| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Department of Water and Agriculture, Swiss Federal Institute for Environmental Science and Technology (EAWAG), Überlandstrasse 133, 8600 Dübendorf, Switzerland
b Department of Environment and Resources, Technical University of Denmark, Bygningstorvet, 2800 Lyngby, Denmark
c Swiss Agency for the Environment, Forests and Landscape (SAEFL), 3003 Bern, Switzerland
* Corresponding author (christian.stamm{at}eawag.ch)
Received for publication July 9, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: ATR, atrazine • EC, electrical conductivity • LOD, limit of detection • LOQ, limit of quantification • SA, sulfonamide antibiotic(s) • SDM, sulfadimidine • SDZ, sulfadiazine • STZ, sulfathiazole
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Sulfonamides (SA), tetracyclines, fluoroquinolones, macrolides, and ß-lactames are the most important groups of veterinary antibiotics. In this paper, we focus on SA because they are important with respect to amounts administered in pig production in Switzerland and the European Union (Ungemach, 2000). In addition, SA seem to be rather persistent in manure, soil, and water (Langhammer, 1989; Thiele-Bruhn, 2003).
The majority of administered SA is excreted with urine and feces in unaltered form or as metabolites (Vree and Hekster, 1985; Langhammer, 1989). Vree and Hekster (1985) observed that on average 50 to 90% of the parent SA were excreted by pigs within 2 to 20 h after administration. After regular administration, Langhammer (1989) measured up to 20 to 40 mg L–1 of sulfadimidine (SDM, synonymous with sulfamethazine) and sulfathiazole (STZ) in fresh pig manure. In stored manure, the same SA were observed at concentrations up to 20 mg L–1 (Langhammer, 1989; Haller et al., 2002; Höper et al., 2003). Based on these concentrations and common manure application rates in Switzerland, up to a few hundred grams of SA per ha may be applied to agricultural soils each year (Burkhardt et al., 2004).
After the manure application on agricultural land, substance properties, soil, and weather conditions are the main factors controlling the loss of SA to surface waters. For herbicides or phosphorus, it is known that transport to surface waters occurs mainly during heavy rainfall and that they get mobilized from the topsoil in surface runoff or preferential flow to subsurface drains (Ahuja, 1986; Ghodrati and Jury, 1992; Leu et al., 2004; Stamm et al., 2002). The few field studies on SA transport in field soils indicate a similar transport behavior (Boxall et al., 2002; Höper et al., 2003), but are not very conclusive.
How much of a substance gets lost by fast transport like surface runoff and preferential flow is strongly influenced by its sorption to the topsoil. A crucial factor affecting this interaction of SA with soil is the pH that controls the speciation of the SA (Langhammer, 1989; Thiele, 2000; Boxall et al., 2002; Thiele-Bruhn et al., 2004). The pKb value of SA, characterizing the transition from the neutral species to the conjugate base, ranges between pH 6 to 7.5 (Table 1). Since sorption of anions is generally rather low in soils, it may be expected that the higher the pH the weaker the SA sorption. This has been confirmed for different soils (Langhammer, 1989; Boxall et al., 2002; Thiele-Bruhn and Aust, 2004) and pH (pH 4: Kd = 30 L kg–1, pH 8: Kd < 1 L kg–1; Boxall et al., 2002). Hence, manure probably increases the SA mobility since its alkaline pH is generally higher than that of the soil solution. Additionally, dissolved organic matter (DOM) in the manure may reduce the affinity of SAs to soil due to sorption competition (Thiele-Bruhn and Aust, 2004) or colloid-facilitated transport (Tolls, 2001).
|
The main goal of the present study was to investigate how different SA species are mobilized after they have been applied as part of a manure application that is typical of practices in Switzerland. To assess the influence of the manure on surface transport, the antibiotics were also applied in aqueous solution as a control. The simultaneous application of the conservative tracer bromide (Br) and the herbicide atrazine (ATR) should reveal how physical mixing and chemical properties influence the fate of SA.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
The spiked concentrations were defined as C0, and the measured concentrations in the filtrate of the manure 24 h after spiking (see below) and in the aqueous solution as C0meas. To compare the transport behavior of the different substances we normalized the measured runoff concentrations (C) to the corresponding measured input concentration C0meas yielding after multiplication by 100 the relative concentration C%.
