Aluminum bioavailability and toxicity to aquatic organisms: Introduction to the special section

Citations Over TimeTop 20% of 2017 papers

Abstract

The ability to accurately predict the aquatic toxicity of aluminum (Al) in natural surface waters has eluded scientists for the past several decades. In the 1980s, the acid rain program, supported by the US Environmental Protection Agency (USEPA), identified Al in low-pH natural waters as a potential concern because of the presence of Al3+. Most of the studies conducted at that time were performed at a pH of 6 or lower and in waters with low dissolved organic carbon (DOC) and hardness concentrations; there was little direct consideration of variables affecting Al bioavailability. However, chemical species of Al differ substantially across the pH range of 6 to 8 found in most natural surface waters. Aluminum can be present as inorganic hydroxy species (Al3+, AlOH2+, Al[OH]2+, Al[OH]30, and Al[OH]4–), as inorganic complexes with fluoride (F–) and sulfate (SO42–), and as weak and strong complexes with organic material 1. In 1988, the USEPA released nationally recommended ambient water quality criteria for Al of 750 and 87 μg/L as acute and chronic criteria, respectively 2. However, these were derived to apply only to waters with pH between 6.5 and 9, and so they were based on a relatively small toxicity database. In 2009, therefore, we assembled a team of biologists, ecotoxicologists, biochemists, chemical engineers, and chemists to help expand this database and identify a means for measuring and predicting the toxicity of Al to aquatic organisms as a function of water chemistry. To address data gaps for pH ranges more representative of typical natural waters (pH 6–8), a series of chronic toxicity tests was performed, initially at pH 6.0 to 6.3 and with some studies at higher pH values, with 9 freshwater species: 2 fish species (the fathead minnow, Pimephales promelas, and zebrafish, Danio rerio), an aquatic oligochaete (Aeolosoma sp.), a rotifer (Brachionus calyciflorus), the great pond snail (Lymnaea stagnalis), an amphipod (Hyalella azteca), a midge (Chironomus riparius), a unionid mussel (Lampsilis siliquoidea), and a plant (duckweed, Lemna minor) 3-5. The species were selected to meet the USEPA requirements for developing ambient water quality criteria or the European guidelines for developing predicted-no-effect concentrations under the Registration, Evaluation, Authorisation, and Restriction of Chemicals (REACH) regulation 6. For the purpose of developing toxicity bioavailability models, multiple tests with a green alga (Pseudokirchneriella subcapitata), a cladoceran (Ceriodaphnia dubia), and a fish (fathead minnow) were performed across a range of DOC, hardness, and pH conditions. These latter data were used to develop a biotic ligand model (BLM) for the prediction of toxicity as a function of water chemistry 7. The linkage between Al bioavailability and interactions with biological membranes is supported by direct evidence of a concentration–response relationship between toxicity and Al accumulation in fish gills. This was assessed using data for Atlantic salmon (Salmo salar), where several thousand measurements of Al on the gills of salmon were evaluated under a variety of water quality conditions 7. For the Al BLM, it was important that the mechanistic framework be extended to consider toxicity resulting from a combination of dissolved and precipitated Al. Toxicity was shown to not correlate with dissolved metal, as is the case for many metals. Although this combination of dissolved plus precipitated Al can be quantified as total Al in toxicity tests, readers are cautioned not to equate bioavailable Al with total Al in natural waters, because for many natural waters, the total Al will be dominated by non-bioavailable Al minerals such as Al-silicates and clays. The Al BLM was developed with the consideration of Al solubility and combined toxic effects attributable to both dissolved and precipitated Al in solution. In cases where solubility is limiting, total Al is considered and a response–additivity calculation is performed to determine the predicted effect concentration that results from both dissolved and precipitated metal. The resulting model determines whether the bioavailable Al species at the biotic ligand (BL) is sufficient to exceed a critical toxicity accumulation level. The BLM was used to normalize the toxicity data set to a common set of water chemistries to develop site-specific species 10% effect concentrations and 50% lower confidence limits of 5th percentile hazard concentrations (μg total Al/L) for selected European Union and Lake Superior (North America) waters as a demonstration of the use of the model 3. The toxicity data sets presented in the present special section also were used to develop a multi-linear regression model (MLR) for the purpose of providing a simplified means to predict toxicity as a function of DOC, hardness, and pH. The MLR models, based on both 10% and 20% effect concentrations, were able to predict values that were within a factor of 2 of observed values in 100% of the cases for P. subcapitata, 91% of the cases of the cases for C. dubia, and 86% to 95% of the cases for P. promelas. The MLR model predictions were shown to compare favorably with measured toxicity responses and provide a way to rapidly calculate site-specific protection values for aquatic species 8. The research performed over the past 6 yr allows for the prediction and measurement of Al toxicity across a range of conditions typically found in natural surface waters in the United States and Europe. The data have been used in dossiers submitted for Al and compounds under the REACH program, and we suggest that the data and models would be useful for setting water quality criteria and standards for Al. William J. Adams Red Cap Consulting, Lake Point, Utah, USA Allison S. Cardwell Oregon State University, Corvallis, Oregon, USA David K. DeForest Windward Environmental, Seattle, Washington, USA Robert W. Gensemer GEI Consultants, Fort Collins, Colorado, USA Robert C. Santore Windward Environmental, Syracuse, New York, USA Ning Wang US Geological Survey, Columbia Environmental Research Center, Columbia, Missouri, USA Eirik Nordheim European Aluminium Association, Brussels, Belgium

Related Papers

• → Evaluation of river water quality monitoring stations by principal component analysis(2005)552 cited
• → Dissolved Organic Matter Kinetically Controls Mercury Bioavailability to Bacteria(2014)194 cited
• Deriving aquatic water quality criteria for inorganic mercury in China by species sensitivity distributions(2012)
• → Validation of Nickel Bioavailability Models for Algae, Invertebrates, and Fish in Chinese Surface Waters(2023)4 cited
• → Neural Network Modeling for Retrieval of Water Quality of Lake Taihu from Field Spectral Measurement(2008)2 cited