Machine Learning for Freshwater Analysis

Researchers from Syracuse University and Texas A&M University have used machine learning to study the extent to which human activity contributes to hydrogeochemical changes in United States rivers. Salt and alkaline levels in freshwater sources can make water undrinkable and negatively impact wildlife, yet increasing alkalinity can have a positive effect by acting to neutralize acidity and absorb carbon dioxide. The model used data from 226 monitoring sites across the country, finding that population density and impervious surfaces were contributing factors to higher salt content. In terms of alkalinity, local climate and hydrogeological conditions were found to be the main influencers rather than human activity, though researchers add that these natural factors can still be adversely impacted by human changes.