8-1112 August 1971 ermining e Fule ol Herbicides in the Ogollolu Aquiler .<:$‘ Colorado Oklahoma New Mexico O O J> '1? J> BIG SPRING Fine textured soils Sandy SOllS % I Little or no water O Miles 100 in Ogallala formation "7» Sand dunes A&M University I The Texas Agricultural Experiment Station I H. O. Kunkel, Acting Director, College Station, Texas Contents Summary Introduction Hydrogeology Wells and Equipment Procedure Results Recharge Test Pumping Test r-l v—l@%\TUl>-Pr-PQ9NJ Discussion r-l IQ Acknowledgment Literature Cited r-I l\'.'> r-I Q9 Appendix Mention of a trademark name or a proprietary product does not constitute a guarantee or warranty of the product by The Texas Agricultural Experiment Station or the U. S. Department of Agriculture and does not imply its approval to the exclusion of other products; that may also be suitable. 2 Summary During the fall of 1969, water contain' nitrate at 11.85 parts per million (ppm) N, i at 0.024 ppm, atrazine at 1.28 ppm and K 0.125 ppm was injected into the Ogall_ through a dual-purpose well. Rechargef for l0 days at an average rate of 360 1? minute (gpm). As the chemicals were inj samples were periodically pumped from M located 30, 66 and 150 feet from the d j well to determine herbicide movement in " In the well 30 feet from the dual-p 5 the concentration of nitrate, picloram .-= reached the injection concentration in 56 to Quantitative analyses for trifluralin co interpreted because of unexpected a y sample containers. After l0 days of rech __ centrations of nitrate, picloram and atr A well 66 feet away were the same as in i water. None of the compounds were det well 150 feet from the dual-purpose well; Ten days after recharging stopped, f purpose well was pumped for 9 days at Ii 3 days at 480 gpm, at which time the nit Q tration of the pumped water equaled the nitrate level of the ground water. Whef ended, the water did not contain herbii tectable quantities, and over 90 percent of l herbicides had been’ recovered. ‘ These results indicate that a dual-p, in a sand aquifer that is accidentally c i, by herbicides will not be a serious hazard? is pumped soon after recharging. Since def wells are normally used for seasonal re t ‘ pumping, the injected water or herbicidesf move very far before pumping began. A“ a well that was contaminated with a herb be used to irrigate a tolerant crop. i a Determining ‘UIFERS ARE RECHARGED by natural or arti- rocesses, they may receive runoff from ‘l lands. In the Texas. High Plains, dual- lls are being used 0n a limited scale to ya water to the Ogallala aquifer. With A re, any contaminant in the playa water directly into the aquifer, thereby creating {tllution hazard. Zwater may contain small quantities of chemicals such as herbicides and insecti- " - et al. (9) analyzed the runoff water from 4 in the Texas High Plains that were t- atrazine, propazine and trifluralin. The ‘ concentrations of the three herbicides .23 and 0.04 ppm, respectively. Other {have found no herbicides or insecticides water, but the playa sediments contained in, aldrin and DDT. Most of the pesticide e attributed to mosquito control in the ‘r than to runoff from agricultural lands. it of the fate of contaminants that enter agricultural engineer, professor, soil scientist and g scientist, USDA Southwestern Great Plains Re- at Bushland. A The Fate oi Herbicides in the Ogallala Aquifer A. D. SCHNEIDER, A. F. WIESE, O. R. Jonrs AND A. C. MATHERS* a recharge well is important in determining quality standards fo-r recharge water. Some materials are filtered or adsorbed by the aquifer within a short distance of the well; others move through the aquifer with the recharge water (3, 7, 8). In the aquifer, degradation of pollutants without sunlight and under anaerobic conditions may be completely different from degradation in surface water. When recharged water is recovered from an aquifer by pumping, highly soluble chemicals are usually recovered with the re- charge water, but less soluble chemicals may remain in the aquifer (7, 8). During fall 1969, water containing three com- monly used herbicides and sodium nitrate was injected into the Ogallala aquifer through a dual-purpo-se well located at the USDA Southwestern Great Plains Re- search Center at Bushland. The dual-purpose well was then pumped long enough to recover essentially all of the recharged water. The results of this study are presented and compared with results from. previous ground-water recharge po-llution studies at the Re- search Center. lPersonal communication from D. M. Wells, Texas Tech Univer- sity Water Resources Center. 