Indicator Bacteria in Subsurface Drain Water Following Swine Manure Application

Agriculturl nd Biosystems Engineering Conference Proceedings nd Presenttions Agriculturl nd Biosystems Engineering 7-2001 Indictor Bcteri in Subsurfce Drin Wter Following Swine Mnure Appliction E. A. Wrnemuende Iow Stte University Rmeshwr S. Knwr Iow Stte University, rsknwr@istte.edu Jmes L. Bker Iow Stte University Jeffery C. Lorimor Iow Stte University Steven K. Mickelson Iow Stte University, estben@istte.edu See next pge for dditionl uthors Follow this nd dditionl works t: http://lib.dr.istte.edu/be_eng_conf Prt of the Bioresource nd Agriculturl Engineering Commons The complete bibliogrphic informtion for this item cn be found t http://lib.dr.istte.edu/ be_eng_conf/3. For informtion on how to cite this item, plese visit http://lib.dr.istte.edu/ howtocite.html. This Conference Proceeding is brought to you for free nd open ccess by the Agriculturl nd Biosystems Engineering t Iow Stte University Digitl Repository. It hs been ccepted for inclusion in Agriculturl nd Biosystems Engineering Conference Proceedings nd Presenttions by n uthorized dministrtor of Iow Stte University Digitl Repository. For more informtion, plese contct digirep@istte.edu.

Indictor Bcteri in Subsurfce Drin Wter Following Swine Mnure Appliction Abstrct Approprite mnure ppliction rtes, timing, nd methods re necessry to mximize nutrient utiliztion by plnts from mnure, while minimizing wter resource pollution potentil, including tht of enteric orgnisms. A field study nd soil column study exmined the response of indictor bcteril densities in subsurfce drin wter to different swine mnure pplictions. The field study focused on the impcts of different mnure mngement regimes on fecl coliform, fecl streptococcus, nd Escherichi coli (E. coli) densities in subsurfce tile drin wter. Eight swine mnure tretments were compred with control tretment where commercil ure mmonium nitrte ws pplied. Mnure tretments included fll injection, spring injection, nd lte winter brodcst t ppliction rtes of 168 kg N/h nd 336 kg N/h. Results indicted tht the highest incidence of significntly elevted bcteril levels occurred where mnure hd been brodcst in lte winter t rte of 336 kg N/h. Disciplines Bioresource nd Agriculturl Engineering Authors E. A. Wrnemuende, Rmeshwr S. Knwr, Jmes L. Bker, Jeffery C. Lorimor, Steven K. Mickelson, nd Stewrt W. Melvin This conference proceeding is vilble t Iow Stte University Digitl Repository: http://lib.dr.istte.edu/be_eng_conf/3

1 This is not peer-reviewed pper. Pper 012197 INDICATOR BACTERIA IN SUBSURFACE DRAIN WATER FOLLOWING SWINE MANURE APPLICATION E.A. Wrnemuende, R.S. Knwr, J.L. Bker, J.C. Lorimor, S. Mickelson, nd S.W. Melvin Written for presenttion t the 2001 ASAE Annul Interntionl Meeting Sponsored by ASAE Scrmento Convention Center Scrmento, Cliforni, USA July 30-August 1, 2001 Abstrct Approprite mnure ppliction rtes, timing, nd methods re necessry to mximize nutrient utiliztion by plnts from mnure, while minimizing wter resource pollution potentil, including tht of enteric orgnisms. A field study nd soil column study exmined the response of indictor bcteril densities in subsurfce drin wter to different swine mnure pplictions. The field study focused on the impcts of different mnure mngement regimes on fecl coliform, fecl streptococcus, nd Escherichi coli (E. coli) densities in subsurfce tile drin wter. Eight swine mnure tretments were compred with control tretment where commercil ure mmonium nitrte ws pplied. Mnure tretments included fll injection, spring injection, nd lte winter brodcst t ppliction rtes of 168 kg N/h nd 336 kg N/h. Results indicted tht the highest incidence of significntly elevted bcteril levels occurred where mnure hd been brodcst in lte winter t rte of 336 kg N/h. The uthors re solely responsible for the content of this technicl presenttion. The technicl presenttion does not necessrily reflect the officil position of the Americn Society of Agriculturl Engineers (ASAE), nd its printing nd distribution does not constitute n endorsement of views which my be expressed. Technicl presenttions re not subject to the forml peer review process by ASAE editoril committees; therefore, they re not to be presented s refereed publictions. Cittion of this work should stte tht it is from n ASAE meeting pper. EXAMPLE: Author's Lst Nme, Initils. 2001. Title of Presenttion. ASAE Meeting Pper No. xx-xxxx. St. Joseph, Mich.: ASAE. For informtion bout securing permission to reprint or reproduce technicl presenttion, plese contct ASAE t hq@se.org or 616-429-0300 (2950 Niles Rod, St. Joseph, MI 49085-9659 USA).

