To reduce the uncertainty associated with pesticide fate model predictions on the large scale, a rapid method is needed that can generate sorption coefficients (Kd values) with sufficient spatial detail. The feasibility of near-infrared spectroscopy (NIRS) to act as such a method was examined, using weak-acidic (2,4-D), weak-basic (atrazine) and zwitterion (glyphosate) herbicides and the natural steroid estrogen (17β-estradiol). A total of 609 horizons in 140 soil profiles were collected in agricultural fields near Brandon, Manitoba and near Saskatoon, Saskatchewan. In both agricultural fields, Kd values in horizons generally increased in the order of 2,4-D < atrazine < 17β-estradiol < glyphosate. Soil organic carbon content (SOC) followed by the soil pH were the major factors controlling the sorption of 2,4-D, atrazine and 17β-estradiol but glyphosate showed very strong sorption to soil particles regardless of measured SOC and soil pH values. For the chemicals studied, Kd values decreased from A to C horizons regardless of the segment of the slope from which the soil samples were collected, with the exception of glyphosate that showed relatively large Kd values in B-horizons illuviated with clay. Both the Zeiss Corona and the Foss 6500 spectrophotometers produced significantly strong predictive models for soil properties and Kd values of 2,4-D, atrazine and 17β-estradiol. However, models for glyphosate Kd values were weak or not significant. Using a test set approach and either soil spectral or soil properties data as independent variables, partial least squares regressions were successfully developed to estimate Kd values for use in the Pesticide Root Zone Model (PRZM) to calculate the herbicide mass leached. The study concluded that the added benefit of NIRS will be most useful if the pesticides under study have small sorption potentials and short half-lives in soil. Regional approaches to predicting Kd values from NIRS spectral data can also be developed if the calibration model is derived by combining a set of fields where each has a similar statistical population characteristic in Kd values. / February 2016
Identifer | oai:union.ndltd.org:MANITOBA/oai:mspace.lib.umanitoba.ca:1993/31014 |
Date | January 2014 |
Creators | Singh, Baljeet |
Contributors | Farenhorst, Annemieke (Soil Science), Malley, Diane (Soil Science) Zvomuya, Francis (Soil Science) Wang, Feiyue (Environment and Geography) Kookana, Rai (Soil Science, The Commonwealth Scientific and Industrial Research Organization, CSIRO) |
Publisher | Journal of Agricultural and Food Chemistry, Geoderma |
Source Sets | University of Manitoba Canada |
Detected Language | English |
Page generated in 0.0018 seconds