Chlortetracycline and tylosin runoff from soils treated with antimicrobial containing manure

Abstract

This study assessed the runoff potential of tylosin and chlortetracycline (CTC) from soils treated with manure from swine fed rations containing the highest labeled rate of each chemical. Slurry manures from the swine contained either CTC at 108 μ g/g or tylosin at 0.3 μ g/g. These manures were surface applied to clay loam, silty clay loam, and silt loam soils at a rate of 0.22 Mg/ha. In one trial, tylosin was applied directly to the soil surface to examine runoff potential of water and chemical when manure was not present. Water was applied using a sprinkler infiltrometer 24-hr after manure application with runoff collected incrementally every 5 min for about 45 min. A biofilm crust formed on all manure-treated surfaces and infiltration was impeded with > 70% of the applied water collected as runoff. The total amount of CTC collected ranged from 0.9 to 3.5% of the amount applied whereas tylosin ranged from 8.4 to 12%. These data indicate that if surface-applied manure contains antimicrobials, runoff could lead to offsite contamination.

Keywords:

• Antimicrobial chemicals
• tylosin
• chlortetracycline

Introduction

About 10 million kg yr^{− 1} of antimicrobial compounds are used in US livestock production.^{[ 1 , 2 , 3 ]} Tylosin and chlortetracycline (CTC), antimicrobial chemicals, are administered to livestock at therapeutic and sub-therapeutic doses to prevent, control, and treat health problems, and enhance growth rates.

In swine (Sus sp.) production, about 55% of all weaned pigs receive chlortetracycline (CTC), and 44% receive tylosin alone or in combination with other drugs.^{[ 4 ]} Therapeutic doses treat, prevent, and control disease with treatment lengths ranging from 27d (respiratory diseases) to 41d (enteric diseases). However, over 50% of antibiotics are given at subtherapeutic doses to promote growth and weight gain ^{[ 4 ]} with a treatment length average of 62d in grower/finisher pigs.

A typical hog can produce 545 kg of manure in a 6 month period.^{[ 5 ]} Tylosin and CTC are poorly absorbed in treated animals and over 90% of the parent compound or bioactive metabolites can be excreted in feces.^{[ 6 , 7 ]} Antibiotic concentrations in manure have been reported as high as 3970 (beef) and 133 (swine) mg kg^{− 1}.^{[ 2 , 8 ]}

Untreated or partially treated animal manures are often land applied to field crops for nutritive and disposal purposes with about 12 % of the US corn crop receiving an average of 13,200 kg manure/ha annually.^{[ 5 ]} Since antimicrobial compounds can be moderately long lived, with the half-life of tetracycline in pig slurry reported to be about 5 months,^{[ 9 ]} some antibacterial residues may remain in the environment for one or more growing seasons in soil fertilized with animal slurries.^{[ 10 ]}

The water solubility of tylosin is 6000 mg/L and CTC is 500 mg/L.^{[ 11 ]} The water solubility of each of these chemicals is very high and based on these data, may be prone to either leach or runoff with water applications. However, soil characteristics and antimicrobial type influence sorption characteristics.^{[ 12 , 13 , 14 , 15 ]} Sorption kinetics of tylosin are influenced by soil type with maximal sorption to a sandy loam soil occurring in 5 min but taking 3 h in a clay soil.^{[ 15 ]} In addition to sorption kinetics, tylosin partition coefficients (K_d) were reported to be 175 L kg^{− 1} when sorbing from water to manure, 7.6 L kg^{− 1} when sorbing to sand, and about 1300 L kg^{− 1} when sorbing to silty clay loam soils.^{[ 14 ]} CTC sorption to soil was found to be an order of magnitude less than sorption of tylosin but still relatively rapid, with maximal sorption to either sand loam or clay occurring within 10 min after application.^{[ 15 ]} Based on these sorption characteristics, little or no movement should occur except perhaps in sand.

Even though sorbed to a great extent, CTC and tylosin have been reported in surface waters in many areas.^{[ 8 , 16 , 17 ]} In a US Geological Survey (USGS) 2001–2004 survey of the Big Sioux River that drains about 13,300 km² of eastern South Dakota, both CTC and tylosin were found at low levels in some samples.^{[ 16 ]} Sioux Falls, the largest city in the region, gets 50% of its drinking water directly from the Big Sioux River. Concern about additional loading of antimicrobials has been expressed because more livestock operations are being developed in the area that do not raise enough animals to be placed under concentrated animal feelimg opreations (CAFO) regulations The majority of these operations land-spread manure. Therefore, the fate of antibacterial chemicals in manure needs to be assessed to understand the risk that these chemicals pose to the water supply and to enable mitigation before problems intensify. The objective of this study was to assess the runoff potential of CTC and tylosin from soil-applied manure.