2 In the soil column study, lechte from intct 20-cm dimeter, 30-cm long soil columns receiving fll nd spring mnure pplictions t 168 kg N/h nd 336 kg N/h ws nlyzed for bcteril densities. Fecl coliform, E. coli, nd enterococci densities in lechte from the columns were determined for four weekly irrigtion events following mnure ppliction. While positive trend between mnure ppliction rte nd bcteril densities in lechte ws observed, this effect ws not generlly sttisticlly significnt t the 10% level. However, n interction between the ppliction rte nd timing ws observed, indicting tht n increse in ppliction rte is more likely to cuse greter increse in bcteril contmintion in subsurfce dringe with spring ppliction thn with fll ppliction. Mnure pplied t 336 kg N/h during the spring often resulted in significntly higher bcteril levels in lechte thn other tretments. Bcteril densities in lechte from fll mnure-pplied soil columns were significntly lower in comprison with bcteril densities in lechte from the spring mnure-pplied soil columns during the second, third, nd fourth irrigtion events. Introduction Where livestock mnure is lnd pplied, the potentil for fecl contmintion of receiving wters exists. Fecl mteril is ssocited with severl pthogens tht pose helth risk to humns. Fecl pthogens tht my become wterborne include: Escherichi coli (E. coli), Slmonell sp., Cmpylobcter sp., Shigell sp., Girdi, nd Cryptosporidium. Becuse it is often difficult nd expensive to detect these pthogenic orgnisms within resonble detection limits, indictor orgnisms re used to detect fecl contmintion nd predict the likelihood of the presence of pthogenic orgnisms. Microbil wter qulity is usully described in terms of common indictor bcteri, such s fecl coliforms, E. coli ( subpopultion of fecl coliforms), fecl streptococci, nd enterococci ( subpopultion of fecl streptococci). Bcteril wter qulity stndrds nd guidelines re most often written in terms of colony forming units (CFU) of fecl coliform or E. coli per 100 ml wter. Bcteril wter qulity determines the suitbility of wter body for both drinking nd recretionl uses. Drinking wter must hve less thn one CFU/100 ml fecl coliforms (zero contmintion in smple volume of 100 ml), nd the mximum llowble limit for fecl coliforms in recretionl wters (limited contct) is 200 CFU/100 ml. Current mnure

3 ppliction guidelines do not prevent the introduction of pthogenic microorgnisms to surfce nd ground wters. Therefore, it is importnt to identify optimum mnure ppliction procedures, which cn minimize bcteril pollution from lnd ppliction while mintining crop yield. Specific mnure ppliction prmeters include ppliction method, timing, nd rte. It is necessry to optimize these ppliction prmeters to mximize mnure benefit, while minimizing the pollution potentil from the use of mnure. This pper will focus on the movement of lnd pplied mnure-borne fecl bcteri to subsurfce drin wter. Specificlly, these studies exmined the impcts of different mnure mngement regimes on four fecl indictor popultions: fecl coliform, Escherichi coli (E. coli), nd enterococci, nd fecl streptococcus, in subsurfce field drin wter nd lechte from intct soil columns. The objective of these studies ws to identify the optimum swine mnure ppliction prmeters in order to minimize impcts to the bcteril qulity of subsurfce drin wter, without diminishing crop yield.

4 Literture Review Microbil Contmintion of Subsurfce Tile Dringe Wter from Field Applictions of Liquid Mnure A field study employed existing tile dringe lines to monitor bcteril leching under norml mngement prctices (Joy et l., 1998). Nldixic cid resistnt E. coli (E. coli NAR) ws used s trcer orgnism to mesure bcteril contmintion of subsurfce dringe reching tile lines following twice-yerly liquid diry mnure ppliction. This study ws performed on lom soil with 4% slope. The field ws plnted in corn nd received conventionl mouldbord plow tillge. The site ws rtificilly drined by 100-mm vitrified cly tile on pproximtely 17-m spcing. Three prllel tile drins were equipped with ccess chmbers consisting of verticl 900-mm plstic pipe. These ccess chmbers were used to obtin wter smples for bcteril nlysis nd quntify flow. Just down-grdient from the ccess chmbers, the tiles drined into surfce ditch. Liquid diry mnure ws inoculted with E. coli NAR nd pplied in spring before plnting nd in fll following hrvest. Mnure ws spred using deflector on the outlet nozzle of the tnker t rte of 56,000 L/h nd hd n initil E. coli NAR density of 7.6 X 10 3 to 1.3 X 10 5 CFU/g mnure. Densities of E. coli NAR in tile wter were determined before, during nd fter mnure pplictions. Smples were obtined from ech tile line ccess chmber, t the point of dischrge into the surfce ditch, nd up nd down strem from the dischrge point. Ech smple ws 200 ml. Smples were nlyzed using membrne filtrtion techniques nd mtec-na gr. At no time ws E. coli NAR detected in the receiving ditch upstrem of the dischrge point, but E. coli NAR ws detected downstrem of the dischrge point. Pek levels of E. coli NAR in tile wter rnged from 1 to 1400 CFU/100mL nd frequently exceeded 100 CFU/100mL. Actul mnure ppliction rte did not ffect E. coli NAR densities in drin wter. Leching of Escherichi coli O157:H7 in Diverse Soils under Vrious Agriculturl Mngement Prctices

5 In study by Gglirdi nd Krns (1999), soil cores receiving rinfll simultion were used to observe leching of E. Coli O157:H7 through three soils under different mnure nd tillge prctices. Bcteril densities in soil nd lechte from intct (simulting no tillge) nd disturbed (simulting tillge) soil cores receiving surfce inocultion of E. Coli O157:H7 with nd without surfce mnure ppliction were compred for three different soils. Soil cores were collected from cly lom, silt lom, nd sndy lom. Intct cores were collected in 177.5-mm sections of 102-mm (inside dimeter) polyvinyl chloride (PVC) pipe. The pipe ws beveled t 45 o on the downwrd fcing edge. A steel plte ws plced over the upwrd fcing edge nd struck with hmmer until 152.5 mm of the pipe ws driven into the soil. Ech core ws then excvted by removing the surrounding soil nd cutting the bottom of the soil core clenly from the remining soil surfce. Ech core ws plced on filter pper inside 133-mm Buchner funnel nd seled in plce. Disturbed cores were creted by plcing cylinders of the sme dimension nd luminum flshing construction onto filter pper inside 133-mm Buchner funnels nd filling the cylinders with soil tht hd been ir dried nd sieved through 5-mm mesh. Soil ws dded nd tmped in 15 mm increments until soil depth equled 152.5 mm. All cores were sturted from the bottom up by drwing reverse osmosis wter through cpillry ction from the bottom of the core until the soil surfce ws wet. Disturbed cores were then removed from the luminum cylinders nd encsed in cement on the verticl sides, eliminting the potentil for preferentil wllflow. The inoculum in this cse ws rifmycin-resistnt derivtive of E. Coli O157:H7 strin B6914 contining plsmid pgfp with genes for green fluorescent protein nd mpicillin resistnce. Orgnism density in inoculum rnged from 3.015 X 10 7 CFU/mL. Mnured cores received 1 ml of inoculum in 50 g of mnure nd non-mnured cores received 1 ml of inoculum directly to the soil surfce. Mnure nd inoculum were spred evenly over the entire core nd rinfll simultions begn immeditely. Eight rinfll simultions of 25.4 mm over 4-hour period were performed on ll soil cores. Simultions occurred dily for the first four dys, nd then every three to four dys therefter. Lechte ws collected in sterile bekers once for ech rinfll, fter grvity