Treatment of the Plots
On eight plots, we applied 3 L m–2 of liquid pig manure and on four plots the same amount of the aqueous solution with a watering can. The applied manure corresponds to 30 m3 ha–1, which is a typical agricultural practice for Switzerland. The spatial arrangement of the different treatments was random. After a contact time of 1 or 3 d after the application (manure: four plots for each contact time, aqueous solution: two plots for each contact time), the plots were irrigated with deionized water. The EC of deionized water (5 µS cm–1) was comparable with natural rainfall. The irrigation rate was 20 mm h–1 and the total amount 30 mm. The irrigation rate corresponds to a heavy rainstorm in this area. One manured plot with a 1-d contact time was irrigated with a smaller amount due to problems with the pressure control at the sprinkler. This plot was not taken into account for the runoff interpretation. We collected surface runoff by a graduated cylinder at intervals of 15 min. For each interval runoff aliquots of 20 mL were taken in brown glass flasks and stored at 4°C.
Analytical Procedures
The manure samples were diluted with water 1:5, simulating the dilution of irrigated manure on the field site, transferred into glass tubes, and centrifuged at 3000 x g for 15 min at 10°C. The supernatant was filtered through a nonsorptive 0.45-µm membrane nylon filter, diluted with water 1:10, adjusted to pH 4.2 with an acetic buffer, and filtered again. Additionally, we extracted SA from the manure by an accelerated solvent extraction (ASE 200 Accelerated Solvent Extractor; Dionex, Sunnyvale, CA) procedure [water to acetonitrile 85:15 (v/v), pH 9.0 with Tris 0.1 M, 100°C, 200 bar, and 5-min static time]. During the validation of this procedure no degradation and adsorption on pure quartz sand was observed. One extraction cycle was sufficient, since a second increased the amount by less than 10%. The substances were quantified according to the same method as for the runoff samples described below.
The runoff samples were filtered through 0.45-µm membrane nylon filters. An aliquot of 180 µL from each sample was filled into brown high performance liquid chromatography (HPLC) vials. To overcome matrix suppression effects, isotope-labeled internal standards for each SA (SDZ, STZ: Toronto Research Chemicals, North York, ON, Canada; SDM: Cambridge Isotope Laboratories, Andover, MA) and ATR (Dr. Ehrenstorfer, Augsburg, Germany) were added (Table 3). The three SA and the metabolite N4–acetyl-sulfadimidine as well as ATR were analyzed with liquid chromatography coupled to mass spectrometry (HPLC: Agilent, Palo Alto, CA; MS: Micromass, Beverly, MA) according to Haller et al. (2002).
|
Mass spectra of SA and ATR and their internal standards were acquired at their specific mass per charge ratios (m/z) in the single ion monitoring (SIM) mode with cone voltages of 60 V (SA) and 40 V (ATR), respectively, using positive ion electrospray mode (ESI+) (Table 3). For a proper quantification above the limits of quantification (LOQ) the intensity of the qualifier ion had to be 90 to 110% of that of the parent ion and the signal to noise ratios (S/N) had to be above 10:1. The limits of detection (LOD) were defined by S/N above 3:1. With the ratios of the quantifying ions varying greatly among the different SA and ATR, the LOQ and LOD for the different substances varied considerably (Table 3).
Bromide was measured by ion chromatography (761 Compact IC; Metrohm, Herisau, Switzerland). The separation was achieved by a Metrosep A Supp 4 column coupled with a pre-column (Metrohm) using water with 5% acetone buffered by 1.7 mM NaHCO3 as the eluent. The flow rate was 1.0 mL min–1, the injection volume 20 µL. We determined the EC by using a conductivity meter. The LOQ was 1.3 µS cm–1.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
A single manure application is not likely to have a permanent effect on runoff due to the influence of, for example, rainfall, biological activity in soil, or plant growth. Bischoff (1984), for example, showed that the manure effect on runoff vanished within 7 d. Therefore, the increased runoff on manured plots after a 3-d contact time was not expected. The drying of organic matter on the soil surface might repel part of the irrigation water decreasing the infiltration capacity (Haynes and Naidu, 1998; Benito et al., 2003).
Transport Behavior
Bromide
The conservative, nonsorbing tracer Br was used in this study to assess the physical mixing process between irrigation water, soil solution, and manure. Hence, the comparison of the Br behavior with the other applied substances allowed us to distinguish physical mixing and adsorption effects affecting surface transport. The actual Br concentrations C0measBr measured in the liquid phase of the manure were almost identical to the nominal spiked value C0Br (C0Br = 22 g L–1, C0measBr = 22.9 g L–1).
On both manured and control plots, the highest CBr% in runoff was always observed within the first 30 min (Fig. 3) . Within 0 to 15 min the CBr% reached average values of 3.7 and 6.9% on single manured plots for 1- and 3-d contact times, respectively. Thereafter, the CBr% values decreased steadily and reached an average level of about 2.0 and 1.5% after 90 min for 1- and 3-d contact times, respectively. The differences between the two contact times are not significant.
|
The observations for Br were confirmed by the EC measurements. For both EC and CBr%, the applied manure or aqueous tracer solution was the main source. Indeed, EC and CBr were highly correlated (manured plots: r = 0.98 ± 0.04, control plots: r = 0.90 ± 0.09). At the end of the irrigation period, the high EC values (400–500 µS cm–1) corresponded to relative concentrations of 1 to 1.3%, values that matched well with those for Br.
