Modelling the fate of dioxins in a trophic network by coupling an ecotoxicological and an Ecopath model

This paper shows the possibility to combine a trophic network model and an ecotoxicological food web model to calculate toxic concentrations of area-specific aquatic species. The trophic network model was derived from an Ecopath application made for a shallow water area (Palude della Rosa), situated in the northern part of the Lagoon of Venice. The organisms-specific output data from the energy model, such as metabolic rates, diet composition and structural characteristics were used as input to the ecotoxicological food chain model. The toxic substance model estimated bioaccumulation (pg/g lipid) of dioxins (e.g. 2378-TCDD, 12378-PeCDD, 123478-HxCDD, 123789-HxCDD, 1234678-HpCDD and 12346789-OCDD) and dibenzofurans (e.g. 2378-TCDF, 12378-PeCDF, 123478-HxCDF, 123678-HxCDF, 123789-HxCDF, 234678-HxCDF, 1234678-HpCDF, 1234789-HpCDF and 12346789-OCDF) for all groups in the trophic network. A sensitivity analysis showed that the toxic concentration in the aquatic species was highly affected by the following parameters: K ow, fraction of organic carbon in the sediment, lipid fraction, and fraction of unassimilated food. The coupling of the two models was validated with data from four sites in the Lagoon (two in the industrial zone, and two in the southern basin). Calibration and comparisons with measured data were conducted for a clam ( Tapes philippinarum), a mussel ( Mytilus edulis) and a fish ( Scardinius eritrophtalamus). Model output showed that the discrepancy between chemical analysis and model estimations were in the range of one order of magnitude for T. philippinarum and M. edulis (except for tetra- and penta-furans, where the overestimation reached two order of magnitude). It was also shown that more monitoring efforts should be made on fish species, to have a set of data useful for the validation of the model.