Materials and Methods

Swine manure collection

Nine male pigs weighing approximately 35 kg were placed in individual stainless steel metabolism crates in order to collect urine and feces. Pigs were fed one of three dietary treatments: control, control + tylosin, and control + CTC. Labeled feeding rates for the two antimicrobial compounds were used; tylosin was mixed into feed at 0.11 g tylosin/ kg feed and CTC was mixed with the feed based on body weight of the pig so that they received about 22 mg CTC/kg body weight daily. The corn-soybean meal based diets met or exceeded the pigs' nutrient requirements and were identical except for the inclusion of antibacterial agents. Feeding of the different diets started on Day 1 and then, starting on Day 6, manure (urine and feces slurry) was collected daily, with collection ending at about Day 16. Daily collected samples were mixed each day in 150 L drums and stored at 4 C until use. For one of the field replications (see below) the solid fecal material was collected but held in a similar manner to the methods for collecting the slurry. Prior to taking samples for field treatment, the samples within a feeding treatment were well mixed.

One-L slurry samples were taken while continuous stirring occurred so that representative solid to liquid ratios could be obtained for field applications. Several replicate slurry samples across feeding trials were filtered and dried. Solid content of the slurry was, on average, 82 g of solid dry matter. After drying several fecal samples to determine percent water, treatments that received fecal solids were weighed so that the final amount of dry material was also about 82 g. Portions of the fresh manure treatments were frozen for laboratory analysis. Samples for field experiments were used “fresh” as soon as practical after collection (1 to 21 d), to represent worst case scenario of land application.

Soils and manure treatment

Four sites with different soil types in the eastern South Dakota landscape were used to assess antimicrobial chemical runoff. Soil types, physical characteristics, permeability and runoff classification are presented in Table 1. Because the study was conducted by confining manure in small (24-cm) ring enclosures, the slope was not considered a variable to runoff, rather, the objective was to examine the infiltration of water and chemical into soil and the runoff that occurred due to the surface applied manure treatment.

Table 1 Subgroup classification, physiochemical characteristics of the Ap horizon, permeability rating, and runoff class of eastern South Dakota soils used in this study. Brandt, Brookings, and Vienna soils were located in Brookings County whereas Houdek soil was located in Moody County.

	(g/kg)
Soil type	Subgroup classification	Sand	Silt	Clay	Organic matter	Permeability rating ^1	Runoff class ^2
Houdek clay loam	Typic Argiustoll	340	370	290	30	Moderately slow	Moderate/High
Vienna silt loam	Calcic Hapludoll	210	550	240	60	Moderately slow	Medium
Brookings silty clay loam	Pachic Hapludoll	70	620	310	60	Moderately slow	Low
Brandt silty clay loam	Calcic Hapludoll	390	383	226	35	Rapid	Low

Three locations (upslope, midslope, and footslope areas) in a Moody County, SD (east-central SD) field were chosen for one study. A Houdek clay loam (fine-loamy, mixed, superactive, mesic typic Argiustoll) was the field soil type with a 3% slope from the upslope to footslope study sites. The permeability of the Houdek soil was rated by the NRCS ^{[ 18 ]} as moderately slow and the runoff class was high (at summit) to moderate (at footslope). In this soil, areas that were trafficked (wheel tracks present) and nontrafficked also were chosen at each location.

Three other sites located in different fields with three different soil types (Table 1) were located in Brookings County, SD. One site had a Vienna silt loam (fine-loamy, mixed, superactive, frigid Calcic Hapludolls) soil type and was located in a field just north of Brookings at the summit of the landscape. The Vienna soil had about a 2% slope, the permeability rating was moderately slow, and the runoff potential was rated as medium.^{[ 19 ]} Another site had a Brookings silty clay loam (fine-silty, mixed, superactive, frigid Pachic Hapludolls) soil type and was located at the toeslope position in a field. The landscape was relatively flat, the permeability moderately slow, and runoff potential was rated as low.^{[ 19 ]} A third location was in a field that had a Brandt silty clay loam (fine-silty, mixed, superactive, frigid Calcic Hapludoll) soil type and was located just east of Brookings, SD. The site had no slope and the Brandt soil permeability was rated as rapid and runoff potential was rated as low. ^{[ 19 ]}

Prior to surface application of manure, 24-cm diameter steel rings were inserted into the soil to a 7.5 cm depth. Treatments were applied to the ring 24-h prior to the runoff event. Applications at the three Brookings County sites were 1-L of manure slurry from each feeding treatment was poured over the soil surface contained in the ring. At the Moody sites, the solid fecal material was added to 1-L of deionized water, agitated to get a slurry-type mixture, and then poured on the soil surface.

