Grazing School

Workshops

Research

Faculty

News

Weather

Contact us

Superintendent
David Davis
21262 Genoa Road
Linneus, MO 64653
Phone: 660 895-5121
FAX: 660 895=5122
Email: DavisDK@missouri.edu

General Description Of Sampling System

Within each rotationally grazed pasture, 12 paddocks were created using portable electric fences and water tanks. Fence lines were always set on the same line and water tanks were usually placed within the 25 ft radius of the water valve. In most pastures, each of these paddocks was a strip approximately 350 to 450 ft long and 90 to 115 ft wide. In two pastures, paddock shape was different due to location of existing permanent fences and water course. All sampling was on an individual paddock basis. While baseline data and ending data was collected from all paddocks and transects, during the 5 grazing seasons, only 4 monitor paddocks were sampled for forage availability during the grazing season due to labor constraints. Paddocks 1 and 12 were not used as monitor paddocks in any pasture due to pre-existing soil and plant conditions occurring at these locations. With the exclusion of paddocks 1 and 12, the four monitor paddocks within each pasture were randomly selected from the remaining ten paddocks. Within each rotational pasture treatment, paddocks were used as additional replications for plant and soil data in the statistical analysis. In the continuously grazed pastures, the sampling stratification followed what would be the same logical paddock subdivision system used in the rotation pastures. The basic difference in sampling method was that, within a rotation pasture, samples were collected from the defined paddock area while in the continuous pastures the samples were collected along a transect line.

Herbage mass and Available forage:

Forage availability was determined before and after each grazing period in four randomly selected paddocks within each rotational grazing pasture. The same 4 paddocks were used throughout the five-year period. Within each paddock, a stratified sampling system was used to divide each paddock into 3 zones with Zone A representing the 1/3 nearest to water, zone B the middle third, and zone C the one third most distant from water. Within each zone, 3 randomly placed .3 m² quadrats were clipped to ½ in. stubble height with hand-held manual grass shears. Prior to clipping, mean sward height was measured using a yard stick within the area to be clipped. The clippings from all three quadrats within a zone were combined in a single bag and labeled with the date, pasture ID, paddock number, zone letter, and whether it was a pre-grazing or post-grazing sample. Post-grazing samples were clipped only in 1996 and 1998. In all other years post-grazing residual was predicted based on the height:yield relationships developed in 1996 and 1998. Data for forage availability presented in this report is the mean of pre- and post-grazing measurements.

In continuously grazed paddocks, nine 0.3m² quadrats were clipped along four transect lines on a bi-weekly basis to monitor forage availability. As with continuous grazing, three quadrats were combined in a single bag and labeled as A, B, or C. Mean sward height was determined at each quadrat site as described for rotational paddocks. Sward height measurements were made along the eight other transects in each continuously grazed pasture at bi-weekly intervals. Forage dry matter availability at the height-only transects was predicted from the height:yield relationship

Clipped samples were brought to the lab and individually weighed. After wet weight is recorded, all three sample bags from a paddock were combined, thoroughly mixed, and a 150 to 200 g subsample collected. The sample was oven dried at 130^o F, ground to pass a 1mm sieve, and was analyzed for nutritive composition. After the paddock was grazed and cattle moved to next paddock residual sward height was measured and residual forage availability was calculated using height:yield relationships developed for each grazing system during the 1996 grazing season.

At each quadrat site, a visual estimate was made of the percentage of dead forage present in the sward. When the samples from each set of three bags was combined and thoroughly mixed prior to sub-sampling, a second visual estimate of dead material was made. The mean of those two estimates is presented as percent dead forage in the report and was used in calculating green forage dry matter yield for each sample. Regression analysis was used to determine the consistency of the two methods of estimating percent dead material and the r-square value was greater than 0.8.

In 2000 no clipped yield estimates were made. Mean sward height measurements were made at the beginning and end of grazing on all 12 paddocks in rotational grazing pastures in all 12 transects in continuous grazing pastures. Forage availability was then predicted based on the height:yield relationships developed in the previous years.

Plant community:

Species composition and stand density data were collected near initiation of grazing each year (April 15 - May 1), midway through the grazing season (July 10-July 25), and after termination of grazing (Oct 1). Within each rotational grazing pasture, data was collected from all paddocks. Species frequency was determined through a step-point sampling system with 50 points per transect and 2 transects per paddock. In continuously grazed pastures, two point-step transects with 50 points each was made approximately 10 ft off the primary transect line on either side. The point-step transect was offset from the primary transect to avoid sites that may have been harvested as clipped quadrats. Percent of species frequency was based on total number of hits for any given species.

Height:Yield prediction:

Prediction equations for forage dry matter yield were developed using linear regression of herbage mass and green forage dry matter yield as dependent variable and mean sward height as independent variable. First and second order equations were calculated. In most cases, fit improvement using the quadratic equation was minor and first order equations were used for simplicity. Broad equations across year, month, grazing method, and stocking rate provided poor fit. Analysis compared equations developed across year, month, and stocking rate within grazing method provided better prediction for rotational stocking but were not significant for continuous stocking. Adequate predictive capability for bulk forage yield for rotational stocking was found by developing equations within stocking method and month, across years and stocking rate. No significant relationships between bulk yield and mean sward height were found for continuous stocking. When dead material in the sward was subtracted from the bulk yield, mean sward height provided acceptable prediction of green forage dry matter yield for both rotational and continuous stocking within grazing method and month. Separate sets of equations were used for pre-grazing and post-grazing predictions.

Soil bulk density:

Soil bulk density was measured in early April, mid-July, and mid-October of each year except April, 1998, when soil conditions were too wet for sampling and October, 1999, when limited labor availability restricted data collection . Within each rotationally grazed paddock or along each transect in continuously grazed pastures, four 3 in. X 3 in. soil cores were extracted, oven dried, and weighed to determine dry soil bulk density during each sampling period.

Soil fertility and nutrient redistribution:

Prior to initiation of grazing in 1996, soil samples were collected from each paddock and transect within the study. Samples were extracted from the surface three inches and the 3 to 6 in. Each sample consisted of 20 cores (.75 x 3 in.). Samples were analyzed for organic matter, soil pH, neutralizable acidity, Bray P₁ phosphorus, exchangeable K, calcium, and magnesium. In spring 2001, a final set of samples will be collected from the same sites and change in nutrient status occurring during the five years of the study will be determined.

Statistical analysis:

Overall treatment effects were determined using the SAS general linear models (GLM) procedure with stocking rate as main plot and grazing method as sub-plot with appropriate error terms for each treatment. Where analysis across years indicated a significant year effect, subsequent analysis was done within years. Time sequence data including plant community and soil bulk density were analyzed as repeated measures across years.