Lake Erie Center

Fate and Transport Studies

Role of retention time and soil depth on the survival transport of Escherichia coli and Enterococcus ssp. in biosolid-amended agricultural soil                     

Danielle M. Long

Biosolids that are added as fertilizer to agricultural fields might contain dangerous levels of Escherichia coli and Enterococcus spp.—indicators of pathogen contamination. This research assessed the effects of soil depth, soil type, and retention time on the reduction of those indicators (and presumably pathogens) in effluent from biosolid-amended soil. Biosolids, amended with E. coli (1011 – 1012 CFU/ 100 ml) and Enterococcus spp. (1010 – 1011 CFU/ 100 ml), were leached through silt loam and sandy loam soils in vertical, flow-through microcosms (5 cm in width and 30, 45, or 60 cm in depth), and bacteria in effluent     were enumerated. Peak densities and total numbers of bacteria transporting through 60 cm of soil were significantly less than through the shorter depths in both soil types. However, the densities of bacteria in effluent from silt loam were significantly less than those from sandy loam, regardless of depth. Despite measurable decreases in densities of bacteria in effluent, values remained above target levels established by the U.S. EPA (576 CFU/100 mL for E. coli and 151 CFU/100 mL for Enterococcus spp.).  

To assess the effect of retention time on bacterial transport and survival, vertical microcosms of the aforementioned lengths were capped at the bottom to prevent leaching. After one or two weeks, the caps were removed and bacteria in resulting effluent were enumerated. The effect of capping under field conditions was also evaluated by creating capped (for two weeks) and uncapped silt loam microcosms that mimicked tiled, agricultural fields (2.4 m in length, 1.2 m in width, and 0.45 m in depth). The retention of effluent in both the silt loam vertical flow-through microcosms and field microcosms for two             weeks yielded peak densities of bacteria below target levels. Effluent from sandy loam soil did not reach target densities.                     

mesocosm microcosms danielle

To expand the applicability of this study to untested soil depths and types, an advection-dispersion model was developed using data obtained from vertical microcosms. Model-predicted densities of bacteria were compared to the observed densities from vertical, flow-through microcosms using a linear regression; the range of R2 values was 0.82–0.97. When the model was applied to the data from field microcosms the R2 values were 0.7 and 0.48 for E. coli and Enterococcus spp., respectively.

Both filtration by soil and capping effectively reduce the densities of indicator bacteria transporting through a soil column, and could be relied upon to achieve target densities in effluent. Agricultural fields with soil depths that are insufficient for filtering bacteria (e.g. < 80 cm of silt loam) could be capped to further reduce densities and reach targets.                     



Predicting Fate and Transport of Fecal Bacteria through Soils Using an Advection-Dispersion Model                     

Ryan Jackwood

Bovine liquid manure (BLM) obtained from concentrated animal feeding operations (CAFOs) often is added to agricultural fields as fertilizer. Unfortunately, pathogens inherent to BLM might contaminate surface and ground waters via field runoff and expedited  transport via field drain tiles. Buffer strips, placed at the edge of agricultural fields, can intercept pathogens that infiltrate into soil. Buffer strip efficiency to intercept pathogens is dependent on soil type and microbial attributes. Parameters for each have been included in models (e.g. the advection-dispersion model) that describe distance-dependent densities of microorganisms being transported with saturated water flow:       

           C(t)/Co = Vo/(Aq(4paL*x)(0.5) exp[-( x-x’ )2/(4αL*x’ )-kpx’]

In this model, C(t)/Co represents the ratio of bacteria that transport a given distance (x) and time (t) through soil from the application source. I hypothesized that (1) an advection-dispersion model could predict distance-dependent densities of fecal indicator bacteria (i.e. Escherichia coli and Enterococcus hirae) commonly present in BLM. (2) Model predictions could be extrapolated to determine the minimum transport distance (depth of proposed buffer strip) required to decrease microbial densities within BLM to the levels suggested by the Ohio Environmental Protection Agency’s Water Quality Standards (WQS), i.e. geometric means of 126 and 35 CFUs/100 mL, respectively. The project’s objectives were to:                                                              

(1)   Formulate a semi-pervious soil that could be used to construct buffer strips using components (sand, clay, and plant material) obtained in northwest Ohio.                     

(2)   Use formulated soil in microcosms of three lengths (30, 45, and 60 cm) to obtain numerical values for the model parameters: fluid velocity, v; dispersivity, αL; initial injection volume, Vo; and single collector efficiency, η.                     

(3)   Obtain values for distance-dependent densities of fecal indicator bacteria in BLM during transport through the microcosms, and determine whether the advection-dispersion model predicted these values by using a correlation analysis.                     

(4)   Use the model to make predictions of distance-dependent densities for bacterial transport through simulated buffer strips and compare the resultant predictions to observed effluent densities from field lysimeters containing formulated soil.                     

(5)   Extrapolate the data to predict transport distances that are needed to sufficiently reduce BLM derived microbes, i.e. buffer strip widths needed to obtain target densities of bacteria.                     

Outcomes for each goal included:

(1)   A soil was formulated by combining sand (60 % by volume), clay (20 % by volume), and dried plant material (20 % by volume). Formulated soil had an infiltration rate of 123 mm/hr and fluid velocity of 0.015 - 0.025 cm/s;                     

(2)   Model parameters had the following values: single collector efficiency (η = 6.16 x 10-2); dispersivity (αL: 30 cm = 2.17 cm; 45 cm = 2.28 cm; 60 cm = 5.41 cm); and initial injection volume (Vo = 0.014 – 13 mL).                     

(3)   The maximum ratio (C/Co) of indicator organisms in effluent from the microcosms decreased with soil depth (E.coli – 30 cm: 0.11, 45 cm: 0.061, 60 cm: 0.0265, E. hirae – 30 cm: 0.11, 45 cm: 0.032, 60 cm: 0.016). The precision of model predictions decreased with system size (R2: E. coli – 30 cm: 0.978, 45 cm: 0.987, 60 cm: 0.741; E. hirae, 30 cm: 0.852, 45 cm: 0.737, 60 cm: 0.638).                     

(4)   Results from a correlation analysis of model predictions and field lysimeter data yielded a difference in model precision between indicator species (E. coli: R2 = 0.4828; E. hirae: R2 = 0.020).                     

 (5)   Lysimeter data were used to predict minimum soil distances needed to decrease bacterial densities to WQS; E. coli and E. hirae required 254 cm and 282 cm of soil distance respectively.                                                              

A buffer strip that contains soil formulated to optimize infiltration rate and provides a minimum of 254 cm of subsurface flow will reduce the threat of contamination from BLM-derived pathogens to surface waters.                     

Last Updated: 3/23/15