Submitted to: Hydrological Processes
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: June 1, 2011
Publication Date: September 1, 2011
Citation: Guber, A.K., Pachepsky, Y.A., Yakirevich, A., Shelton, D.R., Sadeghi, A.M., Goodrich, D.C., Unkrich, C.L. 2011. Uncertainty in modeling of fecal coliform overland transport associated with manure application in Maryland. Hydrological Processes. 25:2393-2404.
Interpretive Summary: Field scale models are required to evaluate the possibility of manure-borne pathogenic and indicator microorganisms reaching surface waters with runoff. Risk-informed management decisions based on modeling results need to include the uncertainty in model predictions. We augmented the USDA-ARS hydraulic model KINEROS2 with the add-on STWIR to develop the first mechanistic field-scale model describing the release and transport of manure-borne microorganisms. The model was successfully calibrated and tested with field experimental data on manure-borne fecal coliform bacteria (FC) obtained from the USDA-ASRS OPE3 experimental watershed. Furthermore, the fluctuations in measured bacterial concentrations across the manured field were used to evaluate the impact of variability on uncertainty in model predictions. Variations in manure-borne FC concentrations were more than four orders of magnitude; this variability was the major source of uncertainty in predicting FC in runoff. Using the average FC concentrations found in manure led to a gross overestimation of the total mass of FC transported to the edge of the field. The results of this work are important for the design and implementation of water conservation practices in that defining an adequate margin of safety should be based on the simulated uncertainty in manure-borne microorganism transport with runoff.
Assessment of the effect of manure application on bacterial concentrations in runoff at the edge of fields requires developing models of the overland microbial release and transport. To provide information for risk-informed decisions, uncertainty in model predictions caused by the variability of bacterial concentrations has to be understood and quantified. The objectives of this work were (a) to develop and calibrate a field-scale model for simulating bacteria overland transport in runoff water, and (b) to assess the uncertainty in model prediction caused by limited information on spatial distribution of bacteria in surface-applied manure. The bacteria transport add-on module STWIR was developed for the event-based runoff and erosion model (KINEROS2) to simulate the convective-dispersive overland transport, bacteria release from manure, reversible attachment-detachment to soil, and surface straining. The model was successfully calibrated with the experimental data on fecal coliforms in runoff and runoff volumes from manured fields at the USDA-ARS Beltsville experimental OPE3 watershed site. The model was first applied in Monte Carlo simulations that mimicked the random spatial distribution of fecal coliforms in applied manure based on the observed - up to four orders of magnitude - variability of manure-borne fecal coliforms. The second series of Monte Carlo simulations was performed to assess the effect of the number of manure samples on the simulation results. Simulation with the mean initial fecal concentration across the field grossly overestimated fecal coliform content in the runoff whereas simulations with median or geometrical mean initial concentrations substantially underestimated fecal coliform content in runoff. The simulations showed that the spatial variation of bacteria concentrations in manure and limited information on spatial distribution of bacteria in surface-applied manure were major sources of uncertainty in predictions of bacteria transport.