Spatial and temporal variability of preferential flow in a subsurface- drained landscape in North-Central Iowa.

Abstract

Preferential flow can have a direct impact on agricultural chemical transport, especially where preferential flow allows a significant fraction of the total rainfall to quickly move to underlying subsurface drains. The spatial and temporal variability of preferential flow was studied from 2006-2008 in a 38.8 hectare row-cropped (corn, soybeans) agricultural field with two topographically-located subsurface drains (north-central Iowa). Continuous measurements of water levels, soil moisture, stream and drain discharge, stream and drain specific conductance (SC) and precipitation were obtained. Persistent, but variable preferential flow was found to exist at the study site throughout the growing season. Four independent mathematical approaches were used to explore the spatial and temporal variability of preferential flow. A specific conductance end-member mixing analysis (EMMA) was performed based on the temporally varying SC in the subsurface drain water for two separate events, with 54 events classified solely on decreases in SC without accounting for drain flow (full calculations in supplementary file: SCanalysis.xlsm) . These 54 events were utilized to calculate the mean onset time of preferential flow and the mean time to maximum preferential flow. The maximum water velocity was calculated based on the initial rise of the observed water level relative to rainfall (full calculations in supplementary file: MaxTransportVelocity.xlsx). The highest maximum water velocity values occurred in the early spring and mid-to-late summer (i.e., higher number of preferential flow pathways), and the slowest maximum water velocity in the fall and early summer (i.e., lower number of preferential flow pathways). Spatial and temporal variability of preferential flow was best assessed with maximum water velocity. A simple unsaturated zone, one-dimensional, dual domain model (source-responsive) was conducted based on the timing and magnitude of the change in ground water levels relative to precipitation events (full calculations in supplementary file: S-Rmodel.xlsm). The source-responsive model described the size of the preferential flow space necessary to accommodate preferential flow. Finally, a simple theoretical model of soil infiltration and flow to subsurface drains (DRAIN-Pro), based on the Green-Ampt and Hooghoudt equations, accounted for preferential flow by estimating the effective vertical and horizontal hydraulic conductivities.