Persistence and Leaching Potential of Microorganisms and Mineral N in Animal Manure Applied to Intact Soil Columns

Abstract

Pathogens may reach agricultural soils through application of animal manure and thereby pose a risk of contaminating crops as well as surface and groundwater. Treatment and handling of manure for improved nutrient and odor management may also influence the amount and fate of manure-borne pathogens in the soil. A study was conducted to investigate the leaching potentials of a phage (Salmonella enterica serovar Typhimurium bacteriophage 28B) and two bacteria, Escherichia coli and Enterococcus species, in a liquid fraction of raw pig slurry obtained by solid-liquid separation of this slurry and in this liquid fraction after ozonation, when applied to intact soil columns by subsurface injection. We also compared leaching potentials of surface-applied and subsurface-injected raw slurry. The columns were exposed to irrigation events (3.5-h period at 10 mm h⁻¹) after 1, 2, 3, and 4 weeks of incubation with collection of leachate. By the end of incubation, the distribution and survival of microorganisms in the soil of each treatment and in nonirrigated columns with injected raw slurry or liquid fraction were determined. E. coli in the leachates was quantified by both plate counts and quantitative PCR (qPCR) to assess the proportions of culturable and nonculturable (viable and nonviable) cells. Solid-liquid separation of slurry increased the redistribution in soil of contaminants in the liquid fraction compared to raw slurry, and the percent recovery of E. coli and Enterococcus species was higher for the liquid fraction than for raw slurry after the four leaching events. The liquid fraction also resulted in more leaching of all contaminants except Enterococcus species than did raw slurry. Ozonation reduced E. coli leaching only. Injection enhanced the leaching potential of the microorganisms investigated compared to surface application, probably because of a better survival with subsurface injection and a shorter leaching path.

INTRODUCTION

Animal manure is widely applied to agricultural soil as a source of nutrients and organic matter. Inappropriate use of animal manure can lead to nitrate pollution of groundwater (1) and eutrophication of surface waters (2), but manure also may introduce pathogenic bacteria and viruses to the soil environment (3, 4, 5). Unless properly regulated, contamination of drinking water, bathing facilities, and fresh produce of leafy and root crops by manure-borne pathogens can cause diseases to humans and wild life (6, 7, 8).

After field application, manure-borne microorganisms can survive for 2 to 3 months at 5 to 25°C (9). Microorganisms can be physically strained in narrow soil pore spaces or water films, or they can attach chemically to soil and immobile slurry particles (10, 11). Gannon et al. (12) found that cell size was the main factor controlling transport of microorganisms in repacked soil. On the other hand, in structured soil this filtering effect can be severely reduced by preferential flow and macropore flow (13, 14), and microorganisms may move with runoff or infiltrating water as free cells and/or attached to soil and manure particles (15, 16, 17). While the environmental fate of manure-borne contaminants has received attention in the past (5, 11), there are recent developments in manure management techniques for improved nutrient use efficiency and odor control, including solid-liquid separation (18, 19) and field application methods, that may alter the environmental fate of some contaminants (20, 21).

In Europe, more than 65% of the manure is managed in liquid form as a slurry (22). Slurry is usually applied to agricultural fields by surface application or, increasingly, by subsurface injection at 6- to 10-cm soil depth to reduce nuisance odor and NH₃ volatilization from the applied slurry (23, 24, 25). These two application methods may represent different risks for leaching of nutrients and microorganisms as a result of the difference in slurry-soil contact (21, 26, 27).

Slurry dry matter (SDM) content affects the redistribution of slurry liquid in the soil after field application (28). Solid-liquid separation techniques typically remove 40 to 60% SDM from raw slurry (29), and this will enhance the infiltration of dissolved and suspended slurry constituents. Soluble and suspended slurry particles (>20% of total SDM) usually remain in the liquid fraction after separation (19), and organic constituents may facilitate transport of contaminants in soils (15, 16, 17, 30). Chemical treatment of slurry may also have an effect on the survival of microorganisms (31) and on the size distribution of slurry particles. Investigating the effect of slurry pretreatment on the leaching potential of manure-borne contaminants is, therefore, important (32).

