Preferential flow and pesticide transport in a clay-rich till: Field, laboratory, and modeling analysis

[1] This study investigates vertical flow and pesticide transport along fractures in water saturated unoxidized clayey till. From two experimental fields, each 40 m(2), 96% and 98%, respectively, of total vertical flow was conducted along fractures in the till, while the remaining 2-4% of flow occurred in the clay matrix at very slow flow rate. An applied dye tracer was observed only along 10-26% of the total fracture length measured on the horizontal surface of the experimental fields. In vertical sections the dyed fracture portions constituted root channels, which penetrated the till vertically along the fractures into the local aquifer at 5 m depth. No dye tracer was observed in the fractures without root channels or in the unfractured clay matrix, suggesting that root growth along the fracture surfaces was the principal agent of fracture aperture enhancement. Using hydraulic fracture aperture values determined from large undisturbed column (LUC) collected from one of the experimental fields, it was estimated that 94% of flow in the fractures was conducted along the fracture root channels, while only 6% of flow was conducted along the fracture sections without root channels. For natural vertical hydraulic gradients (0.8-2.3 at the site), flow rates of 0.8-2 km/d were determined for a fracture root channel, while fracture sections without root channels revealed flow rates of 9-22 m/ d. Corresponding flow rates in the unfractured matrix were 7-19 mm/yr. For infiltrated bromide (nonreactive tracer) and mobile pesticides mecoprop (MCPP) and metsulfuron, very rapid migration (0.28-0.5 m/ d) and high relative breakthrough concentrations (30-60%) into the aquifer were observed to occur along the fracture root channels using a constant hydraulic gradient of 1. Only traces were measured from infiltration of the strongly sorbed pesticide prochloraz. The concentrations of the bromide and pesticides in the monitoring wells were modeled with a discrete fracture matrix diffusion (DFDM) model coupled with a single porosity model (SP) for the till and aquifer, respectively. Using effective fracture spacings and mean fracture apertures for the fracture channel sections as modeling input parameters for the till, the concentrations observed in the wells of the aquifer could be reasonably approximated.