6 dringe hd cesed. Soil smples were collected fter simultion t three evenly spced depths nd from the mnure lyer, if present. Smples were diluted using n isotonic slinephosphte buffer nd plted onto selective gr. After incubtion, green glowing colonies were counted under long-wvelength UV light source. E. Coli O157:H7 ws detected in lechte from ll soils t ll smpling times over the 18 dys of the experiment. Densities rnged from 10 5 to 10 8 CFU on the dy of inocultion nd from 10 4 to 10 6 CFU fter 18 dys. Intct cly cores clogged by the third dy. The number of CFU E. Coli O157:H7 recovered rnged from 0.64 to 30.97 times more thn CFU E. Coli O157:H7 inoculted, with only one tretment (intct sndy lom, no mnure) hving fewer CFU E. Coli O157:H7 recovered thn inoculted. E. Coli O157:H7 generlly replicted better without mnure in disturbed cores nd with mnure for undisturbed cores. Since nutrients were ble to lech fster from disturbed cores, it is believed tht E. Coli O157:H7 hd to compete more with soil microflor for vilble nutrients, especilly where mnure ws pplied. Intct cores, where microsites remined, llowed E. Coli O157:H7 to void excess predtion nd competition. However, there ws no significnt difference between the till nd no-till tretments t the 5% level. Movement of Bcteri in Unsturted Soil Columns with Mcropores In 1998 study by Abu-Ashour nd ssocites, repcked soil columns were used to evlute the effects of mcropores, soil type, tillge, rinfll simultion, nd initil soil moisture on bcteril leching. Investigtors used trcer orgnism, nlidixic cid-resistnt E. coli (E. coli NAR) to monitor bcteril movement through the soil columns. The soils used in this study were silt lom hving n orgnic mtter content of 4.7% nd lom hving n orgnic mtter content of 0.9%. Soils were mechniclly seprted, ir-dried, sieved to 2 mm, nd mixed prior to soil column pcking. Desired initil wter contents were chieved through drying, wetting, nd mixing. One mcropore ws creted in ech desired column using 2.4-mm dimeter rod held verticlly in plce during soil column pcking, nd then removed, leving void with the pproximte dimensions of typicl mcropore. Ech 400-mm X 89-mm dimeter plexiglss soil columns ws pcked to soil depth of 175 mm nd dry bulk density of 1.2 g/cm 3. Soils were compcted in 87.5-mm

7 increments, in order to ensure uniform compction. At the bottom of ech soil column, perforted plexiglss disc overlid with 38-µm stinless steel mesh provided support. Two ports in the bottom provided ccess for lechte smpling nd dringe dischrge. A 6-mm depth of E. coli NAR (10 6 to 10 10 CFU/mL) inoculte ws pplied to the surfce of ech column nd outflow ws quntified nd nlyzed for E. coli NAR. A 10-mm 2-hour rinfll simultion ws performed on some columns 24 hours fter inocultion nd lechte ws gin quntified nd nlyzed. Results indicted tht columns without mcropore retined high percentge of inoculte in the top 2 cm of soil, wheres columns with mcropore lost up to 83% of pplied E. coli NAR in lechte. No E. coli NAR ws found in lechte from soil columns without mcropores. Higher initil soil moisture (40-41%) excerbted bcteril leching, especilly where mcropore ws present. No lechte ws recovered from low initil moisture columns (9-10%). In columns where the top 2 cm ws reworked to destroy the mcropore inlet, downwrd dringe ws retrded nd bcteril migrtion ws reduced. This effect ws more noticeble on wet soils thn on dry. Soil type did not notbly influence bcteril leching in unsturted soil. FIELD STUDY Experimentl Site Description Methodology The experimentl site ws locted t the Iow Stte University s Agronomy nd Agriculturl Engineering Reserch Center west of Ames, Iow on Clrion lom soil, in the Clrion Nicollet Webster Soil Assocition. The soil is generlly well drined nd suited to cultivted crops. The bulk density of on-site Clrion lom is pproximtely 1.4 g/cm 3. The re receives n nnul verge of 82.5 cm of precipittion, with bout 55.0 cm occurring during the spring nd summer months.

8 Experimentl Tretments Eight mnure tretments were compred with commercil N tretment of 168 kg N/h s liquid ure mmonium nitrte (UAN) (Figure 1). Becuse producer is likely to fertilize regrdless of mnure supply, this experimentl design ws chosen in order to provide the bsis for comprison between mnure tretments nd the lterntive commercil N tretment (used s check tretment). Experimentl tretments were divided into three ppliction schedules: fll, lte winter, nd spring. Mnure ws injected in the fll using the stndrd injection nd new slot injection methods. In lte winter, mnure ws brodcst onto frozen ground. In the spring, mnure ws injected using the stndrd injection method. For ech mnure ppliction, recommended ppliction rte of 168 kg N/h ws compred to double ppliction rte of 336 kg N/h. Liquid UAN ws incorported on the commercil plots in the spring t the time of plnting. Tretments Timing Fll Lte Winter Spring Method Inject (stndrd) Inject (new slot) Brodcst Inject (stndrd) Incorporte Rte 168 kg-n/h 336 kg-n/h 168 kg-n/h 336 kg-n/h 168 kg-n/h 336 kg-n/h 168 kg-n/h 336 kg-n/h 168 kg-n/h (UAN) Figure 1. Experimentl tretments on timing, rte, nd method of swine mnure ppliction. Experimentl Lyout The study site ws divided into three experimentl blocks ech hving nine tretment plots (Tble 1), to ccommodte three replictions of the commercil fertilizer tretment nd eight mnure tretments. The resulting 27 individul plots were rrnged in rndomized block design (Figure 2).

