number of degrees of freedom needed to appropriately describe the nutrient system's variability. (C) 200 1 Elsevier Science B.V. All rights reserved.