Sulfonamides and Tracers
To assess the SA concentrations in runoff it is crucial to know how much of the spiked amount was available for transport. The concentrations measured in the liquid phase of the manure C0meas may be assumed to be completely mobile. They were smaller than the nominal spiked concentrations C0SA with values of 30.7 mg L–1 for SDZ and 27.2 mg L–1 for STZ. Hence, about 60% of the spiked SA was directly transport available. We could extract somewhat more SA by the accelerated solvent extraction procedure. The corresponding values for SDZ and STZ were 36.1 and 30.9 mg L–1, respectively. For SDM, which was not spiked, the concentration in the supernatant was 11.0 mg L–1 and after accelerated solvent extraction increased to 14.4 mg L–1. Assuming that the difference between spiked and measured concentrations was due to sorption to the manure solid phase, these values corresponded to a Kd value of 30 to 40 L kg–1 and to a Koc of 60 to 80 L Kg–1 (assuming an organic C content of 50% in the solid fraction). These values are higher than those reported for soils under basic pH conditions (Langhammer, 1989; Tolls, 2001; Thiele-Bruhn and Aust, 2004). The difference between spiked and measured concentrations could also be due to the transformation products. However, the acetyl-conjugate of SDM as a major metabolite was not detected in manure after 3 d, which is consistent with the results of Langhammer (1989).
For ATR, the observed concentration in manure, which was not diluted before centrifugation, was 27.3 mg L–1. This is close to the limit of its water solubility (33 mg L–1). In the diluted (1:5) manure, we measured C0meas = 46.8 mg L–1, which is close to the spiked concentration (C0 = 50 mg L–1). This indicated that the solid phase of ATR was dissolved on the dilution. To be consistent with the SA, we calculated the relative concentration CATR% by normalizing the measured concentrations to C0meas. Degradation of ATR was probably negligible within 3 d (Topp et al., 1996).
Based on their pKb values (Table 1) we expected that the SA losses from the manured plots would be larger compared with the controls because of their higher pH. The results clearly support this expectation. On manured plots with a 1-d contact time the relative runoff concentrations of SA reached averages of maximal 0.3% for SDZ, 0.8% for STZ, and 1.4% for SDM (Fig. 4) . On the controls only 0.16% for SDZ and 0.08% for STZ were determined. After a contact time of 3 d the highest concentrations on single manured plots were 490 µg L–1 for SDZ, 1030 µg L–1 for STZ, and 680 µg L–1 for SDM. These values correspond to relative concentrations of 1.6, 3.8, and 6.3% for SDZ, STZ, and SDM, respectively. These high concentrations occurred during the first 30 min of runoff. Afterward, the values decreased steadily reaching fairly stable levels at about 90 min of runoff time (Fig. 4) (e.g., after 3 d on average <0.2, <0.3, and 0.7% for SDZ, STZ, and SDM, respectively). This temporal evolution of the SA concentration in runoff was qualitatively very similar to the pattern of Br (Fig. 4). The effect of the contact time on the manured plots was also similar for SA and Br. After 3 d more SA were mobilized within the first 30 min (first two samples) than after 1 d.
|
Recently, an empirical model has been proposed to derive Kd values for different SA in soils (Thiele-Bruhn et al., 2004). On one hand, it predicts that sorption increases with increasing organic C content. On the other hand, sorption of different SA is predicted to increase with larger retention times by reversed phase chromatographic separation with HPLC. Accordingly, sorption for the SA should increase in the order SDZ < STZ < SDM and the losses by surface runoff should follow the opposite trend. In reality, the reversed order was observed on the manured plots (Fig. 4). Similar findings have been reported by Höper et al. (2003) in a field study on vertical transport of veterinary antibiotics in arable soils. On the control plots, however, the behavior of SDZ and STZ agreed with the model prediction for a contact time of 1 d. After 3 d, both substances were below LOD (Fig. 4), making an evaluation impossible.
This contrasting behavior of the SA in the controls and the manure treatment demonstrates the influence of the manure on the mobility of these substances. Since manure was observed on the vegetation right before irrigation, part of the manure probably did not interact with soil at all. Hence, the relative mobility of different SA after application to soils may change with time depending on whether the chemistry is still dominated by the manure or the soil solution.
