The number of chemicals in the market is rapidly increasing, while our understanding of the life-cycle impacts of these chemicals lags considerably. To address this, we developed deep artificial neural network (ANN) models to estimate life-cycle impacts of chemicals. Using molecular structure information, we trained multilayer ANNs for life-cycle impacts of chemicals using six impact categories, including cumulative energy demand, global warming (IPCC 2007), acidification (TRACI), human health (Impact2000+), ecosystem quality (Impact2000+), and eco-indicator 99 (I,I, total). The application domain (AD) of the model was estimated for each impact category within which the model exhibits higher reliability. We also tested three approaches for selecting molecular descriptors and identified the principal component analysis (PCA) as the best approach. The predictions for acidification, human health, and the eco-indicator 99 model showed relatively higher performance with R values of 0.73, 0.71, and 0.87, respectively, while the global warming model had a lower R of 0.48. This study indicates that ANN models can serve as an initial screening tool for estimating life-cycle impacts of chemicals for certain impact categories in the absence of more reliable information. Our analysis also highlights the importance of understanding ADs for interpreting the ANN results.

Most existing life-cycle release models for engineered nanomaterials (ENM) are static, ignoring the dynamics of stock and flows of ENMs. Our model, nanoRelease, estimates the annual releases of ENMs from manufacturing, use, and disposal of a product explicitly taking stock and flow dynamics into account. Given the variabilities in key parameters (e.g., service life of products and annual release rate during use) nanoRelease is designed as a stochastic model. We apply nanoRelease to three ENMs (TiO2, SiO2 and FeO x ) used in paints and coatings through seven product applications, including construction and building, household and furniture, and automotive for the period from 2000 to 2020 using production volume and market projection information. We also consider model uncertainties using Monte Carlo simulation. Compared with 2016, the total annual releases of ENMs in 2020 will increase by 34–40%, and the stock will increase by 28–34%. The fraction of the end-of-life release among total release flows will increase from 11% in 2002 to 43% in 2020. As compared to static models, our dynamic model predicts about an order of magnitude lower values for the amount of ENM released from this sector in the near-term while stock continues to build up in the system.

With climate change, the extent, severity, and frequency of droughts around the world are expected to increase. This study analyzed the ability of water districts to meet mandatory urban water conservation targets, which are a common policy response to drought. During California's recent record‐breaking drought, a 25% state‐wide use reduction objective was set and met. However, only 50% of urban water districts analyzed in this study reached their individual conservation target, which offers an opportunity to evaluate the factors associated with successful water use reduction. The findings show that the inclusion of water districts in the polycentric import structure may improve water conservation, but that source diversity may offer water districts a perceived buffer from the need for immediate water use reductions. Drought severity and lower median incomes are associated with greater water conservation, and conservation varies by hydrologic region. This analysis offers insights into institutional design and suggests that local biophysical and economic conditions shape responses in systematic ways that should be addressed by public policy responses to drought.

Species Sensitivity Distribution (SSD) is a key metric for understanding the potential ecotoxicological impacts of chemicals. However, SSDs have been developed to estimate for only handful of chemicals due to the scarcity of experimental toxicity data. Here we present a novel approach to expand the chemical coverage of SSDs using Artificial Neural Network (ANN). We collected over 2000 experimental toxicity data in Lethal Concentration 50 (LC50) for 8 aquatic species and trained an ANN model for each of the 8 aquatic species based on molecular structure. The R2 values of resulting ANN models range from 0.54 to 0.75 (median R2 = 0.69). We applied the predicted LC50 values to fit SSD curves using bootstrapping method, generating SSDs for 8424 chemicals in the ToX21 database. The dataset is expected to serve as a screening-level reference SSD database for understanding potential ecotoxicological impacts of chemicals.

No abstract