Testing statistical methods to predict pesticide drift deposition

Abstract

<p><span>We studied the case of the Andean </span><span>region in Colombia as example of non-mechanized small farming systems in which farmers </span><span>use handheld sprayers to spray pesticides. This is the most common </span>technique to spray <span>pesticide in developing countries. To better understand the spatial distribution of</span> airborne pesticide drift deposits<span> on the soil surface using that spray technique, nine different spatial interpolation </span><span>methods were tested using a surrogate tracer substance (Uranine) i.e. classical approaches </span><span>like the linear interpolation and kriging, and some advanced methods like spatial vine </span><span>copulas, the Karhunen-Loève expansion of the underlying random field, the integrated </span><span>nested Laplace approximation and the Empirical Bayesian Kriging used in ArcMap (GIS). </span><span>This study contributes to</span><span> future </span><span>studies on mass balance and risk assessment related to </span>environmental <span>drift pollution in developing </span><span>countries.</span></p>