|
|
|
© 2007 Plant Management Network. Alternative Methods of Estimating Forage Height and Sward Capacitance in Pastures Can Be Cross Calibrated Edward B. Rayburn and John D. Lozier, P.O. Box 6108, West Virginia University Cooperative Extension, Morgantown 26506-6108; Matt A. Sanderson, Building 3702, Curtin Road, Pasture Systems and Watershed Management Research Unit, USDA-ARS, University Park, PA 16802-3702; Brad D. Smith, 115½ Virginia Avenue, West Virginia University Cooperative Extension, Petersburg 26847-1713; William L. Shockey, 115 W Court St., West Virginia University Cooperative Extension, Kingwood 26537-1192; David A. Seymore, 200 Confederate Rd., P.O. Box 96, West Virginia University Cooperative Extension, Franklin 26807-0096; and Stanley W. Fultz, 330 Montevue Lane, Maryland Cooperative Extension, Frederick 21702-8200 Corresponding author: Edward B. Rayburn. ed.rayburn@mail.wvu.edu Rayburn, E. B., Lozier, J. D., Sanderson, M. A., Smith, B. D., Shockey, W. L., Seymore, D. A., and Fultz, S. W. 2007. Alternative methods of estimating forage height and sward capacitance in pastures can be cross calibrated. Online. Forage and Grazinglands doi:10.1094/FG-2007-0614-01-RS. Abstract A variety of tools are used for measuring pasture height or capacitance. Cross calibrations between these tools would be helpful for extension staff and producers comparing measurements taken with one tool to those taken with an alternative tool. Rotationally and continuously stocked pastures in West Virginia, Pennsylvania, Maryland, and New York were sampled for forage height using a ruler, for compressed height using a falling plate meter and a rising plate meter, and for sward capacitance with a capacitance meter. Thirty to sixty measurements were made across each pasture with each device, with paddock means taken as the measurement for the device. Regressions were run using paired paddock means, testing each device as both the dependent and independent variable, with r² ranging from 0.49 to 0.99. Residual analysis was conducted to evaluate biases due to location and stocking management using the falling plate meter means as the independent variable versus means of the other techniques. No bias in pasture measurements was found due to grazing management. There was a bias due to operator for ruler height and capacitance meter reading. These cross calibrations provide a mechanism for pasture managers to translate pasture heights or capacitance taken with one tool to those taken with another tool. Indirect Methods for Measuring Forage Mass in Pastures The measurement of sward height, compressed height, or sward capacitance is an indirect estimate of forage mass, used for determining research treatment effects and for on-farm pasture budgeting. Clipped pasture sampling is the standard method used for calibrating such tools to estimate forage mass per unit area. However, calibrations are labor-intensive and very site-specific. In research trials, the labor cost of directly measuring forage mass often limits the number of clipped samples taken. As an on-farm method, clipped sampling is not practical due to the time and labor required. However, indirect methods of measuring forage mass appear to be cost effective for improving management efficiency compared to management when forage mass is not known (4,10). Plate meters, sward sticks, and capacitance meters of various designs are used for measuring pasture height, compressed height or capacitance, and can be calibrated to provide an estimate of forage mass. The electronic capacitance meter relies on differences in dielectric constants between air and herbage. It measures the capacitance of the air-herbage mixture (2) and responds mainly to the surface area of the foliage (11). These tools are all quick and easy to use, but the accuracy of these estimates of forage mass is dependent on adequate and proper calibration (7). Where on-farm management or research trials are conducted using different indirect measures of forage availability, cross calibration regressions between methods would be useful and allow conversion of measurements taken with one tool to measurements taken with a second tool for comparison purposes. The purpose of this study was to develop a set of cross calibrations between compressed forage height measured with a standardized falling plate meter (8) and a commercial rising plate meter (3), forage height measured with a ruler, and sward capacitance measured with a commercial electronic capacitance meter (6) so that these methods of measuring pasture height, compressed height, or capacitance can be compared across sites. Cross Calibration Study Rotationally and continuously stocked cool-season grass and grass-legume pastures in West Virginia, Pennsylvania, Maryland, and New York were maintained in the vegetative growth stage and sampled across the growing season for forage height using a