Multi-model ensemble simulated non-point source pollution based on Bayesian model averaging method and model uncertainty analysis

The Utility of Machine Learning Models for Predicting Chemical Contaminants in Drinking Water: Promise, Challenges, and Opportunities

PURPOSE OF REVIEW

This review aims to better understand the utility of machine learning algorithms for predicting spatial patterns of contaminants in the United States (U.S.) drinking water.

RECENT FINDINGS

We found 27 U.S. drinking water studies in the past ten years that used machine learning algorithms to predict water quality. Most studies (42%) developed random forest classification models for groundwater. Continuous models show low predictive power, suggesting that larger datasets and additional predictors are needed. Categorical/classification models for arsenic and nitrate that predict exceedances of pollution thresholds are most common in the literature because of good national scale data coverage and priority as environmental health concerns. Most groundwater data used to develop models were obtained from the United States Geological Survey (USGS) National Water Information System (NWIS). Predictors were similar across contaminants but challenges are posed by the lack of a standard methodology for imputation, pre-processing, and differing availability of data across regions. We reviewed 27 articles that focused on seven drinking water contaminants. Good performance metrics were reported for binary models that classified chemical concentrations above a threshold value by finding significant predictors. Classification models are especially useful for assisting in the design of sampling efforts by identifying high-risk areas. Only a few studies have developed continuous models and obtaining good predictive performance for such models is still challenging. Improving continuous models is important for potential future use in epidemiological studies to supplement data gaps in exposure assessments for drinking water contaminants. While significant progress has been made over the past decade, methodological advances are still needed for selecting appropriate model performance metrics and accounting for spatial autocorrelations in data. Finally, improved infrastructure for code and data sharing would spearhead more rapid advances in machine-learning models for drinking water quality.

New framework for managing the water environmental capacity integrating the watershed model and stochastic algorithm

Integrated calculations of pollution load and water environmental capacity (WEC) are essential for effective water quality management. However, few studies have focused on the dynamic WEC and pollution load in a nonpoint source pollution (NPS)-dominant temperate monsoon watershed under changing rainfall conditions. In this study, a new framework based on the watershed model and WEC calculation with stochastic rainfall input (SR-WEC), was proposed to reveal the dynamic WEC and pollution load under changing rainfall conditions. Stochastic rainfall series was generated by a first-order Markov chain and gamma distribution, and further input into the Soil and Water Assessment Tool (SWAT) to explore the dynamic response of water quality to rainfall. The framework was applied to the Daning River watershed, Three Georges Reservoir Region, China. The results suggested that compared with the new SR-WEC, the traditional return period method with limited observed rainfall input would result in an underestimation of ideal WEC and NPS pollution load by 23% and 48% for TN and 48% and 51% for TP, respectively. Approximately 46% of the annual TN reduction and 51% of the annual TP reduction were concentrated from April to June in a relatively small area. The regression relationships between rainfall and the ideal WEC, pollution load and remnant WEC obtained by the SR-WEC were superior to those of the traditional method, with R² values increasing from 0.005-0.797 to 0.718-0.989. Specific threshold (120 mm/month for the study area) was observed for the effect of rainfall on water quality, beyond which the remnant WEC of organic N would change from decreasing to increasing. The new framework proposed identifies the key periods and areas with consideration of uncertainty of rainfall on water quality, and provides basis for NPS pollution management.