9 Tble 1. Experimentl Tretments Tretment Identifiction Tretment Abbrevition Appliction Timing Appliction Method Appliction Rte (kg N/h) Spring UAN CTL Spring (UAN) Incorporte 168 Fll Inject 1X FI1 Fll Inject 168 Fll Inject 2X FI2 Fll Inject 336 Fll New 1X FN1 Fll New Slot Inject 168 Fll New 2X FN2 Fll New Slot Inject 336 Brodcst 1X WB1 Lte Winter Brodcst 168 Brodcst 2X WB2 Lte Winter Brodcst 336 Spring Inject 1X SI1 Spring Inject 168 Spring Inject 2X SI2 Spring Inject 336 Figure 2. Experimentl lyout of plots for 9 tretments nd 3 replictions.

10 Experimentl Plot Design Ech experimentl plot ws selected to be 7.5 meters wide, to ccommodte n nnul rottion of 5 rows of corn in hlf the plot nd 5 rows of bens in the other hlf, nd 22.5 meters long. Ech plot ws equipped with both subsurfce flow nd surfce runoff collection systems (runoff dt re not presented here). Collection systems were grvity fed to the end of ech plot. All plots were surrounded by erthen berms, to void overlnd flow nd subsequent cross contmintion between plots. Figure 3 shows schemtic of the experimentl plot. subsurfce flow collection tile line slotted pipe for runoff collection 15 cm di. 7.5 m 5 rows corn 5 rows bens 22.5 m 10.5 m runoff collection tnk 3.3 m di. subsurfce dringe collection sump 37.5 cm di. Figure 3. Schemtic digrm of n experime ntl plot (eril view). Mnure Appliction Becuse this reserch required degree of ppliction uniformity nd ccurcy bove tht which could be provided by mnure pplictors currently on the mrket, the reserch tem designed nd built new mnure pplictor for this study (Figure 4). The min gol of the pplictor design ws the bility to ccurtely determine the ppliction rte nd volume of mnure pplied to the plots. This ws essentil in order to effectively evlute the surfce nd subsurfce losses of pollutnts from the treted plots. The min components of the pplictor were two cst iron progressive cvity pumps (Roper 71228) with hrd chrome plted lloy internls. These power tke-off (PTO)-driven pumps were chosen for their bilities to meter mnure precisely nd hndle solid prticles up to 2 cm in dimeter. Running t 700 rpm, ech pump conveyed 41 m 3 /hr (180 gpm). Two pumps were used in the finl design. The pumps were mounted on steel chssis with dul

11 Figure 4. Schemtic of mnure pplictor. wlking tndem djustble wheel bse. The frme for this chssis ws constructed with 10- cm (4-inch) squre steel tubing, nd ech xle hd 5400 kg (12,000 lb) lod rting. The wheels were spced on 229-cm (90-in) centers, such tht they could strddle three 76-cm (30-inch) rows. The pumps were supplied with liquid mnure from 3506-L (925-gllon) polyethylene tnk. During ppliction, the mnure ws recirculted for mixing within the tnk using 3729-W (5-HP) trsh pump. Five-centimeter (2-inch) solid hose ws used for this recircultion. Ech pump ws used to supply mnure to one knife t time. Shutoff vlves locted between the pump nd ech shnk ensured tht ech pump supplied mnure to only one shnk t time. Mnure ws supplied to the shutoff vlves using non-collpsible hose nd PVC tubing. Collpsible hose ws used to convey mnure from the shutoff vlves to the ppliction tubing behind ech shnk. Ech shnk nd knife could be rised nd lowered using hydrulic cylinders (20-cm (8-inch) stroke, 8-cm (3-inch) bore, 7.5kP (2500 psi)) driven by the trctor hydrulic system. Liquid swine mnure from the Iow Stte University Swine Nutrition Frm ws obtined using Better-Bilt 12886-L (3400-gllon) vcuum tnk. The liquid mnure ws gitted using 5966-W (8-HP) trsh pump prior to ppliction.

12 For the winter surfce brodcst ppliction, hose connected to "T" fitting on the trsh pump ws used to fill brodcst ppliction tnk during recircultion. A grduted polyethylene spry tnk ws used with Bnjo pump to pply the mnure volume to within n ccurcy of 5 percent. The mnure ws brodcst mnully using the Bnjo pump nd 7.6-cm (3-in) hose with 2.5-cm (1-in) reducer on the end. Mnure ws uniformly brodcst on ech plot until the desired volume, bsed on nitrogen nlysis, ws pplied. This ppliction procedure ws representtive of mngement used by mny producers to void soil compction. It ws completed lte in the winter fter ny snow hd melted, but while the soil ws still frozen. For the inject tretments, mnure ws injected in two psses of ech plot. During the first pss, the middle two knives were fed by the two pumps for pplying the first hlf of the volume for the predetermined ppliction rte. While pplying mnure using the two middle shnks, the outer two shnks were inctive nd in rised position. Afterwrds, the middle shnks were rised nd the outer two shnks were used to pply the second hlf of the mnure during second pss. This order of opertion reduced soil compction nd tire slippge. Grb smples of the liquid (integrted over the time of ppliction) were lso tken s the mnure ws being pplied. One integrted smple ws collected for ech repliction of ech tretment. Mnure hd n verge initil fecl coliform density of 2,500,000 CFU/100 ml. Instrumenttion Subsurfce dringe ws collected through corrugted plstic subsurfce tile lines, which were instlled t 1.2 m depth nd 10.5 m in length into the field (Figure 3). They were positioned in the middle of ech plot nd perpendiculr to the contour. Ech tile line drined into verticl 37.5 cm dimeter PVC collection sump t the end of ech plot. The collection system dimensions were designed to yield representtive smple t mngeble flow volume. Electric sump pumps were instlled in ech collection sump, which operted utomticlly on flot mechnism to pump subsurfce flow from ech plot through n orifice tube. Orifice pltes diverted 0.2 percent of the totl tile flow into 3.78 liter glss