ruler (inches), compressed forage height using a standardized falling plate meter (inches) (8) and a commercial rising plate meter (centimeters) (3) (Farm Tracker electronic rising plate meter, FarmWorks, P.O. Box 433, Feilding, NZ; dimensions, 14.25-inch diameter and 0.7-lb mass), and for sward capacitance using a commercial electronic capacitance meter (Alistair George Pasture Gauge, Alistair George Manufacturing, Waihi Beach, NZ). A total of 90 paddock measurements were taken, 80 under rotational stocking and 10 under continuous stocking. Each was a mean of 30 to 60 observations. All grazing was conducted using yearling or mature beef or dairy cattle. Pastures were evaluated along established sampling transects and 30 to 60 measurements were taken randomly at regular intervals along the same sampling path. Ruler and falling plate meter heights were taken at the same sampling point, while rising plate meter heights and capacitance meter readings were taken at nearby but independent points. Measurements were averaged by method within each paddock to obtain the paddock mean ruler height, falling plate meter height, rising plate meter height, or capacitance meter reading for that paddock. Pasture ruler height was taken using a yard stick to measure the non-compressed sward height. For ruler height measurements taken in West Virginia, the end of the yard stick was placed on the soil surface and the falling plate was lowered to the surface of the pasture so that three out of the four quadrants of the plate touched a grass or legume leaf. The height of the top of the plate was the measure of ruler height. The plate was then lowered to the pasture surface, allowing the forage to completely support the plate meter. This height was the falling plate meter compressed forage height. On pastures at other locations, ruler height was evaluated subjectively by eye using the yard stick, then the falling plate was lowered to the pasture surface to measure falling plate meter compressed forage height. Regression analysis was used to develop cross-calibration equations between methods used for measuring forage height and capacitance (5). When a measurement method was used as the independent variable, it was used as a standard reference technique assumed to be measured without error. Only regression coefficients significant at P ≤ 0.05 were retained in equations (Table 1). When a regression’s intercept value was not significantly different from zero the regression was run without an intercept. To compare the accuracy of the cross calibrations, regressions using falling plate meter data as the independent variable were used to predict ruler height, rising plate meter height, and capacitance meter readings for each of the respective paddock means. Residuals were calculated by subtracting the predicted value from the observed value for each paddock. Residual values were analyzed by analysis of variance using state (West Virginia, Maryland, Pennsylvania, New York) and grazing method (rotational versus continuous stocking) as factors to determine if there was any bias in cross calibrations due to different measurement methods used in the states, or to pasture conditions due to grazing management (Table 2). Standard deviations about the regression (SDreg) were converted to coefficients of variation (CV) by dividing the SDreg by the mean of the dependent variable. Table 1. Cross calibration regressions between paddocks assessed using a falling plate meter (FPM, inches), a rising plate meter (RPM, cm), a ruler (RHT, inches), and sward capacitance meter (CMR, expressed as pounds of dry matter per acre calculated using a proprietary, non validated calibration equation).
Table 2. Residual analysis of paddock mean ruler height (inches), rising plate meter compressed forage height (cm), and capacitance meter readings predicted using the falling plate meter compressed forage height (inches) cross calibration regressions; values followed by different letters are significantly different at the 0.05 level.
* Mean of residuals. Cross Calibration of Forage Measurement Methods Forage height measured with the ruler as the dependent variable compared to the falling plate meter as the independent variable (Fig. 1) had more scatter than when the rising plate meter was compared to the falling plate meter (Fig. 2). Cross calibration regressions of ruler and rising plate meter with falling plate meter resulted in r² values of 0.96 and 0.99, respectively (Table 1). Ruler heights taken in Pennsylvania tended to be higher than those taken in the other states (Fig 1). When comparing capacitance meter reading as the dependent variable to forage height measured with the falling plate meter as the independent variable (Fig. 3) there was more scatter, with a regression r² of 0.72 (Table 1), than when the ruler (Fig. 1) or rising plate meter (Fig. 2) were used as the dependent variables.
