Machine Learning of Hydrogeochemical Processes along the Columbia River Corridor

Abstract:

Hydrologic exchange between river channels and adjacent subsurface environments is a key process that influences water quality and ecosystem function in river corridors. Predictive numerical models are needed to understand responses of river corridors to environmental change and to support sustainable watershed management. Here we summarize our research work developing and applying machine learning algorithms for (1) hydromorphic classification, which provides a scaling construct enabling extrapolation of outputs from local-scale mechanistic models to reduced-order models applicable at reach and watershed scales; (2) spatial mapping of riverbed substrate grain size distributions and deriving spatially heterogeneous hydraulic property fields; and (3) derivation of reduced order models of hydrologic exchange flows and non-reactive tracer residence time distributions. These form the basis of developing predictive reduced-order models of hydrogeochemical processes along the Columbia River Corridor, which are potentially transferrable to other river reaches and larger scales.