| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Department of Crop Science, North Carolina State University, Box 7620, Raleigh, NC 27695
b USDA-ARS, J. Phil Campbell Sr. Natural Resource Conservation Center, 1420 Experiment Station Road, Watkinsville, GA 30677
c Department of Animal Science, North Carolina State University, Box 7621, Raleigh, NC 27695
* Corresponding author (dfrankln{at}uga.edu)
Received for publication September 13, 2005.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: DRP, dissolved reactive phosphorus • TKP, total Kjeldahl phosphorus • TP, total phosphorus • TSS, total suspended sediments • WEP, water extractable phosphorus
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Grazing cattle can be a source of nutrients and may increase sediment losses from pastures to surface waters in the eastern USA (White et al., 1980; Schepers and Francis, 1982; Byers et al., 2005). However, the impact of grazing livestock on nutrient and sediment export to North Carolina surface waters has not been well quantified. In 2002, it was estimated that more than 705000 ha, or 19% of North Carolina agricultural land area, was used for grazing (USDA, 2002), which makes determining the impact of grazing on water quality critically important. Specifically, the importance of riparian area management to quality of surface waters underscores the need for scientific investigation of environmental impacts when riparian areas are grazed.
Riparian areas unsuitable for row crop production due to topography and seasonal flooding are sometimes grazed. These riparian areas can have relatively high forage productivity because of favorable moisture conditions, even during drier periods. However, poor grazing management can lead to reduced stand density and forage ground cover (Alderfer and Robinson, 1947; Bates et al., 1996; McKenzie, 1997). Poor stands can negatively influence infiltration, runoff, erosion, and sediment deposition (McGinty et al., 1979; Self-Davis et al., 2003) and limit the ecosystem services provided by the riparian area. While there are data relating to runoff volume, sediment, and nutrient export from grasslands in the Southeast, there are few data that report the level of ground cover necessary to protect water quality when livestock graze in or near riparian areas.
Several studies in diverse environments have suggested threshold levels of 70 to 75% ground cover, below which significant runoff and sediment loss can occur (Costin, 1980; Lang, 1979; Mwendera and Saleem, 1997). Lang (1979) indicated that when ground cover was <75%, as estimated by the point or area quadrat method, bare areas began to connect with each other which allowed for faster flow of runoff and less time for infiltration. Also using the quadrat method, Costin (1980) reported that ground cover <70% resulted in greater mean soil loss and runoff rate than ground cover >70%.
Slightly different results were reported by Dadkhah and Gifford (1980) in a Utah study on infiltration rate and sediment production. Cover was established by uprooting vegetation in randomly selected squares of a grid. The researchers reported no difference in infiltration rate on plots with either 50 or 80% ground cover. The researchers suggested that ground cover levels of 50% or greater may be sufficient for adequate watershed protection.
In Arkansas, Self-Davis et al. (2003) examined the effects of forage canopy cover on runoff volume from plots on a 5% slope. Though there was a trend toward greater runoff volume with forage clipping and resulting reduced canopy cover, height of forage did not significantly impact runoff volume from any forage species compared to unclipped plots. Edwards et al. (2000b) examined relationships between tall fescue clipping height and nutrient export from applied feces and urine and reported few total P (TP) concentration differences in runoff between treatments of unclipped forages or forages clipped to heights of 2.5, 10, or 20 cm. As expected, the lowest TP concentrations in runoff were observed from plots that did not receive feces and urine. Low TP concentrations were also reported during rainfall events 21 d after feces and urine application, rather than rainfall events immediately after feces and urine application.
Our objectives were to determine fundamental relationships between ground cover in riparian pastures and the export of sediment and P under simulated grazing conditions (application of beef steer feces and urine and simulated hoof compaction) using simulated rainfall. Specifically, this paper examines the impact of bare ground and low, medium, and high ground cover levels on runoff volume and export of total suspended sediments (TSS), dissolved reactive P (DRP), and total Kjeldahl P (TKP) from two sites of 10 and 20% slope in riparian pasture plots.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Experimental plots were 0.75 by 2.0 m and delineated with 23-cm high flashing inserted to a soil depth of 18 cm to isolate the surface hydrology of the plots. A runoff collection gutter was placed at the down-slope edge of each plot.