13 smpling bottles. This volume ws used to clculte totl flow volumes for flow-weighted verge bcteril densities. Smpling nd Anlysis Subsurfce drin wter smples were tken weekly strting with the onset of flow in the spring or summer nd continuing through the flow seson ending mid fll. A submersible electric pump ws used to collect smples from ech sump. All subsurfce dringe smples tken for bcteril nlysis were pumped directly from the sump into sterile plstic smple bgs. The pump nd smpling equipment ws flushed by pumping few liters through before obtining smple. Smples were nlyzed within 24 hours nd stored t 4 C until they were nlyzed. Anlysis for fecl coliforms (FC), E. coli, nd fecl streptococcus (FS) were done ccording to the membrne filtrtion technique described in Stndrd Methods for the Exmintion of Wter nd Wstewter, 18th edition, plting on m-fc gr for fecl coliforms, m-coli blue broth for E. coli, nd m-enterococcus gr for fecl streptococcus. All FC, E. coli, nd FS densities were recorded in terms of colony forming units (CFU)/ 100 ml. Results In order to detect significnt differences between tretments, log trnsformtion ws performed on bcteril dt nd lest significnt difference test ws performed on the trnsformed dt with α=0.1. In 1998, fecl bcteril densities differed between tretments, lthough trends were not cler (Tble 2). For instnce, the commercil UAN tretment hd the lowest concentrtion of fecl streptococcus, but not for fecl coliform. One possible explntion for this is tht the two species hve different growth requirements. Fctors, such s ph, moisture, temperture, texture, nutrients, nd mcropores, which influence their growth nd movement my vry sptilly between or within plots. It is lso importnt to note tht nnul verges do not reflect short term spikes in bcteril density. Since optiml growth conditions differ between fecl coliforms nd fecl streptococci, it is not surprising tht popultions did not follow identicl trends. The fll inject single rte tretment hd significntly lower fecl coliform levels in subsurfce dringe thn the fll new inject single

14 rte nd winter brodcst double rte tretments. All other tretments were sttisticlly similr t the 10% level. There were no significnt differences in E. coli or fecl streptococci densities between tretments during 1998. Tble 2. Flow weighted verge bcteril densities in subsurfce drin wter during 1998. Flow-weighted Averge Bcteril Density (CFU/100 ml) Tretment Fecl coliform E. coli Fecl streptococci Spring UAN 8.0b 0.7 21.4b Fll Inject 1X 1.2b 0.4 25.2b Fll Inject 2X 3.9b 0.3 111.4b Fll New 1X 65.0 2.3 503.0 Fll New 2X 3.3b 12.7 161.5b Brodcst 1X 7.1b 1.1 42.0b Brodcst 2X 51.4 1.0 152.1b Spring Inject 1X 31.0b 0.5 71.2b Spring Inject 2X 12.7b 7.1 110.3b Bcteril densities during 1999 re given in Tble 3. In mny cses, verge bcteril densities were dominted by single isolted spike in bcteril density. For this reson, it is difficult to correlte nnul verge bcteril densities with tretment. Fecl coliform nd fecl streptococci densities declined slightly overll during 1999, while E. coli densities were similr to those observed during 1998. The double-rte winter brodcst tretment resulted in significntly higher E. coli densities thn the other tretments nd significntly higher fecl streptococci densities thn ll but the fll slot inject double rte tretment, which yielded unexpectedly high fecl streptococci levels. All other tretments were sttisticlly similr t the 10% level. Over the 2-yr durtion of this study, the highest incidence of significntly elevted bcteril levels occurred where mnure hd been brodcst in lte winter t rte of 336 kg N/h, nd where mnure hd been slot injected in fll t 168 kg N/h.

15 Tble 3. Flow weighted verge bcteril densities in subsurfce drin wter during 1999. Flow-weighted Averge Bcteril Density (CFU/100 ml) Tretment Fecl coliform E. coli Fecl streptococci Spring UAN 11.7 0.6b 51.6b Fll Inject 1X 2.0 0.5b 34.2b Fll Inject 2X 4.7 0.8b 33.9b Fll New 1X 33.5 1.0b 38.0b Fll New 2X 12.9 2.0b 147.4b Brodcst 1X 9.6 2.3b 96.3b Brodcst 2X 10.9 15.1 231.7 Spring Inject 1X 35.3 6.3b 44.6b Spring Inject 2X 1.0 1.8b 101.4b The effects of ppliction timing (Tbles 4 nd 5) nd rte (Tbles 6 nd 7) were not significnt t the 10% level. Difficulty mintining biologicl control in the field setting is believed to hve contributed to vribility in the dt set. However, ppliction method (Tbles 8 nd 9) hd some effect. Annul flow-weighted verge fecl streptococcus densities in subsurfce dringe wter were significntly higher where mnure hd been brodcst rther thn injected during the second yer of this study. A possible explntion for this is tht the vilbility of nutrients in the soil nd wter is prmount to bcteril survivl. Bcteri present in mnure generlly hve ccess to high nutrient supply. Enteric orgnisms do not redily dpt to the lower nutrient vilbility in the soil environment postppliction (Klein nd Csid, 1967), nd this contributes to die-off. This effect would be minimized in surfce ppliction. Tble 4. Flow weighted verge bcteril densities in subsurfce drin wter during 1998, ccording to time of ppliction. Flow-weighted Averge Bcteril Density (CFU/100 ml) Timing Fecl coliform E. coli Fecl streptococci Fll 18.4 3.9 200.3 Spring 21.8 3.7 90.8