In order to assess the influence of manure on water and antimicrobial runoff, nonmanured control treatments were used on the Houdek soil. Two g of dry formulation of formulated tylosin (Tradename: Tylan 40; containing 88 g tylosin kg^{− 1}; Elanco, Greenfield, IN) was mixed into 1L of water and applied similar to the manure slurry application. KBr also was added to the Houdek plots with a total of 890 mg Br⁻ per ring.

Sprinkler infiltrometer treatment

Water was applied to the rings using a Cornell Sprinkler Infiltrometer ^{[ 20 ]} about 24-h after application of manure or chemical. The water-filled infiltrometer was placed on top of the steel ring and leveled. The initial height of the water column inside the infiltrometer was recorded. The infiltrometer used a marriotte tube to maintain constant application rate. The application rate was calculated as the original water column height minus the final water column height, divided by the time of application. The application rate was about 230 mm/hr, which is similar to a 50-year, 10-minute storm for eastern SD.^{[ 21 ]} Runoff samples were collected incrementally over a 45 min rain event in brown glass bottles. During the rain event, bottles were changed every 5 min on the Brookings County soils, every 2 min for tests run on Houdek trafficked sites, and every 3 minutes on Houdek nontrafficked sites. Bottles were put on ice and stored in the dark at 5C until analysis.

The Brookings County soils were soil sampled in the ring 48 hrs after water application. Soil was collected from all sites from the 0 to 7.5-cm or 0 to 15-cm depth. In addition, some rings were sampled to 60 cm by 15 cm increments. All soil samples were stored in the dark at 5°C until analysis for either CTC or tylosin.

Determination of chemicals in runoff water

Bromide in runoff water from the Houdek soil was quantified using Br⁻ electrode. Samples extracted for antimicrobials were first treated with 1 g of disodium ethylenediaminetetraacetic acid (EDTA) (Fisher, Pittsburgh, PA), regardless of amount collected during the collection period, for ion sequestration. CTC was extracted and quantified based on methods modified from Reverte et al. ^{[ 22 ]} and tylosin was extracted and quantified based on methods modified from Jacobsen et al.^{[ 23 ]} Water from CTC treatments were acidified with 5N sulfuric acid to a pH of < 3. A 200-ml aliquot of runoff water was vacuum-filtered through a glass fiber filter (Whatman #5, Florham Park, NJ). The filtrate was pulled through preconditioned HLB column (Waters, Milford, MA) under vacuum at a flow rate less than 5 ml/min. After extraction, the columns were eluted with methanol (Fisher, Pittsburgh, PA) and then evaporated to near dryness. The remaining residue was dissolved into 1.5ml of a 90% acetonitrile/10% 100mM ammonium acetate solution.

Soil and manure extraction for antimicrobials

One gram of manure or soil was mixed with 1.2 mL of 1M citric acid buffer (pH = 4.7) in a tube and vortexed for 1 min. Six mL of ethyl acetate (Fisher) was added to the test tube and vortexed again for 1 min. The tubes were then shaken for 15 min and centrifuged at 7000 rpm for 10 min. The organic phase of the supernatant was transferred to a clean tube. Another 6 mL of ethyl acetate was added to the tubes and the solution was vortexed. The tubes were shaken and centrifuged as stated above. This was repeated two more times with the ethyl acetate collected after each extraction. This ethyl acetate solution was evaporated to dryness under a nitrogen gas stream at room temperature. The residue was re-dissolved in 1.5 mL of a 90% acetonitrile/ 10% 100 mM ammonium acetate solution. Samples were stored at 10°C before liquid chromatography/mass spectrometry (LC/MS) analysis.