We quantified the leaching potential (proportion of contaminant leached from land-applied pig slurry after four irrigation events) of Salmonella enterica serovar Typhimurium bacteriophage 28B (phage) as a model organism for viruses and of Escherichia coli and Enterococcus species as model organisms for pathogenic bacteria. The accumulation and leaching of mineral N were also monitored as a measure of net N mineralization and nitrification activity. Leaching experiments were conducted that involved three slurry materials and two slurry application methods. We hypothesized that (i) solid-liquid separation increases the leaching potential of the contaminants in the liquid fraction compared to the raw slurry due to a higher potential exposure to percolating water; (ii) ozonation of the liquid fraction will decrease the leaching potential of all pathogens due to lower survival; and (iii) direct injection of slurry will increase the leaching potential of pathogens compared to surface application due to better survival in the injected slurry.

MATERIALS AND METHODS

Soils.

Intact soil columns of a loamy sand were sampled from a crop rotation with spring barley-winter wheat-spring barley at Foulum Experimental Station (56° 29′ N, 9° 34′ E), Denmark. The plot had not received any animal manure in the previous 2 years. Sampling was done using stainless steel cylinders (length, 20 cm; diameter, 20 cm) as described previously (33). The soil columns were slowly saturated and then drained to a soil water potential of −100 hPa, i.e., close to field capacity for this soil. The soil columns then were sealed and stored at 2°C until used in the experiment. Selected soil characteristics, determined by standard laboratory methods (34), are presented in Table 1.

Table 1.

Selected soil characteristics

Characteristica	Result by soil depth
	10 cm	30 cm
OC (%)	2.0	1.8
Porosity (%)	41	43
ρd (g cm−3)	1.53	1.49
Clay (%)	8	8
Silt (%)	13	14
Sand (%)	79	78
Ksat (mm h−1)	61.1	43.2
pH	6.33	6.33
EC (mS cm−1)	0.047	0.047

^aOC, organic carbon; ρ_d, bulk density; K_sat, hydraulic conductivity at saturation.

Slurries.

Raw slurry (RS) and the liquid fraction of mechanically separated slurry (LS) were collected at a pig farm near Åbøl, Denmark. A 10-liter portion of LS was ozonated at Research Centre Foulum, Denmark, by supplementing ozone at 0.125 liters min⁻¹ until the redox potential reached zero. Slurry samples were stored in blue-cap bottles at 2°C until applied to columns. Selected physicochemical properties of RS, LS, and the ozonated liquid fraction (OLS) are presented in Table 2.

Table 2.

Selected physicochemical properties^a of different slurries

Characteristic	Result by sample type
	RSb	LS	OLS
Density (g cm−1)	1.02jc	1.01j	1.00j
Total solids (%)	5.65j	3.14k	3.09k
Volatile solids (%)	4.17j	1.97k	1.90k
pH	7.10j	7.56k	7.86l
EC (mS cm−1)	19.3j	15.8k	17.9l
NH4-N (g kg−1)	3.01j	2.89j	3.03j
Total N (g kg−1)	4.5j	4.3k	4.2k
E. coli (CFU ml−1)	9.3 × 104j	2.6 × 104k	1.4 × 104l
Phage (PFU ml−1)	1.8 × 106j	1.8 × 106j	1.8 × 106j
Enterococcus sp. (CFU ml−1)	2.9 × 104j	3.2 × 104j	2.7 × 104j

^aSlurries were analyzed prior to application to soil columns.

^bRS, raw slurry; LS, liquid slurry fraction; OLS, ozonated liquid slurry fraction.

^cDifferent letters (j, k, and l) in results indicate significant difference at a level of 0.05.

Prior to application, all slurries were spiked with phage (1.5 × 10⁶ PFU ml⁻¹) as a model organism for pathogenic virus and with 2,6-difluorobenzoic acid (FBA; CAS RN 385-00-2; Sigma-Aldrich, Germany) (2 g liters⁻¹) as a nonreactive tracer. The toxicity of FBA on selected microorganisms was tested before starting the experiment, and no significant effect was found at the concentration used, in accordance with results reported by McCarthy et al. (35).

Experimental design.

All glassware and devices used were sterilized. The experiment was conducted at 10°C to simulate the field temperature usually observed in northern Europe during spring manure application. Soil columns, slurries, and rain water were equilibrated to the experimental temperature before use.

For the leaching experiment, hexaplicate columns were amended with RS by simulated surface application and subsurface injection, respectively. The treated slurry materials LS and OLS were also added to hexaplicate columns, but by subsurface injection only due to a limited capacity for incubation of soil columns in the experimental setup. The slurries were applied at a rate of 50 t ha⁻¹. Injected slurry was placed in a slit with dimensions of 10 (length) by 4 (width) by 9 (depth) cm³ created at the center of the column surface (Fig. 1). For surface application, slurry was applied in a band created by removing the top 2 cm of soil from a circular area of 17-cm diameter at the center of the column surface. Soil removed from a column was subsequently used to cover the slit/band loosely following slurry application.