16 Tble 5. Flow weighted verge bcteril densities in subsurfce drin wter during 1999, ccording to time of ppliction. Flow-weighted Averge Bcteril Density (CFU/100 ml) Timing Fecl coliform E. coli Fecl streptococci Fll 13.3 1.1 63.4 Spring 18.2 4.1 73.0 Tble 6. Flow weighted verge bcteril densities in subsurfce drin wter during 1998, ccording to rte of ppliction. Flow-weighted Averge Bcteril Density (CFU/100 ml) Rte Fecl coliform E. coli Fecl streptococci 168 kg N/h 26.0 1.0 20.0 336 kg N/h 17.8 5.3 7.4 Tble 7. Flow weighted verge bcteril densities in subsurfce drin wter during 1999, ccording to rte of ppliction. Flow-weighted Averge Bcteril Density (CFU/100 ml) Rte Fecl coliform E. coli Fecl streptococci 168 kg N/h 20.1 2.5 53.3 336 kg N/h 7.4 4.9 128.6 Tble 8. Flow weighted verge bcteril densities in subsurfce drin wter during 1998, ccording to method of ppliction. Flow-weighted Averge Bcteril Density (CFU/100 ml) Method Fecl coliform E. coli Fecl streptococci Inject 19.5 3.9 163.8 Brodcst 29.3 1.1 97.1 Tble 9. Flow weighted verge bcteril densities in subsurfce drin wter during 1999, ccording to method of ppliction. Flow-weighted Averge Bcteril Density (CFU/100 ml) Method Fecl coliform E. coli Fecl streptococci Inject 14.9 2.1 66.6b Brodcst 10.2 8.7 164.0

17 Conclusions Becuse of the higher incidence of elevted bcteril levels in subsurfce drin wter below plots receiving winter brodcst mnure ppliction t 336 kg N/h nd the potentil for bcteri to survive longer in lower tempertures nd in the unincorported condition, this tretment should be voided. Becuse brodcst mnure resulted in significntly higher bcteril densities in subsurfce drin wter nd mcropores, which re not disturbed during brodcst opertions, re the min bcteril leching pthwy, mnure should be injected rther thn surfce brodcst. The new slot inject system did not result in lower bcteril contmintion in subsurfce drin wter. However, the higher incidence of significntly elevted bcteril contmintion in the single rte plot thn in the double rte plot suggests the possibility of tile line contmintion. Difficulty mintining biologicl control in the field setting my hve contributed to this vribility nd flsely high bcteril levels. A lbortory study utilizing intct soil columns to monitor fecl bcteril leching under these mnure tretments nd biologiclly controlled conditions is indicted. SOIL COLUMN STUDY Methodology Eighteen soil columns were collected from the Iow Stte University Agronomy nd Agriculturl Engineering Reserch center ner Ames, IA in order to ccommodte three replictions of four mnure tretments nd two control tretments. Soil column tretments re listed in Tble 10. The soil ws Clrion lom in nnul corn nd soyben rottion. Soil columns were extrcted in lte fll, fter the 1999 soyben hrvest, using Giddings probe nd 20-cm bit dpter. The 30-cm long columns were extrcted in 38-cm long sections of sterilized glvnized tubing tht hd been shrpened on the down - fcing edge. In order to detect compction, the verticl distnce between the top edge of the column nd the inside soil surfce ws mesured nd compred to the verticl distnce between the top edge of the

18 column nd the outside soil surfce, prior to extrction of ech soil column. No compction ws detected. Tble 10. Experimentl Tretments. Spring Control Fll Control Spring Inject 1X Spring Inject 2X Fll Inject 1X Fll Inject 2X Not mended Not mended Mnure ppliction t rte of 168kg N/h (150lb N/c) Mnure ppliction t rte of 336kg N/h (300lb N/c) Mnure ppliction t rte of 168kg N/h (150lb N/c) Mnure ppliction t rte of 336kg N/h (300lb N/c) The soil columns were trnsported to growth chmber, simulting the soil temperture t the 10-cm (4-inch) depth during the typicl periods of fll nd spring mnure ppliction. Autoclved screen ws instlled on the bottom of ech column in order to prevent soil loss. The columns were then rrnged in rndom block design in lechte collection pprtus consisting of 25-cm utoclved funnels nd guide tble tht prevented the columns from deviting from the verticl position (Figure 5). They were sturted with 5000 ml of wter nd llowed to drin for 4 dys. After this period, mnure ws incorported to the 10-cm depth. The mnure ws obtined from finishing unit t the Bilslnd Memoril swine frm ner Luther, IA nd ws less thn 7 dys old. Bcteril nlysis reveled fecl coliform density of 2,000,000 CFU/100 ml. The spring soil columns remined under My conditions in the growth chmber following mnure ppliction. The growth chmber temperture ws set to reflect the verge dily minimum nd mximum soil temperture fluctutions t the 10-cm depth, using ten-yer verge from dt collected t the experimentl site from which the columns were extrcted. The temperture regime is illustrted in Figure 6.

19 Figure 5. Soil columns in the lechte collection pprtus. 80 75 70 Mximum Minimum 27 21 Temperture (F) 65 60 55 50 16 10 Temperture (C) 45 40 My 1 My 8 My 15 My 22 My 29 June 5 4 Figure 6. Averge dily soil temperture t the 10 cm (4 inch) depth, Ames IA.