3 HYDROGEOLOGY The main features of the Ogallala Formation at Bushland, Texas, are shown in Figure 1. The forma- tion consists of fine sand and calcium carbonate with some silt and clay. The calcium carbonate occurs pri- marily in nodules and in lenses. The D50 sand size is usually less than 0.25 millimeters (mm), and the uni- formity coefficient varies from about 1.5 to 3.5. (D10, D50 and D60 are the particle sizes in a granular ma- terial such that l0, 50 and 60 percent, respectively, of the material is smaller. The uniformity coefficient is the ratio of the D60 size to the Du, size.) The caprock above the Ogallala Formation consists of indurated calcium carbo-nate about 2 feet thick, and the Triassic Redbed below the formation consists of dense red clay. The permeability of the formation varies little in the horizontal direction, but it varies considerably between different layers in the vertical direction. The estimated coefficient of transmissibility of the aquifer, derived from pumping and recharge tests, is about 15,000 to 20,000 gallons per day per foot (gpd/ft) The specific yield is probably 0.20 or greater. This specific yield is based on water yield- time curves obtained with a neutron moisture meter and detention times obtained from well recharge tests with tracers. DEPTH PULLMAN CLAY LOAM-a. lg” ,1 .A-,‘ ~ ;.,| ~ q . _?I‘.‘.'I.’..-..f\\'l \.._\|',_§_ , ~ ,1 11.. \.i._~ I: I - PLEISTOCENE SEDIMENTS - "I '1'!» sl I ' "4' I J- _—-——--‘....._.i-_ l; ggisome cqcoa; 1.}. Ez-"IN NODULES f -_.-.--¢.-_. FINE SAND 240 DB _B Figure 1. The Ogallala Formation at Bushland, Texas. 4 $556) o 100 v 20o ¢ DUAL-P SCALE-FEET o oasenv ._ I10 Figure 2. Plan view of experimental wells and wot; __ tours (MSL) before recharge began on October 28, l9 Ground water contours before the st i are shown in Figure 2. The slope of the is steeper than the average l0 feet-per- ' east slope for the region. Heavy pumping tion to the east and a limited amount of gr . recharge to the west are believed to be i During June and July 1969, about 30 a water was pumped from “Tell 1, and 20 . this water was recharged into an experim” located 1,200 feet to the west. ’ WELLS AND EQUIPMENT The locations of the dual-purpose we tion wells and recharge equipment are‘ Figures 2 and 3. The recharge water from an irrigation well 2,450 feet northwest. and conveyed through underground con and surface aluminum pipe. " Well 1, a dual-purpose well, is equip turbine pump for pumping and injection recharging. This well was drilled 28 inch eter, cased to the water table with 16-inch‘ and screened through the saturated sand- of continuous-slot spiral well screen. Th gravel packed and developed by pumping . The turbine pump has a maximum capaci gpm, about twice the sustained producti the well. Two injection pipes 2 inches 5 are suspended inside the well casing to c _, . . Water from ~ 3500’ Norlhwesl Shelter for ChemlCfll I’ Metering Equipment . 1g Chamber I [/70 we“ 8 ‘ 1 .4-3O Well 1 /‘l\ o0 \ Well 2 Q em \OWell a ‘ll Q Well 4 lan view of the experimental wells used for ground- ing and the recharge equipment. “ground surface to the water table without ent. This equipment was described by '. <81 ‘observation wells used for pumping water shown in Figure 3. Wells 2, 3 and 4 are g screened through the entire aquifer. Well ch well with a 4-foot-long well screen set eet below the water table. Water level re- * e installed in the six 6-inch observation y: at distances greater than 150 feet from ‘propeller meters were placed in series to e recharge and pumping rates. One of EMICALS MIXED WITH THE RECHARGE WATER these meters was connected to a strip recorder that continuously recorded the flow rate. The chemical metering equipment was designed t0 pump 5 gallons per hour (gal /hr) o-f chemical con- centrate into a mixing chamber in the recharge pipe- line. The metering and agitatio-n system consisted of a l50-gallon concentrate tank, an 8-roller nylon pump, a diaphragm pressure regulator, a positive bypass hose and an orifice plate that delivered 5 gal /hr at 30 pounds per square inch (psi). The positive by- pass discharged into the bottom of the l50-gallon tank to keep the nonsoluble chemicals in suspension. The mixing chamber consisted of successive right- and left-hand flight auger sections welded inside the 8-inch. pipe. The mixing chamber was described by Scalf et al. (8). PROCEDURE The herbicides added to the recharge water were picloram, atrazine and trifluralin; sodium nitrate was also added as a