Reach-scale bankfull channel types can exist independently of catchment hydrology

Byrne, C. F.; Pasternack, G. B.; Guillon, Hervé; Lane, Belize; Solís, Samuel Sandoval

doi:10.1002/essoar.10501068.1

Cited by 4 publications

(18 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Channel types were statistically derived based on multivariate clustering and comparison to field observations to achieve a best fit to all field-surveyed sites. Canals and ditches were excluded from Byrne et al's. (2019) analysis, and while some field sites exist in areas with human influence, channel types are not defined by anthropogenic alteration.…”

Section: Case Studymentioning

confidence: 99%

“…(2019) analysis, and while some field sites exist in areas with human influence, channel types are not defined by anthropogenic alteration. The evidenced natural Water Resources Research 10.1029/2019WR026691 Table 1 Channel Types Identified in the Sacramento Basin by Byrne et al (2019) Channel variability of rivers in the study basin (Byrne et al, 2019; makes it an ideal test bed for testing large-scale geomorphic classification needed for future integrative research (e.g., .…”

Section: Case Studymentioning

confidence: 99%

“…This study aims to predict channel types for 108,943 two hundred-meter-long stream intervals throughout the stream network. As 10 possible channel types have been previously identified (Byrne et al, 2019), a supervised ML classification approach answers the following question: Given a set of predictors, which class (channel type) should be assigned to each stream reach? To achieve this study aim, we developed a three-tiered ML framework ( Figure 1) to (i) define a tractable problem and reduce predictor noise, (ii) assess the performance of models in statistical learning, and (iii) assess the performance of models in predictive modeling.…”

Section: A Multitiered ML Frameworkmentioning

confidence: 99%

“…Although ML approaches can incorporate a large number of predictors and labels, irrelevant or noisy data may deteriorate model performance and increase computational load. Because each channel type was expertly inspected in the unsupervised learning phase by Byrne et al (2019), the labels are treated as noiseless and we refrained from implementing techniques that evaluate and correct for uncertain labeling (e.g., Borut et al, 2010;Garcia et al, 2012Garcia et al, , 2015Sluban et al, 2013). To filter potential predictor noise, the complexity of the problem was assessed by calculating data complexity measures (DCM; Lorena et al, 2018).…”

Section: Define a Tractable Problem And Reduce Predictor Noisementioning

confidence: 99%

“…Byrne et al () previously identified 10 dominant geomorphic channel types present in the Sacramento River basin from data acquired at 290 field sites (Table ). Channel types were statistically derived based on multivariate clustering and comparison to field observations to achieve a best fit to all field‐surveyed sites.…”

Section: Case Studymentioning

confidence: 99%

See 4 more Smart Citations

Machine Learning Predicts Reach‐Scale Channel Types From Coarse‐Scale Geospatial Data in a Large River Basin

Guillon

Byrne

Lane

et al. 2020

Water Resources Research

Self Cite

View full text Add to dashboard Cite

Hydrologic and geomorphic classifications have gained traction in response to the increasing need for basin‐wide water resources management. Regardless of the selected classification scheme, an open scientific challenge is how to extend information from limited field sites to classify tens of thousands to millions of channel reaches across a basin. To address this spatial scaling challenge, this study leverages machine learning to predict reach‐scale geomorphic channel types using publicly available geospatial data. A bottom‐up machine learning approach selects the most accurate and stable model among ∼20,000 combinations of 287 coarse geospatial predictors, preprocessing methods, and algorithms in a three‐tiered framework to (i) define a tractable problem and reduce predictor noise, (ii) assess model performance in statistical learning, and (iii) assess model performance in prediction. This study also addresses key issues related to the design, interpretation, and diagnosis of machine learning models in hydrologic sciences. In an application to the Sacramento River basin (California, USA), the developed framework selects a Random Forest model to predict 10 channel types previously determined from 290 field surveys over 108,943 two hundred‐meter reaches. Performance in statistical learning is reasonable with a 61% median cross‐validation accuracy, a sixfold increase over the 10% accuracy of the baseline random model, and the predictions coherently capture the large‐scale geomorphic organization of the landscape. Interestingly, in the study area, the persistent roughness of the topography partially controls channel types and the variation in the entropy‐based predictive performance is explained by imperfect training information and scale mismatch between labels and predictors.