20 Soil temperture ws chosen over ir temperture for the growth chmber progrm becuse of the semi-exposed condition of the soil columns, which is in contrst to the less exposed condition of similr soil profile in situ. Buffering of ir temperture fluctutions, which significntly ffects soil temperture t depth, ws built in to the growth chmber temperture progrm by setting the growth chmber ir temperture equl to the verge dily soil temperture t the 10 cm depth. In the growth chmber, the verge dily minimum soil temperture occurred during 12 hours of drkness nd ws followed by 12 hours of the verge dily mximum soil temperture during 12 hours of light. Six dys fter mnure ppliction, the first of four irrigtion events took plce on the spring columns. Wter ws irrigted to ponding depth of 5.3 cm (volume = 1700 ml), which is typicl weekly rinfll mount for the first week in My. Weekly rinfll depths were bsed on weekly rinfll dt nd irrigted in single event in order to produce the effects of mcropore flow nd yield enough lechte to perform bcteril nlyses. The lechte ws collected in sterile plstic smple bottles nd nlyzed for fecl coliform, E. coli, nd enterococci using Stndrd Methods 9222D, 9222G, nd 9230C, respectively. This process ws repeted for the second, third, nd fourth irrigtion events. Ponding depth for these events ws 3.7 cm (volume = 1200 ml), 3.4 cm (volume = 1100 ml), nd 3.4 cm (volume = 1100 ml), respectively. Outflow ws quntified in order to provide dt necessry to complete wter budgets on ech column. Averge outflows between tretments were similr. Six dys fter mnure ppliction, fll soil columns were seled nd trnsported to freezer, where they remined for 7 weeks, to simulte over-winter conditions of below freezing tempertures nd snow cover, nd to produce the cell chnges ssocited with freezing nd thwing. After this period, they were trnsported to growth chmber simulting the sme time period s the spring columns. According to field dt, this is the period during which bcteril leching occurs on fll-mnured plots s well s spring mnured plots. Irrigtion events on the fll soil columns begn 2 dys fter trnsport to the growth chmber. The depth nd timing of fll soil column irrigtion events were the sme s the depth nd timing of spring column irrigtion events.

21 A mss evlution ws performed on three representtive soil columns. Prior to ech irrigtion event, these columns were weighed. The mss of outflow ws monitored using volumetric nlysis of lechte smples. The mss dt were used in conjunction with moisture nlysis of the columns fter the completion of the study in order to model the wter budget for ech column. Mss dt re given in Tble 11. Tble 11. Mss blnce for three representtive fll soil columns. Fll control Fll 2X Fll 2X replicte 2 replicte 1 replicte 2 mss (kg) mss (kg) mss (kg) Prior to Event 1 17.91 19.01 17.90 Irrigtion 1.70 1.70 1.70 Dringe wter 0.82 0.78 0.79 Evportion (clculted) 0.67 0.77 0.64 Prior to Event 2 18.12 19.16 18.17 Irrigtion 1.20 1.20 1.20 Dringe wter 0.43 0.42 0.55 Evportion (clculted) 0.71 1.00 0.83 5/10/1999 5/12/1999 5/18/1999 6/7/1999 Prior to Event 3 18.18 18.94 17.99 Irrigtion 1.10 1.10 1.10 Dringe wter 0.43 0.12 0.30 Evportion (clculted) 1.06 1.62 0.71 Prior to Event 4 17.8 18.3 18.08 Irrigtion 1.10 1.10 1.10 Dringe wter 0.36 0.10 0.13 Results nd Discussion Bcteril densities in soil column lechte from irrigtion events 1 through 4 re given in Figures 7 through 10. Generlly, the double rte mnure tretment resulted in slightly higher bcteril densities in soil column lechte. This difference becme more significnt with successive irrigtion events becuse of the higher orgnic mtter present in double rte columns. Orgnic mtter my minimize the stress of between event drying on bcteri. The ppliction rte effect ws sttisticlly significnt t the 10% level for enterococci during event 3. The fll columns yielded similr bcteril densities s the spring columns for event one, nd lower bcteril densities for events 2, 3, nd 4. The ppliction timing effect ws significnt t the 10% level during events 3 nd 4 for fecl coliform, during events 2 nd 3 for E. coli, nd during event 2 for enterococci. This pttern of diverging fll

22 nd spring lechte bcteril densities over time my be the result of decresed vitlity of the fll bcteri due to the freeze-thw cycle. The higher orgnic mtter vilble to bcteri in the double rte columns contributed positively to the survivl of bcteri, prticulrly the fll bcteri. An interction between rte nd timing interction ws significnt for fecl coliform during event 4 nd E. coli during events 3 nd 4. While bcteril densities were higher in lechte from double rte mnure columns during event 1, no significnt differences between tretments were detected during this event. However, bcteril densities from control columns were significntly lower thn in mnured columns, with the exception of enterococci in lechte from the spring control column. Enterococci hve high degree of survivbility in the soil. For this reson, contmintion effects of wildlife ctivity or mnure trnsporttion on the soil column extrction site prior to soil column extrction would be most visible nd most persistent in enterococci densities. With the exception of enterococci, bcteri were not detected in the control columns fter the first irrigtion event, nd lechte bcteril densities from the control columns were lwys significntly less thn those from mnure treted columns. Event 2 resulted in higher bcteril densities in lechte from columns receiving double mnure ppliction rte, lthough this difference ws not significnt. The effect of timing ws significnt however, with E. coli nd enterococci densities significntly lower in fll column lechte thn in spring column lechte. E. coli densities in lechte from spring columns were significntly lower thn those in lechte from the fll double rte columns. Bcteril qulity of lechte resulting from event 3 ws significntly influenced by both timing nd rte, nd ws poorest mong the spring double rte columns. Spring double rte columns resulted in significntly higher E. coli densities thn fll single nd double rte columns. Spring single rte columns resulted in significntly higher E. coli densities in lechte thn fll single rte columns. Other differences between tretments were evident, lthough not sttisticlly significnt t the 10% level. During event 4, spring double rte columns continued to result in the poorest qulity lechte. This tretment resulted in fecl coliform densities in lechte significntly higher thn ll other tretments, nd E. coli densities higher thn spring single rte nd fll double rte tretments.

1000000 23 spring control 100000 fll control spring mnure 1X CFU/100ml CFU/100 ml 10000 1000 100 fll mnure 1X spring mnure 2X fll mnure 2X 10 1 b b b b b Fecl Coliform E.Coli Enterococci Figure 7. Bcteril densities in soil column lechte from event 1. 1000000 CFU/100 CFU/100ml ml 100000 10000 1000 100 b b spring control fll control spring mnure 1X fll mnure 1X spring mnure 2X fll mnure 2X b b b 10 1 b b c c b Fecl Coliform E.Coli Enterococci Figure 8. Bcteril densities in soil column lechte from event 2.