tracer. The commercial names and formulations for the chemicals are listed in Table 1. The herbicides are commonly used, or their wide- spread use is anticipated, and they represent a wide range of solubilities. When nitrate and tritium were injected into Well 1 during a previous study, both materials proved to be accurate tracers (7, 8). Recharge began at 10:20 a.m. October 28 and continued for 10 days, with the chemicals being con- tinuously fed into the recharge water. The recharge rate averaged 360 gallons per minute (gpm) with a variation of —l.9 to +1.1 percent. A total of 5.20 million gallons (43.2 million pounds) of water was injected into Well 1. Each day 0.293 gallons (1108 cubic centimeters) of Tordon 22K, 7.5 pounds of AAtrex, 0.028 gallons (106 cc) of Treflan and 300 pounds of sodium nitrate were mixed with 125 gallons of Ogallala ground water. This mixture was trans- ferred to a 300-gallon tank on a field sprayer for further mixing and was then pumped into the con- centration tank that fed the chemical metering equip- ment. Table 1 shows the total amount and average concentration of the chemicals injected into» Well 1. The chemical feed rate varied considerably during the first 4 days because the chemical pump failed, and the flow rate from a new pump had to be adjusted during the test. ' Total Average iWater Commercial added concentration solubility name Formulation Lb - 3m Ppm, 5.42 l 0.125 400,000 Tordon 22K 2 lb/gal potassium salt of picloram 55.5 1.28 70 AAtrex 80% wettable powder 1.04 0.024 <1 Treflan 4 lb/gal emulsifiable concentrate | 440 11.85‘ 730,000 Fertilizer grade NaNOa, 16% N ‘ppm NO3'N as the average level of n-itrate in the ground water. After recharge ended, Well 1 remained idle for l0 days to simulate the likely interval between re- charging and pumping a dual-p-urpo-se well. Begin- ning at 11:00 a.m. November 17, Well 1 was pumped for 9 days at an average rate of 500 gpm and for 3 additional days at. an average rate of 480- gpm. The pumping rate was reduced on the tenth day because the pumping level was near the depth where the turbine pump would. break suction. iDuring the l2 days of pumping, 8.57 million gallons (71.6 million pounds) of water was pumped from Well l. Because the injected water mixed with the ground water and was displaced by the natural gro-und-water flow, a pumping volume 1% times as great as the recharge volume was required to recover most 0-f the injected water. Since herbicide-s may be adsorbed on the aquifer particles, the sand in the water pumped from Well 1 was continually sampled. A tube that diverted l percent of the flow through a settling tank was in- stalled in the top of the pump column pipe. The sand in the settling tank was removed at 24-hour intervals and frozen until the herbicide concentration could be determined. During the» pumping and recharge cycles, water samples were collected at regular intervals from all wells shown on Figure 3. At Well 1 the recharge water samples were withdrawn just before the water entered the injection pipes in the well cas-ing (Figure 4). During pumping, samples were obtained fro-m a faucet in the pump discharge pipe. The 2-inch well" (Well 8) was sampled with a piston pump operated by a pump jack. Before collecting a sample this well was pumped for 10 minutes to insure that the sample came from the aquifer around the well point. After several hours of pumping, the cone of depression dropped below the well po-int; consequently, this well could not be sampled during the remainder of the pumping test. The 6-inch wells were sampled by lowering the portable submersible pump shown in Figure 5 to the center of the saturated f0 pumping for 30 minutes and then coll j sample. This procedure removed at least t‘ volume of static water in the well casing and" that the water sample came from the aquifer. samples taken at 20-foot depth increments j that the single sample accurately represented t from the aquifer around the well. ’ Duplicate water samples were collected i polyethylene containers. One sample was sen laboratory for analyses, and the other was analysis at a later date, if needed. Nitrate and nitrite were determine-d by p‘ mated, colormetri-c procedures of Kamphake, if and Cohen t; Herbicides in the water samples were det with a Barber-Coleman Model 5360 gas chroma equipped with a radium. 