Spatial Mapping of Riverbed Grain-Size Distribution Using Machine Learning

Abstract

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 and hydraulic conductivity), 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. Such information can be obtained based on the intrinsic link between sediment grain-size distribution and hydraulic properties where sediment texture information is available. However, grain-size measurements are usually spatially sparse and do not have adequate coverage and resolution, particularly for a relatively large domain such as the Hanford Reach of the Columbia River. In this paper, we adopted machine learning (ML) approaches for categorizing and mapping the spatial distributions of riverbed substrate grain size 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 measured substrate sizes 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). An ensemble bagging-based ML technique, Random Forest, was adopted to identify the most influential factors as predictors to develop the predictive models with over-fitting issues addressed. The models were evaluated with respect to each individual substrate size class and the lumped group, and then used to generate the final substrate size maps covering all the grid cells in the numerical modeling domain.

Combined Effects of Stream Hydrology and Land Use on Basin‐Scale Hyporheic Zone Denitrification in the Columbia River Basin

“…In this study, hyporheic exchange flux was estimated based on the NEXSS simulation (Gomez‐Velez et al., 2015), and its flux was highly dependent on streambed sediment grain size/hydraulic conductivity estimates. To reduce the uncertainty of our RCM, future studies should focus on collecting detailed measurements of hydraulic conductivities (Ren et al., 2020; Stewardson et al., 2016) and developing advanced methods characterizing the spatial variation of hydraulic conductivities (Abimbola et al., 2020). In addition, the current model only represented the spatial averaged conditions of HZ denitrification in the CRB, and key model input variables were temporally constant.…”

“…In some cases, both approaches can be combined to gain further insight and predictability. For example, the model can be used to reveal the dominant process or features through variable importance analysis (Ren et al, 2020(Ren et al, , 2021Ward et al, 2022).…”

Automated riverbed material analysis using Deep Learning on underwater images

“…Since the introduced method offers a quick way to provide spatially continuous bed material information of its composition, it may be used to boost the training dataset of predictive, ensemble bagging-based Machine Learning techniques (e.g., Ren et al, 2020) and improve their accuracy. Furthermore, the method can support the implementation of other imagery techniques.…”

“…The method was able to reach a limited execution speed of a few seconds per m 2 and adequately measured the sizes of gravels. Ren et al (2020) applied an ensemble bagging-based Machine Learning (ML) algorithm to estimate GSD along the 70 km long Hanford Reach of the Columbia river. Due to its economic importance, a large amount of measurement data has been accumulated for this study site over the years, making it ideal for using ML.…”

Machine Learning Analysis of Hydrologic Exchange Flows and Transit Time Distributions in a Large Regulated River

“…For most of the simulations reported here, a value of 1.0 × 10 −12 m was applied homogenously across the entire riverbed surface. To test the potential impacts of riverbed heterogeneity on the model predictions, we also considered one case in which different values of conductance were applied as a function of the HU observed at each grid cell location, based on the association of HUs with riverbed substrate size maps (Ren et al, 2020). Table 2 lists the values of the conductance coefficient assigned to each HU for homogeneous and heterogeneous cases.…”