The determination of the real toxicity of sediments in aquatic ecosystems is challenging and necessary for an appropriate risk assessment. Different approaches have been developed and applied over the last several decades. Currently, the joint implementation of chemical, ecological and toxicological tools is recommended for an appropriate and successful toxicity risk assessment. We chose the combination of the toxic unit approach with acute pore water tests (Vibrio fischeri, Pseudokirchneriella subcapitata and Daphnia magna) and whole-sediment exposure tests (V. fischeri, Chironomus riparius), together with invertebrate community composition (multivariate analyses) to detect short and long-term responses of the organisms in four rivers of the Iberian Peninsula. High toxicity was detected in three sites (the downstream sites of the Llobregat and the Júcar, and the most upstream site of the Ebro). We identified organophosphate insecticides and metals as the main variables responsible for this toxicity, particularly in the whole-sediment tests. In particular, chlorpyrifos was mostly responsible for the toxicity (TUs) of D. magna, coinciding with the C. riparius mortality (long-term toxicity) in the mentioned sites, and copper was the main pollutant responsible for the short-term toxicity of P. subcapitata. The combination of the different approaches allowed us to detect ecotoxicological effects in organisms and identify the main contributors to the toxicity in these multi-stressed rivers.
1 January 2014 in Science of The Total Environment
Ecotoxicological risk assessment of chemical pollution in four Iberian river basins (Llobregat, Ebro, Júcar and Guadalquivir) was performed. The data set included more than 200 emerging and priority compounds measured at 77 sampling sites along four river basins studied. The toxic units (TU) approach was used to assess the risk of individual compounds and the concentration addition model (CA) to assess the site specific risk. Link between chemical pollution and aquatic macroinvertebrate communities in situ was examined by using four biological indexes; SPEAR ("Species at Risk Index") as the indicator of decline of sensitive species in relation to general organic (SPEARorganic) and pesticides (SPEARpesticides) pollution; and Shannon and Margalef biodiversity indexes. The results of the study suggested that organic chemicals posed the risk of acute effects at 42% of the sampling sites and the risk of chronic effects at all the sites. Metals posed the acute risk at 44% of the sites. The main drivers of the risk were mainly pesticides and metals. However, several emerging contaminants (e.g. the antidepressant drug sertraline and the disinfectant triclosan) were contributing to the chronic effects risk. When risk associated with metals and organic chemicals was compared, the latter dominated in 2010, mainly due to the presence of highly toxic pesticides, while metals did in 2011. Compounds that are not regulated on the European level were posing the risk of chronic effects at 23% of the sites. The decline of sensitive macroinvertebrate taxa expressed in terms of SPEAR index was correlated with the increase of toxic stress related to organic compounds Biodiversity indexes were negatively correlated with the metals and the urban land use type in the catchment.
Land use type, physical and chemical stressors, and organic microcontaminants were investigated for their effects on the biological communities (biofilms and invertebrates) in several Mediterranean rivers. The diversity of invertebrates, and the scores of the first principal component of a PCA performed with the diatom communities were the best descriptors of the distribution patterns of the biological communities against the river stressors. These two metrics decreased according to the progressive site impairment (associated to higher area of agricultural and urban-industrial, high water conductivity, higher dissolved organic carbon and dissolved inorganic nitrogen concentrations, and higher concentration of organic microcontaminants, particularly pharmaceutical and industrial compounds). The variance partition analyses (RDAs) attributed the major share (10%) of the biological communities' response to the environmental stressors (nutrients, altered discharge, dissolved organic matter), followed by the land use occupation (6%) and of the organic microcontaminants (2%). However, the variance shared by the three groups of descriptors was very high (41%), indicating that their simultaneous occurrence determined most of the variation in the biological communities.
Starting from measured river concentrations, emission factors of 158 organic compounds out of 199 analyzed belonging to different groups of priority and emerging contaminants [pesticides (25), pharmaceuticals and hormones (81), perfluoroalkyl substances (PFASs) (18), industrial compounds (12), drugs of abuse (8) and personal care products (14)] have been estimated by inverse modeling. The Llobregat river was taken as case study representative of Mediterranean rivers. Industrial compounds and pharmaceuticals are the dominant groups (range of 10(4) mg·1000 inhab(-1)·d(-1)). Personal care products, pesticides, PFASs and illegal drugs showed a load approximately one order of magnitude smaller. Considered on a single compound basis industrial compounds still dominate (range of ca. 10(3) mg·1000 inhab(-1)·d(-1)) over other classes. Generally, the results are within the range when compared to previously published estimations for other river basins. River attenuation expressed as the percentage fraction of microcontaminants eliminated was quantified. On average they were around 60-70% of the amount discharged for all classes, except for PFASs, that are poorly eliminated (ca. 20% on average). Uncertainties associated with the calculated emissions have been estimated by Monte-Carlo methods (15,000 runs) and typically show coefficients of variation of ca. 120%. Sensitivities associated with the various variables involved in the calculations (river discharge, river length, concentration, elimination constant, hydraulic travel time and river velocity) have been assessed as well. For the intervals chosen for the different variables, all show sensitivities exceeding unity (1.14 to 3.43), tending to amplify the variation of the emission. River velocity and basin length showed the highest sensitivity value. Even considering the limitations of the approach used, inverse modeling can provide a useful tool for management purposes facilitating the quantification of release rates of chemicals into the aquatic environment.