The relative Br concentrations CBr% varied slightly among the treatments. To remove these effects of physical mixing we normalized the relative concentrations of the reactive compounds (C%) to those of CBr% (Fig. 5) . Remaining differences between the substances can then be attributed to chemical processes affecting the mobilization process into runoff. Assuming that the substances are mobilized from a well-mixed pool at the soil surface, we expected the concentrations of reactive compounds in runoff to increase with time relative to those of the conservative tracer Br. The pool of reactive substances in the aqueous phase should be steadily replenished from the adsorbed phase, whereas the Br pool was only depleted.
|
A possible explanation could be that after 3 d SA had adsorbed to colloids. The colloid-facilitated transport on the sealed soil surface might be faster than the transport of the conservative Br due to the effect of limited matrix diffusion. This phenomenon is known from chromatographic separation as well as for colloid transport in soil (McDowell-Boyer et al., 1986). An alternative hypothesis would be that the irrigation with water of pH = 5.5 increased the affinity of the SA to soil and/or manure particles during the experiment counterbalancing the replenishment of the mobile SA pool. Why this should happen only after 3 d remains unclear. A third explanation could be that the sorption was not reversible at the time scale of the experiment.
The normalization with Br gives the possibility to estimate apparent Kd values for SDZ and STZ in our soil. Thereby we assume that the relative SA concentrations reflect the distribution between the aqueous and the solid phase. Based on mass conservation and assuming linear isotherms as well as a soil to water ratio of 1 in the soil layer wherefrom the SA were mobilized into runoff, Kd values between about 10 and 20 L kg–1 are derived for SDZ and STZ, respectively. For smaller soil to water ratios, which may be more realistic, these values would be even larger. Such values are high compared with results reported in the literature. Nevertheless, it is not completely surprising to observe strong sorption given the pH of the soil (5.7) and the irrigation solution (5.5). At these values, both SA were present as the neutral form thought to be adsorbed significantly stronger than the anions.
Total Losses
Despite the fact that we simulated a worst-case scenario by irrigating wet soil shortly after manure application, on average only 2.6% of Br, 1.7% of ATR, 0.2% of SDZ, 0.4% of STZ, and 0.8% of SDM were lost by runoff after 90 min for a contact time of 1 d (Table 4). On manured plots with a 3-d contact time the losses were somewhat larger. These values were obtained by normalizing to the measured C0measSA instead of the spiked input concentrations, since this nonsorbed SA fraction is directly available for mobilization. Using the spiked mass as reference (C0), the loads of SA from manured plots would be only 60% of the values listed in Table 4.
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
These results demonstrate that one may underestimate SA losses by surface runoff if the influence of manure is not considered. Hence, it is important to understand the manure–water–soil interactions. On the other hand, the data indicate that it may be difficult to transfer results obtained for chemically similar substances like ATR to SA because the latter reach the environment in an application matrix with completely different properties (e.g., with basic pH, solid particles).
Despite enhanced SA losses due to the manure application, the observed loads in runoff were in the same range as ATR. The losses on the control plots, however, where the SA had direct contact with the soil material, were substantially smaller. This indicates that the SA losses to surface waters may decrease rapidly after the chemistry of the manure does not control SA mobility.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Larsbo, K. Fenner, K. Stoob, M. Burkhardt, K. Abbaspour, and C. Stamm Simulating Sulfadimidine Transport in Surface Runoff and Soil at the Microplot and Field Scale J. Environ. Qual., May 1, 2008; 37(3): 788 - 797. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Dolliver and S. Gupta Antibiotic Losses in Leaching and Surface Runoff from Manure-Amended Agricultural Land J. Environ. Qual., May 1, 2008; 37(3): 1227 - 1237. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. A.S. Dolliver and S. C. Gupta Antibiotic Losses from Unprotected Manure Stockpiles J. Environ. Qual., May 1, 2008; 37(3): 1238 - 1244. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Dolliver, S. Gupta, and S. Noll Antibiotic Degradation during Manure Composting J. Environ. Qual., May 1, 2008; 37(3): 1245 - 1253. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Burkhardt and C. Stamm Depth Distribution of Sulfonamide Antibiotics in Pore Water of an Undisturbed Loamy Grassland Soil J. Environ. Qual., March 1, 2007; 36(2): 588 - 596. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. G. Davis, C. C. Truman, S. C. Kim, J. C. Ascough II, and K. Carlson Antibiotic Transport via Runoff and Soil Loss J. Environ. Qual., October 27, 2006; 35(6): 2250 - 2260. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Soil Science Society of America Journal | Journal of Plant Registrations | The Plant Genome | |||
) as the ratio of the total irrigation volume. Marker represents the average of the cumulated runoff divided by the irrigation amount within 15 min. Error bars show standard deviation (control 1 d, 3 d: n = 2; manure 1 d: n = 3; manure 3 d: n = 4).
or +). On controls with a 3-d contact time SDZ was clearly detected but lower than LOQ for the qualifier ion (+). LOD, limit of detection; LOQ, limit of quantification.