Subsurface Sediment Transport in the Shallow Vadose Zone of Fine-Textured Soils With Heterogenous Preferential Flows

Abstract

Subsurface sediment transport in tile-drained landscapes occurs through macropores; however, little is known regarding how heterogeneous preferential flows influence fluxes. We performed laboratory rainfall simulations on 10 intact core lysimeters from a tile-drained field in Indiana, USA to study the impacts of surface and subsurface erosion on sediment leachate in heterogeneous preferential flow paths. Seven rainfall simulations were conducted to assess the impact of rainfall intensity on the leachate of surface eroded sediments (three events), and the impact of antecedent conditions on subsurface eroded sediments (four events). Cumulative sediment yield, linear mixed effects modelling, and hysteresis analyses were performed for all events. Results were presented in a series of four case studies. Results showed that surface sediment leachate concentration and yield were tightly linked to the filtration capacity of lysimeters, with more than 2/3rd of sediment originating from a single lysimeter, despite similar flow leachate volumes from each. Rainfall intensity significantly impacted the transport of surface eroded sediment at the highest intensity. Subsurface sediment erosion from undisturbed macropores was low compared to surface soils, but we found contrasting controls on sediment concentrations at low and high antecedent moistures that were equally important to sediment leachate yields. Disturbed macropores produced comparable sediment yields to surface erosion and behaved similarly to soil pipes in terms of erosion mechanics. Hysteresis results generally highlighted contrasting results for surface and subsurface sources but suggest that the prominence of slow flow, low-concentration leachate sources can alter the interpretation of results in field-scale applications. Our findings underscore an array of processes and pathways for sediment transport in the shallow vadose zone, and results will be useful for evaluating new model formulations.