Fig 1.

Sectioning diagram of soil column (S1, slurry slit; S2 and S3, surrounding soil; and S4, bottom section of the column).

As controls, triplicate columns without irrigation were prepared by subsurface injection of RS and LS in order to examine the redistribution and fate of microorganisms and mineral N as affected by the differences in SDM. Also, triplicate soil columns without slurry amendment but with irrigation were included.

For all except nonirrigated samples there were four separate irrigation events (IE) to simulate rainfall, which occurred 1, 2, 3, and 4 weeks after slurry application. The initial 1-week delay between slurry application and IE1 was chosen to simulate conditions in the field where slurry is typically applied during a dry spell in spring. Artificial rainwater (0.1 mM NaCl, 0.01 mM CaCl₂ [2H₂O], and 0.01 mM MgCl₂ [6H₂O] [VWR, Denmark]) at 10 mm h⁻¹ was applied using a rain simulator (33, 36). During these events the soil columns with or without slurry rested on a glass filter disc of 60- to 100-μm pore size and 1.6-cm thickness (ROBU; Glasfilter Geräte GMBH, Germany). The glass filter disc was mounted on top of a stainless steel plate with only a small interstitial space which was water filled during the experiment. At the bottom of the water-filled space, a water-filled hypodermic needle led the leachate to a blue-cap bottle. The water-filled space and the hanging water column in the hypodermic needle exerted a suction of −12.5 hPa on the soil column's lower boundary to allow for leaching under unsaturated soil conditions (37). The adsorption and filtration properties of the below-column setup were tested to ensure that microorganisms leached from the soil columns would reach the blue-cap bottles. The test showed that concentrations of microorganisms in in-flow and outflow of the below-column setup were similar.

Water and soil sampling.

Leachates were collected when percolation had stopped after each IE. A week after the final irrigation event, IE4, the soil columns were extruded slowly using a pressing device and sectioned to isolate the original slurry hot spot (S1) and three other subsamples (S2, S3, and S4) as indicated in Fig. 1. Immediately after collection, samples were prepared and analyzed for the selected contaminants.

Physicochemical analyses.

Electrical conductivity (EC) and pH of both soil samples and leachates were measured with a Radiometer conductivity meter (Copenhagen, Denmark) and Sentron 3001 pH meter (Roden, The Netherlands), respectively. Subsamples were extracted in 1 M KCl. After filtration (GA55; Advantec, Japan), these extracts, as well as leachates, were analyzed for NH₄-N and NO₃-N on an Auto-analyzer III digital colorimeter (Bran & Luebbe, Germany). Turbidity was measured on an HACH 2100 AN turbidimeter equipped with an EPA filter measuring at wavelengths of 400 to 600 nm (Hach, Loreland, CO). Total organic carbon (TOC) in the leachates was analyzed by a total organic carbon analyzer (TOC-V_CPH; Shimadzu, Duisburg, Germany). FBA concentrations in the leachates were measured by a liquid chromatography-tandem mass spectrometry (LC-MS/MS) technique as described by Juhler and Mortensen (38).

Microbial analyses. (i) Phage.

Phage was enumerated by a double-agar layer method (39). The host strain Salmonella enterica serovar Typhimurium Type 5 was grown in nutrient broth at 37°C for 4 h. Approximately 2 g soil was added to 18 ml maximum recovery diluent (MRD; Oxoid, Denmark) and sonicated for 30 s. Soil samples were then 10-fold diluted in MRD, and fresh leachates were also diluted similarly. One ml diluted sample was mixed with 1 ml broth culture of the host strain and 3 ml soft agar (a mixture of 70% blood agar base [Oxoid, Denmark] and 30% nutrient broth [Oxoid, Denmark]). The mixture was spread on a well-dried blood agar base plate and incubated at 37°C for 18 h. Each sample was analyzed in triplicate. Clear zones (plaques) were counted as PFU. The detection limit for phage was 10 PFU g⁻¹ for soil and 1 PFU ml⁻¹ for leachates.

(ii) E. coli plate counts.

