ESTIMATING HERBICIDE PARTITION-COEFFICIENTS FROM ELECTROMAGNETIC INDUCTION MEASUREMENTS

A potential method for reducing pesticide leaching is to base application rates on the leaching potential of a specific chemical and soil combination. However, leaching is determined in part by the partitioning of the chemical between the soil and soil solution, which varies across a field. Standard methods of measuring the pesticide-soil partitioning coefficient (K-d) are too expensive and slow for routine held mapping. Therefore, alternative methods for mapping K-d must be found if variable application methods are to be successful. We investigated the use of noncontacting electromagnetic induction measurements as surrogate measures of K-d. We measured the partition coefficient for atrazine (2-chloro-4-ethylamino-6-isopropylamino-s-triazine), apparent electrical conductivity by electromagnetic induction (E(m)), and mass fraction of soil organic carbon (f(oc)) on a 250 by 250 m grid with a 25 m spacing. Both K-d and f(oc) were lognormally distributed, while E(m) was poorly described by either a normal or lognormal distribution. Maps of the measured parameters showed similar spatial patterns, having low values on well-drained soils and high values on poorly drained soils. Correlation coefficients between K-d and E(m) and K-d and f(oc) were 0.575 and 0.685, and showed distinct spatial patterns. Spatial structure as indicated by correlograms indicated that each parameter was spatially dependent to distances of about 80 m. Simple relationships of K-d = 176 f(oc) and K-d = exp(0.0336 E(m)) were found between the data. Maps of K-d estimated from f(oc) or E(m) were similar to measured K-d, but more diffuse. Electromagnetic induction measurements failed to predict the observed high K-d values. The advantage of using E(m) measurements to map K-d is that it is a rapid, easy, and inexpensive method once it has been calibrated.