Click on “Download PDF” for the PDF version or on the title for the HTML version.

If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

Uncertainty and Sensitivity Analyses of Visual PRZM-3 Using Monte Carlo Simulation Technique

Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan

Citation:  Paper number  032108,  2003 ASAE Annual Meeting . (doi: 10.13031/2013.14948) @2003
Authors:   Nitin Khandelwal, Hangsheng Lin, James M. Hamlett, Rick L. Day
Keywords:   PRZM-3, Uncertainty, Sensitivity, Monte Carlo Simulation, herbicide leaching

Uncertainty in chemical movement predictions of the Pesticide Root Zone Model-3 due to variability in model parameter estimation for atrazine, metolachlor, and acetochlor applications in Hagerstown clay, Cookport sandy clay loam, Tilsit silt loam, and Washington clay loam soils are identified through a stochastic sensitivity analysis. Due to model complexity and system variability, a stochastic sensitivity analysis based on Monte Carlo simulation in PRZM-3 is chosen. Monte Carlo simulation provides a criterion by which to judge the uncertainties in model predictions due to errors in parameter estimation when the system variability is expressed in probabilistic terms. A stepwise regression approach with P<0.05 yielded organic carbon, pesticide partition coefficient and decay rates of the surface horizons to be the most sensitive parameters for the maximum concentration (Cmax) and annual flux of the herbicides at 1.57 m depth in the soils for 18,000 Monte Carlo simulations. The decay rate of the surface horizon had a negative effect on atrazine Cmax (i.e. a standard deviation increase in decay rate would reduce Cmax by a standard deviation change of 0.4583) in Hagerstown clay soil. The day of occurrence of maximum concentration (Dmax) was more sensitive to the time dependent parameters, which were kept constant in this study, and did not show much sensitivity to the parameters considered in the study. PRZM-3 model outputs showed more sensitivity to soil parameters (top 5 soil parameters ranked among top 10) than to chemical parameters (top 5 chemical parameters ranked among top 15) than to crop/environmental parameters (top 5 crop/environmental parameters ranked among top 35) for the specific combinations of soils and herbicides studied.

(Download PDF)    (Export to EndNotes)