Application of ridge regression to quantify marginal effects of collinear soil properties on phytotoxicity of arsenic, cadmium, lead, and zinc
2009
Soil properties that the mitigate hazardous effects of environmental contaminants through soil chemical sequestration should be considered when evaluating ecological risk from terrestrial contamination. The objective of the present research was to identify predominant soil chemical and physical properties that modify the phytotoxicity of As, Cd, Pb, and Zn to the nonhyper-accumulating higher plant perennial ryegrass (Lolium perenne L.). Phytotoxicity parameters were estimated from a dose-response experiment using the aboveground dry matter growth endpoint and were correlated with an assortment of relevant soil property measurements, with the ultimate goal of developing statistical prediction models for soil-specific adjustments to ecological risk assessments. Significant correlations between soil properties and phytotoxicity estimates were observed for all four contaminants; however, intercorrelation was observed among soil properties, necessitating an alternative to the conventional multiple regression commonly used by ecotoxicologists. Ridge regression, a regression-based technique that suppresses the effects of multicollinearity and enables prediction, was used to assess the marginal contributions of all properties found to mitigate phytotoxicity. Ridge regression models are presented along with two common conventional regression methods and are collectively discussed within the context of the mitigating effects of soil properties on metal/metalloid phytotoxicity. Ridge regression appears to be a powerful alternative to conventional multiple regression for ecotoxicological studies when intercorrelation among predictors is experimentally unavoidable, such as with soil properties.
Afficher plus [+] Moins [-]Mots clés AGROVOC
Informations bibliographiques
Cette notice bibliographique a été fournie par National Agricultural Library
Découvrez la collection de ce fournisseur de données dans AGRIS