Feasibility of Near Infrared Spectroscopy (NIRS) in aiding the understanding of the spatial variability of herbicide and estrogen transport.

Baljeet Singh¹, Annemieke Farenhorst¹ and Diane F. Malley²

¹Department of Soil Science, University of Manitoba, Winnipeg, MB, Canada R3T 2N2

²PDK Projects Inc., Nanaimo, British Columbia, Canada V9V 1L6

Corresponding author: umsingh4@cc.umanitoba.ca

It has been well documented that the outcome of pesticide transport models are uncertain due to lack of sufficient spatial data on sensitive input parameters such as some soil properties and pesticide sorption parameters in the complex soil environment. Near Infrared Spectroscopy (NIRS) has been successfully used in agricultural sciences and industries for several decades for the measurement of many diverse applications. However, its application to spatial assessments of sensitive input parameters of pesticide transport models is just emerging. Within this context, the present study examines the use of NIRS at the field level as a method to increase the availability of spatial data on soil properties and pesticide sorption coefficients. Effective application and integration of NIRS with pesticide fate models will help strengthen the versatility of pesticide fate models when they are used to increase the efficiency of farm pest control management practices while minimizing the risk of pesticide off-site movement. The study involved the collection of samples from 140 soil profiles in two agricultural fields located in western Canada.  In total, 609 soil samples from various horizons to a maximum depth of 100 cm were collected. NIRS Spectral data was obtained by collecting surface reflectance from 25 g of air dried and sieved (<2 mm) soil samples in a glass petri dish.  Scanning was conducted using a 45VIS Corona at a wavelength range of 380 to 1690 nm, and a Foss 6500 spectrophotometer fitted with the Rapid Content Sampler at a wavelength range of 1100 to 2500nm. Calibration data included the quantification of a range of soil characteristics (soil organic carbon content, texture, pH), as well as sorption coefficients of three herbicides (2,4-D, atrazine, glyphosate) and one estrogen (17 β-estradiol). The overall results for the Manitoba and Saskatchewan soil samples were particularly promising because they demonstrate that NIRS can be used to strengthen the availability of spatial data on the input parameters such as soil organic carbon content and Kd values at the field scale.

Key words: Near Infrared Spectroscopy, Spatial variability, herbicide, estrogen, soil, sorption, pesticide root zone model.