.png){kind=icon}
.png){kind=icon}
Return
Evaluating Bedrock Depth Proxies as Representations of Subsurface Controls on Runoff Generation
.png){kind=icon}
OA
AI
L
J
DOI：10.1002/hyp.70621.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Large-sample hydrologic studies increasingly rely on readily available subsurface characteristics despite persistent uncertainty about whether these datasets meaningfully represent controls on runoff generation at the catchment scale. Regional and field-based studies throughout the Southern Appalachians region show that subsurface structure strongly influences storage, connectivity, and runoff response. However, large sample analyses often show that subsurface characteristics, particularly bedrock depth, underperform relative to climate and landscape characteristics. We evaluated six of the currently available large-scale data products representing either bedrock depth or which may be a proxy for bedrock depth in the Southern Appalachians Mountains region. We used a novel variable selection analysis to identify reduced predictor sets from all catchment characteristics and evaluate the predictive influence of these six bedrock depth proxies. Then, many random forest models were fit and interrogated to compare the performance across datasets containing individual proxies, no proxies, and multiple proxies. Most proxies describing bedrock depth produced only marginal and inconsistent improvements in model performance (< 10% increase in R2). Several modelled relationships aligned with regional hydrologic theory, though datasets that intended to represent similar characteristics showed weak agreement. Models incorporating multiple bedrock depth proxies generally performed best, yet the influence of individual proxies diminished when evaluated together, suggesting that relationships between hydrologic signatures and bedrock depth proxies may reflect shared information rather than unique physical controls. Our findings suggest that while bedrock depth and related proxies are important on runoff generation, existing datasets capture only fragments of these controls and may be most effective when considered together.
Keywords:
bedrock depth
catchment attributes
hydrologic signatures
lithology
random forest
subsurface controls
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
2.9
Papers:
614
Citations:
2.2W
Organization
Cited Papers