226 electron capture p and a 6-foot spiral glass column. The colu tained 10% Dow Corning Silicone Oil 200 krom ABS 80/90 mesh P. The carrier gas wa, fied nitrogen at 20 psi, and the injector, col i detector temperatures were 235°, 190° and a respectively. l - ll\ The analyses f0-r atrazine and triflur developed specifically for the study since siml - te-chniques were essential for the more i. herbicide determinations made. The triflu . _ ysis was similar to that proposed by M. (1.15 I The levels of atrazine and trifluralin inj Well l were selected so that the concentrati equal and maximum peak heights with j retention times on the recorder. The analysj two herbicides consisted of extracting 100 cc for l5 minutes with 100 cc of hexane in an Er flask on a mechanical stirrer. One to 2 min A the stirring ended, the hexane and water l The hexane which rose to the top was carefull 2Personal communication. Privy NrrnATsl-l N Figure 4. Well l and the ‘rech pipeline. f . l Portable submersible pump being Ample Well 3. giconcentrated t0 5 cc. Five micro-liters (nliters) jncentrate was injected into the gas chroma- analysis of picloram was a modification o-f pod proposed by Hall et al. One hundred timeters of acidified water was extracted fies with 30 -cc of ethyl-ether. The ether was fed to dryness, and then 1O cc of 12.5 percent i fluoride solution in methyl alcohol was used f picloram acid to the methyl ester of pic- The methyl ester was washed with l0 cc of dissolved in l0 cc of hexane. Then 5 Mliters i'ected into the gas chromatograph. ldard curves were developed by extracting "th known concentrations of the herbicides. analyses were accurate to about one-tenth Vncentrations injected into Well 1. Duplicate herbicide analyses were made on each sample sent to the laboratory. Only the trifluralin analyses that were made immediately proved to be valid because this herbicide adsorbed to the poly- ethylene containers. The amount of picloram in the samples decreased about 20 percent from the first t0 the second analyses. Consequently, the first analysis was generally used for these two herbicides. The atrazine concentrations in the first and second anal- yses were almost equal. RESULTS The nitrate and herbicide concentrations in the wells during pumping and recharging are shown in Figures 6 to l2. These data are presented in more detail in Appendix. Tables 1 to 7. On the figures showing the recharge data, the curves for Well 2 HOURS AFTER RECHARGE BEGAN , Average Recharge '- we" 1 _> Concentration __ i... D........_....._ _.... _ _ _ _ _ _ _._ AA.__-4 + ‘ n + d i l8 + : \ v i‘ ‘ " h. ‘k Well 2 _ | ‘ \ Well 3 _ NITRATE " Figure 6. Nitrate levels in the wells during i """'“"‘_i' _ recharge. L l l 1 | n 1 | __ 24 4 3 72 96 I20 I44 I68 216 24 O are plotted only until the chemical concentration increased to the average recharge concentration. Since Well 8 was sampled to compare the permeability of a thin sand layer with the average permeability of the aquifer, the data for this well are plotted only for the nitrate tracer during the recharge test. Recharge Test The nitrate concentrations in the observation wells generally followed a normal sigmoidal ‘curve as they increased to the average recharge concentra- tion, Figure 6. Well 2 was an exception because it contained 5.71 ppm nitrate-N after only 3 hours of recharge. This probably was caused by a highly per- meable stratum or a thicker gravel pack just above the water table. The screen of Well 1 extended about 3 feet into dewatered sand, and recharge water could have moved quickly through this material with- out having to displace the ground water. The low chemical concentration in Well 1 during the second and third. days. delayed the time required for the nitrate in Well 2 to reach the average recharge con- centration. In Well 8, the nitrate level began to increase in 12 hours, and it reached the average re- charge concentration in 30 hours. Since the screen in Well 8 is only 4 feet long, this curve represents a thin stratum of the saturated formation rather than the entire aquifer. Small amounts o-f recharged nitrate may have reached Well 3 during the first 3 days, but there was no significant increase until 89 hours. After that, the nitrate level in Well 3 increased slowly and reached the average recharge concentration after about 210 hours. None o-f the nitrate was detected in Well 4, 150 feet from the dual-purpose well. Piclo-ram moved through the aquifer at approxi- mately the same rate as the nitrate tracer (Figure 7). It was first detected in Wells 2 and 3 a.t the same time that the nitrate levels began to increase. In Well 2 the picloram level increased to 0.10 ppm in 6 hours ‘l6 We||1 '-> .15 Average Recharge '14 Concentration .13 _ I ______ __ l‘ J ___ 12 _PPM PICLORAM ' b ‘O and declined to less than 0.06 ppm before A the average recharge concentration after 57, During the last 3 days of recharging, the n. level in Well 3 approached the average concentration, but only one sample exce concentration. This was probably the resul high level of picloram in the injected water r 18 hours. 