The cross calibrations using ruler height as the dependent variable had the least precision, with a SDreg of 1.5 and 1.8 inches (CV of 0.20 and 0.24), when falling plate meter and rising plate meter, respectively, were the independent variables (Table 1). When rising plate meter was the dependent variable, the SDreg was 0.6 and 1.0 inches (CV of 0.10 and 0.16) when falling plate meter and ruler, respectively, were the independent variables. With falling plate meter as the dependent variable, the SDreg was 0.4 and 0.8 inches (CV of 0.10 and 0.19) when rising plate meter and ruler, respectively, were the independent variables. The precision (SDreg) of cross calibration for the capacitance meter readings was 353, 385, and 464 (CV of 0.18, 0.20, and 0.24) when rising plate meter, falling plate meter, and ruler, respectively, were independent variables (Table 1). These capacitance meter readings are in terms of forage mass estimated by a proprietary calibration that may not be appropriate for the pastures in this study. In general, the regressions on rising plate meter data accounted for more variability (i.e., had higher r² values) than did regressions on the falling plate meter or ruler. All cross calibration regressions (Table 1) had high r² values due in part to intercept values not significantly different than zero (P ≤ 0.05) being removed by rerunning the regressions without intercepts. All remaining regression coefficients were significant at P ≤ 0.001 except for the intercept coefficient for capacitance meter as a function of falling plate meter reading, which was P ≤ 0.02, and for falling plate meter as a function of rising plate meter reading, which was P ≤ 0.01. Removing non significant intercept coefficients increases the r² value associated with the regression. The standard deviation about the regression (SDreg) is another measure of how well the regression fits the data, and was not greatly affected by removal of intercept coefficients that were not different from zero. Residual analysis of falling plate meter predictions of ruler height, rising plate meter height, and capacitance meter readings found no bias due to grazing management but found bias due to state (Table 2). Ruler height data for the Pennsylvania location was significantly greater than data for other states, with the magnitude of the bias being more than two times the cross calibration SDreg. In Pennsylvania, New York, and Maryland ruler height was measured by eye, while in West Virginia, ruler height was measured as the height at which the falling plate meter contacted the pasture on three out of four quadrants of the plate meter. The subjective ruler heights by eye for the New York and Maryland operators were not significantly different from the West Virginia plate method, while that for the Pennsylvania operator was. Even though there is a documented operator effect on the use of plate meters (1) the operator effect of a subjective measure such as ruler height is likely more pronounced. The capacitance meter readings in Pennsylvania were also greater than those obtained in the other states, the magnitude being comparable to the cross calibration SDreg. This error could be due to the differences in operator use of the meter or due to differences in meters or pasture conditions that affected the measurement of capacitance at the different locations. Conclusions When comparing pasture height data from research or on-farm studies, the falling and rising plate meters gave paddock mean cross calibration values with low error as measured by the SDreg about the cross calibration. Ruler heights were highly reliable for comparisons when conducted by one operator but had higher error across operators. The capacitance meter proved a convenient tool for estimating pasture availability, but non-proprietary calibrations specific to the forages in the region are needed. The cross calibration regressions presented here provide a means of comparing pasture height, compressed pasture height and capacitance obtained using one tool to measured values obtained with the other tools. These cross calibrations provide a means to convert estimates of forage mass obtained with one tool to estimates obtained with the other tools when used on similar forage stands, as was done by Rayburn and Lozier (9). These values can also be used by producers and extension personnel to relate measurements presented in research and extension publications to the tool that they use on their own farm or ranch operations. Acknowledgment This research contributes to the mission of the Northeast Pasture Research and Extension Consortium. Literature Cited 1. Aiken, G. E., and Bransby, D. I. 1992. Observer variability for disk meter measurements of forage mass. Agron. J. 84:603-605. 2. Currie, P. O., Hilken, T. O., and White, R. S. 1987. Evaluation of a single probe capacitance meter for estimating herbage yield. J. Range Manage. 40:537–541. 3. Earle, D. F., and McGowan, A. A. 1979. Evaluation and calibration of an automated rising plate meter for estimating dry matter yield of pasture. Aust. J. Exp. Agric. Anim. Husb. 19:337-343. 4. Fulkerson, W. J., McKean, K., Nandra, K. S., and Barchia, I. M. 2005. Benefits of accurately allocating feed on a daily basis to dairy cows grazing pasture. Aust. J. Exp. Agric. 45:331-336. 5. Hintze, J. L. 1998. NCSS 2000 Statistical System. Number Cruncher Statistical Systems, Kaysville, UT. 84037. 6. Jones, R. J., and Haydock, K. P. 1970. Yield estimation of tropical and temperate pasture species using an electronic capacitance meter. J. Agri. Sci. Cambridge. 75:27-36. 8. Rayburn, E. B., and Rayburn, S. B. 1998. A standardized plate meter for estimating pasture mass in on-farm research trials. Agron. J. 90:238-241. 10. Sanderson, M. A., Rotz, C. A., Fultz, S. W., and Rayburn, E. B. 2001. Estimating forage mass with a commercial capacitance meter, rising plate meter, and pasture ruler. Agron. J. 93:1281-1286. 11. Vickery, P. J., and Nicol, G. R. 1982. An improved electronic capacitance meter for estimating pasture yield: Construction details and performance tests. Tech. Paper 9. CSIRO Animal Res. Lab., Armidale, NSW, Australia. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