Cover Treatments
Upon initiation of this study, four ground cover treatments were established with four replications of each treatment, for a total of 16 plots on each site. Ground cover levels were created by establishing 100% bare soil area with 0% ground cover, 55% bare soil area with 45% ground cover, 30% bare soil area with 70% ground cover, and not altering full (95%) vegetative cover plots. Bare soil areas were created using a 5 by 10 cm rectangular grid and a small hand flame to quickly heat and kill statistically random blocks of vegetative cover. The random blocks created a pattern of varying sizes of bare soil areas as selected blocks connected with each other. After using the flame, herbage residue was raked from the plots. The plots were given several weeks to recover and any remaining residue was removed during subsequent natural rainfall events. Small rectangles of black plastic were placed on the plots in the same random grid pattern to maintain desired ground cover between rain events once the treatments of feces and urine were applied to the plots. Plastic was removed before simulated rain events.
The 0% ground cover treatment was established in each replication to simulate a compacted, cattle heavy use or ‘lounging’ area. Simulated lounging treatments were established by using black plastic to cover the whole plot and solarize all vegetation. A steel compaction device with an impact surface area of 100 cm2 was used to simulate cattle hoof compaction over the entire 0% cover plot, using methods described by Clary (1995). Simulated lounging plots were compacted before the first simulated rain event in the spring and again before the first fall rain event.
To standardize canopy height at each rain event, plots were harvested to a 10-cm stubble height before each rainfall simulation, as well as in July 2003 for plot maintenance. The forage samples from each harvest were weighed and dried (65°C), then analyzed by the North Carolina Department of Agriculture and Consumer Services (NCDA&CS) Agronomic Division for TP by the photometric method described in AOAC method 965.17 (Cunniff, 1995). Percent canopy cover was determined using the line transect method (Laflen et al., 1981) with 40 points after the vegetation was harvested to 10 cm and just before the rainfall simulations.
Rainfall Simulations
Before each simulated rainfall, three soil cores of 0- to 5-cm depth (1.75-cm i.d.) were obtained from each plot. The cores were combined by plot to form a composite sample, which was divided into two subsamples for each plot. One sample was air-dried and ground, while the second was placed in a soil moisture can and dried at 105°C for 24 h to determine gravimetric soil moisture content. Mehlich-3 soil P was determined on the dried and ground sample (Mehlich, 1984).
Three rainfall simulators (Tlaloc 3000, Joern's, West Lafayette, IN) were used to simulate rainfall at an intensity of 70 mm h–1 for a 1-h duration. This is just above a 10-yr, 1-h rain event of about 65 mm h–1 for Raleigh, NC, whereas a 25-yr, 1-h rain event for Raleigh is about 85 mm h–1 (Hershfield, 1961). Simulators were randomly assigned to blocks at each rain event, and the same simulator was used on all plots within a block. Each simulator rained on two plots simultaneously. In April 2003, an initial rainfall simulation was conducted to determine baseline runoff conditions from each plot before application of feces and urine. Deionized water was used as source water for the rainfall simulators. Great care was taken to ensure equal volume and distribution of rainfall from each simulator each time used. Simulators were calibrated by measuring volume and distribution of rain and adjusting the pressure valve at each simulator accordingly before moving simulators onto plots. Rain gauges were placed in plots to verify rainfall rate for each rain event.
Time was recorded as runoff began to drip from the gutter and the initial 125 mL of runoff was collected. When steady flow began, time was again recorded and a timer started so that runoff could be sampled beginning at 5, 10, 15, 20, 25, and 30 min after initiation of steady runoff flow. At each sampling time interval, 500 mL of runoff was obtained. Runoff between samplings was collected into large containers and weighed every 5 min after the initiation of steady runoff state until 30 min of runoff had occurred from each plot. The same methods of rainfall application and runoff collection were used for all rainfall simulations.
Feces and Urine Application
Feces and urine were applied in late May to plots immediately before rainfall simulations at a rate that approximated 10% of the average daily output for mature cattle (Kohn et al., 2005). This equated to a 2.4-kg deposit of moist feces (at 850 g H2O kg–1 wet manure) applied on a 550-cm2 area and a 1-L urine deposit applied over an equal area. The center of the fecal deposit was placed 30.5 cm from the top of the plot and centered between the plot sides. The 1-L urine deposit was placed directly adjacent and down slope of the fecal deposit. Assuming even distribution of feces and urine within a pasture, the two applications to each plot are representative of a stocking rate of about 4 cows ha–1 yr–1. Relatively low agronomic rates of P (<10 kg P ha–1) were applied to plots at each application of feces and urine (Table 1). The amount of WEP in manures has been reported to be highly correlated to runoff P (Kleinman et al., 2002a) and in this study, WEP applied to plots as feces averaged 2.3 kg P ha–1.
|
Following rainfall simulations in May 2003, plots were covered during natural rainfall events to prevent runoff until the final forage harvest of the experiment in November 2003. Rainfall simulations were again conducted 9 to 11 June 2003 to determine runoff constituents from plots, but without additional application of feces and urine. Plots were again treated with feces and urine before rainfall simulations on 9 to 11 Sept. 2003, which were followed by rainfall simulations 6 to 8 Oct. 2003 without additional plot treatment with feces and urine. Feces and urine application in September was on the same location on the plot, but the recognizable fecal material remaining from the spring was manually removed from the plot, weighed, and a 30-g sample removed for TP determination. The remaining fecal mass was manually broken up into about 25 pieces of similar size and distributed in the 2500-cm2 area surrounding the original fecal deposit immediately before the new application of feces and urine in September.