LC/MS analysis of chlortetracycline and tylosin

A 10 μ L aliquot of each sample was injected into a Waters 2695 LC equipped with a Micromass ZQ-MS. The LC had a Waters Nova-Pak C8, 4 μ m, 3.9 × 150 mm column with a Waters Nova-Pak C8 guard column. The flow rate was 1mL/min. Mobile phase A was acetonitrile and mobile phase B was 0.5% formic acid (Fisher) in water with 1mM ammonium acetate. The runtime was 25 min with a mobile phase gradient progressing from 100% B to 50% B/ 50%A at 9 min and holding this ratio for 2 min. The gradient was changed over the next 1 min to 100% A which ran for 3 min. For the remaining time, the gradient was switched back to 100% B and ran for about 12 min for column clean-up. MS analysis was conducted using electrospray ionization at positive ion mode.

Analytical standards of tylosin A, tylosin B, and chlortetracycline were purchased from Sigma-Aldrich (St. Louis, MO) and solutions ranging from 0.6 to 20 μ g/L of each chemical were used for standardization. The masses collected were 479 for CTC, 916 for tylosin A (parent compound), and 772 for tylosin B, a metabolite of tylosin A. Standards, duplicates (at least 10% of samples), and extracted spiked (25 and 50 μ g/L of either analytical standard of tylosin A or CTC) samples, were included in each LC/MS run.

The limit of detection for either tylosin A or CTC in prepared standards was < 0.2 μ g/L. Recoveries of tylosin A or CTC from extracted spiked samples were consistent among extractions and were about 10 and 40% of the applied chemical, respectively, regardless of the original spike concentration. The analytical standard of tylosin B was spiked into both manure and water samples but was not consistently quantifiable from extracted spiked samples although amounts detected in prepared standards were similar to tylosin A. Due to the inconsistent recovery of tylosin B, only tylosin A is discussed. It is also known that metabolites of CTC are excreted in manure, however, these were not measured in this experiment. The software used for analysis was Masslynx 4.0 (Waters, Milford, MA). Chemical recoveries in samples are reported based on correction for the amount recovered from spiked extracted samples.

Data analysis

Two replications were used on the Houdek soil for each traffic treatment. Three replications were used on the other soil types and then repeated in time in 2005 and 2006. Statistical analyses were performed and standard errors and deviations from means were calculated.

Results and discussion

Manure antimicrobial concentrations

After manure collection, slurry samples were extracted and analyzed for CTC and tylosin A concentrations. The CTC concentration was 108 mg/L ± 24. This concentration of CTC was about 14 times greater than concentrations reported from Minnesota swine operations ^{[ 24 ]} and 2.5 times greater than the amount reported in slurry analysis from swine feedlots in Austria.^{[ 25 ]} High CTC levels in this study were attributed to the use of fresh manure rather than manure collected from lagoons or pits.

The tylosin A concentration was 0.11 mg/L ± 0.03 for the manures applied to the Brookings County soils. Only manure from swine fed tylosin was used for the Houdek soil with a concentration of about 0.3 mg/L of tylosin A. The tylosin A concentration of the manure used for the Brookings County soils was similar to and, for the Houdek soil, 3 times greater than the amount estimated in flushwater from confined swine feeding operations reported by Mellon et al.^{[ 2 ]}

Runoff water and antimicrobial recoveries from sprinkler infiltrometer studies

The sprinkler infiltrometer was placed about 100 mm above the soil surface. This low water application elevation meant that water droplets had little chance to gain kinetic energy due to gravity. Thus, the water struck the soil surface with relatively little energy and may have resulted in reduced soil crust formation and surface seal based on typical rainfall intensity.^{[ 26 ]} Therefore, under these conditions, infiltration rates may have been maximized and runoff minimized compared with natural rainfall conditions, even though runoff recovery was very high in manured treatments (see discussion below).

Runoff from the Houdek soil was influenced by position on the slope when no manure was applied and no wheel tracks were present, with infiltration rates of 162 mm/hr at the toeslope position and 234 mm/hr at the shoulder slope (p = 0.01) (data not shown). When manure was applied, the amount of runoff was similar among the three landscape positions and traffic treatments (Table 2) and average infiltration rate across positions was slowed to 31 mm/hr. In all manure treatments with or without traffic, runoff recovery averaged 88% of applied water among all slope positions, whereas runoff recovery from the no manure-no traffic treatment averaged among landscape positions was 21% of the applied water (data not shown). The percentage of bromide collected in the runoff also was influenced by manure and traffic treatments with the rank of high to low percentage as: manure/traffic > no manure/traffic = manure/no traffic > no manure/no traffic (p = 0.02) and averaged 26%, 12%, and 1% of the applied bromide, respectively.

