Herbicide adsorption and organic carbon contents on adjacent low-input versus conventional farms

Little quantitative information is available concerning the long-term impact of alternative framing practices on herbicide adsorption and soil factors that control adsorption. The objective of this study was to evaluate the effects of long-term conventional vs. long-term, low-input farming management practices on herbicide adsorption and soil properties that control adsorption at different landscape positions. Soil samples were collected at depths of 0 to 15 cm along a 350-m boundary between adjacent long-term conventional and low-input farms near Spokane, WA. These samples were characterized for soil organic C (OC) and clay contents, soil pH, and linear adsorption partition coefficients (K(d)) using the ¹⁴C radiolabeled herbicides metribuzin [4,amino-6-11,1-dimethylethyl)-3-(methylthio)-1,24- triazine-5(4,1,1)-one], diuron, [3-(3,4-dichlorophenyl)-1,1-dimethylurea], 2,4-D [(2,4-dichlorophenoxy)acetic acid], and triallate [S-(2,3,3- trichloroally)diisopropylthiocarbamate]. Carbon contents were significantly higher on the low-input farm than the conventional farm at all slope positions, with values at the bottom slope position of 19.4 g kg^-1 on the low-input farm and 12.2 g kg^-1 on the conventional farm. Carbon contents also decreased significantly at upper slope positions, with values at the top slope of 9.0 g kg^-1 on the low-input farm and 6.0 g kg^-1 on the conventional farm. The k(d) values for the four herbicides studied were significantly larger on the low-input farm than the conventional farm at all slope positions. Significant decreases in adsorption occurred on both farms at top-slope positions relative to bottom-slope positions. From 80 to 95% of the increase in K(d) values could be explained by increases in OC contents, but the best-fitting regression line had a large negative intercept for metribuzin, 2,4-D, and diuron. We conclude from this that although the regression model may be used to estimate K(d) values from measured OC contents at moderate to high values of OC, the model has no physicochemical significance and the slope of the model does not equate to a value for K(oc).