Predictive models of benthic invertebrate methylmercury in Ontario and Quebec lakes

摘要: Multivariate analyses on benthic invertebrate methylmercury concentrations ((MeHg)) and water chemistry from 12 Quebec bodies were used to guide the construction of simple, predictive models (MeHg) in 23 lakes Ontario Quebec. Separate for collector-shredder predatory invertebrates constructed using multiple regression assessed their utility by cross- validation. Predatory was negatively related pH positively dissolved organic carbon (DOC) concentration (cross-validation r 2 = 0.31). Collector-shredder only 0.13). Predictive our is similar or surpasses that observed previously published zooplankton MeHg tested against independently collected data. Significant environmental variables contribution overall explanatory power are those found models, suggesting both pelagic food webs, DOC important indicators bioavailability. Although seasonal patterns examined, none detected. These represent an effective means identifying interest researchers reconstructing past archived Resume : Des multidimensionnelles des de methylmercure chez les invertebres ben- thiques et la chimie l'eau milieux aquatiques du nous ont guides dans l'elaboration modeles simples predictifs benthiques lacs l'Ontario Que- bec. Nous avons construit distincts pour collecteurs-dechiqueteurs preda- teurs a l'aide regressions multiples evalue par validation croisee leur utilite prediction. Il y une relation negative entre predateurs le positive avec carbone organique dissous (DOC; 0,31). est seulement en correlation (r 0,13). Les potentiels prediction nos sont semblables ou superieurs ceux zooplancton tires litterature evalues don- nees recoltees facon independante. significatives milieu au pouvoir explicatif global trouves zooplanc- ton, ce qui laisse croire que d'importants indicateurs biodisponibilite MeHg, tant reseaux alimentaires pelagiques benthiques. Bien ayons recherchees, n'avons pas trouve ten- dances saisonnieres invertebres. Ces moyens efficaces identifier interessants chercheurs reconstituer passe partir donnees qualite accumulees archives.