Two g of freshly sieved soil was mixed with 18 ml of 0.01 M phosphate buffer in a glass tube followed by sonication for 20 s (40, 41). Ten-fold dilution series were prepared for both soil and leachate samples, and the diluted samples were plated in triplicate on E. coli petri films (3M A/S, Denmark). After incubation at 37°C for 24 h, characteristic blue colonies were counted as E. coli (CFU). The detection limit for E. coli was 10 CFU g⁻¹ for soil and 1 CFU ml⁻¹ for leachates.

(iii) E. coli DNA extraction.

One hundred ml of each leachate sample was filtered using a 0.2-μm-pore-size polycarbonate filter membrane (GE Osmonics Labstore, Minnetonka, MN) under low suction. The filter membrane was cut into small pieces and then mixed with 0.5 ml of cetyl trimethylammonium bromide extraction buffer, 0.5 ml of phenol-chloroform-isoamyl alcohol (25:24:1, pH 8.0), and 250 mg of zirconia/silica beads and vortexed for 30 s. The mixture was centrifuged at 13,000 × g for 10 min, and the aqueous phase then was transferred to an Eppendorf tube. The aqueous phase was separated from phenol by adding an equal volume of chloroform-isoamyl alcohol (24:1) and centrifuging at 13,000 × g for 5 min. The DNA was precipitated by adding cold ammonium acetate and isopropanol and then centrifuging at 13,000 × g for 10 min. The DNA pellet was washed once with ice-cold 70% ethanol and air dried, and then 25 μl DNase-free water was added. The prepared DNA was used immediately or stored at −20°C until use.

The malate dehydrogenase gene (mdH) of E. coli was chosen for quantitative real-time PCR (qPCR). The specificity of this gene was confirmed by qPCR assays, and it was also ensured that false-positive results or cross-contamination was absent. The primers were designed by PRIMER3 (http://frodo.wi.mit.edu/primer3/) with the sequences mdh1 (forward primer; TGCACGTTTTGGTCTGTCTC) and mdh2 (reverse primer; AGAAGAAACGGGCGTACTGA). The primers were synthesized by DNA Technology Company A/S (Aarhus, Denmark). The qPCR assays were carried out in an Mx3005P thermocycler (Strategene, Denmark) using mdh primers. The PCR mixtures (25 μl) contained 5 μl DNA, 12.5 μl of 2× PCR master mix (Promega, Denmark), 400 nM each primer, and 50,000× diluted SYBR green (Invitrogen, Denmark). The qPCR conditions consist of an initial heat-denaturing step at 94°C for 5 min, followed by 45 cycles of 94°C for 15 s, annealing at 56°C for 20 s, and extension at 72°C for 20 s, followed by an elongation step at 72°C for 3 min. In every qPCR assay, the E. coli standard curve (R² = 0.98) was included in duplicate for absolute quantification. Furthermore, a negative control (5 μl of water) and a positive DNA control (5 μl) of E. coli DNA (2 ng μl⁻¹) were included. The limit of detection for quantitative real-time PCR was 10 CFU ml⁻¹.

(iv) Enterococcus species.

Both soil and leachate samples were diluted in MRD as described for phage dilution. One ml diluted sample was spread on Slanetz and Bartley medium (Oxoid, Denmark). The number of Enterococcus species was determined as typical red-maroon colonies on the Slanetz and Bartley medium following incubation at 44°C for 48 ± 4 h (42). The detection limit for Enterococcus species was 10 CFU g⁻¹ for soil and 1 CFU ml⁻¹ for leachates.

Statistical analysis.

The statistical software R was used for statistical analyses (43). Analysis of variance (ANOVA) of the data was carried out at the 95% confidence level to evaluate differences in leaching of the contaminants and slurry constituents between different treatments.

RESULTS AND DISCUSSION

Effect of slurry type without irrigation.

Figure 2a shows the distribution of contaminants and physicochemical variables among the four sections of columns 5 weeks after injection of LS or RS but without irrigation. To account for between-sample variability, the within-column distributions of contaminants among the four soil sections are presented as percentages of the total recovery in each column. The percentage of the remaining microorganisms and mineral N in the slurry injection zone, S1, was higher with RS than with LS. Conversely, in sections S2 and S3 these percentages were higher with LS than with RS (Fig. 2a). These patterns suggest that the slurry with pathogens and mineral N distributed further into the soil with LS than with RS. Without irrigation, E. coli and Enterococcus species were recovered only in section S1 after application of RS. The main difference between RS and LS was the lower concentration of total and volatile solids in the latter (Table 2), which probably promoted infiltration of the slurry components away from the injection slit. During solid-liquid separation, larger particles are generally removed first, thus particles remaining in LS would, on average, be finer and more mobile than those of RS (18). A higher proportion of mobile to immobile particles in LS compared to that in RS may increase particle-mediated transport of contaminants. Unc and Goss (11) and Pachepsky et al. (44) suggested such a mechanism for the organic matter-facilitated transport of microorganisms.

