Improved Decision Making for Water Lead Testing in U.S. Child Care Facilities Using Machine-Learned Bayesian Networks.

Tap water lead testing programs in the U.S. need improved methods for identifying high-risk facilities to optimize limited resources. In this study, machine-learned Bayesian network (BN) models were used to predict building-wide water lead risk in over 4,000 child care facilities in North Carolina according to maximum and 90th percentile lead levels from water lead concentrations at 22,943 taps. The performance of the BN models was compared to common alternative risk factors, or heuristics, used to inform water lead testing programs among child care facilities including building age, water source, and Head Start program status. The BN models identified a range of variables associated with building-wide water lead, with facilities that serve low-income families, rely on groundwater, and have more taps exhibiting greater risk. Models predicting the probability of a single tap exceeding each target concentration performed better than models predicting facilities with clustered high-risk taps. The BN models' Fß-scores outperformed each of the alternative heuristics by 118-213%. This represents up to a 60% increase in the number of high-risk facilities that could be identified and up to a 49% decrease in the number of samples that would need to be collected by using BN model-informed sampling compared to using simple heuristics. Overall, this study demonstrates the value of machine-learning approaches for identifying high water lead risk that could improve lead testing programs nationwide.

Nitrogen and sulfur deposition reductions projected to partially restore forest soil conditions in the US Northeast, while understory composition continues to shift with future climate change.

Human activities have dramatically increased nitrogen (N) and sulfur (S) deposition, altering forest ecosystem function and structure. Anticipating how changes in deposition and climate impact forests can inform decisions regarding these environmental stressors. Here, we used a dynamic soil-vegetation model (ForSAFE-Veg) to simulate responses to future scenarios of atmospheric deposition and climate change across 23 Northeastern hardwood stands. Specifically, we simulated soil percent base saturation, acid neutralizing capacity (ANC), nitrate (NO3 -) leaching, and understory composition under 13 interacting deposition and climate change scenarios to the year 2100, including anticipated deposition reductions under the Clean Air Act (CAA) and Intergovernmental Panel on Climate Change-projected climate futures. Overall, deposition affected soil responses more than climate did. Soils recovered to historic conditions only when future deposition returned to pre-industrial levels, although anticipated CAA deposition reductions led to a partial recovery of percent base saturation (60 to 72%) and ANC (65 to 71%) compared to historic values. CAA reductions also limited NO3 - leaching to 30 to 66% above historic levels, while current levels of deposition resulted in NO3 - leaching 150 to 207% above historic values. In contrast to soils, understory vegetation was affected strongly by both deposition and climate. Vegetation shifted away from historic and current assemblages with increasing deposition and climate change. Anticipated CAA reductions could maintain current assemblages under current climate conditions or slow community shifts under increased future changes in temperature and precipitation. Overall, our results can inform decision makers on how these dual stressors interact to affect forest health, and the efficacy of deposition reductions under a changing climate.