1 Atrazine also moved. "freely through the '_ with the recharge water and increased to the . recharge concentration in all wells where t I was detected (Figure 8). It was detected at time as the nitrate in Well 2 and about 12 the nitrate in. Well 3. The nitrate and atrazii for Well 2 were similar, but for Well 3the curve was displaced about 24 hours to the the nitrate curve. y Since the trifluralin in the samples f‘ during recharge adsorb-ed to the polyethyli tainers, the data cannot be interpre-ted quan The analyse-s did show, however, that duri days of recharge some trifluralin reached f observation wells where the other herbici detected. Pumping Test A Figure 9 shows the nitrate; concentra‘ Wells 1, 2 and 3 decreasing to the backgro during the pumping test. In Well 3, nitra, to decrease after only 12 hours, but in Wells? it remained nearly constant for 3 days of ning to decrease. The nitrate reached the ba level after 214 hours in Well 3 and during th day in Wells 1 and 2. With a background a 1.66 ppm nitrate-N, the calculated nitrate § was 93 percent. During pumping, the picloram concentrati wells declined until each was essentially f i herbicide (Figure 10). During the first 54 O8 Ob .05 , ‘O4 Figure 7. Picloram levels inithe 1|: ,Q3 p g P|CLQ RAM ing recharge. , ‘~ . ;, .02 ' I .01 o l l l l l I O 24 ' 4 72 9b 120 144 a 168 a 192 21b 240 HOURS AFTER RECHARGE BEGAN 1.7- <_..Well l lb - l.5 - 1.4 > Average Recharge I Concentration 1.3: ..._ ___ 1.2 - 1.1 - 1.0 - ~90 <-Well 3 .70 <—-W6ll 2 .60" _ 50% zine levels in the wells dur- 40r- 30- k i l ATRAZINE PPM ATRAZINE 8 / o S I r “z ' ' “ ‘ ' 15o 1Z4 18a 192 21o 24o HOURS AFTER RECHARGE BEGAN ' l Average Rochatgo / Concentration q NITRATE~ iii i ‘.1 . ‘y 1 .,' PPM - NITRATE - N - n u p u- o- w co ~o 1 a n l | 1 l 1 1 24 48 72 96 I20 144 I68 I92 216 240 264 2B8 HOURS AFTER PUMPING BEGAN o0 Average Recharge / Concentration PICLORAM (I; Z _ PPM PICLQRAM an n ) icloram levels in tile wells dur- o 24 4a 72 9o 12o m we 192 ‘ 21o 2-40 E4 ‘We HQURS AFTER PUMPING BEGAN Ave rage Recha rge Z Concntralion ATRAZINE u; Z '2 o: .90 - 5f .80 - 2 .70 > & .60 - p ~50’ Figure ll. Atrazine levels in the .40- ing pumping. .30 - .20 - J0 - Q l l l l l l l l l l l l 0 24 48 72 96 I20 I44 I68 I92 2l6 240 264 288 HOURS AFTER PUMPING BEGAN pump-ing the picloram in Well 1 exceeded the average recharge concentration. After that, it declined to near the zero level after 220 hours, and at the end of the test, Well 1 was free of picloram. Well 3 con- tained traces of picloram until 216 hours, and Well 2 still contained a trace of picloram at the end of the test. Ninety-three percent of the picloram injected into Well l was recovered in the pumped water. The atrazine level.s in Wells 1, 2 and 3 during the pumping tests are shown in Figure ll. During the first 117 hours of pumping, the atrazine concen- tration in Wells 1 and 2 exceeded the average recharge q concentration. After that, the concentration in Well 1 dropped quickly to 0.25 ppm at 1894 hours and then declined slowly until the end of the test. All three wells, however, contained traces of atrazine at the end “of the test. The measurements indicated that slightlymore- than 100 percent of the injected atrazine was recovered in the pumped water. Because this error is within the accuracy limits of the analyses, it was concluded that essentially all of the atrazine was recovered by pumping. a The trifluralin levels varied erratically during pumping and were difficult to interpret (Figure 12). .07 ' osl .05 - .04 - .03 '- In Wells 2 and 3, trifluralin increased du The j first day and then began to decrease. tration of trifluralin in Well 1 exceeded the. recharge concentration during the first 93 then dropped rapidly to less than 0.005 pp ' hours and remained below this level until the test. All of the wells contained a tra fluralin at the end of the test. The samples , at Well 1 during pumping indicated a i recovery for trifluralin. Part of the variabili, trifluralin data may be due to variation chemical feed rate during recharging. B h fluralin is very insoluble, the flow through t i; cal metering equipment p-robably varied m?“ for the other chemicals. However, accurate t measurements at Well 1 are not available.