Ongoing UGA Research - The Fort Valley Experiment
The offsite impact on water quality in Georgia of applying poultry litter to agricultural fields is uncertain for several reasons. Although we have guidelines for how much poultry litter to apply based on the estimated litter nitrogen (N) content, these guidelines are based on assumed mineralization, volatilization, and leaching rates. Also, our guidelines are based on N and ignore any effect that phosphorus (P) losses might have on water quality. Composting poultry litter might reduce nutrient losses because the N and P could be transformed to more stable organic forms. A field experiment to study N and P losses from poultry litter and composted poultry litter is in its third year at Fort Valley State College. This is a cooperative project lead by Mark Latimore (Extension Agronomist at FVSC) and Miguel Cabrera and David Radcliffe (Soil Science researcher at UGA in Athens). A graduate student from The Netherlands, Willem Vervoort, is also working on the project for his Ph.D dissertation. Funding was provided by the USDA-Water Quality Program and the Southeastern Poultry and Egg Association.

The experiment is being conducted on three one-acre watersheds located at the Fort Valley State College Research Farm. The watersheds are equipped to automatically measure and subsample runoff and subsurface lateral flow. The watersheds have a 1-3 % slope and are surrounded by berms that route all surface runoff through 1-ft H-flumes. A relatively impermeable soil layer at a depth of about three feet causes lateral flow during storm events and tile drains on the downslope borders of each watershed intercept this flow. When flow occurs in the flumes or drains, the flow rate is recorded and subsamples are taken and stored in refrigerated units automatically. Runoff and subsurface samples are analyzed for inorganic (NO3- and NH4+) and total N, and bioavailable and soluble P. The watersheds are planted with Coastal bermudagrass interseeded with the new Georgia 5 fescue that has been developed to grow well in the early spring and late fall.

During the first year of the experiment, the recommended rate of poultry litter (1X), based on the N requirements of a bermudagrass and fescue hay crop, was applied to all three watersheds. This consisted of 4.5 tons/acre of litter, split into spring and fall applications. During the first year, maximum NO3--N concentrations in the subsurface samples were 6.1, 4.3, and 1.1 ppm for watersheds W1, W2, and W3, respectively. Soluble phosphorus concentrations in the runoff samples reached maximum values of 4.9, 3.3, and 1.2 ppm for W1, W2, and W3, respectively. These data show that NO3--N concentrations in subsurface drainage under the 1X treatment did not exceed the EPA drinking water standard of 10 ppm. Phosphorus levels were rather high considering that the EPA has established guidelines of 0.05 and 0.1 ppm total P for lakes and streams, respectively.

One of the lessons learned the first year was that the hydrologic response of the watersheds was not the same. The greatest runoff and subsurface flow occurred on W1, the least on W3, with W2 (which was located midway between W1 and W3) being intermediate. Work in the summer of the second year showed that the experimental area consists of two different soils: the west side where W1 is located is a Cowarts sandy loam, the east side where W3 is located is an Orangeburg sandy loam, and the transition between the two soils occurs where W2 is located. The hydraulic conductivities of the Orangeburg Bt and BC horizons are considerably higher than that of the Cowarts, therefore there is less runoff and lateral flow in this soil. By illustration, runoff accounted for 5.8, 1.9, and 1.6 % of the rainfall on W1, W2, and W3, respectively in the second winter. Subsurface flow accounted for 23.2, 14.1, and 5.3 % of the rainfall on the respective watersheds.

In the second year of the experiment, different treatments were applied to each watershed. W3 continued to receive the 1X rate, while twice the recommended rate (2X) was applied to W2 and the recommended rate plus 22.5 tons/acre of composted litter (1X + C) was applied to W1. The summer was extremely dry and no significant runoff or subsurface flow occurred until after the second part of the split application was completed. Nitrate-N concentrations in subsurface drainage samples reached peak concentrations of 2.9, 3.5, and 4.8 ppm in the 1X, 2X, and 1X + C treatments, respectively. The generally lower concentrations in the second year could be explained by a much larger role of crop uptake in this year, due to better developed stands of bermudagrass and fescue. All NO3--N concentrations remained under the drinking water standard even after addition of 360 lbs/acre of total N with the composted litter. Soluble P peak concentrations in the runoff were 1.6, 3.8, and 8.5 ppm on the 1X, 2X and 1X + C treatments, respectively. Here the addition of about 710 lb/acre of P with the compost treatment substantially increased the concentration of soluble P in the runoff on W1. Analysis of total and bioavailable P levels showed that, for all treatments, total P mainly consisted of soluble P. This suggests that conventional measures, like filter strips and riparian zones, will not lower the concentration in the runoff. Only increasing plant uptake, stabilizing the phosphorus in litter with the use of additives, like alum, or reducing the P in the ration through the use of phytase enzymes may decrease these concentrations.

These results, in general, show that composting works well in reducing the amount of N being lost, but does not reduce the amount of P. Although we only have one year of data with the 2X rate, the data indicates that the current UGA extension recommendation for poultry litter application rates on pastures should not cause any nitrate leaching problems. In fact, twice the recommended rate of litter produced peak NO3--N concentrations less than 10 ppm in the drainage water.

