Sward Height and Yield Relationships in Grazed Pastures
James R. Gerrish1
Estimation of available pasture is a useful aid in making grazing management decisions. Forage yield can be approximated from sward height with reasonable accuracy if precautions are taken to define the sward being estimated. The yield:height relationship in pastures is affected by stand density, species composition, and sward height. Grazing management affects all of these parameters. Differences in predicted yield occur in continuously stocked (CS) versus rotationally stocked (RS) pastures. At shorter heights, RS pastures tended to be higher yielding at comparable heights while CS pastures tended to be higher yielding at taller heights. Greater stand thinning occurred in CS pastures when grazed below 4 in. and more dead material accumulated in CS pastures when allowed to grow above 8 in., compared to RS pastures. Prediction of available forage from height can be reasonably accurate when adjustment is made for species composition, stand density, and sward height.
Introduction: Estimation of available forage dry matter is an important tool for both livestock managers and pasture researchers. The need for balancing forage supply with animal requirements is crucial skill for successful graziers. A simple, inexpensive method for estimating available forage is to measure height and use an estimate of forage yield per acre- inch as a multiplier to calculate standing available forage. However, forage yield per acre-inch is not a constant value but depends upon sward density, species composition, and height. Grazing management can affect all three sward parameters, so there is question whether the same yield:height relationships hold true for continuously and rotationally stocked pastures. Data collected from one grazing season were analyzed to determine how much variance exists in the sward height:forage yield relationship between two different grazing management regimes.
Materials and Methods: Dry matter yield and mean sward height data were collected in a grazing study conducted at the University of Missouri-Forage Systems Research Center. Pastures were either continuously or rotationally stocked at four different stocking rates from April 25 through August 31, 1996. Swards consisted of mixed cool-season grasses and legumes with some presence of both annual and perennial warm-season species. Tall fescue (Festuca arundinacea Schreb.), Kentucky bluegrass (Poa pratensis L.), and orchardgrass (Dactylis glomerata L.) were the primary grasses with red clover (Trifolium pratense L.), white clover (T. Repens L.), and birdsfoot trefoil (Lotus corniculatus L.) being the dominant legumes.
Continuously stocked pastures were sampled bi-weekly along four defined transect lines with sample three sites per transect. Four paddocks out of the 12 paddocks in each rotationally stocked pastures were sampled during each grazing cycle before and after each grazing period. Each paddock had three sample sites. Within each sample site in both CS and RS pastures, three quadrats (3.25 ft2) were clipped at 1/2-in.height. Clipped samples were oven dried (155°F) and total dry matter yield (DMY) was determined. A correction factor for dead material was subtracted from the total yield and dry matter yield of green forage (GDMY) was calculated.
At each quadrat site, mean sward height was determined using a simple yardstick. Green forage DMY for each quadrat was divided by the measured height at that site and GDMY/inch was calculated. Yield X height relationships were determined through linear regression using both linear and quadratic functions.
Results and Discussion: For the type of pasture used in this study, height:yield relationships were significantly different for CS and RS pastures (Figures 1 & 3). At sward heights less than 4 in., CS pastures typically were lower yielding than RS pastures at comparable heights. Above 4-in. height, CS pastures were predicted to have higher yields than RS pastures. This difference is likely due to reduced stand density occurring in CS pastures when grazed below 4 in.. If continuous stocking is light enough to allow growth to be significantly greater than 6 to 8 in., dead material rapidly accumulates in the sward resulting in higher total DMY but reduced GDMY. The correction factor used in this study for dead tissue was a fixed quantity and was not adjusted for the greater dead material percentage at higher sward yields.
Linear prediction equations resulted in much greater variance between CS and RS pasture yields than did the quadratic prediction equations. Whereas linear predicted yield at 12 in. for CS was 900 lb/acre higher than for RS, the difference in quadratic predicted yields at the same height was less than 300 lb/acre. The standard error of estimate tended to be lower for all quadratic equations compared to linear equations.
Measured GDMY/acre-inch was significantly higher for RS pastures at lower sward heights (Figures 2 & 4). Stand density and species composition was determined using a point-step method at three times during the grazing season. At high stocking rate treatments for both CS and RS pastures, more bare ground hits were recorded later in the grazing season. Percentage of bare ground was somewhat greater in CS pastures compared to RS resulting in lower GDMY/acre-inch for CS pastures.
As stand density is a critical factor in the height:yield relationship, any use of height as a predictor of yield should probably be tempered with an adjustment factor for stand density. In a previous study, Gerrish (1982) reported the successful use of multiple regression prediction equations which used both visual estimates of species composition and ground cover coupled with sward height to predict both GDMY and component yields.
In a practical application, producers are not likely to use such complex techniques to estimate forage yield. Another approach to addressing the stand density and species composition factors is to develop a table of forage yield per acre-inch for different pasture types at different stand densities. We have successfully used such a table in field exercises with producers in grazing management education sessions (Table 1). Each pasture type X stand density cell shows a range in expected GDMY/acre-inch. The lower yield can be used for taller swards while the higher yield can be used for shorter swards. As an aid to producer use, NRCS in Missouri has produced sward sticks with this table printed on the sides of a square yard stick. The yard stick can be used for determining stand density with the point-step method, measure sward height, and serve as a reference table for yield information.
Conclusions: The data reported in this paper were collected as basic data in a grazing management study and so do not cover the full range of possibilities in factors affecting height:yield relationships. The results do indicate that grazing management does in fact affect height:yield relationships
Gerrish, J. R., 1982. A prediction model for pasture composition and yield. p. 102. Agronomy Abstracts.
Table 1. Estimated dry matter yield per acre-inch for various forages at three stand density levels.
______________________________________________________________ Stand Density --------------------------------------- Forage 60 - 75% 75 - 90% > 90% ______________________________________________________________ ------------ lb/acre-inch-------------- Tall Fescue + N 250 - 350 350 - 450 450 - 550 Tall Fescue + Legumes 200 - 300 300 - 400 400 - 500 Smooth Bromegrass + 150 - 250 250 - 350 350 - 450 Legumes Orchardgrass + Legumes 100 - 200 200 - 300 300 - 400 Bluegrass + 150 - 250 300 - 400 500 - 600 White Clover Mixed pasture 150 - 250 300 - 400 400 - 500 ______________________________________________________________
Figure 2. Predicted green forage dry matter yield per acre-inch for rotationally stocked pastures.
Figure 3. Predicted green forage dry matter yield for continuously stocked pastures.
Figure 4. Predicted green forage dry matter yield per acre-inch for continuously stocked pastures.
1 Research Assistant Professor, University of Missouri – Forage Systems Research Center, 21262 Genoa Rd., Linneus MO 64653.