‘ Q this. a < Trifluralin was the only herbicide that to the sand (Table 2). The trifluralin con * on the sand was high during the first 4 days _ ing, but it decreased to less than 1/10 the recharge concentration by the end of the 660 pounds of sand was pumped during the consequently, the amount of trifluralin recov TRIFLURALIN Average Recharge / Concentration ‘PPM TRIFLURALIN Figure l2. Trifluralin levels in the wells during pumping. a 1O 7'2 9i» 12o 144 we 192 2w 24o HQURS AFTER PUMPlNG BEGAN ll§l.¢.>_ . . 0 PUMPED FROM WELL 1 AND THE CONCENTRATION ON THE SANDI sand Triflurolin Atrozine Piclorom pumped f gums/day Ppm Ppm _Ppm *1 29,300 .0694 1.18 .009 16,400 .1872 .42 .010 24,600 .0716 .40 .006 23,300 .0912 1.30 .018 29,000 .0600 .36 .011 31,300 .0244 .14 .009 U‘ 20,600 .0280 0 .006 ‘ 30,400 .0120 0 .007 19,600 .0106 0 .009 23,800 .0082 0 0 23,100 .0076 0 0 28,900 .0020 0 0 300,300 . entrotions were calculated using the dry weight of as small—0.0l2 gram. The amount of pic- zttrazine 0n the sand never exceeded the ion in the water pumped during the same i‘ *4 ., . g DISCUSSION Jerbicides picloram, atrazine and trifluralin ough the sand aquifer with the recharge i} ’ each was detected in observation wells , and 66 feet from the dual-purpose well. fno injected nitrate orherbicide was de- ell 4 located 150 feet from the dual-purpose j was expected since the volume of ground "n 150 feet of Well 1 was 31/2 times as great j e of injected water. One or more highly jiiilayersl could have caused the injected water ell 4. Since this did not happen, the zone between the injected water and the ter was narrow. L ll itrate curves for Wells 2 and 8 on Figure 6 I} variation in permeability of the aquifer. ifer were homogeneous, the flow-through the 4-foot well point (Well 8) and the fully well (Well 2) would be equal. The flow- ; e for Well 8 was 20 hours, and for Well 2 ‘hours. Thus, the horizontal permeability (I around the well point is greater than the jrizontal permeability o-f the formation. jected water was displaced to the east by ti.» during the 10-day pause, but this did _t recovering most of the water. After 5 it ping, the nitrate concentration was higher than in-Well-jl. This suggests that the the injected water moved toward Well 2 test. A sample calculation sho-ws, how- j the displacement was small in comparison "rough time for a well is equal to the time at the int of the curve. to the distance the injected water moved from Well l. On Figure 2 the ground water contours show an average» water table slope of about 0.004 foot per foot (ft/ft), and a maximum gradient of 0.0054 ft/ft be- tween Wells 4 and 7. The ‘flow velocity calculated from the maximum gradient, the highest estimate of permeability (25 feet per day per foot), and the po- rosity (0.35) is only l1 feet per month. If water that is injected through a well is to be recovered from. the sam.e well, the ratio of the pumped volume to the recharged volume will increase each day pumping is delayed. Two processes are respon- sible for this-displacement of the injected water by natural flow and mixing between the injected water and ground water. After a pause of several weeks to several months, complete recovery through the same well would no longer be practical or possible. In this study essentially all of the herbicide-s were recovered after a 10-day pause by pumping 12/5 times the volume of recharged water. Thus, the herbicides can be recovered through a dual-purpose well as long as the injected water is not displaced away from the well. The allowable pause will depend o-n local condi- tions such as permeability, porosity and the regional ground-water gradient. In the Texas High Plains where playa water is usually available in May and June and pumping for irrigation continues until September, most of the recharged water would be pumped back within a few months. The results with herbicides are similar to those of an earlier study at the Research Center in which nitrate and DDT were pollution parameters, and tritium was the recharge water tracer (7,8). The ni- trate moved through the aquifer at a rate similar to the tritium tracer, but the DDT was very strongly adsorbed to the Ogallala sand. Unlike the herbicides, there was never any breakthrough of DDT to Well 2 located 30 feet from the dual-purpose well. Three hours after the recharge test ended, a pumping test started, and the DDT concentration in the pumped water was 16 times the average recharge concentration. Within 1 hour the DDT level dropped below the average recharge