USGS1996
The U.S. Geological Survey, in cooperation with the National Soil Tilth Laboratory of the U.S. Department of Agriculture, Agricultural Research Service, conducted a study as part of the multi-scale, inter-agency Management Systems Evaluation Area (MSEA) program to evaluate the effects of agricultural management (farming) systems on water quality. Data on surface flow, tileflow, and streamflow in the Walnut Creek watershed just south of Ames, Iowa, were collected during April 1991-September 1993 at five sites with drainage areas ranging from 366 to 5,130 hectares. Precipitation, flow discharge, and concentration, loads, and yields of nitrate as nitrogen, atrazine, and metolachlor were analyzed to relate the transport of agricultural chemicals to major water-flow processes and to examine flow and transport differences among three subwatersheds. Antecedent conditions and basin-characteristic differences had significant effects on the flow response from the subwatersheds. Monthly streamflow-to-precipitation ratios were greater than 1.0, as a result of snowmelt, and negative when streamflow was lost to the ground-water system in the downstream subwatershed. Dry antecedent conditions resulted in ratios less than 0.3 (July 1992), whereas wet antecedent conditions resulted in ratios from 0.7 to almost 1.0 (July 1993) during months with similar large rainfall amounts. Most of the streamflow from the upland subwatersheds came from tileflow. Surface flow (surface runoff, interflow, and return flow) was highly variable and intermittent, usually lasting for only a few days after a storm, although it could be the dominant source of flow when stormflow was large. Tileflow was less variable and much more persistent, ceasing only after prolonged dry periods. Large quantities of nitrate as nitrogen were transported in Walnut Creek, with concentrations often greater than the Maximum Contaminant Level of 10 milligrams per liter established by the U.S. Environmental Protection Agency for finished drinking water. In the upland subwatersheds, ground-water flow from the tiles appears to have been the primary means of transport to the streams. Concentrations in tileflow and streamflow generally were 4 to 16 milligrams per liter, with the lower concentrations often the result of dilution by surface runoff. Loss ratios, chemical yields expressed as a percentage of average application rates of nitrate as nitrogen for October 1992-September 1993, were about 10 percent for surface flow and more than 100 percent for tileflow from the 366-hectare basin and were more than 200 percent for streamflow from the downstream subwatershed. Concentrations of atrazine and metolachlor in streamflow, typically, were less than the Maximum Contaminant Level of 3.0 micrograms per liter, but were as high as 59 and 80 micrograms per liter, respectively, during stormflow. Concentrations as high as 170 micrograms per liter occurred in tileflow, but these were related to surface flow through surface inlets. The transport of herbicides was extremely variable, with most of the loads occurring during stormflow. Atrazine appeared more susceptible to transport losses to streamflow than did metolachlor. Loss ratios for streamflow from the subwatersheds for April-September periods were 0.3 to 20 percent for atrazine and 0.1 to 2.9 percent for metolachlor. Chemical loss ratios indicated differences in the transport characteristics of the three subwatersheds. The downstream subwatershed, which has steeper terrain, a more-developed natural drainage system, and fewer tiles than the two upland subwatersheds, had the largest loss rates for all three chemicals 206 percent for nitrate as nitrogen (October 1992-September 1993) and 20 percent for atrazine and 2.9 percent for metolachlor (April-September 1993). For May-July 1993, when most of the herbicides were transported, the downstream subwatershed also had the largest cumulative unit discharge and the largest streamflow-to-precipitation ratios.