Table 2 Water application and runoff amounts from three sites in Brookings County. Water was applied with a Cornell sprinkler infiltrometer over a 45-min time period with runoff collected until water no longer ran off the plot. Data are presented as the mean with standard deviation in parentheses.

Manure type	Soil type		Water applied (L)	Water recovered (L)	Recovered (% of applied)
Tylosin	Houdek clay loam ^†	No manure/no traffic	7.6 (1.7)	1.8 (1.3)	21 (13)
		Manure/no traffic	8.7 (1.8)	8.5 (1.5)	88 (5)
		No manure/ traffic	7.2 (2.1)	6.5 (1.2)	88 (8)
		Manure/ traffic	8.1 (1.0)	7.4 (1.2)	90 (23)
	Vienna silt loam		4.0 (1.0)	3.8 (1.0)	95
	Brookings silty clay loam		6.2 (1.8)	5.5 (1.5)	87
	Brandt silty clay loam		5.4 (2.1)	5.0 (1.9)	93
Chlortetracycline
	Vienna silt loam		4.1 (0.4)	3.9 (0.3)	95
	Brookings silty clay loam		5.2 (0.5)	4.7 (0.4)	91
	Brandt silty clay loam		7.7 (2.8)	5.5 (2.0)	71

Although differences in water runoff among Brookings County soils were anticipated based on NRCS infiltration information (Table 1), these three soils had similar high runoff amounts that ranged from 71 to 95% of the applied water (Table 2) with very low infiltration amounts. The high runoff amounts in all areas treated with manure were the result of several factors. First, there were high water application rates to the treatment areas that did not allow for much infiltration. Second, small particulates in the manure slurry most likely sealed most the pores of the soil surface (total suspended solid concentration: 25,000 mg/L for tylosin manure and 35,000 mg/L for CTC manure; authors unpublished data). Third, manure slurry was somewhat water repellent in nature (observed as a crust on the soil surface) most likely due to fatty acids present in the fecal material.^{[ 27 ]} Finally, a green biofilm formed on the surface of the manure slurry and was observed even though the rainfall event occurred only 24-hr after the slurry was applied to the soil surface (author's observation).

Runoff water from Houdek soils when tylosin was applied directly to the soil surface resulted in recovery of 0.1 to 6% of the applied tylosin. The greatest recoveries occurred from the trafficked (average 3%) rather than non-trafficked (average 0.1%) areas from all landscape positions (data not shown). The amount of tylosin lost in runoff water from the non-trafficked areas in this study is similar to the amount reported in runoff by Davis et al. ^{[ 28 ]} when this antimicrobial chemical was applied directly to the surface of a sandy clay loam soil.

When manure containing tylosin was applied to the Houdek soil, the amount of tylosin collected in runoff water averaged 20% (about 0.08 mg) of the amount applied for all except two plots (data not shown). The two plots were located at the shoulder slope position, one with traffic and the other non-trafficked, and averaged less than 2% of the tylosin in the runoff. The runoff samples collected in the first 10 minutes of the start of the runoff events contained all of the tylosin detected. However, since detection of tylosin was low, there may have been more in the subsequent runoff that was not measured. This uncertainty could have been reduced if the samples had been pooled and then the entire amount concentrated rather than analyzing each sample separately.

The percentage of applied tylosin collected in runoff samples from the three Brookings County soils was statistically similar among the three soils types but ranged 19.2 (Brookings silty clay loam) to 28.0 (Vienna silt loam) μ g (Table 3). This accounted for 8.4 to 12.1% of the total amount applied (Table 3). Over 50% of the tylosin loss occurred during the first 10 minutes of runoff (Fig. 1).

Table 3 Average total mass of tylosin A and chlortetracycline measured in runoff from manure treated soils located in Brookings County, SD. Data are presented as the mean with standard deviation in parentheses.

	Soil type
Antimicrobial chemical	Vienna silt loam	Brookings silty clay loam	Brandt silty clay loam
	(μ g)
Tylosin A	28.3 (5.1)	19.2 (6.7)	20.5 (8.9)
Chlortetracycline	2046 (95)	4174 (880)	916 (280)

Fig. 1 Runoff and tylosin losses for each 5-minute collection interval during simulated rainfalls on three Brookings County soil types treated with manure from swine fed rations containing tylosin.

