Development of a machine learning model to estimate biotic ligand model-based predicted no‐effect concentrations for copper in freshwater

The copper biotic ligand model (BLM) has been used for environmental risk assessment by taking into account the bioavailability of copper in freshwater. However, the BLM-based environmental risk of copper has been assessed only in Europe and North America, with monitoring datasets containing all of the BLM input variables. For other areas, it is necessary to apply surrogate tools with reduced data requirements to estimate the BLM-based predicted no-effect concentration (PNEC) from commonly available monitoring datasets. To develop an optimized PNEC estimation model based on an available monitoring dataset, an initial model that considers all BLM variables, a second model that requires variables excluding alkalinity, and a third model using electrical conductivity as a surrogate of the major cations and alkalinity have been proposed. Furthermore, deep neural network (DNN) models have been used to predict the nonlinear relationships between the PNEC (outcome variable) and the required input variables (explanatory variables). The predictive capacity of DNN models in this study was compared with the results of other existing PNEC estimation tools using a look-up table and multiple linear and multivariate polynomial regression methods. Three DNN models, using different input variables, provided better predictions of the copper PNECs compared with the existing tools for four test datasets, i.e., Korean, United States, Swedish, and Belgian freshwaters. The adjusted r2 values in all DNN models were higher than 0.95 in the test datasets, except for the Swedish dataset (adjusted r2 > 0.87). Consequently, the most applicable model among the three DNN models could be selected according to the data availability in the collected monitoring database. Because the most simplified DNN model required only three water quality variables (pH, dissolved organic carbon, and electrical conductivity) as input variables, it is expected that the copper BLM-based risk assessment can be applied to monitoring datasets worldwide.

Application and Validation of a Biotic Ligand Model for Calculating Acute Toxicity of Lead to Moina Dubia in Lakes of Hanoi, Vietnam

It is increasingly being recognized that biotic ligand models (BLMs) are valuable in the risk assessment of metals in aquatic systems. The authors investigated the effect of pH, Ca, Mg, K, Na on the acute toxicity of Pb to Moina dubia, native zooplankton in lakes of Hanoi, Vietnam. Calcium, Magnesium and pH strongly influenced acute Pb toxicity to Moina dubia. Based on this data set, a acute Pb-BLM for Moina dubia was developed according to condition of Hanoi lakes. The developed BLM was shown, in an independent validation with data on acute toxicity test on natural water sets, to be capable of predicting chronic Pb toxicity with 81.3% accuracy. The results proved that BLM can be useful tool for calculating the acute toxicity based on water-quality criteria in lake of Hanoi.

The toxicity of cadmium-copper mixtures on daphnids and microalgae analysed using the Biotic Ligand Model

For the prediction of metals mixture ecotoxicity, the BLM approach is promising since it evaluates the amount of metals accumulated on the biotic ligand on the basis of water chemistry, i.e. species (major cations) competing with metals, and related toxicity. Based on previous work by Farley et al. 2015 (MMME research project), this study aimed at modelling toxicity of Cd:Cu mixtures (0:1–1:1–1:0–1:2 − 1:3 − 2:1–3:1–4:1–5:1–6:1) to the crustacean Daphnia magna (48h immobilization tests) and the microalga Pseudokirchneriella subcapitata (72h growth inhibition tests). The USGS model was chosen, assuming additivity of effects and accumulation of metals on a single site. The assumption that EDTA could contribute to toxicity through metals complexing was also tested, and potential effects due to reduction of ions Ca2+ absorption by metals were considered. Modelling started with parameter values of Farley et al. 2015 and some of these parameters were adjusted to fit modelled data on observed data. The results show that toxicity can be correctly predicted for the microalgae and that the hypothesis of additivity is verified. For daphnids, the prediction was roughly correct, but taking into account CuEDTA led to more realistic parameter values close to that reported by Farley et al. 2015. However, It seems that, for daphnids responses, metals interact either antagonistically or synergistically depending on the Cu:Cd ratio. Furthermore, synergy could not be explained by additional effects linked to a reduction of Ca absorption since this reduction, mainly due to Cd, increased inversely to synergy. Finally, the USGS model applied to our data was able to predict Cu:Cd mixture toxicity to microalgae and daphnids, giving rise to estimated EC50s roughly reflecting EC50s calculated from observed toxicity.