24 1000000 CFU/100ml ml 100000 10000 1000 100 b c bc spring control fll control spring mnure 1X fll mnure 1X spring mnure 2X fll mnure 2X 10 b b b 1 b b d d b b Fecl Coliform E.Coli Enterococci Figure 9. Bcteril densities in soil column lechte from event 3. 1000000 spring control CFU/100 CFU/100ml ml 100000 10000 1000 100 b b b b b b fll control spring mnure 1X fll mnure 1X spring mnure 2X fll mnure 2X 10 1 c c c c Fecl Coliform E.Coli Enterococci Figure 10. Bcteril densities in spring soil column lechte from event 4.

25 The response of bcteril densities in lechte to successive irrigtion events is illustrted in Figures 11 through 13. 1000000 100000 10000 CFU/100 CFU/100ml ml 1000 spring control 100 spring mnure 1X spring mnure 2X fll control 10 fll mnure 1X fll mnure 2X 1 Event 1 Event 2 Event 3 Event 4 Figure 11. Fecl coliform density in soil column lechte. 1000000 100000 10000 CFU/100ml CFU/100 ml 1000 100 10 1 spring control spring mnure 1X spring mnure 2X fll control fll mnure 1X fll mnure 2X Event 1 Event 2 Event 3 Event 4 u Figure 12. E. coli density in soil column lechte.

26 1000 spring control spring mnure 1X spring mnure 2X fll control fll mnure 1X fll mnure 2X 100 CFU/100 CFU/100ml ml 10 1 Event 1 Event 2 Event 3 Event 4 Figure 13. Enterococci density in soil column lechte. It is expected tht fecl coliform densities follow similr pttern to E. coli densities, since E. coli is subset of fecl coliforms. Enterococci re unrelted enteric orgnisms however, with higher degree of survivbility in the soil. This my explin the different pttern of enterococci levels over time nd bckground levels of enterococci in control columns, which received no mnure ppliction. A fster decline in lechte bcteril densities in single rte tretments cn be clerly observed in Figures 11 through 13. Fluctutions in soil column grvimetric moisture content re believed to hve been the mjor fctor contributing to bcteril die-off in this study. These fluctutions cn be observed in Figure 14. It is possible tht more significnt differences resulting from ppliction timing nd rte would be observed under more idel moisture conditions. Sttisticl nlysis of bcteril counts yielded similr results to sttisticl nlysis of bcteril densities. There were no significnt differences in dringe volume between tretments.

27 Grvimentric Soil Moisture Content (%) 30 25 20 15 10 5 Event 1 Event 2 Event 3 Event 4 0 0 100 200 300 400 500 600 Time (hours) Figure 14. Averge grvimetric moisture content of soil columns over time. Conclusions Intct soil columns were used to model the movement of bcteri to subsurfce dringe following fll nd spring swine mnure pplictions t rte of 168 kg N/h nd rte of 336 kg N/h. In lmost every cse, lechte from mnured columns hd significntly higher bcteril densities thn lechte from non-mnured control columns. This suggests tht lnd ppliction of swine mnure is likely to cuse bcteril contmintion of subsurfce drin wter, even t the recommended ppliction rte of 168 kg N/h. Cler differences in bcteril densities were identified between tretments during the second, third, nd fourth irrigtion events following mnure ppliction. Spring ppliction of swine mnure resulted in higher bcteril densities in subsurfce dringe thn fll ppliction during the 5-week period following spring mnure ppliction. Specificlly, the spring 336 kg N/h tretment yielded higher bcteril densities thn other tretments during ll but the first irrigtion event. This suggests tht mnure pplied to the field t rte of 336

28 kg N/h during the spring my contribute significntly more bcteril contmintion to ground wter nd tile dringe thn fll nd spring 168 kg N/h mnure pplictions nd fll 336 kg N/h pplictions. Although few significnt differences were detected between ppliction rtes, the columns tht received 336 kg N/h swine mnure lmost lwys yielded higher bcteril densities in lechte thn the columns tht received 168 kg N/h swine mnure during the sme seson. Additionlly, n interction between the ppliction rte nd timing ws observed, suggesting tht n increse in ppliction rte is more likely to cuse greter bcteril contmintion in subsurfce dringe for spring ppliction thn for fll ppliction. Reserch exmining mnure ppliction impcts to bcteril qulity of drin wter includes field studies utilizing tile drin (Joy et l., 1998), (Culley nd Phillips, 1982), nd lbortory studies using soil columns (Abu-Ashour et l., 1998), soil cores (Gglirdi nd Krns, 2000), or soil blocks (McMurry et l., 1998). Generlly, the lbortory studies hve resulted in more significnt differences between mnure tretments thn the field studies. The results of these studies hve supported this trend. Literture Cited Abu-Ashour, J., D.M. Joy, H. Lee, H.R. Whiteley, nd S. Zelin. 1998. Movement of bcteri in unsturted soil columns with mcropores. Trnsctions of the ASAE. 41(4):1043-1050. Culley, J.L.B. nd P.A. Phillips. 1982. Bcteriologicl qulity of surfce nd subsurfce runoff from mnured sndy cly lom soil. J. Environ. Qul. 11(1):155-158. Gglirdi, J.V. nd J.S. Krns. 2000. Leching of Escherichi coli O157:H7 in diverse soils under vrious griculturl mngement prctices. Appl. Env. Microbiol. 66(3):877-883. Joy, D.M., H. Lee, C.M. Reume, H.R. Whiteley, nd S. Zelin. 1998. Microbil contmintion of subsurfce tile dringe wter from field pplictions of liquid mnure. Cn. Agric. Eng. 40:153-60. Klein, D.A. nd L.E. Csid. 1967. Cn. J. Microboil. 13:1461. McMurry, S.W., M.S. Coyne, nd E. Perfect. 1998. Fecl coliform trnsport through intct soil blocks mended with poultry mnure. J. Environ. Qul. 27:86-92. Smith, M.S., G.W. Thoms, R.E. White, nd D. Ritong. 1985. Trnsport of Escherichi coli through intct nd disturbed soil columns. J. Environ. Qul. 14:87-91.