The percentage of CTC recovered in runoff water ranged from 0.86 to 3.54% of the amount applied. While the percentage of recovery was lower than tylosin, the total mass recovered was much greater and ranged from about 915 to 4200 μ g (Table 3). Ranking of CTC runoff was: Brookings silty clay loam > Vienna silt loam > Brandt silty clay loam with differences significant at p < 0.01 (Table 3). Unlike tylosin that had most of recovered mass collected in the first few minutes of runoff, relatively high amounts of CTC were collected during the entire runoff event from each of the three soils (Fig. 2).

Fig. 2 Runoff and chlortetracycline (CTC) losses for each 5-minute collection interval during simulated rainfalls on three Brookings County soil types treated with manure from swine fed rations containing chlortetracycline.

Several depth increments of soil were collected 24 hr after the rainfall events from the treated areas and extracted for the antimicrobial compound applied. However, only trace amounts of antimicrobials were found in surface soil with none detected at lower depths. These data may indicate that the CTC and tylosin: 1) did not leach into the soil due to the low infiltration rates when manure was present; 2) that most of the antimicrobial remained sorbed to either the manure in the slurry addition or to the surface soil and did not move; or 3) because of poor recovery of antimicrobials even from spiked soil, slurry, and water samples (< 50%), sample sizes were not large enough to detect amounts that may have been present.

Based on sorption coefficients, (K_d or K_f) values of tylosin and CTC, that give indications of the amount of chemical sorbed to the matrix vs the amount that remains in solution, we expected to see differences in the runoff amounts among the soils selected for this study due to differences in organic carbon and clay content. For example, the sorption K_f values of tylosin to loam and clay loam soils have been reported to be about 1300, however the K_d was reported to be about 175 when tylosin was sorbed from water into manure.^{[ 14 ]} Allaire et al.^{[ 15 ]} reported a K_d value for CTC of about 800 for heavy clay soils and 8000 for tylosin K_d value for the same soil type. These high K_d values indicate very high sorption to soil and, therefore, movement from surface soils through ‘normal’ leaching events (without facilitated, channel, or by-pass flow) may be very restricted, although CTC was detected at 25 to 35 cm depth in a sandy loam soil.^{[ 29 ]} Davis et al. ^{[ 28 ]} reported that 50 to 80% of the tylosin and nearly all of the CTC losses during a runoff event with no manure present were associated with sediment. The similarity among runoff amounts from the soils used in this study may be an indication that antimicrobial chemical sorption to manure in the original slurry may be the factor that overrides other movement characteristics and movement may not be directly related to soil properties.

Summary

The data from this study in combination with data from other studies indicate that manure from animals treated with antimicrobial chemicals should be handled carefully. The half-lives of tylosin and CTC in manure applied to sandy loam soil have been reported as 6 to 21 d, respectively,^{[ 29 ]} but in slurry, 50% of the original CTC was present in samples for as long as 5 months ^{[ 9 ]} and may be longer under cold conditions (author's unpublished data). While leaching of these antimicrobial chemicals into and through the soil profile may be limited, their movement in runoff to offsite areas may be an unintended consequence of manure application. This problem may be exacerbated if a crust or biofilm is formed when manure is applied, as infiltration rates dramatically decrease, leading to increased runoff. In addition, the direct soil application of antimicrobial chemicals that are not applied in a manure matrix may not realistically represent their soil movement or runoff potential.

Sensible management practices should be used with manures that contain antimicrobial chemicals. Practices to limit runoff of chemicals from manure should include placing manure in areas that are less prone to runoff; placing manure away from heavy trafficked areas and surface waters; and applying manure to soils with higher clay or organic matter content so that antimicrobial chemicals are sorbed and have limited leaching potential. In addition, incorporating the manure to a shallow depth or injecting the manure below the surface will limit the crusting potential and surface exposure of the manure to rainfall as well as reduce N loss.^{[ 30 ]}

Acknowledgment

This research was partially funded by USGS Grant 104 program No. 06HQGR0120, US Environmental Protection Agency (EPA) grant number CP-97835401-0, EPA-319 fund, and the South Dakota Agricultural Experiment Station. The LC/MS was purchased with support from the National Science Foundation under grant no. 00091948. Thanks to Drs. V. Brözel and S. Gibson. Mitch Olson and Steve Biersbach provided field support for this project.

Notes

¹Permeability ratings are from Natural Resources Conservation Service (NRCS) bulletins.^{[ 18 , 19 ]} The average rapid permeability is 42 μ m/sec and average moderately slow permeability is 2.8 μ m/sec.

²Runoff classifications are reported from NRCS.^{[ 18 , 19 ]}

^†Data presented were pooled over landscape position.