Estimating sample requirements for field evaluations of pesticide leaching

Abstract

A method is presented for estimating the number of samples needed to evaluate pesticide leaching threats to ground water at a desired level of precision. Sample size projections are based on desired precision (exhibited as relative tolerable error), level of confidence (90 or 95%) and estimates of variability (expressed as the coefficient of variation) for selected pesticide and soil characteristics. Summary tables of descriptive statistics are provided as guides for projecting variability of various soil characteristics, pesticide properties (sorption, degradation), pesticide concentration profiles and inorganic solutes. These data were compiled through a comprehensive search of review articles and of reports of laboratory and field studies. The parameters selected are measurements generally required for conducting field pesticide leaching studies. They are also useful for evaluating ground water contamination by calibrating and conducting sensitivity testing with solute transport models. Coefficients of variation for selected soil characteristics were as follows: bulk density, 2 to 17%; organic matter, 42 to 125%; porosity, 4 to 18%; particle size distribution, 3 to 55%; hydrogen ion activity, 2 to 15%; 0.3 bar water content, 18 to 82%; 15 bar water content, 18 to 87%; saturated hydraulic conductivity, 48 to 320%; and infiltration rate, 40 to 97%. Coefficients of variation for selected pesticide characteristics were as follows: transformation rates, 7 to 202%; sorption coefficients, 34 to 56%; and pesticide concentration profiles, 40 to 450%. Variability of pesticide concentrations increased over time after the initial pesticide application. The coefficients of variation for inorganic solutes ranged from 19 to 127%. On the basis of these data, sample size estimates (for constant desired precision and confidence level) for characterizing selected soil characteristics generally are smaller than those required for estimating pesticide concentrations. Fewer samples are projected for estimating pesticide concentrations shortly after application, with progressive increases required in later sampling.