Fig 2.

Distribution among different sections of total retaining phage, E. coli, Enterococcus species, and mineral N and the values of soil water content (SWC) and electrical conductivity (EC) of different sections for different slurries of nonirrigated application, irrigated application, and the two application methods together after 5 weeks of slurry injection (RS, raw slurry; LS, liquid slurry fraction; OLS, ozonated LS; in treatment BG [without slurry application], the microorganisms investigated were neither detected in unamended columns nor leached from the soil; S1, slurry slit; S2 to S3, surrounding soil; and S4, bottom section of the column).

The application of slurry has been found to increase the average soil water-holding capacity (45). Petersen et al. (28) observed a partial retention of the liquid phase of pig slurry around the injection slit, and results indicated that this retention will depend upon soil water content (SWC) at the time of application (28). In accordance with this, SWC and electrical conductivity (EC) in S1 were significantly higher than that in the other sections with both slurries (Fig. 2).

The different patterns of slurry redistribution with RS and LS significantly influenced the survival of microorganisms. The recovery of phage was similar for LS and RS, whereas recoveries of E. coli and Enterococcus species were higher with LS (Table 3). The overall lowest recovery (0.7%) among all contaminants was observed for E. coli in RS and the highest recovery (35%), also for E. coli, in LS. Infiltration would bring organisms from the slurry in better contact with the soil environment, and the observed trends in survival thus indicate that E. coli was surviving better in the soil than if retained around a slurry-saturated injection slit. E. coli and Enterococcus species are both facultative anaerobes, and the physical protection against predation obtained when organisms are carried with infiltrating water toward the smaller pores may be an important factor in determining survival (46, 47, 48).

Table 3.

Total leached TOC and FBA and the percent^a leached and retained of total applied mineral N, phage, E. coli, and Enterococcus species for different treatments

Treatment	Result by treatment type
	Nonirrigated columns (subsurface injection)		Irrigated columns
			Surface applied	Subsurface injection
	RS	LS	RS	RS	LS	OLS
FBA (%)			41.5e ± 3.7	44.1ej ± 2.7	43.2j ± 3.4	42.7j ± 1.7
TOC (mg)			173e ± 14	184ej ± 14	216k ± 13	214k ± 16
Mineral N
Leached (%)			14e ± 3	17ej ± 4	22k ± 2	20jk ± 4
Retained (%)	89c ± 5	97c ± 5	63e ± 3	58ej ± 8	59j ± 5	60j ± 6
Recovered (%)	89c ± 5	97c ± 5	77e ± 0.8	75ej ± 5	81j ± 6	80j ± 6
Phage
Leached (%)			4.1e ± 2.3	10.0fj ± 2.4	15.9k ± 8.4	11.4jk ± 3.1
Retained (%)	3.9c ± 1.0	3.1c ± 0.6	4.9e ± 0.7	5.6ej ± 2.7	3.1k ± 1.3	3.3k ± 1.5
Recovered (%)	3.9c ± 1.0	3.1c ± 0.6	9e ± 1.9	15.7fj ± 1.7	19j ± 8.4	14.7j ± 2.4
E. coli
Leachedb (%)			0.07e ± 0.06	0.21fj ± 0.09	0.61k ± 0.34	0.13j ± 0.08
DNA (%)			8.0e ± 5.4	13.5fj ± 3.1	26.3jk ± 17.7	37.5k ± 24.6
Retainedb (%)	0.7c ± 0.4	35d ± 5	1.56e ± 1.13	2.53ej ± 1.89	11.4k ± 9.5	2.68j ± 1.09
Recoveredb (%)	0.7c ± 0.4	35d ± 5	1.63e ± 1.2	2.7ej ± 1.8	12k ± 9.3	2.8j ± 1.1
Enterococcus sp.
Leached (%)			0.024e ± 0.01	0.12fj ± 0.09	0.17j ± 0.17	0.11j ± 0.17
Retained (%)	3.5c ± 0.4	4.1d ± 0.2	4.1e ± 1.2	12.4fj ± 9.6	17.0j ± 11.0	10.5j ± 8.3
Recovered (%)	3.5c ± 0.4	4.1d ± 0.2	4.1e ± 1.2	12.5fj ± 9.5	17.2j ± 10.9	10.6j ± 8.4

^aAverage concentrations of microorganisms in CFU ml⁻¹ or PFU ml⁻¹ given in Table 2 for each slurry type were used to calculate the results in percentages. Different letters in results indicate significant difference at a level of 0.05. Letters c and d are used for RS and LS in nonirrigated columns, e and f are used for application methods with RS, and j, k, and l are for slurry types.

