Spatial Mapping of Riverbed Grain-Size Distribution Using Machine Learning

Recent alluvial sediments in riverbeds play a significant role in controlling hydrologic exchange flows (HEFs) in river systems. The alluvial layer is usually associated with strong heterogeneity in physical properties (e.g., permeability), which affects local HEFs and therefore biogeochemical processes. The spatial distribution of these physical properties needs to be determined to inform the numerical models used to reveal the realistic hydro-biogeochemical behaviors. One big challenge to mapping the heterogeneous permeability field, particularly for a relatively large domain such as the Hanford Reach of the Columbia River, is that direct measurements of permeability are spatially sparse and do not have adequate coverage and resolution. In this paper, we adopted machine learning (ML) approaches for categorizing and mapping the spatial distributions of riverbed substrate size, which serves as a proxy for permeability (e.g., using Shepherd’s formula calibrated in our previous work), and filling in missing areas of substrate data using the ML models along the reach. Such ML models for substrate size mapping were trained at 13,372 locations using substrate sizes measured by the U.S. Fish and Wildlife Service along with observed and simulated attributes, including bathymetric attributes (e.g., elevation, slope, and aspect ratio) from LIDAR and bathymetric surveys, and hydrodynamic properties (e.g., water depth, velocity, shear stress, and their statistical moments) obtained from the Modular Aquatic Simulation System 2D simulator. A bagging-based ML technique, Random Forest, was applied to identify the most influential factors as predictors to develop the predictive models and generate the final substrate size maps.