Date of Award


Degree Name

Doctor of Philosophy




A 4 km2 agricultural area within the somewhat poorly drained Nottawa Creek Watershed, south central Michigan, was investigated to help understand transport and transformation processes of nutrients and herbicides to surface water. Geology o f the study area is comprised of a thin glacial drift cover overlying shale bedrock. A one layer, two-dimensional numerical triangular finite element groundwater flow model was developed to evaluate regional groundwater flow. Selected horizontal hydraulic conductivities ranged from 3 x 10‘3 m/day to 107 m/day. Field-scale infiltration modeling with the computer program GLEAMS, substantiated with water level measurements, suggest significant recharge events occur during summer, as well as fall and spring. Depth to groundwater varied from about 1.5 to 7.6 meters. Geochemical data were collected for 12 sampling events over 22 months, from December 1999-September 2001, for Nottawa Creek surface water, 9 water table monitoring wells, 4 domestic drift wells, and 8 tile drains. Tile drains were a major contributor o f nitrogen and phosphorus to surface water. Despite this, Nottawa Creek nutrient concentrations were consistently low. Tritium results from the two streambed wells, which were uncontaminated with nitrate, indicate groundwater discharging into Nottawa Creek is post-bomb water. Values o f S15N and 5I80 from nitrate from most wells and tile drains suggest that field applied synthetic fertilizer, composted manure, and liquid manure is transformed or not present. Wells adjacent to an animal waste lagoon and a compost pile exhibit a denitrification or animal waste isotopic signature. Concentrations detected by immunoassay, GC/MS, and LC/MS are reported for triazine and acetamide herbicides. Herbicides and degradation products are transported to surface water by tile drains. Low or non-detect herbicide concentrations in monitoring wells suggest significant degradation or retardation. Immunoassay test results suggest metolachlor ESA was detected in wells, rather than metolachlor. Herbicide testing implied a seasonal concentration detection peak occurring in June-July, which subsequently tapered off dramatically. Low early fall concentrations were substantiated by GC/MS and LC/MS methods in which only metolachlor and alachlor metabolites were detected in streambed wells and Nottawa Creek surface water. Atrazine and other triazine herbicides were not detected for the LLD of 0.05 p/L.

Access Setting

Dissertation-Open Access