^bBased on culturable E. coli.

The percentages of surviving microorganisms recovered in S2 to S4 followed the order Enterococcus species < E. coli < phage. Movement of microorganisms as the slurry infiltrates into the soil after application will be impeded by physical straining and chemical attachment to the soil matrix (10, 49), so size and shape are likely important for their transport in soil. Enterococcus species cells are spherical and approximately 0.5 to 1 μm in size (50), but the cells are organized in chains; E. coli cells are rod shaped and are 0.7 to 1.5 μm long (44, 51); and phages are circular, with a diameter of 0.03 to 0.07 μm (52). This small size may explain why phages were more mobile in soil than the bacteria. Enterococcus species had the lowest mobility, which could be related to the chain organization. Also, water stress tends to make rod-shaped cells more spherical (53), which could have stimulated leaching of E. coli.

Effects of slurry pretreatment with irrigation.

LS had a higher potential for leaching of mineral N, E. coli, phage, and total organic carbon (TOC) than RS, whereas the leaching potential of Enterococcus species was similar for the two slurry types (Table 3). Leaching of the contaminants is presented as percentages of the total applied in the slurry materials. Ozone treatment did not affect the leaching potential of any contaminant except E. coli. Probably only the survival of E. coli was significantly affected by ozone treatment. Nitrate constituted between 92 and >99% of total mineral N in leachates, with the concentrations ranging from 12 to 65 mg liters⁻¹. Nitrogen equivalent to 13 to 24% of the mineral N applied in slurry leached from the columns during the experiment. The leaching potential of phage (10 to 16%) was higher than that of both culturable E. coli (0.1 to 0.6%) and Enterococcus species (0.1 to 0.2%) with all three slurry types (Table 3). These differences are in accordance with the findings from the nonirrigated columns where phage moved further into the soil after application of slurry, followed by E. coli and Enterococcus species.

In contrast to the plate counts of E. coli, there were no differences in the leaching of E. coli when evaluated from mdh DNA copy numbers using qPCR between LS and RS or between LS and OLS (Table 3). Contrary to the plate count results, the mdh gene copy numbers were below the detection limit in leachates from IE4. A steady decrease in E. coli levels was observed in the leachates throughout the experiment with both plate counting (Fig. 3) and qPCR (data not shown). The concentration of culturable E. coli in the leachate decreased 2-fold between IE1 and IE4, and the concentration of DNA decreased 2-fold. The temporal trends of E. coli leaching were apparently similar for both enumeration techniques. The quantification of cell numbers based on DNA (which included culturable and nonculturable or dead cells) was significantly higher (P < 0.001) than the plating-based CFU (culturable cells) of E. coli in the leachates during IE1 to IE3, which indicates that many nonculturable or dead cells of E. coli leached along with culturable E. coli. This result is in agreement with Pedersen and Jacobsen (54), who found a significant difference between the CFU and DNA levels when investigating the survival of microorganisms in an air-dried soil. It is possible that some of this DNA actually derived from dead cells in manure and soil, but it has been argued that the half-life of DNA in environmental samples is very short because of the presence of nucleases (55, 56, 57).

Fig 3.

Percent leached of total applied mineral N, phage, E. coli, and Enterococcus species, as well as FBA concentration, electrical conductivity (EC), total organic carbon (TOC) concentration, and turbidity of leachates for three subsurface-injected slurry types during four irrigation events over a 4-week period (RS, raw slurry; LS, liquid slurry fraction; OLS, ozonated LS).

The leaching potentials of all contaminants changed significantly with time. The leaching of mineral N increased with all three slurry types (Fig. 3), probably reflecting NO₃-N accumulation via nitrification. In contrast, the leaching of all microorganisms decreased rapidly with time, presumably due to inactivation and filtering in the soil (Fig. 3).