concentration and continued to de- crease for 2 more days. The DDT concentrations then became erratic, and the sampling frequency did not permit accurate calculation of the percent recovery. Ninety-four percent of the nitrate and tritium was recovered with the pump-ed water. The trifluralin adsorbed to the sand near the recharge well, but to a lesser degree than the DDT. This herbicide was released from the formation much faster than it was injected, thus resulting in the steep decline in concentration beginning at about 96 hours (Figure 12). Unlike the DDT, essentially all of the trifluralin was recovered from the Ogallala sand. The movement of coliform bacteria through the Ogallala aquifer was also studied when clarified playa water was recharged into the dual-purpose well and three surrounding wells (2, 3). Initially, the playa ll water was recharged into Well 3 and recovered from the aquifer by pumping the dual-purpose well. Almost all of the coliform bacteria in the playa water were filtered by the fine Ogallala sand. In another phase of the study, clarified playa water was recharged into the dual-purpose well, and short intervals. of pumping did not effectively remove coliform bacteria from the surrounding aquifer. Rebhun and Schwarz (6) also reported coliform bacteria contamination after recharging water of drinking quality into wells in a sandstone aquifer. They concluded that suspended organic matter, even though of low concentrations, was filtered near the recharge wells and formed an organic mat. Shortly after recharge ended, conditions became favorable for decomposition of organic matter and bacterial growth. As a result, water pumped after a pause of 2 to 50 days showed high coliform bacteria counts. Similar results would be expected in any aquifer fine enough to filter the suspended organic mate-rial. This herbicide study and the two previous studies at the Research Center offer some guidelines for detennining quality standards for water recharged through wells. The coliform bacteria and DDT were effectively filtered or adsorbed by the fine Ogallala sand. This means that they are not likely to move very far through Ogallala sand. However, the DDT and coliform bacteria were difficult to recover from the aquifer. The herbicides and nitrate moved readily with the recharge water, but they were easily recovered _ by pumping. ACKNOWLEDGMENT This study was a cooperative effort of The Texas Agricultural Experiment Station, Texas AScM Uni- 12 versity, and the Soil and Water Conservation R Division, Agricultural Research Service, U. S. I ment of Agriculture. A LITERATURE CITED Hall, R. C., C. S. Giam and M. G. Merkle. 1970. technique for the determination of picloram a herbicides containing carboxylic acid and ester Analytical Chemistry, V01; 42, No. 3, pp. 423-425. _ Hauser, V. L. and F. B. Lotspeich. Treatment a lake water for recharge through wells. Transacti ASAE, Vol. ll, No. 1, pp._108-l11, 1968. Jones, Ordie R. Movement of coliform bacteria an carbon in the Ogallala aquifer at Bushland, Tex‘ Agr. Exp. Sta. MP-873, February 1968. Jones, O. R. and A. D. Schneider. Comparison of for determining the specific yield of the Ogallala. the Ogallala Aquifer Symposium, Texas Tech U Lubbock, Texas. pp. 118-130, April 30-May 1, '. Kamphake, L. 1., s. A. Hannah and J. M. Cohe mated analysis for nitrate by hydrazine reductio R. A. Taft Sanitary Engineering Center, Cincinna " Rebhun, M. and J. Schwarz. Clogging and con i processes in recharge wells. Water Resources V01. 4, N0. 6, pp. 1207-1217, December 1968. Scalf, M. R., W. J. Dunlap, L. G. McMillion i Keeley. Movement of DDT and nitrate durin water recharge. Water Resources Research, Vol. pp. 1041-1051, October 1969. i‘ Scalf, M. R., v. L. Hauser, L. o. McMillion, W. i and J. W. Keeley. Fate of DDT and nitrate water. Joint report of U. S. Dept. of Interior; Kerr Water Research Center, Ada, Oklahoma, a Dept. of Agriculture, Southwestern Great Plains v Bushland, Texas, April 1968. 