Heinz Zwanziger

Heavy metals in river sediments — A chemometrical evaluation

First results are presented to support decisions within projects of recultivation in the Merseburg region. From two places of the river Luppe eight sediment samples were taken at four different days. The total concentrations of Cd, Cr, Cu, Fe, Mn, Ni, Pb, and Zn were determined by ICP-AES after digestion with aqua regia both in the gross samples and in four particle size fractions obtained by wet sieving. Subsequently each fraction was split into six classes by a sequential extraction to characterise the pattern of heavy metal binding form. Questions about the difference of the regions, the importance of the particle size fractions, the mobility of the metals, and the influence of the seasonal factor were tried to answer by the help of chemometrical methods. It was possible to show that the two sampling places situated only about 7 km from each other have different total metal concentrations as well as different binding form patterns. The trend of the heavy metal enrichment in the smallest particle size fraction can easily be seen in two-dimensional diagrams, but could in general not be confirmed statistically for all metals. The consideration of all four factors together (region, particle size fraction, mobilisation class, and time) shows the significant influence caused by all factors and all interactions. Multivariate covariance analysis is used to include also the influence of sampling conditions.

Zur multivariaten Auswertung von Ringversuchen

SummaryThe advantages and drawbacks of some graphical methods as interpreting aids for cooperative tests are demonstrated using results of slag sample analyses (5 laboratories, 7 components). In cases of low dimensions (components) the method of regular polygons has been found most informative. For higher dimensions modern data analysis techniques, such as principal components analysis (pca), correspondence analysis and discriminant analysis are to be preferred. From pca of correlation matrices one can also derive best choices of feature subsets to construct good polygons. The sensitivity of partial correlation coefficients due to deviating laboratory manners can be assessed by leaving-one-out comparisons.

Characterization of North Sea drilling cuttings by metal binding form patterns and diffuse reflectance spectroscopy (DRIFT)

Samples of marine sediments collected in the vicinity of and remote from an oil production platform have been analysed for several metals (Cr, Cu, Fe, Mn, Ni, and Pb) and for total hydrocarbons. The same samples have been examined by diffuse reflectance IR spectroscopy (DRIFT) after drying, and an attempt has been made to classify the samples according to the patterns of the metals and the reflectance spectra, and to correlate this classification with that based on the sampling site. It is suggested that DRIFT is a faster technique than the conventional solvent-extraction IR method for total hydrocarbons and — while less sensitive — is adequate for screening such samples for significant levels of hydrocarbon pollution. Classification of the samples according to the metals present also distinguishes between the two groups of samples.

Risiko 2. Art und Stichprobenumfang bei statistischen Tests

SummaryComparisons of means or other parameters analytically obtained by statistical tests of Students type are only valuable, if the level of error of the first kind (α) and the error level of the second kind (β) is accounted for as well. Often failing decisions based on the latter are more serious. The risk of second kind is closely related to the difference (d) to be detected and to the replication number or random sample amount (n). Thus an appropriate n has to be maintained to match the statistical test, on the other hand it must not cause uneconomical efforts. Hence the values of n should be determined besides fixing those of α, β, and d prior to experiments.ZusammenfassungBeim Vergleich analytisch ermittelter Parameter (z. B. Mittelwerte) in statistischen Tests (t-Tests) ist neben der Angabe des Risikos 1. Art (Irrtumswahrscheinlichkeit α) auch die Berücksichtigung der Wahrscheinlichkeit β für Fehlentscheidungen 2. Art erforderlich, da diese häufig gravierender sind bzw. die Testaussage sonst u. U. wertlos ist. Das Risiko 2. Art ist eng verknüpft mit der nachzuweisenden Differenz d und dem Stichprobenumfang. Die Irrtumswahr-scheinlichkeit α, die Differenz d und das Risiko 2. Art müssen sinnvoll vorgegeben werden, und der zur Einhaltung dieser Bedingungen erforderliche Stichprobenumfang sollte aus statistischen und ökonomischen Gründen weder unternoch überschritten werden. Deshalb ist es ratsam, den optimalen Stichprobenumfang vor Versuchsbeginn zu ermitteln.