Following IE1 there was a higher leaching of all contaminants with LS than with RS (Fig. 3). The concentrations of the nonreactive tracer, FBA, and TOC and the values of EC and turbidity were higher in the leachate of IE1 with LS than with RS. This indicates that there is, in general, a greater retention of slurry constituents with RS than with LS, probably originating from the lower infiltration of RS into the soil after application (Fig. 2). With both LS and OLS, the FBA leaching peaked during IE1, whereas the tracer peaked during IE2 in columns amended with RS. Total amounts of FBA leached were similar with all slurry types (Table 3). The leaching of TOC was only slightly delayed relative to the leaching of FBA in all treatments (Fig. 3). Dunnivant et al. (58) suggested that leaching of TOC can be delayed with an extended tail due to slow and nonlinear adsorption to soil particles, but the sorption capacity of the sandy loam soil used for this experiment was moderate, as indicated by a cation exchange capacity of 13.1 cmol kg⁻¹ in the low layer of this soil (28). The electrical conductivities (EC) of leachates from LS and OLS treatments were higher than RS during IE2 to IE4. Since the EC of RS was higher than that of LS and OLS (Table 2), it indicates a greater contact between the salts from the liquid slurries and the infiltrating water (Fig. 3).

Both retention and overall recovery of Enterococcus species in leached soil columns was similar for all three slurry materials (Table 3). Phage retention was higher for RS than for LS or OLS; as the overall recoveries for RS, LS, and OLS were similar (Table 3), this was due to reduced leaching of phage with RS. Both retention and recovery of E. coli were higher for LS than for RS because of the higher survival with LS that was also observed in the nonirrigated columns (Table 3). E. coli recovery was lower with OLS than with LS, indicating that viability was reduced by the ozone treatment (Table 3). Recovery of culturable Enterococcus species was similar among slurry treatments, but perhaps survival was affected by ozonation as none leached after IE1 (Fig. 3). With all slurry types, between 80 and 99% of total remaining Enterococcus species was recovered in S1 after four leaching events (Fig. 2b), indicating a very low mobility. It could be due to attachment to slurry solids as proposed by Guber et al. (59) in a study comparing release of Enterococcus species and E. coli from slurry particles. Also, cells organized in chains could have been more exposed to predation if they were filtered out in larger pores during transport through the soil (46, 47).

Total mineral N recoveries in irrigated columns were lower than recoveries in nonirrigated columns (Table 3). In contrast, total numbers of microorganisms recovered in nonirrigated columns were lower than those in irrigated soil except for E. coli in LS. The inactivation of microorganisms was lower in the relatively wet soil of irrigated columns, as indicated by the higher soil recovery of Enterococcus species (Table 3). Leaching of phage during the early irrigation events also led to higher total recovery in irrigated columns (Fig. 3 and Table 3). With mineral N the opposite was the case. It was not clear why the level of mineral N recovered in irrigated columns was lower than that of remaining mineral N in nonirrigated columns. Short-chain fatty acids are the main constituent of DOC in slurry (60), and redistribution of DOC during irrigation events may have stimulated N immobilization and denitrification (61, 62). Recovery of mineral N in columns amended with RS was significantly but only slightly lower than that with LS or OLS, thus N immobilization was probably the main sink for N.

The survival of the microorganisms investigated generally followed the order of E. coli < Enterococcus species < phage. In accordance with this, the phages survived longer than E. coli in slurry-treated soils in a recent field study (63). Also, Enterococcus species remained viable longer than E. coli in soil-slurry mixtures of an experiment by Cools et al. (9), who studied the survival of E. coli and Enterococcus species from pig slurry applied to soil. They found that low temperature (5 versus 15 and 25°C), as well as high moisture content (field capacity versus drier soil), improved survival of these organisms. The relatively low incubation temperature of 10°C in our experiment may thus have increased the survival and leaching of the organisms investigated. Nevertheless, the leaching of bacteria in the current study was low compared to that in other studies where slurry samples or microorganisms in irrigating water were applied to wet soil and/or where irrigation was applied instantly (64, 65, 66). In the present study, slurry was applied to soil at field capacity, which is normal agricultural practice. During initial redistribution in the relatively dry soil, microorganisms may have reached relatively fine pore spaces protected from infiltrating water during subsequent irrigation events. The fact that the first irrigation took place 1 week after slurry application provided additional time for redistribution and inactivation. Our results thus suggest that slurry application during a period without rainfall can reduce the risk of bacterial leaching.