show abstract

Section: Case Studymentioning

confidence: 99%

Section: Case Studymentioning

confidence: 99%

Section: A Multitiered ML Frameworkmentioning

confidence: 99%

Section: Define a Tractable Problem And Reduce Predictor Noisementioning

confidence: 99%

Section: Case Studymentioning

confidence: 99%

See 3 more Smart Citations

Machine Learning Predicts Reach‐Scale Channel Types From Coarse‐Scale Geospatial Data in a Large River Basin

Guillon

Byrne

Lane

et al. 2020

Water Resources Research

Self Cite

View full text Add to dashboard Cite

show abstract

Identifying key channel variability functions controlling ecohydraulic conditions using synthetic channel archetypes

2023

View full text Add to dashboard Cite

Geometric modelling of river channel topography is a method of design synthesis wherein specific 2D geometric elements of river topography, such as the bed profile, cross sectional shape, and channel planform contours, are expressed mathematically in isolation and then combined to produce a 3D heightmap. We utilized the geometric modelling framework to synthesize channel terrains that reveal flow−form−function linkages to investigate what roles variability in bed roughness, thalweg elevation and channel width play in defining hydraulic and ecohydraulic conditions of a channel reach from baseflow to bankfull discharge. To achieve a robust inquiry for a range of settings, this study developed four distinct synthetic channel terrain models for each of three stream reaches of different channel types in the South Fork Eel River in northern coastal California, USA. To test the process-based effects of these diverse terrain synthesis options, we compared the resulting hydraulic patterns and preferred habitat availability for fry/juvenile steelhead trout and coho salmon over a range of discharges. Among thalweg bed undulation, width variation and bed roughness, we found thalweg bed undulation was the key factor affecting the channel ecohydraulic response at baseflow condition. At bankfull condition, thalweg bed elevation had the largest effect in high-order mainstem streams identified by gravelcobble dominated high width-to-depth riffle-pool sequences, width variation had the largest effect in mid-order confined channels with gravel-cobble, high width-to-depth ratio with expansions/contractions and bed roughness had the largest effect in low-order streams with low width-to-depth ratio, high-gradient, cobble-boulder and step-pool/cascade channels.

show abstract

Channel‐reach morphology and landscape properties are linked across a large heterogeneous region

Lane

Guillon

Byrne

et al. 2021

Earth Surf Processes Landf

Self Cite

View full text Add to dashboard Cite

Given the complex array of processes influencing river networks, conceptual frameworks of rivers are critical to our understanding of channel processes and response potential as well as restoration efforts. Yet despite their wide usage, many classifications are based on limited observations over homogenous landscapes, raising questions about their general applicability and quantitative thresholds. Leveraging a large, transect-based morphological field dataset across California, USA, we use data-driven methods to evaluate multivariate patterns in channel morphology and linkages with landscape properties considering a diversity of physioclimatic settings. Emergent patterns highlight the variability in channel form observed across an extensive dataset over heterogeneous but spatially linked watersheds. In general, identified dominant channel attributes and landscape properties align with established channel types defined through expert judgement, but key differences also emerge. Similar to past studies, bed sediment composition and sub-reach depth variability were discriminating channel attributes. The dominance of landscape properties associated with sediment supply or transport capacity suggests that morphological diversity largely reflects these differences as posited by prior classifications. Results also show some channel forms to be largely independent of valley confinement, with several channel bedforms and dominant grain sizes occurring across valley settings. This data analysis study demonstrates the utility of considering channel reaches and landscapes as multidimensional features to elucidate and test established geomorphic understanding over large field datasets.

show abstract

Reach-scale bankfull channel types can exist independently of catchment hydrology

Cited by 4 publications

References 16 publications

Machine Learning Predicts Reach‐Scale Channel Types From Coarse‐Scale Geospatial Data in a Large River Basin

Machine Learning Predicts Reach‐Scale Channel Types From Coarse‐Scale Geospatial Data in a Large River Basin

Identifying key channel variability functions controlling ecohydraulic conditions using synthetic channel archetypes

Channel‐reach morphology and landscape properties are linked across a large heterogeneous region

Contact Info

Product

Resources

About