Evaluating the effect of field manure application on bacterial concentrations in streams requires developing microbial transport models. Testing such models against bacterial monitoring data requires estimating the uncertainty in the data caused by spatial and temporal variability of bacterial concentrations. The objective of this work was to evaluate effects of sampling frequency on estimated concentrations of fecal coliforms (FC) in soil, runoff, and stream water after a spring manure application. Experiments are being carried at the OPE3 ARS research watershed in Beltsville, MD. Aged, stacked dairy manure was broadcast at a four-hectare field equipped with runoff flumes that separated from a perennial first order stream by a riparian zone. Soil at this site is mostly a mixture of sandy loam and loamy sand; extensively being sampled at four locations showing differences in productivity that mostly explained the presence of a network of shallow groundwater pathways. Fecal coliforms in runoff and in stream water were monitored using refrigerated samplers and grab sampling. The FC contents in manure varied within six orders of magnitude, and the distribution was highly asymmetrical. Before manure application, the maximum FC content in soil was found in the least hydrologically active parts of the field. A substantial fecal coliforms increase in soil was observed during the three-week period including sites free of manure application. Microponds created favorable conditions for the FC growth before the field was plowed. A two-orders-of-magnitude FC increase in runoff was observed within a week. Significant diurnal and day-by-day dynamics of FC concentrations in the stream was observed. Because of high spatial and temporal variations, the sampling strategy to define fate and transport of manure-borne fecal coliforms can substantially affect results and may be site-specific.