Effect of application method with irrigation.

Leachate compositions with surface application and subsurface injection of RS are presented in Fig. 4. Slurry injection increased the leaching of phage, E. coli, and Enterococcus species significantly compared to surface application and, for E. coli, with respect to both culturable cells and DNA (Table 3). Bech et al. (27) reported that the average proportion of Salmonella enterica leached was 6.1% after injection and 0.6% after surface application on a silt loam soil, although the difference was not significant due to high variability among replicates. Similarly, significantly more phage leached from intact sandy clay loam soil cores exposed to natural weather conditions when raw slurry was injected rather than being surface applied (67). Injection did not influence the amount of mineral N leached during our experiment (Table 3). A discrete distribution of slurry will, depending on NO₃⁻ availability, stimulate denitrification compared to a more homogeneous distribution (68); the missing effect of slurry distribution confirms that denitrification was not an important sink for mineral N in this study.

Fig 4.

Percent leached of total applied mineral N, phage, E. coli, and Enterococcus species, as well as FBA concentration, electrical conductivity (EC), total organic carbon (TOC) concentration, and turbidity of leachates for subsurface injection and surface application of raw slurry during four irrigation events over a 4-week period.

Surface application reduced leaching of microorganisms compared to injection after each individual IE (Fig. 4). The leaching of mineral N in the beginning of the experiment was low for surface application, which was compensated for by equal or even higher leaching in some columns at the end compared to injection (Fig. 4). This was not the case for microorganisms, probably due to the higher rate of inactivation and greater potential for attachment and straining-related retardation in the soil during transport.

The shorter leaching path with slurry injection most likely accelerated the emergence of contaminants in the leachate (Fig. 4). The pattern of emergence in the leachate of FBA and TOC with the two application methods supported the leaching patterns of the contaminants; the elution of FBA in IE1 for surface-applied slurry was lower than that for injection. Although surface application delayed the peak of both FBA and TOC (Fig. 4), the cumulative leaching of both FBA and TOC was similar for the two application methods (Table 3).

A difference in survival rates could explain the differences in leaching potential between the application methods. The exposure to desiccation may reduce the overall survival of microorganisms after surface application. Phage and Enterococcus species recoveries were higher when applied by injection. The relatively higher leaching in early IEs for injection, where inactivation was probably still low, may have been the major factor behind the higher level of recovery. No significant differences were observed between E. coli recoveries after surface application and injection of slurry (Table 3).

The relative amounts of microorganisms retained in sections S2 to S4 were either higher for injection than surface application or similar for the two application methods, except for phage in S2 (Fig. 2). E. coli (75 to 90%) and Enterococcus species (90 to 94%) were recovered mainly in S1 with both application methods, probably because of low mobility and/or low survival. With both application methods and all slurry types, EC and SWC in S1 remained higher than the background values (Fig. 2), showing that the characteristics of the injection slit environment were partly maintained even after four IEs, in accordance with the field observations of Petersen et al. (28), where gradients in water content around slurry injection slits were maintained following rainfall.

The leaching potential of different application methods may vary with soil structure. The extent of soil-slurry mixing during application is of fundamental importance in controlling the leaching potentials, because a better incorporation of slurry into the bulk soil can remove a major portion of contaminants from active macropore flow paths (21). On the other hand, greater redistribution of slurry constituents in a soil with few preferential flow paths may increase the interaction between matrix flow and contaminants and, hence, leaching. Surface application, being less expensive (23) and less risky with regard to contaminant leaching, may be the best choice for loamy soil types, especially for dilute slurries, and for fields with less risk of surface runoff.

Conclusions.

Initial convective redistribution of slurry constituents after slurry application was more pronounced when using the liquid fraction of slurry after solid-liquid separation of pig slurry compared to untreated pig slurry. More TOC, mineral N, phage, and E. coli leached during four leaching events after application of this liquid fraction. Ozonation of the liquid fraction reduced the leaching potential of E. coli only. A reduced leaching path after subsurface injection, and slightly higher survival of the microorganisms in injected slurry, probably increased the leaching potential of the microorganisms compared to surface application in this study. Even though slurry was applied to soil at around field capacity and in the beginning of a 1-week dry spell, bacteria, virus, and nitrate showed significant leaching potential. From a risk assessment and management perspective, it is necessary to take these results into account. More knowledge is needed about the effects on leaching potentials as a basis for improving slurry management technologies.