7 Wiese, A. F., D. T. Smith and A. D. Schneider. A residue research on the High Plains. Proc. of Conference on Insect, Plant Disease, Weed and B; trol, pp. 149-154, Dec. 15-17, 1970. Texas A&M i, APPENDIX HERBICIDE AND NITRATE CONCENTRATIONS IN APPENDIX TABLE 2. HERBICIDE AND NITRATE CONCENTRATIONS KECHARGE IN WELL 2 DURING RECHARGE Hou-rs offer NOs-N Picloram Atruzine recharge began NOa-N Piclorczm Atrcrzine PP": Ppm 3pm. Ppm 5P1". is"; 12.03 .134 1.48 0.2 1.45 .016 0.67 13.19 .147 1.48 3 5.71 .068 .80 13.11 _.098 1.29 6 5.27 .027 I .87 12.56 .130 1.28 9 5.27 .066 1.07 12.41 1.26 12 5.04 .017 .82 12.48 .116 1.29 15 4.93 .055 .76 12.33 , 1.33 18 4.88 .077 .76 12.10 .159 1.52 21 4.83 .070 2.74 11.95 .151 1.49 24 .117 .85 11.95 .160 1.30 27 ' 5.27 .095, .84 11.44 .156 1.37 3O 6.22 .099 ".46 11.58 1.32 3 33 6.69 .090 A .36 11.36 .141 .93 36 8.08 .107 .46 11.00 .129 .80 42 9.49 .111 .68 10.03 ‘ .128 .81 48 10.37 .124 ".55 9,83 ,123 95 54 _ 10.16 .128 1.01 .129 .95 6O 9.89 .011 .90 9,10 ,098 ,32 66 10.30 .125 .89 9,49 _134 1,13 72 9.89 .121 1.07 12,72 J68 1,74 78 11.29 .142 1.19 15,05 _145 1,67 84 11.00 .121 1.21 13.93 _ .137 1.61 96 11.80 .110 1.25 12,80 _134 1,52 107 11.88 .134 "1.25 1103 J55 L64 119 11.80 .130 1.30 ' = ~11_44 _124 1,09 130 11.88 I .126 1.27 10,36 ,123 1,13 147 12.48 .132 1.14 12,03 ,111 1,37 157 12.18 .104 1.24 13,03 ,129 1_30 168 12.03 .135 1.22 12,25 _094 174 11.95 .128 1.16 12,87 _124 1_33 190 12.80 .095 1.25 11,33 _117 L32 195 12.80 .112 1.27 Y 12,37 093 1_31 216 13.03 .108 1.25 1 12.87 .134 1.33 222 13.03 13.03 . 238 12.80 .138 1.31 12.56 .111 1.39 APPENDIX TABLE 3. HERBICIDE AND NITRATE CONCENTRATIONS IN WELL 3 DURING RECHARGE Hours after recharge begun NOs-N Piclorum Atrazine Ppm Ppm_ Ppm 5 1.54 0 0 11 1.40 0 0 17 1.30 0 0 23 1.30 0 0 29 1.45 0 0 35 1.45 0 0 41 1.45 0 0 47 1.40 0 0 53 1.30 0 0 59 1.30 0 0 65 1.26 0 0 71 1.26 0 0 77 1.26 0 0 83 1.35 0 0 89 1.74 .008 0 95 2.23 .013 .03 105 3.45 .024 .02 120 4.72 .049 .10 129 5.94 .061 .24 147 8.14 .102 .40 156 8.71 .104 .58 167 9.62 .092 .89 175 10.03 .090 .93 I91 10.86 .135 1.00 I97 .101 215 11.88 .111 221 11.95 239 12.10 .108 1.34 13 APPENDIX TABLE 4. HERBICIDE AND NITRATE CONCENTRATIONS IN WELL 8 DURING RECHARGE Hours after recharge begun NOa-N Piclorum Atrozine Ppm Ppm Ppm 0. 0.92 .004 0 3 0.97 0 6 .001 0 9 1.02 0 12 1.06 .003 0 15 3.14 .12 18 5.48 .076 .30 21 7.22 .097 .34 24 7.71 .121 .61 27 9.76 .124 .55 30 10.72 .124 .89 33 11.95 .146 1.04 36 11.73 .108 1.10 42 12.03 .133 1.02 48 11.88 1.01 54 12.25 .130 1.03 60 11.07 .120 1.00 66 10.37 .123 1.09 72 9.89 .125 .85 78 9.62 .104 .82 84 10.03 .098 .82 96 10.72 .124 1.60 107 13.35 .133 1.46 119 12.87 .132 .98 130 12.64 1.51 143 11.80 .099 1.44 147 1.34 157 11.58 .086 1.24 168 12.48 .011 1.17 174 12.80 .129 1.13 190 12.80 .114 1.17 195 12.80 .094 1.25 215 12.80 .099 1.38 222 13.11 238 12.80 .128 1.42 14 APPENDIX TABLE 5. HERBICIDE AND NITRATE CONC IN WELL 1 DURING PUMPING Hours after pumping began NOa-N Picloram Afrozine Ppm m“ 31m .083 11.66 .114 1.51 .167 11.73 .132 1.54 .25 12.18 -~ .132 1.56 .50 12.25 " .158 1.58 .75 12.41 .168 1.55 1.00 12.41 1.51 1.25 12.41 .176 1.55 1.50 12.48 .176 1.44 1.75 12.48 .176 1.46 2.0 12.56 .176 1.46 2.5 .160 1.44 3.0 12.56 .154 1.48 3.5 12.56 .170 1.43 4 12.64 .176 1.38 5 12.64 1.38 6 12.48 .147 1.41 7 12.48 .143 1.39 10 12.56 .139 1.43 12 12.64 .120 1.41 16 12.56 .140 1.43 20 12.64 1.47 24 16.16 .151 1.52 28 10.72 .150 1.54 32 13.68 .130 1.55 40 12.56 .130 1.64 46 12.64 .141 1.61 53 12.33 .141 1.60 60 12.33 .104 1.58 69 12.03 .078 1.56 77 12.18 1.41 83 12.10 .118 1.38 93 11.88 .106 1.36 101 11.15 .102 1.35 117 10.30 .092 1.29 125 9.76 .092 1.09 146 7.89 .077 .81 165 6.05 .057 .58 172 5.32 .048 .51 189 4.02 .037 .26 196 3.34 .028 .22 213 2.53 .008 .16 220 .003 .14 237 1.84 .003 .11 261 1.50 .004 .09 269 1.11 .001 .05 285 1.35 .03 APPENDIX TABLE 6. IN WELL 2 DURING PUMPING HERBICIDE AND NITRATE CON Hours offer w pumping begun NOg-N Piclorom Afrozine Ppm Ppm Err! 0 10.65 .124 1.48 3 11.73 .122 1.47 . 7 12.25 .145 1.54 " 15 12:56 .143 1.56 .1 23 12.48 .148 1.56 31 7.95 .106 1.60 47 12.18 .116 1.52 70 12.18 .106 1.49 95 11.44 .107 1.39 101 11.07 .059 1.35 118 10.16 .056 1.21 125 9.62 .044 1.16 I 146 8.58 .047 1.12 166 7.10 .060 .98 190 5.54 .057 .60 214 4.29 .015 .58 238 3.14 .010 .25 262 2.33 .005 .19 286 1.45 .18 APPENDIX TABLE 7. HERBICIDE AND NITRATE CONCENTRATIONS IN WELL 3 DURING PUMPING Hours offer pumping begun NOa-N Picloram Afrozine Trifluralin Ppm Ppm Ppm Ppm 0 10.79 .124 1.25 2 9.69 .077 1.09 4 9.62 .088 1.01 12 9.76 .086 .99 .0030 16 9.10 .078 .94 .0060 20 8.90 .093 .94 .0070 24 7.34 .083 .98 .0088 32 9.25 .083 .99 .0044 46 5 7.71 .051 1.03 .0040 53 7.53 .042 .97 .0042 70 7.34 .044 .90 .0042 77 5.71 .036 .77 .0058 94 5.04 .026 .48 .0054 101 4.66 .018 .43 .0020 118 3.76 .018 .36 .0014 125 3.29 .028 .26 .0018 146 2.83 .018 .26 .0016 166 2.18 .006 .24 .0014 190 1.54 .002 .19 .0014 214 1.16 0 .12 .0012 238 1.54 0 .11 .0010 262 1.45 0 .09 .0005 286 1.50 0 .05 .0002 Texas Agricultural Experiment Station Texas A8cM University College Station, Texas 77843 l-l. O. Kunkel, Acting Director- Publication