The Eatonton Experiments
Because of the lack of data on N and P fluxes in pastures amended with broiler litter, two field studies were recently initiated at the Central Georgia Branch Station located near Eatonton, Georgia. The objective of the first study is to quantify nitrate leaching, denitrification, ammonia volatilization, and crop N removal in tall fescue pastures fertilized with broiler litter. This study is under the supervision of Miguel Cabrera and is being replicated in Alabama and Tennessee with the cooperation of Wes Wood, Beth Guertal, and Mike Mullen. A graduate student, Lois Braun, is also working on the Georgia portion of this project for her M.S. thesis . Funding for this project was provided by the USDA-Water Quality Program. The objective of the second study is to evaluate the effect of stocking method (continuous versus rotational) on the quality of surface runoff from fescue/bermudagrass pastures fertilized with broiler litter. The investigators in this project are Miguel Cabrera, Carl Hoveland, Mark McCann, Dan McCracken, David Radcliffe, and Larry West. A Ph.D. student , Holly Kuykendall, is also working on this project for her dissertation. Funding was provided by the USDA-Natural Resources Conservation Service.

For the first study, we established six circular plots (130 ft diameter) separated 300 ft from each other on a tall fescue pasture. The circular shape of the plots and their separation is dictated by the method used to measure ammonia volatilization. The plots receive a spring application of broiler litter to provide 65 lb available N/acre. Ammonia volatilization is measured using a micrometeorological method that requires a circular plot with four masts located at 90o angles on the periphery of the plot. Denitrification is determined by measuring the rate of nitrous oxide (N2O) emission with a soil core technique. To determine the magnitude of N leaching, soil percolate at a 3.3-ft depth is collected with three suction cup lysimeters in each plot. Upon collection, percolate samples are frozen until analyzed for NO3-N and NH4-N concentrations.

Broiler litter was first applied on April 26, 1995, and ammonia volatilization was monitored for one month following application. Denitrification and nitrate leaching were monitored throughout the rest of the year. Results from the first year showed an average ammonia loss equivalent to 15% of the applied, available N (i.e., a loss of about 10 lb N/A). Denitrification losses were equivalent to 1.5% of the applied, available N (i.e. a loss of about 1 lb N/A). Nitrate leaching losses during the first year were negligible, in agreement with other studies that have shown minimum nitrate losses at recommended rates of broiler litter application.

For the second study, we established six grazing paddocks (1.8 acres each) in a tall fescue-bermudagrass pasture. Berms (3-ft wide, 1-ft high) were built around each paddock to prevent external surface runoff from entering the paddock and to route internal runoff to a 1.5-ft H-flume, which is instrumented with an ultrasonic sensor to measure water depth. A refrigerated sampler placed next to each flume collects runoff water samples, which are subsequently analyzed for total N, total P, inorganic N, inorganic P, and bioavailable P. The six paddocks are used for three replications of two treatments: continuous stocking and rotational stocking. Both treatments are fertilized equally with poultry litter at a rate equivalent to 150% of the recommended rate (9 ton/A per year). Litter applications are made in March and October of each year. Paddocks are stocked with beef steers and grazed from February or March until December of each year. A minimum of two test steers are maintained on each paddock during the grazing season, with additional animals added as needed to utilize the available forage. In the rotational treatments, steers are allowed to remain three days in each subpaddock. The stocking rate for both treatments is determined based on forage availability in the rotationally stocked paddocks.

Monitoring of runoff water quality started in October 1994 to determine background concentrations of nutrients. On December 14, 1994, two calves were placed in each plot (all plots managed under continuous stocking at this point). The calves grazed the plots until January 25, 1995, when they were removed due to lack of forage. The first broiler litter application (3.8 ton/A) was made on March 16 and 17, 1995, and the stocking treatments (continuous or rotational) were started on April 18, 1995. The second litter application (5.1 ton/A) was made on October 30 and 31, 1995.

Background concentrations of inorganic P (orthophosphate) in runoff water before starting the experiment ranged from 0.1 to 0.4 ppm P (without cattle present). After cattle grazed the pastures for 26 days (but before applying broiler litter), the concentration of orthophosphate in runoff water increased an average of 0.4 ppm . Application of broiler litter increased P in runoff water to values ranging from 2 to 6 ppm P during 1995. During the first year there were no significant differences between stocking treatments on the concentrations of inorganic P in runoff. These values of inorganic P are very similar to the ones obtained in the Fort Valley experiment.

One noticeable difference between the Fort Valley and Eatonton experiments, however, is the volume of runoff measured. Typical rainstorms of 1 to 3 inches during winter at the Eatonton site result in runoff volumes ranging from 15 to 65% of the received rain. Runoff at the Fort Valley site is usually much lower. Consequently, the total amounts of P lost are expected to be lower at the Fort Valley site than at the Eatonton site. We are currently collecting data for the second year of these studies.

Submitted by: Miguel Cabrera and David Radcliffe

last updated: 12 August 1996