Waterworks Archive

The effects of grazing practices on runoff quality

Dwayne Edwards
Department of Biosystems and Agricultural Engineering
University of Kentucky

Beef cattle production is a critical component of Kentucky’s agricultural economy, with over a million cattle marketed annually. Therefore, cattle production managment must insure that any long-term environmental impacts are acceptable to society. Several studies conducted in the western U.S. found that cattle grazing had a negligible effect on quality of runoff and streamflow. Some studies found no difference between grazed and ungrazed pastures with respect to the amount of nitrogen and phosphorus (nutrients with the potential for accelerating eutrophication of downstream water bodies) running off into streams. Other studies were less conclusive, suggesting that high rainfall and grazing near stream banks or at excessive stocking densities could increase stream flow nutrient and fecal bacteria concentrations.

The goals of this study were to (1) compare no grazing, conventional grazing and rotational grazing in terms of runoff quality and (2) assess the effectiveness of grassed filter strips to reduce the effects of grazing on runoff quality. The study used grassed plots to simulate pasture with beef cattle manure applied periodically to replicate grazing. Rainfall simulators constructed as part of this study were used to produce controlled runoff and eliminate dependence on natural rainfall.

The plots were built on a Maury silt loam soil at the University of Kentucky Maine Chance Agricultural Experiment Station. The plots used to test the grassed filter strips were 8 feet wide by 100 feet long with a three percent slope. The plots used to evaluate runoff quality for no grazing, conventional grazing and rotational grazing were only 20 feet long but otherwise the same as the filter strip plots. Each plot was established in Kentucky-31 "tall" fescue prior to the study. The 100 feet plots had gutters installed across the width of the plots at 20 feet intervals down-slope of the upper ends of the plots. The gutters were normally covered, allowing runoff to cross the gutters and continue down the plot, unless runoff samples were being collected. When the gutters were uncovered, runoff flowed down the gutter into an access sump where samples were collected in polyethylene bottles. The 20 feet plots had only one runoff collection gutter installed across the bottoms of the plots.

Five rainfall simulators, each capable of applying from 0-5 inch/hour simulated rainfall to one 20 feet plot, were constructed as a part of this project. Simulated rainfall intensity and frequency was controled by a computer. Each simulator operated independently (simultaneously providing water to separate 20 feet-long plots at separate simulated rainfall intensities), or they were used in series to provide rainfall to the 100 feet plots.

Nine of the 20 feet plots were used in the first goal, with three plots each used to simulate conventional grazing, rotational grazing, and an ungrazed (control) situation. The grazing strategies simulated only manure deposition; there were no attempts to replicate hoof traffic on the plots, and no cattle urine was added to the plots. The conventional grazing strategy was simulated by weekly application of a single, three-pound, beef cattle manure deposit to a randomly-selected location in each plot (equivalent to a stocking density of 1.5 animal units/ac). Rotational grazing was simulated by applying 12 pounds of beef cattle manure every fourth week (equivalent to a stocking density of six animal units/ac for one week, ungrazed for three weeks) in four deposits on each plot in the same plot locations as used in the conventional grazing plots. Simulated rainfall was applied at 4, 8 and 12 weeks after beginning manure application. Runoff samples were collected and analyzed for nitrogen, phosphorus and fecal coliform bacteria.

Three of the 100 feet plots were used to evaluate the performance of grassed filter strips. The upper 40 feet of each plot was set aside to simulate grazed pasture, and the remaining 60 feet served as the filter strip. The grazing was simulated by adding 30 lb beef cattle manure to the lower 3 feet of the 40-feet "pasture." The intent here was not to replicate a realistic grazing situation in terms of manure deposition pattern (this was done during the first goal). Instead, the intent was to ensure that concentrations of nutrients and bacteria entering the filter strips would be high enough to assess filter strip performance. Runoff samples from simulated rainfall applied immediately after the manure were collected at filter lengths of 0 (just down-slope of the manure deposits and prior to entry to the filter strip), 20, 40 and 60 feet, and analyzed for nitrogen, phosphorus, solids and fecal coliform bacteria.

There were no general differences in terms of runoff nitrogen and phosphorus concentrations between the ungrazed control plots and the grazed (either conventional or rotational) plots, as demonstrated in Figure 1. Runoff concentrations of nitrate nitrogen and ortho-phosphorus are seen to be very similar between the conventional graze, rotational graze and control plots. Runoff concentrations of total Kjeldahl nitrogen are slightly elevated for the grazed plots, reflecting presence of more organic nitrogen on those plots. The effects of grazing were most apparent in runoff fecal coliform concentrations. Mean concentration of fecal coliform in runoff from the control plots was 1,500 colony-forming units (cfu)/100 mL, in comparison to 240,000 cfu/100 mL for the conventional graze plots and 180,000 cfu/100 mL for the rotational graze plots. The gross amounts of nitrogen and phosphorus lost in runoff were quite low, typically less than 1 ounce/acre, suggesting that the effects of grazing on downstream waters might be quite small under study conditions.

The buffer strips had no effect on runoff concentrations of nitrate or ammonia nitrogen, because incoming concentrations were too low (typical of freshly-deposited manure). The buffer strips were very effective however (approximately 70 to100%), in reducing total Kjeldahl nitrogen, ortho-phosphorus, solids and fecal coliform concentrations in runoff. Filter strip performance is demonstrated in Fig. 2, which indicates that all significant reduction occurred in the first 20’ of filter strip. The filter strips were even more effective in removing incoming fecal coliforms from incoming runoff, with no FC detected in runoff for buffer strips of 20 feet and longer (even though the mean incoming FC concentration was 1.85 x 10⁵ cfu/100 mL).

In summary, this study indicates that manure deposition by grazing cattle can have a very small impact on runoff nitrogen and phosphorus concentrations under conditions similar to this study, regardless of the grazing method simulated. The effects of manure deposition on fecal coliform concentrations might be worthy of consideration, depending on downstream flow conditions, intended beneficial uses of the downstream waters, and dilution by relatively clean runoff. In any event, the study also suggests that relatively short filter strips can be highly effective for fecal coliform and nutrient removal from runoff. Grassed filter strips thus appear promising as an effective pollution prevention tool that might be considered for highly concentrated manure deposition, poor soil cover and similarly adverse conditions.

Last modified: December 1997

Copyright © 1995-1997 Kentucky Water Resources Research Institute, University of Kentucky