|
(20-08) Integration of spatial analysis techniques in Ecological Risk Assessment: the case study of the edible organism Tapes philipinarum in the Venice lagoon. Critto, Andrea*,1, Micheletti, Christian1, Bertazzon, Stefania2, Marcomini, Antonio1, Zanetto, Gabriele1, 1 University of Venice, Department of Environmental Sciences, Calle Larga S. Marta 2137, Venice, Italy2 University of Calgary, Department of Geography, 2500 University Dr. NW, Calgary, Canada ABSTRACT- The Ecological Risk Assessment (ERA) was applied to the benthic organism Tapes philippinarum, to estimate the risk correlated with the contaminated sediments of the Venice lagoon. In order to develop the assessment, the integration of spatial analysis of environmental data within the ERA procedure was required. The exposure of the clam Tapes ph. to organic (total PCBs, PCB dioxin like, PCDD/Fs) and inorganic (As, Cd, Cr, Cu, Hg, Ni, Pb, Zn) pollutants was characterised by estimating the bioaccumulation inferred from linear regression models. The spatial distribution and correlation of the environmental data used for the model definition (concentration of each pollutant in the surface sediment, total organic carbon, pH and sediment fine fraction; lipid fraction and weight of clam) was undertaken by using geostatistical techniques (variografy and kriging). Standard regression models generally do not take into account the irregularities and discontinuities of spatially distributed variables, that are often characterised by spatial autoregression. These issues were overcome by spatial regression analysis (Cressie, 1993). The basic step was the specification of a neighbourhood structure, or contiguity matrix, which contains information on spatial interactions among observations. Then, a vector of spatial weights was calculated to calibrate the influence of each data-point based on its distance and the underlying spatial autocorrelation model. Spatial regression and related statistical routines were calculated using S-Plus© spatial statistics software (MathSoft, Inc., 1996) and the pseudo R2 (Anselin, 1993) provided the model′s goodness of fit. The spatial regression analysis produced regression models adequate only for organic compounds (pR2 >0.5), based on the comparison with the results obtained by a standardized bioaccumulation model (Gobas, 1994). Finally, maps of bioaccumulation in the clam tissues were obtained by applying the kriging interpolation method, according to the autocorrelation model defined by variogram. Key words: Ecological Risk Assessment, Spatial Analysis, Geostatistic techniques, Bioaccomulation regression models |
