Estimating Bacterial Loadings to Surface Waters from Agricultural Watersheds

Abstract

Fecal bacteria and pathogens are a major source of surface water impairment. In Virginia alone, approximately 73% of impaired waters are impaired due to fecal coliforms (FC). Because bacteria are a significant cause of water body impairment and existing bacterial models are predominantly based upon laboratory-derived information, bacterial models are needed that describe bacterial die-off and transport processes under field conditions. Before these bacterial models can be developed, more field-derived information is needed regarding bacterial survival and transport. The objectives of this research were to evaluate bacterial survival under field conditions and to develop a comprehensive, spatially variable (distributed) bacterial model that requires little or no calibration. Three field studies were conducted to determine die-off or diminution (settling plus die-off) rates of FC and Escherichia coli (EC) over time in: 1) dairy manure storage ponds and turkey litter storage sheds, 2) pasture and cropland soils to which dairy manure was applied, and 3) beef and dairy fecal deposits. The dairy manure storage ponds were sampled just under the pond surface. The FC and EC diminution (settling plus die-off) rates for dairy manure storage ponds were 0.00478 day⁻¹ and 0.00781 day⁻¹, respectively. The five samples collected for turkey litter in storage were inadequate to draw any conclusions. Bacterial die-off rates in cropland and pastureland soils were found to be statistically different from each other at the α = 0.05 level. The FC and EC die-off rates in cropland soils were 0.01351 day⁻¹ and 0.01734 day⁻¹, respectively, while the FC and EC die-off rates in pastureland soils were 0.02246 day⁻¹ and 0.02796 day⁻¹, respectively. Die-off rates for bacteria from dairy heifer, dairy milker, and beef cow fecal deposits were not statistically different from each other. The resulting die-off rate constants for fecal deposits were 0.01365 day⁻¹ and 0.01985 day⁻¹ for FC and EC, respectively. The EC/FC ratio was also evaluated for the fecal deposits and land-applied manure to determine if a quantifiable relationship was discernable. In general the EC/FC ratio declined over time, but no quantifiable relationship was discerned.

The bacterial model simulates die-off, bacterial partitioning between soil and water, and bacterial transport to surface waters in free (in solution) and sediment-adsorbed forms. Bacterial die-off was modeled using Chick's Law, bacterial partitioning was modeled with a linear isotherm equation, and bacterial transport was modeled using continuity and flow equations. The bacterial model was incorporated into the ANSWERS-2000 model, a continuous, distributed, nonpoint source pollution model. The model was tested using data from two plot studies. Calibration was required to improve runoff and sediment predictions. Bacterial model predictions underpredicted bacterial concentrations in runoff with a maximum underprediction error of 92.9%, but predictions were within an order of magnitude in all cases. Further model evaluation, on a larger watershed with predominantly overland flow, over a longer time period, is recommended, but such data were not available at the time of this assessment. The overall conclusions of this research were 1) FC and EC die-off or diminution under the examined field conditions followed Chick's Law, 2) measured die-off rate constants in the field were much less than those cited in literature for laboratory experiments, and 3) for the conditions simulated for two plot studies, the bacterial model predicted bacterial concentrations in runoff within an order of magnitude.