Sample Collection and Analysis
Runoff sample vials were placed in ice and in the dark until transported to the lab. Vacuum filtration (0.45 µm) of 100 mL of runoff sample was used to determine concentration of TSS. Filters were dried at 105°C for 24 h and weighed before and after filtration. Dissolved reactive P was determined by analyzing the filtered sample by the molybdate blue method (Murphy and Riley, 1962). Total Kjeldahl P was determined similarly following Kjeldahl digestion of an unfiltered sample according to USEPA Method 365.4 (USEPA, 1979).
Runoff samples collected at 5-min intervals represented point estimates of concentrations and were plotted vs. cumulative runoff volume. The points were joined with straight lines and the area under the plot was integrated according to the trapezoid rule using the PROC EXPAND procedure (SAS Institute, 1994). This method determined the cumulative mass of sediment and nutrients lost at each collection time; the main focus here being sediment and P export during 30 min of runoff.
The effect of cover, site (a combination of slope and soil differences), and month of rain event to runoff volume, TSS, DRP, and TKP mass export and concentrations was determined by using the PROC GLM procedure (SAS Institute, 1994). Means were separated using Fisher's least significant difference. Unless otherwise noted, differences were considered to be significant at P < 0.05.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Rainfall Rate and Distribution
At simulated rain events, a rainfall rate of 70 mm h–1 was the output target for each rainfall simulator. However, rainfall rate determined from rain gauges at the plots recorded a higher mean rainfall rate of 102 mm h–1 with a standard deviation of 15 mm h–1. Mean rainfall rate and rainfall rate recorded adjacent to feces and urine deposits were included as covariates in initial statistical models to account for any variation, but were not significantly related to runoff volume or cumulative export of TSS, DRP, or TKP during 30 min of runoff and so were not included in statistical models. As the same simulator was used on all plots within a block, the block effect in the statistical model helps account for differences among the rainfall simulators.
Runoff Volume
Cumulative runoff volume for 30 min of runoff was significantly related to cover (P < 0.001) and month of rain event (P < 0.05), but not to site (Table 3). A strong interaction between cover and month of rain event was also evident (P < 0.001). Therefore, in addition to examining mean runoff volumes from the five rain events, volume means by level of cover were also examined at each rain event, averaged over both sites.
|
|
Total Suspended Sediments
Cumulative TSS export during 30 min of runoff was significantly affected by cover (P < 0.001), site (P < 0.01), and rain event (P < 0.05) (Table 3). In addition, there were significant interactions between cover and site (P < 0.001) and between cover and rain event (P < 0.001), but not between site and rain event.
Due to interactions, mean TSS export was examined at each level of cover on each site (Table 4). From bare ground and low cover treatments, mean cumulative TSS export during 30 min of runoff on the site of 20% slope was 2.5-fold greater than export on the 10% slope site. This was expected as other researchers have reported greater soil loss (Mwendera and Saleem, 1997) and reduced trapping of sediment (Jin and Romkens, 2001) with increased slope. Since actual canopy cover values were somewhat higher than established ground cover levels in this study, results were similar to those of Mwendera and Saleem (1997) who suggested critical values for preventing soil losses in Ethiopia at 70% canopy cover at 0 to 4% slope and 85% cover for pasture with 4 to 8% slope. However, at medium and high cover, there were no significant differences in mean TSS export between the two sites in our study. This indicates that the medium cover level (about 80% canopy cover) protected sites on slopes of 20 and 10% equally well. In contrast, increased TSS export was observed from low cover (about 70% canopy cover) plots at the 20% slope site. Because the soils on the two sites differed slightly, the difference in TSS export cannot be solely attributed to slope, but this provides a good starting point for further research. Replication on other sites and soil types will be necessary to more closely examine the impact of slope to TSS export in pastured riparian areas. This is especially important due to the wide variations of slope typical to Piedmont riparian areas.
|
Soil Phosphorus
Soil P has been shown to be an important determinant of the amount of dissolved P in runoff (Pote et al., 1996). In this study, Mehlich-3 soil P generally did not vary significantly across each cover level and rain event, averaging 50 to 60 mg P kg–1 (Table 5). Soil P was included as a covariate in statistical analysis of DRP and TKP export during 30 min of runoff, but was not found to be significantly related to either runoff constituent. However, one noticeable trend in the data was the significant buildup of P in the soil on the bare ground treatment at the end of the study compared to plots at medium and high cover. This was to be expected as there was no vegetation to use applied P.
|
Due to interaction, cumulative DRP export was examined by rain event within each level of cover and averaged across sites (Fig. 2). Mean DRP export from vegetated plots exhibited a similar relationship among rain events; at low, medium, and high cover, mean DRP export was generally greatest during the September rain event and least during the baseline rain event in April. Low export during the baseline rain event was expected, as there was no application of feces or urine to plots. High export during the September event was also expected due to application of feces and urine at that rain event and possibly due to disturbing the remaining fecal deposit from May during the September application. However, DRP export in May did not respond as expected to feces and urine application. There was no greater DRP export in May than that of the baseline event at any level of cover. Export of DRP in May also did not differ from export during June, a rain event which did not include additional feces and urine application. Mean DRP export in May and June was rather evenly distributed between the two events, whereas in the fall, there was a large export of DRP in September and a much smaller export in October at each level of vegetative cover. At medium cover, DRP export in September was 2.5-fold greater than in October, and at high cover the export in September was sixfold greater.
|
In this study, cumulative DRP export during 30 min of runoff (averaged across sites) was also examined by cover level at each rain event, as well as for all rain events combined (Fig. 3). At the baseline event and in October, DRP export from the bare ground treatment was greater than that from low, medium, or high cover levels, which did not differ. In May, there was no difference between DRP export at bare ground and low cover and in June, no differences were observed at any level of cover. In September, DRP export was lower at medium cover than at high cover, which was the only rain event at which medium and high cover differed. However, neither high nor medium cover differed from bare ground or low cover for this rain event.
|
Export of DRP was also plotted against actual canopy cover measurements (excluding the bare ground treatment) to determine if any differences were evident that were not apparent when evaluating DRP export in relation to treatment cover levels. The analysis indicated that canopy cover was not significantly related to export of DRP.
Generally, DRP export was much lower than that seen in other studies in the Southeast, which may in part be a function of moderate Mehlich-3 soil P values and a relatively low rate of P application in the cattle feces and urine. Many other studies of runoff from grasslands in the Southeast involve application of poultry litter, at a higher P application rate, because it is a source for improved pasture productivity. In Georgia, Kuykendall et al. (1999) reported annual losses of 8.6 to 11.5 kg P ha–1 yr–1 from pasture with applied poultry litter, compared to the mean annual DRP export from this study of 0.4 kg P ha–1 yr–1.
Total Kjeldahl Phosphorus
Analysis of TKP export in this study indicated that cover had a substantial impact (P < 0.001) and rain event had a lesser impact (P < 0.01), while site alone did not significantly affect TKP export during 30 min of runoff (Table 3). A strong interaction was also noted between cover and rain event (P < 0.001).
At each rain event with the exception of June, bare ground plots had greater mean TKP export than plots at low, medium, or high cover, which were similar (Fig. 4). In June, there was no difference between mean TKP export at any level of cover. This was likely related to the low runoff volume from bare ground plots during the June rain event (Fig. 1) and lower associated TSS export from those same plots. Soluble P, sediment-bound P, and P in eroded organic matter are constituents of TKP export (Daniel et al., 1994). Thus, TKP at the June event was largely DRP, as is usual when TSS is low (Edwards et al., 2000a). Consistent with TKP export during individual rain events, the bare ground treatment produced the greatest TKP export when all rain events were combined.
|
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. H. Franklin, M. L. Cabrera, H. L. Byers, M. K. Matthews, J. G. Andrae, D. E. Radcliffe, M. A. McCann, H. A. Kuykendall, C. S. Hoveland, and V. H. Calvert II Impact of water troughs on cattle use of riparian zones in the Georgia Piedmont in the United States J Anim Sci, June 1, 2009; 87(6): 2151 - 2159. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. M. Butler, D. H. Franklin, M. L. Cabrera, A. S. Tasistro, K. Xia, and L. T. West Evaluating Aeration Techniques for Decreasing Phosphorus Export from Grasslands Receiving Manure J. Environ. Qual., May 1, 2008; 37(3): 1279 - 1287. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. M. Butler, N. N. Ranells, D. H. Franklin, M. H. Poore, and J. T. Green Jr. Ground Cover Impacts on Nitrogen Export from Manured Riparian Pasture J. Environ. Qual., January 9, 2007; 36(1): 155 - 162. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Soil Science Society of America Journal | Journal of Plant Registrations | The Plant Genome | |||
